Coupled Plasma Filtration Adsorption: Rationale, Technical Development and Early Clinical Experience

Sepsis, Kidney and Multiple Organ Dysfunction

This book has been made possible by the generous support of

Contributions to NephrologyVol. 144

Series Editor

Claudio Ronco Vicenza

Sepsis, Kidney andMultiple OrganDysfunction

Basel · Freiburg · Paris · London · New York ·Bangalore · Bangkok · Singapore · Tokyo · Sydney

Volume Editors

Claudio Ronco Vicenza

Rinaldo Bellomo Melbourne

Alessandra Brendolan Vicenza

55 figures, 7 in color, and 37 tables, 2004

Proceedings of the Third International Course on Critical Care NephrologyVicenza, June 1–4, 2004

Claudio Ronco Rinaldo BellomoDepartment of Nephrology Intensive Care UnitSt. Bortolo Hospital Austin & Repatriation Medical CenterI-36100 Vicenza (Italy) Melbourne, Vic. 3084 (Australia)

Alessandra BrendolanDepartment of NephrologySt. Bortolo HospitalI-36100 Vicenza (Italy)

Bibliographic Indices. This publication is listed in bibliographic services, including Current Contents® andIndex Medicus.

Drug Dosage. The authors and the publisher have exerted every effort to ensure that drug selection anddosage set forth in this text are in accord with current recommendations and practice at the time of publication.However, in view of ongoing research, changes in government regulations, and the constant flow of informationrelating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any change in indications and dosage and for added warnings and precautions. This is particularly importantwhen the recommended agent is a new and/or infrequently employed drug.

All rights reserved. No part of this publication may be translated into other languages, reproduced orutilized in any form or by any means electronic or mechanical, including photocopying, recording, microcopying,or by any information storage and retrieval system, without permission in writing from the publisher.

© Copyright 2004 by S. Karger AG, P.O. Box, CH–4009 Basel (Switzerland) www.karger.comPrinted in Switzerland on acid-free paper by Reinhardt Druck, BaselISSN 0302–5144ISBN 3–8055–7755–9

Library of Congress Cataloging-in-Publication Data

A catalog record for this title is available from the Library of Congress

Contributions to Nephrology(Founded 1975 by Geoffrey M. Berlyne)

V

Contents

IX PrefaceRonco, C. (Vicenza); Bellomo, R. (Melbourne); Brendolan, A. (Vicenza)

Epidemiology and Pathogenesis of ARF, Sepsis and MOF

1 Acute Renal Failure in the Critically Ill: Impact on Morbidityand MortalityHoste, E.A. (Gent); Kellum, J.A. (Pittsburgh, Pa.)

12 Acute Renal Failure in the Intensive Care Unit. Risk FactorsPiccinni, P.; Carraro, R.; Ricci, Z. (Vicenza)

19 Pathophysiology of Ischemic Acute Renal Failure. Inflammation, Lung-KidneyCross-Talk, and BiomarkersBonventre, J.V. (Boston, Mass.)

31 Pathophysiology of Sepsis and Multiple Organ Failure: Pro- versus Anti-Inflammatory AspectsPinsky, M.R. (Pittsburgh, Pa.)

44 Tropical Acute Renal FailureBarsoum, R.S. (Cairo)

The Critically Ill Patients: Pathological Mechanisms

53 Mechanisms Underlying Combined Acute Renal Failure and Acute Lung Injury in the Intensive Care UnitChien, C.-C.; King, L.S.; Rabb, H. (Baltimore, Md.)

63 Cytokine Single Nucleotide Polymorphism. Role in Acute Renal FailureLiangos, O.; Balakrishnan, V.S.; Pereira, B.J.G.; Jaber, B.L. (Boston, Mass.)

76 Mechanisms of Immunodysregulation in SepsisCavaillon, J.-M.; Fitting, C.; Adib-Conquy, M. (Paris)

Fluid, Electrolyte and Acid Base

94 Goals of Resuscitation from Circulatory ShockPinsky, M.R. (Pittsburgh, Pa.)

105 Intravenous Fluids and Acid-Base BalanceBellomo, R.; Naka, T.; Baldwin, I. (Melbourne)

119 Glucose Control in the Critically IllSchetz, M.; Van den Berghe, G. (Leuven)

132 Dysnatremias in the Critical Care SettingMoritz, M.L. (Pittsburgh, Pa.); Ayus, J.C. (San Antonio, Tex.)

Pharmacological Issues in ARF and Sepsis

158 Rasburicase Therapy in Acute Hyperuricemic Renal DysfunctionRonco, C. (Vicenza); Bellomo, R. (Melbourne); Inguaggiato, P. (Cuneo); Bonello, M.; Bordoni, V.; Salvatori, G.; D’Intini, V.; Ratanarat, R. (Vicenza)

166 Diuretics in Acute Renal Failure?Schetz, M. (Leuven)

182 How to Manage Vasopressors in Acute Renal Failure and Septic ShockDan, M.; Rossi, S.; Callegarin, L.; Ronco, C. (Vicenza)

Practical Aspects of CRRT

191 Management of Vascular Catheters for Acute Renal Replacement TherapyD’Intini, V.; Bonello, M.; Salvatori, G.; Ronco, C. (Vicenza)

203 Relationship between Blood Flow,Access Catheter and Circuit Failure during CRRT:A Practical ReviewBaldwin, I.; Bellomo, R. (Melbourne)

Renal Replacement Therapy in the ICU: Consensus and Recommendations from ADQI

214 CRRT: Selection of Patients and Starting CriteriaPalevsky, P.M. (Pittsburgh, Pa.)

Contents VI

222 Fluid Composition for CRRTLeblanc, M. (Montreal)

228 Anticoagulation for Continuous Renal Replacement TherapyDavenport, A. (London)

Which Treatment for ARF in ICU?

239 Peritoneal Dialysis in Acute Renal Failure of Adults:The Under-Utilized ModalityAsh, S.R. (West Lafayette, Ind.)

255 Intermittent Hemodialysis for Acute Renal Failure Patients – An UpdateLameire, N.; Van Biesen, W.; Vanholder, R.; Hoste, E. (Gent)

264 Continuous Renal Replacement TechniquesClark, W.R. (Lawrence, Mass./Indianapolis, Ind.); Ronco, C. (Vicenza)

278 Hybrid Renal Replacement Therapies for Critically Ill PatientsGolper, T.A. (Nashville, Tenn.)

284 Pediatric Acute Renal Failure: Demographics and TreatmentGoldstein, S.L. (Houston, Tex.)

Technical Aspects of CRRT

291 Vascular Access for Extracorporeal Renal Replacement Therapy in the Intensive Care UnitCanaud, B.; Formet, C.; Raynal, N.; Amigues, L.; Klouche, K.; Leray-Moragues, H.; Béraud, J.-J. (Montpellier)

308 Anticoagulation in Continuous Renal Replacement TherapyVargas Hein, O.; Kox, W.J.; Spies, C. (Berlin)

317 Replacement and Dialysate Fluids for Patients with Acute Renal Failure Treated by Continuous Veno-Venous Haemofiltration and/or Haemodiafiltration Davenport, A. (London)

CRRT Information Technology

329 A Practical Tool for Determining the Adequacy of Renal ReplacementTherapy in Acute Renal Failure PatientsPisitkun, T.; Tiranathanagul, K. (Bangkok); Poulin, S.; Bonello, M.; Salvatori, G.; D’Intini, V.; Ricci, Z. (Vicenza); Bellomo, R. (Melbourne); Ronco, C. (Vicenza)

Contents VII

New Frontiers in the Management of ARF, MOF and Sepsis

350 How to Approach Sepsis Today?Vincent, J.-L. (Brussels)

362 High Volume Hemofiltration in Critically Ill Patients:Why,When and How?Tetta, C. (Bad Homburg); Bellomo, R. (Melbourne); Kellum, J. (Pittsburgh, Pa.); Ricci, Z. (Vicenza); Pohlmeier, R.; Passlick-Deetjen, J. (Bad Homburg); Ronco, C. (Vicenza)

376 Coupled Plasma Filtration Adsorption: Rationale,Technical Development and Early Clinical ExperienceBrendolan, A.; Ronco, C.; Ricci, Z.; Bordoni, V.; Bonello, M.; D’Intini, V. (Vicenza); Wratten, M.L. (Mirandola); Bellomo, R. (Melbourne)

387 Plasmapheresis in SepsisBerlot, G. (Trieste); Di Capua, G. (Naples); Nosella, P.; Rocconi, S.; Thomann, C. (Trieste)

395 Author Index

396 Subject Index

Contents VIII

Preface

Multiple epidemiological studies have established and continue to empha-size the fact that sepsis is the dominant syndrome in modern Intensive CareUnits. Severe sepsis occurs in approximately 50 to 100 cases/100,000 people/year and is the most common cause of death in intensive care patients. Severesepsis and septic shock are now also the most common cause of kidney failurein intensive care and the most common cause of severe kidney failure requiringin general renal replacement therapy. This kind of kidney failure, however, israrely seen in isolation. Most commonly, it occurs as part of a syndrome ofmultiple organ failure, where the kidney is one of several organ systems thatbecome profoundly dysfunctional. In this setting, vasodilatory shock is fre-quent, mechanical ventilation is frequent and disorders of bone marrow func-tion, acid-base balance, gastrointestinal activity and cerebral function arecommon. Thus, severe sepsis links kidney function, multiple organ function andpatient outcome from the start to the end.

The care of patients with severe sepsis and/or septic shock is complex andtypically involves a multidisciplinary approach. Critical care specialists typi-cally co-ordinate resuscitation, fluid administration, and mechanical ventila-tion. In conjuction with the nephrologist, they deal with issues of electrolyteand water balance, acid-base control and renal support. Increasingly, renal sup-port focuses on complex approaches to extracorporeal therapy, which requirethe use of sorbents, high-volume plasma water exchange techniques and plas-mafiltration or plasma exchange techniques. In conjunction with the infectiousdisease specialist, critical care physicians and nephrologists co-ordinate anti-biotic or antifungal treatment. This requires important adjustments, which

IX

depend on renal function and the technique of renal support being applied.Accordingly, knowledge of pharmacokinetics and pharmacodynamics becomesessential. Finally, emerging evidence indicates that the resolution of the septicstate and of multiorgan dysfunction might require optimization of the endocrineenvironment through replacement of glucocorticoids in patients with loss ofadrenal functional reserve, the supplementation of vasopressin in selectedpatients with vasodilatory shock and, perhaps more importantly, the restorationof normoglycemia through aggressive insulin administration.

The above considerations make it clear that for patients to receive optimalcare, the treating physician needs a detailed working knowledge of multipleaspects of care so that appropriate multidisciplinary assistance is sought at theright time and new techniques of organ support are applied in a safe, timely andeffective way. In the present book we have combined the contributions ofexperts in various fields to tackle some of the fundamental and complex aspectsof patients care. First we have focussed on the epidemiology of acute renal fail-ure in intensive care and on its role in determining outcome. We then presentrecent advances in the insight into the pathogenesis of ischemic renal failureand of sepsis and multiple organ failure. Because the immune response to infec-tion is central in determining organ injury, the book then focuses on its role indetermining renal and lung injury, on the role of immune mediators in inducingdysregulation of the immune response and on the role of genetics in determin-ing such a response. We then move to the issue of fluid resuscitation, the goalsof resuscitation, the importance of acid-base control and the issues that sur-round glucose control and sodium control in the ICU. Pharmacological aspectsof care involving the use of common medications such as diuretics and vaso-pressors are analyzed and the possible role of uric acid modulation discussed.As extracorporeal therapies are being increasingly used in the care of thesecomplex patients, we focus on important technical aspects of such therapiesincluding vascular catheter management, control of circuit blood flow, antico-agulation, choice of replacement or dialysate fluids, the role of informationtechnology and the selection of patients for treatment. As the choice of treat-ment modality remains controversial, we also discuss different approaches torenal support from intermittent dialysis to continuous therapies and hybridtechniques. Finally, we conclude with a description of advanced extracorporealtechniques of organ support and discuss their role in the management of sepsisand kidney failure in the context of an overall strategy of sepsis management.

The aim of this book is to present all physicians involved in the care of crit-ically ill patients with sepsis and kidney/multiorgan dysfunction with a practi-cal and up-to-date summary of current knowledge and technology as well as afundamental understanding of pathogenesis and likely future developments inthis field. Our endeavour is part of a now long-standing and continuing effort

Preface X

to improve patient outcome through laboratory and clinical research, educationand consensus development. Working on the development of the specialty ofCritical Care Nephrology and of the Acute Dialysis Quality Initiative (ADQI),we hope to move steadily in the direction of improved outcomes for criticallypatients with kidney and multiorgan dysfunction.

We hope this book will serve as a useful tool for consultation, referenceand informative reading for all professionals involved in the care of critically illpatients and that it will represent yet another small step toward improving thestandards of care for such patients worldwide.

Claudio RoncoRinaldo Bellomo

Alessandra Brendolan

Preface XI

Ronco C, Bellomo R, Brendolan A (eds): Sepsis, Kidney and Multiple Organ Dysfunction.Contrib Nephrol. Basel, Karger, 2004, vol 144, pp 1–11

Acute Renal Failure in the Critically Ill:Impact on Morbidity and Mortality

Eric A. Hostea, John A. Kellumb

aIntensive Care Unit, Ghent University Hospital, Gent, Belgium; bThe CRISMA Laboratory (Clinical Research, Investigation, and Systems Modeling of Acute Illness), Department of Critical Care Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pa., USA

The occurrence rate of acute renal failure (ARF) in critically ill patientsvaries in the literature between 1.1 and 31% according to the definition of ARFused and the patient population studied [1–3]. In critically ill patients more than90% of ARF episodes are of ischemic or toxic etiology or a combination ofboth. However, numerous other etiologies of ARF have been identified.Examples include obstruction of the urine outflow tract, acute tubulointerstitialnephritis, acute glomerulonephritis, and atheroemboli [4–9]. The precise diag-nosis of the etiology of ARF is not always obvious or easy to establish; somepatients can therefore be misclassified. Furthermore, the spectrum of etiologieswill be different, depending on the specific patient population studied. In turn,each etiology can be associated with a specific spectrum of comorbidities, andconsequent impact on outcome. For example, the complications and prognosisof contrast media induced ARF are different from that of ischemic ARF in apatient with sepsis, or ARF caused by lupus nephritis.

It is therefore not surprising that the mortality of critically ill patients withARF as reported in recent trials varies between 28 and 82% [3, 10]. In contrastto the former dogma that patients die with and not because of ARF, there is nowample evidence that ARF is itself somehow associated with excess mortality, orin other words ARF has ‘attributable mortality’ [11]. This has been most con-vincingly demonstrated in patients with ARF requiring treatment with renalreplacement therapy (RRT) [1, 12, 13], but also appears to be true for patientswith the less severe form of contrast media-induced ARF [14]. ARF patients areamongst the sickest in the intensive care unit. Patients often develop ARF as aconsequence of conditions which are themselves associated with high mortality,

Epidemiology and Pathogenesis of ARF, Sepsis and MOF

Hoste/Kellum 2

e.g. sepsis, hypotension. Furthermore, ARF may result in a variety of compli-cations that can, in turn, lead to comorbidity or organ dysfunction [12, 14], thusnegatively impacting on patient outcomes in the ICU.

In this review we will discuss many of the complications of ARF, and con-sider their potential impact on patient outcomes in the ICU. While numerousreviews have focused on the clinical effects of RRT, it is important to note thatemerging evidence suggests that even when ARF does not result in RRT, thereis a significant impact on mortality [15]. Thus, it is important to consider thepossible mechanisms of attributable mortality with ARF even when RRT is notemployed. It is this group with which we will concern ourselves.

Complications of ARF

Uremic retention products are compounds that under normal conditionsare excreted by the healthy kidneys. Impaired clearance of these compounds, asin ARF, leads to higher concentrations. When these compounds interfere withorgan function one speaks of ‘uremic toxins’. There are currently more than 90uremic retention products described in the literature for patients with chronicuremia [16, 17]. Whether these compounds fulfill the definition of uremic toxinis less well established [18]. Furthermore, the role that any of these moleculesplays in ARF is less certain. For example, while chronic exposure to moleculessuch as leptin and retinol binding protein is associated with loss of appetiteand neuro-muscular symptoms, it is unknown if these toxicities occur in theacute setting where less time for accumulation has passed. Conversely, uremictoxicity might play an even larger role in acute disease. Indeed there is growingevidence that more intensive treatment of uremia, or a higher dose of RRT,improves survival [10, 19, 20]. Of course, earlier and more intensive RRT mightalso control blood volume, salt and acid-base variables more efficiently and thistoo could affect outcome.

Because the literature on uremic toxicity in ARF is much less abundantthan in chronic uremia, we will focus on the complications of acute renal failure,of which many are a result of uremic retention compounds and their effects onnormal physiology.

Volume OverloadSalt and water retention result in volume overload, this can occur even early

in the course of ARF. Volume overload is one of the main reasons for initiatingRRT [5, 21, 22]. Patients with volume overload will develop edema, ascites, andpleural effusions. As a consequence of volume overload, cardiovascular compli-cations, such as congestive heart failure, hypertension, and pulmonary edema

ARF in the Critically Ill 3

may develop. Respiratory function can be impaired by pleural effusion, and pul-monary edema. That volume overload can lead to many untoward effects hasbeen illustrated in colorectal surgery where a restrictive fluid therapy regimenlead to a significant reduction in comorbidity [23]. Another, less common, com-plication of volume overload is the development of an abdominal compartmentsyndrome (ACS) [24]. ACS is becoming more commonly recognized as a com-plication of volume resuscitation and capillary leak. It has been described inpatients with multiple organ failure and burns [25]. Patients with ACS have anincreased intra-abdominal pressure, resulting in a pro-inflammatory response,and a decreased preload, low cardiac output, hypotension, impaired oxygenationand oliguria [26]. ACS may reduce renal perfusion and further worsen ARF ordelay recovery. Formation of edema may also result in delayed wound healing[27]. Cell swelling interferes with normal cellular function, the most prominentsymptoms are neurologic: impaired consciousness to coma and death [28]. Insummary, volume overload can cause a whole range of complications, and thismay help explain why ARF patients with oliguria have a worse survival, com-pared to non-oliguric ARF patients [2, 4, 7, 29–34].

AcidosisThe kidneys play an important role in preservation of the acid base homeo-

stasis; this through excretion of non-volatile acids and control of cation/anionbalance. When the glomerular filtration rate (GFR) declines, there is accumu-lation of organic anions, e.g. hyperphosphatemia, and other unmeasured anions[35]. Furthermore, there is decreased production of bicarbonate by thedecreased proximal tubular reabsorption and regeneration. Another factor thatis important in the pathogenesis of metabolic acidosis is the decreased buffer-ing capacity, secondary to hypoalbuminemia [35]. Finally, many patients willalso have non-renal reasons for acidosis leading to mixed disorders in acid basebalance. Examples of non-renal etiologies of acidosis are lactic acidosis, respi-ratory acidosis induced by permissive hypercapnia ventilation strategies or, lessfrequently, keto-acidosis. Mild-to-moderate acidosis is therefore a commonfinding in patients with ARF [35, 36].

The effects of metabolic acidosis are diverse [37, 38]. Acidosis can lead tomany untoward effects, especially of the cardiovascular system, and may beassociated with decreased survival [34]. Blood pressure and cardiac output canbe decreased [39, 40], as well as hepatic and renal blood flow [41]. The under-lying mechanisms of these effects are uncertain. Decrease of the �-receptors onthe cell surface of the heart can play a role [42], as well as an increase ofinducible nitric oxide synthase (iNOS) which can lead to vasodilatation [43].Furthermore, it has been demonstrated that hyperchloremic acidosis increaseslung and intestinal injury and decrease gut barrier function [43–46].

Hoste/Kellum 4

Acidosis has also many untoward effects on metabolism. Glucose metab-olism is impaired by the induction of insulin resistance [47] and inhibition ofanaerobic glycolysis. Furthermore, there is induction of protein breakdown anddecreased muscle protein synthesis [48–54]. Hyperchloremic acidosis has pro-inflammatory effects documented in vitro and in animal models. HCl infusionincreases the release of NO and increases the Il-6/Il-10 ratio. In addition, HClincreased NF-�B DNA binding in LPS-stimulated RAW 264.7 cells [55].Acidosis also induces hyperkalemia by a shift of intracellular potassium to theextracellular compartment [56]. Finally, acidosis has a negative impact onmetabolism of the central nervous system, resulting in impaired consciousnessand eventually coma.

Immunological Function, Inflammation, and InfectionPatients with ARF have a high incidence of infectious complications. In one

recent study, the higher the preoperative creatinine, the higher was the incidenceof infections in patients who underwent open-heart surgery [57]. Bloodstreaminfections had an almost 3 times higher occurrence rate in critically ill patientswith ARF compared to patients without ARF [58]. Finally, the occurrence rateof serious infections or sepsis in ARF patients treated with RRT ranges from 15to 58.5% [57, 59, 60]. Thus, immunosupression seems likely in patients withARF [58]. In patients with chronic uremia there is ample evidence that there isa decreased immune response as seen with impaired phagocytic function ofwhite blood cells taken from these patients [61–63]. The underlying mechanismsfor these effects are believed to include malnutrition, uremic toxicity, ironoverload, anemia, and incompatibility of the hemodialyzers. There is less dataconcerning the immune response in the setting of ARF; however, the causativefactors held responsible for the decreased neutrophil function are also presentin patients with ARF. It seems therefore plausible that the phagocytic functionof polymorphonuclear cells in ARF patients is also impaired.

Finally, ARF itself, can lead to a non-infectious, pro-inflammatory responsewith activation of lung macrophages, secretion of pro-inflammatory cytokines,and recruitment of neurtrophils and macrophages with resultant lung injury,as has been demonstrated in animal models of ischemia-reperfusion-inducedARF [64–68].

Electrolyte DisordersHyponatremia is a common finding in patients with ARF [36]. Usually,

hyponatremia is dilutional, either hypervolemic or euvolemic. Underlyingmechanisms for dilutional hyponatremia can be decreased free water clearance(hypervolemic hyponatremia), and hyperosmolar dilution, e.g. as a resultant ofadministration of synthetic colloids, or hyperglycemia (either hypervolemic or


euvolemic after redistribution of fluid from the intracellular to the extracellularcompartment). Free water clearance can already be impaired in an early stageof ARF, before elevations of creatinine or blood urea nitrogen [69].Hypovolemic hyponatremia can be the resultant of losses of hypertonic fluids,e.g. after excessive vomiting or diarrhea, or interstitial or ‘third space’ losses asin burns or muscle injury.

Hyponatremia, especially, acute hyponatremia (duration less than 48 h),interferes with normal cell function. The hypotonic extracellular environmentcauses cell swelling, as the osmotic gradient causes redistribution of water fromthe extracellular to the intracellular compartment. Symptoms include a wholespectrum of primarily neurological symptoms from headache, lethargy to comaand even dead. The more rapid the onset of hyponatremia, the higher the like-lihood of severe symptoms [70].

Like hyponatremia, hyperkalemia is a potentially life-threatening situationas it may be complicated by cardiac arrhythmias. Hyperkalemia is a commoncomplication in patients with ARF. The kidneys play a pivotal role in potassiumhomeostasis; failure of the kidneys to excrete potassium may therefore result inhyperkalemia. Other common causes of hyperkalemia in ARF patients may bethe increased release of potassium from the intracellular compartment, as inrhabdomyolis, tumor lysis syndrome, and hemolysis [71–74], and side effectsof drugs (e.g. non-steroidal anti-inflammatory drugs and COX-2 inhibitors,trimethoprim/sulfamethoxazole, angiotensin receptor blockers, cyclosporin ortacrolimus [75–78], and heparin).

AnemiaAnemia is a consistent finding in ARF patients. Underlying mechanisms

for anemia are decreased synthesis of red blood cells, increased destruction ofred blood cells, and increased blood loss. ARF patients have inappropriate lowlevels of erythropoietin [79, 80]. The uremic state causes increased red bloodcell fragility and destruction [81], and increased blood loss is aggravated byplatelet dysfunction [82, 83]. Anemia may excacerbate cardiovascular disease.Although a hemoglobin level as low as 7 g/dl is well tolerated in a general inten-sive care population, and is even beneficial for the subgroups of patients whoare younger, and less severely ill, higher hemoglobin levels are recommendedfor patients with ischemic heart disease [84, 85].

Drug Prescription

Adequate dosing of drugs is a challenge in patients with ARF, because thepharmocodynamics and pharmacokinetics of drugs may be altered at every

Hoste/Kellum 6

level. Plasma levels can be lower as a result of decreased availability whengiven orally [86], or increased volume of distribution. Alternatively, plasmadrug levels may be higher as a result of reduced albumin binding, decreasedmetabolism and/or decreased renal excretion [87]. Additionally, plasma levelsof metabolites of drugs may be increased for the same reasons, and this mayalso result in increased toxicity. To complicate things even more, none of thevariables mentioned above remains constant during the course of an ARFepisode: GFR changes and RRT affects drug concentrations.

The dosing of an exhaustive list of drugs is reviewed in ‘Drug prescribingin renal failure’, a publication by the American College of Physicians [88]. Wewill limit ourselves to the discussion of the particular aspects of drug dosing of3 groups of drugs, these being the most commonly prescribed to criticallyill patients with ARF: vaso-active agents, antimicrobial agents, and sedatives/analgetics.

Vaso-Active AgentsThe dose of vasopressor agents such as norepinephrine, epinephrine and

dopamine, is prescribed according to their effect on blood pressure and cardiacoutput. There are no dosing restrictions for patients with renal insufficiency.However, for inotropic agents from the class of phosphodiesterase inhibitors, adose reduction is recommended for patients with renal insufficiency. Clearanceof dopamine, epinephrine, and norepinephrine is minimal, and clinicallyinsignificant when continuous veno-venous hemodiafiltration is used as modeof RRT [89]. These data can probably be extrapolated to other forms of RRTcurrently used for critically ill patients with ARF.

Antimicrobial AgentsAlmost all antimicrobial agents are eliminated by the kidneys. Adjustment

of dose and/or interval between 2 doses should therefore always be consideredin patients with ARF in order to prevent toxicity as a result of high plasma lev-els. Another consideration is the increase in distribution volume that ARFpatients will have as a result of edema and low albumin levels. Underdosing ofantimicrobial agents and, therefore, inefficient antimicrobial treatment, is alsoa potential hazard in patients with ARF. In general, the loading dose of anti-microbial agents should be higher while the dosing interval longer, comparedto that given to patients without ARF.

Sedatives and AnalgeticsMost opiates and the older sedatives from the class of the benzodiazepines

(e.g. diazepam and lorazepam) are excreted via the urine. Decreased renal


function will therefore lead to accumulation of the drug or its metabolites,of which some also have sedative properties (e.g. morphine-6-glucuronide,a metabolite of morphine). A dose reduction is therefore recommended forthese drugs in patients with decreased kidney function.

This problem exists to a much lesser extent for propofol and midazolam,drugs that have renal elimination of inactive metabolites, and remifentanyl,a recently developed opiate that is metabolized in plasma and tissues.

Adequate dosing of drugs in order to become adequate drug levels is animportant cornerstone of anti-infectious chemotherapy. Critically ill patientswith ARF are at risk for both low and high drug levels of antimicrobialchemotherapeutic agents, and both are undesirable. High drug levels may resultin drug toxicity, and low drug levels may result in inadequate eradication of theinfectious process, but also increased antimicrobial resistance [90].

Conclusions

It is currently not known why, even in the absence of RRT, ARF leads toan increase in mortality in critically ill patients. However, given the number ofmajor perturbations in physiology that result from renal dysfunction, it seemslikely that the explanation is multifactorial. We do not know if correction ofthese abnormalities will reduce the attributable mortality of ARF or whetherearly RRT, or ‘renal support’ will improve outcome. We speculate that given themultiple effects of even early ARF, efforts to prevent its development or hastenits recovery will significantly improve survival in critically ill patients.

References

1 Chertow GM, Levy EM, Hammermeister KM, Grover F, Daley J: Independent associationbetween acute renal failure and mortality following cardiac surgery. Am J Med 1998;104:343–348.

2 de Mendonca A, Vincent JL, Suter PM, Moreno R, Dearden NM, Antonelli M, Takala J, Sprung C,Cantraine F: Acute renal failure in the ICU: Risk factors and outcome evaluated by the SOFAscore. Intens Care Med 2000;26:915–921.

3 Vivino G, Antonelli M, Moro M, Cottini F, Conti G, Bufi M, Cannata F, Gasparetto A: Risk fac-tors for acute renal failure in trauma patients. Intens Care Med 1998;24:808–814.

4 Brivet F, Kleinknecht D, Loirat P, Landais P, the French Study Group on Acute Renal Failure:Acute renal failure in intensive care units-causes, outcome, and prognostic factors on hospitalmortality: A prospective, multicenter study. Crit Care Med 1996;24:192–198.

5 Guerin C, Girard R, Selli JM, Perdrix JP, Ayzac L: Initial versus delayed acute renal failure in theintensive care unit: A multicenter prospective epidemiological study. Rhone-Alpes Area StudyGroup on Acute Renal Failure. Am J Respir Crit Care Med 2000;161:872–879.

6 Groeneveld A, Tran D, van der Meulen J, Nauta J, Thijs L: Acute renal failure in the medical inten-sive care unit: Predisposing, complicating factors and outcome. Nephron 1991;59:602–610.

Hoste/Kellum 8

7 Liaño F, Pascual J, the Madrid Acute Renal Failure Study Group: Epidemiology of acute renal fail-ure: A prospective, multicenter, community-based study. Kidney Int 1996;50:811–818.

8 Liaño F, Junco E, Pascual J, Madero R, Verde E, the Madrid Acute Renal Failure Study Group:The spectrum of acute renal failure in the intensive care unit compared to that seen in other set-tings. Kidney Int 1998;53(suppl 66):S16–S24.

9 Pascual J, Liaño F, the Madrid Acute Renal failure Study Group: Causes and prognosis of acuterenal failure in the very old. J Am Geriatr Soc 1998;46:721–725.

10 Schiffl H, Lang SM, Fischer R: Daily hemodialysis and the outcome of acute renal failure. N EnglJ Med 2002;346:305–310.

11 Kellum JA, Angus DC: Patients are dying of acute renal failure. Crit Care Med 2002;30:2156–2157.12 Metnitz PG, Krenn CG, Steltzer H, Lang T, Ploder J, Lenz K, Le Gall JR, Druml W: Effect of

acute renal failure requiring renal replacement therapy on outcome in critically ill patients. CritCare Med 2002;30:2051–2058.

13 Hoste EA, Lameire NH, Vanholder RC, Benoit DD, Decruyenaere JM, Colardyn FA: Acute renalfailure in patients with sepsis in a surgical ICU: Predictive factors, incidence, comorbidity, andoutcome. J Am Soc Nephrol 2003;14:1022–1030.

14 Levy EM, Viscoli CM, Horwitz RI: The effect of acute renal failure on mortality: A cohort analy-sis. JAMA 1996;275:1489–1494.

15 Hoste E, Clermont G, Kersten A, Venkataraman R, Kaldas H, Angus DC, Kellum JA: Clinicalevaluation of the new RIFLE criteria for acute renal failure. Crit Care 2004; abstract for the 24thInternational Symposium on Intensive Care and Emergency Medicine, Brussels.

16 Vanholder R, De Smet R: Pathophysiologic effects of uremic retention solutes. J Am Soc Nephrol1999;10:1815–1823.

17 Vanholder R, De Smet R, Glorieux G, Argiles A, Baurmeister U, Brunet P, Clark W, Cohen G,De Deyn PP, Deppisch R, Descamps-Latscha B, Henle T, Jorres A, Lemke HD, Massy ZA,Passlick-Deetjen J, Rodriguez M, Stegmayr B, Stenvinkel P, Tetta C, Wanner C, Zidek W: Reviewon uremic toxins: Classification, concentration, and interindividual variability. Kidney Int 2003;63:1934–1943.

18 Vanholder R, Glorieux G, De Smet R, Lameire N: New insights in uremic toxins. Kidney IntSuppl 2003;84:S6–S10.

19 Conger JD: A controlled evaluation of prophylactic dialysis in post-traumatic acute renal failure.J Trauma 1975;15:1056–1063.

20 Ronco C, Bellomo R, Homel P, Brendolan A, Dan M, Piccinni P, La Greca G: Effect of differentdoses in continuous veno-venous hemofiltration on outcomes of acute renal failure: A prospectiverandomized trial. Lancet 2000;356:26–30.

21 Mehta RL, McDonald B, Gabbai FB, Pahl M, Pascual MTA, Farkas A, Kaplan RM: For theCollaborative Group for the Treatment of ARF in the ICU: A randomized clinical trial of contin-uous versus intermittent dialysis for acute renal failure. Kidney Int 2001;60:1154–1163.

22 Cole L, Bellomo R, Silvester W, Reeves JH: A prospective, multicenter study of the epidemiology,management, and outcome of severe acute renal failure in a ‘closed’ ICU system. Am J Respir CritCare Med 2000;162:191–196.

23 Brandstrup B, Tonnesen H, Beier-Holgersen R, Hjortso E, Ording H, Lindorff-Larsen K,Rasmussen MS, Lanng C, Wallin L, Iversen LH, Gramkow CS, Okholm M, Blemmer T,Svendsen PE, Rottensten HH, Thage B, Riis J, Jeppesen IS, Teilum D, Christensen AM,Graungaard B, Pott F: Effects of intravenous fluid restriction on postoperative complications:Comparison of two perioperative fluid regimens: A randomized assessor-blinded multicenter trial.Ann Surg 2003;238:641–648.

24 Morken J, West MA: Abdominal compartment syndrome in the intensive care unit. Curr Opin CritCare 2001;7:268–274.

25 Balogh Z, McKinley BA, Holcomb JB, Miller CC, Cocanour CS, Kozar RA, Valdivia A, Ware DN,Moore FA: Both primary and secondary abdominal compartment syndrome can be predicted earlyand are harbingers of multiple organ failure. J Trauma 2003;54:848–859; discussion 859–861.

26 Rezende-Neto JB, Moore EE, Melo de Andrade MV, Teixeira MM, Lisboa FA, Arantes RM,de Souza DG, da Cunha-Melo JR: Systemic inflammatory response secondary to abdominal com-partment syndrome: Stage for multiple organ failure. J Trauma 2002;53:1121–1128.


27 Teschan PE: Acute renal failure during the Korean War. Ren Fail 1992;14:237–239.28 Steele A, Gowrishankar M, Abrahamson S, Mazer CD, Feldman RD, Halperin ML: Postoperative

hyponatremia despite near-isotonic saline infusion: A phenomenon of desalination. Ann InternMed 1997;126:20–25.

29 Mehta RL, Pascual MT, Gruta CG, Zhuang S, Chertow GM: Refining predictive models in criti-cally ill patients with acute renal failure. J Am Soc Nephrol 2002;13:1350–1357.

30 Guerin C, Girard R, Selli J-M, Perdrix J-P, Ayzac L: For the Rhone-Alpes Area Study Group onAcute Renal Failure: Initial versus delayed acute renal failure in the intensive care unit: A multi-center prospective epidemiological study. Am J Respir Crit Care Med 2000;161:872–879.

31 Guerin C, Girard R, Selli JM, Ayzac L: Intermittent versus continuous renal replacement therapyfor acute renal failure in intensive care units: Results from a multicenter prospective epidemi-ological survey. Intens Care Med 2002;28:1411–1418.

32 Rasmussen HH, Pitt EA, Ibels LS, McNeil DR: Prediction of outcome in acute renal failure bydiscriminant analysis of clinical variables. Arch Intern Med 1985;145:2015–2018.

33 Bullock ML, Umen AJ, Finkelstein M, Keane WF: The assessment of risk factors in 462 patientswith acute renal failure. Am J Kidney Dis 1985;5:97–103.

34 Chertow GM, Lazarus J, Paganini E, Allgren R, Lafayette R, Sayegh M: Predictors of mortalityand the provision of dialysis in patients with acute tubular necrosis. The Auriculin Anaritide AcuteRenal Failure Study Group. J Am Soc Nephrol 1998;9:692–698.

35 Rocktaeschel J, Morimatsu H, Uchino S, Goldsmith D, Poustie S, Story D, Gutteridge G, Bellomo R:Acid-base status of critically ill patients with acute renal failure: Analysis based on Stewart-Figgemethodology. Crit Care 2003;7:R60.

36 Dolson GM: Electrolyte abnormalities before and after the onset of acute renal failure. MinerElectrolyte Metab 1991;17:133–140.

37 Adrogue HJ, Madias NE: Management of life-threatening acid-base disorders. First of two parts.N Engl J Med 1998;338:26–34.

38 Gunnerson KJ, Song M, Kellum JA: Influence of acid-base balance on patients with sepsis; inVincent JL (ed): Yearbook of Intensive Care and Emergency Medicine 2004. Springer-Verlag, Berlin,2004, pp 58–67.

39 Weil MH, Houle DB, Brown EB Jr, Campbell GS, Heath C: Vasopressor agents: Influence of aci-dosis on cardiac and vascular responsiveness. Calif Med 1958;88:437–440.

40 Kellum JA, Song M, Venkataraman R: Effects of hyperchloremic acidosis on arterial pressure andcirculating inflammatory molecules in experimental sepsis. Chest 2004;125:243–248.

41 Bersentes TJ, Simmons DH: Effects of acute acidosis on renal hemodynamics. Am J Physiol 1967;212:633–640.

42 Marsh JD, Margolis TI, Kim D: Mechanism of diminished contractile response to catecholaminesduring acidosis. Am J Physiol 1988;254:H20–H27.

43 Pedoto A, Caruso JE, Nandi J, Oler A, Hoffmann SP, Tassiopoulos AK, McGraw DJ,Camporesi EM, Hakim TS: Acidosis stimulates nitric oxide production and lung damage in rats.Am J Respir Crit Care Med 1999;159:397–402.

44 Salzman AL, Wang H, Wollert PS, Vandermeer TJ, Compton CC, Denenberg AG, Fink MP:Endotoxin-induced ileal mucosal hyperpermeability in pigs: Role of tissue acidosis. Am J Physiol1994;266:G633–G646.

45 Unno N, Hodin RA, Fink MP: Acidic conditions exacerbate interferon-gamma-induced intestinalepithelial hyperpermeability: Role of peroxynitrous acid. Crit Care Med 1999;27:1429–1436.

46 Pedoto A, Nandi J, Oler A, Camporesi EM, Hakim TS, Levine RA: Role of nitric oxide in acidosis-induced intestinal injury in anesthetized rats. J Lab Clin Med 2001;138:270–276.

47 DeFronzo RA, Beckles AD: Glucose intolerance following chronic metabolic acidosis in man.Am J Physiol 1979;236:E328–E334.

48 Mitch WE, Goldberg AL: Mechanisms of muscle wasting: The role of the ubiquitin-proteasomepathway. N Engl J Med 1996;335:1897–1905.

49 Mitch WE: Mechanisms causing loss of muscle in acute uremia. Ren Fail 1996;18:389–394.50 Bailey JL, Mitch WE: Metabolic acidosis as a uremic toxin. Semin Nephrol 1996;16:160–166.51 Mitch WE: Metabolic acidosis stimulates protein metabolism in uremia. Miner Electrolyte Metab

1996;22:62–65.

Hoste/Kellum 10

52 Bailey JL, England BK, Long RC, Mitch WE: Influence of acid loading, extracellular pH and ure-mia on intracellular pH in muscle. Miner Electrolyte Metab 1996;22:66–68.

53 May RC, Bailey JL, Mitch WE, Masud T, England BK: Glucocorticoids and acidosis stimulateprotein and amino acid catabolism in vivo. Kidney Int 1996;49:679–683.

54 Isozaki U, Mitch WE, England BK, Price SR: Protein degradation and increased mRNAs encod-ing proteins of the ubiquitin-proteasome proteolytic pathway in BC3H1 myocytes require an inter-action between glucocorticoids and acidification. Proc Natl Acad Sci USA 1996;93:1967–1971.

55 Kellum JA, Song M, Li J: Lactic and hydrochloric acids induce different patterns of inflammatoryresponse in LPS-stimulated RAW 264.7 cells. Am J Physiol Regul Integr Comp Physiol 2004;286:R686–R692.

56 Adrogue HJ, Madias NE: Changes in plasma potassium concentration during acute acid-base dis-turbances. Am J Med 1981;71:456–467.

57 Thakar CV, Yared JP, Worley S, Cotman K, Paganini EP: Renal dysfunction and serious infectionsafter open-heart surgery. Kidney Int 2003;64:239–246.

58 Hoste E, Blot S, Lameire N, Vanholder R, De Bacquer D, Colardyn F: Impact of nosocomialbloodstream infection on the outcome of critically ill patients with acute renal failure treated withrenal replacement therapy. J Am Soc Nephrol 2004;15:454–462.

59 Schiffl H, Lang SM, Konig A, Strasser T, Haider MC, Held E: Biocompatible membranes in acuterenal failure: Prospective case-controlled study. Lancet 1994;344:570–572.

60 Fiaccadori E, Lombardi M, Leonardi S, Rotelli CF, Tortorella G, Borghetti A: Prevalence and clin-ical outcome associated with preexisting malnutrition in acute renal failure: A prospective cohortstudy. J Am Soc Nephrol 1999;10:581–593.

61 Vanholder R, Ringoir S: Infectious morbidity and defects of phagocytic function in end-stagerenal disease: A review. J Am Soc Nephrol 1993;3:1541–1554.

62 Vanholder R, De Smet R, Waterloos MA, Van Landschoot N, Vogeleere P, Hoste E, Ringoir S:Mechanisms of uremic inhibition of phagocyte reactive species production: Characterization ofthe role of p-cresol. Kidney Int 1995;47:510–517.

63 Horl WH: Neutrophil function and infections in uremia. Am J Kidney Dis 1999;33:xlv–xlviii.64 Kramer AA, Postler G, Salhab KF, Mendez C, Carey LC, Rabb H: Renal ischemia/reperfusion

leads to macrophage-mediated increase in pulmonary vascular permeability. Kidney Int 1999;55:2362–2367.

65 Rabb H, Chamoun F, Hotchkiss J: Molecular mechanisms underlying combined kidney-lung dys-function during acute renal failure. Contrib Nephrol. Basel, Karger, 2001, pp 41–52.

66 Donnahoo KK, Meldrum DR, Shenkar R, Chung CS, Abraham E, Harken AH: Early renalischemia, with or without reperfusion, activates NFkappaB and increases TNF-alpha bioactivityin the kidney. J Urol 2000;163:1328–1332.

67 Donnahoo KK, Meng X, Ayala A, Cain MP, Harken AH, Meldrum DR: Early kidney TNF-alphaexpression mediates neutrophil infiltration and injury after renal ischemia-reperfusion. AmJ Physiol 1999;277:R922–R929.

68 Donnahoo KK, Shames BD, Harken AH, Meldrum DR: Review article: The role of tumor necro-sis factor in renal ischemia-reperfusion injury. J Urol 1999;162:196–203.

69 Baek SM, Makabali GG, Brown RS, Shoemaker WC: Free-water clearance patterns as predictorsand therapeutic guides in acute renal failure. Surgery 1975;77:632–640.

70 Rose BD, Post TW: Hypoosmolal states-hyponatremia; in Post TW (ed): Clinical Physiology ofAcid-Base and Electrolyte Disorders, ed 5. New York, McGraw-Hill, 2001, pp 696–745.

71 Erek E, Sever MS, Serdengecti K, Vanholder R, Akoglu E, Yavuz M, Ergin H, Tekce M, Duman N,Lameire N: An overview of morbidity and mortality in patients with acute renal failure due tocrush syndrome: The Marmara earthquake experience. Nephrol Dial Transplant 2002;17:33–40.

72 Jeha S: Tumor lysis syndrome. Semin Hematol 2001;38:4–8.73 Arrambide K, Toto RD: Tumor lysis syndrome. Semin Nephrol 1993;13:273–280.74 Gloe D: Common reactions to transfusions. Heart Lung 1991;20:506–512.75 Harris RC Jr: Cyclooxygenase-2 inhibition and renal physiology. Am J Cardiol 2002;89:10D–17D.76 Wrenger E, Muller R, Moesenthin M, Welte T, Frolich JC, Neumann KH: Interaction of spirono-

lactone with ACE inhibitors or angiotensin receptor blockers: Analysis of 44 cases. BMJ2003;327:147–149.


77 Kahan BD: Cyclosporine nephrotoxicity: Pathogenesis, prophylaxis, therapy, and prognosis. AmJ Kidney Dis 1986;8:323–331.

78 Hsu I, Wordell CJ: Hyperkalemia and high-dose trimethoprim/sulfamethoxazole. Ann Pharmaco-ther 1995;29:427–429.

79 Lipkin GW, Kendall RG, Russon LJ, Turney JH, Norfolk DR, Brownjohn AM: Erythropoietindeficiency in acute renal failure. Nephrol Dial Transplant 1990;5:920–922.

80 Lipkin GW, Kendall R, Haggett P, Turney JH, Brownjohn AM: Erythropoietin in acute renal fail-ure. Lancet 1989;i:1029.

81 Nagano N, Koumegawa J, Arai H, Wada M, Kusaka M: Effect of recombinant human erythropoi-etin on new anaemic model rats induced by gentamicin. J Pharm Pharmacol 1990;42:758–762.

82 Eknoyan G, Wacksman SJ, Glueck HI, Will JJ: Platelet function in renal failure. N Engl J Med1969;280:677–681.

83 Weigert AL, Schafer AI: Uremic bleeding: Pathogenesis and therapy. Am J Med Sci 1998;316:94–104.

84 Hébert PC, Yetisir E, Martin C, Blajchman MA, Wells G, Marshall J, Tweeddale M, Pagliarello G,Schweitzer I: The transfusion requirements in critical care investigators for the Canadian criticalcare trials group: Is a low transfusion threshold safe in critically ill patients with cardiovasculardiseases? Crit Care Med 2001;29:227–233.

85 Hébert PC, Wells G, Blajchman MA, Marshall J, Martin C, Pagliarello G, Tweeddale M,Schweitzer I, Yetisir E: Group TTRiCcIftCCCT: A multicenter, randomized, controlled clinicaltrial of transfusion requirements in critical care. N Engl J Med 1999;340:409–417.

86 Craig RM, Murphy P, Gibson TP, Quintanilla A, Chao GC, Cochrane C, Patterson A, Atkinson AJ Jr:Kinetic analysis of D-xylose absorption in normal subjects and in patients with chronic renal fail-ure. J Lab Clin Med 1983;101:496–506.

87 Reidenberg MM: The biotransformation of drugs in renal failure. Am J Med 1977;62:482–485.88 Aronoff G, Berns J, Brier M, Golper T, Morrison G, Singer I, Swan SK, Bennet W: Drug

Prescribing in Renal Failure: Dosing Guidelines for Adults. Philadelphia, American College ofPhysicians, 1999.

89 Bellomo R, McGrath B, Boyce N: Effect of continuous venovenous hemofiltration with dialysison hormone and catecholamine clearance in critically ill patients with acute renal failure. CritCare Med 1994;22:833–837.

90 Pinder M, Bellomo R, Lipman J: Pharmacological principles of antibiotic prescription in thecritically ill. Anaesth Intens Care 2002;30:134–144.

John A. Kellum, MD608 Scaife Hall, The CRISMA Laboratory, Critical Care MedicineUniversity of Pittsburgh, 3550 Terrace Street, Pittsburgh, PA 15261 (USA)Tel. �1 412 647 6966, Fax �1 412 647 3791, E-Mail [email protected]


Acute Renal Failure in the Intensive Care UnitRisk Factors

P. Piccinni, R. Carraro, Z. Ricci

Department of Anesthesiology and ICU, St.Bortolo Hospital, Vicenza, Italy

Acute renal failure (ARF) is a common clinical event affecting 2–5% ofhospitalized patients and up to 10–30% of those in intensive care units (ICU),depending on the population studied and the criteria used to define [1].

The mortality rate in ARF patients remains high [2] despite hemodialysistherapy and substantial improvement of dialysis techniques.

Possible explanations for this finding include the fact that ICU patientstoday are older and more debilitated than previously, and that the same patho-physiologic factors involved in the development of ARF are also incriminatedin the failure of other organs, so that ARF is often part of the multiple organfailure (MOF) syndrome. [3] Acute renal failure can be oliguric (urinary out-put, �400 ml per day) or nonoliguric (�400 ml per day). Patients with nono-liguric acute renal failure have a better prognosis than those with oliguric renalfailure, probably due in large measure to the decreased severity of the insult andthe fact that many have drug-associated nephrotoxicity or interstitial nephritis.The percentage of patients with acute renal failure who require dialysis rangesfrom 20 to 60%. Among the subgroup of patients who survive initial dialysis,less than 25% require long-term dialysis, demonstrating the potential reversibil-ity of the syndrome [3]. Brivet et al. [4], in a prospective, multicenter study,found seven predictive factors for severe ARF patients requiring ICU stayrelated to the acute clinical setting in which ARF occurred and three character-istics of ARF influenced the outcome of patients: (a) hospitalization beforeICU admission or a delayed occurrence of ARF during ICU stay; (b) sepsis, and(c) oliguria.

Levy et al. [5] on an objective to determine if the high mortality in ARFis explained by underlying illnesses in cohort analytic study, found that the

ARF in the Intensive Care Unit 13

mortality in subjects without renal failure was 7%, compared with 34% in thecorresponding index subjects with renal failure.

In ICU some 75% of ARF is a consequence of surgery, often associatedwith sepsis. A small focus of fecal contamination in the abdomen in a patientwith acute renal failure or multiple organ failure will prove fatal and, similarly,a segment of dead bowel also negates recovery.

ARF remains an infrequent but major complication of surgery necessitatingcardiopulmonary bypass (CPB): the likelihood of developing ARF after cardiacsurgery depends on factors associated with poor cardiac performance (particularlywhen separating from CPB) and with the level of baseline renal insufficiency [6].

The outcome of patients with ARF complicating cardiac surgery was par-ticularly poor in those with associated cardiovascular failure [7].

Risk factors for ARF in severe trauma increased by age, ISS �17, thepresence of hemoperitoneum, shock, hypotension, rhabdomyolysis with CPK�10,000 IU/l, presence of acute lung injury requiring mechanical ventilation,and Glasgow Coma Score �10 [8].

A renal injury by exotoxins (e.g. antibiotics, anesthetic agents, contrastmedia, diuretics) and endotoxins (e.g. myoglobin) may also be involved. Inpatients with prerenal azotemia renal injury is more likely to be caused by drugsthat can alter intrarenal hemodynamics, such as NSAIDs, or reach high con-centrations in renal tissue, such as aminoglycosides. Patients with preexistingrenal insufficiency are predisposed to acute renal failure due to radiocontrastagents, aminoglycosides, atheroembolism and cardiovascular surgery. Patientswith both renal insufficiency and diabetes mellitus are at particularly high riskfor toxic reactions to radiocontrast agents. Patients with hyperbilirubinemiaalso appear to be predisposed to acute renal failure. Elderly patients are sus-ceptible to many forms of acute renal failure because the aging kidney losesfunctional reserve and its ability to withstand acute insults is compromised [9].

Synergy between renal hypoperfusion and toxic insults from an increasedrenal concentration of toxins at a time when sodium reabsorption and urinaryconcentration are enhanced and the oxygen supply is reduced are shown infigure 1 [10].

Severe infection include a combination of factors such inadequate perfusionwith consequent inadequate delivery of oxygen, the sepsis-endotoxin-cytokinemediator system and superimposed disseminated intravascular coagulation(malignant intravascular inflammation).

In this condition, the hemodynamic of this kidneys is impaired even with-out systemic hypotension: this is nevertheless a common feature of SIRS evenin presence of high cardiac output (hyperdynamic hypotension) and it is due todecrease of, resistances or decrease of critical closing pressure in the arterialcircuit [11, 12].

Piccinni/Carraro/Ricci 14

Recently, an original study is published on behalf of the SOFA groupof the European Society of Intensive Care Medicine on objective to describerisk factors for the development of ARF in ICU, and the association ofARF with MOF and outcome using the sequential organ failure assessment(SOFA) score.

Forty participating centers in 16 countries enrolled 1,411 patients and con-clude that the most important risk factors for development of ARF present onadmission were acute circulatory or respiratory failure; age more than 65 years,presence of infection, past history of chronic heart failure (CHF) lymphoma orleukemia, or cirrhosis. The presence of infection during the ICU stay increasedthe risk of death by all other factors, especially circulatory failure [13].

In other words, there are conditions favoring deterioration of factors thatrender the kidney susceptible to acute injury, like: high O2 consumption;counter current multiplier; high blood flow; unique glomerular structure; renalblood flow autoregulation.

Systemic hypotension induces loss of autoregulatory flow and repetitiveperiods of low perfusion pressure may induce repetitive ischemia in a region ofthe kidney already prone to hypoxia such as the medulla of the thick ascendinglimb of Henle’s loop (mTAL), see fig. 2.

Heterogeneity of intrarenal blood flow contributes to the pathophysiologyof ischemic acute renal failure. An imbalance between the vasodilator nitric

Medullary oxygen insufficiency

↑ Delivery of Nato macula densa

↓ IGF-I

↑ Medullaryblood flow

↓ mTALtransport

Release ofprostanoids, adenosine,

nitric oxide

NSAIDsCortical

vasoconstriction

Volume depletion

Renal failure

Precipitationof Tamm–Horsfall

protein

BJP

Medullary oxygen sufficiency

Tubularobstructionand damage

Fig. 1. Probable mechanisms leading from medullary hypoxia to renal failure. FromBrezis and Rosen [10], modified.


oxide and the vasoconstrictor endothelin may also impair medullary bloodflow and contribute to tubular-cell damage. In the outer medulla, wheretubules have high oxygen requirements, ischemia causes swelling of tubularand endothelial cells as well as adherence of neutrophils to capillaries andvenules. These changes lead to vascular congestion and decreased blood flow,tipping the tenuous balance between oxygenation and energy demand. This‘medullary anginal syndrome’ is a possible cause of apoptosis for part ofparenchyma which had not undergone previous necrosis, i.e. the infinite storyin biology of programmed cell death [14].

Apoptosis is an evolutionarily conserved and highly regulated program ofcell death, which plays an important role in both normal physiologic processesand, when accelerated, in disease states as well. Many reports on apoptosis havefocused on the role of the executioner cysteine-aspartate proteases termed‘caspases’ that are triggered in response to proapoptotic signals and that resultin disassembly of the cell. Recent studies demonstrated the coexistence ofmultiple parallel apoptotic pathways; in mammalian cells, at least four distinctpathways exist. Knowledge of caspase regulation may allow manipulationof apoptosis. A key caspase involved in the apoptotic pathway is caspase.Inhibition of caspase-3 has been linked to prevention of apoptotic death in vitro,although certain stimuli can induce apoptosis by a caspase-3-independent path-way. A second area of apoptosis research focused on the antiapoptotic protein,BCL [15–17].

Blood flow,4.2ml/min/g

Cortical labyrinths

Medullary rays

Cortex

Renal vein Renal artery

Medullary thickascending limbs

PO2, ~50

mm Hg

PO2, ~10–20 mm HgBlood flow,

1.9 ml/min/g

Macula densa

Outermedulla

Innermedulla

Fig. 2. Anatomical and physiologic features of the renal cortex and medulla. FromBrezis and Rosen [10], modified.


Pathophysiology of Ischemic Acute Renal Failure

The pathophysiology of ischemic acute renal failure, because of prerenalazotemia, can be considered a preischemic state and prerenal azotemia andischemia are common causes of acute renal failure. As in any cell, anoxia inkidney cells results in the depletion of energy stores, collapse of electrolytegradients, disruption of the actin cytoskeleton, activation of phospholipases,and changes in gene expression. Renal hypoxia induces the loss of epithelialpolarity along the proximal tubules and the selective induction of growth-response genes with rapid DNA fragmentation (suggestive of apoptosis)along the medullary thick limbs. Ischemic injury to renal vessels increasesrenovascular reactivity and predisposes patients to secondary ischemic insultsfrom hypotension during the recovery from acute renal failure. Ischemiainduces the expression of histocompatibility antigens on renal tubular cellsand of intercellular adhesion molecules on endothelial cells, which leads tothe local aggregation of neutrophils and platelets. After ischemia, intrarenalcongestion is prominent in the outer medulla because of regional hypoxia andbecause the vasa recta are easily compressed by surrounding tubular edema[18, 19].

Acute Morphologic Lesions in Tubules

Anoxic damage along the tubules is governed by the intrinsic vulnerabil-ity of the various nephron segments and by the tissue gradients of oxygenation.Glomeruli and collecting ducts are relatively resistant to a lack of oxygen. Bycontrast, both proximal and distal tubules (especially medullary thick limbs)are intrinsically susceptible to hypoxia [20]. Nevertheless, the distribution oftubular damage in vivo appears to be determined largely by intrarenal oxygengradients.

Some of the cellular events associated with cell death and the restorationof tubule integrity are shown in figure 3 [21].

Ultimately, the mainstay is the maintenance of systemic arterial pressureat a value compatible with renal perfusion pressure: other factors can be con-comitant but don’t play such a primary role.

From the above comes that the risk factors for ARF in ICU are all theseconditions reducing an effective circulating volume and lead to a furtherdecrease in mean arterial pressure and thus also renal perfusion pressure suchas: hypotension, hypoxia, the presence of sepsis and/or SIRS, fluid restrictionand depletion, nephrotoxicity, positive pressure ventilation, and raised intra-abdominal pressure.


All this speculation confirms the idea that ARF is a ‘broad church’ withheterogeneity of causes and clinical presentation making extremely difficult thecomparison of different study populations at different times ARF in ICUpatients. These factors should be taken into account also when testing newapproaches for the prevention and treatment of ARF and other organ dysfunc-tions in the critically ill [22] because the factors implicated in the developmentof ARF and its associated poor prognosis are not well defined in the literatureand studies aimed at identifying risk factors could assist in our understandingof this disease process.

Fig. 3. Tubular-cell injury and repair in ischemic acute renal failure. From Thadhaniet al. [21], modified.


References

1 Groeneveld AB, Tran DD, Van der Meulen J: Nauta JJ, Thijs LG: Acute renal failure in the medicalintensive care unit: Predisposing, complicating factors and outcome. Nephron 1991;59:602–610.

2 Vilkes B, Mailloux L: Acute renal failure: Pathogenesis and prevention. Am J Med 1986;80:1129.3 Pinsky MR, Matuschak GM: Multiple system organ failure: A unifying hypotesis. J Crit Care

1990;5:108–114.4 Brivet F, Kleinknecht D, Loirat P, Landais P: Acute renal failure in intensive care units – causes,

outcome and prognostic factors: A prospective multicentre study. Crit Care Med 1996;24:192–198.5 Levy E, Viscioli C, Horowitz R: The effect of acute renal failure on mortality. A cohort analysis.

JAMA 1996;275:1489–1494.6 Conlon PJ, Stafford-Smith M, White WD: Acute renal failure following cardiac surgery. Nephrol

Dial Transplant 1999;14:1158–1162.7 Ostermann ME, Taube D, Morgan CJ, Evans TW: Acute renal failure following cardiopulmonary

bypass: A changing picture. Intens Care Med 2000;26:565–571.8 Vivino G, Antonelli M, Moro ML, Cottini F, Conti G, Bufi M, Cannata F, Gasparetto A: Risk fac-

tors for acute renal failure in trauma patients. Intens Care Med 1998;24:808–814.9 Endre ZH: Post cardiac surgery acute renal failure in the 1990s. Aust NZ J Med 1997;25:278–279.

10 Brezis M, Rosen S: Hypoxia of the renal medulla: Its implications for disease. N Engl J Med1995;332:647–655.

11 American College of Chest Physisians/Society of Critical Care Medicine Consensus Conference:Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis.Crit Care Med 1992;20:864–874.

12 Bone RC: Immunologic dissonance: A continuing evolution in our understanding of the systemicinflammatory response syndrome (SIRS) and the multiple organ dysfunction syndrome (MODS).Ann Intern Med 1996;125:6780–6787.

13 de Mendonça A, Vincent JL, Suter PM, Moreno R, Dearden NM, Antonelli M: Acute renal failure inthe ICU: Risk factors and outcome evaluated by the SOFA score. Intens Care Med 2000;26:915–921.

14 Abbis IA, Camerons JS: Epidemiology of acute renal failure in the intensive care unit; in Ronco C,Bellomo R (eds): Critical Care Nephrology. Amsterdam, Kluwer Academic Publishers, 1998, pp133–141.

15 Brady H, Singer G: Acute renal failure. Lancet 1995;346:1533–1540.16 Toda N, Takahashi T, et al: Tin chloride pretreatment prevents renal injury in rats with ischemic

acute renal failure. Crit Care Med 2002;30:1512–1522.17 Tetta C, Mariano F, Buades J, Ronco C, Wratten ML, Camussi G: Relevance of platelet-activating

factor in inflammation and sepsis: Mechanisms and kinetics of removal in extracorporeal treatments.Am J Kidney Dis 1997;5(suppl 4):S57–S65.

18 Majno G, Joris I: Apoptosis, oncosis, and necrosis: An overview of cell death. Am J Pathol 1995;146:3–15.

19 Black SC, Huang JQ, Rezaiefar P, et al: Colocalization of the cysteine protease caspase-3 withapoptotic myocytes after in vivo myocardial ischemia and reperfusion in the rat. J Mol Cell Cardiol1998;30:733–742.

20 Kroemer G: The proto-oncogene Bcl-2 and its role in regulating apoptosis. Nat Med 1997;3:614.21 Thadhani R, Pascual M, Bonventre J: Acute renal failure. N Engl J Med 1996;334:1448–1460.22 Vincent JL: Dear SIRS, I’m sorry to say that I don’t like you. Crit Care Med 1997;371–374.

Pasquale PiccinniDepartment of Anesthesiology and ICU, St. Bortolo HospitalVia Rodolfi 37, IT–36100 Vicenza (Italy)Tel. �39 444 993862 – 3400, Fax �39 444 927567E-Mail [email protected]


Pathophysiology of Ischemic AcuteRenal FailureInflammation, Lung-Kidney Cross-Talk, and Biomarkers

Joseph V. Bonventre

Renal Division, Brigham and Women’s Hospital and Department of Medicine,Harvard Medical School, and the Harvard-Massachusetts Institute of Technology,Division of Health Sciences and Technology, Boston, Mass., USA

Ischemic acute renal failure (ARF) is a syndrome that results from a mis-match between oxygen and nutrient delivery to the nephrons and energydemand of the nephrons. At times there is a clearly defined transient drop intotal or regional blood flow to the kidney which results from compromise of thesystemic circulation; at other times, for example in the setting of sepsis, thereduction in perfusion may not be associated with systemic signs of hypoten-sion or circulatory compromise. ARF is frequently associated with multipleorgan failure and sepsis. Despite advances in preventative strategies and sup-port measures, this syndrome continues to be associated with significant mor-bidity and mortality. In animal models, experimental renal ischemia results inrapid loss of cytoskeletal integrity and cell polarity. There is shedding of theproximal tubule brush border, mislocalization of adhesion molecules and othermembrane proteins such as the Na�K�ATPase, as well as apoptosis andnecrosis [37]. With severe injury, viable and non-viable cells are desquamatedleaving regions where the basement membrane remains as the only barrierbetween the filtrate and the peritubular interstitium. This allows for backleak ofthe filtrate, especially under circumstances where the pressure in the tubule isincreased due to intratubular obstruction resulting from cellular debris in thelumen interacting with proteins such as fibronectin which enter the lumen [41].In contrast to the heart or brain, the kidney can completely recover from anischemic or toxic insult that results in cell death. Surviving cells that remainadherent undergo repair with the potential to recover normal renal function.When the kidney recovers from acute injury it relies on a sequence of events

Bonventre 20

that include epithelial cell spreading and possibly migration to cover theexposed areas of the basement membrane, cell dedifferentiation and prolifera-tion to restore cell number, followed by differentiation which results in restora-tion of the functional integrity of the nephron [2]. In this review I will focusbriefly on three aspects of the pathogenesis of acute renal failure: (1) the roleof inflammation; (2) pulmonary-renal cross-talk, and (3) biomarkers of injury.

Inflammation

The pathogenesis of ischemic acute renal failure has been attributed toabnormal regulation of local blood flow following the initial ischemic episode.Persistent generalized preglomerular vasoconstriction may be a contributing fac-tor; however, we believe that a more important pathophysiological component ofischemic ARF is a reduction in local blood flow to the outer medulla. Despitethe extensive literature suggesting that enhanced vasoconstriction, mediated byvarious endogenous vasoactive agents, contributes to the pathophysiology ofacute renal failure, no vasodilator has proved useful in preventing or treatingARF associated with tubular necrosis in man. In animals protective effects ofagents designed to block vasoconstrictors such as endothelin may be related tothe effects of these agents on neutrophil adhesion or other aspects of leukocyte-endothelial interactions [34]. Vasodilators, such as nitric oxide, also can haveeffects to decrease inflammation. NO inhibits TNF-induced adhesion of neutro-phils to endothelial cells, stimulated by TNF-�, which would also be protective[23]. It has been known for quite some time now that there is less flow to themedullary than the cortical region in the postischemic kidney [39]. In addition,as endothelial cells are injured with resulting cell swelling and increased expres-sion of cell adhesion molecules, so also are leukocytes activated. Enhancedleukocyte-endothelial interactions can result in cell-cell adhesion, which canphysically impede blood flow [5]. Furthermore, these interactions will addition-ally activate both leukocytes and endothelial cells and contribute to the genera-tion of local factors that promote vasoconstriction especially in the presence ofother vasoactive mediators, resulting in compromised local blood flow andimpaired tubule cell metabolism [35]. These leukocyte-endothelial interactionslikely impact the outer medulla to a greater extent than the cortex.

The concept that endothelial cells are a target of post-ischemic renalinjury was suggested as early as 1972 by Leaf and colleagues when theydescribed endothelial swelling and narrowing of the blood vessel lumen asimportant features of post-ischemic injury [12]. There is marked congestionof the outer medulla after ischemia [28]. Recently, evidence for endothelialdysfunction in the cortex has also been presented in studies that demonstrated

Pathophysiology of Ischemic Acute Renal Failure 21

retrograde blood flow through peritubular capillaries upon reperfusionfollowing ischemia [7]. With reperfusion, a partial transient compromise ofthe patency of the peritubular capillaries was also observed. When humanumbilical vein endothelial cells or human embryonic kidney cells, stablyexpressing endothelial nitric oxide synthase, were administered either intra-venously or into the renal artery following ischemia, these cells implantedinto the kidney and resulted in partial functional protection against injury [7].In addition, following prolonged ischemia (60 min) in the rat, peritubular cap-illaries suffer permanent damage [1]. The number of microvessels in the innerstripe of the outer medulla declines, with the reduction in number associatedwith increased tubulointerstitial fibrosis and altered concentrating ability.With ischemia-reperfusion, endothelial cells upregulate integrins, selectins,and members of the immunoglobulin superfamily, including intercellularadhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule (VCAM).Administration of anti-ICAM-1 antibodies prior to renal ischemia-reperfusionprotects the kidney from injury [20]. Kidneys of ICAM-1 knockout mice alsoare protected [21].

There are many mechanisms by which leukocytes potentiate renal injury.Leukocytes are activated by inflammatory mediators, including cytokines,chemokines, eicosanoids, and reactive oxygen species (ROS), which upregulateadhesion molecules that engage counter-receptors on the activated epithelium.Leukocytes are recruited and activated by chemokines, which are upregulatedby the proinflammatory cytokines interleukin-1 (IL-1) and tumor necrosisfactor-� (TNF-�). TNF-�, IL-1, and IFN-�, produce a number of injuriouschanges in proximal tubular epithelial cells. These cytokines also disrupt cell-matrix adhesion dependent on �1 integrin, inducing cell shedding into thelumen.

Leukocyte subgroups are likely to contribute in different ways to ischemia-reperfusion injury. Myeloperoxidase activity is elevated soon after ischemicinsult and may originate from macrophages and/or neutrophils [28]. If neu-trophil accumulation is prevented, however, tissue injury is ameliorated [21].It is possible that neutrophil depletion models, however, may not adequatelydifferentiate involvement of neutrophils from T lymphocytes and macrophages.Later phases of ARF are characterized by infiltration of macrophages andT lymphocytes which predominate over neutrophils. Knock-out mice lackingCD4�/CD8�, a cell adhesion receptor on T lymphocytes, are protected fromischemia-reperfusion injury [32], suggesting a causal role for T lymphocytes inmediating injury. In addition, blockade of T cell CD28-B7 costimulation pro-tects against ischemic injury in rats and significantly inhibits T cell andmacrophage infiltration and activation in situ [10]. The role of the T cell, how-ever, has been recently questioned. Mice deficient in recombination-activating

Bonventre 22

gene (RAG)-1 lack T and B cells and do not produce immunoglobulins or T cellreceptor proteins. In the absence of these cells and their receptors, RAG-1 defi-cient mice are not protected from ARF induced by ischemia. Tubular necrosisand neutrophil infiltration are present to a degree comparable to that seen inwild type mice [30].

Complement may also potentiate leukocyte-endothelial interactions. In anumber of different tissues exposed to ischemia-reperfusion, complement-dependent upregulation of endothelial cell adhesion molecules, with result-ing neutrophil accumulation in the vasculature, has been implicated as amechanism for complement-mediated injury [15]. Some investigators, however,suggest that the primary effect of complement in kidney ischemia-reperfusionis on the epithelial cell due to a direct effect of the membrane attack complexof complement [40].

ROS that are generated during reperfusion and as a result of the inflam-matory response play a major role in cell injury. ROS are generated by activatedinfiltrating leukocytes and by epithelial cells. ROS are directly toxic to tubularepithelial cells, with ROS generating systems mimicking the effects of ischemicinjury [24]. In some cases, scavengers (superoxide dismutase, glutathione,vitamin E) and inhibitors of ROS (deferoxamine) have been found to protectagainst renal injury [26]. The presence of ROS can result in the peroxidation oflipids in cell membranes, protein denaturation and DNA strand breaks. Lipidperoxidation by ROS enhances membrane permeability and impairs functionof membrane enzymes and ion pumps. ROS-induced strand breaks in DNAleads to activation of DNA repair mechanisms, including the nuclear enzymepoly(ADP-ribose) synthetase (PARS). PARS transfers ADP-ribose from nicoti-namide adenine diphosphate (NAD) to nuclear proteins. Following ischemia-reperfusion, PARS is activated, NAD is depleted and generation of cellular ATPis inhibited, worsening the injury. This amplification of injury, also called thePARS suicide, may be the major mechanism by which ROS is directly toxicto cells. Inhibiting PARS protects renal cells against ROS-induced injury, inparticular, resulting from ischemia-reperfusion [8]. ROS also upregulatechemokine expression. Transgenic mice overexpressing the anti-oxidants, intra-cellular and extracellular glutathione peroxidases, are protected against ischemicinjury [19]. These animals have less induction of the chemokines, interleukin-8 (IL-8) and monocyte chemotactic protein-1 (MCP-1), less neutrophil infiltra-tion and less functional injury compared to the wild type controls, suggestingthat the effects of ROS are mediated by chemokines. Exposure of leukocytes tocirculating cytokines reduces their deformability and enhances their tendencyto be sequestered [36]. Sequestered leukocytes can then potentiate injury byfurther generating more ROS and eicosanoids, enhancing inflammation andvascular tone.


Tubule Contribution to Inflammatory Injury

Both the S3 segment of the proximal tubule and the medullary thick ascend-ing limb (MTAL) are located in the outer stripe of the outer medulla. This regionof the kidney is marginally oxygenated under normal conditions and after anischemic insult, oxygenation is further compromised because the return in bloodflow is delayed. Both segments of the nephron contribute to the inflammatoryresponse in ARF [4]. The tubule epithelial cells are known to generate proin-flammatory and chemotactic cytokines such as TNF-�, MCP-1, IL-8, inter-leukin-6 (IL-6), interleukin-1� (IL-1�), and TGF-�, MCP-1, IL-8, RANTESand ENA-78 [6]. Proximal tubular epithelial cells may respond to T lymphocyteactivity through activation of receptors for T cell ligands [22]. When CD40 isligated in response to interaction with CD154, there is MCP-1 and IL-8 produc-tion, TRAF6 recruitment, and MAPK activation in the proximal epithelial cell[22]. CD40 also induces RANTES production by human renal tubular epithelia,an effect which is amplified by production of IL-4 and IL-13 by Th2 cells, a sub-population of T cells [11]. B7–1 and B7–2 can be induced on proximal tubuleepithelial cells in vivo and in vitro. After B7–1 and B7–2 induction, proximaltubule epithelial cells costimulate CD28 on T lymphocytes resulting in cytokineproduction [25].

Protection against Injury by Ischemic Preconditioning

An area of increasing interest is the possibility of rendering an organ resist-ant to subsequent injury by a prior insult or preconditioning maneuver. Ischemicpreconditioning of the kidney confers protection against a subsequent ischemicattack [3]. Identification of the mechanisms responsible for renal ischemic pre-conditioning will likely facilitate our understanding of the pathophysiology ofischemic injury and guide the development of novel therapeutics aimed at mim-icking the protective mechanism(s). Several candidates that could potentiallyserve as mediators of preconditioning have been identified. These include acti-vation of NOS [27] and phosphatidylinositol-3-kinase Akt/PKB pathways,reduction in the relative activation of JNK as compared to ERK1/2 [28, 29], andinduction of heat shock proteins, heme oxygenase, and endoplasmic reticulumstress proteins [16]. A recent review discusses many of these candidate mediatorsof preconditioning in more detail [3].

Of these potential mediators, the NOS pathway is particularly important tomention. Recently, this laboratory demonstrated that iNOS is responsible for acomponent of the long-term protection afforded the kidney by ischemic precon-ditioning [27]. Thirty minutes of prior ischemia results in a prolonged increase

Bonventre 24

in the expression of iNOS and eNOS as well as heat shock protein (HSP)-25.In addition, there is increased interstitial expression of alpha smooth muscle actin,an indicator of long-term renal interstitial changes. Gene deletion of iNOS, butnot eNOS, increased kidney susceptibility to ischemia, as did treatment with phar-macological inhibitors of NO synthesis, including N�-nitro-L-arginine (L-NNA)and L-N6-(1-iminoethyl)lysine (L-NIL), the latter a specific inhibitor of iNOS.When the initial period of ischemia was reduced (15 min), there was lessprotection of the kidney from subsequent ischemia on day 8. Under theseconditions, there was no sustained increase in iNOS or eNOS expression, andprotection was not abolished by L-NIL treatment, suggesting that the residualprotection was not related to iNOS. In addition, renal function was not impairedand expression of interstitial alpha-smooth muscle actin did not change. The dataindicate that iNOS plays an important role in kidney protection afforded by pro-longed ischemic preconditioning and that persistent long-term changes in therenal interstitium may be critical in affording this protection by sustaining iNOSsynthesis. In addition, this study is the first to demonstrate in any organ systemthat partial protection persists up to 12 weeks after an initial ischemic event. TheiNOS-independent protection associated with preconditioning may be related, atleast in part, to an upregulation of HSP-25.

Mitogen-activated protein (MAP) kinases are also likely to be involved inaffording protection following ischemic preconditioning [28]. MAP kinases areubiquitously expressed serine/threonine kinases which, in mammalian cells, playcentral roles in determining if the response to multiple signaling inputs will beproliferation, differentiation or apoptosis. In mammalian cells, three MAP kinasecascades have been identified. The best studied of these include the extracellu-lar signal-regulated kinase (ERK) cascade, involved in cell proliferation anddifferentiation in a variety of cell types. The ERK pathway is activated by growthfactors and many other agonists, including vasoactive peptides. Jun N-terminalkinase (JNK), also known as stress-activated protein kinase (SAPK), and p38are components of the two other MAP kinase cascades. JNK and p38 are activatedby inflammatory cytokines (TNF and IL-1) and with cellular stress, includinggenotoxic stress and osmolar stress. They are only minimally activated by growthfactors. JNK has been implicated in proximal tubular cell injury and is increasedfollowing ATP depletion in vitro and ischemia-reperfusion injury in vivo [31].Interestingly, ERK is activated predominantly in distal cells (MTAL) [33], perhapsexplaining in part the differential susceptibility to injury of the proximal anddistal nephron segments. When the kidney is preconditioned to injury by ischemia,after the second ischemic exposure activation of JNK and p38 is markedlyreduced, as is activation of their upstream MAPK kinases (MKK7, MKK4 andMKK3/6). By contrast, activation of ERK1/2 and its upstream MAPKK acti-vator, MKK3/4, is unaltered by preconditioning. Thus, the relative ratio of


ERK1/2 activation to JNK or p38 activation is enhanced in the preconditionedpostischemic kidney.

Lung-Kidney Cross-Talk

Many of our patients with ARF have multiple organ failure and many are onpositive pressure mechanical ventilation. We have recently found that ventilationof the lung has distal effects on the kidney even if the kidney has not been manip-ulated in the rat [9]. Rats were ventilated with room air at 85 breaths/min for 2 hwith either VT 7 or 20 ml/kg. Positive pressure ventilation resulted in increasedmicrovascular leak in the lung that is dependent on nitric oxide synthase (NOS)expression. Kidney microvascular leak, which was assessed by measuring 24-hour urine protein and tissue Evans blue dye, was found to be markedlyincreased. There was significant microvascular leak in both lung and kidney withlarge VT (20 ml/kg) ventilation. Injection of 0.9% NaCl corrected the hypoten-sion and the decreased cardiac output related to large VT, but it did not attenuatemicrovascular leak in lung and kidney. There was an increase in total proteinuriaand albuminuria resulting from the ventilation (fig. 1). Serum vascular endothe-lial growth factor was significantly elevated in large VT groups. Endothelial NOSbut not inducible NOS expression significantly increased in the lung and kidneytissue with large VT ventilation. The NOS inhibitor, N-nitro-L-arginine methyl

VT 7 VT 20 VT 20NSVT 7 VT 20 VT 20NS0

20

40

70

60

50

30

10

24-H

our

tota

l pro

tein

/cre

at(m

g/g

crea

t)

Alb

umin

uria

(mg/

24h)

0

6

10

161412

8

42

**

††

*No treatment L-NAME

Fig. 1. Effects of mechanical ventilation and N-nitro-L-arginine methyl ester (L-NAME) on urine protein excretion. Twenty-four-hour urine total protein normalized tototal creatinine excretion (a) and albumin (b) are depicted. Ventilation was at 85breaths/min for 2 h at 7 (VT 7) or 20 ml/kg without (VT 20) or with (VT 20 NS) normalsaline (10 ml/kg) administration. Large tidal volume ventilation (VT 20), with or withoutnormal saline administration, caused significant proteinuria compared with the VT 7group (n � 6 in each group). Proteinuria and albuminuria were mitigated by treating theanimals with L-NAME *p � 0.05 vs. VT 7; †p � 0.05 vs. no treatment at VT 7. Modifiedwith permission from Choi et al. [9].

Bonventre 26

ester (L-NAME), attenuated the microvascular leak in lung and kidney and theproteinuria seen with hyperventilation. Endothelial NOS may mediate the sys-temic microvascular leak in this model of ventilation-induced lung injury. Thesefindings may have important implications for the pathophysiology of ischemicARF since enhanced microvascular permeability in the renal parenchyma of theouter medulla could be expected to contribute to outer medullary congestion andimpaired blood flow through the microcapillary bed supplying the tubule struc-tures in that region of the kidney.

Biomarkers

A major shortcoming in our ability to conduct clinical studies to test puta-tive therapeutic agents in ARF, and hence test pathophysiological conceptsreflected by the therapeutic approach, is the lack of a biomarker that will heraldthe disease early enough so that intervention can be introduced at a time whenthere is a reasonable chance to alter the natural history of the disease.Therapeutic intervention is in general delayed in ARF. There is no equivalent of‘troponin’ or creatinine phosphokinase (CPK) which can be used as early mark-ers for ischemic injury in the heart. Biomarkers for ischemic injury could bemonitored in the blood or urine. While several urinary proteins have been eval-uated as potential non-invasive markers of renal injury [38] none of these mark-ers have been used successfully to screen for early renal injury or to identify thesite of injury within the kidney. The availability of sensitive and specific urinemarkers of ARF would lead to improvements in diagnosis, better enable moni-toring of therapy, as well as establishment of prognosis and risk assessment. Inaddition, identification of urinary proteins expressed during ARF may lead tothe identification of novel targets for therapy, or of markers that could be usedto evaluate the effectiveness of therapeutic interventions. Hampel et al. [13] pub-lished urinary protein profiles in 25 patients following administration of radio-contrast material. Using surface-enhanced laser desorption/ionization (SELDI)protein chip array-time of flight mass spectrometry, they demonstrated pertur-bations in the patterns of urinary protein expression following cardiac catheter-ization. The temporal course of protein expression differed between patientswith normal renal function and patients with impaired renal function at baseline.Since the patient population was small and the follow-up time was short, no cor-relation was found between changes in urinary protein expression patterns andthe development of ARF. Furthermore, specific proteins were not identified.Nevertheless, it is quite likely that some of the proteins found could serve asmarkers of impending contrast-induced acute renal failure. While these resultsmust be extended and confirmed by larger studies, they suggest that uroscopy


with mass spectrometry may lead to the identification of markers of ARF withimportant diagnostic and/or predictive implications.

Our laboratory has a long-standing interest in identifying clinically usefulmarkers of acute renal failure. We cloned from rats, mice and humans a noveladhesion molecule, kidney injury molecule 1 (KIM-1 names the human formand Kim-1 names the rodent form), which may serve as a marker of ischemictubular injury [17]. In addition to its upregulation in the postischemic kidney,Kidney injury molecule-1 is upregulated in various models of toxic renalinjury in the rat [18]. In addition, it is expressed on the proximal tubuleepithelial cells in human kidney biopsy sections from patients with acutetubular necrosis [14]. Importantly, there is no staining for KIM-1 in normalhuman kidney, suggesting that KIM-1 is a marker of injured renal tissue.

We recently investigated the utility of KIM-1 as a marker of acute tubulardamage in humans. We found that soluble KIM-1 was detected at higher levelsin the urine of patients with acute tubular necrosis (ATN) compared to patientswith other forms of acute renal failure or chronic renal disease [14]. These find-ings suggest that a soluble form of KIM-1 may be a clinically useful marker forproximal renal tubular injury. We also found that levels of KIM-1 increase andreturn to baseline following aortic cross-clamping in a patient with no overtclinical evidence of ARF. This suggests that KIM-1 may be useful as a markerof ATN early in the course of ARF, before other markers of ARF become evi-dent. We are working to further characterize the function of KIM-1 and to char-acterize its utility as a biomarker for ATN.

Conclusions

Ischemic renal injury is a dynamic process that often exists in the contextof multiple organ failure and involves hemodynamic alterations, inflammationand direct injury to the tubular epithelium followed by a repair process thatrestores epithelial differentiation and function. Inflammation is a significantcomponent of this disease, playing a considerable role in its pathophysiology.Although significant progress has been made in defining the major compo-nents of this process, the complex cross-talk between endothelial cells, inflam-matory cells and the injured epithelium, with each generating and oftenresponding to cytokines and chemokines, is not well understood. In addition,we have not adequately integrated the contributions of other organs in thispathophysiology. A better understanding of the pathophysiology underlyingthis disease with better ways to limit the inflammatory component and/or pro-mote the repair process, will depend upon the identification of biomarkers thatcan help us identify the disease early at a time when intervention can alter its

Bonventre 28

natural history. Progress is being made on multiple fronts but we continue tobe humbled by ARF whose mortality rate has not significantly changed overfour decades.

Acknowledgments

This work was supported by the National Institutes of Health (grants DK 39773, DK38452, NS 10828, DK54741, DK 46267).

References

1 Basile DP, Donohoe D, Roethe K, Osborn JL: Renal ischemic injury results in permanent dam-age to peritubular capillaries and influences long-term function. Am J Physiol 2001;281:F887–F899.

2 Bonventre JV: Dedifferentiation and proliferation of surviving epithelial cells in acute renal failure.J Am Soc Nephrol 2003;14(suppl 1):S55–S61.

3 Bonventre JV: Kidney ischemic preconditioning. Curr Opin Nephrol Hypertens 2002;11:43–48.4 Bonventre JV, Brezis M, Siegel N, Rosen S, Portilla D, Venkatachalam M: Acute renal failure.

I. Relative importance of proximal vs. distal tubular injury. Am J Physiol 1998;275:F623–F632.5 Bonventre JV, Weinberg JM: Recent advances in the pathophysiology of ischemic acute renal fail-

ure. J Am Soc Nephrol 2003;14:2199–2210.6 Bonventre JV, Zuk A: Ischemic acute renal failure: An inflammatory disease? Kidney Int 2004;

in press.7 Brodsky SV, Yamamoto T, Tada T, Kim B, Chen J, Kajiya F, Goligorsky MS: Endothelial dys-

function in ischemic acute renal failure: Rescue by transplanted endothelial cells. Am J Physiol2002;282:F1140–F1149.

8 Chatterjee PK, Zacharowski K, Cuzzocrea S, Otto M, Thiemermann C: Inhibitors of poly (ADP-ribose) synthetase reduce renal ischemia-reperfusion injury in the anesthetized rat in vivo. FASEBJ 2000;14:641–651.

9 Choi WI, Quinn DA, Park KM, Moufarrej RK, Jafari B, Syrkina O, Bonventre JV, Hales CA:Systemic microvascular leak in an in vivo rat model of ventilator-induced lung injury. Am J RespirCrit Care Med 2003;167:1627–1632.

10 De Greef KE, Ysebaert DK, Dauwe S, Persy V, Vercauteren SR, Mey D, De Broe ME: Anti-B7–1blocks mononuclear cell adherence in vasa recta after ischemia. Kidney Int 2001;60:1415–1427.

11 Deckers JG, De Haij S, van der Woude FJ, van der Kooij SW, Daha MR, van Kooten C: IL-4 andIL-13 augment cytokine- and CD40-induced RANTES production by human renal tubular epithe-lial cells in vitro. J Am Soc Nephrol 1998;9:1187–1193.

12 Flores J, DiBona DR, Beck CH, Leaf A: The role of cell swelling in ischemic renal damage andthe protective effect of hypertonic solutions. J Clin Invest 1972;51:118–126.

13 Hampel DJ, Sansome C, Sha M, Brodsky S, Lawson WE, Goligorsky MS: Toward proteomics inuroscopy: Urinary protein profiles after radiocontrast medium administration. J Am Soc Nephrol2001;12:1026–1035.

14 Han WK, Bailly V, Abichandani R, Thadhani R, Bonventre JV: Kidney injury molecule-1 (KIM-1):A novel biomarker for human renal proximal tubule injury. Kidney Int 2002;62:237–244.

15 Homeister JW: Lucchesi BR: Complement activation and inhibition in myocardial ischemia andreperfusion injury. Annu Rev Pharmacol Toxicol 1994;34:17–40.

16 Hung CC, Takaharu I, Stevens JL, Bonventre JV: Protection of renal epithelial cells against oxida-tive injury by endoplasmic reticulum stress preconditioning is mediated by ERK1/2 activation.J Biol Chem 2003;278:29317–29326.


17 Ichimura T, Bonventre JV, Bailly V, Wei H, Hession CA, Cate RL, Sanicola M: Kidney injurymolecule-1 (KIM-1), a putative epithelial cell adhesion molecule containing a novel immunoglob-ulin domain, is up-regulated in renal cells after injury. J Biol Chem 1998;273:4135–4142.

18 Ichimura T, Hung CC, Yang SA, Stevens JL, Bonventre JV: Kidney injury molecule-1: A tissueand urinary biomarker for nephrotoxicant-induced renal injury. Am J Physiol Renal Physiol 2004;286:F552–F563.

19 Ishibashi N, Weisbrot-Lefkowitz M, Reuhl K, Inouye M, Mirochnitchenko O: Modulation ofchemokine expression during ischemia/reperfusion in transgenic mice overproducing humanglutathione peroxidases. J Immunol 1999;163:5666–5677.

20 Kelly KJ, Williams WW, Colvin RB, Bonventre JV: Antibody to intercellular adhesion molecule-1protects the kidney against ischemic injury. Proc Natl Acad Sci USA 1994;91:812–816.

21 Kelly KJ, Williams WW, Colvin RB, Meehan SM, Springer TA, Gutierrez-Ramos JC, Bonventre JV:Intercellular adhesion molecule-1 deficient mice are protected against renal ischemia. J ClinInvest 1996;97:1056–1063.

22 Li H, Nord EP: CD40 ligation stimulates MCP-1 and IL-8 production, TRAF6 recruitment, andMAPK activation in proximal tubule cells. Am J Physiol 2002;282:F1020–F1033.

23 Linas S, Whittenburg D, Repine JE: Nitric oxide prevents neutrophil-mediated acute renal failure.Am J Physiol 1997;272:F48–F54.

24 Malis CD, Weber PC, Leaf A, Bonventre JV: Incorporation of marine lipids into mitochondrialmembranes increases susceptibility to damage by calcium and reactive oxygen species: Evidencefor enhanced activation of phospholipase A2 in mitochondria enriched with n-3 fatty acids. ProcNatl Acad Sci USA 1990;87:8845–8849.

25 Niemann-Masanek U, Mueller A, Yard BA, Waldherr R, van der Woude FJ: B7–1 (CD80) andB7–2 (CD 86) expression in human tubular epithelial cells in vivo and in vitro. Nephron2002;92:542–556.

26 Paller MS: Renal work, glutathione and susceptibility to free radical-mediated postischemicinjury. Kidney Int 1988;33:843–849.

27 Park KM, Byun JY, Kramers C, Kim JI, Huang PL, Bonventre JV: Inducible nitric oxide synthaseis an important contributor to prolonged protective effects of ischemic preconditioning in themouse kidney. J Biol Chem 2003;278:27256–27266.

28 Park KM, Chen A, Bonventre JV: Prevention of kidney ischemia/reperfusion-induced functionalinjury and JNK, p38, and MAPK kinase activation by remote ischemic pretreatment. J Biol Chem2001;276:11870–11876.

29 Park KM, Kramers C, Vayssier-Taussat M, Chen A, Bonventre JV: Prevention of kidneyischemia/reperfusion-induced functional injury, MAPK and MAPK kinase activation, and inflam-mation by remote transient ureteral obstruction. J Biol Chem 2002;277:2040–2049.

30 Park P, Haas M, Cunningham PN, Bao L, Alexander JJ, Quigg RJ: Injury in renal ischemia-reperfusion is independent from immunoglobulins and T lymphocytes. Am J Physiol 2002;282:F352–F357.

31 Pombo CM, Bonventre JV, Avruch J, Woodgett JR, Kyriakis JM, Force T: The stress activated proteinkinases are major c-jun amino-terminal kinases activated by ischemia and reperfusion. J Biol Chem1994;269:26545–26551.

32 Rabb H, Daniels F, O’Donnell M, Haq M, Saba SR, Keane W, Tang WW: Pathophysiologicalrole of T lymphocytes in renal ischemia-reperfusion injury in mice. Am J Physiol 2000;279:F525–F531.

33 Safirstein RL, Bonventre JV: Molecular response to ischemic and nephrotoxic acute renal failure;in Schlündorff D, Bonventre JV (eds): Molecular Nephrology. New York, Marcel Dekker, 1995,pp 839–854.

34 Sanz MJ, Johnston B, Issekutz A, Kubes P: Endothelin causes P-selectin-dependent leukocyterolling and adhesion within rat mesenteric microvessels. Am J Physiol 1999;277:H1823–H1830.

35 Sheridan AM, Bonventre JV: Cell biology and molecular mechanisms of injury in ischemic acuterenal failure [In Process Citation]. Curr Opin Nephrol Hypertens 2000;9:427–434.

36 Suwa T, Hogg JC, Klut ME, Hards J, van Eeden SF: Interleukin-6 changes deformability of neu-trophils and induces their sequestration in the lung. Am J Respir Crit Care Med 2001;163:970–976.

Bonventre 30

37 Thadhani R, Pascual M, Bonventre JV: Acute renal failure. N Engl J Med 1996;334:1448–1460.38 Tolkoff-Rubin NE, Rubin RH, Bonventre JV: Noninvasive renal diagnostic studies. Clin Lab Med

1988;8:507–526.39 Vetterlein F, Pethö A, Schmidt G: Distribution of capillary blood flow in rat kidney during postis-

chemic renal failure. Am J Physiol 1986;251:H510–H519.40 Zhou W, Farrar CA, Abe K, Pratt JR, Marsh JE, Wang Y, Stahl GL, Sacks SH: Predominant role

for C5b-9 in renal ischemia/reperfusion injury. J Clin Invest 2000;105:1363–1371.41 Zuk A, Bonventre JV, Matlin KS: Expression of fibronectin splice variants in the postischemic rat

kidney. Am J Physiol 2001;280:F1037–F1053.

Joseph V. Bonventre, MD, PhDBrigham and Women’s Hospital, Renal DivisionMRB-4, 75 Shattuck Street, Boston, MA 02115 (USA)Tel. +1 617 429 2146, E-Mail [email protected]


Pathophysiology of Sepsis and Multiple Organ Failure: Pro- versusAnti-Inflammatory Aspects

Michael R. Pinsky

Department of Critical Care Medicine, University of Pittsburgh, Pittsburgh, Pa., USA

Many of the common severe illnesses, including sepsis and septic shock,severe trauma, burns or other overwhelming stresses, are characterized by thewidespread activation of host innate immunity response. The hallmark of thisactivation is the release of potent inflammatory mediators into the circulation[1]. This response is often referred to as the systemic inflammatory responsesyndrome (SIRS) [2] to accentuate the non-specific inflammatory nature of theprocess as assessed by measures of circulating active small molecules capableof inducing a generalized inflammatory response on remote organ systems.Furthermore, if sustained, this system-wide inflammatory process may result inloss of the normal vasoregulatory adaptations to stress and metabolic demand.This peripheral vascular paralysis often precedes the development of multipleorgan system dysfunction and death in critical illness. Historical perspectives,recent clinical trials of immunomodulating agents and cellular and moleculardata have created an exciting and productive story that only recently lead to thefirst positive outcome trial of pharmacotherapy for severe sepsis.

Initial Cellular Events of Innate Immunity: Outside-In Signaling

Inflammatory processes require immune cell response, recruitment intothe local area and further activation if the inflammatory response is to be sus-tained. This response involves initial cellular recognition of the stimulant viainnate immune systems. This initial recognition process involves a complexinteraction of host-derived co-factors, such as lipopolysaccharide-binding protein,

Pinsky 32

compliment activation and coating of foreign biological material (bacteria) andrelease of pro-coagulant materials. Many of these activation complexes non-specifically bind to the host universal inflammation receptor CD14 whosetransmembrane domain, the Toll-like receptor (TLR), activates an intracellulartyrosine kinase system that eventually activates the oxidant sensitive pro-inflammatory promoter, nuclear factor-�B (NF-�B). Presently, 10 specific TLRproteins have been identified; however, TLR-2 and TLR-4 appear to be primar-ily involved in the innate immunity aspects of circulating monocytes, polymor-phonuclear leukocytes (PMN) and dendrocytes and TLR-4 appears to betrans-membrane moiety that it associates with in order to initiate outside-in cellsignaling [31]. Simultaneously, expression of novel cell surface receptors andgene induction and new protein synthesis occur. Following this immune com-petent cells and responsive parenchymal cells make novel protein speciesneeded to induce a localized inflammatory response.

The release of immune active mediators stimulates primary target cell(immune cell) responses via similar mechanisms to those described by the pri-mary activation sequence, except now more complex cell surface receptors, co-stimulating factors and parallel mediator systems interactions co-exist.These include activation of the contact system (compliment), fibrinolytic sys-tem (thrombin and activate protein C) and paired cytokine stimuli (TNF-� andIL-1�, or IL-8 and MIP-1). Secondary parenchymal cell response via thesesoluble mediators and formed cell interactions then occur via cell-cell, orparacrine activation.

The low molecular weight protein mediators that initiate, sustain and mod-ulated the inflammatory interactions are called cytokines. With the exception ofpre-pro-IL-1, cytokines have to be synthesized de novo in response to a specificexternal stimulus. Thus, the inflammatory response is a determined processrequiring energy and protein synthesis to occur, post-translational protein pro-cessing to mature the mediators, and secretion to get them into the microenvi-ronment needed for cell-cell signaling.

Based on our understanding of their actions, cytokines are presumably pre-sent to modulate cellular response and metabolism on a local or paracrine level.However, once these cytokines gain access to the blood stream in numberscapable of inducing a systemic immune cell response, then a generalizedinflammatory response develops. It is still not clear to what extent this systemicresponse is adaptive or maladaptive to the host. On the one hand, fever and theinduction of release of acute-phase proteins by the liver and heat shock proteinsinto the cells of the body improve survival in bacterial infections, and malaisemay be a very adaptive symptom in limiting the host activity and excess con-sumption of limited energy resources. However, on several levels the general-ized inflammatory response may not confer survival advantage for the host.

Pathophysiology of Sepsis 33

Clearly, on a local level, a violent inflammatory response is highly effective atcontaining, killing and removing foreign biological material (alive or dead) andthe lack of a vigorous local inflammatory response may impair survival. Thiswas best exemplified by the recent documentation that patients with a mildTNF-� response, as characterized by a specific genotype of a mild TNF-�responder, had a greater likelihood of dying from meningococcal meningitisthan patients who had the vigorous TNF-� responds genotype.

Still, persistent activation, rather than massive pulse activation appears tobe detrimental for survival. Pinsky et al. [3] measured the circulating levelsof TNF-�, IL-1�, and IL-2 and the immunomodulating cytokines IL-6 andinterferon-�. Although levels of IL-1, IL-2 and interferon-� did not correlatewith any aspect of the generalized inflammatory response, severity of illnessnor mortality, sustained elevations of the pro-inflammatory cytokine TNF-�and the immunomodulating cytokine IL-6 did. Thus, rather than the maximalserum levels of any specific cytokines, those patients who will subsequentlydevelop multiple organ dysfunction and may eventually die display a persistentelevation of TNF-� and IL-6 in their blood [3–5].

This observation of a sustained plasma elevation of TNF-� and IL-6 insevere sepsis and its association with a poor outcome initially led to the erro-neous assumption that SIRS reflected an isolated excess pro-inflammatorystate. However, the increased expression of pro-inflammatory cytokines is onlypart of this evolving immunologic picture. First Goldie et al. and then othersdemonstrated that both pro-inflammatory cytokines (TNF-�, IL-1�, IL-6, andIL-8) and anti-inflammatory species (IL-1 receptor antagonist (IL-1ra), IL-10,and the soluble TNF-� receptors I and II (sTNFr-I and sTNFr-II), respectively)co-exist in the circulation, and the anti-inflammatory specifics are markedlyincreased in patients with established SIRS [6, 7].

Thus, SIRS or sepsis may be more accurately described as a malignantform of intravascular inflammation, rather than merely the over expression ofsimply pro- or anti-inflammatory substances. This paradoxical expression in theblood of pro- and anti-inflammatory molecular species creates a cacophony ofunregulated immunologic noise within the internal milieu that if sustainedimpairs host adaptability to stress. Since cytokines appear to be an importantcomponent of the SIRS we shall first consider cytokine actions.

Intracellular Inflammatory Events

The inflammatory response requires an inflammatory stimulus to initiate it.Initiating inflammatory mediators include bacterial and yeast cell walls, such asendotoxin (lipopolysaccharides or LPS), exotoxins and other products. Vascular

Pinsky 34

collagen, damaged endothelial surfaces, the denatured proteins caused by thermalinjury and oxidative stress all stimulate the innate inflammatory response (fig. 1).These inflammatory mediators bind to cell surface receptors inducing a trans-membrane signal transduction and intracellular response via activation of severalspecific gene promoter proteins. Numerous unrelated exogenous stimulants,such as endo- and exotoxins and endogenous stimulants, activated compliment(C5a), Hageman’s factor (XIIa), and products of generalized cell injury can

Hsp70

CD11b upregulation

and activation

Mitochondrial oxidative stress

LPS TNF-�

TNF-�

IL-1�

IF-�

C5a

TLR-4

CD14 Serine kinase and MAP kinase

HSF

NF-�B

I-�B

p65p50

TNF-� mRNA

Hsp70 mRNA

Hsp70

Key intracellular events in innate immunity

�

�

�

�

TNF-�

�

�

�

�

�

�

I-�B�

Oxidative stress

iNOS and COX2 mRNA

Cytochrome C

Fig. 1. A schematic representation of some of the intracellular processes of pro- andanti-intracellular processes operative in sepsis. External signaling, either from mediatorbinding to cell surface receptors of CD11b activation induces a series of redundant and par-allel intracellular kinase reactions through Toll-like receptors (TLR). Part of this initial acti-vation is the up-regulation the cell surface expression of CD11b/CD18. Inflammationprimarily induces an intracellular oxidative stress that can be measured by changes in mito-chondrial membrane potential. Oxidative stress activates nuclear factor �B (NF-�B) byphosphorylation cleavage of its inhibitory sub-unit I-�B as well as stressing the mitochon-dria. I�B is degraded by phosphorylation and ubiquinization and the free remaining activep50-p65 dimer translocates into the nucleus where it binds to promoter sites initiating and/oraugmenting mRNA synthesis of specific genes that code for the synthesis of pro-inflammatorymediators (TNF-�, IL-1, IL-8, iNOS, COX2, etc.). Intracellular oxidative stress also acti-vates heat shock factor (HSF) that, like NF-kB, migrates into the nucleus and binds to its ownpromoter regions that augment the synthesis of mRNA coding for heat shock protein (Hsp).One such Hsp is I-kB and others of the 70s class (Hsp70). Hsp70 down-regulates NF-�B activation by both limiting subsequent mitochondrial oxidation and by limiting NF-�B phosphorylation. Also, Hsp70 by limiting mitochondrial oxidative stress minimizescytochrome release thus decreasing apoptotic stress.


induce immune competent cells (usually modeled as monocytes) to synthesizeTNF-� and IL-1�. Presently, numerous cell surface receptors have been identi-fied [reviewed in reference 31]. One cell-surface receptor with an amazing broadrange of affinities to a variety of different types of pro-inflammatory substancesis CD14. Many foreign substances, such as endotoxin require initial fusing withlipopolysaccharide binding protein before they can bind to CD14. Thus, host-derived factors (usually proteins) are probably necessary for the host recognitionof foreign materials as foreign and initiate the inflammatory response. However,no such intermediate binding appears necessary for stimulation of the inflam-matory response via the intrinsic mediators. CD14 then combines with TLR4 toinduce a trans-membrane activation of a series of serine kinases that releasepotent pro-inflammatory promoter proteins which markedly enhance translationof numerous genes encoding proteins essential for the host inflammatoryresponse. The major oxidative stress proinflammatory promoter protein isnuclear factor kappa B (NF-�B), so named because it was originally describedas the kappa light chain promoter protein for B lymphocytes.

NF-�B is an intracellular species that is resident in the cytosol as a p50-p65 heterodimer attached to an inhibitory sub-unit called I�B. The I�B subunitbinds to the p65 subunit masking is nuclear binding site. Activation of the het-erodimer follows cleavage of I�B by serine kinase-induced I�B phosphoryla-tion. Serine kinase, in turn requires an intracellular oxidative stress to trigger itsactivation via the TLR4 activation. Importantly, oxidative stress is an earlyevent in inflammation. The phosphorylated I�B is rapidly digested. The activep50-p65 heterodimer migrates into the nucleus, binds to various promoter siteson the genome stimulating mRNA synthesis of genes coding for many if notmost of the pro-inflammatory cytokines. A partial list of NF-�B inducible pro-teins is shown in table 1, and includes all the known potent proinflammatorymediators, such as TNF-� and IL-1�, as well as the pro-inflammatory enzymesinducible nitric oxide synthase (iNOS) and cyclooxygenase 2 (COX-2).

NF-�B is an excellent target for modulating the intracellular inflammatoryresponse. In fact, several intracellular mechanisms exist that modulate NF-�Bactivity. Studies of endotoxin tolerance allow for examination of these processes.Exposure to small amounts of endotoxin induces an endotoxin-tolerant state inboth cell culture and animal models about 8 hours later. In this state, subsequentexposure to a previously lethal dose of endotoxin often does not induce the fatalpro-inflammatory state. The endotoxin-tolerate state lasts in decreasing strengthfor about 24–36 h, depending on the species and the initial dose of endotoxin.Interestingly, following the induction of an endotoxin-tolerant state, the initialsteps of pro-inflammatory signal transduction up to cleavage of I�B can occur.However, the liberated NF-�B appears to be dysfunctional. The reasons for thisdysfunction are multiple and not fully defined.

Pinsky 36

Under baseline conditions the body appears to have a distinctly anti-inflammatory flavor. Thus, the initial pro-inflammatory processes operate on asystem that at rest exists as a generalized anti-inflammatory one. For example, theendothelium is anti-thrombotic reducing platelet adhesion. Even the immuneeffector cells carry their preformed pro-inflammatory promoter proteins in adysfunctional form. This resting anti-inflammatory state actually becomes morepronounced during generalized sepsis. Experimentally, it has been shown thatinflammation induces its own counter-regulatory anti-inflammatory response thatin the setting of sustain inflammatory stimuli causes a generalized anergic state,referred to as ‘inflammatory stimuli-induced anergy’ or ‘endotoxin tolerance’ [8].Presumably, inflammatory stimuli-induced anergy minimizes the inflammatoryresponse, preventing a chain-reaction, system-wide activation of the inflamma-tory processes. However, it also limits the host’s subsequent ability to mount anappropriate inflammatory defense to infection. Anti-inflammatory cytokines,such as IL-4, IL-10 and somewhat by IL-1� (but not TNF-�), IL-6 or IL-8, andtransforming growth factor-� (TGF-�) can induce this anergic response.

NF-�B Activation Biology

NF-�B activation requires cleavage of the I�B subunit off the p65 subuit.I�B kinase causes the phosphorylation and subsequent degradation of I�B

Table 1. Some of the known target genes regulated by NF-�B

Cytokines TNF-�TNF-�IL-2IL-6IL-8interferon-�G-CSFGM-CSF

Adhesion molecular endothelial-leukocyte adhesion molecule-1vascular cell adhesion molecule-1

Immunoreceptors immunoglobulin �-light chainT cell receptor A and Bmajor histocompatibility class 1 and 2tissue factor-1

Acute phase proteins complement factor Bcomplement factor C4

Others COX-2lipoxgenase 1


allowing the translocation of NF-�B dimers into the nucleus. I�B binds to theresponsive element of the p65 subunit and has a greater affinity for the p65 sub-unit that does the corresponding response element on the genome. Thus, I�Bexcess can pull active p50-p65 dimers off their promoter sites in the nucleus orprevent their release from the cytosol all together. NF-�B dimers exist in sev-eral forms. However, the two most common forms are the heterodimer p50-p65,which is comprised of the p65 subunit with its active DNA consensus domainbinding site and a smaller subunit p50 subunit devoid the active binding site.Importantly, the p65 subunit with its DNA consensus domain-binding siteallows gene activation once bound to such promoter regions on chromosomes.The p50 monomer has no such activity [9]. The other common NF-�B form isthe p50-p50 homodimer, which does not bind I�B or the genome.

Endotoxin-tolerance induced NF-�B dysfunction reflects excess synthesisof the inhibitor of NF-�B, I�B-� [10] presumably due to the absence of induc-tion of I�B kinase. When endotoxin tolerant cells are challenged with a seconddose of LPS, cytosolic levels of I�B are not reduced, as they are with the initialchallenge, and I�B remains in cytoplasm where it sequesters free NF-�Bdimers [11]. Furthermore, the I�B-� promoter can be up-regulated by NF-�B,thus providing a negative feedback loop for further NF-�B activation [12].Another adaptive mechanism involves the balance of NF-�B p50-p65 to p50-p50 species. The ratio of p65-p50 to p50-p50 determines transcription rates ofgenes coding for the NF-�B inflammatory proteins. Ratios of NF-�B p50-p65heterodimer to p50-p50 homodimer of 1.8 � 0.6 or greater are associated withNF-�B activation inducing mRNA synthesis of these genes, while a ratio of0.8 � 0.1 or less confers lack of mRNA transcription following cleavage of I�B[9]. Other post-translational controls on mRNA and protein trafficking alsooccur but at present have not been identified as being as important as these ini-tial steps.

Down-regulation of NF-�B-related intercellular processes is an importantaspect of the overall intracellular inflammatory response. However, these mech-anisms do not explain why both pro- and anti-inflammatory activation is oftensustained in critical illness. Anti-inflammatory pathways are activated by thesame stimuli that active pro-inflammatory pathways. The cell does have severalanti-inflammatory processes. However, the heat shock protein system is themost prevalent, in terms of mass of protein and scope of its oversight.

Heat Shock Proteins and the Stress Response

The heat shock protein system is a complex, widespread and overarchingbasic cellular defense mechanism against numerous stresses, such as fever,

Pinsky 38

trauma, and inflammation [13]. It is the oldest philogenetic cellar defensemechanism identified. The heat shock proteins constitute a significant propor-tion of the intracellular protein pool. At baseline in the un-stimulated cell theyreflect 2% of the resting cellular protein component, whereas in the stressedcell they can comprise up to 20% of the cellular protein pool. Heat shock pro-teins have several important and varied functions. Their singular distinguishingcharacteristic is that their expression is up regulated in response to intracellularstress. Although originally described as a specific response to heat shock, it wassubsequently demonstrated that many stresses, most notably oxidative stress(H2O2, peroxynitrate, ionized radiation) and even some pharmacologic agentscould induce the up-regulation of the heat shock response.

The exact mechanism by which heat induces the activation of the heatshock response is unknown. However, the initial intracellular signaling mecha-nisms are the release of heat shock factor (HSF) from its cytosolic pool com-plexed with three molecules of heat shock protein 70 (HSP-70). Presumably,oxidative stress induced protein unfolding (denaturation) stimulates HSP-70 tobind to these unfolded proteins and refold them, releasing its co-traveler HSFat the same time. HSF, which like NF-�B, is a protein resident in the cytosolthat is activated by the intracellular oxidative stress to make a promoter proteinto bind to specific genes to stimulate synthesis of heat shock proteins. Heatshock proteins are classified by the molecular size. This classification is notarbitrary because their functions tend to also follow class size. They are cellu-lar chaperones that prevent nascent proteins from being degraded by cytosolicenzymes. They also aid in the subsequent quaternary folding of nascent pro-teins prior to export. As a repair process, they aid in the tertiary and quaternaryrefolding of intracellular protein structure after heat or ionizing radiation-induced unfolding (denaturation).

Heat shock proteins confer a survival advantage to their host. Thermal pre-treatment is associated with attenuated lung damage in a rat model of acute lunginjury induced by intratracheal instillation of phospholipase A2 [14]. Thermal pre-treatment reduces mortality rate and sepsis-induced acute lung injury produced bycecal ligation and perforation [15]. The subsequent increased expression of abroad variety of heat shock proteins confers a non-specific protection from notonly subsequent oxidative stress but also minimizes the cellular response to pro-inflammatory stimuli. Survival, in cold blooded animals given an infectious inocu-lum, is linearly related to body temperature. Finally, if specific heat shock proteinsare depleted, then multiple intracellular signaling processes can be affected.

Both HSF and NF-�B are induced to become active by similar stimuli. Itis not known if all the processes that stimulate HSF activation are the same asthose that stimulate NF-�B activation, but during sepsis and other forms ofintracellular oxidative stress HSF and NF-�B are activated in a parallel fashion.


Mitochondria as Both a Marker and Target of Intracellular Inflammation

The production of reactive oxygen species and an associated oxidativestress on the mitochondria is an initial step in the intracellular activation of theinflammatory pathway. It links mediator binding to the cell surface, ischemia-reperfusion and traumatic injury with inflammatory gene activation [16].Mitochondria are the primary energy sources of the cell. They produce ATP bydriving electrons down a chemi-osmotic gradient via a transmembrane chargeinduced by the Krebs’s cycle inside their inner membranes. Loss of internalmembrane polarization is a cardinal sign of mitochondrial energy failure andresults in leakage of cytochrome C from the mitochondria into the cytosol.Importantly, cytochrome C activates the intrinsic caspase system to initiateapoptosis or programmed cell death. Thus, preventing mitochondrial depolar-ization would be an important cytoprotective mechanism.

The HSP-70 family of heat shock proteins prevents mitochondrial oxida-tive stress and blunts the inflammatory response. HSP-70 also minimizes nitricoxide, oxygen free radical and stretozotocin cytotoxicity [17]. For these andother reasons, HSP-70 is an inducible protective agent in myocardium againstischemia, reperfusion injury and nitric oxide toxicity [18]. Nitric oxide-inducedHSP synthesis blunts the inflammatory response to endotoxin and TNF-� invitro. Evidence that HSP-70 may be active in human sepsis comes from theobservation that higher HSP-70 expression is seen in peripheral mononuclearcells in septic patients. Although not conclusive, these data strongly suggestthat heat shock proteins, and HSP-70 in particular, may be important in modu-lating the intracellular inflammatory signal acting at the level of NK-�B.Importantly, Wong et al. [19] identified a potential heat shock responsive ele-ment in the I-�B� promoter that can be activated by HSF after heat shock. Andthe heat shock response can also modulate NF-�B inactivation by this increasedI-�B� expression. Thus, the primary inhibitor of NF-�B activation, I-�B� isitself a heat shock protein. The interaction between the heat shock protein sys-tem and the pro-inflammatory pathway, however, has yet to be defined.

Is Sepsis Characterized by Excessive Pro- and Anti-InflammatoryActivity? Assessment by Inside-Out Signaling

Both pro- and anti-inflammatory processes are active during acute injurystates. This is observable as intracellular activation of both NF-�B and HSF andextracellular expression of both pro- and anti-inflammatory mediators in thebloodstream. Thus, the immune effector cells and parenchyma receive mixed

Pinsky 40

messages during acute injury. However, the phenotypic response of pro- or anti-inflammatory posturing that they make is often difficult to predict. The normalcellular inflammatory response is essential to survival. It localizes and elimi-nates foreign material, including microorganisms. Similarly, some degree ofsystemic inflammatory response is adaptive. Fever reduces microorganismgrowth, malaise causes the host to rest, and induces both the intracellular heatshock response and the extracellular acute-phase protein release. Both of whichminimize oxidative injury. However, inflammation is also destructive. Localabscess formation causes local necrosis, and generalized inflammation caninduce both increased necrosis and apoptosis leading to multiple systems organfailure [30]. However, sepsis also carries a strong anti-inflammatory response,with expression of anergy and increased susceptibility to nosocomial infection.

All acute severe processes result in the expression in the systemic circula-tion of cytokines. However, it is difficult to assess the degree of inflammatorystimulation by measuring serum cytokine levels. Serum cytokine levels canchange within minutes and may be very different in adjacent tissue compart-ments [20]. Although an excellent marker of disease severity and a good posi-tive predictor of the subsequent development of remote organ systemdysfunction, measuring blood levels of cytokines, such as IL-6 [22], do not aidin defining the pro- to anti-inflammatory balance on in predicting response totherapy.

We shifted our attention to the examination of the functional status of cir-culating immune effector cells. Since polymorphonuclear leukocyte (PMN)activation and localization represent the initial cellular host defense againstinfection, their tight control is essential to prevent widespread non-specificinjury. Subsequently, monocytes localize at the site of inflammation. Theiractivity appears to become the predominant process in both host defense andrepair, especially during the second and third day onward in the course of acuteillness. Thus, inhibition of monocyte immune responsiveness is a powerfulmechanism to downregulate the inflammatory response. Anergy is a cardinalcharacteristic of severe illness and reflects macrophage inhibition of antigenprocessing. Importantly, antigen processing reflects a primary aspect of thiscellular response. In this regard, the cell surface receptor family, HLA-DR, isresponsible for antigen presentation to antigen-processing cells. Immaturemonocytes cannot process antigen and have lower cell surface HLA-DR levels.Volk et al. [21] demonstrated that monocytes require HLA-DR levels �20% fornormal cell meditated immunity. Lower levels of HLA-DR expression confirmimmune suppression. Consistent with the overall theme of increased anti-inflammatory responses in severe sepsis, these workers and others have founda profound decrease in HLA-DR expression on circulating monocytes frompatients with sepsis.


Rosenbloom et al. [22] examined the relation between circulating cytokinelevels and the expression of the strong �2-integrin surface cell adhesion mole-cules on circulating immune effector cells. The activation state of circulatingimmunocytes can be indirectly assessed by measuring the intensity of displayon their �2-integrin surface cell adhesion molecules. They showed that all cir-culating immune effector cells, including PMNs, lymphocytes and monocytesare activated in critical illness. Furthermore, the level of circulating immuneeffector cell activation is proportional to mean circulating IL-6 levels [22].Importantly, the degree of organ dysfunction but not the level of shock severitycorrelated with CD11b expression. Since the level of activation of circulatingPMN, as measured by total PMN count and its display of immature forms, isused clinically as an indicator of the host response to systemic inflammation,assessment of PMN responsiveness should also be a good measure of the pro-and anti-inflammatory balance in severe sepsis. Prior studies have shown thatPMNs can be both overactive [23] and dysfunctional [24], and that their CD11bdisplay can be either decreased [25] or increased [26] in critically ill patients.In essence, measuring immune responsiveness of circulating immune effectorcells can be used as a functional bioassay.

One can also expose native PMN to pro-inflammatory stimuli in vitro andassay their subsequent expression of CD11b. Many proinflammatory stimuli,such as TNF-�, IL-8 and phorbol ester can be used to induce this response. Thede novo display of CD11b on circulating PMNs and its subsequent change inexpression of both total CD11b and its avid form, CBRM1/5 epitope, inresponse to in vitro stimulation to TNF-� or IL-8 characterize the in vivo stateof PMN activation and responsiveness. Circulating PMNs of septic humanshave a similar phenotype characterized by high CD11b and low L-selectinexpression [27]. This is the phenotype of acute activation of the inflammatoryresponse. Thus, severe sepsis is associated with an increased de novo activationof circulating immune effector cells. Paradoxically, however, those samePMNs with a sustained inflammatory state are also impaired in their abilityto upregulate CD11b further or to change surface CD11b to the avid state [28]in response to an ex vivo challenge by exposure to biologically significant lev-els of TNF-�. Furthermore, circulating PMN from subjects with severe sepsishave impaired phagocytosis, reduced oxygen burst capacity and diminishedin vitro adhesiveness [29]. This desensitization to exogenous TNF-� or IL-8 wasnot due to a loss of TNF receptors because the cell surface TNF-� receptordensity is not reduced on the cells of septic patients. Importantly, the hypo-responsiveness is extended across all circulating PMN and monocytes in criti-cally ill patients [27].

The NF-�B activity and in vitro responsiveness to LPS of peripheral bloodmonocytes from septic subjects displays a similar pattern to that seen with

Pinsky 42

endotoxin tolerance [30]. The cause of the reduced nuclear translocation of NF-�B is apparently not due to a negative feedback interaction because cyto-solic I�B are not increased in sepsis. Rather one sees an increase in the pro-portion of the inactive p50-p50 species relative to the active p65-p50 species,suggesting that alteration of NF-�B synthesis per se is the primary regulatoryfunction seen in the immune suppression of sepsis. In one study [30], survivorshad higher levels of NF-�B than non-survivors, again suggesting that althoughdownregulation of inflammation is a normal aspect of sepsis, and excessiveinhibition of the process is associated with a poor prognosis.

Acknowledgement

This work was supported in part by the federal government (GM61992–03, NHLBI K-24 HL67181–01A2, and NRSA 2-T32 HL07820–06).

References

1 Schlag G, Redl H: Mediators of injury and inflammation. World J Surg 1996;20:406–410.2 Bone RC: Toward a theory regarding the pathogenesis of the systemic inflammatory response syn-

drome: What we do and do not know about cytokine regulation. Crit Care Med 1996;24:163–172.3 Pinsky MR, Vincent JL, Deviere J, Alegre M, Kahn RJ, Dupont E: Serum cytokine levels in human

septic shock: Relation to multiple-system organ failure and mortality. Chest 1993;103:565–575.4 Thijs LG, Hack CE: Time course of cytokine levels in sepsis. Intensive Care Med 1995;21

(suppl 2):S258–S263.5 Blackwell TS, Christman JW: Sepsis and cytokines: Current status. Br J Anaesth 1996;77:110–117.6 Goldie AS, Fearon KC, Ross JA, Barclay GR, Jackson RE, Grant IS, Ramsay G, Blyth AS,

Howie JC: Natural cytokine antagonists and endogenous anti-endotoxin core antibodies in sepsissyndrome. The Sepsis Intervention Group. JAMA 1995;274:172–217.

7 Vanderpoll T, Malefyt RD, Coyle SM, Lowry SF: Anti-inflammatory cytokine responses duringclinical sepsis and experimental endotoxemia – Sequential measurements of plasma soluble inter-leukin (IL)-1 receptor type Ii, IL-10, and IL-13. J Infect Dis 1997;175:118–122.

8 Cavaillon JM: The nonspecific nature of endotoxin tolerance. Trends Microbiol 1995;3:320–324.9 Ziegler-Heitbrock HWL, Wedel A, Schraut W, Strobel M, Wendelgass P, Sterndorf T, Bauerle PA,

Haas JG, Riethmuller G: Tolerance to lipopolysaccharide involves mobilization of nuclear factor�B with predominance of p50 homodimers. J Biol Chem 1994;269:17001–17004.

10 Larue KEA, McCall CE: A liable transcriptional repressor modulates endotoxin tolerance. J ExpMed 1994;180:2269–2275.

11 Kohler NG, Joly A: The involvement of an LPS inducible I�B kinase in endotoxin tolerance.Biochem Biophys Res Commun 1997;232:602–607.

12 De Martin R, Vanhove B, Cheng Q, Hofer E, Csizmadia V, Winckler H, Bach FH: Cytokine-inducible expression in endothelial cells of an I�B�-like gene is regulated by NF�B. EMBO J1993;12:2773–2779.

13 Jäättelä M, Wising D: Heat shock proteins protect cells from monocyte cytotoxicity: Possiblemechanism of self-protection. J Exp Med 1993;177:231–236.

14 Villar J, Edelson JD, Post M, Mullen BM, Slutsky AS: Induction of the heat stress proteins is asso-ciated with decreased mortality in an animal model of acute lung injury. Am Rev Respir Dis 1993;147:177–181.


15 Villar J, Ribiero SP, Mullen BM, Kuliszewski M, Post M, Slutsky AS: Induction of the heat shockresponse reduces mortality rate and organ damage in a sepsis-induced acute lung injury model.Crit Care Med 1994;22:914–921.

16 Polla BS, Jacquier-Sarlin MR, Kantengwa S, Mariethoz E, Hennet T, Russo-Marie F, Cossarizza A:TNF-� alters mitochondrial membrane potential in L929 but not in TNF�-resistance L929.12cells: Relationship with the expression of stress proteins, annexin 1 and superoxide dismutaseactivity. Free Rad Res 1996;25:125–131.

17 Bellmann K, Wenz A, Radons J, Burkart V, Kleemann R, Kolb H: Heat shock induces resistancein rat pancreatic islet cells against nitric oxide, oxygen radicals and streptozototocin toxicity in vitro.J Clin Invest 1995;95:2840–2845.

18 Malyshev IY, Malugin AV, Golubeva LY, Zenina TA, Manukhina EB, Mikoyan VD, Vanin AF:Nitric oxide donor induces HSP70 accumulation in the heart and in cultured cells. FEBS Lett1996;391:21–23.

19 Wong HR, Ryan M, Wispé JR: Stress response decreases N-F�B nuclear translocation andincreases I-�B� expression in A549 cells. J Clin Invest 1997;99:2423–2428.

20 Boutten A, Dehoux MS, Seta N, Ostinelli J, Venembre P, Crestani B, Dombret MC, Durand G,Aubier M: Compartmentalized IL-8 and elastase release within the human lung in unilateral pneu-monia. Am J Respir Crit Care Med 1996;153:336–342.

21 Docke WD, Syrbe U, Meinecke A, Platzer C, Makki A, Asadullah K, Klug C, Zuckermann H,Reinke P, Brunner H, von Baehr R, Volk HD: Improvement of monocyte function: A new thera-peutic approach; in Reinhart K, Eyrich K, Sprung C (eds): Sepsis: Current Perspectives inPathophysiology and Therapy. Berlin, Springer, 1994, pp 473–500.

22 Rosenbloom AJ, Pinsky MR, Bryant JL, Shin A, Tran T, Whiteside T: Leukocyte activation inthe peripheral blood of patients with cirrhosis of the liver and SIRS: Correlation with seruminterleukin-6 levels and organ dysfunction. JAMA 1995;274:58–65.

23 Trautinger F, Hammerle AF, Poschl G, Micksche M: Respiratory burst capability of polymor-phonuclear neutrophils and TNF-alpha serum levels in relationship to the development of septicsyndrome in critically ill patients. J Leukocyte Biol 1991;49:449–454.

24 McCall CE, Grosso-Wilmoth LM, LaRue K, Guzman RN, Cousart SL: Tolerance to endotoxin-induced expression of the interleukin-1 beta gene in blood neutrophils of humans with the sepsissyndrome. J Clin Invest 1993;91:853–861.

25 Nakae H, Endo S, Inada K, Takakuwa T, Kasai T: Changes in adhesion molecule levels in sepsis.Res Comm Mol Pathol Pharmacol 1996;91:329–338.

26 Lin RY, Astiz ME, Saxon JC, Rackow EC: Altered leukocyte immunophenotypes in septic shock.Studies of HLA-DR, CD11b, CD14, and IL-2R expression. Chest 1993;104:847–853.

27 Rosenbloom AJ, Pinsky, MR, Napolitano C, Nguyen T-S, Levann D, Pencosky N, Dorrance A,Ray BK, Whiteside T: Suppression of cytokine mediated �2-integrin activation on circulating neu-trophils in critically ill patients. J Leukocyte Biol 1999;66:83–89.

28 Diamond MS, Springer TA: A Subpopulation of Mac-1 (CD11b/CD18) Molecules MediatesNeutrophil Adhesion to ICAM-1 and Fibrinogen. J Cell Biol 1993;120:545–556.

29 Terregino CA, Lubkin C, Thom SR: Impaired neutrophil adherence as an early marker of systemicinflammatory response syndrome and severe sepsis. Ann Emerg Med 1997;29:400–403.

30 Adib-Conquy M, Adrie C, Moine P, Ashnoune K, Fitting C, Pinsky MR, Dhainaut J-F, Cavaillon J-M: NF-AB expression in mononuclear cells of septic patients resembles that observedin LPS-tolerance. Am J Respir Crit Care Med 2000;162:1877–1883.

31 Lien E, Ingalls RR: Toll-like receptors. Crit Care Med 2002;30:S1–S11.

Michael R. Pinsky, MD606 Scaife Hall, 3550 Terrace StreetPittsburgh, PA 15261 (USA)Tel. �1 412 647 5387, Fax �1 412 647 8060E-Mail [email protected]


Tropical Acute Renal Failure

Rashad S. Barsoum

Cairo Kidney Center, Cairo, Egypt

Acute renal failure (ARF) is one of the most challenging medical problemsin the tropics. Typical weather conditions and compromised public health stan-dards impose a highly polluted bio-ecological environment, which leads to ahigh prevalence of both primary and secondary infection. Late or misdiagnosisand inadequate management are often responsible for important complications,including ARF. Another key factor is the lack of professional primary medicalcare in remote areas in the tropics. Millions of people are treated by WitchDoctors, Sangomas and quacks, who often use traditional medications withobscure composition and side effects. These agents have been frequentlyblamed in the pathogenesis of toxic ARF, in addition to other organ damage.Secondary and tertiary care are also compromised by the lack of expertise,equipment and funds. All these factors make of ARF an issue of major impor-tance and grave prognosis in tropical countries.

Epidemiology

There are no reliable statistics about the incidence of ARF in the trop-ics. Based on regional sporadic publications, it is possible to estimate the aver-age annual incidence around 150 cases per million population [1]. This iscertainly an underestimate, reflecting only those referred to large hospitalsstaffed with enthusiastic physicians who are keen to report. Nevertheless,these cases constitute the major source of our information on tropical ARF.As shown in table 1, medical causes of ARF are far more common in thetropics, as opposed to the predominance of surgical causes in the developedworld.

Tropical Acute Renal Failure 45

Medical Causes of Tropical ARF

Infection and chemical intoxication are the two major causes of medical ARFin the tropics. Their relative incidence varies in different countries, dependingon the local bioecology and social behavior. Figure 1 shows some exampleswhich illustrate this discrepancy.

Infection

Primary tropical infections can lead to acute renal injury in four ways:(a) Direct invasion of the renal parenchyma by microbial agents; (b) Inductionof an immune response that leads to glomerular or tubulointerstitial inflamma-tion; (c) Induction of hemodynamic disturbances that lead to tubular necrosis,and (d) Iatrogenic renal injury associated with treatment or prophylaxis againsttropical infections.

Direct Parenchymal InvasionMany bacterial agents may cause acute hematogenous ‘pyelonephritis’.

The lesions are mostly interstitial. They may be diffuse or focal, the latter takingthe form of micro-abscesses as with septicemic melioidosis, or typical solitaryabscesses as with typhoid. Diffuse interstitial inflammation may be encoun-tered in leptospirosis, diphtheria, scrub typhus, tuberculosis and leprosy [1].It is unusual for these forms of renal invasion to cause ARF, although a tran-sient oliguria and marginal retention of non-protein nitrogen is not uncom-mon. However, ARF may occur if the infection is associated with significant

Table 1. Principal causes of ARF in the Tropics com-pared to that in the developed world

South (%) North (%)

Surgical and post-traumatic1 10–15 65–70Obstetric2 15–25 1–3Infections 45–50 5–10Exogenous toxins 15–20 10–30

1Mostly urologic and abdominopelvic in the tropics,cardiopulmonary in the North [1, 2].

2Septic abortion, pregnancy toxemia [2–4].

Barsoum 46

hemodynamic effects as with septicemia, intravascular hemolysis, rhabdomyo-lysis or disseminated intravascular coagulation (DIC).

Viral interstitial nephropathies are more likely to cause ARF by direct inva-sion. Typical examples include Dengue fever and Hantan viral disease, both ofwhich can cause an acute hemorrhagic syndrome with reversible ARF. Directglomerular invasion by viral particles has been suggested to explain theglomerular lesions with HIV [5], HCV [6] and other infections.

Of the tropical parasitic infections Kala-azar may cause febrile ARF, par-ticularly in immunocompromized patients [7]. Leishmanial amastigotes tend toreside inside the interstitial monocytes and evade their phagocytic function; yetthe monocytes continue to release cytokines that mediate an acute interstitialnephritis.

Immune-Mediated Renal LesionsImmune complex-mediated acute post-infectious glomerulonephritis,

occasionally associated with transient mild to moderate renal insufficiency, hasbeen described with streptococcal infections of the skin in African kids asopposed to the pharynx in the developed world. Other tropical infections thatmay cause acute nephritis include typhoid, paratyphoid and leptospirosis.

Crescentic glomerulonephritis is more often associated with acute orrapidly progressive renal failure. This has been described with HIV, HAV,HCV, Mycoplasma pneumoniae lung infection, post-streptococcal glomeru-lonephritis, staphylococcal and other bacterial infections of the endocardium

0

10

20

30

40

50

60

70

80

Per

cent

of c

ases

Medical Obstetric Surgical

South Africa

UKUSAEgyptSingapore

ThailandIndiaGhana

Fig. 1. Main causes of ARF in selected tropical countries, compared to that in the USAand the UK.


and visceral abscess. It may occur with secondary syphilis and lepromatous lep-rosy [8]. Many of these conditions may be associated with type II or III mixedcryoglobulinemia.

Crescentic glomerulonephritis may also be pauci-immune. Patients withsevere streptococcal or falciparum malarial infection have evidence of systemicor renal limited pANCA-positive vasculitis, in which case necrotizing glomerularlesions are characteristic [9]. Similar lesions are occasionally seen in hepatitis Bviral infection [10], where ANCAs are not detected.

A distinct type of immune-mediated ARF is the postdysenteric hemolyticuremic syndrome, which is notoriously common in children in North India andBangladesh [11]. This syndrome is identical with the D� HUS seen with certainEscherichia coli infections in the West.

Hemodynamically Mediated ARFPyrexial illness in the warm and humid tropical climate leads to exces-

sive fluid loss with perspiration and sweating. Many infections are associatedwith vomiting and/or diarrhea, which add to the dehydration. Shrinkage ofblood volume becomes particularly dangerous in kids. Although renal failurein these cases is usually pre-renal, the presence of infective toxemia (e.g. incholera) and renal interstitial lesions often add up leading to more seriousconsequences. Hyperpyrexia may nonspecifically lead to rhabdomyolysis,which augments the renal injury even in the absence of significant circula-tory impediment. Certain infections may cause an outrageous release of cat-echolamines (e.g. tetanus), thereby amplifying the consequences of renalischemia [12].

The acute systemic effects of disseminated bacterial infection may lead toperipheral blood pooling and capillary leakage with subsequent reduction inblood volume, renal ischemia and acute tubular necrosis. Several mediatorshave been incriminated in this scenario including nitric oxide, TNF-�, dilatoryprostaglandins, kinins and free oxygen radicles. Tubular injury may be inducedby intrinsic pigments as hemoglobin and bilirubin, as well as microbial toxinsor medications. ARF is further propagated by tubular obstruction by detachedviable tubular cells, necrotic cell debris and pigment casts.

Sludging of red cells and platelets is often encountered in the renal micro-circulation, due to the associated hemoconcentration. This factor is consider-ably augmented in falciparum malaria, owing to increased stickiness of the redcells and platelets. Increased mechanical fragility of the parasitized red cellsleads to intravascular hemolysis, which adds up to the pathogenesis of ARF inmalaria [9].

Many other infections may lead to intravascular hemolysis in patientswith G6PD deficiency, as viral hepatitis, typhoid and typhus. The same and

Barsoum 48

other infections may also lead to DIC, which further compromises the renalcirculation [1].

Iatrogenic Renal InjuryRenal injury in acute tropical infections may be attributed to treatment

rather than to the primary disease. Typical examples are different antibiotics,notoriously amphotericin B, aminoglycosides, cephalosporines and van-comycin. Nephrotoxicity is considerably augmented when these agents are usedin combination, and when diuretics or non-steroidal anti-inflammatory agentsare added.

Another interesting example of iatrogenic ARF is the vasculitic ARFoccurring with interferon treatment for HCV infection [13]. This complicationis often encountered with the first few injections, and is characterized by severesystemic effects including pyrexia, hypotension, leukopoenia and thrombocy-topenia. The renal lesion is dominated by interstitial edema, cellular infiltrationand variable degrees of tubular atrophy. Most patients respond to steroid treat-ment, which has lead to the inclusion of these agents in several interferon-basedprotocols in viral hepatitis.

ARF may also occur as a result of vaccination. Anaphylactic and serum-sickness types of immune response leading to ARF have been described withmany vaccines. A recent alarming report from Brazil incriminates vaccinationfor transmitting yellow fever, resulting in fatal hemorrhagic ARF [14].

Intoxication

ARF may be attributed to exposure to a long list of occupational or environ-mental toxins, which are beyond the scope of this review. Of specific epidemi-ological and clinical interest in the tropics is the renal injury resulting fromaccidental exposure or therapeutic use of certain poisons of animal or plant originas traditional medicines.

Toxins of Animal OriginThe prototype of these poisons is the bite of snakes, usually Viperidae or

Elapidae occurring mainly in West Africa, the Indian subcontinent, South-EastAsia, New Guinea and Latin America. ARF occurs in 5–80% of cases [15]being attributed to the induction of massive release of monokines, complementactivation, hemolysis and intravascular coagulation. Snake venom may alsolead to a widespread vasculitis which involves the glomerular capillaries.Specific antibody response may lead to immune complex formation and sec-ondary glomerular injury.


Less common are scorpion or jellyfish stings as well as spider or centipedebites, often encountered in southern America, North Africa and the MiddleEast. These may lead to ARF in 6–10% of cases by inducing hemodynamicsequelae similar to those of snakebites. Scorpion stings are associated withmassive catecholamines release, which significantly contributes to renalischemia. Insect stings may also lead to ARF particularly in sub-Saharan Africa.

Oral ingestion of raw carp bile is a fairly common habit in South East Asia,Taiwan, China and Korea. The stuff is traditionally believed to improve visualacuity, stop cough, decrease body temperature and lower blood pressure.Ingestion of large quantities may lead to gastrointestinal toxicity that may beassociated with ARF in up to 54%, hematuria in 77% and jaundice in 62% ofpatients [16]. The mechanism of ARF seems to be multifactorial, involvingrenal ischemia, tubular toxicity and bile pigment nephropathy.

Toxins of Plant OriginMushroom poisoning is the best known of this group. Accidental ingestion

of certain wild mushrooms (Amanita or Galerina species) leads to gastroin-testinal and hepatic toxicity. ARF occurs 10–14 h after exposure. The latter isusually severe, with up to 50% mortality.

People in Indonesia, Malaysia and southern Thailand frequently useDjenkol beans as digestives. Toxicity follows ingestion of the raw bean in largeamounts. ARF occurs in about 50% of patients [17].

Hair dyes have been incriminated in the pathogenesis of ARF in Turkey,Northern Africa and the Sudan. Accidental toxicity from Henna, a traditionalcosmetic material used in the Middle East, has been described in the Turkey,being attributed to absorption through the skin. Suicidal ingestion of a similarcosmetic compound, paraphenylenediamine (PPD) was reported to induce asystemic inflammatory response, with respiratory distress and ARF in Morocco[18]. Combination of PPD and Henna seems to be even more toxic, with areported mortality of 40% in the Sudan [19].

Other poisons of plant origin, frequently used as traditional medicines forthe relief of cough, colics, impotence and many other common conditions maylead to ARF [20]. These include Callilepis laureola, Semecarpus anacardium,Securidaca longipedunculata, Euphorbia matabelensis and Crotalaria laburni-folia. The renal damage associated with these poisons is attributed to directnephrotoxicity as well as to the non-specific effects of gastrointestinal andhepatic toxicity.

Intoxication with Industrial ChemicalsAccidental exposure to many industrial chemicals may induce ARF. Such

accidents are particularly common in tropical countries owing to the poor control

Barsoum 50

over the manufacture process, storage and waste disposal. Major accidents havebeen reported with copper sulphate in India and Bangladesh, naphthalene inNigeria and many others.

Special Clinical Considerations in Tropical ARF

The clinical profile of ARF in the tropics is modified by a number of fac-tors. Most significant is the overlap between the manifestations of ARF andthose of severe systemic infection or intoxication. In many cases, there is severetoxemia, shock, dehydration, DIC, hepatocellular or pulmonary injury, centralnervous system involvement, etc. These may impose difficulties in the selectionof certain treatment modalities as hemodialysis in the presence of shock or ableeding tendency.

The complexity of pathogenetic factors involved such as hemolysis, rhab-domyolysis and DIC often leads to aggravation of the uremic syndrome. Mostpatients reaching the hospital are already severely anemic, thrombocytopenic,severely acidotic and critically hyperkalemic. Coma, convulsions, pericarditis,ascending paralysis, urea frost and similar clinical signs that are now seldomseen in the West, are an everyday experience in tropical ARF. Sky high serumcreatinine concentrations and incredible serum potassium levels of 9 or 10 mEq/lassociated with all the classical ECG signs may be seen when there is significantrhabdomyolysis.

All that usually occurs in a patient with background chronic disease ofvariable severity. Many natives in the tropics would be living with multiple chronicparasitic infestations and significant protein-calorie malnutrition. The addedcatabolic effects of the new infection or intoxication, and subsequently that ofthe uremic state result in a considerable negative nitrogen balance that has amajor impact on outcome.

Finally, there is the effect of delayed referral. Lack of adequate informationin the primary care units, patients’ reluctance to leave their own territories to betreated in central hospitals, inefficient and unsafe transfer vehicles may lead toa very high mortality from ARF that has been estimated in a consensus meetingto approach 80%. This is in sharp contrast to the reported mortality from referralhospitals that ranges between 20 and 30% [21].

Management of Tropical ARF

As with conventional ARF, the most effective intervention is preven-tion. Awareness of the potential renal complications of injury, surgery, obstetric


complication, infections and medications is the most cost-effective single mea-sure. Unfortunately, very little attention is paid by different medical specialitiesto this important issue.

Timely intervention at the onset of ARF is usually rewarded by excellentresults. Adequate attention to a primary infection, prompt control of hyperpyrexia,correction of shock, adequate hydration, avoidance of nephrotoxic drugs or drugcombinations are simple but most effective. There is no evidence that inductionof diuresis by osmotic or loop diuretics would change the outcome, nor does theuse of dopamine, calcium channel blockers or ACE inhibitors.

Adequate care of nutrition is of vital importance. Patients who are unableto eat or drink must be supported by enough fluids, electrolytes, calories andamino acids parenterally. This may be difficult with severe oliguria, when thetotal permissible fluid intake is restricted. An undoubted blessing, in thisrespect, is that patients treated in the tropical climate may lose a lot of fluid insweat and insensible routes, which provides more room for intravenous fluidadministration.

Taking the multifactorial nature of ARF into consideration, dialysis is oftenneeded as soon as the patient has reached the hospital. Fortunately, facilities forperitoneal dialysis, acute hemodialysis and even continuous renal replacementtherapy are now available in most tropical countries, though critically restrictedto teaching or army hospitals. In addition to logistic difficulties in transfer andacceptance of patients in these hospitals, there may be technical problems isimplementing dialysis and other extracorporeal therapies due to vascular accessproblems, bleeding tendencies, circulatory instability etc. It is hoped that the useof peritoneal dialysis finds its way to smaller territorial hospitals where promptrenal replacement therapy can be implemented when needed.

The challenge facing the medical profession will continue for many decades.Patients will continue to die of ARF until the value of proper education of theprimary care physician is appreciated and implemented. Unfortunately, thissimple fact is lost within the jungle of politics, inappropriate diversion of funds,and inequity of health care.

References

1 Barsoum R, Sitprija V: Tropical nephrology; in Schrier RW, Gottaschalk CW (eds): Diseases ofthe Kidney, ed 4. Boston, Little, Brown, 1996, pp 2221–2286.

2 Kleinknecht D: Epidemiology of acute renal failure; in Cantarovich F, Rangoonwala B, Verho M(eds): Progress in Acute Renal Failure. New Jersey, Marion Roussel, 1998, pp 11–22.

3 Firmat J, Zucchini A, Martin R, Aguirre C: A study of 500 cases of acute renal failure (1978–1991).Ren Fail 1994;16:91–99.

4 Randeree IG, Czarnocki A, Moodley J, Seedat YK, Naiker IP: Acute renal failure in pregnancy inSouth Africa. Ren Fail 1995;17:147–153.

Barsoum 52

5 Cohen AH: HIV-associated nephropathy: Current concepts. Nephrol Dial Transplant 1998;13:540–542.

6 Horikoshi S, Okada T, Shirato I, et al: Diffuse proliferative glomerulonephritis with hepatitis Cvirus-like particles in paramesangial dense deposits in a patient with chronic hepatitis C virus hepati-tis. Nephron 1993;64:462–464.

7 Caravaca F, Munoz A, et al: Acute renal failure in visceral leishmaniasis. Am J Nephrol 1991;11:350–352.

8 Barsoum R: The Kidney in tropical infections; in El-Nahas AM, Harris K, Anderson S (eds):Mechanisms and Clinical Management of Chronic Renal Failure, ed 2. Oxford, Oxford UniversityPress, 2000, pp 371–400.

9 Barsoum RS: Malarial acute renal failure. J Am Soc Nephrol 2000;11:2147–2154.10 Lai KN, Lai FM, Chan KW, et al: The clinicopathologic features of hepatitis B virus-associated

glomerulonephritis. Q J Med 1987;63:323–333.11 Srivastava RN, Moudgil A, Bagga A, Vasudev AS: Hemolytic uremic syndrome in children in

northern India. Pediatr Nephrol 1991;5:248–288.12 Daher EF, Abdulkader RC, Motti E, et al: Prospective study of tetanus-induced acute renal

dysfunction: Role of adrenergic overactivity. Am J Trop Med Hyg 1997;57:610–614.13 Dimitrov Y, Heibel F, Marcellin L, et al: Acute renal failure and nephrotic syndrome with alpha

interferon therapy. Nephrol Dial Transplant 1997;12:200–203.14 Vasconcelos PF, Luna EJ, Galler R, et al: Serious adverse events associated with yellow fever

17DD vaccine in Brazil: A report of two cases. Lancet 2001;358:91–97.15 Warrell DA: Venomous bites and stings in the tropical world. Med J Aust 1993;159:773–779.16 Park SK, Kim DG, Kang SK, et al: Toxic acute renal failure and hepatitis after ingestion of raw

carp bile. Nephron 1990;56:188–193.17 Eiam-ong S, Sitprija V, Saetang P, et al: Djenkol bean nephrotoxicity in Southern Thailand. Proc

First Asia Pacific Congress on Animal, Plant and Microbial Toxins, Singapore, 1989, p 628.18 Bourquia A, Jabrane AJ, Ramdani B, Zaid D: Systemic toxicity of paraphenylenediamine. 4 cases.

Presse Med 1988;17:1798–1800. 19 Devecioglu C, Katar S, Dogru O, Tas MA: Henna-induced hemolytic anemia and acute renal fail-

ure. Turk J Pediatr 2001;43:65–66.20 Kadiri S, Arije A, Salako BL: Traditional herbal preparations and acute renal failure in south west

Nigeria. Trop Doct 1999;29:244–246. 21 Barsoum RS: Dialysis in developing countries; in Jacobs C, Kjellstrand CM, Koch KM,

Winchester JF (eds): Replacement of Renal Function by Dialysis, ed 4. Dordtrecht, Kluwer, 1996,pp 1433–1445.

Professor Rashad S. BarsoumThe Cairo Kidney Center, 3, Hussein El-Memar Street, Antik-khana PO Box 91, Bab-El-Louk, 11513, Cairo (Egypt)Tel. �20 2 5761673, Fax �20 2 5769749, E-Mail [email protected]


Mechanisms Underlying CombinedAcute Renal Failure and Acute LungInjury in the Intensive Care Unit

Chu-Chun Chien, Landon S. King, Hamid Rabb

Department of Medicine, Johns Hopkins University, Baltimore, Md., USA

Acute renal failure (ARF) is associated with high morbidity and mortalityin intensive care unit patients [1]. It is increasingly recognized that ARF is anindependent risk factor for death, and that ARF is a systemic disease. With theavailability of dialysis, much of the mortality associated with ARF occurs dueto extrarenal complications such as circulatory collapse, cardiopulmonary fail-ure, and gastrointestinal bleeding [2, 3]. Lung dysfunction is particularly com-mon, and astute clinicians have long observed that there is a predisposition tocombined lung and kidney dysfunction in the critically ill patient. When ARFand acute lung injury (ALI) are combined, the mortality rate exceeds 80% [4].We will briefly review the mechanisms underlying combined ARF and ALI(fig. 1). Classic renal-pulmonary syndromes such as Goodpasture’s disease,Wegeners and other autoimmune diseases, though important considerations in thedifferential diagnosis of a patient with combined ARF and ALI [5, 6], will not bediscussed in this chapter. In addition, the pulmonary consequences of dialysis willnot be discussed [7, 8].

Pulmonary Effects of Acute Renal Failure

ARF leads to fluid retention, which in turn contributes to venous conges-tion and impaired pulmonary gas exchange [9]. Acidosis stimulates compen-satory hyperventilation and increases the work of breathing, potentiallyaccelerating respiratory muscle fatigue [10]. However, the impact of renal fail-ure on lung function extends beyond simply fluid overload and its sequelae.

The Critically Ill Patients: Pathological Mechanisms

Chien/King/Rabb 54

The uremic milieu likely contributes to the development or exacerbation oflung injury and dysfunction [11], and this can be improved with renal transplan-tation [12, 13]. A marked decrease in lung diffusion capacity in uremic patientshad been noted for decades [14–16], and this abnormality correlated well with theseverity of renal impairment. However, many have argued that the ‘uremic lung’was merely a consequence of volume overload, and that aggressive fluid removalwith dialysis resolves the pulmonary changes. The important clinical associationbetween combined ARF and ALI, as well as controversy in this area has led tostudies to directly test causal links between ARF and ALI (table 1).

Increased pulmonary vascular permeability is the hallmark of non-cardiogenic pulmonary edema. It was hypothesized that ARF would directly leadto an increase in pulmonary vascular permeability, and hence predispose to adultrespiratory distress syndrome (ARDS). Rats undergoing renal ischemia/reperfu-sion injury (IRI) demonstrated a significant increase in pulmonary vascular per-meability to albumin compared to sham-operated animals at 24 h that peaked48 h after reperfusion [17]. Lungs from rats subjected to renal IRI had a markedincrease in red blood cells in the interstitium with accompanying edema, as wellas alveolar hemorrhage and sludging of red blood cells in the microvasculature.

Fig. 1. Interaction between injured kidney and lung. ARF leads to an increase in pul-monary vascular permeability and inflammation during ARF, which is in part, macrophage-mediated. Lung epithelial sodium channel, aquaporin-5 and Na-K-ATPase are downregulatedduring ARF, which may be mediated by cytokines or uremic toxins. Conversely, mechanicalventilation leads to a decrease in renal epithelial sodium channel, aquaporin-2 and Na-K-ATPase. Potential mediators of these changes include hemodynamic factors, cytokines andleukocytes. Modified from Rabb H, Chamoun F, Hotchkiss J: Molecular mechanisms under-lying combined kidney-lung dysfunction during acute renal failure; in Ronco C, Bellomo R,La Grecca G (eds): Blood Purification in Intensive Care. Basel, Karger, 2001.

Acute lung injury

Acute renal failure

MediatorsHemodynamicsCytokines?Leukocytes?

MediatorsMacrophagesCytokines?Uremic toxins?

ConsequencesInflammation and increased vascular permeabilityPulmonary edemaAbnormalities in ENaC, AQP5, Na-K-ATPase

Consequences Increased vascular perm Proteinuria Abnormalities in ENaC, AQP2, Na-K-ATPase Inflammation Epithelial injury and apoptosis

Acute Renal Failure and Lung Injury in the Critically Ill Patient 55

Although the literature suggested that neutrophils participate in the pathogene-sis of ARDS, neutrophil depletion did not alter the effect of ARF on lung func-tion. Macrophages were targeted with CNI-1493, an agent that inhibitsmacrophage function by blocking both cytokine-inducible arginine uptake andthe p38 mitogen-activated protein (MAP) kinase signal transduction pathway.Rats in ARF receiving CNI-1493 displayed significant protection, with a reduc-tion in the increase in pulmonary vascular permeability following IRI. The CN1-1493 treated animals demonstrated interstitial edema, but alveolar hemorrhagewas absent. No effect on serum creatinine was observed.

In another study, bilateral ischemic renal injury or bilateral nephrectomy,but not unilateral ischemia-reperfusion, led to down-regulation of Na-K-ATPase and aquaporin-5 (AQP5) levels in the lung [11]. Decreased pulmonaryabundance of the epithelial sodium channel (ENaC) was also observed in thissetting, and was associated with renal injury. The clinical relevance ofdecreased membrane transporters in the distal lung is supported by the obser-vation of augmented pulmonary edema with amiloride or ouabain administra-tion in a pulmonary IRI model [18].

Recent data indicate that lung injury occurs rapidly following renalischemia. Mice with ischemic ARF had evidence of marked pulmonary edemaand neutrophil accumulation at 4 and 8 h after ischemia, associated with an

Table 1. Effects of experimental ARF on lung

Renal injury model Animal Lung changes Reference species No.

30 min IRI or bilateral rat downregulated pulmonary 11nephrectomy ENaC, Na-K-ATPase, AQP5

30 min IRI rat increased pulmonary vascular 17permeability, interstitial edema and alveolar inflammation, partially macrophage-mediated

40 min IRI mouse increased wet to dry weight 19ratio and upregulation of signaling process�-MSH inhibited many effects

30 min IRI mouse increased expression of 20adhesion molecules and neutrophil infiltration

IRI � Ischemia-reperfusion injury; ENaC � epithelial sodium channel; AQP5 �aquaporin-5; �-MSH � �-melanocyte-stimulating hormone.

Chien/King/Rabb 56

upregulation of proinflammatory signaling molecules in the lung. Treatmentwith �-melanocyte-stimulating hormone (�-MSH) attenuated both the ARFand associated lung changes in this model [19]. The increase in pulmonary neu-trophils during ARF is likely modulated in part by upregulation of leukocyteadhesion molecules [20].

The factors that link ARF and ALI are unknown. Circulating cytokines havebeen implicated in the pathogenesis of lung injury following distant organ IRI[21–23]. Macrophage-derived proinflammatory cytokines including tumor necro-sis factor-� (TNF-�) [24–26], interleukin-1 (IL-1) [27, 28], interleukin-6 (IL-6)[29] have been implicated. The type II alveolar epithelial cells might be the tar-get of these macrophage-derived inflammatory mediators [30]. Remote effects ofreperfusion injury have also been attributed to complement activation [31–33],and arachidonate derivatives such as thromboxane [34, 35] and leukotriene B4[36, 37]. In view of recent data implicating T cells in ARF [38, 39], the T cell isan attractive circulating candidate to mediate the ARF-ALI association.

In an effort to reduce the systemic inflammatory responses that can con-tribute to the high mortality associated with ARF, continuous renal replacementtherapy is being explored as a therapeutic intervention beyond the conventionaluse for renal replacement [40, 41]. Innovative approaches with plasmapheresisare also promising for ARF [42].

Renal Effects of Mechanical Ventilation

Patients requiring mechanical ventilation often have renal dysfunction thatmay be a consequence of the underlying illness. However, it is increasingly rec-ognized that mechanical ventilation itself can damage the kidney. Several clin-ical reports and experimental studies have shown that mechanical ventilationhas significant influence on renal function.

Positive-pressure ventilation alters venous return, cardiac preload, pul-monary vascular resistance, and cardiac afterload. A decrease in several param-eters of renal function, including glomerular filtration rate (GFR), renal bloodflow (RBF), and free-water clearance were noted using positive pressure venti-lation [43]. The interaction between mechanical ventilation and renal functionhas been extensively examined in the setting of positive end expiratory pressure(PEEP). In an anesthetized canine model, institution of 10 cm of PEEPmarkedly decreased urine flow rate, urinary sodium excretion, and creatinineclearance [44]. Prompt deterioration in renal function following the initiation ofPEEP [45], continuous positive airway pressure (CPAP) [46] and continuouspositive-pressure ventilation (CPPV) [47] has been observed repeatedly in var-ious animal models. Both animal and human studies indicate that positive


intrathoracic pressure might negatively affect renal function by decreasingvenous return and compromising cardiac output [48]. This effect may be wors-ened by concurrent volume depletion [49], consistent with the observation thatintravascular volume expansion may decrease the injury; however, the effects ofpositive intrathoracic pressure on RBF have not been consistent.

A comparison of intermittent mechanical ventilation (IMV) and continu-ous mechanical ventilation (CMV) in anesthetized canines with normal lungfunction indicated that IMV may have lesser negative influence on renalfunction [50], potentially a consequence of decreased mean intra-thoracic pres-sure that limits the retention of water and salt that occurs during prolongedmechanical ventilation [50, 51].

Positive pressure ventilation may change the expression of several media-tors central to renal function. Mechanical ventilation reduces production of renaldopamine, potentially an independent factor mediating renal injury. Exami-nation of the effects of IMV in awake humans with healthy lungs revealed adecrease in urine output and an increase in antidiuretic hormone (ADH) follow-ing institution of mechanical ventilation, without a change in free-water clear-ance [52]. PEEP has also been shown to increase antidiuretic hormone andaldosterone levels in some [53], but not all [54], animal studies, and may con-tribute to the decreasing urine output observed after institution of this modality.In some studies this effect can be ameliorated by adequate fluid resuscitation[55–57]. The rapid restoration of diuresis and natriuresis suggests the change ofantidiuretic hormone in this setting may be a consequence of a decreased atrialtransmural filling pressure caused by a positive intrathoracic pressure [58]. Thistopic has been comprehensively reviewed recently [59].

Ischemia remains the major cause of ARF in hospitalized patients, andrecent studies suggest that renal injury is also regulated by circulating cytokinessuch as TNF-� [60–62] and growth factors such as hepatocyte growth factor(HGF) [63, 64] produced by extrarenal organs. Mechanical ventilation, espe-cially with prior lung injury, produces similar changes in inflammatory mole-cules in kidney, such as TNF-�, IL-6 [29, 65], and vascular endothelial growthfactor (VEGF) [29, 65] (table 2). In a mouse model of 4 h of mechanicalventilation, marked decreases in renal ENaC, aquaporin-2 (AQP2) and Na-K-ATPase were found – similar to what occurs after direct ischemic injury tokidneys.

Recent experiments have demonstrated that mechanical ventilation with-out change in blood pressure or central venous pressure caused flattening ofepithelial cells in the canine kidney [66]. In a rabbit model of injurious mechan-ical ventilation, renal tubular apoptosis as well as biochemical markers of renaldysfunction were found, supporting a causal association between mechanicalventilation and the development of remote renal injury [67]. High tidal volume

Chien/King/Rabb 58

ventilation in rats produced a higher percentage of collapse of Bowman’s space[68] and proteinuria resulted from renal microvascular leak [69].

Conclusion

The persistent high mortality associated with ARF requires nephrologists tobroaden mechanistic studies beyond the pathophysiology of isolated ARF andmetabolites of dialysis approach, to the more complex arena of inter-organ fail-ure. Advances in our understanding of inflammation and lung mechanics mayprovide a basis for investigation of these critically ill patients with ARF that will

Table 2. Effects of experimental acute lung injury on kidney

Lung injury model Animal Renal changes Reference species No.

4 h ventilation � HCl mouse ventilation alone increases VEGF 29acid aspiration plus high VT

ventilation increases IL-6, VEGFR-2

5 h ventilation mouse increased neutrophil infiltration 65increased Il-1, ICAM-1, and IL-6decreased AQP2, ENaCincreased AQP1 in high VT

ventilation

6 h ventilation � LPS dog neutrophil margination in the 66and HCl peritubular capillaries

tubular cell sloughing

4 and 8 h ventilation � rabbit increased epithelial apoptosis index 67HCl

1 h ventilation rat high VT group with more 68glomerular collapse and perivascular edema

2 h ventilation rat high VT group with renal vascular 69leak and proteinuria L-NAME attenuated this effect

HCl � Hydroxyl chloride; VEGF � vascular epithelial growth factor; IL-6 � interleukin-6;VEGFR-2 � vascular epithelial growth factor receptor 2; IL-1 � interleukin-1; ICAM-1 �intracellular adhesion molecule 1; AQP2 � aquaporin-2; ENaC � epithelial sodium channel;AQP1 � aquaporin-1; LPS � lipopolysaccharide; VT � tidal volume; L-NAME � N-nitro-L-arginine methyl ester.


lead to new treatment strategies. It is likely that an important developmental linkbetween lung and kidney also contributes to this problem. Integrative physio-logic approaches on the bench and prospective cross-specialty clinical trials willbe required to make the necessary inroads in this lethal syndrome.

Acknowledgements

CC was supported by a training grant from Chang-Gung Memorial Hospital, Taipei,Taiwan. HR was supported by NIDDK R01 DK54770 and ROTRF 696313559. LK was sup-ported by NHLBI R01 HL70217 and American Heart Association Grant Aid. Both HR andLK are supported by NHLBI SCCOR grant HL073944.

References

1 Vivino G, Antonelli M, Moro ML, Cottini F, Conti G, Bufi M, Cannata F, Gasparetto A: Risk fac-tors for acute renal failure in trauma patients. Intens Care Med 1998;24:808–814.

2 Chertow GM, Lazarus JM, Paganini EP, Allgren RL, Lafayette RA, Sayegh MH: Predictors ofmortality and the provision of dialysis in patients with acute tubular necrosis. The AuriculinAnaritide Acute Renal Failure Study Group. J Am Soc Nephrol 1998;9:692–698.

3 Chertow GM, Levy EM, Hammermeister KE, Grover F, Daley J: Independent association betweenacute renal failure and mortality following cardiac surgery. Am J Med 1998;104:343–348.

4 Mehta RL, Pascual MT, Gruta CG, Zhuang S, Chertow GM: Refining predictive models in criti-cally ill patients with acute renal failure. J Am Soc Nephrol 2002;13:1350–1357.

5 Falk RJ, Jennette JC: ANCA small-vessel vasculitis. J Am Soc Nephrol 1997;8:314–322.6 von Vigier RO, Trummler SA, Laux-End R, Sauvain MJ, Truttmann AC, Bianchetti MG: Pulmonary

renal syndrome in childhood: A report of twenty-one cases and a review of the literature. PediatrPulmonol 2000;29:382–388.

7 Grassi V, Malerba M, Boni E, Tantucci C, Sorbini CA: Uremic lung. Contrib Nephrol. Basel,Karger, 1994, vol 106, pp 36–42.

8 Prezant DJ: Effect of uremia and its treatment on pulmonary function. Lung 1990;168:1–14.9 Mehta RL, Clark WC, Schetz M: Techniques for assessing and achieving fluid balance in acute

renal failure. Curr Opin Crit Care 2002;8:535–543.10 Tarasuik A, Heimer D, Bark H: Effect of chronic renal failure on skeletal and diaphragmatic mus-

cle contraction. Am Rev Respir Dis 1992;146:1383–1388.11 Rabb H, Wang Z, Nemoto T, Hotchkiss J, Yokota N, Soleimani M: Acute renal failure leads to dys-

regulation of lung salt and water channels. Kidney Int 2003;63:600–606.12 Kalender B, Erk M, Pekpak MA, Apaydin S, Ataman R, Serdengecti K, Sariyar M, Erek E: The

effect of renal transplantation on pulmonary function. Nephron 2002;90:72–77.13 Chan CH, Lai CK, Li PK, Leung CB, Ho AS, Lai KN: Effect of renal transplantation on pul-

monary function in patients with end-stage renal failure. Am J Nephrol 1996;16:144–148.14 Lee HY, Stretton TB, Barnes AM: The lungs in renal failure. Thorax 1975;30:46–53.15 Herrero JA, Alvarez-Sala JL, Coronel F, Moratilla C, Gamez C, Sanchez-Alarcos JM, Barrientos A:

Pulmonary diffusing capacity in chronic dialysis patients. Respir Med 2002;96:487–492.16 Moinard J, Guenard H: Membrane diffusion of the lungs in patients with chronic renal failure. Eur

Respir J 1993;6:225–230.17 Kramer AA, Postler G, Salhab KF, Mendez C, Carey LC, Rabb H: Renal ischemia/reperfusion

leads to macrophage-mediated increase in pulmonary vascular permeability. Kidney Int 1999;55:2362–2367.

Chien/King/Rabb 60

18 Khimenko PL, Barnard JW, Moore TM, Wilson PS, Ballard ST, Taylor AE: Vascular permeabilityand epithelial transport effects on lung edema formation in ischemia and reperfusion. J Appl Physiol1994;77:1116–1121.

19 Deng J, Hu X, Yuen PS, Star RA: �-Melanocyte-stimulating hormone inhibits lung injury afterrenal ischemia-reperfusion. Am J Respir Crit Care Med 2004 Jan 7 [Epub ahead of print].

20 Boydstun I, Lam VD, Hawkins EP, Welty SE: Adhesion molecule induction in lung following renalischemia reperfusion injury. J Am Soc Nephrol 1997;8:584A.

21 Welborn MB III, Moldawer LL, Seeger JM, Minter RM, Huber TS: Role of endogenous inter-leukin-10 in local and distant organ injury after visceral ischemia-reperfusion. Shock 2003;20:35–40.

22 Koike K, Moore EE, Moore FA, Franciose RJ, Fontes B, Kim FJ: CD11b blockade prevents lunginjury despite neutrophil priming after gut ischemia/reperfusion. J Trauma 1995;39:23–27.

23 Moore FA: The role of the gastrointestinal tract in postinjury multiple organ failure. Am J Surg1999;178:449–453.

24 Rabinovici R, Zhang D, Su Y, Luo X, Zhao Q, Yang JH: MOB-1 and TNF-alpha interact to inducemicrovascular lung injury. Shock 2002;18:261–264.

25 McCarter MD, Mack VE, Daly JM, Naama HA, Calvano SE: Trauma-induced alterations inmacrophage function. Surgery 1998;123:96–101.

26 Beierle EA, Vauthey JN, Moldawer LL, Copeland EM: Hepatic tumor necrosis factor-alpha pro-duction and distant organ dysfunction in a murine model of obstructive jaundice. Am J Surg 1996;171:202–206.

27 Shimabukuro DW, Sawa T, Gropper MA: Injury and repair in lung and airways. Crit Care Med2003;31:S524–S531.

28 Welk B, Malloy JL, Joseph M, Yao LJ, Veldhuizen AW: Surfactant treatment for ventilation-inducedlung injury in rats: Effects on lung compliance and cytokines. Exp Lung Res 2001;27:505–520.

29 Gurkan OU, O’Donnell C, Brower R, Ruckdeschel E, Becker PM: Differential effects of mechan-ical ventilatory strategy on lung injury and systemic organ inflammation in mice. Am J Physiol2003;285:L710–L718.

30 Crestani B, Aubier M: Inflammatory role of alveolar epithelial cells. Kidney Int Suppl 1998;65:S88-S93.

31 Guo RF, Lentsch AB, Sarma JV, Sun L, Riedemann NC, McClintock SD, McGuire SR, Van RooijenN, Ward PA: Activator protein-1 activation in acute lung injury. Am J Pathol 2002;161:275–282.

32 Wada K, Montalto MC, Stahl GL: Inhibition of complement C5 reduces local and remote organinjury after intestinal ischemia/reperfusion in the rat. Gastroenterology 2001;120:126–133.

33 Kuhnle GE, Kiefmann R, Sckell A, Kuebler WM, Groh J, Goetz AE: Leukocyte sequestration inpulmonary microvessels and lung injury following systemic complement activation in rabbits.J Vasc Res 1999;36:289–298.

34 Kukkonen S, Heikkila L, Verkkala K, Mattila S, Toivonen H: Thromboxane receptor blockadedoes not attenuate pulmonary pressure response in porcine single lung transplantation. J HeartLung Transplant 1996;15:409–414.

35 Turnage RH, Kadesky KM, Bartula L, Myers SI: Pulmonary thromboxane release followingintestinal reperfusion. J Surg Res 1995;58:552–557.

36 Fantini GA, Conte MS: Pulmonary failure following lower torso ischemia: Clinical evidence fora remote effect of reperfusion injury. Am Surg 1995;61:316–319.

37 Klausner JM, Paterson IS, Kobzik L, Valeri CR, Shepro D, Hechtman HB: Leukotrienes but notcomplement mediate limb ischemia-induced lung injury. Ann Surg 1989;209:462–470.

38 Rabb H, Daniels F, O’Donnell M, Haq M, Saba SR, Keane W, Tang WW: Pathophysiological roleof T lymphocytes in renal ischemia-reperfusion injury in mice. Am J Physiol 2000;279:F525–F531.

39 Burne MJ, Daniels F, El Ghandour A, Mauiyyedi S, Colvin RB, O’Donnell MP, Rabb H:Identification of the CD4(�) T cell as a major pathogenic factor in ischemic acute renal failure.J Clin Invest 2001;108:1283–1290.

40 Scheel P, Eustace J, Rabb H: Continuous dialysis as systemic therapy in the critically ill patient?Crit Care Med 2003;31:988–989.

41 Ronco C, Bellomo R, Ricci Z: Continuous renal replacement therapy in critically ill patients.Nephrol Dial Transplant 2001;16(suppl 5):67–72.


42 Stegmayr BG, Banga R, Berggren L, Norda R, Rydvall A, Vikerfors T: Plasma exchange as rescuetherapy in multiple organ failure including acute renal failure. Crit Care Med 2003;31:1730–1736.

43 Murdaugh HV Jr, Sieker HO, Manfredi F: Effect of altered intrathoracic pressure on renal hemo-dynamics, electrolyte excretion and water clearance. J Clin Invest 1959;38:834–842.

44 Hall SV, Johnson EE, Hedley-Whyte J: Renal hemodynamics and function with continuouspositive-pressure ventilation in dogs. Anesthesiology 1974;41:452–461.

45 Steinhoff HH, Samodelov LF, Trampisch HJ, Falke KJ: Cardiac afferents and the renal responseto positive pressure ventilation in the dog. Intens Care Med 1986;12:147–152.

46 Marquez JM, Douglas ME, Downs JB, Wu WH, Mantini EL, Kuck EJ, Calderwood HW: Renalfunction and cardiovascular responses during positive airway pressure. Anesthesiology 1979;50:393–398.

47 Fewell JE, Bond GC: Renal denervation eliminates the renal response to continuous positive-pressure ventilation. Proc Soc Exp Biol Med 1979;161:574–578.

48 Mitaka C, Nagura T, Sakanishi N, Tsunoda Y, Amaha K: Two-dimensional echocardiographic eval-uation of inferior vena cava, right ventricle, and left ventricle during positive-pressure ventilationwith varying levels of positive end-expiratory pressure. Crit Care Med 1989;17:205–210.

49 Shinozaki M, Muteki T, Kaku N, Tsuda H: Hemodynamic relationship between renal venous pres-sure and blood flow regulation during positive end-expiratory pressure. Crit Care Med 1988;16:144–147.

50 Steinhoff HH, Kohlhoff RJ, Falke KJ: Facilitation of renal function by intermittent mandatory ven-tilation. Intens Care Med 1984;10:59–65.

51 Baratz RA, Ingraham RC: Renal hemodynamics and antidiuretic hormone release associated withvolume regulation. Am J Physiol 1960;198:565–570.

52 Khambatta HJ, Baratz RA: IPPB, plasma ADH, and urine flow in conscious man. J Appl Physiol1972;33:362–364.

53 Kaczmarczyk G, Jorres D, Rossaint R, Krebs M, Unger V, Falke K: Extracellular volume expan-sion inhibits antidiuretic hormone increase during positive end-expiratory pressure in consciousdogs. Clin Sci (Lond) 1993;85:643–649.

54 Teba L, Dedhia HV, Schiebel FG, Blehschmidt NG, Lindner WJ: Positive-pressure ventilationwith positive end-expiratory pressure and atrial natriuretic peptide release. Crit Care Med 1990;18:831–835.

55 Costa KN, Carvalho WB, Kopelman BI, DiDo R: Dosage of atrial natriuretic peptide in pediatricpatients submitted to mechanical ventilation. Rev Assoc Med Bras 2000;46:320–324.

56 Rossaint R, Jorres D, Nienhaus M, Oduah K, Falke K, Kaczmarczyk G: Positive end-expiratorypressure reduces renal excretion without hormonal activation after volume expansion in dogs.Anesthesiology 1992;77:700–708.

57 Ramamoorthy C, Rooney MW, Dries DJ, Mathru M: Aggressive hydration during continuouspositive-pressure ventilation restores atrial transmural pressure, plasma atrial natriuretic peptideconcentrations, and renal function. Crit Care Med 1992;20:1014–1019.

58 Andrivet P, Adnot S, Sanker S, Chabrier PE, Macquin-Mavier I, Braquet P, Brun-Buisson C:Hormonal interactions and renal function during mechanical ventilation and ANF infusion inhumans. J Appl Physiol 1991;70:287–292.

59 Pannu N, Mehta RL: Mechanical ventilation and renal function: An area for concern? Am J KidneyDis 2002;39:616–624.

60 Donnahoo KK, Meng X, Ao L, Ayala A, Shames BD, Cain MP, Harken AH, Meldrum DR:Differential cellular immunolocalization of renal tumour necrosis factor-alpha production duringischaemia versus endotoxaemia. Immunology 2001;102:53–58.

61 Donnahoo KK, Meng X, Ayala A, Cain MP, Harken AH, Meldrum DR: Early kidney TNF-alphaexpression mediates neutrophil infiltration and injury after renal ischemia-reperfusion. Am JPhysiol 1999;277:R922–R929.

62 Meldrum KK, Meldrum DR, Meng X, Ao L, Harken AH: TNF-alpha-dependent bilateral renalinjury is induced by unilateral renal ischemia-reperfusion. Am J Physiol 2002;282:H540–H546.

63 Liu Y: Hepatocyte growth factor and the kidney. Curr Opin Nephrol Hypertens 2002;11:23–30.64 Matsumoto K, Mizuno S, Nakamura T: Hepatocyte growth factor in renal regeneration, renal dis-

ease and potential therapeutics. Curr Opin Nephrol Hypertens 2000;9:395–402.

Chien/King/Rabb 62

65 Tremblay LN, Miatto D, Hamid Q, Govindarajan A, Slutsky AS: Injurious ventilation induces wide-spread pulmonary epithelial expression of tumor necrosis factor-alpha and interleukin-6 messen-ger RNA. Crit Care Med 2002;30:1693–1700.

66 Chien CC, Haas M, Jakobe A, McVerry B, Simon BA, Rabb H: Mechanical ventilation associatedlung injury in canines causes renal histologic changes consistent with early acute tubular necro-sis. J Am Soc Nephrol 2003;14:354A.

67 King L, Doddo OJ, Becker P, Haas M, Chien CC, Burne-Taney MJ, Rabb H: Mechanical ventila-tion in mice leads to renal inflammation and dysregulation of transporters. J Am Soc Nephrol2003;14:564A.

68 Valenza F, Sibilla S, Porro GA, Brambilla A, Tredici S, Nicolini G, Miloso M, Tredici G, Gattinoni L:An improved in vivo rat model for the study of mechanical ventilatory support effects on organsdistal to the lung. Crit Care Med 2000;28:3697–3704.

69 Choi WI, Quinn DA, Park KM, Moufarrej RK, Jafari B, Syrkina O, Bonventre JV, Hales CA:Systemic microvascular leak in an in vivo rat model of ventilator-induced lung injury. Am J RespirCrit Care Med 2003;167:1627–1632.

H. Rabb, MDJohns Hopkins University HospitalRoss 970, 720 Rutland Ave, Baltimore, MD 21205 (USA)Tel. �1 410 502 1555, Fax �1 410 614 5129, E-Mail [email protected]


Cytokine Single NucleotidePolymorphismRole in Acute Renal Failure

Orfeas Liangos, Vaidyanathapuram S. Balakrishnan, Brian J. G. Pereira, Bertrand L. Jaber

Dialysis Research Laboratory, Division of Nephrology, Tufts-New England Medical Center, Boston, Mass., USA

Acute renal failure (ARF) is a serious complication and carries highmortality rate [1, 2], which has not improved despite advances in intensive careand dialytic support. The lack of improvement in patient outcomes calls fornovel preventive and therapeutic approaches.

Mediators of the Systemic Inflammatory Response Syndrome in ARF

Sepsis or the broader syndrome, the systemic inflammatory responsesyndrome (SIRS), represents an overwhelming, host response to a variety oftriggers such as gram-negative bacterial endotoxin [3]. During the initial phase ofsepsis, pro-inflammatory cytokines are released, including tumor necrosisfactor-� (TNF-�) and interleukin-1� (IL-1�), which in turn stimulate therelease of other biologically active molecules, resulting in vascular endothelialdamage and organ dysfunction. This initial phase is followed by a compen-satory anti-inflammatory response (CARS) in which anti-inflammatory cytokinesare produced, which offset pro-inflammatory responses [4]. Interleukin-10 isthe most potent anti-inflammatory cytokine, and its release inhibits productionof TNF-� and IL-1� [5–7].

Acute renal failure often develops during sepsis, and pro-inflammatorycytokines released in response to endotoxin, organ ischemia-reperfusion, comple-ment activation, and through other cytokines and immune modulators have been

Liangos/Balakrishnan/Pereira/Jaber 64

shown to play a role in the development of endotoxin-mediated ARF [8, 9] (fig. 1).Leukocyte activation with expression of adhesion molecules, and the productionof oxygen-free radicals, arachidonic acid metabolites, platelet-activating factor,nitric oxide, endothelins, and heat-shock proteins follows [10] whereby creatinga pro-inflammatory cascade that contributes to injury of the glomerular andperitubular vasculature [11]. TNF-receptor-1 deficient mice are resistant toendotoxin-induced ARF, arguing for a critical role for TNF-� [9]. In addition,bacterial products can activate neutrophils in situ and IL-8-mediated chemotaxisfurther recruits neutrophils to sites of inflammation [12]. Morphological studies ofacute tubular necrosis in humans have shown neutrophils in the vasa recta and theinterstitium and support a role for inflammation in the pathogenesis of ARF [13].

Ischemic and Nephrotoxic Injury in ARF

Increasing evidence supports an important role for inflammatory mecha-nisms in both ischemic [11, 14] and nephrotoxic [15, 16] renal injury. Animalmodels of ischemic ARF have demonstrated an activation of renal myeloperoxi-dase (MPO) and increased gene expression of intercellular adhesion molecule-1

Sepsis

Ischemia-reperfusion Complement activationEndotoxin release

Cellular activation

Kidney injury

Oxygen freeradicals

Nitric oxide

Endothelins

Proteases

Arachidonicacid metabolites

Platelet-activating factor

↑ Urinary markers(NAG, KIM-1… )

↑ Circulating markers(serum creatinine)

Genetic influence Genetic influenceTNF-� IL-10

�

�

Fig. 1. Schematic representation of the inflammatory response to sepsis and resultingkidney injury. NAG � N-acetyl-�-D-glucosaminidase; KIM � kidney injury molecule 1.Modified from Jaber et al. [63], with permission.

Cytokine Gene Polymorphism in Acute Renal Failure 65

(ICAM-1) and IL-6 [17]. A mouse model of cisplatin nephrotoxicity showsincreased TNF-� and IL-1� mRNA expression and increased urinary TNF-�[16]. Inhibitors of TNF-� such as pentoxifylline and anti-TNF-� monoclonalantibodies have been shown to ameliorate the pro-inflammatory response.Similarly, TNF-�-deficient mice are resistant to cisplatin nephrotoxicity, arguingfor a central role for TNF-� [16].

On the other hand, exogenous administration of IL-10 inhibits TNF-� andICAM-1 mRNA expression in mouse kidneys following ischemic and nephrotoxicinjury [15], and antibodies to ICAM-1 protect the kidney against ischemicinjury [18].

Cytokines and ARF in Humans

Increased plasma and urinary cytokine levels have been shown to correlatewith urinary levels of N-acetyl-�-D-glucosaminidase (NAG), a bio-marker ofrenal proximal tubular dysfunction following surgery with cardiopulmonarybypass (CPB) [19]. However, it is unclear as to whether urinary cytokines arefiltered by the glomerulus or produced in situ. More studies, perhaps involvingthe recently described soluble form of human kidney injury molecule-1 (KIM-1),a highly specific bio-marker for renal proximal tubular injury [11] may shedlight on a potential pathogenetic link between urinary cytokines and early tubu-lar injury.

In regards to plasma cytokine levels, IL-6 and IL-10 have been shown todifferentiate surviving from non-surviving patients with ARF [20]. In addition,elevated serum levels of soluble TNF receptors among patients with septicshock were shown to predict ARF and mortality [21].

The above data support the role of cytokines as important mediators ofkidney parenchymal injury, but the role of genetic factors affecting cytokine pro-duction in the inflammatory response cascade and their implications for ARFfollowing ischemic or nephrotoxic injury have remained largely unexplored.

Human Gene Polymorphism and Biological Diversity

Among humans, there is a remarkable homogeneity of genetic information,with two unrelated individuals sharing greater than 99.9 percent of their DNAsequences [22]. Variations observed in the remaining 0.1% of the humangenome have become the subject of intense investigation [23]. This variability,known as gene polymorphism, is the basis of biological diversity and somegenotypic variations have been shown to correlate with specific phenotypes


relevant to human disease [24]. It is not clear, however, whether many of thesegenetic variants are causal for the diseases in question or simply located in prox-imity to other, yet unknown pathogenic genetic factors, an occurrence known aslinkage disequilibrium.

Polymorphism of human genes occurs at one or more of the following sites(fig. 2): (1) the promoter or 5�-flanking region affecting transcriptional activity[25]; (2) the exon(s) or the gene coding sequences affecting gene expression orfunction through changes in structure, binding or trafficking of the gene product;(3) the intron(s) or the gene intervening sequences where changes may lead todefects in RNA and mRNA processing, and (4) the 3�-untranslated (3�-UTR)region further downstream which may affect gene expression through alteredRNA half-life or influencing ribosomal translation of mRNA. On a molecularlevel, gene polymorphism can be categorized into: (1) single nucleotide poly-morphism (SNP); (2) variable number of tandem repeats (VNTR) or minisatellitepolymorphism, and (3) microsatellite polymorphism [25], of which the singlenucleotide polymorphism is the most common and consists of single nucleotidesubstitution. SNP in the promoter region may change transcriptional activityand thereby be of functional relevance.

Identification of Genes Relevant to Human Disease

There are two approaches to identifying genes relevant to human disease:Linkage analyses and association studies. Linkage analyses identify co-inheritanceof a specific phenotype with a region of the genome in families. These analyses

Direction of transcription

5' 3'

Initiator codon Terminationcodon

Start oftranscription

Untranslatedregion

Promoterregion • Alteration in transcriptional activity

Introns (intervening sequences)• Defect in RNA processing • Defect in mRNA processing

• Alteration in mRNA half-life

• Alteration in ribosomal mRNA translation

Exons (coding sequences) • Silent • Alteration in gene expression • Alteration in gene function (structure, binding or trafficking of protein product

Fig. 2. Sketch of the structure of a human gene with potential sites of polymorphism.


are successful in identifying novel genes for ‘monogenic’ diseases, and envi-ronmental factors play a minimal role in disease expression. One such exampleis polycystic kidney disease.

Association studies identify susceptibility genes for common ‘polygenic’diseases. This is known as the ‘candidate gene’ approach and relies on biologicalplausibility to identify highly likely candidate genes. In this analytical approach,environmental factors play a critical role in disease expression. Examples wouldinclude the study of genes encoding for cytokines in sepsis.

Association Studies in Acute Infectious and Inflammatory Disorders

In recent years, the study of cytokine gene polymorphism has become thesubject of intense interest, as these genetic markers may be potential determi-nants of susceptibility to disease incidence, severity of illness and clinical out-comes. For inflammatory conditions, polymorphisms of the TNF and IL-10genes are of particular interest, particularly those associated with altered geneexpression. Associations between these polymorphisms and acute inflammatorystates are summarized (table 1).

Tumor Necrosis Factor Locus

The TNF-� gene is located on the short arm of chromosome 6. Polymor-phism within the promoter region of the TNF-� gene at positions �238 (G to A)and �308 (G to A) have been described. The �308 A-allele, also known as theTNF-�2 allele, has been associated with high promoter activity [26, 27] andenhanced TNF-� production [28].

The gene coding for TNF-�, also known as lymphotoxin-� (LT-�), lies nextto the TNF-� gene. Single nucleotide polymorphism at position �1069 (G to A)in the first intron of the TNF-� gene has been characterized [29]. The ‘A’ vari-ant is the TNF-�2 allele whereas the less frequent ‘G’ variant is known as theTNF-�1 allele. TNF-�2 homozygous individuals produce significantly higheramounts of TNF-� and IL-1� [30, 31]. There is strong linkage disequilibriumbetween the TNF-��308 (G to A) and TNF-� also called NcoI polymorphisms,whereby heterozygosity for both TNF polymorphisms is associated with higherTNF-� production [32]. This finding highlights the importance of linkage dis-equilibrium in studies examining the association between polymorphism of asingle candidate gene and severity of disease.

Polymorphism of the TNF-� gene has been studied extensively and hasbeen associated with adverse clinical outcomes among patients with sepsis


[33, 34] as well as patients with community-acquired pneumonia [35]. Otherstudies have linked TNF-�2 homozygous status with higher mortality of andincreased susceptibility to severe sepsis [36]. The TNF-� �250 AA-genotypehas been associated with a greater risk of septic shock in community-acquiredpneumonia [37]. Both the TNF-� �250 G and the TNF-� �308 G-alleles havealso been associated with prolonged mechanical ventilation following coronaryartery bypass graft (CABG) surgery [38]. The �308 A-allele (TNF�2) has alsobeen associated with an increased incidence of ARF among neonates [39] andwith higher TNF-� production and a higher risk of death among adult patientswith dialysis-requiring ARF (fig. 3) [40].

These data support the hypothesis that TNF-� gene polymorphisms linkedto increased production of pro-inflammatory cytokines may have important

Table 1. Association between polymorphism of TNF, IL-10 genes andsepsis or acute renal failure (modified from Jaber et al. [63] with permission)

Polymorphic allele Acute illness Reference

TNF-a�308 A-allele (TNF-�2) sepsis/septic shock [33, 34, 36]�308 A-allele (TNF-�2) meningococcal disease [64]�308 A-allele (TNF-�2) neonatal acute renal failure [39]�308 A-allele (TNF-�2) increased mortality in [40]

dialysis-requiring acute renal failure

�238 A-allele increased risk of death in [35]community-acquired pneumonia

TNF-b (LT-a)�250 AA genotype increased risk of septic shock [35, 37]

in community-acquired pneumonia

�1069 (NcO1) allele-2/2 septic shock following acute [65]biliary pancreatitis

�1069 (NcO1) allele-2/2 increased mortality in [66]severe sepsis

�1069 (NcO1) allele-2/2 susceptibility to severe [36]post-traumatic sepsis

IL-10�1082 GA genotype meningococcal disease [42, 67]�1082 G-allele severity of illness in [43]

community-acquired pneumonia

�592 A-allele mortality in sepsis [44]


implications in the severity of ARF, and, consequently, the requirement for dial-ysis and hospital mortality.

The Interleukin-10 Gene

The IL-10 gene is located on the long arm of chromosome 1. The IL-10promoter region is polymorphic with a single-base-pair substitution at position�1082 (G to A) [41]. In vitro studies have identified three phenotypic secretionlevels based on allelic substitutions: GG, GA and AA genotype, for high, inter-mediate and low producer status, respectively [27, 41]. Additional polymor-phisms at positions �819 (C to T) and �592 (C to A) have been described.Polymorphisms at these two sites are in linkage disequilibrium.

Interleukin-10 �1082 polymorphic alleles have been associated with suscep-tibility to meningococcal disease and adverse clinical outcomes [42], and severityof illness in community-acquired pneumonia [43]. In dialysis-requiring ARF, car-riers of the IL-10 �1082 G-allele had higher IL-10 production and lower risk ofdeath after adjustment for severity of illness scores, especially when combined tothe TNF-� low producer genotype (�308 A-allele carrier) (fig. 3) [40]. The IL-10�592 A-allele has been associated with higher mortality in sepsis [44].

Low TNF-�, high IL-10

Low TNF-�, low IL-10

High TNF-�, high IL-10

High TNF-�, low IL-10

1.0

0.8

0.6

0.4

0.2

0.00 10 20 30

Cum

ulat

ive

surv

ival

Time (days)

Fig. 3. Cumulative survival in a cohort of patients with dialysis requiring ARF stratifiedby the combination of TNF-� and IL-10 gene-polymorphism and adjusted for APACHE IIscore. Modified from Jaber et al. with permission [40].


Limitations of Gene Polymorphism-Association Studies

Polymorphism-association studies are typically designed in a case-controlformat. They frequently have a small sample size, and should be interpreted withcaution. This is particularly true for diseases that are common and polygenic innature and where a single candidate gene hypothesis lacks biological plausibility.Small, single center studies may also be limited in their external validity regard-ing ethnicity and geographical boundaries [45]. Survival bias is another concernin case control studies evaluating genetic markers as predictors of mortality.Proper matching for age and follow up time or using prospective cohort studydesigns can overcome these limitations [46].

The Future in ARF

Future developments in ARF will focus on development and establishmentof early bio-markers of kidney injury such as NAG and KIM-1. There is also aneed to develop non-invasive tools that help detect ARF-induced inflammation.The ultra-small super-paramagnetic iron oxide (USPIO) enhanced MRI is onesuch example. This technology can detect inflammation caused by experimentalcerebral ischemia, arthritis, nephrotoxic nephritis, and renal transplant rejection[47]. The USPIO particles are internalized by leukocytes, and once internalized,MRI signal intensity decreases on T2-weighted images. This extremely promisingtool awaits further studies in humans.

Novel risk assessment and stratification tools to help allocate resources arealso likely to be the focus of future research efforts, and cytokine genotype pro-files are a good example for such use. Finally, targeted therapies, including theuse of cytokine modulating strategies are another example of future interventionsin ARF. Indeed, due to their importance in inflammatory diseases, the modifi-cation of cytokine responses through therapeutic intervention remains a focusof intense investigation. A summary of modulation of two particular cytokines,TNF-� and IL-10, is briefly reviewed.

Down-Regulation of TNF-�

Glucocorticoids, pentoxifylline and thalidomide inhibit TNF-� synthesisat distinct points of its biosynthetic pathway. Glucocorticoids interfere with NF-�Bwhich in turn decreases expression of TNF-� activity [48], and reduces post-transcription protein translation [49]. The xanthine derivative pentoxifylline,reduces TNF-� mRNA expression in monocytes [50], which is partly mediated


by generation of cyclic adenosine monophosphate (cAMP) which inhibits NF-�B-mediated transcription [51]. Rolipram, a specific inhibitor of monocytepredominant phosphodiesterose inhibitor type-4, is a 500-fold more potentinhibitor of TNF-� synthesis in human monocytes, compared with pentoxifylline[52]. Thalidomide also selectively inhibits biosynthesis of TNF-� [53].Interleukin-10 reduces expression of TNF-� [54], and its administration tohealthy volunteers inhibits endotoxin-induced rise in body temperature and releaseof TNF-� [55]. Both monoclonal antibodies to TNF-� (Inflixmab, Adalilumab)as well as soluble TNF receptors (sTNFR) (Etanercept) can bind and neutralizethis cytokine. Finally, somatostatin is also capable of down-regulating cell-surfaceTNF receptor expression in human macrophages, which become less responsiveto TNF-� [56]. These therapeutic approaches of immune modulation in inflam-matory diseases could play a role in ameliorating the inflammatory componentin the pathophysiology of ARF.

Upregulation of IL-10

Cyclic AMP stimulates IL-10 gene transcription via activation of cAMPresponse element binding proteins [57]. The phosphodiesterase inhibitors theo-phylline and rolipram induce IL-10 expression [58, 59]. The xanthine derivativespentoxifylline and caffeine, however, decrease IL-10 whereas dexamethasoneincreases its release [60]. Calcineurin inhibitors also increase IL-10 mRNAexpression and synthesis [61]. Finally, glucocorticoids have been shown tomodulate the inflammatory response syndrome in favor of IL-10, resulting in adecrease in TNF-�, interleukin-6 (IL-6) and interleukin-8 (IL-8) levels and anincrease in IL-10 levels [62].

Summary and Conclusions

Although the pathogenesis of ARF is heterogeneous and results froma combination of different environmental influences and host responses, there isoverwhelming data to suggest a common pathway that involves pro- and anti-inflammatory molecules, which in turn, determines the extent of tissue injury.The number of recognized cytokine gene polymorphisms is growing daily. Thedevelopment of cytokine gene mapping may help identify patients in whom anexcessive systemic inflammatory response may follow a therapeutic interven-tion (e.g. CABG, contrast administration), and who may be at increased risk fordeveloping acute organ dysfunction. Through these advances, tools may bedeveloped to better understand, prevent and treat ARF.


Genetic epidemiology studies may help characterize the importance ofgenetic markers in the development of ARF. This would require large prospec-tive cohort studies aimed at examining associations between genetic markers,urinary (urine KIM-1 and NAG) and circulating (serum creatinine) markers ofkidney injury. Once firmly established, the association of a particular geneticprofile and outcome could be used to risk stratify patients for the developmentof ARF (fig. 4). Ultimately, cytokine-modulating therapies could be employedon the basis of genotypic risk stratification with the goal to prevent kidneyinjury or minimize its deleterious effects on patient outcome.

References

1 Hou SH, Bushinsky DA, Wish JB, Cohen JJ, Harrington JT: Hospital-acquired renal insufficiency:A prospective study. Am J Med 1983;74:243–248.

2 Chertow GM, Levy EM, Hammermeister KE, Grover F, Daley J: Independent association betweenacute renal failure and mortality following cardiac surgery. Am J Med 1998;104:343–348.

3 Conference ACoCPSoCCMC: Definitions of sepsis and organ failure and guidelines for the useof innovative therapies in sepsis. Crit Care Med 1992;13:818–822.

4 Opal S, DePalo V: Anti-inflammatory cytokines. Chest 2000;117:1162–1172.5 De Waal Malefyt F, Abrams J, Bennett B, Figdor CG, de Vries JE: Interleukin 10 (IL-10) inhibits

cytokine synthesis by human monocytes: An autoregulatory role of IL-10 produced by monocytes.J Exp Med 1991;174:1209–1220.

6 Fiorentino DF, Zlotnik A, Mosmann TR, Howard M, O’Garra A: IL-10 inhibits cytokine productionby activated macrophages. J Immunol 1991;147:3815–3822.

7 Perianayagam M, Jaber B, Guo D, King A, Pereira B, Balakrishnan V: Defective interleukin-10synthesis by peripheral blood mononuclear cells among hemodialysis patients. Blood Purif 2002;20:543–550.

8 Fouqueray B, Philippe C, Herbelin A, Perez J, Ardaillou R, Baud L: Cytokine formation within ratglomeruli during experimental endotoxemia. J Am Soc Nephrol 1993;3:1783–1791.

9 Cunningham P, Dyanov H, Park P, Wang J, Newell K, Quigg R: Acute renal failure in endotoxemiais caused by TNF acting directly on TNF receptor-1 in kidney. J Immunol 2002;168:5817–5823.

Spectrum ofenvironmentalpredisposition

High Low

Spectrum of geneticpredisposition

Low High

e.g. TNF-� -308, IL-10 -1082 SNP…

e.g. radiocontrast, sepsis, surgery…

Risk of acuterenal failure

Fig. 4. Acute renal failure risk stratification: A model for genetic and environmentalinteraction.


10 Dybdahl B, Wahba A, Lien E, Flo TH, Waage A, Qureshi N, Sellevold OF, Espevik T, Sundan A:Inflammatory response after open heart surgery: Release of heat-shock protein 70 and signalingthrough toll-like receptor-4. Circulation 2002;105:685–690.

11 Bonventre J, Weinberg J: Recent advances in the pathophysiology of ischemic acute renal failure.J Am Soc Nephrol 2003;14:2199–2210.

12 Zoja C, Angioletti S, Donadelli R, Zanchi C, Tomasoni S, Binda E, Imberti B, Te Loo M,Monnens L, Remuzzi G, Morigi M: Shiga toxin-2 triggers endothelial leukocyte adhesion andtransmigration via NF-kappaB dependent up-regulation of IL-8 and MCP-1. Kidney Int 2002;62:846–856.

13 Solez K, Morel-Maroger L, Sraer J-D: The morphology of acute tubular necrosis in man: Analysisof 57 renal biopsis and a comparison with glycerol model. Medicine 1979;58:362–376.

14 Burne M, Daniels F, El Ghandour A, Mauiyyedi S, Colvin R, O’Donnell M, Rabb H: Identificationof the CD4� T cell as a major pathogenic factor in ischemic acute renal failure. J Clin Invest2001;108:1283–1290.

15 Deng J, Kohda Y, Chiao H, Wang Y, Hu X, Hewitt SM, Miyaji T, McLeroy P, Nibhanupudy B, Li S,Star RA: Interleukin-10 inhibits ischemic and cisplatin-induced acute renal injury. Kidney Int2001;60:2118–2128.

16 Ramesh G, Reeves W: TNF-alpha mediates chemokine and cytokine expression and renal injuryin cisplatin nephrotoxicity. J Clin Invest. 2002;110:835–842.

17 Burne-Taney M, Kofler J, Yokota N, Weisfeldt M, Traystman R, Rabb H: Acute renal failure afterwhole body ischemia is characterized by inflammation and T cell-mediated injury. Am J Physiol2003;285:F87–F94.

18 Kelly KJ, Williams WWJ, Colvin RB, Bonventre JV: Antibody to inter-cellular adhesion molecule1 protects the kidney against ischemic injury. Proc Natl Acad Sci USA 1994;91:812–816.

19 Gormley SM, McBride WT, Armstrong MA, Young IS, McClean E, MacGowan SW, Campalani G,McMurray TJ: Plasma and urinary cytokine homeostasis and renal dysfunction during cardiacsurgery. Anesthesiology 2000;93:1210–1216;discussion 1215A.

20 Himmelfarb J, Sezer M, Shyr Y, Freedman S, Ikizler T, Becker K, Spratt D, Group PS: A prospec-tive analysis of pro-inflammatory and anti-inflammatory cytokines in patients with acute renalfailure (abstract). J Am Soc Nephrol 2002;13:5A.

21 Iglesias J, Marik PE, Levine JS, Investigators NIS: Elevated serum levels of the type I and type IIreceptors for tumor necrosis factor-alpha as predictive factors for ARF in patients with septicshock. Am J Kidney Dis 2003;41:62–75.

22 Venter J, Adams M, Myers E: The sequence of the human genome. Science 2001;291:1304–1351.23 Guttmacher A, Collins F: Genomic medicine: Genomic medicine – A primer. N Engl J Med

2002;347:1512–1520.24 Haukim N, Bidwell J, Smith A, Keen L, Gallagher G, Kimberly R, Huizinga T, McDermott M,

Oksenberg J, McNicholl J, Pociot F, Hardt C, D’Alfonso S: Cytokine gene polymorphism inhuman disease: On-line databases, supplement 2. Genes Immun 2002;3:313–330.

25 Nussbaum R, McInnes R, Willard H (eds): Genetic variation in individuals: Mutation and poly-morphism; in Thompson & Thompson Genetics in Medicine, ed 6. Philadelphia, Saunders, 2001,pp 79–94.

26 Abraham LJ, Kroeger KM: Impact of the -308 TNF promoter polymorphism on the transcriptionalregulation of the TNF gene: Relevance to disease. J Leukoc Biol 1999;66:562–566.

27 Morse H, Olomolaiye O, Wood N, Keen L, Bidwell J: Induced heteroduplex genotyping of TNF-alpha, IL-1beta, IL-6 and IL-10 polymorphisms associated with transcriptional regulation.Cytokine 1999;11:789–795.

28 Wilson AG, Symons JA, McDowell TL, McDewitt HO, Duff GW: Effects of a polymorphism inthe human tumor necrosis factor-� promoter on transcriptional activation. Proc Natl Acad SciUSA 1997;94:3195–3199.

29 Abraham L, Du D, Zahedi K, Dawkins R, Whitehead A: Haplotypic polymorphisms of the TNFBgene. Immunogenetics 1991;33:50–53.

30 Pociot F, Molvig J, Wogensen L, Worsaae H, Dalboge H, Baek L, Nerup J: A tumour necrosis fac-tor beta gene polymorphism in relation to monokine secretion and insulin-dependent diabetesmellitus. Scand J Immunol 1991;33:37–49.


31 Messer G, Spengler U, Jung M, Honold G, Blomer K, Pape G, Riethmuller G, Weiss E:Polymorphic structure of the tumor necrosis factor (TNF) locus: An NcoI polymorphism in the firstintron of the human TNF-beta gene correlates with a variant amino acid in position 26 and a reducedlevel of TNF-beta production. J Exp Med 1991;173:209–219.

32 Heesen M, Kunz D, Bachmann-Mennenga B, Merk H, Bloemeke B: Linkage disequilibrium betweentumor necrosis factor (TNF)-alpha-308 G/A promoter and TNF-beta NcoI polymorphisms:Association with TNF-alpha response of granulocytes to endotoxin stimulation. Crit Care Med2003;31:211–214.

33 Mira J-P, Cariou A, Grall F, Delclaux C, Losser M-R, Heshmati F, Cheval C, Monchi M, Teboul J-L,Riche F, Leleu G, Arbibe L, Mignon A, Delpech M, Dhainaut J-F: Association of TNF2, a TNF-[alpha] promoter polymorphism, with septic shock susceptibility and mortality: A multicenterstudy. JAMA 1999;282:561–568.

34 Tang GJ, Huang SL, Yien HW, Chen WS, Chi CW, Wu CW, Lui WY, Chiu JH, Lee TY: Tumornecrosis factor gene polymorphism and septic shock in surgical infection. Crit Care Med 2000;28:2733–2736.

35 Wunderink RG, Waterer GW, Cantor RM, Quasney MW: Tumor necrosis factor gene polymor-phisms and the variable presentation and outcome of community-acquired pneumonia. Chest2002;121:87S.

36 Majetschak M, Flohe S, Obertacke U, Schroder J, Staubach K, Nast-Kolb D, Schade FU, Stuber F:Relation of a TNF gene polymorphism to severe sepsis in trauma patients. Ann Surg 1999;230:207–214.

37 Waterer G, ElBahlawan L, Quasney M, Zhang Q, Kessler L, Wunderink R: Heat shock protein70–2 � 1267 AA homozygotes have an increased risk of septic shock in adults with community-acquired pneumonia. Crit Care Med 2003;31:1367–1372.

38 Yende S, Quasney MW, Tolley E, Zhang Q, Wunderink RG: Association of tumor necrosis factorgene polymorphisms and prolonged mechanical ventilation after coronary artery bypass surgery.Crit Care Med 2003;31:133–140.

39 Treszl A, Toth-Heyn P, Kocsis I, Nobilis A, Schuler A, Tulassay T, Vasarhelyi B: Interleukingenetic variants and the risk of renal failure in infants with infection. Pediatr Nephrol 2002;17:713–717.

40 Jaber B, Rao M, Guo D, Balakrishnan V, Perianayagam M, Freeman R, Pereira B: Cytokine pro-moter gene polymorphisms and mortality in acute renal failure. Cytokine 2004;25:212–219.

41 Crawley E, Kay R, Sillibourne J, Patel P, Hutchinson I, Woo P: Polymorphic haplotypes of theinterleukin-10 5� flanking region determine variable interleukin-10 transcription and are associatedwith particular phenotypes of juvenile rheumatoid arthritis. Arthritis Rheum 1999;42:1101–1108.

42 van der Pol WL HT, Vidarsson G, van der Linden MW, Jansen MD, Keijsers V, de Straat FG,Westerdaal NA, de Winkel JG, Westendorp RG: Relevance of Fc-gamma receptor and interleukin-10polymorphisms for meningococcal disease. J Infect Dis 2001;184:1548–1555.

43 Gallagher P, Lowe G, Fitzgerald T, Bella A, Greene C, McElvaney N, O’Neill S: Association ofIL-10 polymorphism with severity of illness in community acquired pneumonia. Thorax 2003;58:154–156.

44 Lowe PR, Galley HF, Abdel-Fattah A, Webster NR: Influence of interleukin-10 polymorphisms oninterleukin-10 expression and survival in critically ill patients. Crit Care Med 2003;31:34–38.

45 Ridker P, Stampfer M: Assessment of genetic markers for coronary thrombosis: Promise and pre-caution. Lancet 1999;353:687–688.

46 Newman TB, Browner WS, Hulley SB: Enhancing causal inference in observational studies; inHulley SB, Cummings SR (eds): Designing Clinical Research. Baltimore, Williams & Wilkins,1988, pp 98–109.

47 Jo S, Hu X, Kobayashi H, Lizak M, Miyaji T, Koretsky A, Star R: Detection of inflammation fol-lowing renal ischemia by magnetic resonance imaging. Kidney Int 2003;64:43–51.

48 Auphan N, Didonato JA, Rosette C, Helmberg A, Karin M: Immunosuppression by glucocorti-coids: Inhibition of NF-kB activity through induction of IkB synthesis. Science 1995;270:286–290.

49 Han J, Thompson P, Beutler B: Dexamethasone and pentoxifylline inhibit endotoxin-inducedcachectin/tumor necrosis factor synthesis at separate points in the signaling pathway. J Exp Med1990;172:391–394.


50 Doherty GM, Jensen JC, Alexander H, Buresh CM, Norton JA: Pentoxifylline suppression of tumornecrosis factor gene transcription. Surgery 1991;110:192–198.

51 Parry G, Mackman N: Role of cyclic AMP response element-binding protein in cyclic AMP inhi-bition of NF-kappaB-mediated transcription. J Immunol 1997;159:5450–5456.

52 Semmler J, Wachtel H, Endres S: The specific type IV phosphodiesterase inhibitor rolipram sup-presses tumor necrosis factor-production by human mono-nuclear cells. Int J Immunopharmacol1993;15:409–413.

53 Moreira AL, Sampaio EP, Zmuidzinas A, Frindt P, Smith KA, Kaplan G: Thalidomide exerts itsinhibitory action on tumor necrosis factor alpha by enhancing mRNA degradation. J Exp Med1993;177:1675–1680.

54 Moore K, de Waal M, Coffman R, O’Garra A: Interleukin-10 and the interleukin-10 receptor. AnnuRev Immunol 2001;765:683–765.

55 Pajkrt D, Camoglio L, Tiel-van Buul M, de Bruin K, Cutler D, Affrime M, Rikken G, van der Poll T,ten Cate J, van Deventer S: Attenuation of proinflammatory response by recombinant human IL-10 in human endotoxemia: Effect of timing of recombinant human IL-10 administration.J Immunol 1997;158:3971–3977.

56 Bermudez LE, Wu M, Young LS: Effect of stress-related hormones on macrophage receptors andresponse to tumor necrosis factor. Lymphokine Res 1990;9:137–145.

57 Brenner S, Prosch S, Schenke-Layland K, Riese U, Gausmann U, Platzer C: cAMP-inducedInterleukin-10 promoter activation depends on CCAAT/enhancer-binding protein expression andmonocytic differentiation. J Biol Chem 2003;278:5597–5604.

58 Mascali J, Cvietusa P, Negri J, Borish L: Anti-inflammatory effects of theophylline: Modulationof cytokine production. Ann Allergy Asthma Immunol 1996;77:34–38.

59 Escofier N, Boichot E, Germain N, Silva P, Martins M, Lagente V: Effects of interleukin-10 andmodulators of cyclic AMP formation on endotoxin-induced inflammation in rat lung. FundamClin Pharmacol 1999;13:96–101.

60 van Furth A, Seijmonsbergen E, Langermans J, van der Meide P, van Furth R: Effect of xanthinederivates and dexamethasone on Streptococcus pneumoniae-stimulated production of tumor necrosisfactor alpha, interleukin-1 beta (IL-1 beta), and IL-10 by human leukocytes. Clin Diagn Lab Immunol1995;2:689–692.

61 Cho M, Kim W, Min S, Min D, Min J, Lee S, Park S, Cho C, Kim H: Cyclosporine differentiallyregulates interleukin-10, interleukin-15, and tumor necrosis factor a production by rheumatoidsynoviocytes. Arthritis Rheum 2002;46:42–51.

62 Tabardel Y, Duchateau J, Schmartz D, Marecaux G, Shahla M, Barvais L, Leclerc JL, Vincent JL:Corticosteroids increase blood interleukin-10 levels during cardiopulmonary bypass in men. Surgery1996;119:76–80.

63 Jaber BL, Liangos O, Pereira BJ, Balakrishnan VS: Polymorphism of immunomodulatory cytokinegenes: Implications in acute renal failure. Blood Purif 2004;22:101–111.

64 Nadel S, Newport MJ, Booy R, Levin M: Variation in the tumor necrosis factor-6 promoter regionmay be associated with death from meningococcal disease. J Infect Dis 1996;174:878–880.

65 Zhang D, Li J, Jiang Z, Yu B, Tang X: Association of two polymorphisms of tumor necrosis factorgene with acute severe pancreatitis. J Surg Res 2003;112:138–143.

66 Stuber F, Petersen M, Bokelmann F, Schade U: A genomic polymorphism within the tumor necro-sis factor locus influences plasma tumor necrosis factor-alpha concentrations and outcome ofpatients with severe sepsis. Crit Care Med 1996;24:381–384.

67 Keijsers V, Verweij C, Westendorp R, et al: IL-10 promotor haplotypes: Discrepancy between lackof functionality in transfection assays and skewed distribution in meningococcal disease. GenesImmun 1999;1:78–79.

Bertrand L. Jaber, MDDepartment of Medicine, Caritas St. Elizabeth’s Medical Center736 Cambridge Street, Boston, MA 02135 (USA)Tel. �1 617 562 7832, Fax �1 617 562 7797, E-Mail [email protected]


Mechanisms of Immunodysregulation in Sepsis

Jean-Marc Cavaillon, Catherine Fitting, Minou Adib-Conquy

UP Cytokines and Inflammation, Institut Pasteur, Paris, France

Tissue injury, blood loss, hypoxia, transfusion, bacterial translocation,microbial infection, and cell activation by microbial products occur in patientswith systemic inflammatory response syndrome (SIRS) (e.g. trauma, hemor-rhage, burns, surgery or sepsis). These events contribute to the inflammatoryresponse and affect the quality of the immune status. In addition, drugs (e.g.anesthetics, opioids…) also influence immune responses (fig. 1). Depressedimmune status including decreased blood cell counts, low expression of surfacemarkers (e.g. MHC class II antigen), altered natural killer (NK) cell activity,diminished cellular cytotoxicity, reduced antigen presentation, poor prolifera-tion in response to mitogens and depressed cytokine production, are seenin vitro, and illustrated in vivo by anergy to skin test antigens. These observa-tions led Bone [1] to coin the concept of ‘compensatory anti-inflammatoryresponse syndrome’ or CARS. Bone postulated that when the SIRS responsepredominates, the syndrome is associated with an organ dysfunction andcardiovascular compromisation leading to shock; in contrast, when CARS pre-dominates, it is characterized by anti-inflammatory responses associated with asuppressive effect on the immune system, also known as ‘immunoparalysis’.However, this term is far too excessive and we will see that the alteration of theimmune response is not a generalized phenomenon. It was initially acceptedthat the SIRS response occurred first and was followed in some patients by theCARS response. However, it is most probable that the two syndromes occurconcomitantly [2]. Although alterations in immune responses are probablyassociated with an enhanced sensitivity to nosocomial infections, there is noclear demonstration that they are directly responsible for poor outcome insepsis. Furthermore, the mechanisms behind the maintenance of the sustainedsuppression of the immune function remain incompletely understood.

Mechanisms of Immunodysregulation in Sepsis 77

Assessment of Immunodysregulation

Circulating Leukocytes and HLA-DR ExpressionThe absolute number of circulating cells is modified during sepsis.

Analysis of lymphocyte subsets reveal a decreased number of circulating NKcells, CD4�, and CD8� lymphocytes, and an increase of B lymphocytes [3]and of regulatory T cells (CD4�/CD25�) [4]. Within monocytes, the subset ofCD14low/CD16� cells is increased [5]. In a mouse model of sepsis, anenhanced number of NK-T cells found in the spleen has been associated withimmune suppression [6].

Abnormal antigen presentation has been observed in SIRS patients [7],and decreased HLA-DR expression on monocytes may contribute to this defect[8]. In a study by Hershman et al. [9], there was a significant decrease of

Anesthetics

Blood loss

Tissue injuryTransfusion

Hypoxia

Bacterial translocation

Inflammatory response

Immune status

Immune dysregulation

Lymphocyte proliferation

NK activity

Cytokine production

Lymphocyte and monocyte

populations changes

Anergy to skin test antigens

Apoptosis

Trauma—Hemorrhage—Burns—Surgery—Sepsis

Cell surface markers changes

(e.g. HLA DR)

Microbial agents

Opioids

Fig. 1. Numerous events occurring during systemic inflammatory response syndromecontribute to modify the immune status as illustrated by various altered immune responsive-ness that can be monitored either in vitro, ex vivo or in vivo.

Cavaillon/Fitting/Adib-Conquy 78

HLA-DR expression on peripheral blood monocytes of trauma patients as com-pared to that of healthy volunteers. In those who developed sepsis, HLA-DRexpression took 3 weeks to return to normal and in the patients who did not sur-vive, this expression never returned to normal. HLA-DR antigen expressioncorrelated directly with the clinical status and identified a group of patients athigh risk of infection and death following trauma. Low HLA-DR expression isnow widely recognized as a good marker of the intensity of the immune depres-sion and of increased risk of bacterial infection [10]. In an elegant study,Fumeaux and Pugin [11] showed that the presence of circulating IL-10 in septicplasma was partially responsible of the reduced HLA-DR expression followingintracellular sequestration of the molecule.

Lymphocyte ProliferationImpaired lymphocyte proliferation in SIRS patients was reported more

than three decades ago [12]. The impairment was proportional to the severity ofthe injury. In vitro, lymphocyte proliferative response to antigens, mitogens andmixed lymphocyte reaction are all significantly decreased. Longer depressionof lymphocyte responses and lower responses were observed in patients whobecame infected and developed sepsis [13–15]. Increased level of apoptosis isnot directly associated with lowered T-cell proliferative response [16]. However,a profound apoptosis of B and T lymphocytes has been found in humans[17, 18], and it has been established that prevention of lymphocyte apoptosisimproves survival in a sepsis model in mice [19].

Natural Killer Cell ActivitySignificantly depressed NK cell activity has been described in burn and

trauma patients. In the study by Maturana et al. [20], patients with septic shockhad also a markedly lower NK activity than healthy controls. Pre-incubation ofperipheral blood lymphocytes with either interferon-� (IFN�) or interleukin-2(IL-2) enhanced NK cell activity in healthy controls but not in patients with sep-sis indicating the difficulty to reverse the depressed immune responsiveness [21].

Neutrophil FunctionsApoptosis of circulating neutrophils (PMN) is delayed in patients with

SIRS or sepsis [22], and mRNA expression of Mcl-1, an anti-apoptotic mem-ber of the Bcl-2 family, is upregulated in PMN of patients with sepsis [23]. Inaddition, some of the activities mediated by PMN appear to be altered. This isthe case of phagocytosis, bactericidal activity [24] and migration [25]. Thereduced responsiveness of PMN to chemoattractant agents may reflect theaction of nitric oxide [25], the decreased expression of certain chemokinereceptors [26], or a deactivation occurring in the blood stream after interacting


with large amounts of circulating chemokines as assessed, for example, by thehuge amounts of IL-8 found associated to PMN in septic patients [27].

Delayed HypersensitivityThe in vitro evidence of immune depression is also reflected in vivo by

tests of delayed type hypersensitivity. Several years ago, Christou et al. [28]skin tested surgical patients with recall antigens prior to operation. Patients whohad normal skin test responses were of similar age and had equal degrees ofsurgical procedures performed compared with those patients who were anergic(i.e. had depressed skin test responses). Post-operatively, sepsis, mortality anddeath due to sepsis were significantly higher in the anergic population, recon-firming the hypothesis that skin test anergy pre-operatively is a signal ofincreased risk for septic complications and death in such patients. These authors[29] reported that surgical patients who were anergic to a battery of five skintest antigens had a two-fold higher rate of postoperative infection than thosewho reacted to only two antigens and were more than fivefold likely to die inthe post-operative period. The hallmark of anergy is a lack of recruited specificT cells at the site of injection in the skin [30].

Ex vivo Cytokine ProductionA reduced capacity of circulating leukocytes from septic patients to pro-

duce pro-inflammatory cytokines as compared to cells from healthy controlshas been regularly reported in ex vivo experiments. The very first observationon the hyporeactivity of circulating cells in septic patients was demonstratedin peripheral blood lymphocytes: Wood et al. [31] reported a decreased IL-2production upon phytohemagglutinin (PHA) stimulation. Interferon-� (IFN�)production is affected in sepsis as well as in patients with severe injury [32].It is often suggested that the depressed response mainly affects the produc-tion of Th1 cytokines (IL-2, IFN�) whereas production of Th2 cytokineswould be upregulated. However, we demonstrated in sepsis and in SIRSpatients that the production of Th2 cytokines (IL-5, IL-10) could also bealtered and that the nature of the triggering agent itself influences the obser-vation [33]. Similarly, in patients resuscitated after cardiac arrest, identifiedas a sepsis-like syndrome, we showed that both concanavalin A-induced IL-10 (Th2) and IFN� (Th1) productions were significantly decreased inex vivo assays [34].

Similarly, ex vivo cytokine production (IL-1�, IL-1ra and IL-8) by LPS-activated neutrophils from septic patients is reduced as compared to PMN ofhealthy controls [35–37]. Most interestingly, van der Poll’s group [38] demon-strated that injection of endotoxin in healthy human volunteers led to hypo-reactive blood derived-PMN in terms of production of the chemokines IL-8,


GRO-� and ENA-78, when further challenged in vitro with either LPS, heat-killed Streptococcus pneumoniae or heat-killed Pseudomonas aeruginosa.

The reactivity to LPS stimulation has been particularly studied in isolatedmonocytes and in whole-blood assays. Monocytes from septic patients have adiminished capacity to release TNF�, IL-1�, IL-1�, IL-6, IL-10 and IL-12 [32,39–42], whereas this is not the case for IL-1ra [40] and G-CSF [43]. In humanvolunteers, intravenous endotoxin suppresses the cytokine response of periph-eral blood mononuclear cells (PBMC) when activated in vitro by LPS [44].Most interestingly, it was shown that the reduced production of cytokinesreflects a reduced number of cytokine-producing cells [45].

Many parameters of immmunoparalysis observed in SIRS patients arereminiscent of the endotoxin tolerance phenomenon, which characterizes therefractoriness of cells or whole animals to respond to a second endotoxin chal-lenge, shortly after a first encounter. Cross-tolerance has been regularlyreported in experimental models of endotoxin tolerance between LPS and otherToll-like receptors (TLR), ligands [46–48] or whole bacteria [49, 50]. Similarly,in sepsis patients and in LPS-injected human volunteers a reduced responsive-ness to gram-positive bacteria and superantigens have been reported [44,51–53]. However, this is not always the case and we showed that in SIRSpatients, cells normally responded to certain stimuli [54]. Revisiting the cross-tolerance with gram-positive bacteria in a mouse model of endotoxin tolerance,we showed that the phenomenon was only transient, especially in the bloodcompartment [55].

In addition to the nature of the stimuli, there are other observations thatallow ascertaining that the cellular hyporeactivity is not a global phenomenon.For example, we showed that the release of ‘macrophage migration inhibitoryfactor’ (MIF) by circulating mononuclear cells was enhanced in sepsis[Maxime et al., submitted]. Furthermore, during sepsis and SIRS, cells derivedfrom tissues are either fully responsive to ex vivo stimuli, or even primed, incontrast to cells derived from hematopoietic compartments (blood, spleen…),which are hyporeactive. This dichotomy illustrates the concept of compartmen-talization, which occurs within the body during sepsis [56]. Thus, during sys-temic inflammation, SIRS and CARS seem to be present concomitantly [2]:SIRS predominating within the inflamed tissues, while in the blood the leuko-cytes show hyporeactivity.

Concept of Compartmentalization

The measurement of inflammatory mediators is in agreement with theconcept of a compartmentalization of inflammation. For example, iNOS activity


was found to be restricted to the nidus of infection in patients undergoing a sep-tic shock after cellulitis [57]. TNF and IL-6 concentrations were higher in cere-brospinal fluids than in plasma of patients with meningitis [58]. After chesttrauma, significantly higher levels of IL-1� and IL-8 were found in broncho-alveolar lavage fluids than in plasma, whereas anti-inflammatory mediators(sTNFRI and II, IL-1ra) were present both locally and systematically [59].Measurement of pro-inflammatory cytokines and anti-inflammatory mediatorsin pleural fluids also suggested the occurrence of both systemic and compart-mentalized response in septic as well as in non-septic patients [60]. Finally, thecompartmentalized cytokine production was elegantly demonstrated byDehoux et al. [61] who showed higher levels of inflammatory cytokines inbroncho-alveolar lavage (BAL) fluids recovered from involved lung of patientswith unilateral pneumonia as compared to the contralateral, non-involved lung.In this study, the authors showed that the productions of IL-1, IL-6, and TNF byalveolar macrophages activated with LPS were reduced as compared to theresponses obtained with cells recovered from healthy controls. However, incontrast to this report, which deals with a local infectious process, and in con-trast to the hyporeactivity of the circulating leukocytes, it has often been shownthat cells derived from the inflamed tissues are activated and primed and fullyresponsive to ex vivo stimulation. In acute respiratory distress syndrome(ARDS) patients, alveolar macrophages displayed an enhanced capacity of pro-ducing IL-1 after ex vivo stimulation with LPS alone or together with IFN�[62]. Interestingly, in human volunteers, Smith et al. [63] reported that intra-venous endotoxin primed alveolar macrophages for an enhanced in vitro LPS-induced production of IL-1, TNF and PGE2. We analyzed the reactivity ofalveolar macrophages derived from baboons that had undergone unilateral lungirradiation. In the days following irradiation, not only the spontaneous produc-tion of IL-8 and TNF was enhanced, but those induced by LPS, Staphylococcusand Streptococcus were increased as compared to the response obtained withthe alveolar macrophages recovered before irradiation [2].

The study of peritoneal macrophages obtained from continuous ambula-tory peritoneal dialysis patients revealed that LPS-activated cells released sig-nificantly more IL-1� during peritonitis as compared with the infection-freeperiod [64]. In women with endometriosis, spontaneous and LPS-induced pro-duction of TNF�, IL-6, IL-8, IL-10, IL-12 and nitric oxide (NO) by peritonealmacrophages was higher than in controls [65, 66]. In patients with inflamma-tory bowel disease, adherent lamina propria mononuclear cells activated withpokeweed mitogen or with a combination of LPS and IFN� displayed an upreg-ulated production of TNF�, IL-1�, IL-1� and IL-6 [67].

The activation of intracellular signaling pathways has also been studied.Schwartz et al. [68] observed an increased activation of NF-�B in alveolar


macrophages from patients with ARDS. Moine et al. [69] subsequently showeddecreased cytoplasmic levels of p50, p65 and c-Rel in alveolar macrophagesfrom patients with ARDS, consistent with an enhanced migration of liberatedNF-�B dimers from the cytoplasm to the nucleus. In murine models of hemor-rhage or of endotoxinemia, activation of NF-�B, CREB, MEK-1/2 and Erk2were found in alveolar neutrophils but not in blood neutrophils [70, 71].

It is worth noting that cells derived from inflammatory foci may be far lesssensitive to anti-inflammatory mediators. This was elegantly demonstrated byPang et al. [72] who showed, in chronic bronchial sepsis patients, that the pro-duction of IL-8 by PMN derived from sputum was not significantly reduced bythe addition of increasing amounts of IL-10 in contrast to the inhibitory effectof IL-10 on circulating PMN. Similarly, we showed in cystic fibrosis patientsthat sputum-derived PMN that produce high levels of IL-8 were poorly down-regulated by dexamethasone in contrast to circulating PMN [73].

Mechanisms of Hyporesponsiveness of Monocytes

Desensitizing Agents in PlasmaThe presence of deactivating or immunosuppressive agents within the

blood stream may contribute to the hyporeactivity of circulating leukocytes. Inthe late 70’, it was reported that sera of burn patients were able to suppress theproliferative response of normal cells [74]. Prins et al. [75] showed that serafrom septic patients had the capacity to downregulate the TNF production byactivated monocytes from healthy donors. The fact that ‘septic plasma’ behaveas an immunosuppressive milieu [76] is illustrated in human volunteers by thecapacity of endotoxin to induce plasma inhibitors [77]. Most interestingly, inseptic patients this suppressive effect was significantly reduced after passage ofplasma through a resin and after incubation with anti-IL-10 antibodies [78]. IL-10 was identified as a major functional deactivator of monocytes in humanseptic shock plasma [79]. TGF� was also shown in animal models of hemor-rhagic shock and of sepsis to be the causative agent of the depressed splenocyteresponsiveness [80]. Monocytes from immunocompromised trauma patientsseem to be a source of TGF� [81], and TGF� released by apoptotic T cells con-tributes to this immunosuppressive milieu [82]. In addition, there is accumulat-ing evidence for a strong interaction between components of the nervous andthe immune systems. Numerous neuromediators have been shown to behave asimmunosuppressors. Catecholamines, found to be at higher concentrations instressful situations [83], suppress the activity of immunocompetent cells,inhibit TNF production and favor IL-10 release. Alpha-melanocyte-stimulatinghormone also contributes to immunosuppression by inducing IL-10 production


by human monocytes [84]. In addition, vasoactive intestinal peptide and pitu-itary adenylate cyclase-activating polypeptide directly inhibit endotoxininduced pro-inflammatory cytokine secretion [85]. SIRS is also associated withan activation of the hypothalamus-pituitary-adrenal axis which leads to therelease of glucocorticoids, well known for their potent ability to limit cytokineproduction [86]. Prostaglandins are produced during sepsis and can also con-tribute to the down regulation of cytokine production [87]. Finally, the levels ofcirculating heat shock proteins (HSP) are elevated in SIRS and sepsis patients[88]. Since it has been shown that over-expression of HSP can inhibit LPSinduced production of cytokines [89], one can postulate that HSP may play arole in desensitizing circulating cells.

Endotoxin-Neutralizing Molecules in PlasmaAs mentioned previously, the reduced capacity of monocytes to produce

inflammatory cytokines has been particularly established using LPS as a trig-gering agent. Since numerous recent studies analyzed the hyporeactivity phe-nomenon with whole blood samples, it is possible that endotoxin-neutralizingmolecules interfered in these studies. Indeed, we showed that the hyporeactiv-ity to LPS was both, an intrinsic property of circulating monocytes, as well asthe reflection of a specific neutralizing activity within the plasma of SIRSpatients [90]. It has been reported that plasma of septic patients content largeamounts of LPS-binding protein (LBP) which can either inhibit the LPS mole-cules [91], or transfer LPS to lipoproteins [92] known for their inhibitory activ-ity towards LPS [93]. Furthermore, sera from septic patients contain greatamounts of soluble CD14 which also favors the shuttle of LPS towards lipopro-teins [94]. More recently, enhanced levels of ubiquitin, an 8.6-kD proteininvolved in intracellular function, have been found in serum of sepsis andtrauma patients and shown to specifically inhibit TNF induction by LPS [95].

Toll-Like Receptor ExpressionToll-like receptors (TLR) are a family of receptors that recognize compo-

nents of bacteria, virus, parasites and fungi and induce a pro-inflammatoryresponse by several cell types. So far, 10 human TLRs differing in their speci-ficity for microbial components have been cloned, which respond to variouscomponents, including LPS from Gram-negative bacteria, lipopeptides ofGram-positive cell walls, bacterial DNA, and flagella. TLR4 was identified asthe receptor for LPS and requires the presence of an extracellular accessoryprotein called MD-2. CD14 physically associates with LPS complexed withLPS binding protein (LBP) and transfers the endotoxin to the TLR4 and MD-2dimer. Each component of this complex is required for efficient LPS-inducedsignaling.


Recent studies reported a downregulation of surface expression of TLR4 inendotoxin-tolerant macrophages [96, 97]. Therefore, it was of interest to investi-gate the expression of this molecule on the surface of monocytes from SIRSpatients. A decreased expression of TLR4 but not TLR2, on CD14-positive cellswas found in trauma patients compared with healthy subjects [54]. However, thislower expression of TLR4 was not sufficient to explain the decreased capacityof the cells to respond to some stimuli. Indeed, while LPS-induced TNF wasdecreased, this was not the case of LPS-induced IL-1ra and IL-10. This laterobservation suggests that the defect may occur at the level of the signaling path-ways within the cell leading to TNF production, rather than at the initiation of thesignaling cascade on the cell surface. In contrast to our observation, an enhancedexpression of TLR4 was found on monocytes of SIRS patients [98], and at thepresent time we don’t have any explanations for these divergent findings.

Nuclear Factor-kBNuclear factor-�B (NF-�B) is critical for maximal expression of many

cytokines involved in the pathogenesis of inflammation. Activation and regula-tion of NF-�B are tightly controlled by a group of inhibitory proteins (I�B),which maintain NF-�B in the cytoplasm of effector cells. Blackwell et al. [99]investigated the role of NF-�B in the mechanism of endotoxin tolerance in a ratalveolar macrophage cell line, which was made endotoxin-tolerant. This treat-ment induced a state of tolerance such that subsequent exposure to high-doseLPS resulted in decreased production of cytokines as compared to LPS-sensitive cells. This decreased cytokine production was associated with animpaired activation of NF-�B and a depletion of both p65 and p50 forms. Thisstudy suggested that endotoxin tolerance may be mediated by limiting theamount of NF-�B available for activation and thus, inhibiting transcription ofNF-�B-dependent genes. On the other hand, Ziegler-Heitbrock et al. [100]demonstrated that endotoxin tolerance in monocyte cell line was associatedwith an increase of the inactive p50 homodimer of NF-�B and a decrease of thep50p65 active heterodimer. Accordingly, we studied NF-�B expression anddimer characteristics in mononuclear cells of patients with severe sepsis ormajor trauma and of healthy controls [101, 102]. The expression of p65p50 het-erodimer was significantly reduced for all patients as compared to controls. Thep50p50 homodimer was reduced in the survivors of sepsis. In addition, subse-quent in vitro stimulation of PBMC with LPS did not induced further NF-�Bnuclear translocation: the survivors of sepsis showed low expression of bothp65p50 and p50p50, while non-survivors of sepsis showed a predominance ofthe inactive homodimer and a low p65p50/p50p50 ratio when compared to con-trols. In the later group of patients there was a reverse correlation betweenplasma IL-10 levels and the p65p50/p50p50 ratio after in vitro LPS stimulation


(r � �0.8, p � 0.04). The reduced expression of nuclear NF-�B was not due toits inhibition by I�B�, since very low expression of I�B�, and low levels of p65and p50 were found in the cytoplasm of PBMC from sepsis patients when com-pared to controls. These results demonstrate that upon LPS activation, PBMCof SIRS patients show patterns of NF-�B expression that resemble thosereported during LPS tolerance: global downregulation of NF-�B in survivors ofsepsis, presence of large amounts of the inactive homodimer in the non-survivorsof sepsis [101]. In trauma patients, we observed a long-term reduction of bothp65/p50 heterodimers and p50/p50 homodimers and a reduced p65p50/p50p50ratio after LPS stimulation in vitro [102]. Other mechanisms of NF-�B regula-tion have been reported in a rat model of sepsis for which the reduced capacityof PMN to produce TNF was associated with an increased generation of I�B�caused by the intrinsic generation of C5a [103].

Downregulation of TLR-Associated Signaling PathwaysThere are still very few studies, which have addressed in humans the link

between immunodepression seen in circulating cells in SIRS patients and alter-ation of some signaling pathways. In the past few years, numerous intracellularmolecules that negatively regulate LPS-activated signaling pathways have beendiscovered (table 1). This is the case of Toll-interacting protein (Tollip), an adap-tor protein found to associate with the cytoplasmic TIR domain of IL-1R, TLR2and TLR4 and to potently suppress the activity of IL-1 receptor associated kinase(IRAK) after TLR activation [104]. A splice variant of MyD88, termed MyD88s,induced upon LPS activation has been described. This molecule is defective in itsability to induce IRAK phosphorylation and behaves as dominant-negativeinhibitor [105]. Single immunoglobulin IL-1R-related molecule (SIGIRR) andST2, members of the TLR/IL-1R superfamily, are negative modulators of the sig-naling induced by IL-1 or TLR ligands in different types of cells [106a, 106b].RanGTPase plays a major role in nuclear trafficking and reduces NF-�B accu-mulation in the nucleus. As a consequence it acts as a dominant-negative regula-tor of LPS-induced TNF [107]. IRAK-M prevents the dissociation of IRAK andIRAK-4 from MyD88 and the formation of IRAK-TRAF6 complexes, and is anegative regulator of TLR signaling. Interestingly, endotoxin tolerance is signifi-cantly reduced in IRAK-M-deficient mice [108]. It has been recently reportedthat monocytes from septic patients when stimulated with LPS ex vivo, expressIRAK-M mRNA more rapidly than cells from healthy donors [109]. Similarly toIRAK-M�/� mice, endotoxin tolerance cannot be observed in mice deficient forsuppressor of cytokine signaling-1 (SOCS-1) [110]. In contrast to IRAK-M andSOCS-1 that function during the second or continuous exposure to stimulation,phosphatidylinositol 3-kinase (pi3K) can behave as an early inhibitor of TLRsignaling [111]. Learn et al. [112] reported in septic patients that the repressed


production of IL-1� and the selective elevation of secreted form of IL-1Ra inresponse to LPS was linked to a probably altered IRAK signaling pathway and amaintained efficient pi3K dependent signaling pathway.

In addition to negative regulators directly induced upon activation by TLRligands, other molecules induced by IL-10 can be involved such as hemeoxygenase-1 [113], Bcl-3 [114], Stat-3 [115], or SOCS-3 [116]. Indeed, in anexperimental model of sepsis, an upregulation of SOCS-3 was noticed inmacrophages and neutrophils present in spleen, lung and peritoneal cavity [117].

Table 1. Negative regulators of LPS-induced TNF production

Mediators Mechanisms Found in mouse models (M) or in human (H) SIRS

p50/p50 NF-�B prevents action of bioactive H, Mhomodimer p65/p50 heterodimer

IRAK-M prevents dissociation of IRAK and IRAK-4 from

MyD88 and formation of IRAK- H, MTRAF6 complexes

Pi3K early negative signal, favors HIL-1Ra production

MyD88s dominant-negative inhibitor of n.d.MyD88

Tollip associates with the cytoplasmic n.d.TIR domain

SIGIRR interacts with IL-1R, TLR4, 5 n.d.and 9 and TRAF6

ST2 sequestration of adaptators n.d.MyD88 and TIRAF/Mal

RanGTPase reduces NF-�B accumulation in n.d.the nucleus

SOCS-1 inactivation of STAT-1 and NF-�B MSOCS-3 induced by IL-10 MBcl-3 induced by IL-10 n.d.Stat-3 induced by IL-10 MHO-1 induced by IL-10 H

HO-1 � Heme oxygenase-1; IRAK � IL-1 receptor-associated kinase; NF-�B � nuclearfactor-�B; pi3K � phosphatidylinositol 3-kinase; SIGIRR � single immunoglobulin IL-1R-related molecule; SOCS � suppressor of cytokine signaling; STAT-1 � signal transducerand activator of transcription; Tollip � Toll-interacting protein; TRAF-6 � TNF receptor-associated factor-6.

n.d. � Not determined (see text for references).


Reversal of Monocytes HyporeactivityIFN� and GM-CSF can restore the responsiveness of monocytes from

SIRS and sepsis patients [118–120]. These observations are reminiscent of thewell-known capacity of these cytokines to reverse endotoxin tolerance [121,122]. Whether an in vivo treatment by these cytokines could have beneficialeffects remains to be fully established. Indeed, it is still unclear whether thisimmune alteration is responsible of poor outcome or only a reflect of a physio-logical adaptation. In this context, it is worth to mention the fact that extracor-poreal immunotherapy with IL-2 lowered mortality of surgical infection [123].

Alteration or Physiological Adaptation?

Sepsis and non-infectious SIRS are paradoxically associated with an exac-erbated production of cytokines, as assessed by their presence in biologicalfluids, and a diminished ability of circulating cells to produce cytokine uponin vitro activation. This may represent a protective response against an over-whelming dysregulation of the pro-inflammatory process, but on the other handit may alter the immune status (‘endogenous immunosuppression’) leading to anincreased risk of subsequent nosocomial infections [56]. However, cellularhyporeactivity is not a global phenomenon and some signaling pathways are unal-tered and allow the cells to respond normally to certain stimuli. Furthermore,during sepsis and SIRS, cells derived from tissues or inflammatory foci areeither fully responsive to ex vivo stimuli or even primed, in contrast to cellsderived from hematopoietic compartments (blood) that are hyporeactive. Thefirst analysis of intracellular signaling pathways within leukocytes of sepsispatients revealed that some alterations are associated with cellular hyporeactiv-ity. Thus, the immunodysregulation reported in sepsis and SIRS patients, oftenillustrated by a diminished capacity of leukocytes to respond to LPS, is not ageneralized phenomenon and SIRS is associated with a compartmentalizedresponsiveness, which involves either anergic or primed cells.

References

1 Bone RC, Grodzin CJ, Balk RA: Sepsis: A new hypothesis for pathogenesis of the disease process.Chest 1997;121:235–243.

2 Cavaillon J-M, Adib-Conquy M, Cloëz-Tayarani I, Fitting C: Immunodepression in sepsis and SIRSassessed by ex vivo cytokine production is not a generalized phenomenon: A review. J EndotoxinRes 2001;7:85–93.

3 Holub M, Kluckova Z, Beneda B, et al: Changes in lymphocyte subpopulations and CD3�/DR�expression in sepsis. Clin Microbiol Infect 2000;6:657–660.

4 Monneret G, Debard AL, Venet F, et al: Marked elevation of human circulating CD4�CD25�regulatory T cells in sepsis-induced immunoparalysis. Crit Care Med 2003;31:2068–2071.


5 Fingerle G, Pforte A, Passlick B, Blumenstein M, Strobel M, Ziegler-Heitbrock HWL: The novelsubset of CD14�/CD16� blood monocytes is expanded in sepsis patients. Blood 1993;82:3170–3176.

6 Rhee R, Carlton S, Lomas J, et al: Inhibition of CD1d activation suppresses septic mortality:A role for NK-T cells in septic immune dysfunction. J Surg Res 2003;115:74–81.

7 Polk HJ, George CD, Wellhausen S, et al: A systematic study of host defense processes in badlyinjured patients. Ann Surg 1986;204:282–299.

8 Livingston D, Appel S, Wellhausen S, Polk H: Depressed interferon gamma production and mono-cyte HLA-DR expression after severe injury. Arch Surg 1988;123:1309–1312.

9 Hershman MJ, Cheadle WG, Wellhausen SR, Davidon P, Polk HC: Monocyte HLA-DR antigenexpression characterizes clinical outcome in the trauma patients. Br J Surg 1990;77:204–207.

10 van den Berk JMM, Oldenburger RHJ, van den Berg AP, et al: Low HLA DR expression on mono-cytes as a prognostic marker for bacterial sespsis after liver transplantation. Transplantation 1997;63:1846–1848.

11 Fumeaux T, Pugin J: Role of interleukin-10 in the intracellular sequestration of human leukocyteantigen-DR in monocytes during septic shock. Am J Respir Crit Care Med 2002;166:1475–1482.

12 Salo M, Merikanto J, Eskola J, Nieminen S, Aho AJ: Impaired lymphocyte transformation afteraccidental trauma. Acta Chir Scand 1970;145:367–372.

13 Baker C, Miller C, Trunkey D: Predicting fatal sepsis in burn patients. J Trauma 1979;19:641–648.14 Keane R, Birmingham W, Shatney C, Winchurch R, Munster A: Prediction of sepsis in the multi-

traumatic patients by assays of lymphocyte responsiveness. Surg Gynecol Obstet 1983;156:163–167.15 Levy EM, Alharbi SA, Grindlinger G, Black PH: Changes in mitogen responsiveness lymphocytes

subsets after traumatic injury: Relation to development of sepsis. Clin Immunol Immunopathol1984;32:224–233.

16 Pellegrini J, De A, Kodys K, Puyana J, Furse R, Miller-Graziano C: Relationships between T lym-phocyte apoptosis and anergy following trauma. J Surg Res 2000;88:200–206.

17 Hotchkiss RS, Tinsley KW, Swanson PE, et al: Sepsis-induced apoptosis causes progressive pro-found depletion of B and CD4� T lymphocytes in humans. J Immunol 2001;166:6952–6963.

18 Hotchkiss RS, Karl IE: The pathophysiology and treatment of sepsis. N Engl J Med 2003;348:138–150.

19 Hotchkiss RS, Tinsley KW, Swanson PE, et al: Prevention of lymphocyte cell death in sepsisimproves survival in mice. Proc Natl Acad Sci USA 1999;96:14541–14546.

20 Maturana P, Puente J, Miranda D, Sepulveda C, Wolf ME, Mosnaim AD: Natural killer cell activ-ity in patients with septic shock. J Crit Care 1991;6:42–45.

21 Puente J, Carvajal T, Parra S, et al: In vitro studies of natural killer cell activity in septic shockpatients: Response to a challenge with alpha-interferon and interleukin-2. Int J Clin PharmacolTher Toxicol 1993;31:271–275.

22 Jimenez M, Watson R, Parodo J, et al: Dysregulated expression of neutrophil apoptosis in the sys-temic inflammatory response syndrome. Arch Surg 1997;132:1263–1269.

23 Harter L, Mica L, Stocker R, Trentz O, Keel M: Mcl-1 correlates with reduced apoptosis in neu-trophils from patients with sepsis. J Am Coll Surg 2003;197:964–973.

24 van Dijk W, Verbrugh H, van der Tol M, et al: Interactions of phagocytic and bacterial cells inpatients with bacteremia caused by gram-negative rods. J Infect Dis 1980;141:441–449.

25 Benjamim C, Ferreira S, Cunha F: Role of nitric oxide in the failure of neutrophil migration insepsis. J Infect Dis 2000;182:214–223.

26 Cummings CJ, Martin TR, Frevert CW, et al: Expression and function of the chemokine receptorCXCR1 and CXCR2 in sepsis. J Immunol 1999;162:2341–2346.

27 Marie C, Fitting C, Cheval C, et al: Presence of high levels of leukocyte-associated interleukin-8upon cell activation and in patients with sepsis syndrome. Infect Immun 1997;65:865–871.

28 Christou NV: Host defense mechanism in surgical patients: A correlation study of the delayedhypersensitivity skin test response, granulocyte function and sepsis. Can J Surg 1985;28:39–49.

29 Christou NV, Meakins JL, MacLean LD: The predictive role of delayed hypersensitivity in preop-erative patients. Surg Gynecol Obstet 1981;152:297–301.

30 Christou N, Meakins J, Gordon J, et al: The delayed hypersensitivity response and host resistancein surgical patients: 20 years later. Ann Surg 1995;222:534–546.


31 Wood J, Rodrick M, O’Mahony J, et al: Inadequate interleukin 2 production: A fundamentalimmunological deficiency in patients with major burns. Ann Surg 1984;200:311–320.

32 Ertel W, Keel M, Neidhardt R, et al: Inhibition of the defense system stimulating interleukin-12interferon-� pathway during critical illness. Blood 1997;89:1612–1620.

33 Muret J, Marie C, Fitting C, Payen D, Cavaillon J-M: Ex vivo T-lymphocyte derived cytokine pro-duction in SIRS patients is influenced by experimental procedures. Shock 2000;13:169–174.

34 Adrie C, Adib-Conquy M, Laurent I, et al: Successful cardiopulmonary resuscitation after cardiacarrest as a ‘sepsis like’ syndrome. Circulation 2002;106:562–568.

35 McCall CE, Grosso-Wilmoth LM, LaRue K, Guzman RN, Cousart SL: Tolerance to endotoxin-induced expression of the interleukin-1ß gene in blood neutrophils of humans with the sepsis syn-drome. J Clin Invest 1993;91:853–861.

36 Marie C, Muret J, Fitting C, Losser M-R, Payen D, Cavaillon J-M: Reduced ex vivo interleukin-8production by neutrophils in septic and non-septic systemic inflammatory response syndrome.Blood 1998;91:3439–3446.

37 Marie C, Muret J, Fitting C, Payen D, Cavaillon J-M: IL-1 receptor antagonist production duringinfectious and noninfectious systemic inflammatory response syndrome. Crit Care Med 2000;28:2277–2283.

38 Schultz MJ, Olszyna DP, de Jonge E, Verbon A, van Deventer SJH, van der Poll T: Reducedex vivo chemokine production by polymorphonuclear cells after in vivo exposure of normalhumans to endotoxin. J Infect Dis 2000;182:1264–1267.

39 Muñoz C, Misset B, Fitting C, Bleriot JP, Carlet J, Cavaillon J-M: Dissociation between plasmaand monocyte-associated cytokines during sepsis. Eur J Immunol 1991;21:2177–2184.

40 Van Deuren M, Van Der Ven-Jongekrijg H, Demacker PNM, et al: Differential expression ofproinflammatory cytokines and their inhibitors during the course of meningococcal infections.J Infect Dis 1994;169:157–161.

41 Marchant A, Alegre M, Hakim A, et al: Clinical and biological significance of interleukin-10plasma levels in patients with septic shock. J Clin Immunol 1995;15:265–272.

42 Randow F, Syrbe U, Meisel C, et al: Mechanism of endotoxin desensitization: Involvement ofinterleukin-10 and transforming growth factor-�. J Exp Med 1995;181:1887–1892.

43 Weiss M, Fisher G, Barth E, et al: Dissociation of LPS-induced monocytic ex vivo production of gran-ulocyte colony stimulating factor and TNF� in patients with septic shock. Cytokine 2000;13:51–54.

44 Granowitz EV, Porat R, Mier JW, et al: Intravenous endotoxin suppresses the cytokine response ofperipheral blood mononuclear cells of healthy humans. J Immunol 1993;151:1637–1645.

45 Spolarics Z, Siddiqi M, Siegel J, et al: Depressed interleukin-12-producing activity by monocytescorrelates with adverse clinical course and a shift toward Th2-type lymphocyte pattern in severelyinjured male trauma patients. Crit Care Med 2003;31:1722–1729.

46 Kreutz M, Ackerman U, Hauschildt S, et al: A comparative analysis of cytokine production andtolerance induction by bacterial lipopetides, lipopolysaccharides and Staphylococcus aureus inhuman monocytes. Immunology 1997;92:396–401.

47 Crabtree TD, Jin L, Raymond DP, et al: Preexposure of murine macrophages to CpG oligonu-cleotide results in a biphasic tumor necrosis factor alpha response to subsequent lipopolysaccha-ride challenge. Infect Immun 2001;69:2123–2129.

48 Lehner MD, Morath S, Michelsen KS, Schumann RR, Hartung T: Induction of cross-tolerance bylipopolysaccharide and highly purified lipoteichoic acid via different toll-like receptors indepen-dent of paracrine mediators. J Immunol 2001;166:5161–5167.

49 Cavaillon JM, Pitton C, Fitting C: Endotoxin tolerance is not a LPS-specific phenomenon: Partialmimicry with IL-1, IL-10 and TGF�. J Endotoxin Res 1994;1:21–29.

50 Nose M, Uzawa A, Nomura M, et al: Control of endotoxin shock by the dried preparation of lowvirulent Streptococcus pyogenes OK-432. Cell Immunol 1998;188:97–104.

51 Cavaillon J-M, Muñoz C, Marty C, et al: Cytokine production by monocytes from patients withsepsis syndrome and by endotoxin-tolerant monocytes; in Levin J, Alving CR, Munford RS, StützPL (eds): Bacterial Endotoxin: Recognition and Effector Mechanisms. Amsterdam, Elsevier, 1993.

52 Astiz ME, Saha DC, Brooks K, Carpati CM, Rackow EC: Comparison of the induction of endo-toxin tolerance in endotoxemia and peritonitis by monophosphoryl lipid A and lipopolysaccharide.Circulatory Shock 1993;39:194–198.


53 Wilhelm W, Grundmann U, Rensing H, et al: Monocyte deactivation in severe human sepsis or fol-lowing cardiopulmonary bypass. Shock 2002;17:354–360.

54 Adib-Conquy M, Moine P, Asehnoune K, et al: Toll-like receptor (TLR) 2-, TLR4- and TLR9-mediated TNF and IL-10 production are not similarly affected during systemic inflammation.Am J Resp Crit Care Med 2003;168:158–164.

55 Fitting C, Dhawan S, Cavaillon JM: Compartmentalisation of endotoxin tolerance. J Infect Dis2004;189:1295–1303.

56 Munford RS, Pugin J: Normal response to injury prevent systemic inflammation and can beimmunosuppressive. Am J Respir Crit Care Med 2001;163:316–321.

57 Annane D, Sanquer S, Sébille V, et al: Compartmentalised inducible nitric-oxide synthase activityin septic shock. Lancet 2000;355:1143–1148.

58 Waage A, Halstensen A, Espevik T, Brandzaeg P: Compartmentalization of TNF and IL-6 inmeningitis and septic shock. Mediators Inflam 1993;2:23–25.

59 Keel M, Ecknauer E, Stocker R, et al: Different pattern of local and systemic release of pro-inflammatory and anti-inflammatory mediators in severely injured patients with chest trauma.J Trauma 1996;40:907–914.

60 Marie C, Losser MR, Fitting C, Kermarrec N, Payen D, Cavaillon J-M. Cytokines and solublecytokines receptors in pleural effusions from septic and nonseptic patients. Am J Respir Crit CareMed 1997;156:1515–1522.

61 Dehoux MS, Boutten A, Ostinelli J, et al: Compartmentalized cytokine production within thehuman lung in unilateral pneumonia. Am J Respir Crit Care Med 1994;150:710–716.

62 Jacobs RF, Tabor DR, Burks AW, Campbell GD: Elevated interleukin-1 release by human alveolarmacrophages during adult respiratory distress syndrome. Am Rev Respir Dis 1989;140:1686–1692.

63 Smith PD, Suffredini AF, Allen JB, Wahl LM, Parrillo JE, Wahl SM: Endotoxin administration tohumans primes alveolar macrophages for increased production of inflammatory mediators. J ClinImmunol 1994;14:141–148.

64 Fieren MWJA, Van Den Bemd GJ, Bonta IL: Endotoxin-stimulated peritoneal macrophagesobtained from continuous ambulatory peritoneal dialysis patients show an increased capacity torelease interleukin-1� in vitro during infectious peritonitis. Eur J Clin Invest 1990;20:453–457.

65 Rana N, Braun DP, House R, Gebel H, Rotman C, Dmowski WP: Basal and stimulated secretion ofcytokines by peritoneal macrophages in women with endometriosis. Fertil Steril 1996;65:925–930.

66 Wu MY, Ho HN, Chen SU, Chao KH, Chen CD, Yang YS: Increase in the production of IL-6, IL-10 and IL-12 by LPS stimulated peritoneal macrophages from women with endometriosis.Am J Reprod Immunol 1999;41:106–111.

67 Rugtveit J, Nilsen EM, Bakka A, Carlsen H, Brandtzaeg P, Scott H: Cytokine profiles differ innewly recruited and resident subsets of mucosal macrophages from inflammatory bowel disease.Gastroenterology 1997;112:1493–1505.

68 Schwartz MD, Moore E, Moore FA, et al: Nuclear factor-kappa B is activated in alveolar macrophagesfrom patients with acute respiratory distress syndrome. Crit Care Med 1996;24:1285–1292.

69 Moine P, McIntyre R, Schwartz MD, et al: NF-�B regulatory mechanisms in alveolar macrophagesfrom patients with acute respiratory distress syndrome. Shock 2000;13:85–91.

70 Shenkar R, Abraham E: Mechanisms of lung neutrophil activation after hemorrhage or endotox-emia: Roles of reactive oxygen intermediates, NF-�B and cyclic AMP response element bindingprotein. J Immunol 1999;163:954–962.

71 Abraham E, Arcaroli J, Shenkar R: Activation of extracellular signal-regulated kinases, NF-�B,and cyclic adenosine 5�-monophosphate response element binding protein in lung neutrophilsoccurs by differing mechanisms after hemorrhage or endotoxemia. J Immunol 2001;166:522–530.

72 Pang G, Ortega M, Zighang R, Reeves G, Clancy R: Autocrine modulation of IL-8 production bysputum neutrophils in chronic bronchial sepsis. Am J Respir Crit Care Med 1997;155:726–731.

73 Corvol H, Fitting C, Chadelat K, et al: Distinct cytokine production by lung and blood neutrophilsfrom children with cystic fibrosis. Am J Physiol 2003;284:L997–L1003.

74 Constantian M: Association of sepsis with an immunosuppressive polypeptide in the serum ofburn patients. Ann Surg 1978;188:209–215.

75 Prins JM, Kuijper EJ, Mevissen ML, Speelman P, van Deventer SJ: Release of tumor necrosis fac-tor alpha and interleukin 6 during antibiotic killing of Escherichia coli in whole blood: Influence


of antibiotic class, antibiotic concentration, and presence of septic serum. Infect Immun 1995;63:2236–2242.

76 Cavaillon J-M: ‘Septic Plasma’: An immunosuppressive milieu. Am J Respir Crit Care Med 2002;166:1417–1418.

77 Spinas G, Bloesch D, Kaufmann M, Keller U, Dayer JM: Induction of plasma inhibitors of inter-leukin 1 and TNF-alpha activity by endotoxin administration to normal humans. Am J Physiol1990;259:R993–R997.

78 Ronco C, Brendolan A, Lonnemann G, et al: A pilot study of coupled plasma filtration with adsorp-tion in septic shock. Crit Care Med 2002;30:1250–1255.

79 Brandtzaeg P, Osnes L, Øvstebø R, Joø GB, Westwik AB, Kierulf P: Net inflammatory capacityof human septic shock plasma evaluated by a monocyte-based target cell assay: Identification ofinterleukin-10 as a major functional deactivator of human monocytes. J Exp Med 1996;184:51–60.

80 Ayala A, Knotts JB, Ertel W, Perrin MM, Morrison MH, Chaudry IH: Role of interleukin-6 andtransforming growth factor-beta in the induction of depressed splenocyte responses following sep-sis. Arch Surg 1993;128:89–94.

81 Miller-Graziano CL, Szabo G, Griffey K, Mehta B, Kodys K, Catalano D: Role of elevated mono-cyte transforming growth factor production in post-trauma immunosuppression. J Clin Immunol1991;11:95–102.

82 Chen W, Frank M, Jin W, Wahl S: TGF-beta released by apoptotic T cells contributes to animmunosuppressive milieu. Immunity 2001;14:715–725.

83 Jones SB, Westfall MV, Sayeed MM: Plasma catecholamines during E. coli bacteremia in con-scious rats. Am J Physiol 1988;254:R470–R477.

84 Luger TA, Kalden DH, Scholzen TE, Brzoska T: �-melanocyte-stimulating hormone as a media-tor of tolerance induction. Pathobiology 1999;67:318–321.

85 Delgado M, Pozo D, Martinez C, et al: Vasoactive intestinal peptide and pituitary adenylatecyclase-activating polypeptide inhibit endotoxin-induced TNF� production by macrophages:In vitro and in vivo studies. J Immunol 1999;162:2358–2367.

86 Annane D, Cavaillon JM: Corticosteroids in sepsis: From bench to bedside? Shock 2003;20:197–207.

87 Choudhry MA, Ahmad S, Ahmed Z, Sayeed MM: Prostaglandin E2 down-regulation of T cell IL-2production is independent of IL-10 during Gram negative sepsis. Immunol Lett 1999;67:125–130.

88 Njemini R, Lambert M, Demanet C, Mets T: Elevated serum heat-shock protein 70 levels inpatients with acute infection: Use of an optimized enzyme-linked immunosorbent assay. ScandJ Immunol 2003;58:664–669.

89 Ding X, Fernandez-Prada C, Bhattacharjee A, Hoover D: Over-expression of hsp-70 inhibits bacte-rial lipopolysaccharide-induced production of cytokines in human monocyte-derived macrophages.Cytokine 2001;16:210–219.

90 Cavaillon JM, Adrie C, Fitting C, Adib-Conquy M: Endotoxin tolerance: Is there a clinical rele-vance? J Endotoxin Res 2003;9:101–107.

91 Zweigner J, Gramm HJ, Singer OC, Wegscheider K, Schumann RR: High concentration oflipopolysaccharide-binding protein in serum of patients with severe sepsis or septic shock inhibitthe lipopolysaccharide response in human monocytes. Blood 2001;98:3800–3808.

92 Vreugdenhil ACE, Snoeck AMP, van’t Veer C, Greve JWM, Buurman WA: LPS-binding proteincirculates in association with apoB-containing lipoproteins and enhances endotoxin-LDL/VLDLinteraction. J Clin Invest 2001;107:225–234.

93 Cavaillon JM, Fitting C, Haeffner-Cavaillon N, Kirsch SJ, Warren HS: Cytokine response bymonocytes and macrophages to free and lipoprotein-bound lipopolysaccharide. Infect Immun1990;58:2375–2382.

94 Kitchens RL, Thompson PA, Viriyakosol S, O’Keefe GE, Munford RS: Plasma CD14 decreasesmonocyte responses to LPS by transferring cell-bound LPS to plasma lipoproteins. J Clin Invest2001;108:485–493.

95 Majetschak M, Krehmeier U, Bardenheuer M, et al: Extracellular ubiquitin inhibits the TNF-alpharesponse to endotoxin in peripheral blood mononuclear cells and regulates endotoxin hypore-sponsiveness in critical illness. Blood 2003;101:1882–1890.


96 Nomura F, Akashi S, Sakao Y, et al: Endotoxin tolerance in mouse peritoneal macrophagescorrelates with down-regulation of surface Toll-like receptor 4 expression. J Immunol 2000;164:3476–3479.

97 Medvedev AE, Kopydlowski KM, Vogel SN: Inhibition of lipopolysaccharide-induced signaltransduction in endotoxin-tolerized mouse macrophages: Dysregulation of cytokine, chemokine,and Toll-like receptor 2 and 4 gene expression. J Immunol 2000;164:5564–5574.

98 Calvano J, Agnese D, Um J, et al: Modulation of the lipopolysaccharide receptor complex (CD14,TLR4, MD-2) and toll-like receptor 2 in systemic inflammatory response syndrome-positivepatients with and without infection: Relationship to tolerance. Shock 2003;20:415–419.

99 Blackwell TS, Blackwell TR, Christman JW: Induction of endotoxin tolerance depletes nuclear fac-tor-�B and suppresses its activation in rat alveolar macrophages. J Leuk Biol 1997;62:885–891.

100 Ziegler-Heitbrock HWL, Wedel A, Schraut W, et al: Tolerance to lipopolysaccharide involvesmobilization of nuclear factor-�B with predominance of p50 homodimers. J Biol Chem 1994;269:17001–17004.

101 Adib-Conquy M, Adrie C, Moine P, et al: NF-�B expression in mononuclear cells of septic patientsresembles that observed in LPS-tolerance. Am J Respir Crit Care Med 2000;162:1877–1883.

102 Adib-Conquy M, Asehnoune K, Moine P, Cavaillon J-M: Longterm impaired expression ofnuclear factor-�B and I�B� in peripheral blood mononuclear cells of patients with major trauma.J Leuk Biol 2001;70:30–38.

103 Riedemann N, Guo R, Bernacki K, et al: Regulation by C5a of neutrophil activation during sep-sis. Immunity 2003;19:193–120.

104 Zhang G, Ghosh S: Negative regulation of toll-like receptor-mediated signaling by Tollip. J BiolChem 2002;277:7059–7065.

105 Janssens S, Burns K, Tschopp J, Beyaert R. Regulation of interleukin-1- and lipopolysaccharide-induced NF-kappaB activation by alternative splicing of MyD88. Curr Biol 2002;12:467–471.

106a Wald D, Qin J, Zhao Z, et al: SIGIRR, a negative regulator of Toll-like receptor-interleukin 1receptor signaling. Nat Immunol 2003;4:920–927.

106b Brint EK, Xu D, Liu H, et al: ST2 is an inhibitor of interleukin 1 receptor and Toll-like receptor 4signaling and maintains endotoxin tolerance. Nat Immunol 2004;publ online.

107 Yuan Q, Zhao F, Chung S, et al: Dominant negative down-regulation of endotoxin-induced tumornecrosis factor alpha production by Lps(d)/Ran. Proc Natl Acad Sci USA 2000;97:2852–2857.

108 Kobayashi K, Hernandez LD, Galan JE, Janeway CAJ, Medzhitov R, Flavell RA: IRAK-M is anegative regulator of Toll-like receptor signaling. Cell 2002;110:191–202.

109 Escoll P, del Fresno C, Garcia L, et al: Rapid up-regulation of IRAK-M expression following asecond endotoxin challenge in human monocytes and in monocytes isolated from septic patients.Biochem Biophys Res Commun 2003;311:465–472.

110 Nakagawa R, Naka T, Tsutsui H, et al: SOCS-1 participates in negative regulation of LPSresponses. Immunity 2002;17:677–687.

111 Fukao T, Koyasu S: PI3K and negative regulation of TLR signaling. Trends Immunol 2003;24:358–363.

112 Learn CA, Boger MS, Li L, McCall CE: The phosphatidylinositol 3 kinase pathway selectivelycontrols sIL-1ra not interleukin-1� production in the septic leukocytes. J Biol Chem 2001;276:20234–20239.

113 Lee T, Chau L: Heme oxygenase-1 mediates the anti-inflammatory effect of interleukin-10 inmice. Nat Med 2002;8:240–246.

114 Kuwata H, Watanabe Y, Miyoshi H, et al: IL-10-inducible Bcl-3 negatively regulates LPS-inducedTNF-alpha production in macrophages. Blood 2003;102:4123–4129.

115 O’Farrell A-M, Liu Y, Moore KW, Mui ALF: IL-10 inhibits macrophage activation and prolifera-tion by distinct signaling mechanisms: Evidence for Stat3-dependent and-independent pathways.EMBO J 1998;17:1006–1018.

116 Berlato C, Cassatella MA, Kinjyo I, Gatto L, Yoshimura A, Bazzoni F: Involvement of suppressorof cytokine signaling-3 as a mediator of the inhibitory effects of IL-10 on lipopolysaccharide-induced macrophage activation. J Immunol 2002;168:6404–6411.

117 Grutkoski P, Chen Y, Chung C, Ayala A: Sepsis-induced SOCS-3 expression is immunologicallyrestricted to phagocytes. J Leukoc Biol 2003;74:916–922.


118 Döcke WD, Randow F, Syrbe U, et al: Monocyte deactivation in septic patients: Restoration byIFN� treatment. Nat Med 1997;3:678–681.

119 Flohé S, Lendemans S, Selbach C, et al: Effect of granulocyte-macrophage colony-stimulatingfactor on the immune response of circulating monocytes after severe trauma. Crit Care Med 2003;31:2462–2469.

120 Williams M, Withington S, Newland A, Kelsey S. Monocyte anergy in septic shock is associatedwith a predilection to apoptosis and is reversed by granulocyte-macrophage colony-stimulatingfactor ex vivo. J Infect Dis 1998;178:1421–1433.

121 Randow F, Döcke WD, Bundschuh DS, Hartung T, Wendel A, Volk HD: In vitro prevention andreversal of lipopolysaccharide desensitization by IFN�, IL-12 and GM-CSF. J Immunol 1997;158:2911–2918.

122 Adib-Conquy M, Cavaillon J-M: IFN� and GM-CSF prevent endotoxin tolerance in human mono-cytes by promoting IRAK expression and its association to MyD88, and not by modulating TLR4expression. J Biol Chem 2002;277:27927–27934.

123 Pal’tsev A, Ovechkin A, Zakharova N, et al: Cytokines in the treatment of a generalized surgicalinfection. Anesteziol Reanimatol 2000;2:27–30.

Jean-Marc CavaillonUP Cytokines and InflammationInstitut Pasteur, 28, rue Dr Roux, FR–75015 Paris (France)Tel. �33 1 45 68 82 38, Fax �33 1 40 61 35 92, E-Mail [email protected]


Goals of Resuscitation from Circulatory Shock

Michael R. Pinsky

Department of Critical Care Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pa., USA

Cardiovascular insufficiency commonly occurs in critically ill patients andmay be the reason for intensive care unit (ICU) admission or a complication oftreatment or disease during ICU stay. If cardiovascular insufficiency is associatedwith inadequate O2 delivery to the tissues, then the patient is said to be in shock.The management of shock-like states depends on both the etiology of the con-dition and attaining certain cardiovascular targets of blood flow, O2 delivery andperfusion pressure. Thus, the physician caring for such individuals needs to havea rational approach to both the diagnosis of shock and its resuscitative therapy.

Shock States

Shock is a circulatory a state characterized by inadequate delivery of nutri-ents and O2 delivery to meet the metabolic requirements of the tissues. Shocketiologies have been traditionally divided into four broad categories based onpathophysiologic mechanisms [1]: hypovolemic, cardiogenic, distributive, andobstructive. Although resuscitation of patients in shock may restore adequateglobal O2 transport, it may not restore blood flow to damaged vascular beds dueto the associated alterations in autoregulation and microvascular blood flow.

Clearly, any given patient may demonstrate qualities that blur the separationof shock into these specific states. Still, the most common quality of a shock-likestate is hypovolemia, as characterized by hypovolemic shock. In hypovolemicshock, the vascular volume is near or below the stressed volume, resulting in aprofound decrease in venous return. Independent of any changes in vasomotortone or cardiac pump function, a loss of 30–40% of the circulating volume willlead to marked hypotension and organ hypoperfusion [2]. Although not clearly

Fluid, Electrolyte and Acid Base

Goals of Resuscitation 95

documented in humans, in animal models of hemorrhagic shock, a greater thana 40% loss of intravascular volume can result in irreversible shock and death ifnot treated effectively with fluid resuscitation in less than two hours [3]. Thenormal physiologic response to a hypovolemic stress is to increase sympatheticoutput that increases vascular tone and thus reduces unstressed volume. Theresultant effect is to increase upstream pressure in the venous reservoirs increasingvenous return. The associated increased sympathetic tone increases heart rate,contractility and cortisol secretion. Furthermore, the associated release of aldos-terone from the adrenal cortex increases sodium and water reabsorption in thekidney. Although these endocrinological responses are effective in restoringintravascular volume in the long run, over the immediate resuscitative intervalthey only serve to prevent sodium excretion.

Cardiogenic shock reflects any process wherein inadequate ventricular pumpfunction induces hypoperfusion. Cardiac arrhythmia, ischemia, infarction, valvu-lar dysfunction, or myocardial failure are all capable of inducing cardiogenicshock. Myocardial failure may also be secondary to systemic metabolic processessuch as hypoxia, acidemia, hypothyroidism or hyperthyroidism. Only a smallpercentage of subjects with cardiogenic shock will demonstrate increased cardiacoutput in response to isolated inotropic during administration. Thus, patientsdemonstrating cardiogenic shock need to have their etiology quickly and accu-rately identified if appropriate etiology-specific therapy is to be given.

Distributive shock represents a state in which the body looses its ability toautoregulate blood flow. Loss of vascular integrity often occurs in response tosepsis (malignant intravascular inflammation), adrenal cortical insufficiency,hypocalcemia and with the infusion of vasodilator drugs. Presently, physician’sinterest in distributive shock focuses on systemic inflammatory states, such assevere sepsis, burns, pancreatitis and trauma. In these conditions, cytokine acti-vation and release into the systemic circulation commonly occurs and inducesa generalized inflammatory response characterized by up-regulation of vascularendothelial inducible nitric oxide synthase (iNOS), release of arachidonic acidmetabolites and primary vascular smooth muscle dysfunction [4–6]. Moreover,localized production of nitric oxide, O2 radicals, prostaglandins, platelet aggre-gating factor, and up-regulation of CD11b/CD18 receptors on leukocytes leadto loss of capillary integrity, adherence of circulating leukocytes to the vascularendothelium, and localized disruption of the microcirculation [4, 6, 8–10].Importantly, one needs not evoke infection to see this systemic response. Severetrauma also displays a similar sustained inflammatory mediator activation state.

Non-cardiogenic obstructive shock includes all causes of diminished cardiacoutput secondary to compression on the vascular system or obstruction to bloodflow other than from valvular disease, such as tension pneumothorax, pericardialtamponade, constrictive pericarditis, pulmonary embolism, and acute pulmonary

Pinsky 96

hypertension with acute cor pulmonale. Pulmonary embolism and acute pul-monary hypertension deserve special mention, since both cause shock by induc-ing right heart failure and diminished left ventricular end diastolic volume. Anacute rise in pulmonary artery pressure from massive pulmonary emboli orhypercarbic respiratory failure leads to dilation of the right ventricle. This inturn displaces the intraventricular septum toward the left ventricle, and by theprocess of ventricular interdependence, decreases left ventricular diastolic com-pliance. The resulting decrease in preload leads to diminished cardiac output.Thus, treating such patients with fluid resuscitation will induce only more rightventricular dilation and worsen the obstructive state.

Goals of Therapy

The individual who is in shock may exhibit a wide range of clinical signs(associated with increased sympathetic tone). There is no universal set of vari-ables that may be assessed to determine the presence or absence of shock. Thedetermination of shock is often made clinically from the history and physicalexamination. Patients often present with tachypnea and dyspnea as early signsof cardiovascular deterioration. They may appear pale, diaphoretic, with cold,clammy skin, or may appear warm, erythematous, and dry. The blood pressuremay be normal, elevated, or low, and the heart rate may be tachycardic, brady-cardic, and rarely normal. Urine output is often diminished, and mentationimpaired. On laboratory evaluation, a metabolic acidosis may be present withelevation of serum lactate, excesses chloride loading or be unexplained [11, 12].Except in acute hypovolemic shock the metabolic acidosis is in excess to thatexplained by anaerobic metabolism (lactic acidosis) alone. However, this isneither sensitive nor specific for the presence of shock [13]. The arterial blood gasanalysis may reveal a respiratory alkalosis, metabolic acidosis or a combinationof the two. Since the signs of circulatory shock can be non-specific, it followsthat the end-points of therapy may also be difficult to define.

Resuscitation of shock can be divided into primary and secondary periods.The primary period is the time from initial evaluation through the first round ofresuscitation. The goals during this period are cardio-pulmonary-cerebral resus-citation [14]. Basic and advanced life support principles are utilized with theinitial goal being attainment of an adequate coronary and cerebral perfusionpressure and perfusion of the tissues with oxygenated blood. This encompassesestablishment of an adequate airway and, if necessary, mechanical ventilation,restoration of productive cardiac rhythm and forward blood flow, and attain-ment of a mean arterial blood pressure �60 mm Hg [15]. Without reaching thisimmediate goal, all other resuscitative goals are of questionable value and thus


should not be considered alone. Once mean arterial blood pressure is adequateto maintain cerebral and myocardial perfusion, then the secondary period ofresuscitation begins. The goals of this period are: (1) establishment of an adequateorgan perfusion pressure for all organs; (2) establishment of adequate organblood flow, and (3) establishment of adequate O2 transport to the metabolicallyactive tissues. The first two goals are reached by utilizing volume expansion andvasoactive agents, often using data acquired by invasive hemodynamic monitor-ing via a pulmonary artery (PA) catheter. Although the utility of the PA catheterin treating patients in shock has been questioned [16], data derived from the PAcatheter is often critical in establishing the correct etiological diagnosis.

Improving O2 content, cardiac output and vascular responsiveness accom-plish the third goal in acute resuscitation. Delays in resuscitation of the previouslyhealthy subject now in shock may render them not responsive to aggressive ther-apy later on, in an analogy to the irreversible shock model described by Wiggersin the 1920s.

In support of this hypothesis is the recent large single center clinical trialof early goal-directed therapy for the management of severe sepsis. This studydocumented a marked improved outcome when subjects were aggressivelytreated for circulatory shock in the Emergency Department during the initial 6 hof hospitalization rather than waiting for them to be transferred to the ICU forbetter monitoring [17]. That study focused on rapidly achieving not only anadequate mean arterial pressure, central venous pressure and urine output, butalso an adequate degree of tissue perfusion, as assessed by superior vena cavalSO2 (SsvcO2). This study is in stark contract to the negative results of numerouslarge studies that aimed to improve survival for subjects once transferred intothe ICU using similar resuscitation end-points [18–21]. These negative studiesdo not mean that resuscitation is ineffective to support life in the patient inshock. They merely demonstrate that there is no specific level of TO2 or mixedvenous O2 saturation that one must attain to insure a good outcome [16, 22].Clearly, what are needed are real time measures of tissue wellness and metabolicfunction, such as gastric tonometry [23], or functional measures of cardiovas-cular responsiveness [24]. Unfortunately, such non-invasive metabolic monitoringdevices have yet to be validated as superior to non-specific measures presentlyavailable [25]. Potentially, the lack of documented benefit of these many non-invasive measures reflects inadequate study design.

Hemoglobin and the Optimal Hematocrit

Assuming that one has established an adequate mean arterial pressure, thenone needs to insure that O2 transport (TO2) out of the heart to the body is adequate

Pinsky 98

to meet the metabolic demands. Traditionally, resuscitation guidelines havemirrored the determinants of TO2. TO2 is equal to the product of cardiac outputand arterial O2 content (CaO2). CaO2 in turn is equal to the amount of O2

adsorbed onto hemoglobin and dissolved in the plasma. Since 1,200 times moreO2 is carried on hemoglobin than dissolved in the plasma one usually ignoresthe plasma component of TO2.

The ideal hemoglobin concentration is unknown, but based on experiencewith Jehovah’s Witnesses, it seems the lowest limit tolerable below which evenotherwise healthy people suffer cardiovascular collapse is 2–5 g/dl [26–29].Many patients have done well during and after surgery despite very low hemo-globin and hematocrit levels as long as they can maintain adequate tissue per-fusion and cardiac output. This infers that the body is able to augment cardiacoutput to sustain TO2. All else being equal, there seems to be little evidence toresuscitate a stable patient with a hematocrit �20% with red blood cells [30].Rather, packed red blood cells (PRBCs) should be transfused to specificallycorrect evidence of tissue ischemia or hypoxia or to reverse a high output-induced myocardial ischemia or failure state. Animal data suggest the maximumsystemic extraction ratio for O2 is about 70%, with each organ having its ownO2 requirements [31]. If O2 transport fails to meet this minimum, then O2

consumption will be dependent on organ blood flow [12, 31, 32]. Although we donot know the optimal hemoglobin concentration, aiming for adequate organperfusion and O2 delivery with PRBC transfusions is reasonable up to a hemo-globin concentration of about 7 gm/dl [33, 34]. Above this, any additional benefitin O2 delivery is offset by increases in blood viscosity. Moreover, stored bloodis depleted of 2,3-diphosphoglycerate, causing it to be very oxygen-avid, furtherlimiting O2 utilization [35].

Fluid Resuscitation

Patients presenting with most forms of shock, excluding acute left ven-tricular failure with cardiogenic pulmonary edema, are initially responsive tointravascular volume replacement. Initial resuscitation, therefore should includeintravascular volume replacement as part of a diagnostic and treatment strategy.For example, failure to show any significant increase in mean arterial pressureafter rapid bolus infusion of crystalloid despite an increase in right atrial pres-sure usually signifies the need for vasoactive drug support. An adequate fluidchallenge is defined as one in which either end-organ blood pressure or cardiacoutput increases (volume responsive); or in which either heart rate decreases orleft ventricular filling pressure or central venous pressure increases withoutchanges in systemic blood flow measurements (volume resistant).


The choice of crystalloid or colloid solutions to treat a patient in shock isbased more on religion than on science. Crystalloids are solutions that containsodium as their major osmotically active particle [36]. Lactated Ringer’s andnormal saline (0.9% NaCl) are examples of frequently used isotonic crystalloids.However, normal saline is still slightly hypertonic and carries a profoundlyelevated chloride content relative to intravascular fluid. Hyperchloremic meta-bolic acidosis is a common occurrence in the ICU and is usually due to salineresuscitation. It is unclear, however, if the hyperchloremic metabolic acidosisinduced by normal saline adversely affects outcome. Colloids are fluids with largemolecular weight substances that do not readily pass across capillary walls [36].Examples include albumin solutions, dextran, and hetastarch. Hetastarch gelatinsare carried in a normal saline vehicle, thus they share the chloride load prob-lems with normal saline. However, since one often gives less colloid thancrystalloid, this concern is less important. Newer gelatins include put hetastarchin a lactated Ringer’s-like vehicle, thus minimizing the chloride loading.

Recent interest in excess mortality in subjects given albumin has arisen.However, a large double blind multicenter trail of crystalloid versus albumin ispresently underway in Australia to address this and other issues. Thus, conclu-sions about the risks and benefits of albumin infusions will need to wait untilthis trial is concluded. Infused albumin has a plasma half-life of approximatelysixteen hours, but its effect lasts for approximately twenty-four hours. 25%albumin expands the vascular volume by translocation of interstitial fluid, suchthat for each 100 ml of 25% albumin given, the vascular volume will increaseby 450 ml. Other albumin containing solutions have a near-physiologic colloidosmotic pressure, and therefore expand the vascular volume by 1 ml for eachml given.

Dextran is a synthetic colloid that is rarely used because of its associatedside effects [37, 38]. Hetastarch is also a synthetic colloid solution. It comes asa 6% solution in normal saline (Hesban) or lactated Ringers (Hextend) andfunctions similarly to 5% albumin solutions. It has a very long plasma half-life(17 days) and unlike dextran, is non-immunogenic [39]. Potential problemswith hetastarch infusions arise when massive resuscitation is given (�1,500 ml)wherein one may see an elevation of serum amylase, osmotic diuresis, and increasebleeding tendency [39].

Colloids have several potential benefits, such as greater sustained intra-vascular volume presence and less edema formation, while being much moreexpensive. Despite all the potential benefits of using colloid as part of resusci-tation, there is no proven benefit in outcome. A meta-analysis of eight studiesin sepsis showed an overall 5.7% relative decrease in mortality in patients resus-citated with crystalloids alone, and a 12.3% difference in favor of crystalloid intrauma patients. A 7.8% difference was found in favor of colloid use in medical

Pinsky 100

patients [40]. Due to the relatively few studies included in this analysis, it isdifficult to make any conclusions regarding benefit in survival by use of colloidversus crystalloid.

Vasopressors

Since the goal of the primary resuscitation period is to obtain an adequatecerebral and coronary perfusion pressure, if this cannot be done with volumetherapy alone, or if the blood pressure is profoundly depressed, then use of avasoactive agent is required. These agents may also be necessary during thesecondary resuscitation period to maintain adequate organ perfusion pressureand cardiac output in the setting of inflammatory mediator-induced pathologicalvasodilation. Selective use of agents may augment blood pressure and cardiacoutput to achieve the desired goals, but which agent is used often does not matteras long as the desired goals are achieved.

Vasopressors induce their response by stimulating the �-adrenergic recep-tors on vascular smooth muscle cells. �-Adrenergic receptor stimulationaugments contractility and induces vasodilation and tachycardia. Several vaso-pressor agents are available with varying degrees of �-, �- and vasopressin-2receptor activity. Phenylephrine is a pure �-agonist and isoproterenol is a pure�-agonist [41, 42]. All other vasoactive agents have varying degrees of effecton both � and � types of receptors. The most potent vasoactive agent is epi-nephrine, since this will stimulate all adrenergic receptors. Dopamine is oftenused in states of shock to preserve renal function; however, no study hasdemonstrated a proven protective effect of dopamine on renal function [43, 44]and a recent large multicenter trail powered to show no effect documented or noeffect of low dose dopamine on the development of renal failure in postopera-tive surgical ICU patients [45]. In septic shock, dopamine has not been shownsuperior to other agents such as dobutamine and norepinephrine in obtainingtarget values for O2 delivery and O2 consumption [46]. Furthermore, dopamineincreased splanchnic O2 requirements over that of norepinephrine [47]. Finally,other agents such as dopexamine may produce similar hemodynamic profilesand renal preservation in patients with reduced cardiac index following coronaryartery bypass surgery [48]. Furthermore, in a retrospective study of similarlymatched subjects given norepinephrine versus all other vasopressor, Martin etal. [49] demonstrated that those subjects treated with norepinephrine had amarkedly reduced mortality.

Sepsis is associated with a reduction in vasopressin-2 receptor activity,making vascular smooth muscle cells less responsive to �-adrenergic stimulation.Infusions of low does vasopressin may reverse this desensitization. However,


higher does of vasopressin induce profound splanchnic vasoconstriction. Thus,it is not clear that vasopressin infusions will improve outcome in distributiveshock states. A large multicenter clinical trial of vasopressin is on going. Weshall need to await the results of this trial before making any conclusions aboutthe use of vasopressin in shock.

What Are the Targets for Cardiovascular Resuscitation?

Based on the clinical trials of early goal-directed therapy, vasopressor-induced ischemia and inflammation and unclear benefits from hemoglobin-based resuscitation the following approach seems reasonable. First, in thehypotensive patient restoration of a mean arterial pressure to �60 mm Hg in apreviously normotensive subject by the rapid infusion of volume and, if needed,vasopressor, is indicated to prevent and/or reverse cerebral and coronaryischemia. Second, rapid diagnosis of the etiology of shock should be undertaken.Except in the setting of obstructive or cardiogenic shock, volume resuscitationis the mainstay of initial therapy. The choice of fluid is optional, but lactatedRinger’s solution seems like a reasonable choice if blood products are not goingto be infused simultaneously. The choice of vasopressor agent is still notdefined, but norepinephrine is gaining increased usage because of its therapeu-tic profile. In the setting of cardiogenic or obstructive shock, although inotropicagents may transiently increase cardiac output, every effort needs to be done tomake the diagnosis and treat the primary problem because few of the etiologiesthat induce either of these two types of shock respond to inotropic drug infusionalone. The infusion of packed red blood cells to treat patients with concomitantanemia and circulatory shock is problematic. In the setting of reduced cardiacreserve, red blood cells may be life saving, whereas in the otherwise healthypatient, there is little data to defend transfusing to a hematocrit above 25%.Clearly, the attempt in a resuscitated ICU patient to sustain TO2 at some definedelevated level is not justified.

Conclusion

Resuscitation from circulatory and respiratory failure represent mainstaysof emergency and critical care management. Importantly, no amount of resus-citative effort will be successful in promoting patient survival if the primaryreason for the shock state is not identified and treated, independent of resusci-tation. Having said that, aggressive resuscitation to normal functional levelsof blood flow and organ perfusion pressure during the first 6 h following the

Pinsky 102

development of shock improves outcome both in patients with trauma or sepsis.However, clinical studies have demonstrated that restoration of total blood flowto supranormal levels in subjects with established shock that has been presentfor over 6 h does not improve survival. Still, some defined clinical targets areessential in these patients as well to prevent further organ injury due to ischemiaand its associated inflammatory response. Thus, the rapid restoration of normalhemodynamics by conventional means, including fluid resuscitation and surgicalrepair, results in a better long-term outcome than inadequate or delayed resus-citative efforts. Clear initial targets for resuscitation are a mean arterial pressure�65 mm Hg, and a cardiac output and O2 transport to the body adequateenough to prevent tissue hypoperfusion. The level of cardiac output needed toachieve this goal is probably different among subjects and within subjects overtime. Indirect signposts of adequate perfusion, such as venous O2 saturation,mentation, urine output and local measures of tissue blood flow are useful inmonitoring this response.

Acknowledgement

This work was supported in part by the federal government (GM61992–03, NHLBI K-24 HL67181–01A2, and NRSA 2-T32 HL07820–06).

References

1 Hinshaw LB, Cox BG: The Fundamental Mechanisms of Shock, Plenum Press, New York, 1972. 2 Calcagni DE, Bircher NG, Pretto E: Resuscitation: Blood, blood component, and fluid therapy; in

Grande CM, et al (eds): Textbook of Trauma Anesthesia and Critical Care. St. Louis, Mosby-YearBook, 1993, p 400.

3 Rush BF: Irreversibility in post-transfusion phase of hemorrhagic shock. Adv Exp Med Biol 1971;23:215–221.

4 Calandra T, Glauser MP: Cytokines and septic shock. Diagn Microbiol Infect Dis 1990;13: 377–381.5 Mason JW, et al: Plasma kallikrein and Hageman factor in gram-negative bacteremia. Ann Intern

Med 1970;73:545–551.6 Pinsky MR, Vincent JL, et al: Serum cytokine levels in human septic shock: Relation to multiple-

system organ failure and mortality. Chest 1993;103:565–575.7 Rosenbloom AJ, Pinsky MR, et al: Leukocyte activation in the peripheral blood of patients with

cirrhosis of the liver and SIRS: Correlation with serum interleukin-6 levels and organ dysfunction.JAMA 1995;274:58–65.

8 Salin ML, McCord JM: Free radicals and inflammation: Protection of phagocytosing leukocytesby superoxide dismutase. J Clin Invest 1975;56:1319–1323.

9 Bernard GR, et al: Prostacyclin and thromboxane A2 formation is increased in human sepsis syn-drome. Am Rev Resp Dis 1991;144:1095–1101.

10 Snyder SH, Bredt DS: Biologic roles of nitric oxide. Scient Am 1992;266:68–77.11 Cilley RE, et al: Low oxygen delivery produced by anemia, hypoxia, and low cardiac output.

J Surg Res 1991;51:425–433.


12 Cain SM: Oxygen delivery and uptake in dogs during anemic and hypoxic hypoxia. J Appl Physiol1977;42:228–234.

13 Cohen RD, Woods F: Lactic acidosis revisited. Diabetes 1983;32:181–191.14 Safar P: Cerebral resuscitation after cardiac arrest: Research initiatives and future directions. Ann

Emerg Med 1993;22:324–349.15 Fazekas JF, Kleh J, Parrish AE: The influence of shock on cerebral hemodynamics and metabolism.

Am J Med Sci 1955;229:41–45.16 Sandham JD, Hull RD, Brant RF, Knox L, Pineo GF, Doig CJ, Laporta DP, Viner S, Passerini L,

Devitt H, Kirby A, Jacka M, and Canadian Critical Care Clinical Trials Group: A Randomized,Controlled Trial of the Use of Pulmonary-Artery Catheters in High-Risk Surgical Patients. N EnglJ Med 2003;348:5–14.

17 Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, Peterson E, Tomlanovich E: Earlygoal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001;345:1368–1377.

18 Hayes MA, Timmins AC, Yau EH, Palazzo M, Hindo CJ, Watson D: Evaluation of systemic oxygendelivery in the treatment of the critically ill. N Engl J Med 1994;330:1717–1722.

19 Gattinoni L, Brazzi L, Pelosi P, Latini R, Tognoni G, Pesenti A, Fumagalli R: A trial of goal-orientedhemodynamic therapy in critically ill patients. N Engl J Med 1995;333:1025–1032.

20 McKinley BA, Kozar RA, Cocanour CS, Valdiva A, Sailors RM, Ware DN, Moore FA: Normal ver-sus supranormal oxygen delivery goals in shock resuscitation: The response is the same. J Trauma2002;53:825–832.

21 Velmahos GC, Demetriades D, Shoemaker WC, Chan LS, Tatevossian R, Wo CCJ, Vassiliu P,Cornwell III EE, Murray JA, Roth B, Belzberg H, Asensio JA, Berne TV: Endpoints of resuscitationof critically injured patients: Normal or supranormal? A prospective randomized trial. Ann Surg2000;232:409–418.

22 Miller MJ, Cook W, Mithoefer J: Limitations of the use of mixed venous pO2 as an indicator of tissuehypoxia. Clin Res 1979;27:401A.

23 Gutierrez G, Palizas F, Doglio G, et al: Gastric intramucosal pH as a therapeutic index of tissueoxygenation in critically ill patients. Lancet 1992;339:195–199.

24 Pinsky MR: Functional hemodynamic monitoring: Asking the right question. Intensive Care Med2002;28:386–388.

25 Pinsky MR: Beyond global oxygen supply-demand relations: In search of measures of dysoxia. IntensCare Med 1994;20:1–3.

26 Olugbenga A, et al: Management of severe anemia without transfusion in a pediatric Jehovah’sWitness patient. Crit Care Med 1994;22:524–528.

27 Parker RI: Aggressive non-blood product support of Jehovah’s Witnesses. Crit Care Med 1994;22:381–382.

28 Koenig HM, et al: Use of recombinant human erythropoietin in a Jehovah’s Witness. J Clin Anesth1993;5:244–247.

29 Viele MK, Weiskopf RB: What can we learn about the need for transfusion from patients whorefuse blood? The experience with Jehovah’s Witnesses. Transfusion 1994;34:396–401.

30 Hebert PC, Wells G, Blajchman MA, Marshall J, Martin C, Pagliarello G, Tweeddale M,Schweitzer I, Yetsir E: A multicenter, randomized, controlled clinical trial of transfusion require-ments in critical care. Transfusion Requirements in Critical Care Investigators, Canadian CriticalCare Trials Group. N Engl J Med 1999;340:409–417.

31 Schlichtig R, et al: Tolerance of unusually low mixed venous oxygen saturation: Adaptations inthe chronic low cardiac output syndrome. Am J Med 1986;80:813–818.

32 Schwartz S, et al: Sequential hemodynamic and oxygen transport responses in hypovolemia, andhypoxia. Am J Physiol 1981;241:H864–H871.

33 Welch HG, et al: Prudent strategies for elective red blood cell transfusion. Ann Intern Med1992;116:393–402.

34 Carson JL, et al: Severity of anemia and operative mortality and morbidity. Lancet 1988;i:727–729.

35 Weisel RD, et al: Adverse effects of transfusion therapy during abdominal aortic aneurysmectomy.Surgery 1978;83:682–690.

Pinsky 104

36 Rainey TG, Read CA: Pharmacology of colloids and crystalloids; in Chernow B (ed): ThePharmacologic Approach to the Critically Ill Patient, ed 3. Baltimore, Williams & Wilkins, 1994,pp 272–290.

37 Imm A, Carlson RW: Fluid resuscitation in circulatory shock. Crit Care Clins 1993;9:313–333.38 Thoren L: Dextran as a plasma volume substitute. Prog Clin Biol Res 1978;19:265–282.39 Waxman K, et al: Blood and plasma substitutes: Plasma expansion and oxygen transport properties.

World J Med 1985;143:202–206.40 Velanovich V: Crystalloid versus colloid fluid resuscitation: A meta-analysis of mortality. Surgery

1989;105:65–71.41 Bisonni RS, et al: Colloids versus crystalloids in fluid resuscitation: An analysis of randomized

control trials. J Fam Pract 1991;32:387–390.42 Almog Y, Breslow MJ: A rational approach to using vasopressors in the ICU. J Crit Ill 1995;10:

171–183.43 Duke GJ, Bersten AD: Dopamine and renal salvage in the critically ill patient. Anaesth Intens Care

1992;20:277–302.44 Duke GJ, Briedis JH, Weaver RA: Renal support in critically ill patients: Low-dose dopamine or

low-dose dobutamine? Crit Care Med 1994;22:1919–1925.45 Bellomo R, Chapman M, Finfer S, Hickling K, Myburgh J: Low-dose dopamine in patients with

early renal dysfunction: A placebo-controlled randomised trial. Australian and New ZealandIntensive Care Society (ANZICS) Clinical Trials Group. Lancet 2000;356:2139–2143.

46 Hannemann L, et al: Comparison of dopamine to dobutamine and norepinephrine for oxygendelivery and uptake in septic shock. Crit Care Med 1995;23:1962–1970.

47 Marik PE, Mohedin M: The contrasting effects of dopamine and norepinephrine on systemic andsplanchnic oxygen utilization in hyperdynamic sepsis. JAMA 1994;272:1354–1357.

48 Olsen NV, Lund J, et al: Dopamine, dobutamine, and dopexamine: A comparison of renal effectsin unanesthetized human volunteers. Anesthesiology 1993;79:685–694.

49 Martin C, Viviand X, Leone M, Thirion X: Effect of norepinephrine on the outcome of septicshock. Crit Care Med 2000;28:2758–2765.

Michael R. Pinsky, MD606 Scaife Hall, 3550 Terrace Street, Pittsburgh, PA 15261 (USA)Tel. �1 412 647 5387, Fax �1 412 647 8060, E-Mail [email protected]


Intravenous Fluids and Acid-BaseBalance

Rinaldo Bellomoa, Toshio Nakab, Ian Baldwinc

aDepartment of Intensive Care and Department of Surgery, bAustin Hospital,University of Melbourne, and cRMIT University, Melbourne, Australia

The administration of intravenous fluids is one of the most common treat-ments given in the Intensive Care Unit. The rationale for such treatment is thatcritically ill patients have fluid losses that they cannot compensate for, that theyhave relative or absolute reductions in circulating blood volume, which in turnresults in decreased organ flow, and that such organ ischemia is responsible forincreased morbidity and mortality. According to this biologic rationale, ongo-ing fluid losses should be replaced to maintain fluid homeostasis and relative orabsolute deficiencies in effective circulating blood volume should be preventedor rapidly corrected. There is also agreement that insensible fluid losses andisotonic fluid losses through urine or from other body cavities should bereplaced with a judicious mixture of water and crystalloid solutions. All intra-venous fluids, however, have a profound impact on acid-base physiology andacid-base balance that is not widely acknowledged. Furthermore, there is con-troversy on the biochemical, physiological and clinical meaning of their acid-base effects. Critical care physicians need to appreciate such controversy inorder to understand the physiological consequences of their choices.Accordingly, this article focuses on the acid-base effects of intravenous fluids,the mechanisms responsible for such effects, the possible clinical consequencesof intravenous fluids-induced changes in acid-base balance and the implicationsof the available evidence in this field of medicine for clinicians.

Crystalloid Solutions and Acid-Base Balance

Crystalloid solutions contain water and electrolytes only. They haveosmotic pressure but lack oncotic pressure. They are easily prepared and cheap.

Bellomo/Naka/Baldwin 106

Although they are typically seen as a uniform group of solutions with little toseparate them from one another (unlike colloids), this view, particularly fromthe acid-base balance point of view, is simplistic.

There are, in fact, profound physiological differences between the twomajor groups of near-isotonic crystalloid solutions: the chloride-rich solu-tion and the so-called ‘balanced’ solutions (i.e. normal saline vs. Plasmalyte,Hartmann’s solution or Ringer’s lactate solution) especially in terms of theiracid-base effect during massive resuscitation [1] and there are even more pro-found differences between isotonic and hypertonic crystalloid solutions [2]. Inthis manuscript we will focus on near-isotonic solutions. Therefore, when dis-cussing crystalloids, even though normal saline is seen as the prototype, oneshould, firstly of all be aware of such differences.

The major difference between these crystalloids lies in the different chlo-ride content or, as we shall see later, their so-called ‘strong ion difference’.Saline solutions, of course, have an ‘unphysiologic’ or ‘supranormal’ concen-tration of chloride relative to sodium (i.e. the chloride concentration is equal tothat of sodium) while the other ‘buffered’ solutions (lactate fluids such asRinger’s in the USA and Hartmann’s in Europe or fluids with other buffers suchas acetate or gluconate as in the case of plasmalyte) have a concentration ofchloride relative to sodium, which more closely approximates that seen inplasma water (typically 38 mmol/l less than sodium). To use more recent acid-base physiology ideology [3–6], saline and saline-derived fluids have a strongion difference of 0 mEq/l, while ‘buffered’ fluids more closely approximatethe physiological strong ion difference of 42 mEq/l. In order to understand thephysiological significance of the strong ion difference of a particular fluid,one needs to understand the theory behind the Stewart-Figge approach to acid-base analysis.

Acid-Base Analysis Using the Stewart-Figge Methodology

Recently, quantitative acid-base analysis using the Stewart-Figge’sapproach [3, 4] has been introduced as a more useful way of approaching theeffects of intravenous fluids on acid-base balance.

This method first involves first calculating the apparent strong ion differ-ence (SIDa) (all concentrations in mEq/l):

SIDa � [Na�] � [K�] � [Mg2�] � [Ca2�] – [Cl�] – [lactate]

and then the taking into account the role of weak acids (CO2, albumin, andphosphate) in the balance of electrical charges in plasma water.

Intravenous Fluids and Acid-Base Balance 107

The weal acid component is expressed through the calculation of the effec-tive strong ion difference (SIDe; pCO2 in mm Hg, albumin in g/l and phosphatein mmol/l):

SIDe � 1,000 � 2.46 � 10�11 � pCO2/(10�pH)� [Alb] � (0.12 � pH � 0.631) �[Phos] � (0.309 � pH � 0.469).

Once weak acids are quantitatively taken into account, the SIDa to SIDedifference should equal zero unless there are unmeasured charges (anions). Suchcharges are then described by the strong ion gap (SIG): SIG � SIDa � SIDe.

The component of albumin and phosphate is defined as the total concen-tration of non-volatile weak acid (Atot). [Atot], along with SID and pCO2, is anindependent determinant of [H�] or pH. According to the Stewart-Figgeapproach, metabolic acidosis can then result from a reduction in the SID or froman increase in Atot, and respiratory acidosis can result from a gain in pCO2. Thechanges in each of these variables can be quantified to express how much eachone is responsible (in mEq/l) for the findings on blood analysis. According tothis paradigm, three forces are responsible for acid-base status: the SID, the CO2

tension and Atot. Bicarbonate concentration, hydrogen ion activity and, there-fore, pH are dependent on the above three forces. The source of hydrogen ionsis water, which becomes more dissociated as the SID falls, or the pCO2 rises orthe Atot rises. According to this paradigm, the administration of fluids with aSID of 0 should change the SID in plasma downward, thereby inducing a iatro-genic low SID-induced, non-anion gap, non lactic acidosis due to increased dis-sociation of plasma water and release of hydrogen ions [6]. On the other hand,the administration of fluids with a near normal SID should have a minimal or noimpact on the SID of plasma and, thereby, induce no acidosis [6].

As we shall discuss later, events are actually somewhat more compli-cated, especially when the system is tested to the extreme during continuoushemofiltration.

The Conventional Ideology

There is no argument with the fact that intravenous fluids induce a meta-bolic acidosis, as this observation has been made innumerable times [7–9].However, there is much controversy on why this phenomenon occurs. As wehave seen above, one possible explanation is that provided by the Stewart-Figgeparadigm. Another is provided by the conventional understanding of acid-basephysiology, which has dominated teaching for decades and still dominatesworldwide. This conventional view of how fluids might affect acid-base status


states that chloride-rich fluids induce acidosis simply by ‘diluting’ the buffers(esp. bicarbonate); conversely, the administration of ‘buffered’ solutions pro-vides a metabolic source for bicarbonate generation such that the dilutionaleffect of fluids is compensated for by the transformation of lactate or acetateinto bicarbonate and the restoration of a physiological concentration of bicar-bonate. Although in the opinion of the authors, this approach does not satisfac-torily explain a variety of clinical observations, no experiments have beendevised so far to allow researchers to demonstrate that one explanation is ‘true’or ‘false’.

The Biochemical Consequences

Irrespective of the mechanisms responsible for intravenous fluid inducedacidosis, it is important for clinicians to understand the biochemical character-istics typically seen in this setting. Such understanding will avoid incorrectdiagnoses and treatments, especially in the critically ill where these disordersare complex [10, 11] or in those patients with acute renal failure, where thatcomplexity increases [12, 13]. When crystalloid solutions with a SID of 0 aregiven intravenously in sufficient quantities and rapidly enough, patients willdevelop the following biochemical changes: an increase in serum chloride con-centration (this increase can be of up to 10 mmol/l if sufficient amounts aregiven), a mild increase in serum sodium (typically 1–2 mmol/l) and a decreasein serum albumin concentration (typically down to 20–25 g/l). As a conse-quence of these changes, a Base Deficit develops of about 4–5 mEq/l (BaseExcess of �4 or �5 mEq/l). This metabolic acidosis is associated with a fall inserum bicarbonate of about 4–5 mmol/l and a normal anion gap and is mild tomoderate in severity. The expected effect of a SID change of about 10 mEq/l(from 42 to 32–34 mEq/l) is attenuated by the alkalinizing effects of hypoalbu-minemia (decreased Atot). This acid-base picture can be quantitatively pre-dicted and neatly demonstrated when cardiac surgery patients becomeconnected with the cardiopulmonary bypass pump and the pump has beenprimed with a solution containing a SID of 0. In this setting, as the pump primeenters the circulation and is distributed across the plasma volume, all of theabove changes occur almost instantaneously [14–16] (fig. 1–4).

The Physiological Consequences of Iatrogenic Acidosis

Until recently, there was limited evidence to suggest a physiologicaleffect for this mild to moderate iatrogenic hyperchloremic acidosis. However,


recent experimental work has sought to investigate whether different kinds ofacidoses induce different physiological changes. Studies in rats by Kellum’sgroup have shown that hyperchloremic acidosis even of a moderate degree(5 mEq/l) decreases blood pressure and increases nitrate/nitrite levels. If severe(base deficit of 15 mEq/l), hyperchloremic acidosis induced hypotension but didnot directly affect nitrate/nitrite levels [17]. The same group compared iatrogeniclactic acidosis to hyperchloremic acidosis and found that these two forms of

�3

�2.5

�2

�1.5

�1

�0.5

0

0.5

1

Pre-CPB Post-CPB

mE

q/l

Base excess

Fig. 1. Change in base excess from baseline (pre-CPB) to immediately after initiationof cardiopulmonary bypass (post-CPB) with a pump prime that has an SID of 0.

Fig. 2. Change in bicarbonate from baseline (pre-CPB) to immediately after initiationof cardiopulmonary bypass (post-CPB) with a pump prime that has an SID of 0.

19

20

21

22

23

24

25

Pre-CPB Post-CPB

mm

ol/l

Bicarbonate


acidosis led to quite markedly different physiological immune responses: hyper-chloremic is pro-inflammatory (increased NO release, increased Il-6:IL10 ratiosand increased NFkB DNA binding) while lactic acidosis is anti-inflammatory(decreased NO release, increased IL-10 levels and decreased NFkB DNAbinding) [18]. Thus, the choice of intravenous fluids appears to not only affectacid-base status but also the patients’ immune response.

Fig. 3. Change in SID from baseline (pre-CPB) to immediately after initiation ofcardiopulmonary (post-CPB) bypass with a pump prime that has an SID of 0.

Fig. 4. Change in serum albumin concentration from baseline (pre-CPB) to immedi-ately after initiation of cardiopulmonary bypass (post-CPB) with a pump prime that has anSID of 0.

33

34

35

36

37

38

39

Pre-CPB Post-CPB

SID

mE

q/l

0

5

10

15

20

25

30

35

Pre-CPB Post-CPB

g/l

Albumin


The Clinical Implications

As describe above, iatrogenic acidosis clearly has biochemical and physi-ological effects. What remains unclear is whether it has clinical consequences.Nonetheless, evidence exists to suggest that, when clinicians do not appreciatethat the acidosis that is developing, for example, in a trauma patient receivinglarge amounts of saline is iatrogenic, they may be believe that the patient hasinternal bleeding or vital organ ischemia. They may then proceed to invasiveinterventions or to further unnecessary fluid administration or both, which cancause major injury to the patient. Whether the administration of a particularkind of fluid instead of another results in increased survival will have to waitfor the official release of the results of the SAFE trial comparing fluid resusci-tation with albumin to fluid resuscitation with saline in ICU patients [19]. Theresults of this trial should become available in 2004.

Colloids

Colloid solutions always contain water and electrolytes. They are typicallyisotonic. They also contain a class of agents, which because of their size conferoncotic pressure to the solution. Such agents can be natural (albumin) or syn-thetic (dextran, starches, gelatins). They can be iso-oncotic or hyperoncotic(20% albumin). These agents are diverse in terms of molecular size, molecularproperties, pharmacokinetics, pharmacodynamics, hemodynamic effect, sideeffects and acid-base consequences. Their common property in respect to crys-talloids is due to the oncotic pressure effect, which induces greater short-term(hours) intravascular permanence of the administered fluid. Accordingly, usingthe paradigms presented earlier in this article, similar acid-base effects could beexpected to those seen with crystalloids containing different values of strongion difference except for two additional considerations: the specific effect ofthe colloidal solutes on acid-base balance and the additional effect of greaterinitial intravascular permanence, which might translate into a more pronouncedeffect on acid-base physiology immediately after infusion [14].

The Controversy

Given all of these options (crystalloids with a SID of 0, crystalloids with aSID close to the physiological level, colloids with variable SID and variablecontents of charged ions) there has been controversy from the start on whichfluids to choose to resuscitate patients. Firstly, a crystalloid school (typically


trauma-based and in North American centers) and a colloid school (typicallyamong those dealing with older patients and in European centers) have beenvehemently arguing about the merits of one or the other approach for more than50 years [20, 21]. Secondly, more recently the debate has gained the additionaldimension of chloride-rich vs. chloride-poor or unphysiologic SID vs. physio-logic SID fluids.

Most information available at this time comes from comparisons of twoclasses of fluids in terms of their physiological effects. Furthermore, only verysmall, grossly statistically under-powered studies have reported on mortalityand major morbidity.

In response to these deficiencies in the data available, meta-analyses havebeen conducted first comparing mortality with crystalloids to mortality withcolloids [20] and then comparing albumin alone to crystalloids [21]. However,because these meta-analyses are based upon small, often poorly conductedtrials with physiological outcome as the primary outcome, they have beenheavily criticized and their conclusions only represent starting points for futureinvestigations. In terms of acid-base physiology, no randomized studies existcomparing the clinically different effects of using fluids of different SIDs. Thus,clinicians are left with having to make thoughtful decisions in individual casestaking into account the limited information available. At this time and untilfurther information emerges, it seem reasonable to avoid aggravating hyper-chloremic acidosis, when already present, by administering more chloride richfluids. It also seems reasonable to administer a mixture or a choice of fluids thatwould not markedly contribute to metabolic acidosis, if one already exists. Inpatients that already show hyperlactatemia, it seems undesirable to administerlactate-buffered fluids which might aggravate the acidosis or, in the very least,potentially create diagnostic confusion by increasing the blood lactate concen-tration. The choice of colloids or crystalloids remains a matter of debate, whichmay be partly resolved by the publication of the results of a recently completedrandomized controlled trial comparing albumin to saline. This trial might alsoprovide new and important data on the differential acid-base effects of thesetwo solutions and the links between acid-base balance and clinical outcomes.

Effect of Replacement Fluids and Their Composition

The administration of intravenous fluids as fluid replacement duringcontinuous hemofiltration represents a unique situation from the point of viewof the effect of intravenous fluids on acid-base physiology because of severalreasons: (1) up to 6 or even 8 liters may be administered per hour; (2) in mostcases, more than 45 l of fluids are administered each day; (3) fluids of different


compositions are available for such therapy especially from the point of view oftheir ‘buffer’ content [22], and (4) acute renal failure patients already have com-plex acid-base disorders even before the fluids are administered [23].

In this regard, lactate, acetate, and bicarbonate have all been used as‘buffers’ (or SID generators according to Stewart) during RRT. Citrate has beenused as ‘buffer’ and anticoagulation. These ‘buffers’ affect acid-base balance,thus we must understand their physiological characteristics.

Bicarbonate has the major advantage of being the most physiologic anionequivalent. However, the production of a commercially available bicarbonate-based solution is not easy because of the formation of calcium and magnesiumsalts during long-term storage. Furthermore, the cost of this solution is approx-imately three times higher than that of other ‘buffer’ solutions. Accordingly,acetate and lactate have been used widely for fluid replacement therapy. Undernormal conditions, acetate is rapidly converted on a 1:1 basis to CO2 and thenbicarbonate by both liver and skeletal muscle. Lactate is also rapidly convertedin the liver on a 1:1 basis [24, 25].

Studies of acetate-based solutions, however, appear to exert a negativeinfluence on the mean arterial blood pressure and cardiac function in the criti-cally ill [26–29]. Morgera et al. [29] compared acid-base balance betweenacetate- and lactate-buffered replacement fluids and reported that the acetate-buffered solution was associated with a significant lower pH and bicarbonatelevels than the lactate-buffered solution. However, the acetate-buffered solutionhad 9.5 mmol/l less ‘buffer’ than the lactate-buffered solution.

Thomas et al. [30] studied the effects of lactate- versus bicarbonate-buffered fluids. Hemofiltration fluids contained either 44.5 mmol/l of sodiumlactate or 40.0 mmol/l of sodium bicarbonate with 3 mmol/l of lactate (43 mEq/lof buffer). Lactate rose from approximately 2 to 4 mmol/l when lactate-basedfluids were given but not with bicarbonate. Both therapies resulted in a similarimprovement in metabolic acidosis.

Tan et al. [31] also studied the acid-base effect of CVVH with lactate orbicarbonate buffered solution. The lactate-buffered solution had an SID of 46vs. 35 mEq/l for the bicarbonate fluid. From the Stewart-Figge point of view,the lactate-buffered solution should have led to a greater amount of alkalosis.However, this study showed a significant increase in plasma lactate levels inthese critically ill patients (lactate intolerance) and a decrease in base excesswith the lactate-buffered solution. According to the Stewart-Figge paradigm,lactate, if not metabolized and still present in blood, should act as a stronganion, which would have the same acidifying effect of chloride. Accordingly,iatrogenic hyperlactatemia can cause a metabolic acidosis. This effect can, ofcourse, also be explained by the failure to convert exogenous lactate intobicarbonate.


Citrate has been used for regional anticoagulation. During this procedure,citrate is administered to the circuit before the filter and chelates calcium, thusimpeding coagulation. Once citrate enters the circulation, it is metabolized toCO2 and then bicarbonate on a 1:3 basis thus 1 mmol of citrate yields 3 mmolof CO2 and then bicarbonate.

Under these circumstances, citrate acts as the ‘buffer’ as well as the antico-agulant. If the method described by Mehta et al. [32] is applied, approximately48 mmol/h of ‘bicarbonate equivalent’ is given as citrate. This rate of alkaliadministration may result in metabolic alkalosis (up to 25% of cases). Attentionmust be paid in patients with liver disease who may not be able to metabolizecitrate. In these patients, citrate may accumulate and result in severe ionizedhypocalcemia and metabolic acidosis because the citrate anion (C6H5O7

3–) acts asan unmeasured anion and increases the SIG which has acidifying effects.

When oxidizable anions are used in the replacement fluids, the anion(acetate, lactate and citrate) must be completely oxidized to CO2 and H2O inorder to generate bicarbonate. If the metabolic conversion of non-bicarbonateanions proceeds without accumulation, their buffering capacity is equal to thatof bicarbonate. Thus, the effect on acid-base status depends on the ‘buffer’concentration rather than on the kind of ‘buffer’ used. When the metabolicconversion is impaired, the increased blood concentration of the anions leadsto an increased strong anion in lactate or unmeasured anions for acetate andcitrate: all lower the SIDa and acidify blood. The nature and extent of theseacid-base changes is governed by the intensity of plasma water exchange/dialysis, by the ‘buffer’ content of the replacement fluid and by the metabolicrate for these anions.

High-Volume HemofiltrationContinuous hemofiltration has been shown to remove cytokines from the

circulation of septic ARF patients [33]. However, studies of standard intensityCRRT have shown limited effects on physiological variables and outcomes incritically ill patients [34]. For these reasons, there has been a move to so-calledhigh-volume hemofiltration with the goal of increasing the possible beneficialimpact of CRRT. Recently, such high volume hemofiltration (HVHF) has beenapplied to the treatment of ARF with improved survival [35] and in septicshock patients with favorable hemodynamic results [36]. However, if com-mercial lactate-buffered replacement fluid is used during HVHF, patientsmight receive more than 270 mmol/h of exogenous lactate. This lactate loadcould easily overcome endogenous lactate metabolism even in healthy subjects[37] and result in progressive hyperlactatemia. Hyperlactatemia has beenreported during lactate-buffered fluids in critically ill patients with ARF treatedwith intermittent hemofiltration and a lactate load of 190–210 mmol/h [38].


Such hyperlactatemia might induce a metabolic acidosis. Cole et al. [36] stud-ied the effect of HVHF on acid-base balance. HVHF with lactate-bufferedreplacement fluids (6 liters/h of lactate-buffered fluids) induced iatrogenichyperlactatemia. Plasma lactate levels increased from a median 2.51 mmol/l toa median of 7.3 mmol/h at 2 h. This change was accompanied by a significantdecrease in bicarbonate and base excess (fig. 5). However, such hyperlac-tatemia only had a mild and transient acidifying effect. A decrease in chlorideand SIDe and the removal of unmeasured anions (decrease in SIG) all rapidlycompensated for this effect. Thus, the final effect was that HVHF induced onlya minor change in pH from 7.42 to 7.39 at 2 h. In the period from 2 to 8 h, theblood lactate concentration remained stable at around 7–8 mmol/l, while com-pensatory effects continued, which restored bicarbonate levels to 27.2 mmol/land pH to 7.44 by 8 h of treatment.

Although the chloride concentration in the replacement fluid was highcompared to the serum chloride level, a progressive decrease in chloride wasobserved. This might be due to chloride losses in excess of gains. Uchino et al.[37] examined the sieving coefficient for chloride during HVHF and showed asieving coefficient for chloride �1. Another possible explanation forhypochloremia would be the intracellular movement of chloride in response tometabolic acidosis (chloride shift). A decrease in SIDe was explained by theaggregate minor changes in PaCO2, albumin and phosphate. The changes inSIG appeared most likely to be due to simple filtration of unmeasured anion.

Consequently, HVHF with lactate-buffered fluids induced a marked hyper-lactatemia but did not induce a progressive acidosis. However caution should beexerted in particular patients who have marked pretreatment hyperlactatemia

Fig. 5. Changes in lactate, bicarbonate and base excess concentrations from baselinefollowing initiation of high-volume hemofiltration.

�3

�2

�1

0

1

2

3

4

5

6

0 2 6 8

Cha

nge

in m

Eq

/l

Hours of treatment

LactateBicarbonate

Base excess


(�5 mmol/l), liver dysfunction, or where the intensity of HVHF exceeds6 liters/h of plasma water exchange. Bicarbonate use is warranted in suchpatients.

Conclusion

Intravenous fluids given in sufficient amounts or rapidly enough or bothcan markedly alter acid-base status. During hemofiltration and especially dur-ing high-volume hemofiltration, replacement fluid solutions containing‘buffers’ such as lactate, acetate, bicarbonate and citrate are given rapidly andin large amounts and can have a variable effect on acid-base balance, depend-ing on the dose and rate of metabolic disposition of the ‘buffer’. Evidence isaccumulating that these iatrogenic acid-base alterations have physiological con-sequences and that the choice of fluids affects such physiological changes.Several observations also suggest that misdiagnosis of these disorders can leadto patient injury. Critical care physicians must understand the nature, origin andmagnitude of the alterations in acid-base status caused by intravenous fluidtherapy, if they wish to provide safe and effective care to their patients.

References

1 Coran AG, Ballantine TV, Horwitz DL, Herman CM: The effect of crystalloid resuscitation inhemorrhagic shock on acid-base balance: A comparison between normal saline and ringer’s lactatesolution. Surgery 1971;69:874–880.

2 Prough DS, Whitley JM, Taylor CL, Deal DD, De Witt DS: Regional cerebral blood flow follow-ing resuscitation from hemorrhagic shock with hypertonic saline: Influence on subdural mass.Anesthesiology 1991;75:319–327.

3 Stewart PA: Modern quantitative acid-base chemistry. Can J Physiol Pharmacol 1983;61:1444–1461.

4 Figge J, Mydosh T, Fencl V: Serum proteins and acid-base equilibria: A follow-up. J Lab Clin Med1992;120:713–719.

5 Bellomo R, Ronco C: New paradigms in acid-base physiology. Curr Opin Crit Care 1999;5:427–428.

6 Story D, Bellomo R: The acid-base physiology of crystalloid solutions. Curr Opin Crit Care 1999;5:436–439.

7 Kellum JA, Bellomo R, Kramer DJ, Pinsky MR: Etiology of metabolic acidosis during salineresuscitation in endotoxemia. Shock 1998;9:1–5.

8 Prough DS, Bidani A: Hyperchloremic metabolic acidosis is a predictable consequence of intra-operative infusion of 0.9% saline. Anesthesiology 1999;90:1247–1249.

9 Skellett S, Mayer A, Durward A, Tibby SM, Murdoch IA: Chasing the base deficit: Hyperchloremicacidosis following 0.9% saline fluid resuscitation. Arch Dis Child 2000;83:514–516.

10 Story D, Poustie S, Bellomo R: Quantitative physical chemistry analysis of acid-base disorders incritically ill patients. Anaesthesia 2001;56:530–533.

11 Story DA, Morimatsu H, Bellomo R: Strong ions, weak acids and base excess: A simplified Fencl-Stewart approach to clinical acid-base disorders. Br J Anaesth 2004;92:54–60.


12 Rocktaeschel J, Morimatsu H, Uchino S, Goldsmith D, Pousie S, Story DA, Gutteridge G,Bellomo R: Acid-base status od critically ill patients with acute renal failure: Analysis based onStewart-Figge methodology. Crit Care 2003;7:60–66.

13 Rocktaschel J, Morimatsu H, Uchino S, Ronco C, Bellomo R: Impact of continuous veno-venoushemofiltration on acid-base balance. Int J Artif Organs 2003;26:19–25.

14 Hayhoe M, Bellomo R, Liu G, Mc Nicol L, Buxton B: The etiology and pathogenesis ofCPB-associated metabolic acidosis during polygeline prime. Intens Care Med 1999;25:680–685.

15 Liskaser F, Bellomo R, Hayhoe M, et al: Role of pump prime in the etiology and pathogenesis ofcardiopulmonary bypass-associated acidosis. Anesthesiology 2000;93:1170–1173.

16 Hayhoe M, Bellomo R: The pathogenesis of acid-base changes during cardiopulmonary bypass.Curr Opin Crit Care 1999;5:464–467.

17 Kellum JA, Song M, Venkataraman R: Effects of hyperchloremic acidosis on arterial pressure andcirculating inflammatory molecules in experimental sepsis. Chest 2004;125:243–248.

18 Kellum JA, Song M, Li J: Lactic and hydrochloric acids induce different patterns of inflammatoryresponse in LPS-stimulated RAW 264.7 cells. Am J Physiol Regul Integr Comp Physiol 2003[Epub ahead of print].

19 Finfer S, Bellomo R, Myburgh J, Norton R: Efficacy of albumin in critically ill patients. BMJ2003;326:559–560.

20 Choi PT-L, Yip GY, Quinonez LG, Cook DJ: Crystalloids vs. colloids in fluid resuscitation:A systematic review. Crit Care Med 1999;27:200–210.

21 Cochrane Injuries Group Albumin Reviewers: Human albumin administration in critically illpatients: Systematic review of randomised controlled trials. BMJ 1998;317:235–240.

22 Rocktäschel J, Morimatsu H, Uchino S, Ronco C, Bellomo R: Impact of continuous veno-venoushemofiltration on acid-base balance. Int J Artif Organs 2003;26:19–25.

23 Rocktaeschel J, Morimatsu H, Uchino S, Goldsmith D, Poustie S, Story D, Bellomo R: The complexnature of acid-base disorders in critically ill patients with acute renal failure. Crit Care 2003;7:R60–R66.

24 Bellomo R, Ronco C: The pathogenesis of lactic acidosis in sepsis. Curr Opin Crit Care 1999;5:452–457.

25 Bellomo R: Bench-to-bedside review: Lactate and the kidney. Crit Care 2002;6:322–326.26 Heering P, Ivens K, Thumer O, Brause M, Grabensee B: Acid-base balance and substitution fluid

during continuous hemofiltration. Kidney Int 1999;56:s37-s40.27 Davenport A, Will E, Davison AM: The effect of lactate-buffered solutions on the acid-base status

of patients with renal failure. Nephrol Dial Transplant 1989;4:800–804.28 Mansell MA, Morgan SH, Moore L, Kong CH, Laker MF, Wing AJ: Cardiovascular and acid-base

effects of acetate and bicarbonate haemodialysis. Nephrol Dial Transplant 1987;1:229–232.29 Morgera S, Heering P, Szentandrasi T, Manassa E, Heintzen M, Willers R, Passlick-Deetjen J,

Grabensee B: Comparison of a lactate- versus acetate-based hemofiltration replacement fluid inpatients with acute renal failure. Renal Fail 1997;19:155–164.

30 Thomas AN, Guy JM, Kishen R, Geraghty IF, Bowles BJM, Vadgama P: Comparison of lactateand bicarbonate buffered haemofiltration fluids: Use in critically ill patients. Nephrol DialTransplant 1997;12:1212–1217.

31 Tan HK, Uchino S, Bellomo R: The acid-base effects of continuous hemofiltration with lactate orbicarbonate buffered replacement fluids. Int J Artif Organs 2003;26:477–483.

32 Mehta RL, McDonald B, Aguilar M, Ward DM: Regional citrate anticoagulation for continuousarteriovenous hemodialysis in critically ill patients. Kidney Int 1990;38:976–981.

33 Bellomo R, Tipping P, Boyce N: Continuous veno-venous hemofiltration with dialysis removescytokines from the circulation of septic patients. Crit Care Med 1993;21:522–526.

34 Cole L, Bellomo R, Hart G, Journois D, Davenport P, Tipping P, Ronco C: A phase II randomizedcontrolled trial of continuous hemofiltration in sepsis. Crit Care Med 2002;30:100–106.

35 Ronco C, Bellomo R, Homel P, et al: Effect of different doses in CVVH on outcomes of acuterenal failure: A prospective randomized trial. Lancet 2000;356:26–30.

36 Cole L, Bellomo R, Journois D, Davenport P, Baldwin I, Tipping P: High volume hemofiltrationin human septic shock. Intensive Care Med 2001;27:978–986.


37 Levraut J, Ciebera JP, Jambou P, Ichiai C, Labib Y, Grimaud D: Effect of continuous veno-venoushemofiltration with dialysis on lactate clearance in critically ill patients. Crit Care Med 1997;25:58–62.

38 Cole L, Bellomo R, Baldwin I, Hayhoe M, Ronco C: The impact of lactate-buffered high volumehemofiltration on acid-base balance. Intens Care Med 2003;29:1113–1120.

39 Uchino S, Cole L, Morimatsu H, Goldsmith D, Ronco C, Bellomo R: Solute mass balance duringisovolaemic high volume haemofiltration. Intens Care Med 2003;29:1541–1546.

Prof. Rinaldo BellomoDepartment of Intensive Care, Austin and Repatriation Medical Centre Heidelberg, Vic. 3084 (Australia)Tel. �61 3 9496 5992, Fax �61 3 9496 3932, E-Mail [email protected]


Glucose Control in the Critically Ill

M. Schetz, G. Van den Berghe

Department of Intensive Care Medicine, University Hospital Gasthuisberg, Leuven, Belgium

The subject of glucose and insulin metabolism after injury began 150 yearsago with the observation by Reyboso of glucosuria after ether anesthesia and byClaude Bernard (1877) of hyperglycemia during hemorrhagic shock. Since thenit has become clear that hyperglycemia, glucose intolerance and insulin resist-ance are features that are common to all types of severe acute illness and injury,even in patients without diabetes [1, 2]. This ‘stress diabetes’ used to be inter-preted as an adaptive stress response and as such important for survival. Insulinwas only administered if the glycemia exceeded 200 mg/dl.

Insulin is the most important anabolic hormone known today. It promotesoverall anabolic reactions for all three energy and structural components in thebody: carbohydrates, proteins and fat. In addition, insulin has also been shownto have anti-inflammatory properties [3]. Recently, a large prospective random-ized trial established the beneficial effect of treating even moderate hyper-glycemia with insulin in critically ill patients [4]. In this review, the results ofthis trial and possible underlying mechanisms will be discussed.

Glucose Regulation in Critical Illness

The alterations in carbohydrate metabolism during stress, inflammation orinfection include enhanced peripheral glucose uptake (largely non-insulinmediated and most prominent in tissues involved in the immune response andwound healing), hypermetabolism with augmented glycolysis and glucose oxi-dation (in order to satisfy the increased metabolic demands of the tissuesinvolved in the repair processes) and increased gluconeogenesis and depressedglycogenesis (to maintain the availability of glucose for tissues that are obligateglucose consumers such as wound, brain, erythrocytes and immune cells).

Schetz/Van den Berghe 120

In addition, the enhanced gluconeogenesis is resistant to inhibition by insulinand high glucose levels, permitting to reroute the glucose towards high meta-bolic priorities at the expense of insulin-dependent territories. The major site ofinsulin resistance appears to be the muscle (although it might also exist inadipose tissue and the heart), where it results in reduced insulin-stimulated glu-cose uptake. Insulin resistance in the liver reduces the suppressibility of gluco-neogenesis by exogenous glucose (reviewed in [5, 6]).

Stress hyperglycemia results from the integrated action of hormonal,cytokine and nervous ‘counter-regulatory’ signals on glucose metabolic path-ways. The hormones involved include glucagons, adrenaline, noradrenaline andcortisol. The proinflammatory cytokines affect glucose homeostasis both indi-rectly by stimulating counter-regulatory hormone secretion and directly byincompletely understood post-receptor mechanisms inducing insulin resistancein liver and muscle [7, 8].

The development of modern intensive care medicine has enabled a dramaticincrease in the short-term survival of previously lethal conditions such as multi-ple trauma, extensive burns, major surgery and severe sepsis. Many of thesepatients nowadays enter the chronic phase of critical illness, characterized byimmune paralysis, a wasting syndrome and persistent organ dysfunction. Themechanisms regulating hyperglycemia during protracted critical illness remainrelatively unclear. Growth hormone, cortisol, catecholamine and cytokine levelsare usually decreased in this phase of protracted critical illness [9].

The putative adaptive nature of stress hyperglycemia lead Mizock [5] in1995 to the following conclusion: ‘If one accepts the concept of hyperglycemiaof injury or infection as beneficial by promoting cellular glucose uptake, thenmodest degrees of hyperglycemia should be tolerated without efforts to lowerblood glucose to normal values of 90–120 mg/dl. The level of glycemia shouldbe high enough to maximize cellular glucose uptake without causing hyper-osmolarity. A glucose concentration of 160–200 mg/dl has been recommendedto achieve this goal and is probably acceptable to most clinicians’. In addition,moderate hyperglycemia was often viewed as a buffer against the occurrence ofhypoglycemia and brain damage, which was feared by many clinicians todevelop under tighter glucose control. However, in 2001 the same author wasforced to correct his opinion in view of the increasing evidence of adverseeffects of even moderate hyperglycemia [6].

Adverse Effects of Hyperglycemia

It is generally accepted that hyperglycemia �200 mg/dl should be avoidedbecause it induces osmotic diuresis and fluid shifts that may result in hypovolemia,

Glucose Control in ICU 121

electrolyte abnormalities and hyperosmolar non-ketotic coma. Increased glucoselevels have been shown to enhance the risk of postoperative infection, as wellin diabetics [10, 11] as in non-diabetics [2]. A decrease in the phagocytosis bypolymorphonuclear neutrophils [12], a blunted oxidative burst of leukocytesfollowing exposure to high concentrations of glucose [13, 14], decreased intra-cellular bactericidal activity and opsonic activity [12, 15] and nonenzymaticglycosylation of immunoglobulins [16] might be amongst the underlyingmechanisms of this increased susceptibility to infections.

In patients with type 1 diabetes the diabetes control and complications trial(DCCT) demonstrated that maintenance of blood glucose concentrations closeto the normal range resulted in a highly significant decrease of the progressionrates of diabetic retinopathy, nephropathy and peripheral and autonomic neu-ropathy [17]. A large clinical trial likewise provided evidence for the impor-tance of tight glycemic control for the prevention of complications in patientswith type 2 diabetes [18]. Mortality following acute myocardial infarctionseems to be adversely affected by the presence of diabetes [19], or by hyper-glycemia on admission in patients without previously diagnosed diabetes [20].In the DIGAMI Study intensive insulin therapy in diabetic patients admitted tothe hospital with an acute myocardial infarction resulted in a significantlyimproved 30-day and long-term survival [21, 22] and also the risk for re-infarctionand new cardiac failure was remarkably reduced [23]. Admission hypergly-cemia appears to be an independent predictor of mortality and poor neuro-logical recovery after stroke [24] and traumatic brain injury [25, 26]. In patientswith severe burns an association between hyperglycemia and mortality wasalso noted [27].

Intensive Insulin Therapy in the ICU

Recently, an extensive, prospective, randomized, controlled clinical trialstudied the effects of strict glycemic control on mortality and morbidity of crit-ically ill patients [4]. Previous studies by the same investigator had revealedhigh levels of insulin-like growth factor-binding protein-1 (IGFBP-1) inpatients with protracted critical illness. In addition these increased IGFBP-1levels appeared to be an independent predictor of mortality [28]. Since IGFBP-1is transcriptionally repressed by insulin [29], its high levels were hypothesizedto reflect lack of insulin effect in the liver, which together with the risks asso-ciated with hyperglycemia, generated the rationale for this intervention studytargeting normoglycemia with exogenous insulin in critically ill patients.

The results of this trial for the first time challenged the classical dogmathat stress-induced hyperglycemia is beneficial to these patients. Over a


one-year period, a total of 1,548 patients, admitted to the intensive care unit pre-dominantly after extensive or complicated surgery or trauma and requiringmechanical ventilation, were enrolled in the study. The patients were randomlysubdivided into two groups. One group received ‘intensive insulin therapy’ tokeep blood glucose levels tightly between 80 and 110 mg/dl (4.5 and6.1 mmol/l) by exogenous insulin infusion. In the ‘conventional treatment’group insulin was only administered if blood glucose levels exceeded 220 mg/dl(12 mmol/l). When normoglycemia was strictly maintained by intensive insulintherapy a marked reduction in intensive care unit mortality was observed ascompared with the conventional treatment. The effect was particularly presentin the patients with ICU stay of more than 5 days, where mortality decreasedfrom 20.2 to 10.6% (p � 0.005) (fig. 1). Even conventionally treated patientswith only moderate hyperglycemia (110–150 mg/dl or 6.1–8.3 mmol/l) had asignificantly higher mortality rate than the patients with strict glycemic controlbelow 110 mg/dl (6.1 mmol/l) [30]. Although the study included a large numberof cardiac surgery patients, the beneficial effect of tight glucose control waspresent in the different diagnostic subgroups [4].

Hos

pita

l sur

viva

l (%

)

80

90

100

70

All patients p�0.01

0 100 200

Long-stay patients p�0.02

100

90

80

70

Days

0 50 100 150 200

Intensive

Conventional

ICU

sur

viva

l (%

)

80

90

100 All patients p�0.005

Days

0 20 40 60 80 100 120

Intensive

Conventional

0 40 80 120

Long-stay patients p�0.007

100

90

80

Fig. 1. Kaplan-Meier cumulative survival plots for intensive care and in-hospital sur-vival showing the effect of intensive insuline treatment in the whole study population(n � 1,548) and in the 451 patients with intensive care stay of more than 5 days. Adaptedwith permission from Van den Berghe et al. [4]. Copyright 2001, Massachusetts MedicalSociety. All rights reserved.


Intensive insulin therapy also improved several morbidity-related factors.Thus, the need for prolonged ventilatory support, the duration of intensive carestay, the number of blood transfusions and the incidence of blood stream infec-tions and excessive inflammation were all reduced. Even more striking was thehighly significant decrease in the development of acute renal failure and criti-cal illness polyneuropathy associated with intensive insulin therapy [4] (fig. 2).

Harmful Effect of Hyperglycemia or Beneficial Effect of Insulin?

Multivariate logistic regression analysis indicated that the lowered bloodglucose level rather than the insulin dose was related to reduced mortality, crit-ical illness neuropathy, bacteremia and inflammation. Whether the observedbeneficial effect is due to normoglycemia per se or to another concurrent meta-bolic effect of insulin can not be inferred from this trial. For acute renal failurerequiring renal replacement therapy, not the actual glucose level but theinsulin dose was an independent negative predictor. This might be related to thereduced renal elimination of insulin and the extracorporeal elimination ofglucose [30].

RR

R (%

)

20

40

60

0

**

**

****

******

�Dialysis/CVVH

�Bloodstream infections

�Antibiotics �10 days

� ICU stay �14 days

�Mechanical ventilation �14 days

�Critical illness polyneuropathy

Fig. 2. Relative risk reduction of prolonged (�14 days) ICU stay, prolonged (�14 days)mechanical ventilation, critical illness polyneuropathy, requirement of renal replacementtherapy, prolonged (�10 days) antibiotics and bacteremia in patients with strict glycemiccontrol compared with conventional treatment.


It is conceivable that insulin played a direct role in the functional improve-ment of the insulin-sensitive organs. Since in normal individuals the heart andskeletal muscles are responsible for the majority of the insulin-stimulated glu-cose uptake and, in addition, hyperglycemic conditions aggravate musclecatabolism [31], amelioration of these processes could partially explain thebeneficial effects of intensive insulin therapy on the duration of mechanicalventilation of the critically ill patients in the intensive insulin therapy trial.Subsequent analysis of muscle biopsies of non-survivors in the ‘insulin in crit-ical illness study’ indeed suggests that, in comparison to the conventional treat-ment group, skeletal muscle steady-state mRNA levels of GLUT-4 (the glucosetransporter that is only present in tissues where glucose uptake is mediated byinsulin) and hexokinase II (HXK-II) (a rate-controlling step in intracellularinsulin-stimulated glucose metabolism) are higher in patients following inten-sive insulin therapy. This suggests that peripheral glucose uptake is stimulatedin the latter group of patients [32].

The liver is the major site for gluconeogenesis and is an important insulin-sensitive organ that could be involved in the improved outcome of the patientsintensively treated with insulin. However, the analysis of serum and hepatic geneexpression levels of insulin-like growth factor binding protein-1 (IGFBP-1)(a marker of hepatic insulin sensitivity) and gene expression levels of phospho-enolpyruvate carboxykinase (PEPCK) (the rate-limiting enzyme in gluconeoge-nesis in the liver) revealed that neither of them is regulated by insulin in criticallyill patients. This may indicate that controlling gluconeogenesis was not the majorfactor responsible for the normalization of blood glucose levels with exogenousinsulin in the critically ill [33]. However, true glucose kinetics can only be esti-mated by glucose turnover studies. Such a study, using a well-designed caninemodel of critical illness, was recently performed. When a sublethal hypermeta-bolic infection was induced in dogs, hepatic glucose uptake was decreased andappeared to be unresponsive to insulin administration. In contrast, the peripheraluptake of glucose did respond to insulin infusion. Contrary to our findings,insulin therapy suppressed hepatic glucose production. This apparently occurredthrough inhibition of glycogenolysis rather than diminished hepatic uptake ofgluconeogenic amino acids and gluconeogenesis [34].

Another major insulin-responsive organ is the adipose tissue. Similar to thediabetic patient [35], abnormal serum lipid profiles are observed in the criticallyill patient [36–38]. Most characteristically, triglyceride levels are elevatedwhereas the levels of circulating high-density lipoprotein (HDL) cholesterol andlow-density lipoprotein (LDL) cholesterol are low [39]. On the other hand, circu-lating small dense LDL particles, that presumably are more pro-atherogenic thanthe medium and large LDL particles [40], are increased [41]. Interestingly, thisdyslipidemia could partially be restored by intensive insulin therapy, with almost


complete obliteration of the hypertriglyceridemia and a substantial increase in,but not normalization of, the serum levels of HDL and LDL cholesterol [32]. Therole of triglycerides in energy provision and the coordinating position of thelipoproteins in transportation of lipid components (cholesterol, triglycerides,phospholipids, lipid-soluble vitamins) are well established [42]. In addition, it hasrecently been shown that lipoproteins can scavenge endotoxins and by doing soare able to prevent death in animal models [43, 44]. Multivariate logistic regres-sion analysis demonstrated that the improvement of the deranged lipidemiaexplained a significant part of the beneficial effect on mortality and organ failureand, surprisingly, surpassed the effect of glycemic control and insulin dose [32].In the same way, the effect of intensive insulin therapy on inflammation, reflectedby a lowering of the serum C-reactive protein (CRP) concentrations [45], was nolonger independently related to the outcome benefit when the changes in lipidmetabolism were taken into account [32]. This observation is suggestive of a linkbetween the anti-inflammatory effect of intensive insulin therapy and its amelio-ration of the lipid profile. However, a mechanistic explanation for the dominanteffect of serum lipid correction still needs to be delineated.

Role of Hyperglycemia in Critical Illness-Associated Renal Failure

Presumably, a different pathophysiology is involved in diabetic and criticalillness-associated nephropathy. Whereas diabetic nephropathy mainly affects theglomerulus, in the critically ill acute tubular necrosis is the major pathophysio-logical mechanism of renal failure. The prevention of renal dysfunction and fail-ure is of crucial importance in critically ill patients, since it has repeatedly beenshown to be an independent predictor of mortality [46–48]. Strikingly, the num-ber of critically ill patients that required renal replacement therapy was reducedby 42% when they were intensively treated with insulin, as compared to the con-ventionally treated patients. Hence, intensive insulin therapy emerged as aneffective preventive measure for acute renal failure in critical illness. Whetherthis is directly related to a metabolic or anti-inflammatory effect of insulin orsecondary to the reduced incidence of bacteremia and sepsis is not clear.

Role of Hyperglycemia in Critical Illness-Associated Neuropathy

The most frequent presentation of neuropathies in the diabetic patient is dis-tal sensory neuropathy with the classic stocking distribution [49, 50]. Patients


with protracted critical illness often suffer from a diffuse axonal polyneuropathy[50], which presents as a tetraparesis with muscle atrophy. Even though thecourse of critical illness polyneuropathy is self-limited in most cases and a goodrecovery should be expected once the underlying critical illness is overcome, itseverely delays weaning from the ventilator and impairs early mobilization of thepatient [51]. Sepsis, the use of high dose corticosteroids as well as the use ofneuromuscular blocking agents are all factors that have been implicated in theetiology of critical illness polyneuropathy. Yet, the exact pathogenesis of thiscomplication is still not understood [52]. Recently, strong indications becameavailable that emphasize the importance of blood glucose levels in relation to thedevelopment of critical illness polyneuropathy. First, Bolton [53] described astrong link between the risk of critical illness polyneuropathy on one hand andincreased blood glucose and decreased serum albumin levels, which are bothmetabolic manifestations of multiple organ failure and sepsis, on the other hand.Sepsis, and the accompanying release of cytokines, was considered to be thecausal factor. In addition, the Leuven study on intensive insulin therapy inthe ICU convincingly demonstrated that strict maintenance of glycemia withinthe normal range by infusion of insulin has an important preventive effect on theincidence of critical illness polyneuropathy, which is associated with a decreasein duration of mechanical ventilation of protracted critically ill patients [4].

Immune System Impairment and Risk of Infections caused by Hyperglycemia

The recent Leuven insulin in ICU study also provided a causal linkbetween hyperglycemia and a higher risk of serious infections, regardless of aprevious history of diabetes. Indeed, the occurrence of blood stream infectionswas reduced by almost 50% and sepsis-associated mortality was largely pre-vented when critically ill patients were intensively treated with exogenousinsulin to keep glucose levels within the normal range. These observations sug-gest that insulin-titrated blood glucose control enhances the immune system[4]. Improved capacity to clear bacterial invaders was recently shown to medi-ate this benefit in a novel rabbit model of prolonged critical illness [54].

Effects of Critical Illness and Insulin on Inflammation and Coagulation

Critical illness also resembles diabetes mellitus in the activation of theinflammatory cascade. Intensive insulin therapy proved to be of high value in


the prevention of excessive inflammation in critically ill patients [4, 45].This finding was confirmed in an experimental rabbit model of prolongedcritical illness [54]. The exact mechanisms explaining the anti-inflammatoryeffect of insulin have not yet been unravelled. Suppression of the secretion andantagonism of the harmful actions of tumor necrosis factor-�, macrophagemigration-inhibitory factor and superoxide anions have all been suggested(summarized in [3]).

Furthermore, diabetes mellitus and critical illness both are hypercoagula-ble states [55, 56]. Putative causes in diabetes include vascular endotheliumdysfunction [57], increased blood levels of several clotting factors [58, 59],elevated platelet activation [60, 61] and inhibition of the fibrinolytic system[62]. Levels of the anticoagulant protein C are also decreased [63]. Looking atthe similarities between critical illness and diabetes [64, 65] and the powerfulpreventive effect of intensive insulin therapy on septicemia, multiple organ fail-ure and mortality [4], it is very important to also investigate the influence ofthis simple and cheap metabolic intervention on the balance between coagula-tion and fibrinolysis in the critically ill. Also the effect of insulin on endothe-lial function deserves further investigation.

Possible Dangers of Intensive Insulin Therapy

This risk of hypoglycemia is a major concern of intensive insulin therapyduring critical illness, the more so as clinical symptoms of the autonomicresponse (sweating, tachycardia, tremor) and central nervous symptoms likedizziness, blurred vision, altered mental acuity, confusion and eventually con-vulsions may be masked by concomitant diseases and by inherent intensive caretreatments such as sedation, analgesia and mechanical ventilation. The braincan be irreversibly damaged in the situation of severe hypoglycemia, whenglucose levels drop below 30 mg/dl (�1.67 mmol/l), or in case of persistenthypoglycemia.

In the Leuven study, the risk of hypoglycemia increased form 0.8 to 5.2%.However, with the algorithm used, these episodes of hypoglycemia were alwaysrapidly diagnosed and treated and thus did not detectably cause any seriousadverse events or permanent damage. It is also important to note that hypo-glycemia occurred in the stable phase, mostly after the first week and was oftendue to inadequate insulin dose reduction during interruption of enteral feeding[4]. Limiting the risk of hypoglycemia is of utmost importance, however, andthis requires a number of precautions. These include, among others: (1) pointof care measurement of blood glucose in order to shorten the delay betweenmeasurement and infusion adjustment; (2) using a method for blood glucose


measurement that is accurate, particularly in the low range, and (3) adequatetraining of the nursing and medical staff.

Conclusion

Trauma, burns and critical illness are accompanied by the development ofhyperglycemia, which results from the combined action of hormonal, cytokineand nervous ‘counter-regulatory’ signals on glucose metabolic pathways.Additionally, the major insulin-sensitive organs become resistant to insulin dur-ing critical illness, as manifested by increased serum insulin levels, impairedperipheral glucose uptake and elevated hepatic glucose production. Recently,Van den Berghe et al. [4] demonstrated the beneficial effects of strictly main-taining normoglycemia in ICU patients by administration of intensive insulintherapy. A remarkable reduction in mortality of ICU patients was observed, par-ticularly of those with prolonged critical illness. Furthermore, the intensiveinsulin therapy to a large extent was able to protect against acute renal failureand the development of critical illness polyneuropathy, and partially counter-acted the deranged serum lipid profile, excessive inflammation and impairedimmunity seen in critically ill patients.

References

1 Thorell A, Nygren J, Ljungqvist O: Insulin resistance: A marker of surgical stress. Curr Opin ClinNutr Met Care 1999;21:69–78.

2 McCowen KC, Malhotra A, Bistrian BR: Stress-induced hyperglycaemia. Crit Care Clin 2001;17:107–124.

3 Das UN: Is inslin an anti-inflammatory molecule? Nutrition 2001;17:409–413.4 Van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, Vlasselaers D,

Ferdinande P, Lauwers P, Bouillon R: Intensive insulin therapy in critically ill patients. N Engl JMed 2001;345:1359–1367.

5 Mizock BA: Alterations in carbohydrate metabolism during stress: A review of the literature. Am JMed 1995;98:75–84.

6 Mizock BA: Alterations in fuel metabolism in critical illness: Hyperglycaemia. Best Pract ResClin Endocrinol Metab 2001;15:533–551.

7 Grimble RF: Inflammatory status and insulin resistance. Curr Opin Clin Nutr Met Care 2002;5:551–559.

8 Marette A: Mediators of cytokine-induced insulin resistance in obesity and other inflammatorysettings. Curr Opin Clin Nutr Met Care 2002;5:377–383.

9 Van den Berghe G, de Zegher F, Bouillon R: Clinical review 95: Acute and prolonged critical ill-ness as different neuroendocrine paradigms. J Clin Endocrinol Metab 1998;83:1827–1834.

10 Pozzilli P, Leslie RD: Infections and diabetes: Mechanisms and prospects for prevention. DiabetMed 1994;11:935–941.

11 Funari AP, Zerr KJ, Grunkemeier GL, Starr A: Continuous intravenous insulin infusion reducesthe incidence of deep sternal wound infection in diabetic patients after cardiac surgical proce-dures. Ann Thorac Surg 1999;67:352–360.


12 Rassias AJ, Marrin CA, Arruda J, Whalen PK, Beach M, Yeager MP: Insulin infusion improvesneutrophil function in diabetic cardiac surgery patients. Anesth Analg 1999;88:1011–1016.

13 Nielson CP, Hindson DA: Inhibition of polymorphonuclear leukocyte respiratory burst by elevatedglucose concentrations in vitro. Diabetes 1989;38:1031–1035.

14 Perner A, Nielsen SE, Rask Madsen J: High glucose impairs superoxide production from isolatedblood neutrophils. Intens Care Med 2003;29:642–645.

15 Rayfield EJ, Ault MJ, Keusch GT, Brothers MJ, Nechemias C, Smith H: Infection and diabetes:The case for glucose control. Am J Med 1982;72:439–450.

16 Black CT, Hennessey PJ, Andrassy RJ: Short-term hyperglycemia depresses immunity throughnonenzymatic glycosylation of circulating immunoglobulin. J Trauma 1990;30:830–832.

17 The Diabetes Control and Complications Trial Research Group: The effect of intensive treatmentof diabetes on the development and progression of long-term complications in insulin-dependentdiabetes mellitus. N Engl J Med 1993;329:977–986.

18 UK Prospective Diabetes Study (UKPDS) Group: Intensive blood-glucose control with sulphony-lureas or insulin compared with conventional treatment and risk of complications in patients withtype 2 diabetes (UKPDS 33). Lancet 1998;352:837–853.

19 Mukamal KJ, Nesto RW, Cohen MC, Muller JE, Maclure M, Sherwood JB, Mittleman MA:Impact of diabetes on long-term survival after acute myocardial infarction: Comparability of riskwith prior myocardial infarction. Diab Care 2001;24:1422–1427.

20 Bolk J, van der Ploeg T, Cornel JH, Arnold AE, Sepers J, Umans VA: Impaired glucose metabolismpredicts mortality after a myocardial infarction. Int J Cardiol 2001;79:207–214.

21 Malmberg K, Ryden L, Efendic S, Herlitz J, Nicol P, Waldenstrom A, Wedel H, Welin L: Randomizedtrial of insulin-glucose infusion followed by subcutaneous insulin treatment in diabetic patients withacute myocardial infarction (DIGAMI study): Effects on mortality at 1 year. J Am Coll Cardiol 1995;26:57–65.

22 Malmberg K: Prospective randomised study of intensive insulin treatment on long term survival afteracute myocardial infarction in patients with diabetes mellitus. DIGAMI (Diabetes Mellitus, InsulinGlucose Infusion in Acute Myocardial Infarction) Study Group. BMJ 1997;314:1512–1515.

23 Malmberg K, Ryden L, Hamsten A, Herlitz J, Waldenstrom A, Wedel H: Effects of insulin treat-ment on cause-specific one-year mortality and morbidity in diabetic patients with acute myocar-dial infarction. DIGAMI Study Group. Diabetes Insulin-Glucose in Acute Myocardial Infarction.Eur Heart J 1996;17:1337–1344.

24 Williams LS, Rotich J, Qi R, Fineberg N, Espay A, Bruno A, Fineberg SE, Tierney WR: Effectsof admission hyperglycaemia on mortality and costs in acute ischemic stroke. Neurology 2002;59:67–71.

25 Rovlias A, Kotsou S: The influence of hyperglycaemia on neurological outcome in patients withsevere head injury. Neurosurgery 2000;46:335–342; discussion 342–343.

26 Gore DC, Chinkes D, Heggers J, Herndon DN, Wolf SE, Desai M: Association of hyperglycemiawith increased mortality after severe burn injury. J Trauma 2001;51:540–544.

27 Young B, Ott L, Dempsey R, Haack D, Tibbs P: Relationship between admission hyperglycemiaand neurologic outcome of severely brain-injured patients. Ann Surg 1989;210:466–472.

28 Van den Berghe G, Wouters P, Weekers F, Mohan S, Baxter RC, Veldhuis JD, Bowers CY,Bouillon R: Reactivation of pituitary hormone release and metabolic improvement by infusion ofgrowth hormone-releasing peptide and thyrotropin-releasing hormone in patients with protractedcritical illness. J Clin Endocrinol Metab 1999;84:1311–1323.

29 Baxter RC: Changes in the IGF-IGFBP axis in critical illness. Best Pract Res Clin EndocrinolMetab 2001;15:421–434.

30 Van den Berghe G, Wouters PJ, Bouillon R, Weekers F, Verwaest C, Schetz M, Vlasselaers D,Ferdinande P, Lauwers P: Outcome benefit of intensive insulin therapy in the critically ill: Insulindose versus glycemic control. Crit Care Med 2003;31:359–366.

31 Gore DC, Chinkes DL, Hart DW, Wolf SE, Herndon DN, Sanford AP: Hyperglycemia exacerbatesmuscle protein catabolism in burn-injured patients. Crit Care Med 2002;30:2438–2442.

32 Mesotten D, Swinnen JV, Vanderhoydonc F, Wouters PJ, Van den Berghe G: Contribution of cir-culating lipids to the improved outcome of critical illness by glycemic control with intensiveinsulin therapy. J Clin Endocrinol Metab 2004;89:219–226.


33 Mesotten D, Delhanty PJD, Vanderhoydonc F, Hardman KV, Weekers F, Baxter RC, Van denBerghe G: Regulation of insulin-like growth factor binding protein-1 during protracted critical ill-ness. J Clin Endocrinol Metab 2002;87:5516–5523.

34 Donmoyer CM, Chen SS, Lacy DB, Pearson DA, Poole A, Zhang Y, McGuinnes OP: Infectionimpairs insulin-dependent hepatic glucose uptake during total parenteral nutrition. Am J Physiol2003;284:E574–E585.

35 Taskinen MR: Pathogenesis of dyslipidemia in type 2 diabetes. Exp Clin Endocrinol Diab 2001;109:S180-S188.

36 Lanza-Jacoby S, Wong SH, Tabares A, Baer D, Schneider T: Disturbances in the composition ofplasma lipoproteins during gram-negative sepsis in the rat. Biochim Biophys Acta 1992;1124:233–240.

37 Khovidhunkit W, Memon RA, Feingold KR, Grunfeld C: Infection and inflammation-inducedproatherogenic changes of lipoproteins. J Infect Dis 2000;181:S462–S472.

38 Carpentier YA, Scruel O: Changes in the concentration and composition of plasma lipoproteinsduring the acute phase response. Curr Opin Clin Nutr Metab Care 2002;5:153–158.

39 Gordon BR, Parker TS, Levine DM, Saal SD, Wang JC, Sloan BJ, Barie PS, Rubin AL: Low lipidconcentrations in critical illness: Implications for preventing and treating endotoxemia. Crit CareMed 1996;24:584–589.

40 Kwiterovich PO: Lipoprotein heterogeneity: Diagnostic and therapeutic implications. Am J Cardiol2002;90:1i-10i.

41 Feingold KR, Krauss RM, Pang M, Doerrler W, Jensen P, Grunfeld C: The hypertriglyceridemiaof acquired immunodeficiency syndrome is associated with an increased prevalence of low den-sity lipoprotein subclass pattern B. J Clin Endocrinol Metab 1993;76:1423–1427.

42 Tulenko TN, Sumner AE: The physiology of lipoproteins. J Nucl Cardiol 2002;9:638–649.43 Harris HW, Grunfeld C, Feingold KR, Rapp JH: Human very low density lipoproteins and chy-

lomicrons can protect against endotoxin-induced death in mice. J Clin Invest 1990;86:696–702.44 Harris HW, Grunfeld C, Feingold KR, Read TE, Kane JP, Jones AL, Eichbaum EB, Bland GF,

Rapp JH: Chylomicrons alter the fate of endotoxin, decreasing tumor necrosis factor release andpreventing death. J Clin Invest 1993;91:1028–1034.

45 Hansen TK, Thiel S, Wouters PJ, Christiansen JS, Van den Berghe G: Intensive insulin therapyexerts anti-inflammatory effects in critically ill patients and counteracts the adverse effect of lowmannose-binding lectin levels. J Clin Endocrinol Metab 2003;88:1082–1088.

46 Levy EM, Viscoli CM, Horwitz RI: The effect of acute renal failure on mortality: A cohort analy-sis. JAMA 1996;275:1489–1494.

47 Metnitz PGH, Krenn CG, Steltzer H, Lang T, Ploder J, Lenz K, LeGall JR, Druml W: Effect ofacute renal failure requiring renal replacement therapy on outcome in critically ill patients. CritCare Med 2002;30:2051–2058.

48 de Mendonca A, Vincent JL, Suter PM, Moreno R, Dearden NM, Antonelli M, Takala J, Sprung C,Cantraine F: Acute renal failure in the ICU: Risk factors and outcome evaluated by the SOFAscore. Intens Care Med 2000;26:915–921.

49 Boulton AJ: Clinical presentation and management of diabetic neuropathy and foot ulceration.Diab Med 1991;8:S52–S57.

50 Hund E: Neurological complications of sepsis: Critical illness polyneuropathy and myopathy.J Neurol 2001;248:929–934.

51 Leijten FS, De Weerd AW, Poortvliet DC, De Ridder VA, Ulrich C, Harink-De Weerd JE: Criticalillness polyneuropathy in multiple organ dysfunction syndrome and weaning from the ventilator.Intens Care Med 1996;22:856–861.

52 Bolton CF, Young GB: Critical Illness Polyneuropathy. Curr Treat Options Neurol 2000;2:489–498.

53 Bolton CF: Sepsis and the systemic inflammatory response syndrome: Neuromuscular manifesta-tions. Crit Care Med 1996;24:1408–1416.

54 Weekers F, Giulietti A, Michalaki M, Coopmans W, Van Herck E, Mathieu C, Van den Berghe G:Metabolic, endocrine and immune effects of stress hyperglycaemia in a rabbit model of prolongedcritical illness. Endocrinology 2003;in press.

55 Carr ME : Diabetes mellitus: A hypercoagulable state. J Diabetes Complications 2001;15:44–54.


56 Calles-Escandon J, Garcia-Rubi E, Mirza S, Mortensen A : Type 2 diabetes: One disease, multi-ple cardiovascular risk factors. Coron Artery Dis 1999;10:23–30.

57 Williams E, Timperley WR, Ward JD, Duckworth T: Electron microscopical studies of vessels indiabetic peripheral neuropathy. J Clin Pathol 1980;33:462–470.

58 Patrassi GM, Vettor R, Padovan D, Girolami A: Contact phase of blood coagulation in diabetesmellitus. Eur J Clin Invest 1982;12:307–311.

59 Carmassi F, Morale M, Puccetti R, De Negri F, Monzani F, Navalesi R, Mariani G: Coagulationand fibrinolytic system impairment in insulin dependent diabetes mellitus. Thromb Res 1992;67:643–654.

60 Hughes A, McVerry BA, Wilkinson L, Goldstone AH, Lewis D, Bloom A: Diabetes, a hyperco-agulable state? Hemostatic variables in newly diagnosed type 2 diabetic patients. Acta Haematol1983;69:254–259.

61 Garcia Frade LJ, de la Calle H, Alava I, Navarro JL, Creighton LJ, Gaffney PJ: Diabetes mellitusas a hypercoagulable state: Its relationship with fibrin fragments and vascular damage. ThrombRes 1987;47:533–540.

62 Carmassi F, Morale M, Puccetti R, De Negri F, Monzani F, Navalesi R, Mariani G: Coagulationand fibrinolytic system impairment in insulin dependent diabetes mellitus. Thromb Res 1992;67:643–654.

63 Vukovich TC, Schernthaner G: Decreased protein C levels in patients with insulin-dependenttype I diabetes mellitus. Diabetes 1986;35:617–619.

64 Garcia Frade LJ, Landin L, Avello AG, Martin Yerro J, Navarro JL, Creighton LJ, Gaffney PJ:Changes in fibrinolysis in the intensive care patient. Thromb Res 1987;47:593–599.

65 Mavrommatis AC, Theodoridis T, Economou M, Kotanidou A, El Ali M, Christopoulou-Kokkinou V,Zakynthinos SG: Activation of the fibrinolytic system and utilization of the coagulation inhibitors insepsis: Comparison with severe sepsis and septic shock. Intens Care Med 2001;27:1853–1859.

Miet Schetz, MD, PhDDepartment of Intensive Care MedicineUniversity Hospital Gasthuisberg, Herestraat 49, BE–3000 Leuven (Belgium)Tel. �32 16344021, Fax �32 16344015, E-Mail [email protected]


Dysnatremias in the Critical Care Setting

Michael L. Moritza, J. Carlos Ayusb

aDivision of Nephrology, Department of Pediatrics, Children’s Hospital of Pittsburgh, University of Pittsburgh School of Medicine, Pittsburgh, Pa., and bDivision of Nephrology, Department of Medicine, University of Texas Health Science Center at San Antonio, San Antonio, Tex., USA

Under normal circumstances, the human body is able to maintain plasmasodium levels in the normal range (135–145 mEq/l). Many conditions canimpair the body’s ability to maintain a normal plasma sodium by either inter-fering with free access to water or buy impairing renal concentration and dilu-tion. When the serum sodium departs from the normal range, the cells of thebody are subjected to an injurious osmotic stress which can lead to either aninflux or efflux of water. The most serious complications of dysnatremias arerelated to central nervous system dysfunction, such as cerebral edema do tohyponatremia and cerebral dehydration from hypernatremia. Even when thereis no apparent central nervous system manifestations, dysnatremias are a majorcomorbidity factor.

In no setting is it more difficult to prevent and treat dysnatremias thanin the critical care setting. Patients frequently have multiorgan systemdysfunction where both access to fluids and renal water handling is impaired.Fluid management can be difficult as it must to be incorporated into themanagement of acute and chronic illnesses. The most severe dysnatremiaswith central nervous system manifestation will be managed in the inten-sive care unit where an immediate diagnosis and therapy is required. Thisreview will focus on the common clinical reasons for encountering a dysna-tremia in the critical care setting and on the prevention and treatment ofdysnatremias.

Dysnatremias in the Critical Care Setting 133

Hyponatremia

Pathogenesis of HyponatremiaHyponatremia, defined as a serum �135 mEq/l, is a common disorder that

occurs in both the outpatient and inpatient setting. The body’s primary defenseagainst developing hyponatremia is the kidney’s ability to generate a dilute urineand excrete free water. Rarely is excess ingestion of free water alone the cause ofhyponatremia, as an adult with normal renal function can typically excrete over15 liters of free water per day. It also rare to develop hyponatremia from excessurinary sodium losses in the absence of free water ingestion. In order for hypo-natremia to develop, it typically requires a relative excess of free water in con-junction with an underlying condition that impairs the kidney’s ability to excretefree water (table 1). Renal water handling is primarily under the control of argi-nine vasopressin (AVP) which is produced in the hypothalamus and released fromthe posterior pituitary. AVP release impairs water diuresis by increasing the per-meability to water in the collecting tubule. There are osmotic, hemodynamic andnon-hemodynamic stimuli for AVP release. In most cases of hyponatremia thereis a stimulus for vasopressin production which results in impaired free waterexcretion. The body will attempt to preserve the extracellular volume at theexpense of the serum sodium, therefore a hemodynamic stimulus for AVP pro-duction will override and inhibitory effect of hyponatremia [1]. There are numer-ous stimuli for AVP production (table 1) that occur in hospitalized patients whichmake virtually any hospitalized patient at risk for hyponatremia.

Table 1. Disorders in impaired renal water excretion

Effective circulating volume depletionGastrointestinal losses: vomiting, diarrheaSkin losses: cystic fibrosisRenal losses: salt wasting nephropathy, diuretics, cerebral salt wasting, hypoaldosteronism

Edemetous states: heart failure, cirrhosis, nephrosis, hypoalbuminemia

Thiazide diureticsRenal failure

AcuteChronic

Non-hypovolemic states of ADH excessSIADHCortisol deficiencyHypothyroidism

Moritz/Ayus 134

Diagnostic Approach

Before embarking on an aggressive therapeutic regimen, it is vital to con-firm that hyponatremia is in fact associated with hypo-osmolality.Hyponatremia can be associated with either a normal or an elevated serumosmolality (fig. 1). The most common reasons for this are hyperglycemia,severe hyperproteinemia or hyperlipidemia. Hyperglycemia results in hyper-osmolality with a translocation of fluid from the intracellular space to theextracellular space, resulting in a 1.6 mEq/l fall in the serum sodium for every100 mg/dl elevation in the serum glucose concentration above normal. Severehyperlipidemia, hypercholesterolemia and hyperproteinemia can cause a dis-placement of plasma water, which will result in a decreased sodium concen-tration (pseudohyponatremia) with a normal serum osmolality [2]. Serumsodiums are currently measured by either direct or indirect-reading ion-selective electrode potentiometry. The direct method will not result in pseudo-hyponatremia, as it measures the activity of sodium in the aqueous phase ofserum only. The indirect method on the other hand can result in pseudo-hyponatremia as the specimen is diluted with a reagent prior to measurement[3]. The indirect method is currently performed in approximately 60% ofchemistry labs in the United States; therefore, pseudohyponatremia remains anentity that clinicians need to be aware of [4]. If hyponatremia is associatedwith hypo-osmolality (true hyponatremia), the next step is to measure the uri-nary osmolality to determine if there is an impaired ability to excrete freewater (urineosm �100mosm/kg).

The information that is most useful in arriving at a correct diagnosis ofhyponatremia is a detailed history of fluid balance, weight changes, medica-tions (especially diuretics), and underlying medical illnesses. Hyponatremia isusually a multifactorial disorder and a detailed history will identify sources ofsalt and water losses, free water ingestion, and underlying illnesses that cause anonosmotic stimulus for vasopressin production. An assessment of the volumestatus on physical examination and the urinary electrolytes can be extremelyhelpful, but both can be misleading [5]. In patients in whom hyponatremia isdue to salt losses, such as diuretics, signs of volume depletion may be absent onphysical examination, as the volume deficit may be nearly corrected due to oralintake of hypotonic fluids if the thirst mechanism is intact.

In general a urinary sodium concentration less than 25 mEq/l is consistentwith effective circulating volume depletion, while a urine sodium greater than25 mEq/l is consistent with renal tubular dysfunction, use of diuretics or the syn-drome of inappropriate antidiuretic hormone secretion (SIADH) [6]. Numerousfactors can affect the urine sodium, making interpretation difficult, therefore thetiming of the urinary measurements in relation to dosages of diuretics,


�280mOsm/kg H2O

�100mOsm/kg H2O

Effective circulatoryvolume depletion

Yes

No

No

Urine Na �25mEq/L• SIADH• Reset osmostat

Urine Na �25mEq/L• Repeat algorithm

Renal insufficiencyHypothyroidGlucocorticoid deficiencyPost-operativeSpinal fusionPain/stress/nauseaPositive pressure ventilation

Urine osmolality

�100mOsm/kg H2O• Psychogenic polydypsia• Water intoxication in infants• Reset osmostat

Plasma osmolality

�280mOsm/kg H2O • Hyperglycemia• Mannitol• Pseudohyponatremia � Hyperlipidemia � Hyperproteinemia

Serum Na �135 mEq/L

Urine Na �25mEq/L• Extrarenal losses• Edemetous states

Urine Na �25mEq/L• Salt wasting nephropathy• Mineralocorticoid-deficiency• Cerebral salt wasting• Diuretics• Osmotic diuresis

Fig. 1. Diagnostic approach to hyponatremia.

Moritz/Ayus 136

intravenous fluid boluses, or fluid and sodium restriction are also important. Ifsome cases assessment of volume status by the measurement of a central venouspressure may be helpful [7].

Hospital Acquired Hyponatremia and Its Prevention

Hyponatremia is a common electrolyte disorder occurring in about 3% ofhospitalized patients [8]. The incidence of hyponatremia in the intensive careunit can be as high as 30% [9]. Hyponatremia is a major predictive factor ofmortality in hospitalized patients with heart failure [10] and communityacquired pneumonia [11]. The mortality in hospitalized patients with hypo-natremia is over seven times higher than in those with out [12]. Over two thirdsof all hyponatremia is hospital acquired due to a non-osmotic stimuli for AVPproduction [8, 13]. Hospitalized patients have numerous stimuli for AVP pro-duction, therefore all hospitalized patient should be considered at risk for devel-oping hyponatremia and prophylactic measures must be taken.

The most important factor resulting in hospital acquired hyponatremia isthe administration of hypotonic fluids to a patient who has a compromised abil-ity to maintain water balance [14–18]. In adults this will usually occur in thepostoperative period. While a healthy male adult can excrete at least 15 liters offluid a day and maintain sodium homeostasis, it has been shown that in a womenas few as 3–4 liters of hypotonic fluid over 2 days can result in fatal hypona-tremic encephalopathy in the postoperative setting [16, 19]. Hyponatremia caneven develop if excessive near-isotonic saline, lactated Ringer’s solution, isadministered in the postoperative period [20]. Thus, the most important measurewhich can be taken to prevent hyponatremia is to avoid using hypotonic fluidspostoperatively and to administer isotonic saline, 0.9% sodium chloride, unlessotherwise clinically indicated. The serum sodium should be measured daily inany patient receiving continuous parenteral fluid.

Hospital acquired hyponatremia is commonly seen in the post-operativesetting [21]. Postoperative patients develop hyponatremia due to a combinationof nonosmotic stimuli for ADH release, such as subclinical volume depletion,pain, nausea, stress, edema-forming conditions, and administration of hypo-tonic fluids. ADH levels are universally elevated postoperatively when com-pared to preoperative values [22, 23]. Premenopausal females are most at riskfor developing hyponatremic encephalopathy postoperatively [19].

Hospital acquired hyponatremia is of particular concern in children as thestandard of care in pediatrics has been to administer hypotonic fluids containing0.2–0.45% sodium chloride as maintenance fluids [24]. The safety of thisapproach has never been established. Hospitalized children have numerous


nonosmotic stimuli for vasopressin production which place them at risk fordeveloping hyponatremia [25]. There are over fifty reported cases of neurologicmorbidity and mortality in the past ten years resulting from hospital-acquiredhyponatremia in children receiving hypotonic parenteral fluids [18]. Over half ofthese cases occurred in the postoperative setting in previously healthy childrenundergoing minor elective surgeries [26, 27]. Hyponatremia is especially dan-gerous in children with underlying CNS injury such as encephalitis, with mildhyponatremia (sodium �130 mEq/l) resulting in cerebral herniation [28, 29]. Wehave recently argued that isotonic saline, 0.9% sodium chloride, should be theparenteral fluid of choice in pediatric patients unless there are ongoing freewater losses or a free water deficit [18].

Hyponatremic Encephalopathy

Clinical SymptomsA major consequence of hyponatremia is the influx of water into the intra-

cellular space resulting in cellular swelling, which can lead to cerebral edemaand encephalopathy. The clinical manifestations of hyponatremia are primarilyneurologic and related to cerebral edema caused by hypo-osmolality (table 2).The symptoms of hyponatremic encephalopathy are quite variable between indi-viduals with the only consistent symptoms being headache, nausea, vomiting,emesis, and weakness. As the cerebral edema worsens, patients then developbehavioral changes, and impaired response to verbal and tactile stimuli.Advanced symptoms are signs of cerebral herniation, with seizures, respiratoryarrest, dilated pupils and decorticate posturing. Hyponatremic encephalopathy isimportant to recognize early, as it accounts for one third of the seizures encoun-tered in the ICU setting [30]. Not all patients have the usual progression in symp-toms and advanced symptoms can present with sudden onset.

Table 2. Anatomic and biochemical changes and clin-ical symptoms of hyponatremic encephalopathy

Anatomic and biochemical changes Clinical symptoms

Brain swelling headachenauseavomiting

Pressure on a rigid skull seizuresExcitatory amino acidsTentorial herniation respiratory arrest

Moritz/Ayus 138

Neurogenic Pulmonary EdemaA common yet often unrecognized symptom of hyponatremic encephalopa-

thy is neurogenic pulmonary edema [14, 31, 32]. Neurogenic pulmonary edemais a well described yet under diagnosed condition that occurs as a complication ofsevere CNS injury [33]. There is typically a rapid onset of pulmonary edema fol-lowing the development of cerebral edema. No cardiac etiology is found and thereis a complete and rapid resolution of respiratory systems following appropriatetreatment of hyponatremic encephalopathy with hypertonic saline. If not recog-nized early the condition is almost universally fatal [14]. The pathophysiologicmechanism of neurogenic pulmonary edema is unclear, but appears to be due to(1) increased microvascular permeability to proteins [34], and (2) a sympatheticdischarge resulting in pulmonary vasoconstriction with increased pulmonaryhydrostatic pressure [35]. The incidence of neurogenic pulmonary edema com-plicating hyponatremic encephalopathy is uncertain, but 15% of patients withsevere hyponatremia do have radiographic evidence of pulmonary edema [32].Hyponatremic encephalopathy should be considered in any patient presentingwith a non-cardiogenic pulmonary edema.

Brains Cell Volume RegulationThe brain’s adaptation to hyponatremia initially involves a loss of blood and

cerebral spinal fluid. This is followed by volume regulatory decrease whichoccurs by the extrusion sodium, potassium and organic osmolytes in order todecrease the brain osmolality [36]. Various factors can interfere with successfulbrain adaptation and may a play a more important role than the absolute changein serum sodium in predicting whether a patient will suffer hyponatremicencephalopathy. Elevated AVP levels appear to be a contributing factor to thedevelopment of cerebral edema as AVP is know to increase water content in theabsence of hyponatremia, and impair brain regulatory volume mechanisms[37, 38]. The major factors that interfere with brain adaptation are physical fac-tors related to age, hormonal factors related to gender, and hypoxemia [39].

Risk Factors for Developing Hyponatremic Encephalopathy (table 3)

AgeChildren under 16 years of age are at increased risk for developing hypo-

natremic encephalopathy due to their relatively larger brain to intracranialvolume ratio as compared to adults [27, 40]. A child’s brain reaches adult sizeby 6 years of age, whereas the skull does not reach adult size until 16 years ofage [41, 42]. Consequently, children have less room available in their rigid


skulls for brain expansion and are likely to develop brain herniation fromhyponatremia at higher serum sodium concentrations than adults. Children willhave a high morbidity from symptomatic hyponatremia unless appropriate ther-apy is instituted early [18, 26–29]. After the third decade of life the brain beginsto atrophy, with the steepest reduction in brain volume occurring after 50 yearsof age [43, 44]. The brain volume of an eighty year old is approximately that ofa young child. Consequently, the elderly are at the lowest risk of developingcentral nervous system manifestation of hyponatremia.

GenderRecent epidemiological data have clearly shown that menstruant women

are at substantially higher risk for developing permanent neurological sequelaeor death from hyponatremic encephalopathy than men or postmenopausalfemales [15, 16, 19, 45]. The relative risk of death or permanent neurologicdamage from hyponatremic encephalopathy is approximately 30 times greaterfor women compared to men, and approximately 25 times greater for menstru-ant females than postmenopausal females [19]. Menstruant females candevelop symptomatic hyponatremia at serum sodium values as high as128 mEq/l [16]. Hyponatremic encephalopathy in menstruant females primar-ily occurs in healthy females following elective surgeries while receiving hypo-tonic fluids [16, 19]. Premenopausal women are at high risk for developinghyponatremic encephalopathy due to the inhibitory effects of sex hormones andthe effects of vasopressin on the cerebral circulation, which in the female ani-mal model as opposed to the male are characterized by cerebral vasoconstric-tion and hypoperfusion to brain tissue [40, 46].

HypoxiaHypoxemia is a major risk factor for developing hyponatremic encepha-

lopathy. The occurrence of a hypoxic event such as respiratory insufficiency

Table 3. Risk factors for developing hyponatremicencephalopathy

Risk factor Pathophysiologic mechanism

Children increase brain to intracranial volume ratioFemales sex steroids (estrogens) inhibit brain

adaptationincrease vasopressin levelscerebral vasoconstrictionhypoperfusion of brain tissue

Hypoxemia impaired brain adaptation

Moritz/Ayus 140

is a major factor militating against survival without permanent brain damagein patients with hyponatremia [16]. The combination of systemic hypoxemiaand hyponatremia is more deleterious than is either factor alone becausehypoxemia impairs the ability of the brain to adapt to hyponatremia, leadingto a vicious cycle of worsening hyponatremic encephalopathy [47].Hyponatremia leads to a decrement of both cerebral blood flow and arterialoxygen content [14]. Patients with symptomatic hyponatremia can develophypoxemia by at least two different mechanisms: neurogenic pulmonaryedema or hypercapnic respiratory failure [14]. Respiratory failure can be ofvery sudden onset in patients with symptomatic hyponatremia [16, 31]. Themajority of neurologic morbidity seen in patients with hyponatremia hasoccurred in patients who have had a respiratory arrest as a feature of hypona-tremic encephalopathy [16, 19, 27, 45, 48]. Recent data has shown thathypoxia is the strongest predictor of mortality in patients with symptomatichyponatremia [32].

Hyponatremic Encephalopathy in the Outpatient Setting

Symptomatic hyponatremia can be particularly difficult to recognize in theout-patient setting, as the most common symptoms, namely headache, nausea,vomiting and confusion, can be attributed to other causes. Also, neurogenicpulmonary edema which can be a presenting feature of hyponatremicencephalopathy can be erroneously attributed to cardio-pulmonary disease.Various conditions can result in hyponatremic encephalopathy in the outpatientsetting. The most common causes would be psychogenic polydypsia, thiazidediuretics, and water intoxication in infants. There are some new and unusualpresentations of hyponatremic encephalopathy that have been recently reportedin the outpatient setting. Ayus et al. recently reported on hyponatremic ence-phalopathy occurring in marathon runners, with a presenting symptom ofnoncardiogenic pulmonary edema [31]. All patients had been taking NSAIDs.Patients treated with hypertonic saline had prompt resolution of symptomswithout neurologic sequelae. Fatal hyponatremic encephalopathy has also beenreported following colonoscopy [49, 50]. This appears to be due to a com-bination of large quantities of polyethylene glycol used for bowel preparationin conjunction with increased ADH levels from bowel manipulation.Hyponatremic encephalopathy has been reported to present with hip frac-tures in elderly women, resulting from an unexpected fall in the home [45].Symptomatic hyponatremia has also been reported with the recreationaldrug 3,4-methylenedioxymetamphetamine (Ecstasy) [51]. This results fromincreased vasopressin secretion and excess water ingestion. Symptomatic


hyponatremia can be particularly difficult to recognize in the outpatient settingas the most common symptoms, namely headache, nausea, vomiting and con-fusion, can be attributed to other causes.

Thiazide Diuretics

Thiazide diuretics are an important cause of symptomatic hyponatremiain the outpatient and hospitalized setting [52]. The main reason for this is thatthiazide diuretics act at the distal convoluted tubule and therefore do notimpair urinary dilation as loop diuretics do. Thiazide diuretics lead to hypo-natremia by increasing renal sodium and potassium losses; this in turn leads tovolume depletion, increased vasopressin production and impaired free waterexcretion [53]. Volume depletion may not be clinically apparent due to waterretention. Thiazide-induced hyponatremia typically occurs within two weeksof the start of diuretic [52]. The elderly are most at risk for thiazide-inducedhyponatremia, presumably due to an age-related decline in GFR [54]. Thiazidediuretics also exacerbate the hyponatremia seen in edema forming states.Treatment of thiazide-induced hyponatremia can be challenging as withdrawalof the diuretic can result an over correction of the serum sodium due to a freewater diureses.

Syndrome of Inappropriate Vasopressin Production (SIADH)

SIADH is one of the most common causes of hyponatremia in the hospitalsetting and frequently leads to severe hyponatremia (plasma Na �120 mEq/l)[8]. It is caused by elevated ADH secretion in the absence of an osmotic orhypovolemic stimulus [55]. SIADH can occur due to a variety of illnesses, butmost often occurs due to central nervous system disorders, pulmonary disor-ders and medications (table 4) [56]. Among the latter, the chemotherapeuticdrugs vincristine and cytoxan, and the antiepileptic drug carbamazapine areespecially common. SIADH is essentially a diagnosis of exclusion as can beseen from figure 1. Before SIADH can be diagnosed, diseases causingdecreased effective circulating volume, renal impairment, adrenal insuffi-ciency, and hypothyroidism must be excluded. The hallmarks of SIADH are:mild volume expansion with low to normal plasma concentrations ofcreatinine, urea, uric acid, and potassium; impaired free water excretion withnormal sodium excretion which reflects sodium intake [57]; and hyponatremiawhich is relatively unresponsive to sodium administration in the absence offluid restriction.

Moritz/Ayus 142

SIADH is usually of short duration and resolves with treatment of theunderlying disorder and discontinuation of the offending medication. Fluidrestriction is the cornerstone to therapy, but is a slow method of correction, andis frequently impractical in infants who receive most of their nutrition as liq-uids. All intravenous fluids should be of a tonicity of at least normal saline, andif this does not correct the plasma sodium, 3% sodium chloride may be givenas needed. If a more rapid correction of hyponatremia is needed the addition ofa loop diuretic in combination with hypertonic saline is useful [58, 59]. Agentswhich produce diabetes insipidus such as demeclocyline can be used if SIADHpersists for greater than a month and is unresponsive to fluid restriction,increased sodium intake and loop diuretics [59]. Vasopressin 2 receptor antag-onists are a promising therapy that are currently under investigation but are notapproved for clinical use [60].

Cerebral Salt Wasting (CSW)

Hyponatremia is frequently encountered in the neurosurgical setting and inpatients with CNS injury. This has usually been attributed to SIADH, a condi-tion whose hallmark is euvolemia, with the corner stone of management beingfluid restriction. More recently it has become apparent that an increasing num-ber of neurosurgical patients with hyponatremia have a distinct clinical entitycalled cerebral salt wasting [61, 62], a condition whose hallmark is renalsodium loss leading to extracellular volume depletion, with the cornerstone of

Table 4. Common causes of SIADH

CarcinomasBronchogenic carcinomasOat cell of the lungDuodenum PancreasNeuroblastoma

MedicationsVincristineIntravenous cytoxanCarbamazepineOxcarbazepineSeritonin reuptake inhibitors

Central nervous system disordersInfection: meningitis, encephalitisNeoplasmsVascular abnormalitiesPsychosisHydrocephalusPost-pituitary surgeryHead trauma

Pulmonary disordersPneumoniaTuberculosisAsthmaPositive pressure ventilationPneumothorax


management being volume expansion and salt supplementation [63]. Becausethese two diseases have many clinical similarities, it can be difficult to confirma diagnosis of CSW. It is essential to be able to distinguish between these twoconditions as their management is completely different and fluid restrictionwould be harmful in CSW.

The pathogenesis of CSW is not completely understood, but it appears tobe due to the release of natriuretic peptides, such as atrial natriuretic peptide,brain natriuretic peptide and c-type natriuretic peptide [63]. These peptidesappear to lead to a natriurersis via of complex mechanism of (1) hemodynamiceffects leading to an increased GFR; (2) inhibition of the rennin-angiotensinsystem, and (3) inhibition of the secretion and action of AVP [64]. This com-plex mechanism can lead to biochemical features that are indistinguishable toSIADH with a low uric acid, plasma renin, aldosterone and vasopressin levels,despite volume depletion [65]. The only distinguishing feature between CSWand SIADH is extracellular volume depletion. This can be particularly difficultto assess in CSW and the biochemistries may not be helpful. Central venouspressure or pulmonary capillary wedge pressures may be useful.

From a practical standpoint the administration of normal saline should bean adequate prophylaxis against developing clinically significant hypona-tremia, �130 mEq/l, in SIADH. If clinically significant hyponatremia developsin patient with a CNS disorder receiving only normal saline, than the diagnosisof CSW should be strongly considered. If there are no signs of extracellular vol-ume depletion than a brief period of fluid restriction could be tried. If there aresigns of volume depletion or a lack of response to fluid restriction than thepatient should be managed as CSW. Patients with CSW should be volumeexpanded with normal saline, followed by sufficient quantities of normal salineand 3% NaCl to main fluid balance and a normal serum sodium. The adminis-tration of fluodrocortisone may be beneficial as aldosterone production is rela-tively decreased in CSW [66].

Treatment of Hyponatremic Encephalopathy

Despite the controversies surrounding the optimal treatment of hypo-natremic encephalopathy there are two aspects generally accepted by expertsin the field: (1) treatment should be directed based on the neurological involve-ment and not the absolute serum sodium, and (2) hypertonic saline is notindicated in the asymptomatic patient who is neurologically intact, regardlessof the serum sodium [25, 48, 67–72]. In general, correction with hypertonicsaline is unnecessary and potentially harmful if there are no neurologic mani-festations of hyponatremia. Symptomatic hyponatremia, on the other hand, is a

Moritz/Ayus 144

medical emergency. Once signs of encephalopathy are identified prompt treat-ment is required in a monitored setting before imaging studies are performed.The airway should be secured and endotrachial intubation and mechanical ven-tilation may be necessary. Fluid restriction alone has no place in the treatmentof symptomatic hyponatremia. If symptomatic hyponatremia is recognized andtreated promptly, prior to developing a hypoxic event, the neurological outcomeis good [45, 48, 70, 73].

Patients with symptomatic hyponatremia should be treated with hyper-tonic saline (3%, 513 mEq/l) using an infusion pump (table 5). The rate ofinfusion should raise the plasma sodium by about 1 mEq/l per hour until either(1) the patient is alert and seizure free; (2) the plasma sodium has increased by20 mEq/l, or (3) a serum sodium of about 125–130 mEq/l has been achieved,whichever occurs first [15, 39, 48, 69, 70, 72–75]. If the patient is activelyseizing or with impending respiratory arrest the serum sodium can be raisedby as much as 4–8 mEq/l in the first hour or until seizure activity seizes [75].Recent studies have demonstrated that the optimal rate of correction of symp-tomatic hyponatremia is approximately 15–20 mEq in 48 h, as patients withcorrection of hyponatremia in this range have a much lower mortality and animproved neurological outcome compare to those with a correction of lessthan 10 mEq in 48 h [32, 45, 48]. Assuming that total body water comprises50% of total body weight, 1 ml/kg of 3% sodium chloride will raise the plasmasodium by about 1 mEq/l. In some cases furosemide can also be used toprevent pulmonary congestion and to increase the rate of serum sodiumcorrection.

Table 5. Treatment of hyponatremia

Most important step is prevention: avoidance of hypotonic fluidadministrationMeasure plasma osmolality to confirm hypo-osmolalitySymptomatic hyponatremia (headache, nausea, emesis, weakness)Start treatment with hypertonic saline infusion (513 mM): use an infusionpump in an intensive care unit setting

Monitor serum sodium every 2 h until the patient is stable and symptom free

Stop hypertonic saline when the patient is symptom free or serum sodium is increased by 20 mmol/l in the initial 48 h of therapy

Avoid hyper- or normonatremia during the initial 5 days of therapy, particularly in alcoholic or liver disease patients

Asymptomatic hyponatremiaFluid restrictionTherapy of underlying disorder


Risk Factors for Developing Cerebral Demyelination (table 6)

Cerebral demyelination is a rare complication which has been associatedwith symptomatic hyponatremia [76]. Animal data has demonstrated thatcorrection of hyponatremia by �25 mEq/l in 24 h can result in cerebraldemyelination [77]. This has resulted in a mistaken belief that a rapid rate ofcorrection is likely to result in cerebral demyelination [78]. Studies haveshown that rate of correction has little to do with development of cerebraldemyelinating lesions, and that lesions seen in hyponatremic patients aremore closely associated with other comorbid factors or the magnitude ofcorrection in serum sodium [32, 48, 79, 80]. In one prospective study it wasobserved that hyponatremic patients who develop demyelinating lesions hadeither (a) been made hypernatremic inadvertently; (b) had their plasmasodium levels corrected by greater then 25 mmol/l in 48 h; (c) suffered ahypoxic event, or (d) had severe liver disease [48]. Some data has suggestedthat azotemia may decrease the risk of developing cerebral demyelination[81]. It has also been demonstrated that if there is an overcorrection of hyper-natremia that a therapeutic re-lowering can ameliorate the symptoms ofcerebral demyelination [82].

Cerebral demyelination can be asymptomatic or can manifest in confusion,quadriplegia, pseudobulbar palsy, and a pseudocoma with a ‘locked-in stare’[83]. When symptoms of cerebral demyelination occur following an overcor-rection of hypernatremia it is typically a delayed phenomena that is best diag-nosed on MRI approximately 14 days following correction [16, 84, 85]. Thelesions of cerebral demyelination can be pontine or extrapontine. The lesionstypically develop many days after the correction of hyponatremia, but can beseen in the absence of any sodium abnormalities [86]. In fact, the primary causeof brain damage in patients with hyponatremia is not cerebral demyelination,but cerebral edema and herniation [16, 19, 27, 45, 48]. Most brain damageoccurs in untreated patients and is not a consequence of therapy.

Table 6. Risk factors for developingcerebral demyelination in hyponatremicpatients

Development of hypernatremiaIncrease in serum sodium exceeding 25 mmol/l in 48 h

HypoxemiaSevere liver diseaseAlcoholismCancerSevere burnsMalnutritionHypokalemia

Moritz/Ayus 146

Hypernatremia

Hypernatremia is defined as a serum sodium greater than 145 mEq/l. Inboth children and adults hypernatremia is primarily seen in the hospital settingoccurring in individuals that have restricted access to water for a variety of rea-sons [87–89]. In most instances these patients are either debilitated by an acuteor chronic illness or neurologic impairment, or are at the extremes of age.Hypernatremia is a particularly common problem in the intensive care unit asmost patients are either intubated or moribund and have restricted access to flu-ids [90]. Additional contributing factors for hypernatremia in the intensive caresetting are excess sodium administration, renal concentrating defects, gastroin-testinal fluid losses and dialysis related complications (fig. 2). Unlike mildchronic hyponatremia, which may be physiologic in certain edematous dis-eases, a serum sodium greater than 145 mmol/l should always be consideredabnormal and evaluated thoroughly in order to prevent the development of sig-nificant hypernatremia.

Pathogenesis

The body has two defenses to protect against developing hypernatremia:the ability to produce a concentrated urine and a powerful thirst mechanism.ADH release occurs when the plasma osmolality exceeds 275–280 mosm/kgand results in a maximally concentrated urine when the plasma osmolality

Central DINephrogenic DIDiureticsTubulopathyRecovering acute renal failureHyperglycemiaHigh solute feedsMannitol

FeverHigh ambient temperatureExerciseBurnsRespiratory illness

GastroenteritisOsmotic diarrheaLactuloseCharcoal-sorbitolColostomy/ileostomyMalabsorptionVomiting

Neurologic impairmentHypothalamic disorderRestricted access to fluidsFluid restrictionIneffective breast feeding

Hypertonic NaClNaHCO3Normal saline, blood productsHigh solute feedingSodium ingestionSodium polystereneImproper dialysis solution

Serum Na �145mEq/l

Evaluate for contributing factors

Water losses

Decreased fluid intake Excess sodiumadministrationGastrointestinalInsensibleRenal

Fig. 2. Diagnostic approach to hypernatremia.


exceeds 290–295 mosm/kg. Thirst is the body’s second line of defense, butprovides the ultimate protection against hypernatremia. If the thirst mechanismis intact and there is unrestricted access to free water, it is rare for someone todevelop sustained hypernatremia from either excess sodium ingestion or a renalconcentrating defect.

Diagnosis

Hypernatremia is usually multifactorial and a systematic approach isrequired to determine the contributing factors (fig. 2). A serum sodium, glu-cose, and osmolality must be evaluated. An elevated serum sodium is alwaysassociated with hyperosmolality, and should be considered abnormal. In casesof significant hyperglycemia, the serum sodium will be depressed due to theassociated translocation of fluids from the intracellular to extracellular space.Once the diagnosis of hypernatremia is established, a detailed history andreview of fluid intake should be taken to determine if the patient has an intactthirst mechanism, has restricted access to fluids, or is not being provided ade-quate free water in intravenous fluids. Urine volume should be measured andcompared to fluid intake, and the urine osmolality and electrolytes should bedetermined to assess if the renal concentrating ability is appropriate and toquantify the urinary free water losses. A less than maximally concentrated urine(less than 800 mosm/kg) in the face of hypernatremia is a sign of a renal con-centrating defect, as hypernatremia is a maximal stimulus for ADH release. Inpatients with hypernatremia the following should be evaluated: gastrointestinallosses, dermal losses from fever or burns, diet history (including tube feedings),medication history (including diuretics) and sources of exogenous sodium.

Clinical Manifestations of Hypernatremia

Hypernatremia results in an efflux of fluid from the intracellular space tothe extracellular space to maintain osmotic equilibrium. This leads to transientcerebral dehydration with cell shrinkage. Brain cell volume can decrease by asmuch as 10–15% acutely, but then quickly adapts [91]. Within one hour thebrain significantly increases its intracellular content of sodium and potassium,amino acids and unmeasured organic substances called idiogenic osmoles.Within one week the brain regains approximately 98% of its water content.If severe hypernatremia develops acutely, the brain may not be able to increaseits intracellular solute sufficiently to preserve its volume, and the resultingcellular shrinkage can cause structural changes. Cerebral dehydration from

Moritz/Ayus 148

hypernatremia can result in a physical separation of the brain from themeninges leading to a rupture of the delicate bridging veins and intracranial orintracerebral hemorrhages [92, 93]. Venous sinus thrombosis leading to infarc-tion can also develop [94]. Acute hypernatremia has also been shown to causecerebral demyelinating lesions in both animals and humans [91, 95–97]. Patientswith hepatic encephalopathy are at the highest risk for developing demyelinat-ing lesions [98].

Children with hypernatremia are usually agitated and irritable but canprogress to lethargy, listlessness and coma [99]. On neurologic examinationthey frequently have increased tone, nuchal rigidity and brisk reflexes.Myoclonus, asterixis and chorea can be present; tonic-clonic and absenceseizures have been described. Hyperglycemia is a particularly common conse-quence of hypernatremia in children. Severe hypernatremia can also result inrhabdomyolysis [100]. While earlier reports showed that hypocalcemia wasassociated with hypernatremia, this has not been found in more recent literature[89]. In adults hypernatremia primarily manifests as central nervous systemdepression [88]. Adults with hypernatremia are rarely alert and most have con-fusion with abnormal speech and obtundation with stupor or coma. The degreeof central nervous system depression appears to correlate with the severity ofhypernatremia.

Mortality

The mortality associated with hypernatremia is high in both children andadults. Hypernatremia is associated with a mortality rate of 15% in children;this rate is estimated to be 15 times higher than the age-matched mortality inhospitalized children without hypernatremia [89]. The high mortality is unex-plained. Most of the deaths are not directly related to central nervous systempathology and appear to be independent of the severity of hypernatremia.Recent studies have noted that patients who develop hypernatremia followinghospitalization and patients with a delay in treatment have the highest mortal-ity [89, 101, 102]. Approximately 40% of the deaths in children occurred whilepatients were still hypernatremic. The mortality in adults is between 40 and70%, much higher than in age-matched hospitalized patients [87, 88, 102]. It isnot clear how hypernatremia contributes to mortality in adults, but it does notappear to be affected by age or the severity of hypernatremia. The main predic-tive factors of death in adults with hypernatremia are the duration of hyper-natremia and degree of neurologic impairment [88, 102].

A subset of patients that have a particularly high morbidity and mortality arethose with end-stage liver disease [98, 103]. Patients with hepatic encephalopathy


frequently have hypernatremia from an osmotic diarrhea from the oral adminis-tration of lactulose to treat hyperammonemia. The mortality in patients withhepatic encephalopathy due to lactulose induced hypernatremia is between 40and 80% [103, 104]. Patients with liver disease who develop hypernatremia arealso at high risk for developing cerebral demyelination [76, 95].

Treatment

The goal of therapy for hypernatremia is to both correct the serum sodiumand simultaneously maintain a normal circulatory volume. The cornerstonein the management of hypernatremia is providing adequate free water to correctthe serum sodium. The method of correction largely depends on the etiologyof the hypernatremia and the renal concentrating ability. In cases of hyperna-tremic dehydration, the free water deficit can not be assessed by physicalexamination as the majority of the water losses are primarily intracellular. Signsof volume depletion are less pronounced in patients with hypernatremia due tobetter preservation of the extracellular volume [25]. A simple way of estimat-ing the minimum amount of fluid necessary to correct the serum sodium is bythe following equation:

Free water deficit (ml) � 4 ml F lean body weight (kg)� [desired change in serum Na mEq/l].

Larger amounts of fluid will be required depending on the fluid composi-tion. To correct a 3-liter free water deficit, approximately 4 liters of 0.2%sodium chloride in water or 6 liters of 0.45% sodium chloride in water wouldbe required, as they contain approximately 75 and 50% free water, respectively.The calculated deficit does not account for insensible losses or ongoing urinaryor gastrointestinal losses. Maintenance fluids, which include replacement ofurine volume with hypotonic fluids, are given in addition to the deficit.Glucose-containing fluids should be limited as they can result in significanthyperglycemia [105, 106]. If there are signs of circulatory collapse, fluid resus-citation with normal saline or colloid should be instituted before correcting thefree water deficit. Oral hydration should be instituted as soon as it can be safelytolerated. Plasma electrolytes should be checked every two hours until thepatient is neurologically stable.

The rate of correction of hypernatremia is largely dependent on the severityof the hypernatremia and the etiology. Due to the brain’s relative inability toextrude unmeasured organic substances called idiogenic osmoles, rapid correc-tion of hypernatremia can lead to cerebral edema [91]. While there are no defin-itive studies that document the optimal rate of correction that can be undertaken

Moritz/Ayus 150

without developing cerebral edema, empirical data have shown that unless symp-toms of hypernatremic encephalopathy are present, a rate of correction notexceeding 1 mEq/h or 15 mEq/24 h is reasonable [107–109]. In severe hyperna-tremia (�170 mEq/l) serum sodium should not be corrected to below 150 mEq/lin the first 48–72 h [108]. Seizures occurring during the correction of hyperna-tremia are not uncommon in children, and may be a sign of cerebral edema[110–112]. They can usually be managed by slowing the rate of correction or bygiving hypertonic saline to increase the serum sodium a few milliequivilents. Theseizures are usually self-limited and not a sign of long-term neurologic sequelae[107, 108]. Patients with acute hypernatremia, corrected by the oral route, can tol-erate a more rapid rate of correction with a much lower incidence of seizures[110, 113]. The type of therapy is largely dependent on the etiology of the hyper-natremia and should be tailored to the pathophysiologic events involved in eachpatient. Certain forms of therapy for hypernatremia require special mention.

Central Diabetes insipidus

Central diabetes insipidus (CDI) is an important cause of hypernatremia inthe intensive care setting which must be recognized early as it requires specifictherapy [114]. CDI results from inadequate AVP secretion. CDI in the intensivecare setting typically presents with abrupt polyuria with a free water diuresis.Severe hypernatremia can develop in an individual who has restricted access tofluids and is receiving sodium containing parenteral fluids. Common causes ofCDI in the intensive care setting include brain tumors, pituitary surgery, centralnervous system infections, deceleration injury, head trauma, and cerebralhemorrhages or infarcts. Because patient with CDI can conserve sodium appro-priately, they typically do not manifest signs of volume depletion. A urine osmo-lality which is not maximally concentrated in the presence of hypernatremiasuggests a renal concentrating defect. In CDI the urine osmolality is typicallyless than the plasma osmolalilty. The simplest way to distinguish CDI fromnephrogenic DI is to administer desmopressin (dDAVP) a V2 receptor agonist. InCDI there will typically be a greater than 50% increase in urine osmolality inresponse to dDAVP. The administration of dDAVP either subcutaneously orintranasally is the preferred treatment for this condition.

Hypernatremia in the Edematous Patient

While hypernatremia is usually associated with volume depletion, anincreasing number of patients in the intensive care setting have hypernatremia


with edema [115]. This typically occurs in patients with either multisystemorgan failure or edema forming conditions such as congestive heart failure,cirrhosis or acute renal insufficiency. These patients initially present with a nor-mal serum sodium and become increasingly edematous following the adminis-tration of large amounts of volume in the form of saline, colloid, or bloodproducts to restore circulatory volume. Hypernatremia then develops if thepatient has either urinary or gastrointestinal free water losses in combinationwith fluid restriction and ongoing saline administration. The free water diuresisis usually due to loop diuretics, renal insufficiency, an osmotic diuresis ortubular dysfunction from medications. Gastrointestinal free water losses areusually from lactulose administration. This clinically scenario must be recog-nized early as the hypernatremia can be prevented if sodium is removed fromall continuous infusions including parenteral nutrition and sufficient free in theform of 5% dextrose in water is administered to maintain a normal serumsodium. It may not be possible to correct hypernatremia in the edematouspatient with free water alone if there is severe renal insufficiency or markedfluid overload leading to congestive heart failure or pulmonary congestion. Inthis situation, dialytic therapy may be required to correct both fluid overloadand hypernatremia.

Dialytic Therapies in the Hypernatremic Patient

When hypernatremia can not be adequately managed with free water sup-plementation alone, dialysis may be required [116–118]. This would be the casein severe renal failure, fluid overload and possibly salt poisoning. Dialytic ther-apies provide an option to correct both fluid and electrolyte disorders simulta-neously. Different dialytic therapies can be used and each has unique issues thatmust be considered.

Peritoneal DialysisPeritoneal dialysis is frequently employed in small children where

hemodialysis access is difficult or when hemodialysis or hemofiltration thera-pies are unavailable or contraindicated. Peritoneal dialysis with standarddianeal solution, sodium concentration 132 mEq/l, will not result in a correc-tion of hypernatremia unless free water is administered to the patient [119].Peritoneal dialysated ultrafiltrate is hypotonic in relation to the patient’s serumsodium, and will result in worsening hypernatremia if the patient is fluidrestricted or receiving isotonic fluids. In order to achieve sodium removal inexcess of free water removal, a modified dialysate solution must be prepared bythe pharmacy which (1) has a reduced sodium concentration in order to allow

Moritz/Ayus 152

the diffusion of sodium across peritoneal membrane, and (2) has an increaseddextrose concentration, compared to commercial dialysate solution, in order tomaintain a hypertonic solution that will allow ultrafiltration [118].

HemodialysisHemodialysis has been successfully used in the management of hyper-

natremia in adults and is available in all modern hospital settings [116, 117].Hemodialysis allows the correction of serum sodium in two ways (1) dialyticclearance of serum sodium, and (2) by allowing ultrafiltration during dialysisto replace with free water following dialysis. Before dialyzing a patient withhypernatremia, there must be an estimation of how rapid the serum sodium willfall during dialysis as rapid correction of hypernatremia can result in cerebraledema. The fall in serum sodium can be estimated for the KT/V for the dialysistreatment assuming the K is equal to the blood flow. For example, a dialysistreatment with a KT/V of 1.2 will result in a 66% reduction between thepatient’s serum sodium and the sodium concentration of the dialysate bath. Insevere hypernatremia both dialysis prescription and dialysate sodium concen-tration may have to be modified to prevent a rapid fall in serum sodium.

Continuous Venovenous HemofiltrationContinuous venovenous hemofiltration is the preferred dialytic therapy for

a patient with hypernatremia and fluid overload. It allows for both a slow cor-rection of the hypernatremia and fluid removal can be adjusted as needed tobest suite the patient. When correcting hypernatremia it is best to set thereplacement fluid sodium concentration at 140 mEq/l and adjust the replace-ment fluid rate for an appropriate fall in serum sodium. If the serum sodium isfailing to correct, free water can be administered to the patient without havingto change the replacement fluid composition.

References

1 Dunn FL, Brennan TJ, Nelson AE, Robertson GL: The role of blood osmolality and volume inregulating vasopressin secretion in the rat. J Clin Invest 1973;52:3212–3219.

2 Turchin A, Seifter JL, Seely EW: Clinical problem-solving: Mind the gap. N Engl J Med 2003;349:1465–1469.

3 Weisberg LS: Pseudohyponatremia: A reappraisal. Am J Med 1989;86:315–318.4 Bruns DE, Ladenson JH, Scott MG: Hyponatremia. N Engl J Med 2000;343(12):886–887; author

reply 888.5 Musch W, Thimpont J, Vandervelde D, Verhaeverbeke I, Berghmans T, Decaux G: Combined frac-

tional excretion of sodium and urea better predicts response to saline in hyponatremia than dousual clinical and biochemical parameters. Am J Med 1995;99(4):348–355.

6 Chung HM, Kluge R, Schrier RW, Anderson RJ: Clinical assessment of extracellular fluid volumein hyponatremia. Am J Med 1987;83(5):905–908.


7 Damaraju SC, Rajshekhar V, Chandy MJ: Validation study of a central venous pressure-based pro-tocol for the management of neurosurgical patients with hyponatremia and natriuresis. Neurosurgery1997;40(2):312–316; discussion 316–317.

8 Anderson RJ, Chung HM, Kluge R, Schrier RW: Hyponatremia: A prospective analysis of its epi-demiology and the pathogenetic role of vasopressin. Ann Intern Med 1985;102(2):164–168.

9 DeVita MV, Gardenswartz MH, Konecky A, Zabetakis PM: Incidence and etiology of hypona-tremia in an intensive care unit. Clin Nephrol 1990;34(4):163–166.

10 Lee DS, Austin PC, Rouleau JL, Liu PP, Naimark D, Tu JV: Predicting mortality among patientshospitalized for heart failure: Derivation and validation of a clinical model. JAMA 2003;290(19):2581–2587.

11 Fine MJ, Auble TE, Yealy DM, Hanusa BH, Weissfeld LA, Singer DE, Coley CM, Marrie TJ,Kapoor WN: A prediction rule to identify low-risk patients with community-acquired pneumonia.N Engl J Med 1997;336(4):243–250.

12 Tierney WM, Martin DK, Greenlee MC, Zerbe RL, McDonald CJ: The prognosis of hyponatremiaat hospital admission. J Gen Intern Med 1986;1(6):380–385.

13 Natkunam A, Shek CC, Swaminathan R: Hyponatremia in a hospital population. J Med 1991;22(2):83–96.

14 Ayus JC, Arieff AI: Pulmonary complications of hyponatremic encephalopathy: Noncardiogenicpulmonary edema and hypercapnic respiratory failure. Chest 1995;107(2):517–521.

15 Ayus JC, Arieff AI: Brain damage and postoperative hyponatremia: The role of gender. Neurology1996;46(2):323–328.

16 Arieff AI: Hyponatremia, convulsions, respiratory arrest, and permanent brain damage after electivesurgery in healthy women. N Engl J Med 1986;314(24):1529–1535.

17 Aronson D, Dragu RE, Nakhoul F, Hir J, Miller A, Boulos M, Zinder O, Green J, Mittleman MA,Markiewicz W: Hyponatremia as a complication of cardiac catheterization: A prospective study.Am J Kidney Dis 2002;40(5):940–946.

18 Moritz ML, Ayus JC: Prevention of hospital-acquired hyponatremia: A case for using isotonicsaline. Pediatrics 2003;111(2):227–230.

19 Ayus JC, Wheeler JM, Arieff AI: Postoperative hyponatremic encephalopathy in menstruant women.Ann Intern Med 1992;117(11):891–897.

20 Steele A, Gowrishankar M, Abrahamson S, Mazer CD, Feldman RD, Halperin ML: Postoperativehyponatremia despite near-isotonic saline infusion: A phenomenon of desalination. Ann InternMed 1997;126(1):20–25.

21 Chung HM, Kluge R, Schrier RW, Anderson RJ: Postoperative hyponatremia. A prospective study.Arch Intern Med 1986;146(2):333–336.

22 Grant PJ, Hampton KK, Primrose J, Davies JA, Prentice CR: Vasopressin and haemostatic responsesto inguinal hernia repair under local anaesthesia. Blood Coagul Fibrinolysis 1991;2(5):647–650.

23 Wilson J, Grant PJ, Davies JA, Boothby M, Gaffney PJ, Prentice CR: The relationship betweenplasma vasopressin and changes in coagulation and fibrinolysis during hip surgery. Thromb Res1988;51(4):439–445.

24 Dabbagh D, Atiya B, Fleischmann LE, Gruskin AB: Fluid and electrolyte therapy; in Burg DP,Ingelfinger JR, Wald ER, Polin RA (eds): Gellis & Kagan’s Current Pediatric Therapy, ed 16.Philadelphia, Saunders, 1999, pp 860–870.

25 Moritz ML, Ayus JC: Disorders of water metabolism in children: Hyponatremia and hypernatremia.Pediatr Rev 2002;23(11):371–380.

26 Halberthal M, Halperin ML, Bohn D: Lesson of the week: Acute hyponatraemia in children admit-ted to hospital: Retrospective analysis of factors contributing to its development and resolution.BMJ 2001;322(7289):780–782.

27 Arieff AI, Ayus JC, Fraser CL: Hyponatraemia and death or permanent brain damage in healthychildren. BMJ 1992;304:1218–1222.

28 McJunkin JE, de los Reyes EC, Irazuzta JE, Caceres MJ, Khan RR, Minnich LL, Fu KD, Lovett GD,Tsai T, Thompson A: La Crosse encephalitis in children. N Engl J Med 2001;344(11):801–807.

29 Moritz ML, Ayus JC: La Crosse encephalitis in children. N Engl J Med 2001;345(2):148–149.30 Wijdicks EF, Sharbrough FW: New-onset seizures in critically ill patients. Neurology 1993;43(5):

1042–1044.

Moritz/Ayus 154

31 Ayus JC, Arieff AI: Noncardiogenic pulmonary edema in marathon runners. Ann Intern Med2000;133(12):1011.

32 Nzerue C, Baffoe-Bonnie H, Dail C: Predictors of mortality with severe hyponatremia. J Am SocNephrol 2002;13:A0728.

33 Fontes RB, Aguiar PH, Zanetti MV, Andrade F, Mandel M, Teixeira MJ: Acute neurogenic pul-monary edema: Case reports and literature review. J Neurosurg Anesthesiol 2003;15(2):144–150.

34 McClellan MD, Dauber IM, Weil JV: Elevated intracranial pressure increases pulmonary vascularpermeability to protein. J Appl Physiol 1989;67(3):1185–1191.

35 Maron MB: Pulmonary vasoconstriction in a canine model of neurogenic pulmonary edema. J ApplPhysiol 1990;68(3):912–918.

36 McManus ML, Churchwell KB, Strange K: Regulation of cell volume in health and disease. N EnglJ Med 1995;333(19):1260–1266.

37 Vajda Z, Pedersen M, Doczi T, Sulyok E, Stodkilde-Jorgensen H, Frokiaer J, Nielsen S: Effects ofcentrally administered arginine vasopressin and atrial natriuretic peptide on the development ofbrain edema in hyponatremic rats. Neurosurgery 2001;49(3):697–704; discussion 704–695.

38 Doczi T, Laszlo FA, Szerdahelyi P, Joo F: Involvement of vasopressin in brain edema formation:Further evidence obtained from the Brattleboro diabetes insipidus rat with experimental subarach-noid hemorrhage. Neurosurgery 1984;14(4):436–441.

39 Ayus JC, Arieff AI: Pathogenesis and prevention of hyponatremic encephalopathy. EndocrinolMetab Clins N Am 1993;22(2):425–446.

40 Arieff AI, Kozniewska E, Roberts TP, Vexler ZS, Ayus JC, Kucharczyk J: Age, gender, and vaso-pressin affect survival and brain adaptation in rats with metabolic encephalopathy. Am J Physiol1995;268:R1143–R1152.

41 Sgouros S, Goldin JH, Hockley AD, Wake MJ, Natarajan K: Intracranial volume change in child-hood. J Neurosurg 1999;91(4):610–616.

42 Xenos C, Sgouros S, Natarajan K: Ventricular volume change in childhood. J Neurosurg 2002;97(3):584–590.

43 Courchesne E, Chisum HJ, Townsend J, Cowles A, Covington J, Egaas B, Harwood M, Hinds S,Press GA: Normal brain development and aging: Quantitative analysis at in vivo MR imaging inhealthy volunteers. Radiology 2000;216(3):672–682.

44 Takeda S, Matsuzawa T: Age-related brain atrophy: A study with computed tomography. J Gerontol1985;40(2):159–163.

45 Ayus JC, Arieff AI: Chronic hyponatremic encephalopathy in postmenopausal women:Association of therapies with morbidity and mortality. JAMA 1999;281(24):2299–2304.

46 Fraser CL, Swanson RA: Female sex hormones inhibit volume regulation in rat brain astrocyteculture. Am J Physiol 1994;267:C909–C914.

47 Vexler ZS, Ayus JC, Roberts TP, Fraser CL, Kucharczyk J, Arieff AI: Hypoxic and ischemichypoxia exacerbate brain injury associated with metabolic encephalopathy in laboratory animals.J Clin Invest 1994;93(1):256–264.

48 Ayus JC, Krothapalli RK, Arieff AI: Treatment of symptomatic hyponatremia and its relation tobrain damage. A prospective study. N Engl J Med 1987;317(19):1190–1195.

49 Ayus JC, Levine R, Arieff AI: Fatal dysnatraemia caused by elective colonoscopy. BMJ 2003;326(7385):382–384.

50 Cohen CD, Keuneke C, Schiemann U, Schroppel B, Siegert S, Rascher W, Gross M, Schlondorff D:Hyponatraemia as a complication of colonoscopy. Lancet 2001;357(9252):282–283.

51 Holden R, Jackson MA: Near-fatal hyponatraemic coma due to vasopressin over-secretion after‘ecstasy’ (3,4-MDMA). Lancet 1996;347(9007):1052.

52 Sonnenblick M, Friedlander Y, Rosin AJ: Diuretic-induced severe hyponatremia: Review andanalysis of 129 reported patients. Chest 1993;103(2):601–606.

53 Abramow M, Cogan E: Clinical aspects and pathophysiology of diuretic-induced hyponatremia.Adv Nephrol Necker Hosp 1984;13:1–28.

54 Clark BA, Shannon RP, Rosa RM, Epstein FH: Increased susceptibility to thiazide-inducedhyponatremia in the elderly. J Am Soc Nephrol 1994;5(4):1106–1111.

55 Bartter FC, Schwartz WB: The syndrome of inappropriate secretion of antidiuretic hormone. AmJ Med 1967;42:790–806.


56 Zerbe R, Stropes L, Robertson G: Vasopressin function in the syndrome of inappropriate antidi-uresis. Annu Rev Med 1980;31:315–327.

57 Cooke CR, Turin MD, Walker WG: The syndrome of inappropriate antidiuretic hormone secretion(SIADH): Pathophysiologic mechanisms in solute and volume regulation. Medicine (Baltimore)1979;58(3):240–251.

58 Hantman D, Rossier B, Zohlman R, Schrier R: Rapid correction of hyponatremia in the syndromeof inappropriate secretion of antidiuretic hormone: An alternative treatment to hypertonic saline.Ann Intern Med 1973;78(6):870–875.

59 Perks WH, Walters EH, Tams IP, Prowse K: Demeclocycline in the treatment of the syndrome ofinappropriate secretion of antidiuretic hormone. Thorax 1979;34(3):324–327.

60 Saito T, Ishikawa S, Abe K, Kamoi K, Yamada K, Shimizu K, Saruta T, Yoshida S: Acute aquaresisby the nonpeptide arginine vasopressin (AVP) antagonist OPC-31260 improves hyponatremia inpatients with syndrome of inappropriate secretion of antidiuretic hormone (SIADH). J ClinEndocrinol Metab 1997;82(4):1054–1057.

61 Wijdicks EF, Vermeulen M, ten Haaf JA, Hijdra A, Bakker WH, van Gijn J: Volume depletion andnatriuresis in patients with a ruptured intracranial aneurysm. Ann Neurol 1985;18(2):211–216.

62 Sivakumar V, Rajshekhar V, Chandy MJ: Management of neurosurgical patients with hyponatremiaand natriuresis. Neurosurgery 1994;34(2):269–274; discussion 274.

63 Harrigan MR: Cerebral salt wasting syndrome. Crit Care Clin 2001;17(1):125–138.64 Levin ER, Gardner DG, Samson WK: Natriuretic peptides. N Engl J Med 1998;339(5):321–328.65 Rabinstein AA, Wijdicks EF: Hyponatremia in critically ill neurological patients. Neurology 2003;

9(6):290–300.66 Ishikawa SE, Saito T, Kaneko K, Okada K, Kuzuya T: Hyponatremia responsive to fludrocortisone

acetate in elderly patients after head injury. Ann Intern Med 1987;106(2):187–191.67 Gowrishankar M, Lin SH, Mallie JP, Oh MS, Halperin ML: Acute hyponatremia in the perioperative

period: Insights into its pathophysiology and recommendations for management. Clin Nephrol1998;50(6):352–360.

68 Gross P: Treatment of severe hyponatremia. Kidney Int 2001;60(6):2417–2427.69 Lauriat SM, Berl T: The hyponatremic patient: Practical focus on therapy. J Am Soc Nephrol

1997;8(10):1599–1607.70 Sarnaik AP, Meert K, Hackbarth R, Fleischmann L: Management of hyponatremic seizures in

children with hypertonic saline: A safe and effective strategy. Crit Care Med 1991;19(6):758–762.71 Soupart A, Decaux G: Therapeutic recommendations for management of severe hyponatremia:

Current concepts on pathogenesis and prevention of neurologic complications. Clin Nephrol1996;46(3):149–169.

72 Verbalis JG: Adaptation to acute and chronic hyponatremia: Implications for symptomatology,diagnosis, and therapy. Semin Nephrol 1998;18(1):3–19.

73 Hantman D, Rossier B, Zohlman R, Schrier R: Rapid correction of hyponatremia in the syndromeof inappropriate secretion of antidiuretic hormone: An alternative treatment to hypertonic saline.Ann Intern Med 1973;78(6):870–875.

74 Fraser CL, Arieff AI: Epidemiology, pathophysiology, and management of hyponatremic ence-phalopathy. Am J Med 1997;102(1):67–77.

75 Worthley LI, Thomas PD: Treatment of hyponatraemic seizures with intravenous 29.2% saline.Br Med J (Clin Res Ed) 1986;292(6514):168–170.

76 Norenberg MD, Leslie KO, Robertson AS: Association between rise in serum sodium and centralpontine myelinolysis. Ann Neurol 1982;11(2):128–135.

77 Kleinschmidt-DeMasters BK, Norenberg MD: Rapid correction of hyponatremia causes demyeli-nation: Relation to central pontine myelinolysis. Science 1981;211(4486):1068–1070.

78 Sterns RH: Treating hyponatremia: Why haste makes waste. South Med J 1994;87(12):1283–1287.79 Ayus JC, Krothapalli RK, Armstrong DL: Rapid correction of severe hyponatremia in the rat:

Histopathological changes in the brain. Am J Physiol 1985;248(5 Pt 2):F711–F719.80 Ayus JC, Krothapalli RK, Armstrong DL, Norton HJ: Symptomatic hyponatremia in rats: Effect

of treatment on mortality and brain lesions. Am J Physiol 1989;257(1 Pt 2):F18–F22.81 Soupart A, Penninckx R, Stenuit A, Decaux G: Azotemia (48 h) decreases the risk of brain damage

in rats after correction of chronic hyponatremia. Brain Res 2000;852(1):167–172.

Moritz/Ayus 156

82 Soupart A, Ngassa M, Decaux G: Therapeutic relowering of the serum sodium in a patient afterexcessive correction of hyponatremia. Clin Nephrol 1999;51(6):383–386.

83 Wright DG, Laureno R, Victor M: Pontine and extrapontine myelinolysis. Brain 1979;102(2):361–385.

84 Brunner JE, Redmond JM, Haggar AM, Kruger DF, Elias SB: Central pontine myelinolysis andpontine lesions after rapid correction of hyponatremia: A prospective magnetic resonance imagingstudy. Ann Neurol 1990;27(1):61–66.

85 Kumar SR, Mone AP, Gray LC, Troost BT: Central pontine myelinolysis: Delayed changes on neu-roimaging. J Neuroimaging 2000;10(3):169–172.

86 Leens C, Mukendi R, Foret F, Hacourt A, Devuyst O, Colin IM: Central and extrapontine myelinol-ysis in a patient in spite of a careful correction of hyponatremia. Clin Nephrol 2001;55(3):248–253.

87 Palevsky PM, Bhagrath R, Greenberg A: Hypernatremia in hospitalized patients. Ann Intern Med1996;124(2):197–203.

88 Snyder NA, Feigal DW, Arieff AI: Hypernatremia in elderly patients. A heterogeneous, morbid,and iatrogenic entity. Ann Intern Med 1987;107(3):309–319.

89 Moritz ML, Ayus JC: The changing pattern of hypernatremia in hospitalized children. Pediatrics1999;104(3 Pt 1):435–439.

90 Polderman KH, Schreuder WO, Strack van Schijndel RJ, Thijs LG: Hypernatremia in the intensivecare unit: An indicator of quality of care? Crit Care Med 1999;27(6):1105–1108.

91 Ayus JC, Armstrong DL, Arieff AI: Effects of hypernatraemia in the central nervous system andits therapy in rats and rabbits. J Physiol 1996;492(Pt 1):243–255.

92 Finberg L, Luttrell C, Redd H: Pathogenesis of lesions in the nervous system in hypernatremicstates. II. Experimental studies of gross anatomic changes and alterations of chemical compositionof the tissues. Pediatrics 1959;23(1 Pt 1):46–53.

93 Luttrell CN, Finberg L: Hemorrhagic encephalopathy induced by hypernatremia. I. Clinical, lab-oratory, and pathological observations. AMA Arch Neurol Psychiatry 1959;81(4):424–432.

94 Grant PJ, Tate GM, Hughes JR, Davies JA, Prentice CR: Does hypernatraemia promote thrombosis?Thromb Res 1985;40(3):393–399.

95 Clark WR: Diffuse demyelinating lesions of the brain after the rapid development of hypernatremia.West J Med 1992;157(5):571–573.

96 Brown WD, Caruso JM: Extrapontine myelinolysis with involvement of the hippocampus in threechildren with severe hypernatremia. J Child Neurol 1999;14(7):428–433.

97 Soupart A, Penninckx R, Namias B, Stenuit A, Perier O, Decaux G: Brain myelinolysis followinghypernatremia in rats. J Neuropathol Exp Neurol 1996;55(1):106–113.

98 Fraser CL, Arieff AI: Hepatic encephalopathy. N Engl J Med 1985;313(14):865–873.99 FINBERG L: Pathogenesis of lesions in the nervous system in hypernatremic states. I. Clinical

observations of infants. Pediatrics 1959;23(1 Pt 1):40–45.100 Abramovici MI, Singhal PC, Trachtman H: Hypernatremia and rhabdomyolysis. J Med 1992;

23(1):17–28.101 Moritz ML: Hypernatremia in hospitalized patients. Ann Intern Med 1996;125(10):860.102 Mandal AK, Saklayen MG, Hillman NM, Markert RJ: Predictive factors for high mortality in

hypernatremic patients. Am J Emerg Med 1997;15(2):130–132.103 Warren SE, Mitas JA, 2nd., Swerdlin AH: Hypernatremia in hepatic failure. JAMA 1980;243(12):

1257–1260.104 Nelson DC, McGrew WR Jr, Hoyumpa AM Jr: Hypernatremia and lactulose therapy. JAMA 1983;

249(10):1295–1298.105 Mandell F, Fellers FX: Hyperglycemia in hypernatremic dehydration. Clin Pediatr (Philad)

1974;13(4):367–369.106 Stevenson RE, Bowyer FP: Hyperglycemia with hyperosmolal dehydration in nondiabetic infants.

J Pediatr 1970;77(5):818–823.107 Banister A, Matin-Siddiqi SA, Hatcher GW: Treatment of hypernatraemic dehydration in infancy.

Arch Dis Child 1975;50(3):179–186.108 Rosenfeld W, deRomana GL, Kleinman R, Finberg L: Improving the clinical management of

hypernatremic dehydration: Observations from a study of 67 infants with this disorder. ClinPediatr (Philad) 1977;16(5):411–417.


109 Pizarro D, Posada G, Levine MM: Hypernatremic diarrheal dehydration treated with "slow" (12-hour) oral rehydration therapy: A preliminary report. J Pediatr 1984;104(2):316–319.

110 Hogan GR, Pickering LK, Dodge PR, Shepard JB, Master S: Incidence of seizures that followrehydration of hypernatremic rabbits with intravenous glucose or fructose solutions. Exp Neurol1985;87(2):249–259.

111 Hogan GR, Dodge PR, Gill SR, Pickering LK, Master S: The incidence of seizures after rehydra-tion of hypernatremic rabbits with intravenous or ad libitum oral fluids. Pediatr Res 1984;18(4):340–345.

112 Hogan GR, Dodge PR, Gill SR, Master S, Sotos JF: Pathogenesis of seizures occurring duringrestoration of plasma tonicity to normal in animals previously chronically hypernatremic. Pediatrics1969;43(1):54–64.

113 Pizarro D, Posada G, Villavicencio N, Mohs E, Levine MM: Oral rehydration in hypernatremicand hyponatremic diarrheal dehydration. Am J Dis Child 1983;137(8):730–734.

114 Verbalis JG: Disorders of body water homeostasis. Best Pract Res Clin Endocrinol Metab2003;17(4):471–503.

115 Kahn T: Hypernatremia with edema. Arch Intern Med 1999;159(1):93–98.116 Pazmino PA, Pazmino BP: Treatment of acute hypernatremia with hemodialysis. Am J Nephrol

1993;13(4):260–265.117 Chai J, Diao L, Sheng Z, Guo Z, Gao W, Jia X: Heparin-free hemodialysis in the treatment of

hypernatremia in severely burned patients. Burns 2000;26(7):634–637.118 Moritz ML, del Rio M, Crooke GA, Singer LP: Acute peritoneal dialysis as both cause and treat-

ment of hypernatremia in an infant. Pediatr Nephrol 2001;16(9):697–700.119 Nolph KD, Hano JE, Teschan PE: Peritoneal sodium transport during hypertonic peritoneal dialysis:

Physiologic mechanisms and clinical implications. Ann Intern Med 1969;70(5):931–941.

Juan C. Ayus, MD1967 Haddon St., 3705 Fifth Ave, Houston, TX 77019–5762 (USA)Tel. �1 713 942 8434, Fax �1 713 942 9342, E-Mail [email protected]


Rasburicase Therapy in AcuteHyperuricemic Renal Dysfunction

Claudio Roncoa, Rinaldo Bellomob, Paola Inguaggiatoc,Monica Bonelloa, Valeria Bordonia, Gabriella Salvatoria, Vincenzo D’Intinia, Ranistha Ratanarata

aDepartment of Nephrology, St. Bortolo Hospital, Vicenza, Italy; bDepartment of Intensive Care Austin & Repatriation Medical Center, Heidelberg Victoria, Melbourne, Australia; cDepartment of Nephrology, S. Croce-Carle Hospital, Cuneo, Italy

Among the different causes of acute renal failure (ARF) in the hospitalizedpatient, cancer may represent an important factor leading to kidney dysfunc-tion. The underlying pathological condition generated by a neoplastic disorder,depends very much on the type of cancer: Solid tumors involving the kidneymay cause Acute renal failure by obstruction, parenchymal disruption and vas-cular thrombosis; solid tumors (not of hematological origin) not primarilyinvolving the kidney may cause ARF especially in the context of the tumor lysissyndrome (primary or secondary to therapy) [1]. Among these we may list solidtumors, small cell lung cancer, metastatic breast cancer, metastatic medul-loblastoma, metastatic seminoma, ovarian cancer and hepatoblastoma [2]. Finally,hematological disorders such as high-grade non-H lymphoma, acute lymphoidleukemia, acute myeloid leukemia and chronic myeloid leukemia in theblastic phase, may cause ARF especially during hemolytic crisis, and tumorlysis syndrome (TLS) [3]. Paraproteinemia might represent another importantdisorder leading to renal dysfunction during multiple myeloma [4].

The mechanisms by which renal parenchyma can be damaged includehemodynamic changes and the presence of circulating toxins. The additionalpresence of obstruction or compression may further contribute to the syndrome.Whatever is the primary cause of the renal insult, a final common pathwayleading to ARF includes intrarenal vasoconstriction and medullary ischemia,tubular obstruction, decreased glomerular filtration and back-leak of ultrafiltratein the interstitial space. Several other mechanisms may take part in the process

Pharmacological Issues in ARF and Sepsis

Rasburicase Therapy in Acute Hyperuricemic Renal Dysfunction 159

of renal damage including loss of cellular polarity, loss of tight junction gatefunction, loss of cell substrate adhesion, exfoliation of viable cell from tubularbasement membrane, aberrant cell-cell adhesion. Finally, the presence of alteredgene expression and the process of cellular dedifferentiation lead to lethal cellinjury with the two mechanisms of necrosis and apoptosis [5].

The mechanisms of cell necrosis in ARF are a severe depletion of ATP stores,a reduced activity of membrane transport, cell swelling, increase in intracellularfree calcium, and activation of phospholipases and proteases. Apoptosis mayalso be enhanced in ARF due to the increased concentration of physiologic acti-vators, the loss of renal growth factors, the impaired cell matrix adhesion andthe loss of cell-cell adhesion. The presence of cytotoxic agents such as reactiveoxygen species, intracellular free calcium, pharmacological agents and chemother-apy or the effect of physical factors such as hyperthermia or irradiation, mayfurther contribute to enhance the cellular apoptotic patterns.

Uric Acid:A New Toxin?

The presence of a neoplastic disorder and associated acute illness maysometimes lead to the presence of immunodysfunction, septic complications andmultiple organ dysfunction. In these settings the patient becomes critically illand the clinical picture is sustained by a series of circulating biochemical medi-ators of inflammation and endothelial damage. Among these substances, in cancerpatients there is a high chance of high circulating levels of uric acid [6].

Hyperuricemia is a potentially serious complication in patients with neo-plasms characterized by rapid cell proliferation and destruction.

Hyperuricemia is caused by the increased purine metabolism due to theenhanced catabolism of nucleic acid, result of the increased cellular turnover(primary TLS), or by the aggressive cancer chemotherapy regimens, that worsencell lysis and release of purine metabolites (secondary TLS).

Hyperuricemia may lead to acute renal failure through different mecha-nisms (fig. 1).

Mechanical obstruction of the tubular lumen due to precipitation of uratecrystals is one mechanism. A direct toxicity on renal tubular cells has also beenpostulated: this may be caused by a direct endothelial damage mediated by nitricoxide, an increased local oxidant stress on renal tissue and finally by abnormalplatelet activation. Finally, the potential effect of high level of uric acid on localand systemic inflammation is another emerging pathway: Uric acid levels corre-late with cytokine levels, they stimulate the synthesis of monocyte chemoattractantprotein 1 (MCP1) and they stimulate the activation of monocytes with increasedproduction of tumor necrosis factor and interleukin-1� (table 1).

Ronco/Bellomo/Inguaggiato/Bonello/Bordoni/Salvatori/D’Intini/Ratanarat 160

Moreover, uric acid is able to prevent the inactivation of superoxide dis-mutase by hydrogen peroxide that contributes to oxidant stress in vivo; highuric acid levels dramatically increase the vascular homeostasis of superoxidedismutase [7].

Uric acid A new

pathogenetic factor in ARF

Mechanical obstruction

Direct toxicity

Local and systemic inflammation

Afferent arteriole

Glome- rulus

Arcuatearteryand vein

Proximal convolutedtubule

Loop of Henle

Collec- tingduct

Vasarecta

Fig. 1. Uric acid is involved in acute renal damage with three different mechanisms:mechanical obstruction of the tubular lumen, direct toxicity on renal cells, and induction ofinflammation response.

Table 1. Uric acid – A new pathogenetic factor in ARF

• Tubular obstruction (precipitation of urate crystals)• Tissue damage

– Endothelial dysfunction (NO mediated)– Oxidant stress and cell dysruption– Platelet activation

• Acute inflammation– Correlates with circulating cytokines– Stimulates synthesis of MCP1 (monocyte

chemoattractant protein 1)– Stimulates monocyte production of IL-1� and TNF-�


Paradoxically, uric acid has also been reported to increase oxidative damage,to enhance platelet adhesiveness and to stimulate vascular smooth muscle cellgrowth [8].

In vitro, uric acid is effective in scavenging free radicals and can chelatetransition metal ions.

It has been also hypothesized that uric acid might be an evolutionaryantioxidant substitute for the loss of the ability to synthesize ascorbate. Hence,the overall functional importance of urate in vivo remains unknown.

Rationale for Rasburicase Therapy

The leading cause of acute hyperuricemic states is the TLS. In this settingand in other hyperuricemic disorders, the precipitation of urate crystals in thetubular lumen may cause obstruction while the precipitation in the interstitialtissue may cause inflammation and edema. Often these disorders are accompa-nied by disorders of phosphate metabolism with formation of calcium phos-phate and its precipitation in the renal tissue. All these events may contribute toan acute renal dysfunction. Even in the absence of a hyperuricemic state as aprimary insult to the kidney, the development of such disorder, might representa concomitant factor aggravating other previous or simultaneous insults. Forthis reason, all efforts should be made to protect the kidney for multiple insultsand thus to prevent hyperuricemic states [9]. The strategy includes the identifi-cation of a patient at risk, the actuation of preventive measures and finally theimplementation of a therapeutic plan in the early phases of the syndrome.Among the first, we count fluid resuscitation and maintenance of blood volumeand pressure. This ensures the perfusion of the kidney even in the presence ofa loss of autoregulation. In the second step, the maintenance of a urine flow isimportant to reduce the chance of obstruction and tubular damage. In spite oftheir satisfactory results in animal models, most renoprotective drugs do notappear to be efficacious in human clinical conditions. In the final group ofpreventive-therapeutic measures, we can hypothesize the utilization of drugslowering uric acid concentration.

In the therapy of hyperuricemic states and renal failure, allopurinol hasbeen used but unfortunately it does not degrade the uric acid already present, itblocks production of new uric acid but it also causes the accumulation ofxanthine and hypoxanthine, potentially toxic substances, it inhibits the degra-dation of the chemotherapeutic agents 6-mercaptopurine and azathioprine,increasing their chemotoxicity and finally it may worsen renal failure.

Urate oxidase is the enzyme that oxidizes uric acid in allantoin, i.e. five toten times more soluble than uric acid. Such enzyme is not present in humans


and other primates. It has been extracted from Aspergillus flavus in 1968, andcommercialized in France since 1975 and in Italy since 1984 as Uricozyme®.This molecule reduces uric acid plasma concentration more effectively thanallopurinol and it corrects renal dysfunction more rapidly than allopurinol;however, being non-recombinant it is associated with acute hypersensitivityreactions (urticaria, bronchospasm and/or hypoxemia occur in 5% of patients,even without any history of allergy) and has a production process with a lowyield [10].

Rasburicase is the recombinant form of urate oxidase. The recombinant ver-sion has been obtained from a modified Saccharomyces cerevisiae and recentlycommercialized as Fasturtec®. It is well tolerated at a dose of 0.2 mg/kg/day i.v.(T1/2 of 21.2 � 12 h). Rasburicase has been shown to decrease uric acid from 9.7 to 1.0 mg/dl after one injection. It presents less allergic reactions thanUricozyme® but nevertheless, caution in patients with a history of severe allergies

Allopurinol

Xanthine

Xanthine oxidase

Uric acid (urinary excretion)

Allantoin (urinary excretion)

Purine pathway Purine catabolism

Urate oxidase

HN

NH

O

O NH

NHH

HN

O

NH2

N

O HN

O H

O

Fig. 2. Urate oxidase is the enzyme that oxidizes uric acid in allantoin, i.e. five to tentimes more soluble than uric acid. Instead, allopurinol inhibits xanthine oxidase which cata-lyzes the conversion of xanthine in uric acid.


0

�100

�80

�60

�40

�20

Rasburicase dose level (mg/kg)

0.00 0.05 0.10 0.15 0.20

% R

educ

tion

at 2

h• Analysis of uric acid concentrations showed that the rate of decline in uric acid increases with increased dose of rasburicase (0.05mg/kg to 0.20mg/kg)

• At 2h after the start of infusion, linear regression of percentage change from baseline was statistically significant (p�0.0001)

Fig. 3. Urate oxidase has a dose-dependent mechanism of action. At 2 h after the admin-istration of 0.2 mg/kg of rasburicase, the reduction rate of uric acid was 80% from baseline.

0

2

4

6

8

10

12

0 12 36 60 84 108 132 144 0 12 36 60 84 108 1320

1.5

1.0

0.5

Uric

aci

d/p

hosp

horu

s (m

g/d

l)

Cre

atin

ine

(mg/

dl)

Rasburicase Allopurinol

Time from start of treatment (h) Time from start of treatment (h)

Phosphorus

Uric acid

Creatinine

Creatinine

Phosphorus

Uric acid

Fig. 4. Urate oxidase is more efficient than allopurinol in reducing uric acid plasmalevels. Moreover, urate oxidase does not induce hyperphosphoremia and is associated withameliorated renal function in the first 5 days after the start of treatment.


is suggested. There is no need of urine alkalinization. The molecule should beavoided in pregnancy and G6PDH deficiency syndromes [11].

In figure 2, the mechanism of action of urate oxidase and the structure ofthe molecule is depicted.

Rasburicase has been shown to reduce the levels of uric acid much fasterthan allopurinol and it seems to have a uricolytic effect correlated with the dose(fig. 3). The drug tested in children with malignancies resulted 6 times moreefficient than allopurinol [9, 12].

Allantoin excretion increases in parallel with the reduction of uric acidconcentration. The amount of allantoin excreted in urine is a direct measure ofthe removal of both the plasma uric acid present before chemotherapy as wellas the removal of any additional uric acid produced by chemotherapy-inducedTLS.

The drug appears to be beneficial as well in preventing an increase in con-centration of uric acid. Furthermore, it has been shown to present an effect oncreatinine levels as compared to allopurinol (fig. 4).

Conclusion

The presence of elevated concentrations of uric acid seems to represent anadditional risk factor in acute renal failure or in the case of initial renal dys-function. The early treatment of hyperuricemic states appears to be beneficialon renal function and we might hypothesize that such approach may representan important protective measure for the kidney. Further studies should elucidatethe mechanisms involved in this setting and should clarify the clinical rationalefor this type of therapy.

References

1 Jeha S: Tumor lysis syndrome. Semin Hematol 2001;38(suppl 10):4–8.2 Baeksgaard L, Sorensen JB: Acute tumor lysis syndrome in solid tumors: A case report and review

of the literature. Cancer Chemother Pharmacol 2003;51:187–192.3 Yang SS, Chau T, Dai MS, Lin SH: Steroid-induced tumor lysis syndrome in a patient with

preleukemia. Clin Nephrol 2003;59:201–205.4 Sile S, Wall BM: Acute renal failure secondary to spontaneous acute tumor lysis syndrome in

myelofibrosis. Am J Kidney Dis 2001;38:E21.5 Arrambide K, Toto RD: Tumor lysis syndrome. Semin Nephrol 1993;13:273–280.6 Ames BN, Cathcart R, Schwiers E, Hochstein P: Uric acid provides an antioxidant defense in

humans against oxidant- and radical-caused aging and cancer: A hypothesis. Proc Natl Acad SciUSA 1981;78:6858–6862.

7 Hink UH, Santanam N, Dikalov S, McCann L, Nguyen AD, Parthasarathy S, Harrison DG, Fukai T:Peroxidase properties of extracellular superoxide dismutase: Role of uric acid in modulatingin vivo activity. Arterioscler Thromb Vasc Biol 2002;22:1402–1408.


8 Rao GN, Corson MA, Berk BC: Uric acid stimulates vascular smooth muscle cell proliferation byincreasing platelet-derived growth factor A-chain expression. J Biol Chem 1991;266:8604–8608.

9 Cairo MS: Prevention and treatment of hyperuricemia in hematological malignancies. ClinLymphoma 2002;3(suppl 1):S26–S31.

10 Mahmoud HH, Leverger G, Patte C, Harvey E, Lascombes F: Advances in the management ofmalignancy-associated hyperuricemia. Br J Cancer 1998;77(suppl 4):18–20.

11 Pui CH, Relling MV, Lascombes F, Harrison PL, Struxiano A, Mondesir JM, Ribeiro RC,Sandlund JT, Rivera JK, Evans WE, Mahmoud HH: Urate oxidase in prevention and treatment ofhyperuricemia associated with lymphoid malignancies. Leukemia 1997;11:1813–1816.

12 Wolf G, Hegewisch-Becker S, Hossfeld DK, Stahl RA: Hyperuricemia and renal insufficiencyassociated with malignant disease: Urate oxidase as an efficient therapy? Am J Kidney Dis 1999;34:E20.

Claudio Ronco, MDDepartment of Nephrology, St. Bortolo HospitalViale Rodolfi, IT–36100 Vicenza (Italy)Tel. �39 0444993869, Fax �39 0444993949, E-Mail [email protected]


Diuretics in Acute Renal Failure?

Miet Schetz

Department of Intensive Care Medicine, University Hospital Gasthuisberg, Leuven, Belgium

Oliguria is generally recognized as a bad prognostic sign in patients withacute renal failure (ARF) [1–4]. The temptation to increase urine output in patientswith or at risk for ARF is therefore great. A survey by the European Workgroup ofCardiothoracic Intensivists reported that 11 of 38 centers employed furosemidecontinuously for ‘renoprotection’ and 34 used furosemide bolus injectionswhen diuresis decreased [5]. In a recent observational study in the US, 59% ofthe patients with ARF received diuretics at the time of nephrology consultation[6] and in a large international epidemiological study 70% of the patientsreceived diuretics before the start of renal replacement therapy [Bellomo et al.,unpubl. data]. Despite this ubiquitous use, it is not clear whether loop diureticsbeneficially affect renal function in ARF. This chapter provides evidence for theopposite being nearer to the truth.

Loop Diuretics in ARF: Experimental Evidence

Perfusion of an isolated rat kidney with furosemide in a hypoxic solutionresults in decreased damage to the medullary thick ascending limb of the loopof Henle (mTAL segment) and the S3 segment of the proximal tubule [7]. Theeffects of furosemide in experimental models of toxic tubular necrosis aresomewhat controversial. In glycerol-induced [8], uranyl-nitrite-induced [9] aswell as in myoglobin-induced ARF [10], no effect of furosemide on renal functioncould be established although some studies find an increase of urine output.Other investigators, however, found either a protective effect in HgCl2-inducedARF [11, 12], or a deleterious effect in glycerol-induced [10, 13, 14] orgentamicin-induced [15] ARF. Uffermann et al. [16] showed that the deleteri-ous effect of furosemide in HgCl2-induced ARF could be prevented by the

Diuretics in ARF 167

replacement of urinary sodium losses. In experimental models of ischemic ARFno effect [17–19], renoprotection [15, 20–26] or even deleterious effects [11, 27]have been found. In contrast nephropathy a protective effect of pretreatmentwith furosemide has also been established [28]. On the other hand, furosemideappeared not to be able to prevent the endotoxin-induced decrease of renalblood flow (RBF) in dogs [29]. In these experimental models furosemide ismostly administered before or immediately after the toxic or ischemic insult.

Loop Diuretics in ARF: Clinical Evidence

In healthy volunteers most authors observed an increase of glomerular fil-tration rate (GFR) with furosemide, although this is not a consistent finding[30–32]. Several studies evaluated the effect of prophylactic administration ofloop diuretics in patients at risk of ARF (table 1). In an uncontrolled observationin ICU patients, Lukas did not observe changes in GFR, RBF or RBF distribu-tion. However 9% of the patients developed hypotension [33]. More recent trialswith better design investigated the prophylactic use of loop diuretics in the set-ting of radiocontrast examination and in the perioperative setting. Weinsteinprospectively randomized 18 patients with chronic renal insufficiency undergo-ing contrast examination: the control group received fluid administration aloneand the study group a combination of furosemide and fluids. He found moredeterioration of renal function in the furosemide group, which is not unexpectedin view of a more pronounced negative fluid balance [34]. Another well-designedclinical trial randomized 78 patients with chronic renal insufficiency requiringcardiac angiography in three groups: the control group received 0.45% NaCl ata rate of l ml/kg/h starting 12 before and continued until 12 h after the contrastadministration, whereas the two treatment groups received a similar fluid regi-men combined with either 80 mg furosemide or 25 g mannitol, respectively, 30and 60 min before the contrast administration. Despite comparable changes inbody weight, the increase of serum creatinine and the incidence of ARF, definedas an increase of serum creatinine of at least 0.5 mg/dl, were significantly higherin the furosemide group. Mannitol also did not show a protective effect [35].A prospective, randomized, controlled trial studied 100 patients with serum cre-atinine exceeding 1.8 mg/dl undergoing coronary angiography. The controlgroup received placebo combined with fluid therapy at a rate of 150 ml/h. Theother patients received a combination of fluids, furosemide 1 mg/kg anddopamine, with addition of mannitol if the left filling pressure appeared to belower than 20 mm Hg. After the procedure the urinary losses of all patients werecompensated with intravenous fluid administration. Despite a higher urine out-put, the administration of diuretics appeared to have no effect on the incidence

Schetz 168

of renal insufficiency or the requirement for dialysis [36]. Lassnigg et al. [5] per-formed a placebo-controlled double-blind randomized trial on the effect ofdopamine and furosemide in 126 patients with normal preoperative renal func-tion undergoing cardiac surgery. Furosemide administration was started at a rateof 0.5 �g/kg/min at induction (or 2.5 mg/h for an 80-kg patient) and continueduntil 48 h postoperatively. Patients in the furosemide group had a significantlyhigher urine output but this appeared to be at the cost of a more pronouncedpostoperative increase of serum creatinine. A reduction of preload was notdeemed responsible for this finding since the increased urine output was com-pensated with fluid administration and the groups did not differ with regard toblood pressure or filling pressures [5]. In summary, clinical trials on the prophy-lactic use of loop diuretics in patients at risk for ARF do not suggest a beneficialeffect on renal function and even provide some evidence for a deleterious effect.

Table 1. Clinical trials on the prophylactic use of loop diuretics in ARF

Author n Setting Design Outcome parameter Mortality

Lukas 54 ICU patients uncontrolled no effect on GFR, RBF NR[33] 9% hypotension

Weinstein 34 radiocontrast PRC increase of Screat body weight NR[34] CRI Fluid 2 �mol/l �1.3

Fluid � F 37 �mol/l* �0.7

Solomon 105 radiocontrast PRC increase of Screat body weight no effect[35] CRI saline 0.1 mg/dl �0.49

saline � F 0.5 mg/dl* �0.78saline � M 0.3 mg/dl �0.23

Stevens 66 radiocontrast PRPC increase of Screat fluid I/O no effect[36] CRI saline � 0.51 mg/dl 1.3

placebo 0.56 mg/dl 0.91saline � F �M � D

Lassnigg 104 cardiac surgery PRPC increase of Screat no effect[5] nl RF placebo 0.14 mg/dl

F 0.40 mg/dl*D 0.22 mg/dl

PRC � Prospective randomized controlled; PRPC � prospective randomized placebo-controlled; CRI �chronic renal insufficiency; RF � renal function; UO � urine output; F � furosemide; M � mannitol;D � dopamine; GFR � glomerular filtration rate; RBF � renal blood flow; Screat � serum creatinine; NR � notreported.

*p � 0.05.


Two older prospective randomized trials demonstrate a beneficial effect offurosemide on urinary output in patients with established oliguric ARF. However,a beneficial effect on renal function, the need for renal replacement therapy ormortality was not found [37, 38]. Such effects are only mentioned in retrospec-tive or uncontrolled trials [39–41]. A more recent prospective, double-blind,placebo-controlled randomized trial investigated the effect of furosemide in 92 patients with established oliguric ARF, excluding patients with pre- and post-renal failure. The study group received furosemide at a dose of 3 mg/kg every6 h (during 21 days or until recovery of renal function or death), combined with100 ml mannitol 20% every 6 h and low-dose dopamine in continuous infusion.The diuretic regimen resulted in a significant increase of urine output, howeverwithout effect on the recovery of renal function, the need for dialysis or survival.48% of the patients in the furosemide group increased their urine output versus23% in the placebo group. These non-oliguric patients (that appeared to be lessseverely ill than the oliguric group) had a better survival, however without dif-ference between the diuretic and placebo group [42]. In 121 patients undergoingmajor thoraco-abdominal or vascular surgery furosemide 1 mg/h appeared tohave no effect on the postoperative decrease of creatinine clearance, nor on themortality or the length of stay in the ICU. This low dose of furosemide even didnot increase urine output, but was associated with significantly more hypokalemia[43]. Cotter et al. [44] randomized 20 patients with congestive heart failure inthree groups: a combination of dopamine with furosemide 80 mg/day (group A)or 5 mg/kg/day (group B) or a high dose furosemide (10 mg/kg/day) alone(group C). Despite a similar urine output and decrease of body weight, creatinineclearance deteriorated significantly in group B and C, an effect that the authorsattributed to a decreased blood pressure. The incidence of hypokalemia was alsosignificantly higher in the groups receiving larger doses of furosemide. Sirivellaet al. [45] compared the effect of a combined infusion of dopamine, furosemideand mannitol with intermittent administration of diuretics in 100 patients that‘despite an adequate cardiac output’ developed oliguric renal failure after car-diac surgery. The combined infusion resulted in a significantly higher diuresis,an improved time course of serum creatinine and a significantly decreased needfor dialysis. The beneficial effect appeared to be most pronounced when theinfusion was started within 6 h after the start of oliguria. The doses of intermit-tent diuretics were rather high (1.5–3 mg/kg furosemide every 4–6 h) and nodata are provided on the actually administered doses.

A recent retrospective survey in 326 intensive care patients with ARFrequiring nephrology consultation concluded that the use of diuretics at the timeof consultation was associated with an increased risk of death and nonrecoveryof renal function. The increased risk was borne largely by patients who wererelatively unresponsive to diuretics [6]. However, analysis of a prospectively

Schetz 170

derived international database of patients requiring renal replacement therapyfor ARF, could not confirm these findings [Bellomo, unpubl. data].

In conclusion, although most clinical studies report an increase of urineoutput, the evidence for a beneficial effect of loop diuretics on renal function islimited. Patients responding to loop diuretics with an increased urine output areprobably characterized by a less severe form of renal failure, rather than repre-senting a beneficial effect of the treatment. Some clinical studies even point toa deleterious effect of loop diuretics on renal function. On the other hand itshould be acknowledged that none of the available studies has enough statisticalpower. This is also the conclusion of two recent meta-analyses [46, 47].

Pathophysiology of ARF

In order to understand the effect of loop diuretics on renal function, a shortreminder of the pathophysiology of renal hypoperfusion is required. The nor-mal blood supply to the kidney is very heterogeneous with approximately 90%entering the cortex and 10% the medulla. These regional differences are relatedto the flow-dependent functions of the cortex (glomerular filtration and tubularreabsorption), whereas the function of the medulla, that consists in concentratingthe urine, requires a limited blood supply to prevent the washout of the inter-stitial osmotic gradient. In addition, the active tubular transport that is requiredto generate this gradient requires a high O2 consumption. The heterogeneousperfusion explains why, despite the large global blood supply, the kidney is verysusceptible to hypoperfusion and why ischemic damage is mainly found inthe tubular segments that are located in the outer medulla: the pars recta of theproximal tubulus (S3 segment) and the medullary thick ascending limb of theloop of Henle (mTAL) [48, 49].

The regulation of RBF is very complex and depends on the balance betweenvasoconstricting (the sympathoadrenal axis, the renin-angiotensin aldosteronsystem (RAAS) and arginine vasopressin) and vasodilator influences (vasodi-lating prostaglandins, mainly PGI2 and PGE2, atrial natriuretic peptide, kininsand nitric oxide (NO)) [50]. Both sympathetic stimulation and angiotensin IIpreferentially cause vasoconstriction in the cortex [51, 52] redistributing theblood flow to the deeper salt-retaining nephrons and the vulnerable medulla.A decreased RBF with corticomedullary redistribution has also been observedin sepsis [53].

A second mechanism controlling RBF and GFR is renal autoregulation,involving both myogenic autoregulation and the tubuloglomerular feedback(TGF) [50]. The TGF finetunes RBF and GFR by eliciting afferent vasocon-striction in response to an increased sodium chloride delivery at the macula


densa in the distal tubule, whereas decreased NaCl delivery decreases theafferent tone. The apical membrane receptor(s) involved in this sensing mech-anism have not been completely elucidated, but both the Na�-K�-2Cl– cotrans-porter (NK2CC) and the K-channel ROMK are unequivocally implicated.Experimental evidence suggests that adenosine is the mediator of the vasocon-strictive response via an adenosine-1 receptor (A1AR). The magnitude of theTGF response depends on the ambient level of angiotensin II. A second com-ponent of the juxtaglomerular apparatus is the macula densa control of reninsecretion, with a decreased distal tubular solute delivery stimulating reninsecretion. COX-2 mediated synthesis of prostaglandin E2 appears to be theunderlying mechanism of this increased renin release [54, 55]. As well theextrinsic systems, causing corticomedullary redistribution of blood flow bycortical vasoconstriction with or without medullary vasodilation, as the TGFcan be seen as protective mechanisms against excessive loss of salt and wateror, by diminishing the filtered solute load, as a tool to prevent nephron oxygendeficiency [56, 57].

With mild renal hypoperfusion, GFR can be maintained by compensatorymechanisms, consisting of afferent vasodilation (mainly mediated by prosta-glandins, NO and the myogenic and TGF component of autoregulation) andefferent vasoconstriction (mediated by angiotensin II) [58]. Prerenal failure,characterized by a decreased GFR with preserved tubular function, results whenthese compensatory mechanisms fail. Adaptive mechanisms to avoid medullaryhypoxia and ischemic tubular damage in prerenal failure are paracrine vasodila-tory mediators, such as adenosine (via adenosine-2 receptor), prostaglandinsand NO, that dilate medullary vessels and decrease solute reabsorption and theassociated O2 consumption [59]. The previously mentioned corticomedullaryredistribution of RBF not only increases blood supply to the vulnerable medullarytubular segments but also decreases GFR. This reduces solute delivery, decreasesmedullary work and restores the balance between O2 delivery and consumption[48, 60]. Prerenal failure has therefore been called ‘acute renal success’ becausethe reduction of glomerular filtration prevents further damage to the vulnerablemedullary tubular segments [56].

With ongoing renal hypoperfusion ischemic damage of tubular cells oracute tubular necrosis (ATN) ensues. Both vascular and tubular factors con-tribute to the further decrease of GFR in ATN [61, 62]. Vascular factors are apersisting renal vasoconstriction, a decrease of the permeability of the glomerularcapillaries and medullary congestion. Tubular factors are tubular obstruction bycellular debris and backleak of glomerular filtrate. The TGF is thought to playan important role in the vascular pathogenesis of ATN: damage to more proximaltubular segments increases the NaCl load to the macula densa causing afferentvasoconstriction [62–64].

Schetz 172

Rationale for Administering Loop Diuretics in ARF

Theoretically, there are several reasons to expect a beneficial effect of loopdiuretics on renal function in ARF:

(1) The diuretic effect increases tubular flow with a flushing effect preventingtubular obstruction [26]. Since the swollen obstructed tubuli can hamper venousoutflow from the medulla, prevention of tubular obstruction might also improveRBF [66].

(2) Loop diuretics may increase RBF by two mechanisms: induction ofcyclooxygenases (COX) with increased production of vasodilating prostaglandins[22, 30, 66–69] and interruption of the TGF by inhibition of the NK2CC thatinitiates the signaling at the macula densa [70–73].

(3) The inhibition of NK2CC by loop diuretics reduces active sodiumtransport. The associated reduction of O2 consumption may prevent ischemicdamage of the vulnerable medullary tubulus segments [7, 60, 74]. In an animalmodel of ARF characterized by severe medullary hypoxia, pretreatment withfurosemide prevented renal ATP depletion [75].

(4) Apoptosis has been shown to play an important role in the pathophysi-ology of ARF [76]. Recent evidence suggests that COX-2 expression inhibitsTNF-induced apoptosis in renal mesangial cells [77]. Since loop diureticsinduce COX-2 expression, they might also prevent apoptosis.

How to Explain the Absence of a Beneficial Effect?

Several explanations can be put forward to explain the absence of a protectiveeffect of diuretics on renal function in ARF.

(1) Excessive preload reduction activating both the RAAS and the adren-ergic system will result in renal vasoconstriction. Replacement of the urinarylosses has been shown to prevent furosemide-induced decreases in GFR [78, 79].In addition to the urinary losses, prostaglandin-mediated venodilation alsocontributes to a decreased preload [80, 81]. Maintenance of an adequate pre-load is therefore of utmost importance when administering loop diuretics in anattempt to prevent ARF. The optimal method to establish an ‘adequate preload’remains, however, a matter of debate [82–87]. This makes it difficult to verifya comparable filling status in patient assigned to different treatment arms ofrandomized trials.

(2) Loop diuretics do not improve RBF, or at least not under any circum-stances. Although some (mostly older) animal studies find an increase of RBFafter the administration of loop diuretics [20, 88–91], other experimental obser-vations suggest the opposite [79, 92–97]. The use of different methods to


measure RBF, differences in animal species, in the administered doses of loopdiuretics and in the experimental conditions may have contributed to thesecontradictory results. Greven et al. [98] found and increase of RBF in controlanimals and a decrease in glycerol-treated rats. Burke et al. [99] found thatfurosemide only increases RBF when the renal perfusion pressure is above thelower limit of autoregulation and suggested that at lower perfusion pressuresrenal vascular resistance is already close to a minimum value. Furosemide alsodoes not seem to affect the LPS-induced decline of RBF in anesthetizeddogs [29]. In an animal model of ischemic ARF, the protective effect of furo-semide was not influenced by a COX-inhibitor, questioning the role of vaso-dilatory prostaglandins in this setting [22]. A similar observation was madewith regard to the bumetanide-induced increase of RBF in the isolated perfusedrat kidney [91].

Most human studies measure RBF by PAH clearance, which is probablynot a valid method in critical illness and ARF [100]. In healthy volunteers,furosemide increases RBF [30, 67, 101], an effect that can be suppressed byCOX inhibitors [30]. However, no change is noted in critically ill patients[33, 102, 103] and a decrease is seen in patients with congestive heart failure[104]. This might be related to the inability of prostaglandins to counteractstrong vasoconstrictive effects. Loop diuretics indeed may provoke vasocon-striction by stimulating renin release and angiotensin II formation [101,105–107], an effect that is secondary to a decreased renal perfusion pressureand preload reduction resulting in adrenergic stimulation. Loop diuretics alsohave a direct effect at the macula densa where they diminish the NaCltransport-dependent renin-inhibitory signal to the granular cells, probablyinvolving COX-2 dependent prostanoids [55, 108, 109]. It is interesting tonote that in conscious animals a furosemide-induced renal vasoconstrictionor decrease of renal perfusion was not noted after the administration ofan angiotensin-converting enzyme inhibitor (ACEI) [79, 94, 96], pointing toa role for angiotensin II as a mediator of renal vasoconstriction duringfurosemide infusion. In addition, furosemide has been shown to stimulate therelease of the thromboxane [68].

Loop diuretics inhibit the TGF, which might indeed be protective inpatients who, as a result of ischemic damage to the proximal tubulus, have ahigh solute load at the macula densa. However, very high tubular concentrationsof furosemide have been shown to be necessary for complete inhibition of theTGF [110]. On the other hand, in patients with prerenal failure, the same TGFmight be protective against further ischemic damage of the vulnerable medul-lary segments.

In conclusion, the final effect of loop diuretics on RBF is the resultant ofseveral partially opposite effects. Added to the already disturbed balance between

Schetz 174

intrarenal vasodilator and vasoconstrictor substances, induced by the underlyingpathology, it is not surprising that the net effect of loop diuretics on RBC isdifficult to predict in patients.

(3) The absence of a beneficial effect of loop diuretics on renal functioncould also be explained by an adverse effect on intrarenal blood flow distri-bution. This intrarenal distribution is even more difficult to measure thantotal RBF. Using laser Doppler flowmetry, which is probably the most reli-able method, Dobrowolski et al. [112] confirmed the results of earlier inves-tigations [74, 88, 90, 111], and demonstrated a more pronounced decrease ofmedullary compared with cortical blood flow after furosemide administration[112, 113]. The furosemide-induced decrease in medullary blood flow can beattenuated by saline loading, an ACEI [113] or an AT(1) receptor antagonist[104], again pointing to a role of the RAAS in the medullary vasoconstric-tion. Castrop et al. [114] showed that furosemide, though increasing COXsexpression in the cortex, decreased medullary COXs mRNA. The differentialeffect of furosemide on cortical and medullary blood flow might endangerthe medullary O2 balance and shift a prerenal failure to an acute tubularnecrosis.

(4) Furosemide adversely affects autoregulation. Already in 1977, it wasshown that furosemide, by inhibiting the TGF, abolishes autoregulation of renalblood flow [115], a finding that has subsequently been confirmed [97, 116–118].The increased lower limit of autoregulation might explain why the prophylacticuse of furosemide can have a deleterious effect [5].

(5) It is possible that tubular obstruction is not prevented or remediedbecause the induced diuresis only relies on a decreased reabsorption in non-obstructed tubuli. An increased tubular aggregation of Tamm-Horsfall proteinshas even been described with loop diuretics [119].

Toxicity of Loop Diuretics

Loop diuretics may be associated with toxicity, including ototoxicity [120,121], interstitial nephritis, dermatological symptoms and arrhythmias due tofurosemide-induced hypokalemia [42, 44]. An even more important disadvantageof the use of loop diuretics is, however, the loss of important diagnostic para-meters due to furosemide-induced diuresis and natriuresis. Not only the urinaryindices, that allow to distinguish between prerenal and renal failure, becomeunreliable but also the urine output as parameter of cardiac output and tissueperfusion is lost. The loop diuretic-induced increase of diuresis often induces afeeling of therapeutic success and further measures to improve renal perfusionand prevent renal dysfunction are omitted.


In recent years, it has become evident that continuous administration ofloop diuretics results in an improved efficiency with decreased toxicity and shouldtherefore be preferred above bolus administration [122–128].

Conclusion

An increase of urine output is the only irrefutably proven effect of loop diuret-ics in ARF. This, of course, may have beneficial effects on the cardiorespiratorysystem, facilitates fluid management of critically ill patients and creates space foradequate nutrition. In addition, the loop-diuretic-induced increase of urinary out-put yields prognostic information, since it points to a less severe renal insult [129].

Table 2. Clinical trials on the therapeutic use of loop diuretics in ARF

Author n Setting Design Outcome parameter Outcome Mortalityparameter

Kleinknecht 66 established PR duration oliguria n dialysis NR[37] ARF F boluses 17d 20

controls 17d (p � 0.2) 22 (NS)

Brown 58 posttrauma PR increased UO n dialysis [38] postsurgery F bolus 6% 12 57%

ARF F bolus � 3 g/24 h 86% (p � 0.005) 10 (NS) 64% (NS)

Shilliday 92 oliguric ARF PRPC increased UO dialysis[42] Fluid 23% 40% 50%

Fluid � F 48% (p0.02) 31% (p0.87) 66% (NS)

Hager 121 postoperative PRPC decrease of ClCr[43] major surgery Placebo 19% 5%

F 17% (NS) 9% (NS)

Cotter 20 congestive PRC Change of Clcreat[44] heart failure D � low dose F �3 ml/min

D � medium dose F �25 ml/minHigh dose F �17 ml/min (p0.007)

Sirivella 100 cardiac surgery PR dialysis[45] postoperative intermittent F 90%

ARF infusion of 6.7% (p?)D � F � M

PRC � Prospective randomized controlled; PRPC � prospective randomized placebo-controlled; CRI � chronicrenal insufficiency; UO � urine output; F � furosemide; M � mannitol; D � dopamine; GFR � glomerular filtra-tion rate; RBF � renal blood flow; ARF � acute renal failure; n dialysis � number of patients requiring dialysis;NR � not reported.

Schetz 176

A beneficial effect of loop diuretics on renal function has, however, notbeen demonstrated, although it should be acknowledged that the available trialsdo not have adequate statistical power. In established ARF loop diuretics areprobably administered too late to be of any help. However, their prophylacticadministration in humans has also not been proven protective, presumablybecause they disturb renal autoregulation and the protective corticomedullaryredistribution. It remains to be established whether a restricted temporary gapexists in which loop diuretics might be effective. However, since the majorityof the available clinical trials does not show benefit and even suggest harm, alarge prospective randomized trial can only be justified if clinically relevantanimal models suggest a beneficial effect.

References

1 Anderson RJ, Linas SL, Berns AS, Henrich WL, Miller TR, Gabow PA, Schrier RW: Nonoliguricacute renal failure. N Engl J Med 1977;296:1134–1138.

2 Brivet FG, Kleinknecht DJ, Loirat P, Landais PJ: Acute renal failure in intensive care units: Causes,outcome, and prognostic factors of hospital mortality; a prospective, multicenter study. FrenchStudy Group on Acute Renal Failure. Crit Care Med 1996;24:192–198.

3 Liano F, Pascual J: Epidemiology of acute renal failure: A prospective, multicenter, community-based study. Madrid Acute Renal Failure Study Group. Kidney Int 1996;50:811–818.

4 Guerin C, Girard R, Selli JM, Perdrix JP, Ayzac L: Initial versus delayed acute renal failure in theintensive care unit: A multicenter prospective epidemiological study. Rhone-Alpes Area StudyGroup on Acute Renal Failure. Am J Respir Crit Care Med 2000;161:872–879.

5 Lassnigg A, Donner E, Grubhofer G, Presterl E, Druml J, Hiesmayr M: Lack of renoprotective effectsof dopamine and furosemide during cardiac surgery. J Am Soc Nephrol 2000;11:97–104.

6 Mehta RL, Pascual MT, Soroko S, Chertow GM: Diuretics, mortality, and nonrecovery of renalfunction in acute renal failure. JAMA 2002;288:2547–2553.

7 Heyman SN, Rosen S, Epstein FH, Spokes K, Brezis ML: Loop diuretics reduce hypoxic damageto proximal tubules of the isolated perfused rat kidney. Kidney Int 1994;45:981–985.

8 Kôlling B, Greven J: Untersuchungen zur Wirksamkeit von Furosemid bei experimentellen akutenNierenversagen. Res Exp Med Berl 1978;173:89–93.

9 Lindner A, Cutler RE, Goodman G: Synergism of dopamine plus furosemide in preventing acuterenal failure in the dog. Kidney Int 1979;16:158–166.

10 Bidani AK, Churchill PC, Packer W: Theophylline-induced protection in myoglobinuric acuterenal failure: Further characterization. Can J Physiol Pharmacol 1987;65:42–45.

11 Middeke M, Drasch G, Holzgreve H: Partial nephrectomy and furosemide in toxic and ischemicnonoliguric acute renal failure in rats. Res Exp Med Berl 1985;18:227–232.

12 Kurtz TW, Hsu CH: Effects of furosemide diuresis on mercuric chloride-induced acute renal failurein the rat. Nephron 1986;43:279–282.

13 Greven J, Klein H: Renal effects of furosemide in glycerol induced acute renal failure of the rat.Pflügers Arch 1976;365:81–87.

14 Lopez-Novoa JM, Rodicio-Diaz JL, Hernando-Avendano L: Negative effect of furosemide pre-treatment in glycerol induced acute renal failure. Biomedicine 1977;26:117–122.

15 de Rougemont D, Oeschger A, Konrad L, Thiel G, Torhorst J, Wenk M, Wunderlich P, Brunner FP:Gentamicin-induced acute renal failure in the rat: Effect of dehydration, DOCA-saline andfurosemide. Nephron 1981;29:176–184.

16 Ufferman RC, Jaenike JR, Freeman RB, Pabico RC: Effects of furosemide on low-dose mercuricchloride acute renal failure in the rat. Kidney Int 1975;8:362–367.


17 Papadimitriou M, Milionis A, Sakellariou G, Metaxas P: Effect of furosemide on acute ischaemicrenal failure in the dog. Nephron 1978;20:157–162.

18 Lewis RM, Rice JH, Patton MK, Barnes JL, Nickel AE, Osgood RW, Fried T, Stein JH: Renalischemic injury in the dog: Characterization and effect of various pharmacologic agents. J Lab ClinMed 1984;104:470–479.

19 Källskog Ö, Nygren K, Wolgast M: Failure of loop diuretics to improve the long term outcome ofischemic damage in the kidney. Ups J Med Sci 2001;106:151–160.

20 De Torrente A, Miller PD, Cronin RE, Paulsin PE, Erickson AL, Schrier RW: Effects of furosemideand acetylcholine in norepinephrine-induced acute renal failure. Am J Physiol 1978;235:F131–F136.

21 Patak RV, Fadem SZ, Lifschitz MD, Stein JH: Study of factors which modify the development ofnorepinephrine-induced acute renal failure in the dog. Kidney Int 1979;15:227–237.

22 Kramer HJ, Schuurmann J, Wassermann C, Dusing R: Prostaglandin-independent protection byfurosemide from oliguric ischemic renal failure in conscious rats. Kidney Int 1980;17:455–464.

23 Hanley MJ, Davidson K: Prior mannitol and furosemide infusion in a model of ischemic acute renalfailure. Am J Physiol 1981;241:F556–F564.

24 Sinsteden TD, O’Neil TJ, Hill S, Lifschitz MD, Stein JH: The role of high-energy phosphate innorepinephrine-induced acute renal failure in the dog. Circ Res 1986;59:93–104.

25 Kim SJ, Lim YT, Kim BS, Cho SI, Woo JS, Jung S, Kim YK: Mechanism of reduced GFR in rabbitswith ischemic acute renal failure. Ren Fail 2000;22:129–141.

26 Bayati A, Nygren K, Kallskog O, Wolgast M: The effect of loop diuretics on the long-term outcomeof post-ischaemic acute renal failure in the rat. Acta Physiol Scand 1990;139:271–279.

27 Dubourg L, Drukker A, Guignard JP: Failure of the loop diuretic furosemide to improve renal func-tion of hypoxemic vasomotor nephropathy in the newborn rabbit. Pediatr Res 2000;47:504–508.

28 Heyman SN, Brezis M, Greenfeld Z, Rosen S: Protective role of furosemide and saline inradiocontrast-induced acute renal failure in the rat. Am J Kidney Dis 1989;14:377–385.

29 Yao K, Ina Y, Nagashima K, Ohno T, Karasawa A: Effect of the selective adenosine A1-receptorantagonist KW-3902 on lypopolysaccharide-induced reductions in urine volume and renal bloodflow in anesthetized dogs. Jpn J Pharmacol 2000;84:310–315.

30 Passmore AP, Copeland S, Johnston GD: The effects of ibuprofen and indomethacin on renal functionin the presence and absence of furosemide in healthy volunteers on a restricted sodium diet. Br J ClinPharmacol 1990;29:311–319.

31 Guidi E, Colussi G, Rombola G, Airaghi C, Minetti E, Malberti F: Evaluation of the tubuloglomeru-lar feedback system in human subjects. Exp Nephrol 1995;3:61–64.

32 Epstein FH, Prasad P: Effects of furosemide on medullary oxygenation in younger and older subjects.Kidney Int 2000;57:2080–2083.

33 Lucas CE, Zito JG, Carter KM, Cortez A, Stebner FC: Questionable value of furosemide in pre-venting renal failure. Surgery 1977;82:341–320.

34 Weinstein JM, Heyman S, Brezis M: Potential deleterious effect of furosemide in radiocontrastnephropathy. Nephron 1992;62:413–415.

35 Solomon R, Werner C, Mann D, D’Elia J, Silva P: Effects of saline, mannitol, and furosemide toprevent acute decreases in renal function induced by radiocontrast agents. N Engl J Med 1994;331:1416–1420.

36 Stevens MA, McCullough PA, Tobin KJ, Speck JP, Westveer DC, Guido-Allen DA, Timmis GC,O’Neill WW: A prospective randomized trial of prevention measures in patients at high risk forcontrast nephropathy: Results of the P.R.I.N.C.E. Study. Prevention of Radiocontrast InducedNephropathy Clinical Evaluation. J Am Coll Cardiol 1999;33:403–411.

37 Kleinknecht D, Ganeval D, Gonzalez-Duque LA, Fermanian J: Furosemide in acute oliguric renalfailure: A controlled trial. Nephron 1976;17:51–58.

38 Brown CB, Ogg CS, Cameron JS: High dose furosemide in acute renal failure: A controlled trial.Clin Nephrol 1981;15:90–96.

39 Cantarovich F, Galli C, Benedetti L, Chena C, Castro L, Correa C, Perez-Loredo J, Fernandez JC,Locatelli A, Tizado J: High dose frusemide in established acute renal failure. Br Med J 1973;4:449–450.

40 Minuth AN, Terrell JB Jr, Suki WN: Acute renal failure: A study of the course and prognosis of104 patients and of the role of furosemide. Am J Med Sci 1976;271:317–324.

Schetz 178

41 Benoit O, Lebleu J, Noel C, Dracon M, Dequiedt P, Lelievre G, Tacquet A: Aspects actuels de l’insuffisance renale aigue. Sem Hop 1982;58:2499–2503.

42 Shilliday IR, Quinn KJ, Allison ME: Loop diuretics in the management of acute renal failure:A prospective, double-blind, placebo-controlled, randomized study. Nephrol Dial Transplant 1997;12:2592–2596.

43 Hager B, Betschart M, Krapf R: Effect of postoperative intravenous loop diuretic on renal func-tion after major surgery. Schweiz Med Wochenschr 1996;126:666–673.

44 Cotter G, Weissgarten J, Metzkor E, Moshkovitz Y, Litinski I, Tavori U, Perry C, Zaidenstein R,Golik A: Increased toxicity of high-dose furosemide versus low-dose dopamine in the treatmentof refractory congestive heart failure. Clin Pharmacol Ther 1997;62:187–193.

45 Sirivella S, Gielchinsky I, Parsonnet V: Mannitol, furosemide, and dopamine infusion in postopera-tive renal failure complicating cardiac surgery. Ann Thor Surg 2000;69:501–506.

46 Kellum JA: The use of diuretics and dopamine in acute renal failure: A systematic review of theevidence. Crit Care 1997;1:53–59.

47 Kellum JA, Leblanc M, Venkataraman R: Acute renal failure. Clin Evidence 2002;8:829–848.48 Brezis M, Rosen S: Hypoxia of the renal medulla: Its implications for disease. N Engl J Med

1995;332:647–655.49 Sheridan AM, Bonventre JV: Pathophysiology of ischemic acute renal failure. Contrib Nephrol.

Basel, Karger, 2002, vol 132, pp 7–21.50 Sladen NS, Landry D: Renal blood flow regulation, autoregulation, and vasomotor nephropathy.

Anesthesiol Clins N Am 2000;18:791–807. 51 Evans RG, Madden AC, Denton KM: Diversity of responses of renal cortical and medullary blood

flow to vasoconstrictors in conscious rabbits. Acta Physiol Scand 2000;169:297–308.52 Guild SJ, Eppel GA, Malpas SC, Rajapakse NW, Stewert A, Evans RG: Regional responsiveness

of renal perfusion to activation of the renal nerves. Am J Physiol 2002;283:R1177–R1186.53 Garrison RN, Wilson MA, Matheson PJ, Spain DA: Nitric oxide mediates redistribution of intrarenal

blood flow during bacteremia. J Trauma 1995;39:90–97.54 Stichtenoth DO, Wagner B, Frolich JC: Effect of selective inhibition of the inducible cyclooxygenase

on renin release in healthy volunteers. J Investig Med 1998;46:290–296.55 Schnermann J: The juxtaglomerular apparatus: From anatomical peculiarity to physiological rele-

vance. J Am Soc Nephrol 2003;14:1681–1694.56 Thurau K, Boylan JW: Acute renal success: The unexpected logic of oliguria in acute renal failure.

Am J Med 1976;61:308–314.57 Schurek HJ, Johns O: Is tubuloglomerular feedback a tool to prevent nephron oxygen deficiency?

Kidney Int 1997;51:386-392.58 Myers BD, Moran SM: Hemodynamically mediated acute renal failure. N Engl J Med 1986;314:

97–105.59 Pallone TL, Zhang Z, Rhinehart K: Physiology of the renal microcirculation. Am J Physiol 2003;

284:F253–F266.60 Brezis M, Rosen S, Silva P, Epstein FH: Renal ischemia: A new perspective. Kidney Int 1984;26:

375–383.61 Nissenson AR: Acute renal failure: Definition and pathogenesis. Kidney Int Suppl 1998;66:

S7–S10.62 Bankir L, Kriz W, Goligorsky M, Nambi P, Thomson S, Blantz RC: Vascular contributions to

pathogenesis of acute renal failure. Ren Fail 1998;20:663–677.63 Mason J: Tubulo-glomerular feedback in the early stages of experimental acute renal failure. Kidney

Int Suppl 1976;6:S106–S114.64 Kribben A, Edelstein CL, Schrier RW: Pathophysiology of acute renal failure. J Nephrol 1999;12

(suppl 2):S142–S151.65 de Rougemont D, Brunner FP, Torhorst J, Wunderlich PF, Thiel G: Superficial nephron obstruction

and medullary congestion after ischemic injury: Effect of protective treatments. Nephron 1982;31:310–320.

66 Katayama S, Attallah AA, Stahl RA, Bloch DL, Lee JB: Mechanism of furosemide-inducednatriuresis by direct stimulation of renal prostaglandin E2. Am J Physiol 1984;247:F555–F561.


67 MacKay IG, Muir AL, Watson ML: Contribution of prostaglandins to the systemic and renal vas-cular response to furosemide in normal man. Br J Clin Pharmacol 1984;17:513–519.

68 Liguori A, Casini A, Di-Loreto M, Andreini I, Napoli C: Loop diuretics enhance the secretion ofprostacyclin in vitro, in healthy persons, and in patients with chronic heart failure. Eur J ClinPharmacol 1999;55:117–124.

69 Castrop H, Klar J, Wagner C, Hocherl K, Kurtz A: General inhibition of renocortical cyclooxyge-nase-2 expression by the renin-angiotensin system. Am J Physiol 2003;284:F518–F524.

70 Wright FS, Schnermann J: Interference with feedback control of glomerular filtration rate byfurosemide, triflocin, and cyanide. J Clin Invest 1974;53:1695–1708.

71 Schlatter E, Salomonsson M, Persson AE, Greger R: Macula densa cells sense luminal NaClconcentration via furosemide sensitive Na�2Cl-K� cotransport. Pflügers Arch 1989;414:286–290.

72 Casellas D, Moore LC: Autoregulation and tubuloglomerular feedback in juxtamedullary glomeru-lar arterioles. Am J Physiol 1990;258:F660–F669.

73 Johnston PA, Kau ST: The effect of loop of Henle diuretics on the tubuloglomerular feedback mech-anism. Methods Find Exp Clin Pharmacol 1992;14:523–529.

74 Brezis M, Agmon Y, Epstein FH: Determinants of intrarenal oxygenation. I. Effects of diuretics.Am J Physiol 1994;267:F1059–F1062.

75 Burke TJ, Malhotra D, Shapiro J: Effects of enhanced oxygen release from hemoglobin by RSR13in an acute renal failure model. Kidney Int 2001;60:1407–1414.

76 Bonegio R, Lieberthal W: Role of apoptosis in the pathogenesis of acute renal failure. Curr OpinNephrol Hypertens 2002;11:301–308.

77 Ishaque A, Dunn MJ, Sorokin A: Cyclooxygenase-2 inhibits tumor necrosis factor alpha-mediatedapoptosis in renal glomerular mesangial cells. J Biol Chem 2003;278:10629–10640.

78 Tucker BJ, Blantz RC: Effect of furosemide administration on glomerular and tubular dynamicsin the rat. Kidney Int 1984;26:112–121.

79 Bak M, Shalmi M, Petersen JS, Poulsen LB, Christensen S: Effects of angiotensin-converting enzymeinhibition on renal adaptations to acute furosemide administration in conscious rats. J PharmacolExp Ther 1993;266:33–40.

80 Silke B: Haemodynamic impact of diuretic therapy in chronic heart failure. Cardiology 1994;84(suppl 2):115–123.

81 Pickkers P, Dormans TP, Russel FG, Hughes AD, Thien T, Schaper N, Smits P: Direct vasculareffects of furosemide in humans. Circulation 1997;96:1847–1852.

82 Pinsky MR: Clinical significance of pulmonary artery occlusion pressure. Intens Care Med 2003;29:175–178.

83 Pinsky MR: Functional hemodynamic monitoring. Intens Care Med 2002;28:386–388.84 Pinsky MR: Hemodynamic monitoring in the ICU. Contrib Nephrol. Basel, Karger, 2001, vol 132,

pp 92–113.85 Michard F, Teboul JL: Predicting fluid responsiveness in ICU patients: A critical analysis of the

evidence. Chest 2002;121:2000–2008.86 Marik PE: Pulmonary artery catheterization and esophageal Doppler monitoring in the ICU. Chest

1999;116:1085–1091.87 Marik PE: Assessment of intravascular volume: A comedy of errors. Crit Care Med 2001;29:

1635–1636.88 Birtch AG, Zakheim RM, Jones LG, Barger AC: Redistribution of renal blood flow produced by

furosemide and ethacrynic acid. Circ Res 1967;21:869–878.89 Ludens JH, Williamson HE: Effect of furosemide on renal blood flow in the conscious dog. Proc

Soc Exp Biol Med. 1970;133:513–515.90 Lameire N, Ringoir S: The influence of furosemide on the distribution of renal blood flow in the

normal and hypotensive dog. Kidney Int 1974;5:455.91 Castrop H, Schweda F, Schumacher K, Wolf K, Kurtz A: Role of renocortical cyclooxygenase-2

for renal vascular resistance and macula densa control of renin secretion. J Am Soc Nephrol 2001;12:867–874.

92 Yoshida M, Suzuki-Kusaba M, Satoh S: Participation of the prostaglandin system in furosemide-induced changes of renal function in anesthetized rats. Renal Physiol 1987;10:25–32.

Schetz 180

93 Christensen S, Petersen JS: Effects of furosemide on renal haemodynamics and proximal tubularsodium reabsorption in conscious rats. Br J Pharmacol 1988;95:353–360.

94 Janssen BJ, Eerdmans PH, Smits JF: Mechanisms of renal vasoconstriction following furosemidein conscious rats. Naunyn Schmiedebergs Arch Pharmacol 1994;349:528–537.

95 Tenstad O, Williamson HE: Effect of furosemide on local and zonal glomerular filtration rate inthe rat kidney. Acta Physiol Scand 1995;155:99–107.

96 de Wildt SN, Smith FG: Effects of the angiotensin converting enzyme (ACE) inhibitor, captopril,on the cardiovascular, endocrine, and renal responses to furosemide in conscious lambs. Can JPhysiol Pharmacol 1997;75:263–270.

97 Pires SL, Julien C, Chapuis B, Sassard J, Barres C: Spontaneous renal blood flow autoregulationcurves in conscious sinoaortic baroreceptor-denervated rats. Am J Physiol 2002;282:F51–F58.

98 Greven J, Klein H: Renal effects of furosemide in glycerol induced acute renal failure of the rat.Pflügers Arch 1976;365:81–87.

99 Burke TJ, Duchin KL: Glomerular filtration during furosemide diuresis in the dog. Kidney Int1979;16:672–680.

100 Murray PT, Le Gall JR, Mranda DDR, Pinsky MR, Tetta C: Physiologic endpoints (efficacy) foracute renal failure studies. Curr Opin Crit Care 2002;8:519–525.

101 MacDonald TM, Craig K, Watson ML: Frusemide, ACE inhibition, renal dopamine andprostaglandins: Acute interactions in normal man. Br J Clin Pharmacol 1989;28:683–694.

102 Epstein M, Schneider NS, Befeler B: Effect of intrarenal furosemide on renal function andintrarenal hemodynamics in acute renal failure. Am J Med 1975;58:510–516.

103 Bradley VE, Shier MR, Lucas CE, Rosenberg IK: Renal hemodynamic response to furosemide inseptic and injured patients. Surgery 1976;79:549–554.

104 Chen HH, Redfield MM, Nordstrom LJ, Cataliotti A, Burnett JC Jr: Angiotensin II AT1 receptorantagonism prevents detrimental renal actions of acute diuretic therapy in human heart failure. AmJ Physiol 2003;284:F1115–F1159.

105 Meyer P, Menard J, Papnicolaou N, Alexandre J, Devaux C, Milliez P: Mechanism of renin releasefollowing furosemide diuresis in rabbit. Am J Physiol 1968;215:908–915.

106 Weber PC, Scherer B, Larsson C: Increase of free arachidonic acid by furosemide in man as thecause of prostaglandin and renin release. Eur J Pharmacol 1977;41:329–333.

107 Lorenz JN, Weihprecht H, Schnermann J, Skott O, Briggs JP: Renin release from isolated juxta-glomerular apparatus depends on macula densa chloride transport. Am J Physiol 1991;260:486–493.

108 Kammerl MC, Nüsing RM, Richthammer W, Krämer BK, Kurtz A: Inhibition of COX-2 counter-acts the effects of diuretics in rats. Kid Intern 2001;60:1684–1691.

109 Mann B, Hartner A, Jensen BL, Kammerl M, Kramer BK, Kurtz A: Furosemide stimulates maculadensa cyclooxygenase-2 expression in rats. Kidney Int 2001;59:62–68.

110 Mason J, Kain H, Welsch J, Schnermann J: The early phase of experimental acute renal failure.VI. The influence of furosemide. Pflügers Arch 1981;392:125–133.

111 Engelman RM, Gouge TH, Smith SJ, Stahl WM, Gombos EA, Boyd AD: The effect of diureticson renal hemodynamics during cardiopulmonary bypass. J Surg Res 1974;16:268–276.

112 Dobrowolski L, B-dzynska B, Sadowski J: Differential effect of frusemide on renal medullary andcortical blood flow in the anaesthetised rat. Exp Physiol 2000;85:783–789.

113 Dobrowolski L, Badzynska B, Grzelec-Mojzesowicz M, Sadowski J: Renal vascular effects offrusemide in the rat: Influence of salt loading and the role of angiotensin II. Exp Physiol 2001;86:611–616.

114 Castrop H, Vitzthum H, Schumacher K, Schweda F, Kurtz A: Low tonicity mediates a downregu-lation of cyclooxygenase-1 expression by furosemide in the rat renal papilla. J Am Soc Nephrol2002;13:1136–1144.

115 Duchin KL, Peterson LN, Burke TJ: Effect of furosemide on renal autoregulation. Kid Int 1977;12:379–386.

116 Wittmann U, Nafz B, Ehmke H, Kirchheim HR, Persson PB: Frequency domain of renal autoreg-ulation in the conscious dog. Am J Physiol 1995;269:F317–F322.

117 Just A, Wittmann U, Ehmke H, Kirchheim HR: Autoregulation of renal blood flow in the con-scious dog and the contribution of the tubuloglomerular feedback. J Physiol 1998;506:275–290.


118 Nishiyama A, Majid DS, Walker M, Miyatake A, Navar LG: Renal interstitial ATP responses tochanges in arterial pressure during alterations in tubuloglomerular feedback activity. Hypertension2001;37:753–759.

119 Lameire N, Vanholder R: Pathophysiologic features and prevention of human and experimentalacute tubular necrosis. J Am Soc Nephrol 2001;12(suppl 17):S20–S32.

120 Rybak LP: Ototoxicity of loop diuretics. Otolaryngol Clins N Am 1993;26:829–844.121 Humes HD: Insights into ototoxicity: Analogies to nephrotoxicity. Ann NY Acad Sci 1999;884:15–18.122 Rudy DW, Voelker JR, Greene PK, Esparza FA, Brater DC: Loop diuretics for chronic renal insuf-

ficiency: A continuous infusion is more efficacious than bolus therapy. Ann Intern Med 1991;115:360–366.

123 Martin SJ, Danziger LH: Continuous infusion of loop diuretics in the critically ill: A review of theliterature. Crit Care Med 1994;22:1323–1329.

124 Yelton SL, Gaylor MA, Murray KM: The role of continuous infusion loop diuretics. AnnPharmacother 1995;29:1010–1014.

125 Dormans TP, van Meyel JJ, Gerlag PG, Tan Y, Russel FG, Smits P: Diuretic efficacy of high dosefurosemide in severe heart failure: Bolus injection versus continuous infusion. J Am Coll Cardiol1996;28:376–382.

126 Klinge JM, Scharf J, Hofbeck M, Gertlig S, Bonakdar S, Singer H: Intermittent administration offurosemide versus continuous infusion in the postoperative management of children followingopen heart surgery. Intensive Care Med 1997;23:693–697.

127 Luciani GB, Nichani S, Chang AC, Wells WJ, Newth CJ, Starnes VA: Continuous versus inter-mittent furosemide infusion in critically ill infants after open heart operations. Ann Thorac Surg1997;64:1133–1139.

128 Pivac N, Rumboldt Z, Sardelic S, Bagatin J, Polic S, Ljutic D, Naranca M, Capkun V: Diuretic effectsof furosemide infusion versus bolus injection in congestive heart failure. Int J Clin Pharmacol Res1998;18:121–128.

129 Star RA: Treatment of acute renal failure. Kidney Int 1998;54:1817–1831.

Miet Schetz, MD, PhDDepartment of Intensive Care MedicineUniversity Hospital Gasthuisberg, Herestraat 49, BE–3000 Leuven (Belgium)Tel. �32 16344021, Fax �32 16344015, E-Mail [email protected]


How to Manage Vasopressors in AcuteRenal Failure and Septic Shock

Maurizio Dana, Sandra Rossia, Luca Callegarina, Claudio Roncob

Departments of aAnesthesia and Intensive Care, and bNephrology, Ospedale di Vicenza, Vicenza, Italy

Acute renal failure (ARF) occurs frequently in patients with septic shockand carries a marked increase in mortality [1].

Significant oliguria should prompt immediate attention to the adequacy ofcirculating blood volume and of mean arterial pressure (MAP) before the risingof serum creatinine concentration and the signs of established ARF [2, 3].

Shock is defined as a condition of inadequate supply or utilization of meta-bolic substrate and oxygen at tissues resulting in functional compromise (lacticacidosis) and eventually cellular damage (necrosis or apoptosis).

Septic shock is characterized by peculiar features: hyperdynamic circula-tory state with high cardiac output (CO), loss of vasomotor tone and low totalperipheral resistance (TPR).

The hyperdynamic stage develops after volume expansion, which is the firstline therapy for arterial hypotension associated with shock. Without volumeexpansion septic shock would be characterized by low CO, elevated systemicvascular resistance (SVR), and decreased venous return (VR) due to a combina-tion of increased venous capacitance, splanchnic blood pooling, and effectivevolume depletion (i.e. increased vascular permeability, fluid sequestration) [4].To maintain or to increase CO the preload has to be restored and increased abovenormal values because of a concomitant septic myocardial depression [5, 6].

Despite the high global flow and globally decreased systemic vascularresistance occurring after volume resuscitation, there is a persistent cellulardysoxia. Therefore, it has been concluded that this flow is non-efficient for themetabolizing tissues of the periphery (distributive shock).

Indeed focal vasoconstriction occurs in some peripheral vessels, particu-larly in the mesenteric, pulmonary and renal circulation and part of the blood

Vasopressors and Septic Shock 183

flow is shunted to the venous compartment past dysfunctioning and collapsedmicrocirculatory functional units [7, 8].

To confirm that the relation between global flow and effective regional flowto organ and tissues is not linear, it has been shown that strategies of increasingfurther CO and the delivery of O2 to the periphery do not improve survival [9].

Blood flow to the periphery is regulated by a complex interplay betweenautoregulatory control (proportional to local metabolic demand) and global auto-nomic control, which alter peripheral vasomotor tone to optimize inter-organsblood flow distribution. It is well known that arteriolar vessels of different organshave different sensitivity to sympathetic nerve stimulation and to catecholamines[10, 11]. Autonomic control is intended to preserve flow to tissues most suscep-tible to ischemia by sacrificing, if needed, tissues with better oxygen extractioncapacities. Therefore, there is hierarchy into preserving flow: the brain and theheart are at the top and the kidneys right after.

The general rule is that to maintain organ perfusion blood arterial pressureis important: below a ‘critical’ pressure (related to age, diseases and vasculatureconditions) the regional autoregulation of flow is not preserved.

Another important issue is that microcirculation is a heterogeneous andextremely complex network of vessels with different architecture, differentautonomic/metabolic regulation of flow and different rheological and resistiveproperties related to the different structure and function of organs and tissues.This implies a heterogeneous responsiveness to the circulatory derangement ofthe shock state. Furthermore, blood flow to an organ is not equivalent to theeffective capillary flow in all regions of that organ.

Blood flow in the kidney is autoregulated over a wide range of arterialpressure but, out of this range, a steep relation exists between pressure and flow.This means that, out of the autoregulative range, flow drops in the kidney morethan in other organs.

Given its peculiar vascular architecture, the outer medullary region of thekidney is at higher risk of ischemia than the other parenchyma [12].

Since arterial pressure is the primary factor determining renal blood flowindependently from CO, acute hypotension has to be treated without delay.

The role of vasopressors to restore adequate mean arterial pressure (MAP)to achieve optimal renal resuscitation and the physiological sound underlyingtheir use will be reviewed in this chapter.

Sepsis-Induced Circulatory Dysfunction

Sepsis induces a self-amplifying cascade of inflammatory cytokines andmediators with vasoactive properties such as nitric oxyde (NO), prostaglandins,

Dan/Rossi/Callegarin/Ronco 184

endothelins, leukotrienes, platelet-activating factor (PAF) [13]. Inducible nitricoxide [iNO] appears to have a central role in the determination of septic vascu-lar dysfunction and the inhibition of its synthesis improves blood pressure andincreases the blunted responsiveness to vasopressors [14–16]. This mediator ispivotal to the disregulation of vasomotor tone resulting in distributive shock.After this considerations NO would appear to be the target of a modern resusci-tation therapy for septic shock. Unfortunately inhibition of NO sinthasi [NOS]failed to demonstrate normalization of global oxygen extraction and in somestudies revealed numerous adverse effects. NO it is not only a mediator inducedby inflammation [iNO] but, first of all is a physiologic mediator with two con-stitutive isoforms [cNO]: the endothelial [eNO], regulating tissue perfusion,and neuronal [nNO], regulating nonadrenergic, non-cholinergic peripheralneurotransmission.

It is worth to appreciate that the consequences of the distributive shock inthe microcirculation are, as the word says, heterogeneously distributed owing tothe following phenomena: adherence of activated polymorphonuclear leuko-cytes to vascular endothelium, O2 free radical-mediated injury to endothelialcells, thrombotic random obstruction of vessels and increased vascular perme-ability. Vascular critical closing pressure will be modified in the various organsand also in regions of the same organ.

Given the wealth of data regarding the occurrence of endothelial dysfunc-tion in sepsis that induces the concomitant presence of randomly distributedvasoconstriction together with general vasodilatation, the protective role ofeNO needs to be taken into great consideration. NO synthesis prevents severehypoperfusion of pancreas, small and large bowel, while its inhibition mightlead to the appearance of patchy dysoxya [17].

An Hydraulic Point of View

The function of the left ventricle is to pump blood to the tissues so that theymay be perfused with oxygen and nutrients according to their metabolic needs.Whereas blood flow through the tissues is continuous, or nearly so, the heart’soutput is intermittent. Pulsation of pressure and flow are wasteful work and arethe consequences of the design limitation of the heart: the left ventricle needs aphase of pause from contraction (diastole) in order to provide its own perfusionthrough the coronary circulation. The arterial system, that links the heart to thecapillary circulation, subserves two functions: the first, as a low resistances cir-cuit, and the second as a cushion to flow pulsation at its input (Windkesselmodel). Ventriculo-arterial interaction is optimized when the percentage ofenergy dissipated in pulsatile compared to total external work is 10–20%.


If the pressure and flow in the arterial system are interdependent, the arterialpressure wave must include information regarding flow. From a mechanical pointof view, one can usefully think in terms of its mean value and its fluctuationaround this mean, rather than in terms of systolic and diastolic pressure. There isa mean pressure during the systolic phase and a mean pressure during diastolicphase and a mean arterial pressure (MAP) summarizing pressure fluctuations dur-ing the entire cardiac cycle. When the arterial system works efficiently, like in thenormal young human there is small fluctuation around MAP (fig. 1). Surprisingly,MAP decreases very little (1–3 mm Hg) along the entire length of the arterial sys-tem from aorta to the most peripheral arterioles, but, at arterial termination, it fallsabruptly over a very short distance. This phenomenon is due to the fact that theentire surface area of the post-arteriolar network is much increased compared tothat of proximal arteries. The mean pressure at the input of the post-arteriolar net-work is the pressure necessary to drive blood flow through collapsible vessels.

To reverse hypotension we have to restore a global vasomotor tone withoutinducing harmful vasoconstriction. Accordingly, how and to which extent weincrease mean arterial pressure is the real key point rather than what vasopres-sor we use for this purpose. Our treatment should be targeted to reproduce opti-mal ventriculo-arterial coupling avoiding large systolic and diastolic fluctuationaround the mean because only in this way we can rely on a fair relation betweenpressure and flow.

A further important consideration in using vasopressors is that pressure andflow waveform are not superimposable. An important contribution to determine

MSP

MDP

PPMSP

MP

MP

MDP

a b

Fig. 1. Diagrammatic representation of the arterial pressure wave contour in abdomi-nal aorta: fluctuations around a mean during systole and diastole in two different conditions:(a) normal fluctuation or optimal cushioning (Windkessel), and (b) increased or non optimalsystolic cushioning. PP � Pulse pressure; MSP � mean systolic pressure; MDP �mean diastolic pressure; MP � mean pressure.


the pressure waveform and the pressure/flow coupling, is the timing and ampli-tude of wave reflection. This phenomenon must have an important physiologicrole in the mammalian arterial system in transforming the intermittent flow ofthe heart. Wave reflection exists because the arterial system is not composed bylong and uniform tubes with a remote end, but rather by vessels with differentphysical properties and branching at different angles. Architecture (length,branching, diameter), structure (wall, diameter), function (elastic properties,vascular tone) of the arterial system are optimized in normal young adult toincrease diastolic pressure. An oscillatory phenomenon such as pressure andflow wave will be repetitively reflected to same extent wherever there is a dis-continuity or an impedance mismatch. Looking from the proximal arterial sys-tem, from which departs all the vessels for major vital organs, there are manyreflecting sites: branching points, areas of alteration of arterial distensibility,and high resistance arterioles. In the mathematical model, the extreme cases arewhen a tube with pulsatile liquid flow is completely blocked or opens into alarge reservoir. In the first condition, we have total positive wave reflection(closed system) in the later we have total negative reflection (open system). Inthe real arterial system, we are obviously dealing with intermediates conditions.In aged patients with vascular diseases, increased wave reflection occurs insystole rather than in diastole, causing systolic hypertension and decreasedventricular ejection that results in a disproportioned pressure versus volumework (worse ventriculo-arterial coupling). This pathophysiologic process iswell described [18].

What happens to the arterial system in the condition of profoundly reducedvasomotor tone (open system condition) is not described so far. Since we knowthat vasodilators used to treat hypertension modify wave reflection, we mayinfer that vasoactive mediators causing septic shock will do the same.

The product of pressure by flow has dimension of power. While the prod-uct of instantaneous pressure by flow is the total hydraulic power transferredfrom the left ventricle to the systemic circulation, the product of mean pressureby mean flow is the part of this power converted into the steady flow. Steadycontinuous flow is the one really effective for organ perfusion [19]. The differ-ence between total and steady power represents the pulsatile power, which is thewasteful part of power. As we know that MAP equals CO time TPR and alsothat CO is normal or increased during sepsis; the only way to restore adequateMAP is to manipulate TPR.

The problem is that, looking at the pulsatile circulation, we want to mod-ify the steady circulation. We assume that the MAP we calculate from the wave-form analysis by integrating the area under the pulsation is the pressure drivingeffective flow to the periphery. We should at least take into account that thatMAP is the mean between a mean pressure during systole and a mean pressure


during diastole and that pressure fluctuations around MAP represents howmuch hydraulic power is wasted rather than converted in efficient flow to theperiphery.

Cholley et al. [20] suggest that, in the hemodynamic derangement of septicshock, an alteration of the mechanical properties of the large arteries induces amarked increase in percent pulsatile power. Pressure should be monitored moreproximally than in the radial artery (this vessel has different reflection phenom-enon than the aortic tract from which depart the splanchnic vasculature) andvasopressor therapy should be titrated to obtain the appropriate mean pressurewith the smallest systolic and diastolic fluctuation. To match this result adequatepreload (stroke volume) is necessary before increasing vasomotor tone withcatecholamines.

Vasopressor Therapy for Septic Shock

Optimal hemodynamic treatment in septic shock means to support the car-diovascular function for providing adequate flow to metabolizing tissues. Earlyreversal of hypotension with the aid of vasopressors is a crucial point for main-taining flow particularly for the most vulnerable organs, brain, heart and kidneys.

Vasopressin is an endogenous polypeptide hormone which produces vaso-constriction at a precapillary level. It has been recently proposed for the treat-ment of unresponsive shock, but, since its harmful effect in the splanchnicvasculature [12, 13], it is not to be considered standard therapy until prospec-tive randomized study will demonstrate a survival advantage. Such a study iscurrently underway [21].

Although the complex pathophisiology of distributive shock, catecholaminesare so far the mainstay treatment.

Natural and synthetic catecholamines increase inotropism by beta receptorsstimulation and modify vasomotor tone by �- and �-receptors; since the mecha-nism of adrenergic stimulation is the same, the difference between them residesexclusively in the degree to witch they stimulate the two kind of receptor.

Norepinephrine is an endogenous cathecolamine with powerful vasocon-strictor effect secondary to �1-receptor agonism with absence of �2-receptorstimulation (�-vasodilatation). It also has a �1-positive inotropic effect. Norepi-nephrine seems to have the good characteristics to treat the distributive shock. Itsaction, although qualitatively not different than that of epinephrine on the vascu-lature (�1-agonism), is more selective because it is devoid of �2 stimulation.

Epinephrine is an endogenous cathecolamine with powerful inotropic,chronotropic, vasoconstrictive effect because unselectively stimulates all adren-ergic receptors.


Dopamine is an endogenous indirect acting cathecolamine (norepinephrineprecursor) which activates also a dopaminergic receptor in the renal and splanch-nic vasculature producing selective vasodilatation of these arterial beds. Thiseffect is dominant only at low doses (up to 3 �g/kg/min), at increasing doses�1 �2 (4–10 �g/kg/min) and �1 (�10 �g/kg/min) effects predominate.

Dobutamine is a synthetic cathecolamine with �1 �2 agonism, then exert-ing only inotropic and vasodilator effect.

Phenylephrine is a synthetic cathecolamine with selective �1 effect andvirtually absent �-receptor stimulation.

Strategies for Treatment of Septic Shock

Volume expansion must be the first line treatment. Only when and ade-quate preload has been assured, the use of catecholamines should be consideredto restore an appropriate arterial pressure [23].

Although randomized trials are missing, there is general agreement on theuse of adrenergic agents. The same agreement is lacking on the choice of onecatecholamine versus another.

The reasons for this are quite obvious: these agents do not modify directlythe pathophysiologic mechanism responsible for the loss of vasomotor tone(fig. 2). We relay on this indirect and simplistic strategy because other therapiesdirected to NO synthesis regulation, the core of the pathophysiologic process,have not shown to be effective. Catecholamines, although with different modu-lation, act on the same class of �- and/or �-receptors. Inappropriate stimulationof these receptors have potential detrimental effects. Major concern has always

Elastic arteries ⇒ muscular arteries ⇒ arterioles

Capillary network

CatecholaminesSeptic vasodilatation

No

Fig. 2. Diagrammatic representation of the sites of action of catecholamines and of thesite of septic vasodilatation: the heart on the left, the systemic circulatory system from majorelastic arteries to muscular arteries, arterioles and finally capillary network from the left tothe right.


been addressed to the fact that intense � stimulation, such that of norepineph-rine, might induce visceral hypoperfusion/ischemia, leading to MODS, loss ofgut mucosal integrity and finally worsening of microcirculatory dysfunction.However, recent animal studies showed an increase in renal blood flow whennorepinephrine was used to restore blood pressure in a model of canine septicshock [24–26]. Moreover, retrospective studies showed that patients with septicshock had a better response if treated with norepinephrine rather than withother agents [27].

In this complex scenario of loss of vasomotor tone, possible down-regulationof �-adrenergic responsiveness and contemporary potential visceral ischemiainduced by vasopressors, the key point is not which catecholamine to use buthow to titrate its dose and modulate its receptors action.

References

1 Brivet FG, Kleinknecht DJ, Loirat P: Acute renal failure in intensive care units: Causes, outcome,and prognostic factors of hospital mortality: A prospective multicenter study. Crit Care Med1996;24:192–198.

2 Brezis M, Rosen S: Hypoxia of the renal medulla: Its implications for disease. N Engl J Med1995;332:647–655.

3 Myers BD, Moran SM: Hemodynamically mediated acute renal failure. N Engl J Med 1986;314:97–105.

4 Parrillo JE: Septic shock in humans: Advances in understanding of pathogenesis, cardiovasculardysfunction, and therapy. Ann Intern Med 1990;113:227–242.

5 Parker MM, Suffredini AF, Natanson C, et al: Response of left ventricular function in survivorsand non survivors of septic shock. J Crit Care 1989;4:19–25.

6 Parker MM, McCarthy KE, Ognibene FP, et al: Right ventricular dysfunction and dilatation, sim-ilar to left ventricular changes, characterizes the cardiac depression of septic shock in humans.Chest 1990;97:126–132.

7 Ince C, Sinaasappel M: Microcirculatory oxygenation and shunting in sepsis and shock. Crit CareMed 1999;27:1369–1377.

8 Ince C: Microcirculatory weak units: An alternative explanation. Crit Care Med 2000;28:3128–3129.

9 Gattinoni L, Brazzi L, Pelosi P, et al: A trial of goal-oriented hemodynamic therapy in critically illpatients. N Engl J Med 1995;333:1025–1032.

10 Daemen MJ, De May JG: Regional heterogeneity of arterial structure changes. Hypertension1995;25:464–473.

11 O’Rourke MF, Safar ME, Dzau VJ: Arterial Vasodilatation: Mechanism and Therapy. London,Edward Arnold, 1993.

12 Heyman SN, Fuchs S, Brezis M: The role of medullary ischemia in acute renal failure. NewHorizons 1995;3:597–607.

13 Bone RC, Grodzin CJ, Balk RA: Sepsis: A new hypothesis for pathogenesis of the disease process.Chest 1997;112:235–243.

14 Landry DW, Lewin HR, Gallant EM, et al: Vasopressin pressor hypersensitivity in vasodilatoryseptic shock. Crit Care Med 1997;25:1279–1282.

15 Landry DW, Lewin HR, Gallant EM, et al: Vasopressin deficiency contributes to the vasodilationof septic shock. Circulation 1997;95:1122–1125.

16 Murray PT, Wylam ME, Umans JG: Nitric oxide and septic vascular dysfunction. Anesth Analg2000;90:89–101.


17 Mulder MF, Van Lambalgen AA, Huisman E, et al: Protective role of NO in the regional hemo-dynamic change during acute endotoxemia in rats. Am J Physiol 1994;266:1558–1564.

18 Nichols WW, O’Rourke MF: McDonald’s Blood Flow in Arteries, ed 4. London, Arnold, 1998,pp 284–293.

19 Nichols WW, O’Rourke MF: McDonald’s Blood Flow in Arteries, ed 4. London, Arnold, 1998,pp 294–295.

20 Cholley B, Lang R, Berger D, et al: Alterations in systemic arterial mechanical properties duringseptic shock: Role of fluid resuscitation. Am J Physiology 1995;269:H375–H384.

21 Russell JA, Cooper DJ, Walley KR, et al: Vasopressin and septic shock trial (VASST): baseline char-acteristics and organ dysfunction in vasopressor dependent patients with septic shock. Am J RespCrit Care Med 2003;167:A548.

22 Bellomo R, Ronco C: The Use of Inotropic and Vasopressor Agents in Patients at Risk ofRenal Dysfunction. Critical Care Nephrology. Amsterdam, Kluwer Academic Publishers, 1998,pp 1133–1137.

23 Bellomo R, Kellum JA, Wisniewski SR, et al: Effects of norepinephrine on the renal vasculaturein normal and endotoxemic dogs. Am J Resp Crit Care Med 1999;28:2758–2765.

24 Di Giantomasso D, Morimatsu H, May CN, Bellomo R: Intrarenal blood flow distribution inhyperdynamic septic shock: Effect of norepinephrine. Crit Care Med 2003;31:2509–2513.

25 Di Giantomasso D, May CN, Bellomo R: Vital organ blood flow durino hyperdynamic sepsis.Chest 2003;124:1053–1059.

26 Martin C, Viviand X, Leone M, Thirion X: Effect of norepinephine on the outcome of septicshock. Crit Care Med 2000;28:2758–2765.

27 Martin C, Papazian L, Perrin G, et al: Norepinephrine or dopamine for the treatment of hyperdy-namic septic shock? Chest 1993;103:1826–1831.

Dr. Maurizio DanDepartment of Anesthesia and Intensive Care MedicineOspedale di Vicenza, Viale Rodolfi 37, IT–36100 Vicenza (Italy)Tel. �39 0444993895, E-Mail [email protected]


Management of Vascular Catheters forAcute Renal Replacement Therapy

V. D’Intini, M. Bonello, G. Salvatori, C. Ronco

Department of Nephrology, Dialysis and Transplantation San Bortolo Hospital,Vicenza, Italy

Central venous catheters (CVC or DLHDC – double lumen hemodialysiscatheters) have facilitated the implementation of extracorporeal renal replace-ment therapy (RRT). The use of CVCs has grown exponentially and now is afundamental and basic prerequisite in the management of a broad range of bothrenal and non-renal clinical scenarios. Such catheters have revolutionized clin-ical practice permitting the application of new therapeutic technologies aidingmanagement of the patient with acute and chronic problems.

For acute clinical situations, a CVC is fundamental for immediate vascu-lar access in order to implement RRT. The clinician has at his disposal a widerange of catheters with different characteristics to optimize therapy.

Such is their widespread utilization that up to 60% of new patients com-mencing chronic hemodialysis dialyze through a CVC initially and an overallprevalence of 30% of chronic patients are using a catheter for hemodialysisaccess [1].

The ideal temporary catheter must be easy and quick to insert, be simpleand safe to use for nursing staff, avoid major risks both at the time of insertionand during continued use, avoid long-term damage to central vessels, be madeof biocompatible material to prevent thrombosis, and be capable of providingadequate blood flows (100–400 ml/min) with low re-circulation at low hydro-static pressures.

Despite advances in design and material, catheters are associated with sig-nificant acute and long term morbidity. Life-threatening complications duringcatheter insertion such as hemorrhage and pneumothorax, thrombosis, localand systemic infection represent a potential risk for the patient and should begiven special attention.

Practical Aspects of CRRT

D’Intini/Bonello/Salvatori/Ronco 192

Most catheters used today are inserted by the Seldinger technique intojugular, subclavian and femoral veins. Ultrasound guided localization of thevein is recommended. Standard non-cuffed DLHDCs are generally indicatedfor temporary access. If a protracted treatment is foreseen, tunnelled cuffedDLHDCs should be utilized for their low infection risk profile.

Characteristics of Catheters

The ultimate DLHDC must be biocompatible, easy to insert and use, befunctional and durable, carry a low risk of infection and thrombosis and mustbe inexpensive [2]. Structurally it should be designed to optimize size, flexi-bility, strength and compliance. A compromise of maximal luminal diameter formaximal blood flow, with compact dimensions to minimize complicationsis needed as well as a balance of flexibility and rigidity to maintain lumenpatency. DLHDCs can be made from polyethylene, Teflon, silicone,polyurethane and polyurethane/polycarbonate copolymers. Each material hasindividual characteristics with advantages and disadvantages. Polyethyleneis more rigid compared to silicone. Silicone is softer and thus a more flexiblematerial, but more difficult to insert through ligamentous or fibrotic tissues andmore prone to early mechanical failure because of lumen compression.However, a major advantage of flexible catheters is that the tip of the cathetermay be left in the right atrium without danger of cardiac perforation.Polyurethane is intrinsically strong thus constructed with thin walls, preservingrigidity in the longitudinal axis while avoiding lumen collapse at high nega-tive pressures. It also has thermoplastic properties, becoming softer at bodytemperature.

Finally different materials are subject to chemical disruption by variousproducts which can affect material performance. Alcohols including isopropylalcohol and ointments containing polyethylene glycol (mupirocin or betadineointment) can weaken polyurethane catheters considerably. Catheters createdfrom copolymer materials are resistant to such chemicals and may be the prin-cipal material used for future DLHDCs.

Insertion Techniques

Insertion techniques have clearly improved contributing to reduced inser-tion-related complications and increasing catheter function duration. Insertionsite (jugular, subclavian and femoral veins) is dictated by the clinical situation.Subclavian cannulation is associated with a high incidence of stenosis and is

Vascular Catheters for ARF 193

contraindicated if the risk of chronic dialysis is foreseen [3]. Real-time ultrasoundguided cannulation is associated with less complications and has been stronglyrecommended by the DOQI committee on vascular access [4]. Such technologyreduces the number of needle passes, failed placements, and insertion-relatedcomplications. Inexperienced operators increase their success rate to 95% withthe use of ultrasound guidance [5]. Ultrasound imaging is a valuable tool in car-ing for dialysis patients because about 30% have significant vein abnormalities,such as total occlusion, non-occlusive thrombus, stenosis and anatomic varia-tion [6]. Post-insertion chest X-ray after internal jugular or subclavian insertionconfirms the position of the catheter tip in the superior vena cava and allowsevaluation for possible pneumothorax and hemothorax.

The length of time a catheter should be left in situ remains controversial.The rate of infection for internal jugular non-cuffed catheters suggests that theyshould be used for no more than 3 weeks [4]. Femoral catheters left in place for7–14 days did not develop any complication whereas the number of complica-tions rose significantly in patients whom femoral catheters was left in place forover 21 days [7].

Instruction on the technique of insertion of DLHDCs is beyond the scopeof this review and can be found in various texts [8].

Catheter Positioning

Based on opinion and DOQI guidelines, fluoroscopy or chest X-raycontrol of correct catheter position and tip location should be done prior to use.Improper tip placement is a common cause of poor flow. The tip of a semi-rigidcatheter should extend to the superior vena cava, 1–2 cm above the right atrium.Shorter catheters may be plagued by excessive re-circulation and longercatheters risk atrial perforation. The optimal length for a right internal jugularcatheter is approximately 15 cm. Femoral catheters should be longer in orderto reach the inferior vena cava to minimize re-circulation. The optimal orienta-tion of the catheter tip is important for proper function and good flow.Silicone catheters allow higher blood flows when the distal tip is located in theright atrium. Fluoroscopy is required to optimize cuffed tunnelled catheterplacement.

Catheter-Related Morbidity

The risk of insertion-related complications varies according to the skill ofthe operator, site of insertion, and use of imaging equipment such as ultrasound


and fluoroscopy. Complications can be immediate or delayed. Early dysfunc-tion is usually related to mechanical problems (e.g. inappropriate positioning,or kinking), while late dysfunction (�2 weeks) is often caused by thromboticproblems such as partial or total obstructive thrombosis of the catheter lumen,thrombosis or stenosis of the cannulated vein, external sheath formation on thecatheter distal end and internal catheter clotting. In the latter case, the partial orcomplete occlusion of the lumen and distal and/or lateral perforations greatlyincrease the extracorporeal resistances and reduce the effective blood flowaccordingly.

Immediate Complications

Severe immediate complications are almost entirely associated with use ofthe anatomic landmark technique (blind insertion technique). The success ofthe landmark guided placement of the catheter assumes normal vascularanatomy, the vein being patent and of normal caliber. The average complica-tions reported in the literature are reported in table 1. These risks are greater incritical care patients because of coexistent coagulopathies, thrombocytopenia,liver disturbances and drug related bleeding problems. With insertion of afemoral catheter, bleeding is usually easily controlled unless severe deep dam-age to the arterial wall has occurred. Puncture of the femoral artery along itssuperficial path is rarely associated with more than a small hematoma, butheparin free dialysis is advised for a minimum of 24 h. Severe femoral arterydamage or rupture, particularly if deep to the inguinal ligament, may causeuncontrollable hemorrhage into the retroperitoneal space, requiring urgent sur-gical intervention. Both subclavian and internal jugular vein line insertion havea low risk of uncontrolled arterial bleeding in case of accidental arterial punc-ture. Visualization of the anatomy, with a hand held ultrasound device, prior tothe procedure has further decreased the incidence of arterial puncture, although

Vascular injuriesHemothorax

HydrothoraxCarotid artery punctureSubclavian artery puncture

PneumothoraxAir embolismWire or catheter embolismArrythmias

Table 1. Central line insertion complications


a small risk is still recognized. Hemostasis can usually be obtained by pressureover the puncture site when the carotid artery has been punctured by the finderor introducer needle. If the large bore dialysis catheter has been introduced intothe arterial system, the catheter should not be removed until the surgical teamis present, as bleeding is less easily controlled once the catheter has beenremoved. Subclavian artery damage is more difficult to control and is com-monly associated with a hemothorax. Urgent exploration by a vascular surgeonwith repair of damaged vessel is required.

Local bruising is rarely a problem unless the bleeding leads to the forma-tion of a hematoma. This can become secondarily infected or, if the bleeding isarterial, form a false aneurysm. If during subclavian and internal jugular veincatheter insertion the patient complains of shortness of breath the procedureshould be stopped and a chest X-ray obtained. Rarer complications include airembolism, damage to the superior vena cava, right common carotid arteryfistula, right atrial thrombus, lymphorrhea, pericardial tamponade and arrhyth-mias occurring as a result of attempted catheterization (the latter is usually pre-cipitated by the guidewire irritating the sino-atrial node or conducting tissues).The use of ultrasound-guided cannulation has resulted in a substantial decreasein procedural complications. Randolph et al., performed a meta-analysis of theliterature and concluded that ultrasound guidance decreased central venouscatheter placement failure (relative risk (RR) 0.32), decreased complicationsfrom catheter placement (RR 0.22), and decreased the necessity of multiplecatheter placement attempts (RR 0.60) when compared to landmark cannula-tion techniques [5].

There exist meta-analyses comparing complications with the internal jugu-lar or the subclavian central venous approach [9]. Six trials (2,010 catheters)found significantly more arterial punctures with jugular catheters comparedwith subclavian (3.0 vs. 0.5%, RR 4.70). In six trials (1,299 catheters), therewere significantly fewer malpositions with the jugular access (5.3 vs. 9.3%, RR0.66). In ten trials (3,420 catheters), the incidence of hemato- or pneumothoraxwas 1.3 vs. 1.5%. The authors concluded that there are more arterial puncturesbut less catheter malpositions with the internal jugular compared with the sub-clavian access, and that there is no difference in the incidence of hemo- orpneumothorax.

Long-Term Complications

These are more often caused by thrombotic problems such as partial ortotal obstructive thrombosis of the catheter lumen, thrombosis or stenosis of thecannulated vein, and catheter-related infections.


Thrombotic Occurrences

The incidence of catheterized vein thrombosis varies from 20 to 70%according to the sites and diagnostic modalities used. Central vein stenosis andthrombosis are more common with a subclavian catheter. Recently, in prospec-tive studies, stenosis/thrombosis was found to complicate up to 28% of subcla-vian dialysis catheters, and infection further increased the risk [10]. For thisreason, subclavian catheters should be avoided if possible particularly if chronichemodialysis is foreseen. Acute catheters in the jugular vein caused thrombo-sis/ stenosis in only 2% of cases surveyed with ultrasound in one study [11].Femoral venous catheterization is associated with a greater risk of thromboticcomplication than subclavian catheterization in ICU patients [12].

Several factors contribute to the thrombogenicity on the catheter.Schematically, they are represented by the catheter type (material and compo-sition, flexibility, aspect and surface treatment), insertion mode, host vein(including diameter, local hemodynamics), duration of use, and clotting statusof the patient (hyperfibrinogenemia, inflammatory disease, thrombocytosis,previous thrombosis). The prothrombotic risk of a patient may be reduced byadministering anticoagulants (e.g. low-molecular-weight heparin) and/orantiplatelet drugs. Most cases of thrombosis can only be attributed to activationof the coagulation cascade by a relatively bioincompatible device.

Most catheter flow problems are related to intrinsic thrombosis.Intraluminal thrombus, catheter tip thrombus, and fibrin sheath thrombus arethe principal types of intrinsic thrombosis.

Intraluminal thrombus usually occurs when an inadequate volume ofheparin is instilled into the catheter after dialysis or when heparin escapes fromthe catheter lumen between dialysis sessions. Blood can enter into the catheterand form an intraluminal thrombus. When this occurs, the catheter may becomecompletely occluded. Urokinase instillation characteristically resolves thethrombosis [4]. Many catheters have side holes at the tip of the arterial limbwhich does not retain heparin. A tip thrombus may be occlusive or it may act asa valve. Preventive measures that are commonly used to avoid intraluminalthrombosis are largely ineffective because of the presence of the side holes.Urokinase instillation usually resolves this problem.

Fibrin sheath thrombus is the most common type of thrombus. The term‘fibrin sheath’ refers to a sleeve of fibrin that surrounds the catheter starting atthe point where it enters the vein. The fibrin sheath is only loosely attached tothe catheter. As the sheath extends downward, it eventually closes over the tipof the catheter. It is probable that all central venous catheters become encasedin a layer of fibrin within a few days of insertion. The incidence of catheter dys-function secondary to fibrin sheath has been reported to be 13 to 57% [13].


Chronic systemic anticoagulation (e.g. warfarin) may be beneficial in prevent-ing this [14].

Prevention of Catheter Thrombosis

Proper flushing and heparinization of the catheter with concentratedheparin following dialysis will serve to decrease but not eliminate the risk ofthrombosis. Each side of the hemodialysis catheter has a fill volume inscribedon the clear portion of the catheter just below the cap. This volume is differentfor both lumens. The venous volume is generally 0.1–0.2 cm3 greater than thearterial volume. However, a 0.1–0.2 cm3 overfill is recommended. Chronic anti-coagulation with either warfarin or low-molecular-weight heparin has gainedanecdotal support for preventing both lumen thrombus and fibrin sheath for-mation on hemodialysis catheters [14].

Treatment of Catheter Thrombosis

Thrombosed, non-cuffed catheters can be exchanged over a guide wire ortreated with urokinase. Physical methods have been tried to recannulate blockedcatheters, from forceful saline flush to brush or drill devices with variable success.Urokinase remains the most effective method successfully dissolving 70–90% ofthrombus. Tissue plasminogen activator (tPA) may be superior to urokinase. Ifcatheter exchange still demonstrates catheter dysfunction, a fibrin sheath may bepresent. This is diagnosed radiologically and can be stripped using a snarecatheter. The snare is introduced through the femoral vein and advanced up tothe level of the dialysis catheter. The report success of this procedure ranges from92 to 98%. Fibrin sheath stripping generally results in asymptomatic embolizationof the fibrin sheath. Advantages of this technique include a good success rate,safety and preservation of the catheter with a patency duration that is reasonable.

Catheter-Related Infections

Temporary catheters have higher rates of infection than tunnelled dialysiscatheters (bacteremia 6.2 vs. 1.8 per 1,000 catheter days and exit site infection3.6 vs. 1.4 per 1,000 catheter days in not cuffed or tunnelled catheter, respec-tively). In the intensive care unit, non-tunnelled catheters carry a bacteremiarisk estimated between 3 and 10% and increases with duration exponentially atboth the femoral and internal jugular site [15].


The femoral site has a greater risk of infection than the internal jugular site.The instantaneous risk (hazard) of bacteremia at the femoral site increased after1 week compared with 3 weeks at the internal jugular site. In addition, if an exitsite infection occurred and the catheter was left in place, the risk of bacteremiarose from 2% at 24 h from onset of exit site infection to 13% at 48 h. These find-ings support the NKF-K/DOQI guidelines to limit acute dialysis catheter use inthe femoral vein and internal jugular vein to 1 and 3 weeks, respectively, and toremove catheters immediately if an exit site infection occurs.

Factors increasing infection risk include skin and nasal colonization withstaphylococcus, catheter hub colonization, duration of catheterization, throm-bosis, frequency of catheter manipulation, diabetes mellitus, iron overload,immunodeficiency, use of a transparent dressing (promotes skin colonization)and the conditions of catheter placement. Povidone and mupirocin ointmentswith dry gauze dressings have been shown to significantly reduce the risk ofbacteremia from acute dialysis catheters in randomized controlled trials, how-ever increase the risk of fungal infections in immunocompromised patients.

Infections can develop at the exit site, within the tunnel with tunnelledcatheters and cause catheter related bacteremia (CRB). The therapeuticapproach to each of these is different. Exit site infection is defined as localizedinfection of the skin and soft tissue around the exit site. Erythema, purulentdischarge and local tenderness are typically present. Fever and other signs ofsystemic infection are absent. With a temporary catheter, removal of thecatheter is warranted with replacement at an alternate site. With tunnelled dial-ysis catheters, bacteremia has not been shown to be clearly associated with exitsite infection. Most exit-site infections are caused by gram-positive organismsincluding Staphylococcus aureus and Staphylococcus epidermidis, althoughother bacteria can be involved. Exit-site infection can usually be treated effec-tively with oral or intravenous antibiotics. In more severe cases or those that failto respond to antibiotics, revision of the catheter with creation of a new exit siteremote from the infected area may resolve the exit-site infection. If these meas-ures fail, the tunnelled catheter may ultimately need to be removed.

Tunnel infections are invasive soft tissue infections that extend along thesubcutaneous tunnel toward the vein. These typically involve the cuff and exitsite, although in some cases they may appear only in the tunnel proximal to thecuff, with no drainage or communication to the exit site. Tenderness, swelling,and erythema along the catheter tract are typical, with purulent drainage fromthe exit site. Fever and other signs of systemic infection are often present, andbacteremia may occur. The catheter should be removed and antibiotics givenparenterally.

Catheter related bacteremia (CRB) from acute catheters should be treatedwith immediate catheter removal and appropriate antibiotics. The decision to


remove a tunnelled catheter for suspected CRB must often be made clinicallyprior to blood culture results and individualized based on the severity of sepsis,co-morbid illnesses, status of permanent arteriovenous access, and alternativesfor venous access. S. aureus is the most common organism representing 33–80%of positive cultures. The prevalence of nasal carriage of S. aureus among patientsundergoing hemodialysis has ranged from 30 to 60%. Reduction of nasal car-riage of S. aureus has resulted in a decrease in yearly bloodstream infections dueto S. aureus. The prolonged empirical use of vancomycin should be avoidedbecause of the risk of inducing vancomycin resistant enterococcus. Initial ther-apy with vancomycin and an aminoglycoside antibiotic (or a �-lactam with lac-tamase inhibitor or a quinolone) until culture results are available is prudent.Rapid conversion to appropriate antibiotics based on culture sensitivities isneeded not only to prevent emergence of resistant organisms, but also to avoidototoxicity. Antibiotic therapy should be continued for a minimum of 3 weeks.Blood cultures should be repeated a week following therapy to ensure that theinfection has been eradicated. A major management dilemma is whether toattempt salvage of the tunnelled catheter if sepsis is controlled and alternativeaccess is difficult. Recent reports show only 32–37% are salvageable. If signif-icant symptoms persist beyond 48 hours despite effective antibiotics the tun-nelled catheter should be removed [16].

A biofilm surrounding the catheter may predispose the CVC to infectionand failure to eradicate it despite appropriate antibiotics. Instillation of concen-trated antibiotic solutions into the lumen of silicone vascular catheters (antibi-otic lock) in vitro eliminates the biofilm, suggesting a useful clinicalapplication. Studies in tunnelled DLHDCs have suggested that the biofilm canbe eradicated by an antibiotic lock solution instilled into the catheter lumen,permitting bacteriologic cure without replacing the catheter [17]. These studieshave prompted a recent consensus panel to recommend use of antibiotic locksfor clinical management of uncomplicated bacteremia related to tunnelledDLHDCs. However, there are no large, prospective studies assessing the effi-cacy of this approach in treating hemodialysis patients.

Catheter Performances and Recirculation

Blood flow through the catheter is directly proportional to the pressure gen-erated by the blood pump across the catheter and inversely proportional to resis-tance in the catheter itself. Modern blood pumps have a limited capacity togenerate pressure, and high pressures can be damaging to red cells, so a majorgoal of catheter design has been to reduce the resistance to flow as much as pos-sible. The resistance to flow is directly related to length and inversely proportional


to the fourth power of the catheter diameter, thus the manufacturer can modifylength and diameter to reduce the pressures.

All CVCs must be able to provide optimal dynamics at all blood flowrequirements. Various design characteristics are essential to reduce the internalresistance, limit the risk of dysfunction through parietal suction and prevent itsobstruction by the (internal or external) formation of a fibrin sleeve. The nega-tive pressure recorded on the arterial side reflects resistance to blood suction. Itdepends on the blood flow applied, on the degree of venous collapse, blood pres-sure and on the patient’s intra-vascular fluid status. Negative pressure should notexceed �300 mm Hg to prevent the risk of parietal vascular lesions and/or bloodhemolysis. The positive pressure recorded on the venous side reflects resistanceto venous return. It depends on the blood flow and the degree of obstruction ofthe venous catheter. With the catheters in use today, venous pressure approxi-mates one half that of the blood flow displayed (such that for a blood flow rateof 300 ml/min, venous pressure is close 150–175 mm Hg). Thus, any change inarterial and/or venous blood pressures regimen must be indicative of catheterdysfunction that should be followed by suitable corrective actions. Oliver et al.,in acute renal insufficiency patients demonstrated that early catheter mechani-cal dysfunction was more frequently associated with double-lumen catheters(non-tunnelled) in the subclavian vein than with those inserted in other sites[11]. Specific anatomic abnormalities (fibrous band, inter-osseous mobile zone)can create kinking or stricture along the subclavian pathway.

Catheter re-circulation can reduce dialysis efficacy. It depends on the siteof catheter insertion and blood flow prescribed. Re-circulation is much lessimportant in continuous RRT modalities since lower flows are required.Femoral catheters, particularly short ones, exhibit a high re-circulation rateaveraging 20% (5–38%). Internal jugular and subclavian catheters have a muchlower re-circulation rate averaging 10% (5–15%). Catheter re-circulation isincreased by reversing connecting lines (blood re-circulation rates reach20–30%), but whether re-circulation in catheters relates to poor cardiac output,valvular insufficiency, use of vasopressor, or other clinical states is not known.An extra 20–30 min of dialysis time was recommended to compensate for theadverse effect of reversing the catheters on solute kinetics. Cardiopulmonaryre-circulation is defined as flow of dialyzed blood back through the dialyzer viathe central blood circuit (heart and lungs) without equilibration with blood inthe rest of the body. Cardiopulmonary re-circulation accounts for much of therebound in solute concentration following dialysis: approximately 30% of thetotal rebound of urea and less for other solutes. It might be speculated that inpatients with isolated central venous catheters there is negligible cardiopul-monary re-circulation, so the efficiency of dialysis could be slightly higher thanthat achievable with arteriovenous fistulas. This concept however still remains


to be proven but it could partially offset the lower clearance achievable throughcatheters due to their lower blood flow rates.

Care and Maintenance of the Catheters

Dialysis catheters, like all intravenous lines, should be kept clean and cov-ered at all times. To avoid bacterial contamination, the line should not be usedfor administration of medications, parenteral nutrition or accessed for bloodsampling, except during an emergency. The use of dry gauze dressing withpovidone iodine and mupirocin ointment at the catheter exit site, rather than atransparent dressing, can reduce the incidence of exit site infections, especiallyin patients who have nasal carriage of S. aureus. Dressing should be changed atleast twice weekly, or earlier if any discharge or moisture is noted at the exitsite. In the patients with an allergy to povidone iodine, alternate agents such aspolyantimicrobial gel can be substituted. The glycol constituents of ointmentshould not be used on polyurethane catheters.

A surgical mask worn by the patient and nurse any time the catheter isaccessed reduces the spread of infectious droplets and reduces contaminationof the catheter site. When not in use the lumens should be filled with 1 mlheparin (5,000 units per ml) mixed with 2 ml normal saline and heparin shouldalways be removed and discarded prior to using the line. Antibiotic use and ananticoagulant lock of the catheter appear to be efficient ways to prevent endo-luminal bacterial contamination in high risk patients.

Conclusions

Double-lumen catheters are an essential and convenient way for imple-menting RRT. Semirigid polyurethane catheters are the first choice for short-term use. Hemocompatibile, flexible silicone catheters which are less damagingto blood vessels seem better suited for medium and long term use. Acutecatheters should be placed in the femoral or internal jugular veins if possible.The subclavian vein site should be avoided to reduce overall catheter-relatedcomplications, particularly subclavian stenosis. Ultrasound should be used tosurvey local anatomy prior to insertion and/or it should be used in real time toreduce complications and increase the success rate of insertion. Catheter man-agement with dry gauze dressing with or without antibiotic ointments and erad-ication of nasal carriage of S. aureus can prevent infectious complications.Dysfunction through thrombosis remains the Achilles heel but chronic antico-agulation may decrease the incidence.


References

1 Pisoni LR, Young EW, Dykstra DM, Greenwood RN, Hecking E, Gillespie B, Wolfe RA, GoodkinDA, Held PJ: Vascular access use in Europe and the United States: Results from DOPPS. KidneyInt 2002;61:305–316.

2 Ash SR: The evolution and function of central venous catheters for dialysis. Sem Dial 2001;14:416–424.

3 Beenen L, van Leusen R, Deenik B, Bosch FH: The incidence of subclavian vein stenosis usingsilicone catheters for hemodialysis. Artif Organs 1994;18:289–292.

4 NKF-K/DOQI Clinical Practice Guidelines for Vascular Access: UPDATE 2000. Am J Kidney Dis2001;37:S137–S181.

5 Randolph AG, Cook DJ, Gonzales CA, Pribble CG: Ultrasound guidance for placement of centralvenous catheters: A meta-analysis of the literature. Crit Care Med 1996;24:2053–2058.

6 Forauer AR, Gloockner JF: Importance of US findings in access planning during jugular veinhemodialysis catheter placements. J Vasc Interv Radiol 2000;11:233–238.

7 Weyde W, Wikiera I, Ginger M: Prolonged cannulation of the femoral vein is a safe method oftemporary vascular access for hemodialysis. Nephron 1998;80:86.

8 McGee DC, Gould MK: Preventing complications of central venous catheterization. N Engl J Med2003;348:1123–1133.

9 Ruesch S, Walzer B, Tramèr MR: Complications of central venous catheters: Internal jugular ver-sus subclavian access. A systematic review. Crit Care Med 2002;30:454–460.

10 Hernandez D, Diaz F, Rufino M, Lorenzo V, Perez T, Rodriguez A, De Bonis E, Losado M,Gonzales-Posada JM, Torres A: Subclavian vascular access stenosis in dialysis patients: Naturalhistory and risk factors. J Am Soc Nephrol 1998;9:1507–1510.

11 Oliver MJ: Acute dialysis catheters. Semin Dial 2001;14:432–435.12 Merrer J, De Joughe B, Golliot F, Lefrant JY, Raffy B, Barre E, Rigaud JP, Casciani D, Misset B,

Bosquet C, Outin H, Brun-Buisson C, Nitenberg G: Complications of femoral and subclavianvenous catheterisation in critically ill patients. JAMA 2001;286:700–707.

13 Schwab SJ, Beathard G: The hemodialysis catheter conundrum: Hate living with them, but can’tlive without them. Kidney Int 1999;56:1–17.

14 Obialo CI, Conner AC, Lebon LF: Maintaining patency of tunnelled hemodialysis catheters.Scand J Urol Nephrol 2003;37:172–176.

15 Oliver MJ, Callery SM, Thorpe KE, Schwab SJ, Churchill DN: Risk of bacteremia, from tem-porary hemodialysis catheters by site of insertion and duration of use: A prospective study.Kidney Int 2000;58:2543–2545.

16 Mermel LA, Farr BM, Sherertz RJ, Raad II, O’Grady N, Harria JS, Crafen DE: Guidelines for themanagement of intravascular catheter-related infections. Clin Infect Dis 2001;32:1249.

17 Krishnasami Z, Carlton D, Bimbo L, Taylor ME, Balkovetz DF, Barker J, Allon M: Managementof hemodialysis catheter-related bacteremia with an adjunctive antibiotic lock solution. Kidney Int2002;61:1136–1142.

Vincenzo D’IntiniFRACP, San Bortolo Hospital, Viale Rodolfi, Vicenza (Italy)Tel. �39 0444993869, Fax �39 0444993949, E-Mail [email protected]


Relationship between Blood Flow,Access Catheter and Circuit Failureduring CRRT:A Practical Review

Ian Baldwin, Rinaldo Bellomo

Department of Intensive Care and Department of Medicine, Austin Hospital,Heidelberg, Vic., Australia

Continuous therapy for acute renal failure in the ICU setting places newdemands on the extracorporeal circuit (EC) originally designed for shortertime frame intermittent therapy. In addition, both nurses and physicians oftenneed to compromise on measures taken to maintain and promote the optimalfunction of the EC with the needs of the critically ill patient. For example,physicians may not be able to prescribe anticoagulant drugs at suggested dosesin a post operative or trauma patient. Nurses may not be free to change apatient’s position in bed for comfort, pressure care, and physiotherapy of thechest, due to access catheter placement and the need to ensure patency. Boththese issues can have a significant affect on optimal functional life of the EC,preventing it clotting. If the EC is failing frequently, treatment can be ineffec-tive and costs increase with nursing and the need for new EC components [1].This paper presents a discussion of blood flow mechanics during CRRT andhow these factors may be important in the setting of clot formation and failureof the EC during CRRT.

Blood Pumping for Extracorporeal Circuits

Despite clot formation in the EC being the major impediment to success-ful operation of the CRRT system [2–7] investigations into the flow of bloodthrough the EC of a CRRT system and into genesis of clot formation due toflow problems has been limited. This is not the case in the context of car-diopulmonary bypass (CPB) systems where a larger EC and membrane is used

Baldwin/Bellomo 204

during heart surgery [8–10]. The characteristics of this EC are different to theEC for CRRT and have been well investigated [11–13]. Current CPB systemsare very efficient [8] and do not clot during use primarily due to the short timeframe of approximately 45–90 min [9], the large access catheters used and theability to administer higher doses of heparin [10]. In the CRRT setting, how-ever, blood flow is much slower (�200 ml/min), the access catheter is notplaced via a surgical approach direct into exposed vessels, the system isintended to be continuous (�24 h) in function and the critically ill patients arenot able to tolerate high doses of heparin. In comparison, CPB is a wellresearched, standardized, controlled, efficient technique with the above statedsystem factors favoring success. In CRRT, however, the mechanical factorsrelated to blood flow in the EC are different and have not been investigated. TheEC and components are those designed for intermittent therapy used in a con-tinuous setting. In preference, attention has mainly focused on the performanceassessment of anti-clotting drugs with the result that blood flow mechanics, theconsequences of flow reduction, small dual lumen access catheters and the dif-ferent access catheter sites used in the critically ill patient have been over-looked. There are many resistance points in the EC which have the potential toimpede blood flow and cause clot formation [4], e.g. the tubing, the shape andcontour of the venous air trap chamber, the connection points of the EC to theaccess catheter, the access catheter itself and the blood flow-pump mechanismgenerating blood flow. The cessation of blood flow in the EC due to clotting isa complex process and there may be interrelationships between resistancepoints. Clotting occurring quickly and unexpectedly, such as when anticoagu-lation is used or there is an abnormal clotting profile such as in liver failure,may be related to three factors:• Inadequate performance of anticoagulant drugs in this patient.• Mechanical factors within the EC involving a dysfunction of blood flow.• Both of these.

The significance of this insight may be in that if the focus is often on inad-equate performance of anticoagulant drugs, high doses or unnecessary changesto anticoagulant methods may be prescribed inappropriately. This occurs clini-cally when circuit clotting (time to failure or filter life) is thought to be due toinadequate anticoagulant drug effect, although the cause may be mechanical.Bedside diagnosis of mechanical factors may be difficult and is sometimesmade after alternative diagnoses are excluded. For example, a well-functioningEC is in use with anticoagulant drugs at therapeutic levels, and a patient withpoorly clotting blood in whom there would be no valid hematological reason forcircuit clotting, the EC suddenly clots. The relationship between pump perfor-mance, the EC tubing circuit, the access catheter and nursing activities duringtherapy may be as important as the effects of anticoagulant drugs. Roller pumps

Blood Flow, Access Catheter and Circuit Failure during CRRT 205

persist in current CRRT machine designs due to their perceived efficiency andsimplicity. This is despite reports of inefficiency [14, 15], damaging affect onred blood cells and platelets, and the fact that there is no accepted measure ofperformance during operation [16, 17].

Therefore until a new blood pump design is tested and validated to be bet-ter, an understanding of the potential failings of current roller pump technologyis useful for clinicians managing CRRT.

Roller Pumps and Segment Tubing

Blood roller pumps have been extensively studied in the context of cardiacbypass (CPB) surgery and their role in this form of extracorporeal circuit hasplayed an important role in the development of cardiac surgery [12]. Cooley [12]reviewed development of the roller pump, indicating that the first form of this waspatented in 1855 by Porter and Bradley. This was a two roller cam displacementpump suggested for many medical applications where pumping fluids is requiredand was further expanded upon to comprise three and four roller mechanisms inan extracorporeal context. Head et al. [13] described the addition of the extrarollers and how they decreased the stroke volume of the pump and were requiredto rotate faster to achieve the same flow and as a result of this increased the levelof blood haemolysis or blood cell damage. Setting the diameter of the roller camsin the pump housing and therefore the occlusiveness of the blood pump tubingwas thought to be a major cause of such hemolysis problems [13, 18, 19] and alsoaffected the efficiency of the blood pump in generating flow. Total tubing occlu-sion has been reported to be causative of hemolysis [20] but necessary to gener-ate effective blood flow whereas near occlusion or non-occlusive setting has beenreported to provide reduced blood flow [13, 19, 21] related to resistance down-stream of the roller; blood can slip back past the roller. In a review of roller pumpsin CPB, Noon et al. [21] stated that:

‘the optimization of occlusion setting in the CPB context would require anin-line flow meter and a pump with occlusion setting adjustable during operation.Rotary speed and roller occlusion could then be reciprocally adjusted to providethe maximum degree of under occlusion consistent with the required flow rates’.

In more simple terms, this would mean that the occlusion of the rollercould be widened during operation until measured flow started to fall. Thiswould set the correct occlusion for blood pumping, and at a level that wouldminimize haemolysis. Reed and Stafford [16] reinforced this concept. Theycited the example of a nonocclusive setting of a pump and the resistance ofvascular access cannulae affecting flow rate. In this situation, if an inlineflowmeter was not used, a patient may have a blood flow at below acceptable

Baldwin/Bellomo 206

rates with the pump operator being completely oblivious to the problem. Thisexperience is very important to users of CRRT machines. Roller pump occlu-sion is a necessary part of machine maintenance. How many users of CRRTmachines, particularly those that are independent of a dialysis service wheretechnicians are more focussed to machine maintenance attend to this?However, if the pump occlusion is correct, vascular access resistance maycause blood flow to drop and the blood pump pressure alarms along with theoperator not detect this. Anecdotal experience of the authors concluded thiswas occurring during our CRRT and may have contributed to premature filterclotting. This conclusion is derived from anecdotal observations such as:• Observing blood flow into venous air chambers to be poor despite the cor-

rect pump speed setting. This observation implies experience teaches theobservant; ‘what looks right’; a steady stream, not a dribble.

• Clotting of filter prematurely (2–3 h) despite anticoagulation drugs.• Observing air bubbles in the blood filled tubing moving forwards and then

backwards implying there is reduced net flow forwards in the EC.

Blood Pumps and Flow Doppler Assessment of Blood Flow

Depner et al. [14] assessed the pressure effects on roller pump blood flowduring dialysis. This report presented results from both in vitro and in vivowork. The in vitro study utilized either normal saline or haematocrit adjustedanticoagulated human blood at 37�C. The flow was measured with an ultrasonicprobe clamped around the pump outlet line. Clamps were applied to the distalends of both the inlet and outlet lines varying the pressures.

The in vivo work involved the measurement of blood flow in 64 patientsduring routine dialysis. In this in vivo study, the ultrasonic probe was placed onthe inlet side of the dialysis membrane. The investigators concluded that onlyinlet (pre-pump) pressure affected blood flow. This was due to low pressure ofblood resulting in poor filling of the pump housing tubing, and the subsequentstroke of the pump being ‘dry’ or less in stroke volume. Such an effect was notaccording to a proportional relationship, but rather the flow dropped sharplywhen the pressure inside the tubing fell to sub atmospheric levels (less than150 mm Hg) and the tubing collapses closing off the internal lumen and its abil-ity to refill. In the clinical context of CRRT this would suggest that blood flowwould abruptly stop without any warning signs, highlighting the need for reli-able access catheter blood flow/pressure. High pump outlet (post-pump) pres-sures did not effect flow, but did create a water hammer effect in that the tubingpulsated or moved in response to revolutions of the roller pump and there wasa corresponding reversal of flow as each roller completed its stroke and


occlusion was lost. This meant that the blood leaked backwards past the camcompressing the tubing. There was no change in the resultant average flow.Many physicians and nurses experienced with CRRT have observed this, butmay not be aware of the mechanism other than it appeared to be related toimminent filter clotting or high EC pressures.

In vivo assessment of blood flow indicated that measured flow was accuratewhen the prescribed roller pump speed was slow. A slow rate of pump roller rota-tion means that the blood pump housing tubing has more time for refill beforecam occlusion occurs again. The tubing recoil is complete following the previ-ous compression stroke and the tubing has time to refill with incoming blood.A slow pump speed was less than 50 ml/min. At high blood pump speeds(�100 ml/min) the measured flow was consistently lower than that set or pre-scribed. This is presumed to be due to the reduction in ‘refill time’ created by thefaster cam rotation.

In their discussion, the authors outlined an important mechanism in relationto the roller pump tubing segment which would explain differences between pre-scribed and actual flows. The proposed mechanism for flow reduction is based onthe collapse of the segment tubing into the roller pump due to a fall in inlet pres-sure. This fall in pressure can be caused by poor supply or availability of bloodinto the pump tubing. This creates impairment in the elastic recoil of the tubingin the pump housing after the compression stroke has occurred and the tubingsegment feeding the pump effectively runs dry, therefore it does not re-expand tothe correct size, and its volume drops. This loss of recoil or re-expansion of tub-ing causes a ‘decrease in pump output that is proportional to the reduction incross-sectional area’. The collapsed tubing may then fill from the reverse direc-tion after the roller completes its stroke and a subsequent drop in output occurs.

Access Catheters

In 1996, Uldall [22] published a review of vascular access for CRRT stat-ing that the femoral, subclavian, and internal jugular veins were the sites ofchoice for CRRT. The author concluded that the femoral vein was the suggestedchoice in immobile and generally supine patients. Most of the material presentedwas in the context of ease of insertion, and safety, given the risks as a medicalprocedure. Wynckel et al. [23] concur with this review in many respects but sug-gest that ‘prolonged femoral site catheterization increases both risks of throm-bosis and infection’. In both these reviews there were no comments with respectto the ongoing nursing care and management required and therefore no consid-eration of the site utilized and its suitability for the subsequent hours or days oftreatment providing reliable blood flow.

Baldwin/Bellomo 208

Bellomo and Ronco [24] comment on double lumen catheters making spe-cific reference to the fact that they must be capable of high blood flows with allpatient postures, physiotherapy, and when turning patients. They conclude thatthe safest and most expedient insertion site in the critically ill patient was thefemoral site. This is of particular concern when there is a need to establish theaccess quickly. The femoral vein is closer to the skin surface and makes fastaccess possible. Again, this is based on anecdotal clinical experience by thesephysicians and no scientific assessment.

Baldwin and Elderkin [2] related similar anecdotal experience withrespect to the preferred access site, indicating that the femoral site was moredifficult for the application of an adhesive dressing; was not easily observedwhen bed sheets were over the patient, and did limit the possibility of sittingthe patient upright or out of bed. These activities in a patient with a catheter inthe femoral site would be likely to kink off the catheter and stop blood flow.Sieffert and Sheppard [25] reviewed the nursing implications of vascularaccess sites, stating that any site should not be a cause of poor flow and mustnot impede the nursing management and progress of the patient in any way.They also made a comparison of the three sites indicating that the femoral sitewas susceptible to disturbed flow and was difficult to visualize under the bed-clothes. Furthermore, these authors noted that catheterization from the leftside of the chest can cause occlusion of the catheter as it tended to rest againstthe vessel wall and occlude the arterial lumen side holes. The authors statedthat this was evidenced by the low arterial pressure alarm occurring each timethe patient coughed during spontaneous breathing causing a change in thoracicpressure, and a reduction in access blood flow. These comments were, again,based on anecdotal experience of the authors and reinforce that in the litera-ture there was no independent assessment of dual lumen access catheters andthe effects of nursing care for a given access site. There is one study by Tapsonet al. [26] which compares the performance of two brands of catheter placedin the subclavian position. Their results revealed that from 223 patient treat-ments in 29 patients with the two brands of catheter, 11 catheters were removeddue to ‘catheter failure’ [p.197] or insufficient arterial flow, unresponsive tocatheter repositioning. This study did not refer to the patient activities duringthe treatments and implied that these catheter failures could not be explainedscientifically.

Access Site

Sands et al. [15] compared prescribed and delivered blood flow during 208dialysis treatments using an ultrasonic sensor to measure flow. Twenty-three of


the treatments were performed in patients with a dual lumen catheter for vas-cular access. The remaining 185 treatments were performed in patients whosevascular access was in the form of an arteriovenous fistula. All of the 23 treat-ments utilizing the vascular access catheter were associated with a blood flowwhich was less than that prescribed. There was a variety of pump speed settingstested (ranging from 200 to 500 ml/min) indicating the error was not speeddependent. Of the entire group being assessed, 22.8% had a delivered bloodflow which was at least 10% less than prescribed. It was not clear in the methodof this study how the data was collected in the context of the time period of thetreatment. It appeared that the readings taken were at the start of the treatmentand were a single data point analysis. In the discussion the investigators suggestthat vascular access catheters were ‘particularly problematic’ and confirm ‘thewell known difficulties with providing adequate dialysis on a long-term basiswith indwelling catheters’. Despite any criticism of this study, the findings doreflect that there is a real difficulty drawing blood from, or returning blood viathese access catheters. The authors suggest that this problem is particularlyhighlighted when they are in place for periods of weeks or months in patientsbeing treated in the chronic dialysis setting.

The experience described does reveal that access catheters used duringCRRT often fail to provide adequate blood to the blood pump they serve. Thismay be clearly obvious and is attended to accordingly. However, there may alsobe prolonged periods where the failure is not obvious and the blood pump failswithout any awareness but the reduced pump output becomes the cause of ECclotting and failure. This is where continuous and direct ultrasound blood flowmeasurement may have a place in future CRRT machines.

Blood Flow Monitoring

Blood flow monitoring cannot be performed with current CRRT tech-nology. This ultrasonic direct measurement has the potential to be linked tothe blood and fluids pumps via intelligence software automatically makingadjustment and alarm alerts for operators, or as previously cited, make adjust-ment to the roller cam occlusion and or pump speed when flow reduces dueto increases in downstream resistance. Furthermore, in addition, advisoryalarms may sound and UF/dialysate/replacement fluids would decrease tomaintain the blood flow UF ratio. An advisory for ‘pump calibration’ mayalso appear.

We have developed the technique for continuous ultrasound blood flowmonitoring [27] and are now further validating its usefulness. The following isa brief review of the method.

Baldwin/Bellomo 210

Blood Flow Probe

The ultrasound flow probe was manufactured to our tubing specificationsby Neomedix systems (Neomedix, Ithaca, N.Y., USA) and is designed to inter-face with a Transonic HT 109 flow monitor (Neomedix).

This probe is a four-crystal beam miniature device which clamps aroundthe blood tubing illuminating the blood flow between the opposing walls of thetubing. A microchip processor receiving the Doppler signals calculates transittime for blood passing through the probe head and displays a measure of flowon a monitor (digital only) and computer screen (wave and digital) (fig. 1). Theprobe provides 200 samples per second to the software as millivoltage to gen-erate a flow wave. This high frequency response ensures that instantaneous flowis accurate. The millivolt signal can then be cross-calibrated or zeroed againsta known flow rate or no flow respectively for display in ml/min.

Doppler Flow Measurement Unit and Computer Interface

Data is acquired from this flow monitor by a lap top computer via anRS232 port. A Windows based data acquisition and playback software program(WinDaq Software, Transonic Systems Inc., Ithaca, N.Y., USA) converts thisdata to a waveform for recording and playback. This waveform indicates a peakand trough to the blood flow consistent with the pulsatile nature of the rollerpump rotation. Flow reduction can then be detected from this wave as it occursin association with the digital display for max., min., and mean [27]. This wave

Set up blood flow monitoring

Bloodpump

* Flow monitorMini doppler

* HT 109 transonic flow meter–Neomedix systems

Flow monitor

Computer

Lap top PC.

Fig. 1. Uncompressed view of the normal flow wave patterns seen during peristalticpump action for the maintenance of blood flow during CRRT.


display has the same usefulness as current hemodynamic monitors for patientarterial blood pressure (fig. 1).

EC Failure

Figure 2 shows a long period of blood flow wave where the view is com-pressed using a software ‘option’ for ‘view’. Point A indicates normal flow setat 200 ml/min. Point B indicates a nursing patient positional change whereaccess occlusion occurs and flow drops severely for a short time and then recov-ers. Point C indicates a period of deteriorating flow over approximately 45 min(with no operator awareness) until point D where the filter clots and blood flowstops completely with no recovery possible.

Conclusion

There is an important relationship between the access catheter, the bloodpump and the hemofilter. This relationship may be even more important in

Fig. 2. Compressed blood flow wave revealing onset of a flow reduction and subse-quent associated filter clotting.

Flow reductionbegins here

Flow reduction identified in compressed wave view

Flow reduction period

Peak flow 198mls/min

DCBA

Failure

Trough flow 83mls/min

Baldwin/Bellomo 212

respect of the functional time before the EC or filter clots. Many clinicians maybe treating patients where they are unaware that blood flow is incorrect fromtheir prescription and/or are making unnecessary changes to anticoagulation inresponse to EC failure. Blood flow may be as important as other measures toprevent clotting in the EC during routine CRRT. Blood flow monitoring can beconsidered a logical next step in the design of future CRRT machines.

References

1 Baldwin IC, Bridge NP, Elderkin TD: Nursing Issues, Practices, and Perspectives for theManagement of Continuous Renal Replacement Therapy in the Intensive Care Unit. Critical CareNephrology. Dordrecht, Kluwer, 1998, pp 1309–1327.

2 Baldwin I, Elderkin T: Continuous hemofiltration: Nursing perspectives in critical care. NewHorizons 1995;3:738–747.

3 Webb AR, Mythen MG, Jacobsen D, Mackie IJ: Maintaining blood flow in the extracorporealcircuit: Heamostasis and anticoagulation. Intens Care Med 1995;21:84–93.

4 Favre H, Martin PY, Stoermann C: Anticoagulation in continuous extracorporeal renal replacementtherapy. Semin Dial 1996;9:112–118.

5 Martin P-Y, Chevrolet JC, Suter P, Favre H: Anticoagulation in patients treated by continuousvenovenous hemofiltration: A retrospective study. Am J Kid Dis 1994;24:806–812.

6 Mehta R: Anticoagulation strategies for continuous renal replacement therapies: What works? AmJ Kid Dis 1996;28(suppl 3):806–812.

7 Mehta RL, Dobos GJ, Ward DM: Anticoagulation in continuous renal replacement procedures.Semin Dial 1992;5:61–68.

8 Connolly MW, Guyton RA: Cardiopulmonary bypass and intra operative protection; in Schlant RC,Alexander RW (eds): Hurst’s The Heart, ed 8. New York, McGraw Hill, 1994, vol 141, chap 22,pp 2443–2450.

9 Phillips SJ: Emergent cardiopulmonary bypass. Ann Thorac Cardiovasc Surg 1993;55:1281–1282.10 Bell PE, Diffee GT: Cardiopulmonary bypass. Principles, nursing implications. AORN J 1991;53:

1480–1496.11 Noon GP, Harrell JE, Feldman L, Petrson J, Kent PM, De Bakey ME: Development and evaluation

of pulsatile roller pump and tubing for cardiac assistance. Artif Org 1983;7:49–54.12 Cooley D: Development of the roller pump for use in the cardiopulmonary bypass circuit. Tex

Heart Inst J 1987;14:113–118.13 Head LR, Coenen JP, Angola E, Nogueira C, Mendelsohn D, Kay EB: Operation of the roller

pump for extracorporeal circulation. J Thorac Cardiovasc Surg 1960;39:210–220.14 Depner TA, Rizwan S, Stasi TA: Pressure effects on roller pump blood flow during hemodialysis.

ASAIO Trans 1990;36:m456–m459.15 Sands J, Glidden D, Jacavage W, Jones B: Difference between delivered and prescribed flow in

hemodialysis. ASIO J 1996;42:M717–M719.16 Reed CC, Stafford TB: Cardiopulmonary Bypass, ed 2. Houston, Texas Medical Press, 1985, p 376.17 Ronco C, Brendolan A, Bellomo R: Current technology for continuous renal replacement thera-

pies. Crit Care Nephrol 1998;1269–1308.18 Bernstein EF, Gleason LR: Factors influencing hemolysis with roller pumps. Surgery 1967;61:

432–442.19 Hodges PC, Cardozo R, Thevenet A, Lillehei CW: Comparison of relative merits of occlusive and

nonocclusive pumps for open-heart surgery. J Thorac Surg 1958;36:470–478.20 Tamari Y, Lee-Sensiba K, Leonard EF, Tortolani A: A dynamic method for setting roller pumps

non-occlusively reduces hemolysis and predicts retrograde flow. ASAIO J 1997;43:39–52.21 Noon GP, Kane LE, Feldman L, Peterson JA, DeBakey ME: Reduction of blood trauma in roller

pumps for long term perfusion. World J Surg 1985;9:65–71.


22 Uldall R: Vascular access for continuous renal replacement therapy. Semin Dial 1996;9:93–97.23 Wynckel A, Melin JP, Hanrotel C, Chanard J: Catheter and line requirements for continuous hemofil-

tration; in Journois D (ed): Continuous Hemofiltration in the Intensive Care Unit. Amsterdam,Harwood Academic Publishers, 1997, chap 18.

24 Bellomo R, Ronco C: Circulation of the continuous artificial kidney: Blood flow, pressures, clear-ances and the search for the best. Blood Purif 1997;15:354–365.

25 Sieffert E, Sheppard M: Continuous hemofiltration: The nursing perspective; in Journois D (ed):Continuous Hemofiltration in the Intensive Care Unit. Amsterdam, Harwood Academic Publishers,1997, chap 12.

26 Tapson JS, Hoenich NA, Wilkinson R, Ward MK: Dual lumen subclavian catheters for hemodial-ysis. Int J Artif Org 1985;8:195–200.

27 Baldwin I, Bellomo R, Koch B: A technique for the monitoring of blood flow during continuoushemofiltration. Intens Care Med 2002;28:1361–1364.

Ian BaldwinDepartment of Intensive CareAustin Hospital, Studley Rd, Heidelberg, Victoria 3084 (Australia)Tel. �1 61 3 9496 3340, Fax �1 61 3 9496 3932, E-Mail [email protected]


CRRT: Selection of Patients and Starting Criteria

Paul M. Palevsky

Renal Section, VA Pittsburgh Healthcare System and Renal-Electrolyte Division, Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pa., USA

Acute renal failure (ARF) is a common complication of critical illness witha mortality in excess of 50% [1]. Despite more than a half-century of experiencein the use of hemodialysis and other forms of renal support in the managementof ARF, consensus on the optimal management of renal replacement therapy inARF does not exist. In particular, there are no standard guidelines for the use ofcontinuous renal replacement therapies (CRRT), resulting in wide variations inpractice patterns across regions and between individual centers within regions.

The First Acute Dialysis Quality Initiative (ADQI) consensus conferencewas held in New York in August, 2000 with a goal of reviewing the availableevidence regarding optimal management of CRRT in order to make evidence-based practice recommendations and to delineate key issues for future research[2]. Among the topics covered by the participants of this conference was theselection of patients for acute extracorporeal renal support in ARF with partic-ular emphasis on the timing of initiation and selection of patients for CRRT [3].Specific issues addressed included indications for renal replacement therapy inpatients with ARF, timing of initiation of acute extracorporeal renal support,patient and institutional characteristics for selecting CRRT for the managementof ARF, and non-renal indications for CRRT. In this review, the consensusfindings on the selection of patients and criteria for initiation of CRRT arediscussed in light of studies published in the past 3 years.

Patient Selection for Acute Extracorporeal Renal Support

Therapy for intrinsic ARF is primarily supportive, with no effective pharma-cologic therapy for renal failure due to acute tubular necrosis [4, 5]. Prior to the

Renal Replacement Therapy in the ICU: Consensus andRecommendations from ADQI

CRRT: Selection of Patients and Starting Criteria 215

development of dialysis as an effective renal replacement therapy, the mostcommon causes of death in acute renal failure were directly related to uremia,electrolyte disturbances (primarily intractable metabolic acidosis and hyper-kalemia), volume overload and hemorrhagic diatheses. Although mortalityrates in ARF remain high despite the advent of dialysis, the causes of death havechanged, with sepsis, gastrointestinal bleeding, cardiovascular and pulmonarydysfunction and withdrawal of life-support now being most common [4, 6, 7].

The timing for initiation of renal support in acute renal failure remainscontroversial. In its initial use, in the decade following World War II, hemodial-ysis was applied to patients with advanced symptoms of renal failure, includingclinical uremia, severe hyperkalemia and pulmonary edema [8–10]. Althoughcontrol of uremic symptoms and volume overload were achieved, a clear reduc-tion in mortality could not be demonstrated [10]. Teschan et al. [10] introducedthe concept of ‘prophylactic’ dialysis, initiated prior to the onset of overt symp-toms, in the treatment of ARF in military causalities during the Korean conflict.Multiple studies in the intervening four decades have attempted to define theappropriate timing for initiation of renal support in ARF.

In their landmark report, Teschan et al. [10] described a prospective,uncontrolled series of 15 patients with oliguric ARF in whom dialysis wasinitiated ‘prophylactically’, before the BUN reached 100 mg/dl (36 mmol/l). Allcause mortality was 33% with mortality due to hemorrhage or sepsis of 20%.Although no control group was studied, the authors reported that the resultscontrasted dramatically with their own past experience in patients in whomdialysis was not initiated until ‘conventional’ indications were present.

Similar conclusions were reached in a series of retrospective studies pub-lished in the 1960s and early 1970s. Easterling and Forland [11] reported on aseries of 45 patients with ARF initiated on dialysis prior to the onset of symp-toms and concluded that early initiation of therapy to prevent uremic symptomswas desirable. However, their study lacked a control group, and they wereunable to draw any conclusions regarding improved survival with early dialy-sis. In a retrospective analysis of 33 patients with postoperative ARF treatedwith hemodialysis during two different periods, Parsons et al. demonstrated asurvival of 75% in patients in whom dialysis was initiated when the BUN wasbetween 120 and 150 mg/dl (43–54 mmol/l) as compared to 12% survival inpatients in whom dialysis was not initiated until the BUN was greater than200 mg/dl (71 mmol/l) [12]. Similarly, in a retrospective series of 162 patients,Fischer et al. [13] observed a mortality of 57% in patients with ARF in whomdialysis was initiated when the BUN reached 150 mg/dl (54 mmol/l) orwhen clinical deterioration was first observed, as compared to 74% mortality inpatients in whom dialysis was not initiated until the BUN was greater than200 mg/dl.

Palevsky 216

In the largest of these retrospective studies, Kleinknecht et al. [14] reportedon 500 patients with ARF treated between 1966 and 1970. Patients receiving‘prophylactic’ dialysis, defined as early and frequent treatment to maintain pre-dialysis BUN less than 93 mg/dl (33 mmol/l), had a mortality of 29% ascompared to 42% mortality in patients in whom dialysis was initiated only ifthe BUN was greater than 163 mg/dl (58 mmol/l) or if severe electrolyte distur-bances were present. A marked reduction in mortality due to sepsis and gas-trointestinal bleeding was observed in the more aggressively dialyzed group.

The first prospective evaluation of ‘prophylactic’ dialysis in ARF wasreported by Conger [15] in 1975. He described 18 patients with post-traumaticARF sustained during the Vietnam War and treated on the Naval Hospital ShipUSS Sanctuary between April and October 1970. Patients were alternatelyassigned to an intensive dialysis regimen to maintain the pre-dialysis BUN andcreatinine at less than 70 mg/dl (25 mmol/l) and 5 mg/dl (440 �mol/l), respec-tively, or a non-intensive regimen in which dialysis was not initiated until theBUN approached 150 mg/dl (54 mmol/l) and the creatinine reached 10 mg/dl(885 �mol/l), or the patient developed clinical indications for dialysis (hyper-kalemia, volume overload or uremic encephalopathy). Survival was 64% (5 of8 patients) in the intensive treatment group as compared to 20% (2 of 10 patients)in the non-intensive dialysis group. In addition, complications of hemorrhage(36 vs. 60%) and gram negative sepsis (50 vs. 80%) were less frequent in theintensive treatment group.

Expanding on this study, Gillum et al. [16] studied 34 patients with ARFwho were randomized to receive either intensive hemodialysis, to keep theBUN �60 mg/dl (21 mmol/l) and the serum creatinine �5 mg/dl (440 �mol/l),or non-intensive dialysis, allowing the BUN to reach 100 mg/dl (36 mmol/l) andthe serum creatinine to reach 9 mg/dl (795 �mol/l). Patients were stratified basedon etiology of ARF (trauma-surgery or medical) and randomized in a pairedfashion when the serum creatinine reached 8 mg/dl (705 �mol/l). Although notassessing only timing of initiation of therapy, mortality was higher in the inten-sively dialyzed group (58.8 vs. 47.1%), but given the small sample size, thiswas not statistically significant (p � 0.73). Hemorrhagic and septic complicationswere more common in the non-intensively dialyzed group; however, these differ-ences also did not reach statistical significance.

Gettings et al. [17] have reported the results of a retrospective analysis ofearly (BUN �60 mg/dl; 21 mmol/l) versus late (BUN �60 mg/dl; 21 mmol/l)initiation of continuous venovenous hemofiltration (CVVH) in 100 adult patientswith post-traumatic acute renal failure. The 41 patients who were ‘early’ starterswere younger (40.5 � 17.9 vs. 48.0 � 18.9 years) but otherwise comparableto the 58 ‘late’ starters. Patients had similar Injury Severity Scores (early:33.0 � 13.5; late: 37.2 � 15.0) and Glasgow Coma Scores (early: 11.8 � 3.8;


late: 12.5 � 3.7) on admission. No other indices of severity of illness werereported. CRRT was initiated on day 10.5 � 15.3, when the BUN was42.6 � 12.9 mg/dl (15.2 � 4.6 mmol/l) in the ‘early’ group as compared to day19.4 � 27.2, when the BUN was 94.5 � 28.3 mg/dl (33.8 � 10.2 mmol/l) inthe ‘late’ group. Survival was 39.0% in the ‘early’ group as compared to 20.3%in the late group.

Based on these data, the ADQI conference concluded that no specific rec-ommendations on the timing of initiation of renal replacement therapy could bemade beyond the conventional criteria of pulmonary edema not responsive todiuretics, hyperkalemia, metabolic acidosis and uremic complications [3].However, because the consequences of these complications are likely to be moresevere in critically ill patients with ARF, they recommended that renal replace-ment therapy should usually begin before these manifestations are present [3].They further recommended that the issue of timing of initiation of renal replace-ment therapy should be the subject of further research.

Since the consensus conference, one additional study has been publishedcomparing early (within 12 h of fulfilling criteria for ARF based on presence ofoliguria or a creatinine clearance of less than 20 ml/min; n � 35) or late (BUNgreater than 112 mg/dl (40 mmol/l), potassium greater than 6.5 mmol/l or pul-monary edema; n � 36) initiation of continuous hemofiltration [18]. The meanBUN was 48 mg/dl (17.1 mmol/l) in the early treatment group and 104.4 mg/dl(37.4 mmol/l) in the late treatment group. No difference in survival was observedbetween the two groups, with 28-day survivals of 68.8% in the early initiationgroup and 75.0% in the late initiation group (p � 0.80). This study was, how-ever, markedly underpowered. Thus, the data remain insufficient to make addi-tional recommendations regarding the timing of initiation of renal replacementtherapy in ARF.

Patient Selection for CRRT in ARF

It is generally accepted that hemodynamically unstable patients tolerateCRRT better than intermittent therapies. The initial development of these ther-apies was to permit renal support to hemodynamically unstable patients whowere unable to tolerate conventional hemodialysis. The slower, more gradualremoval of fluid and solute during CRRT results in less hemodynamic compro-mise than occurs with intermittent hemodialysis. This prescribing pattern isborne out in both retrospective and prospective observational comparisons ofCRRT and intermittent hemodialysis [19, 20]. In a single-center retrospectivecomparison of CRRT and intermittent hemodialysis, Swartz et al. [19] observedthat patients treated with CRRT had a higher acuity of illness, with lower mean

Palevsky 218

systolic blood pressure, lower platelet count and greater percentage of patientsrequiring ventilatory support despite less advanced renal failure (lower BUNand creatinine) at the time of initiation of renal support.

Similar findings were observed in a prospective, multicenter, observa-tional study comparing continuous and intermittent renal replacement therapythat was published after the consensus conference [20]. The mean SAPS IIscore at time of initiation of renal support was 62 � 18 in 354 patients treatedwith CRRT as compared to 58 � 22 in 233 patients treated with intermittenttherapy. Cardiovascular dysfunction was present in 85.9% of CRRT patients ascompared to 53.6% of patients treated with intermittent therapy. Similarly, resp-iratory failure requiring ventilatory support was present in 92.7% of patientstreated with CRRT as compared to 71.2% of patients initially treated with inter-mittent therapy. The mean number of failed organs was 3.6 � 1.1 in patientsinitially treated with CRRT as compared to 2.8 � 1.2 in patients initially treatedwith intermittent therapy. Thus, these two studies suggest that there is greaterutilization of continuous renal support in patients with greater acuity of illness.

The question of whether CRRT offers a survival advantage over intermit-tent hemodialysis in the management of ARF remains unresolved. Althoughunadjusted mortality was higher in patients treated with CRRT in the two obser-vational studies discussed above, modality of renal replacement therapy had noprognostic value with regard to mortality risk after risk adjustment [19, 20].

Only one large randomized controlled trial comparing continuous to inter-mittent therapy has been published in a peer-reviewed journal [21]. In this studyof 166 patients with ARF randomized to CRRT or intermittent hemodialysis, anintention-to-treat analysis found a 28-day all-cause mortality of 59.5% inpatients randomized to CRRT as compared to 41.5% in patients randomized tointermittent therapy with in-hospital mortality rates of 65.5 and 47.6%, respec-tively. However, this study was flawed by unbalance randomization, whichresulted in significantly higher APACHE III scores and a significantly greaterpercentage of patients with liver failure in the CRRT group. Using multivari-able stepwise logistic-regression analysis, hepatic failure, APACHE III scoreand Organ System Failure (OSF) score were all independently related to ICUmortality. In this analysis, modality of therapy was not independently associ-ated with mortality, with an odds of death associated with randomization toCRRT of 1.58 (95% CI: 0.7–3.3). Similarly, a time-to-event analysis using aCox proportional hazards model yielded a similar conclusion (hazard ratio withCRRT: 1.35; 95% CI: 0.89 to 2.06). Despite the higher mortality in the CRRTgroup, surviving patients initially treated with CRRT had higher rates of recov-ery of renal function.

Two meta-analyses have been published since the consensus conferencecomparing intermittent to continuous therapy in ARF [22, 23]. Kellum et al. [22]


analyzed 13 studies encompassing a total of 1,400 patients with ARF. Only 3 ofthe 13 studies were prospective randomized trials. Overall, there was no differ-ence in mortality in the pooled analysis. Adjusting for study quality and sever-ity of illness, the authors calculated an adjusted relative risk of death in patientstreated with CRRT of 0.72 (95% CI: 0.60–0.87). In a subanalysis of six studiesin which patients treated with CRRT and intermittent therapy had similarbaseline severity of illness, the unadjusted relative risk of death with CRRT was0.48 (95% CI: 0.34–0.69). In the second meta-analysis, Tonelli et al. [23] foundno difference in outcome associated with the modality of renal support insix randomized controlled trials encompassing 624 patients. The difference inconclusions between these two studies reflects the different criteria for studyinclusion.

Based on the data available at the time of the consensus conference, nofirm recommendations for patient selection for CRRT could be made, except inpatients who have, or are at risk, for cerebral edema [3]. In this subgroup ofpatients, data suggest that treatment with continuous hemofiltration is lesslikely to be associated with increased intracranial pressure and decreasedcerebral perfusion pressure as compared to intermittent therapy [24–27]. Theobservational study [20] and two meta-analyses [22, 23] published since theconsensus conference have not provided sufficient additional data to alter thisconclusion. A large prospective randomized controlled trial comparing inter-mittent to continuous renal replacement therapy will be required to resolvethis issue.

Non-Renal Indications for CRRT

There are multiple case-series of CRRT use in settings other than ARF,including acute intoxications [28], cardiac failure [29] and ARDS [30]. Significantinterest has focused on the use of CRRT in the management of sepsis and sys-temic inflammation [31, 32]. For example, in trauma patients without renal fail-ure, CVVH has been demonstrated to be associated with decreases in bothTNF-� and IL-6 levels [33] and was associated with improved hemodynamicstability [34]. It is uncertain, however, whether the cytokine clearance is due toconvective clearance or membrane adsorption [35]. The clinical relevance ofthe cytokine clearance that has been observed has also been questioned, as theclearances achievable with CRRT are markedly lower than their biologicalturnover [36]. Newer modalities using plasma filtration and adsorption mayprovide greater promise [37]. At present, however, the evidence remains insuf-ficient to recommend the use of CRRT for non-renal indications outside ofappropriately designed clinical trials [3].

Palevsky 220

References

1 Liano F, Junco E, Pascual J, Madero R, Verde E: The spectrum of acute renal failure in the intensivecare unit compared with that seen in other settings. The Madrid Acute Renal Failure Study Group.Kidney Int 1998;66(suppl):S16–S24.

2 Kellum JA, Mehta RL, Angus DC, Palevsky P, Ronco C: The first international consensus conferenceon continuous renal replacement therapy. Kidney Int 2002;62:1855–1863.

3 Bellomo R, Angus DC, Star RA: The Acute Dialysis Quality Initiative. II. Patient selection for CRRT.Adv Ren Replace Ther 2002;9:255–259.

4 Thadhani R, Pascual M, Bonventre JV: Acute renal failure. N Engl J Med 1996;334:1448–1460.5 Star RA: Treatment of acute renal failure. Kidney Int 1998;54:1817–1831.6 Turney JH: Why is mortality persistently high in acute renal failure? Lancet 1990;335:971.7 Woodrow G, Turney JH: Cause of death in acute renal failure. Nephrol Dial Transplant 1992;7:

230–234.8 Kolf WJ: First clinical experience with the artificial kidney. Ann Intern Med 1965;62:608–619.9 Merrill JP, Smith S, Callahan EJ, Thorn GW: Use of artificial kidney. II. Clinical experience.

J Clin Invest 1950;29:425–438.10 Teschan PE, Baxter CR, O’Brien TF, Freyhof JN, Hall WH: Prophylactic hemodialysis in the treat-

ment of acute renal failure. Ann Intern Med 1960;53:992–1016.11 Easterling RE, Forland M: A five year experience with prophylactic dialysis for acute renal failure.

Trans Am Soc Artif Intern Organs 1964;10:200–208.12 Parsons FM, Hobson SM, Blagg CR, McCracken BH: Optimum time for dialysis in acute reversible

renal failure: Description and value of improved dialyzer with large surface area. Lancet 1961;i:129–134.

13 Fischer RP, Griffen WOJ, Reiser M, VClark DS: Early dialysis in the treatment of acute renal failure.Surg Gynecol Obstet 1966;123:1019–1023.

14 Kleinknecht D, Jungers P, Chanard J, Barbanel C, Ganeval D: Uremic and non-uremic complicationsin acute renal failure: Evaluation of early and frequent dialysis on prognosis. Kidney Int 1972;1:190–196.

15 Conger JD: A controlled evaluation of prophylactic dialysis in post-traumatic acute renal failure.J Trauma 1975;15:1056–1063.

16 Gillum DM, Dixon BS, Yanover MJ, et al: The role of intensive dialysis in acute renal failure. ClinNephrol 1986;25:249–255.

17 Gettings LG, Reynolds HN, Scalea T: Outcome in post-traumatic acute renal failure whencontinuous renal replacement therapy is applied early vs. late. Intens Care Med 1999;25:805–813.

18 Bouman CS, Oudemans-Van Straaten HM, Tijssen JG, Zandstra DF, Kesecioglu J: Effects of earlyhigh-volume continuous venovenous hemofiltration on survival and recovery of renal function inintensive care patients with acute renal failure: A prospective, randomized trial. Crit Care Med2002;30:2205–2211.

19 Swartz RD, Messana JM, Orzol S, Port FK: Comparing continuous hemofiltration with hemodial-ysis in patients with severe acute renal failure. Am J Kidney Dis 1999;34:424–432.

20 Guerin C, Girard R, Selli JM, Ayzac L: Intermittent versus continuous renal replacement therapyfor acute renal failure in intensive care units: Results from a multicenter prospective epidemio-logical survey. Intens Care Med 2002;28:1411–1418.

21 Mehta RL, McDonald B, Gabbai FB, et al: A randomized clinical trial of continuous versus inter-mittent dialysis for acute renal failure. Kidney Int 2001;60:1154–1163.

22 Kellum JA, Angus DC, Johnson JP, et al: Continuous versus intermittent renal replacement therapy:A meta-analysis. Intens Care Med 2002;28:29–37.

23 Tonelli M, Manns B, Feller-Kopman D: Acute renal failure in the intensive care unit: A systematicreview of the impact of dialytic modality on mortality and renal recovery. Am J Kidney Dis 2002;40:875–885.

24 Davenport A, Will EJ, Davison AM, et al: Changes in intracranial pressure during machine andcontinuous haemofiltration. Int J Artif Organs 1989;12:439–444.


25 Davenport A, Finn R, Goldsmith HJ: Management of patients with renal failure complicated bycerebral oedema. Blood Purif 1989;7:203–209.

26 Davenport A, Will EJ, Davison AM: Early changes in intracranial pressure during haemofiltrationtreatment in patients with grade 4 hepatic encephalopathy and acute oliguric renal failure. NephrolDial Transplant 1990;5:192–198.

27 Davenport A, Will EJ, Davison AM: Continuous vs. intermittent forms of haemofiltration and/ordialysis in the management of acute renal failure in patients with defective cerebral autoregulationat risk of cerebral oedema. Contrib Nephrol. Basel, Karger, 1991, vol 93, pp 225–233.

28 Bellomo R, Kearly Y, Parkin G, Love J, Boyce N: Treatment of life-threatening lithium toxicity withcontinuous arterio-venous hemodiafiltration. Crit Care Med 1991;19:836–837.

29 Biasioli S, Barbaresi F, Barbiero M, et al: Intermittent venovenous hemofiltration as a chronictreatment for refractory and intractable heart failure. ASAIO J 1992;38:M658–M663.

30 DiCarlo JV, Alexander SR, Agarwal R, Schiffman JD: Continuous veno-venous hemofiltration mayimprove survival from acute respiratory distress syndrome after bone marrow transplantation orchemotherapy. J Pediatr Hematol Oncol 2003;25:801–805.

31 Ronco C, Tetta C, Mariano F, et al: Interpreting the mechanisms of continuous renal replacementtherapy in sepsis: The peak concentration hypothesis. Artif Organs 2003;27:792–801.

32 Tetta C, D’Intini V, Bellomo R, et al: Extracorporeal treatments in sepsis: Are there new perspec-tives? Clin Nephrol 2003;60:299–304.

33 Sanchez-Izquierdo JA, Perez Vela JL, Lozano Quintana MJ, Alted Lopez E, Ortuno de Solo B,Ambros Checa A: Cytokines clearance during venovenous hemofiltration in the trauma patient.Am J Kidney Dis 1997;30:483–488.

34 Sanchez-Izquierdo Riera JA, Alted E, Lozano MJ, Perez JL, Ambros A, Caballero R: Influence ofcontinuous hemofiltration on the hemodynamics of trauma patients. Surgery 1997;122:902–908.

35 Bouman CS, van Olden RW, Stoutenbeek CP: Cytokine filtration and adsorption during pre- andpostdilution hemofiltration in four different membranes. Blood Purif 1998;16:261–268.

36 Sieberth HG, Kierdorf HP: Is cytokine removal by continuous hemofiltration feasible? Kidney Int1999;72(suppl):S79–S83.

37 Bellomo R, Tetta C, Brendolan A, Ronco C: Coupled plasma filtration adsorption. Blood Purif2002;20:289–292.

Paul M. Palevsky, MDRoom 7E123 (111F-U), VA Pittsburgh Healthcare SystemUniversity Drive Division, Pittsburgh, PA 15240 (USA)Tel. �1 412 688 6000 ext 5932, Fax �1 412 688 6908, E-Mail [email protected]


Fluid Composition for CRRT

Martine Leblanc

University of Montreal, Nephrology and Critical Care, Maisonneuve-Rosemont Hospital, Montreal, P.Q., Canada

Electrolyte and Acid-Base Disturbances in CRRT Patients

Critically ill patients with acute renal failure often present increased potas-sium and phosphate serum concentrations as well as metabolic acidemia. Serumsodium level may vary but can be found low in cases of massive fluid overload.

To facilitate our understanding of electrolytes and acid-base imbalance incritical illness, the ‘strong ion approach’ is very useful [1]. A strong ion is con-sidered almost completely dissociated in a solution; strong cations in bloodplasma are sodium (Na�), potassium (K�), calcium (Ca2�) magnesium (Mg2�),whereas strong anions are chloride (Cl�), and lactate until metabolized. Thestrong ion difference corresponds to the net charge balance of all strong ionspresent in a solution; the apparent strong ion difference in blood plasma canbe calculated as: [(Na� � K� � Ca2� � Mg2�) � (Cl� � lactate)]. Since the‘normal’ apparent strong ion difference of human blood plasma is 40–42 mmol/l,it can be concluded that unmeasured anions are normally present [2]. However,in critical illness, the apparent strong ion difference is often found lower, in the30–35 mmol/l range, as a consequence of an underlying metabolic acidosis withexcessive unmeasured anions, and of a low serum albumin [1].

More than 15,000 mmol of CO2, as much as 4,500 mmol of lactic acid, and100–200 mmol of other nonvolatile acids are produced daily by the human bodyfrom metabolism and nutrition [2]. During critical illness, volatile and non-volatile acid load may increase as a result of enhanced catabolism. Althoughseveral pathways counteract the acid burden, nonvolatile acid elimination isreduced in renal failure and a normal strong ion difference can not be restoredby the failing kidneys [3]. In renal failure, a mixed type of metabolic acidosisboth hyperchloremic and high-anion gap is commonly found [4]; the apparent

Fluid Composition for CRRT 223

strong ion difference of renal failure patients is usually reduced. Chloride andunmeasured anions in blood plasma (phosphates, sulfates, various organic acidsnot completely oxidized, and other unknown molecules) just accumulate, beingnot eliminated by kidneys. A high anion gap metabolic acidosis in criticallyillness may indicate lactate accumulation, although hyperlactatemia does notnecessarily translate into acidemia.

Fluids for CRRT

Besides azotemia control, a major aim of CRRT is the restoration of phys-iologic electrolyte concentrations and acid-base balance in blood plasma. Thechoice of fluids for CRRT will be determinant in the achievement of such a goal,and not only their composition but also the operational conditions of CRRT willhave an impact. Fluids used for CRRT include dialysate and replacement/sub-stitution solutions. Solutions generally used for CRRT contain electrolytesin concentrations similar to those unbound to proteins in blood plasma. Thestrong ion difference of those solutions has a significant impact on the acid-basebalance, especially when high rates or massive amounts are delivered.

Some of the commercially available solutions and their content are presentedin table 1. It can be appreciated that strong ion differences of those solutions arerelatively close to the strong ion difference of normal human blood plasma. Forrare or very specific cases, customized fluids may be necessary, but this iscertainly not a major concern.

As the main alkalinazing anion, lactate and bicarbonate are most frequentlyused in CRRT solutions. Buffer anion in CRRT fluids is usually in a concentration

Table 1. Composition of some of the commercially available fluids forCRRT (ion concentrations in mmol/l)

Hemosol Hemosol B0 Premixed NormocarbLG2 (Gambro) dialysate (Vaughan)(Gambro) (Baxter)

Na� 142 140 142 140K� 2 0 2 0Cl� 109.5 109.5 117 106.5Ca2� 1.75 1.75 1.75 0Mg2� 0.75 0.5 0.75 0.75Phosphate 0 0 0 0HCO3

� 0 32 32 35Lactate� 40 3 32 0

Leblanc 224

above 25–30 mmol/l [5]. Controlled trials have demonstrated that lactate orbicarbonate buffered solutions for CRRT have a similar efficacy to correctmetabolic acidosis [6–8]. However, serum lactate was found slightly higherwith lactate solutions without clear evidence of detrimental impact on outcome.Nonetheless, this may complicate the clinical interpretation of blood lactatelevels. Generally, the use of bicarbonate-buffered fluids has been associatedwith an improved control of metabolic acidosis [9].

Lactate solutions are usually well tolerated since the human body canrapidly metabolize large amounts of lactate (at a rate of 100 mmol/h for as muchas 2,000 mmol/day). It is converted by the liver into bicarbonate in a 1:1 ratioand usually does not remain in circulation as a strong ion for a long time.Although lactate solutions are well tolerated during CRRT when infused atmoderate flows of 2–3 liters/h, the infusion of large amounts of D,L-lactatecould possibly increase catabolism and induce cerebral dysfunction [10]. Inaddition, in patients who are limited in lactate metabolism (liver failure, severeshock), lactate solutions should be avoided. In such circumstances, infused lac-tate remains in circulation as a strong anion and endogenous bicarbonate is lostthrough the effluent, both contributing to a metabolic acidosis. Bicarbonate-buffered fluids are preferred in such cases. However, since bicarbonate solutionsfor CRRT are not widely available, they may have to be prepared locally by theinstitution.

Solutions used for CRRT may else contain citrate as the main alkalinizinganion [5, 11]. Citrate is converted to bicarbonate by the liver and by the musclein a 1:3 ratio. Citrate is used for regional anticoagulation of the extracorporealcircuit during CRRT and is particularly appealing for patients at risk for bleeding.Most studies on citrate-CRRT show correction of the metabolic acidosis [12].However, since one molecule of citrate provides 3 molecules of bicarbonate,metabolic alkalosis may be a consequence.

Acetate has also been proposed as the alkalinizing anion, since it is metab-olized into bicarbonate by the liver and muscle in a 1:1 ratio and will contributeto the restoration of the strong ion difference. However, acetate has not beenfrequently used in CRRT fluids, and potential uneventful effects of hyperac-etatemia observed previously with intermittent hemodialysis, are unknown inCRRT. In 84 critically ill patients with acute renal failure treated with CVVH,lactate was associated with a significant increase in serum bicarbonate and pHafter 48 h when compared to acetate [13].

Supraphysiologic sodium concentration in the dialysate is a well-knowntool to improve hemodynamic stability during intermittent hemodialysis; it isunclear if enhancing sodium concentration in CRRT fluids could also beof benefit. Since most CRRT solutions have no or low potassium concentra-tions and do not contain phosphate, potassium and/or phosphate addition or


supplementation will frequently be required after the initial phase [10]. In ourcenter, we usually add potassium chloride and potassium phosphate to reach aconcentration of phosphate of 1 mmol/l and a concentration of potassium of4 mmol/l in the solutions (dialysate and replacement). These additions are madeonly when serum phosphate is below 1.5 mmol/l and serum potassium below5 mmol/l (table 2). We never encountered any problem with salt formation orcrystallization even in bicarbonate solutions. We recently reported our resultsbut they are unpublished at this time.

The presence or absence of dextrose or glucose in CRRT fluids can beappraised from different perspectives. First, one may consider that loosing glucoseand calories through the effluent is not desirable and that a physiologic glucoseconcentration (near 5 mmol/l) should be offered in the dialysate and replacement.On the other hand, it is certainly not desirable to use solutions with supraphys-iologic glucose or dextrose content since this may induce hyperglycemia. Indeed,it has been recently shown that a tight glycemic control is associated with a betteroutcome in critical illness [14]. Also, dextrose-containing solutions may enhancelactate production and may be detrimental in presence of brain injury [15].

We may see in the near future more sophisticated CRRT solutions, forexample containing nutrients or anti-oxidants. More research is definitelyrequired in that area. Whatever solution is used, electrolytes and acid-base

Table 2. Our local protocol for the addition of potassium and phosphate to HemosolLG2 and Hemosol B0

If serum phosphate is below 1.5 mmol/lAdd 2 ml of potassium phosphate (racemic mixture of KH2PO4/K2HPO4) to a 5-liter bagof Hemosol LG2 or Hemosol B0; this is a fixed amount and maximum per bag andcorresponds to a phosphate concentration of 1.2 mmol/l in the solution

If potassium phosphate has been added and serum potassium is below 5 mmol/lAdd 2 mmol of potassium chloride (KCl) per 5-liter bag of Hemosol LG2 or 12 mmolof KCl per 5-liter bag of Hemosol B0; this is a fixed amount and maximum per bag andcorresponds to a potassium concentration of 4.2 mmol/l in the solution

If potassium phosphate has not been added and serum potassium is below 5 mmol/lAdd 10 mmol of potassium chloride (KCl) per 5-liter bag of Hemosol LG2 or 20 mmol ofKCl per 5 -liter bag of Hemosol B0; this is a fixed amount and maximum per bag and cor-responds to a potassium concentration of 4.0 mmol/l in the solution

If potassium phosphate has not been added and serum potassium is above 5 mmol/lDo not add any potassium chloride (KCl) to Hemosol LG2; if potassium is below7 mmol/l, add only 10 mmol of KCl per 5-liter bag of Hemosol B0 (corresponding to apotassium concentration of 2 mmol/l in the solution); if serum potassium is above7 mmol/l, do not add any KCl to Hemosol B0

Leblanc 226

balance should be assessed frequently during CRRT, especially if high-volumeexchanges are performed.

Influence of CRRT Prescription

Diffusion, the main exchange mechanism in CVVHD, is excellent to removesmall molecules but is much more limited for middle-sized solutes. The cationsand anions involved in the strong ion difference are available to be exchangedby diffusive fluxes.

Convection is the exchange mechanism used during CVVH. Most ionsconsidered in the strong ion difference equation pass easily through the membranewith convective fluxes and have sieving coefficients near one. Thus, diffusionand convection are efficient processes to restore electrolytes and acid-basebalance, and can be considered equivalent from that perspective.

However, during convection, large amounts of bicarbonate can be lost inthe ultrafiltrate, inducing rapidly a metabolic acidosis if not replaced appropri-ately by the substitution fluid. For example, during CVVH at an ultrafiltrationrate of 3 liters/h, as much as 50 mmol of HCO3

� can be lost hourly. The amountof bicarbonate lost in the effluent can be easily assessed since both the sievingcoefficient and effluent rate are known. If the replacement consists in normalsaline, a hyperchloremic acidosis would rapidly occur. Therefore, substitutionfluids must be relatively similar to normal blood plasma, and must be alkalinizing.Nevertheless, since the composition of replacement fluids can be flexible, thealkalinizing performance of convective modalities may be superior and theseapproaches may be more versatile. This can be demonstrated by acetate-freebiofiltration that corresponds to a hemodiafiltration in which bicarbonate isabsent from the dialysate, bicarbonate being administered exclusively via replace-ment. This method offers the possibility to titrate more precisely the amount ofgiven buffer and has been used to treat severe metabolic acidosis.

Continuous renal replacement therapies provide a constant restoration ofelectrolytes and acid-base balance, avoiding large fluctuations. However, sinceflow rates during CRRT are lower than during intermittent modalities, acid-basebalance restoration is progressive and usually occurs over 36–48 h in most cases.As well, at usual CRRT rates in average-size patients, a metabolic steady stateis expected to occur after 3 or 4 days; this may apply to several electrolytes [16].

When applying higher flows, for example to reach an ultrafiltration rate of35 ml/kg/h, as recommended by Ronco et al. [17], acid-base correction andmetabolic steady state may both occur sooner. In high-volume CVVH, becauseof the potential massive influx of lactate anions, bicarbonate fluids are recom-mended. On the other hand, a metabolic alkalosis may be induced by high-volume


CVVH if using replacement fluids containing too much bicarbonate, and itsconcentration may have to be reduced under certain circumstances.

References

1 Kellum JA: Metabolic acidosis in the critically ill: Lessons from physical chemistry. Kidney Int1998;53(suppl 66):S81–S86.

2 Leblanc M, Kellum JA: Biochemical and biophysical principles of hydrogen ion regulation; inRonco C, Bellomo R (eds): Critical Care Nephrology, ed 1. Dordrecht, Kluwer Academic Publishers,1998, pp 261–277.

3 Warnock DG: Uremic acidosis. Kidney Int 1988;34:278–287.4 Wallia R, Greenberg A, Piraino B, Mitro R, Puschett JB: Serum electrolytes in end-stage renal dis-

ease. Am J Kid Dis 1986;8:98–104.5 Macias WL, Clark WR: Acid-base balance in continuous renal replacement therapy. Semin Dialy

1996;9:145–151.6 Thomas AN, Guy JM, Kishen R, Geraghty IF, Bowles BJ, Vadgama P: Comparison of lactate and

bicarbonate buffered haemofiltration fluids: Use in critically ill patients. Nephrol Dial Transplant1997;12:1212–1217.

7 Heering P, Ivens K, Thumer OM, Grabensee B: Acid-base balance and substitution fluid duringcontinuous hemofiltration. Kidney Int 1999;56(suppl 72):S37–S40.

8 Zimmerman D, Cotman P, Ting R, Karanicolas S, Tobe SW: Continuous veno-venous haemodialysiswith a novel bicarbonate dialysis solution: Prospective cross-over comparison with a lactate bufferedsolution. Nephrol Dial Transplant 1999;14:2387–2391.

9 Schetz M, Leblanc M, Murray P: The Acute Dialysis Quality Initiative. VII. Fluid composition andmanagement in CRRT. Adv Renal Replacement Ther 2002;9:282–289.

10 Veech RL: The untoward effects of the anions of dialysis fluid. Kidney Int 1988;34:587–597.11 Palevsky P: Continuous renal replacement therapy component selection: Replacement fluid and

dialysate solutions. Semin Dial 1996;9:107–111.12 Palsson R, Niles JL: Regional citrate anticoagulation in continuous venovenous hemofiltration in

critically ill patients with a high risk of bleeding. Kidney Int 1999;55:1991–1997.13 Morgera S, Heering P, Szentandrasi T, Manassa E, Heintzen M, Willers R, Passlick-Deetjen J,

Grabensee B: Comparison of a lactate- versus acetate-based hemofiltration replacement fluid inpatients with acute renal failure. Renal Failure 1997;19:155–164.

14 van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, Vlasselaers D,Ferdinande P, Lauwers P, Bouillon R: Intensive insulin therapy in the critically ill patients. N EnglJ Med 2001;345:1359–1367.

15 Sieber FE, Traystman RJ: Special issues: Glucose and the brain. Crit Care Med 1992;20:104–114.16 Clark WR, Mueller BA, Kraus MA, Macias WL: Extracorporeal therapy requirements for patients

with acute renal failure. J Am Soc Nephrol 1997;8:804–812.17 Ronco C, Bellomo R, Homel P, Brendolan A, Dan M, Piccinni P, La Greca G: Effects of different

doses in continuous veno-venous haemofiltration on outcomes of acute renal failure: A prospectiverandomised trial. Lancet 2000;356:26–30.

Martine Leblanc MDNephrology and Critical Care, Maisonneuve-Rosemont Hospital5415 de l’Assomption, Montreal, PQ (Canada)Tel. �1 514 252 3400, ext 3733, Fax �1 514 255 3026E-Mail [email protected]


Anticoagulation for Continuous Renal Replacement Therapy

Andrew Davenport

University College and Royal Free Hospital Medical School, andCenter for Nephrology, The Royal Free Hospital, London, UK

Over the last three decades there have been great strides in the developmentof continuous forms of renal replacement therapy (CRRT), from a techniqueproviding ultrafiltration aiding fluid balance, to an effective treatment for con-trolling solute, fluid, electrolyte and acid-base balance in the critically ill patientwith multi-organ failure. More recently, it has been reported that the intensityof CRRT might improve the outcome of patients with acute renal failure incritical illness [1]. However, for CRRT to provide efficient effective treatmentsuperior to that achieved with standard daily intermittent hemodialysis, theCRRT circuit has to operate continuously. Although there are many earlierpublications reporting circuit lives in excess of 48 h, more recent studies havesuggested an average circuit life of 16–18.6 h/day [2, 3]. In clinical practice pre-mature circuit clotting leads to wasted time, as one circuit has to be dismantled,and a new one set up, primed and then connected to the patient. Thus the patientdoes not receive CRRT but discontinuous treatment, due to this so called circuit‘down time’, which can vary from 3 to 8 h/day [2, 3].

In the intensive care unit, most patients with acute renal failure have asso-ciated sepsis with an inflammatory state resulting in activation of leukocytes,macrophages, platelets and coagulation protein pathways [4]. Passage throughthe extracorporeal circuit results in the formation of leukocyte-platelet aggre-gates and platelet micro-thrombi, resulting in premature thrombosis, whichtypically occurs on the dialyzer/hemofilter membrane, the venous air detectorchamber and venous access catheter [5]. Ideally, treatments designed to reduceinflammatory cell activation and leukocyte-platelet interactions would help tomaintain the integrity of the CRRT circuit. Some centers advocate the use offresh frozen plasma, which may have a role in replacing complement inhibitors

Anticoagulation for CRRT 229

and capsaises, and may have some analogous effect to the use of immunoglob-ulin in controlling acute vascular rejection in T and B cell positive cross-matchallografting [6]. Whereas in current clinical practice anticoagulants targetedagainst thrombin generation are most commonly used.

Systemic Anticoagulants

Unfractionated HeparinUnfractionated heparin remains the most commonly used extracorporeal

systemic anticoagulant for CRRT, with an action time of 3–5 min. Paradoxically,rather than reducing platelet activation, heparin activates platelets by complex-ing with platelet factor 4. In addition the effect of heparin is often decreasedin the intensive care unit patient, as key co-factors such as antithrombin andheparin cofactor II levels are often reduced.

Heparin is not removed by dialysis or hemofiltration, but is eliminated ina dose dependent manner, mainly by the liver, but also by the kidney. Thus, thehalf-life is increased in renal failure to 40–120 min. As heparin is a highly neg-atively charged molecule, it can be adsorbed to the plastic of the circuit tubingduring priming, and priming doses vary from center to center, from 1,000 to10,000 IU. Provided there is no risk of bleeding, a loading dose of heparin(10–20 U/kg) is given at the start of CRRT, followed by a maintenance infusionof 3–20 U/kg/h [4].

It is important that the effect of heparin is monitored, as individualpatients differ in their response. The whole blood clotting time (WBCT) andthe activated coagulation time (ACT), are the most common bedside tests ofheparin anticoagulation. Fresh unanticoagulated whole blood samples shouldbe rapidly placed into a glass tube at 37�C. The ACT is similar, but contains anactivator of the intrinsic coagulation system. Both are prone to error, andrequire regular quality control [7]. In addition, the results are dependent uponthe level of coagulation factors, platelets and hematocrit. Thus, the results ofWBCT and ACT taken from the same patient will differ during hemofiltrationwith post-dilutional fluid replacement if taken before and after filter, simplydue to ultrafiltration increasing the hematocrit and platelet concentration caus-ing shorter post-filter times. The activated partial thromboplastin time (aPTT)is a laboratory test on plasma separated from citrated blood, and should bemeasured in conjunction with a prothrombin time, which although littleaffected by heparin, provides valuable information about the levels of coagu-lation factors. Centers differ not only in which monitoring tests are performedand their frequency, but also the site at which samples are taken [7, 8].Sampling immediately prior to the arterial port of the hemofilter/hemodialyzer

Davenport 230

target whole blood activated clotting time (ACT) of around 140–180 s, or alaboratory aPTT of 100–140 s.

Filter patency has not been proven to be determined either by the totalheparin dose, or by aPTT or other clotting studies [7]. Although van der Weteringet al. [9] reported a reduced filter patency rate with systemic aPTT times lessthan 35 s. They also reported that with a systemic aPTT of 15–35 s theincidence of de novo patient hemorrhage was 2.9 per 1,000 h CRRT, whichincreased to 7.4 at an aPTT of 45–55 s [9]. Even using low-dose heparin(500 U/h) for CRRT in patients at risk of hemorrhage has not been proven toreduce the risk of bleeding complications [10].

The main complication of heparin is hemorrhage. Fortunately, the half-lifeis relatively short, and heparin activity can be quickly reversed with protamine,1 mg given for every 1,000 U heparin.

Rarely heparin administration can result in an acute allergic reaction,usually due to pork sensitivity. Occasionally patients can develop immune-mediated heparin-induced thrombocytopenia (HIT) [11]. This is more likelyfollowing cardiac or vascular surgery, when the patient has been exposed tolarge doses of heparin, and often presents with repeated CRRT circuit clotting[11]. The standard laboratory method for detecting heparin dependent anti-bodies uses a platelet aggregation assay, utilising platelets from normal healthydonors, the patient’s plasma and the same heparin preparation, as administeredto the patient. This screening test may be negative in up to 50% of cases, and ifHIT is clinically suspected, then a more sensitive ELISA test using PF4 com-plexed with heparin should be performed.

Low-Molecular-Weight HeparinsLow-molecular-weight heparins (LMWHs) are glycosaminoglycans with a

molecular weight of around 5 kD and cannot bind to both antithrombin andthrombin simultaneously, thus losing antithrombin activity compared to stan-dard heparin. As inactivation of factor Xa does not require direct heparin bind-ing, LMWH by activating antithrombin, retains anti-Xa activity.

The currently available LMWHs, dalteparin, enoxaparin, nadroparin,reviparin, and tinzaparin differ in size, half-life and biological activity, but allhave a half-life much greater than for unfractionated heparin, with enoxaparinhaving the longest at 27.7 h [4]. Compared to standard heparin, LMWHs havebeen reported to be more effective in reducing fibrin deposition on dialyzermembranes, and extracorporeal clotting, but also less hemorrhagic complica-tions [4]. This appears to be due to both a quicker onset of action that standardheparin, and also less leukocyte and platelet activation [12].

When LMWHs were first used in CRRT most centers either used aloading dose followed by a continuous infusion, or gave further bolus doses


6 hourly [13]. Although LMWHs are relatively small molecules, they are notsignificantly cleared during CRRT [7]. More recently, LMWHs have beenshown to be effective when started as a continuous infusion without a loadingdose (e.g. dalteparin 600 IU/h, tinzaparin 400–800 IU/h ) achieving a meananti-factor Xa activity of 0.49 IU/ml (therapeutic range for systemic anti-coagulation 0.3–0.8 IU/ml) within 1 h of starting CRRT [14]. However, moststudies have shown that when low doses of LMWH are used for CRRT,although the risk of hemorrhage has been reduced, filter life and/or patencyhas similarly been reduced, whereas if higher doses are used the risk of hem-orrhage is increased [7]. Thus, LMWH when used either as a fixed or variabledose regime for CRRT does not offer any benefit over standard unfractionatedheparin [6, 15].

LMWH by activating antithrombin, retains anti-Xa activity, and specialassays are required to determine the inhibition of factor Xa. Commercial kitsfor testing factor Xa activity must not contain exogenous antithrombin, so thatthe LMWH dose can be titrated and correlated against anticoagulant activity.The recommended anti-factor Xa activity range for CRRT is 0.2–0.4 U/ml, andthis has been shown to allow successful treatment of critically ill patients at riskof hemorrhage [15].

If bleeding does occur, then this may be more severe than standard heparin,due to the prolonged half-life. Protamine may only have a moderate effect, andwill depend upon the individual LMWH used, and in severe cases fresh frozenplasma and/or recombinant factor VII may be required. HIT may rarely developwith LMWHs [11], and if so then LMWHs must be discontinued.

DanaparoidDanaparoid, is a mixture of glycosaminoglycans: 84% heparan sulphate,

12% dermatan sulphate and 4% chondroitin sulphate, derived from porcineintestinal mucosa. Danaparoid exerts its anticoagulant effect predominantly byactivating antithrombin, primarily against FXa but also against thrombin. Asdanaparoid has a minimal effect on platelets, it has been successfully used inthe management of patients with heparin-induced thrombocytopenia, althoughthere is a potential cross reactivity of �5% [11]. Prior to starting danaparoidin cases of HIT, laboratory testing should be undertaken to exclude cross-reactivity, as very occasionally in vivo cross-reactivity has been reported [11].

An initial bolus dose is required for CRRT, 2,500 U (35 U/kg) of dana-paroid, if there is no hemorrhagic risk, followed by an initial infusion of400 U/h, which is then adjusted (usually between 200 and 400 U/h) to thedesired anti-Xa (0.4–0.6) [11]. Although others have reported that lower dosesof danaparoid, initial bolus of 750 U followed by a mean infusion of 138 U/h,have been effective maintaining an anti-Xa activity of 0.2–0.6 [16].

Davenport 232

The major problem with danaparoid is its prolonged half-life. If bleedingoccurs, there is no simple antidote and patients may require supportive treat-ment with activated factor VII concentrate or fresh frozen plasma [4].

Direct Thrombin Inhibitors

Recombinant HirudinRecombinant hirudin (lepirudin and desirudin) binds to both the active and

exosite of thrombin, and is an irreversible thrombin antagonist licensed fortreating patients with HIT. Hirudin is excreted renally and the plasma half-lifeis increased in renal failure from 1–2 to 36–75 h [17]. Thus, the dose has to bemodified for CRRT with a bolus of 0.01 mg/kg followed by an infusion of0.005 mg/kg/h for continuous hemofiltration [18]. Hirudin therapy is associatedwith an increased risk of hemorrhage, and there is no simple antidote. The dos-ing of hirudin is made more complex, particularly during CRRT, with greaterclearances achieved with polysulphone and polyamide high flux membranescompared to polyacrylonitrile [19]. In addition, a varying number of patientstreated continuously with hirudin develop anti-hirudin antibodies, which donot effect the activity of hirudin, but increase the plasma half-life by reducingelimination [11].

Thus there has to be careful monitoring of r-hirudin therapy. Unfortunatelythe relationship between plasma hirudin concentration and the activated partialthromboplastin time (aPTT) is not linear [11]. Thus maintaining a aPTT of1.5–2.0 does not necessarily reduce the risk of hemorrhage. Other more spe-cific assays of r-hirudin have been advocated, including a chromogenic assay todetermine plasma concentrations, and directly measuring thrombin activitywith a viper venom clotting test – the ecorin clotting time [11], but are notroutinely available. In cases of hirudin overdosage and/or hemorrhage, highvolume hemofiltration can be used to reduce the plasma hirudin levels.

ArgatrobanArgatroban is derived from L-arginine, and reversibly binds to the active

enzymatic site of thrombin. Unlike r-hirudin, argatroban is metabolised hepat-ically and excreted in the bile, with a half-life of 40–50 min. In addition, theaPTT has been shown to have a good correlation with argatroban plasma levels[20]. In the intensive care setting argatroban has become the anticoagulant ofchoice for the treatment of patients with HIT.

As argatroban is a reversible inhibitor with a relatively short half-life, sohas to be administered by constant infusion, starting at a dose of 2 �g/kg/min(reduced to 0.5 �g/kg/min for those with liver disease), and titrating the dose to


achieve an aPTT ratio of 2–2.5. In clinical practice, an average dose of0.9 �g/kg/min has been reported for CRRT [21].

Regional Anticoagulants

CitrateAlthough regional citrate anticoagulation has been used for more than

20 years, it is only in the last few years that there has been an increasing usageof citrate for CRRT. Anticoagulation with citrate induces a degree of complex-ity, as if trisodium citrate is used, then a specialised calcium free, low or zeromagnesium, and reduced or bicarbonate free dialysate is required. Calcium thenhas to be infused centrally to restore the plasma ionised calcium concentration[4]. Citrate, by complexing calcium and reducing calcium concentration, notonly prevents activation of the coagulation cascades but also platelets duringpassage through the dialyzer, but not complement or leukocyte activation. Thus,membrane fouling and deposition of fibrin and platelets is much reduced whencitrate is used compared to LMWH and standard heparin. The half-life of thecalcium-citrate complex is minutes, and thus citrate is a regional anticoagulant,allowing the successful dialysis of patients at risk of bleeding [22].

Bedside monitoring is possible by using WBACT (200–250 s), howevermost centers adjust the citrate infusion according the post dialyzer calciumconcentration (target 0.25–0.35 mmol/l) [25]. As initially there were no com-mercially available calcium-free fluids for dialysate or replacement solutions,individual centers developed different protocols. The rate of the citrate infusionis dependent upon the blood flow, and the modality of CRRT, hemofiltrationand/or dialysis. Thus during hemofiltration with a blood flow of 180 ml/minthe citrate infusion was started at 13.6 mmol/h [23], whereas with continu-ous hemodialysis with a blood flow of 137 ml/min, the citrate infusion was17.5 mmol/l [24], and for hemodiafiltration at 125 ml/min, citrate was started at25 mmol/h [25].

Thereafter, the rate of citrate infusion is adjusted according to the calciumconcentration, being reduced if the calcium is less than 0.25 mmol/l, and corre-spondingly increased when greater than 0.36 mmol/l [25]. More recently, citratehas been used as acid citrate dextrose and centers have successfully usedcommercial solutions containing both calcium and bicarbonate [26].

Citrate dialysis has been reported to result in citrate intoxication, eitherwhen citrate is not metabolised rapidly, if there is hepatic failure or musclehypoperfusion, or during isolated ultrafiltration, when citrate is not beingdialysed out. Failure to adequately metabolise citrate results in an increasedtotal calcium to ionized calcium ratio (�2.5), due to the accumulation of the

Davenport 234

calcium-citrate complex. Not surprisingly this has been most commonlyreported in patients with liver failure [27], and should be managed by decreas-ing the citrate infusion [25].

During CRRT hypernatremia can occur due to the sodium load if trisodiumcitrate is used, and thus many centers have developed specialised hyponatremicdialysates [22]. Similarly, as each citrate molecule is metabolised through tothree bicarbonates, patients are at risk of developing a metabolic alkalosis.To compensate many centers use specialised dialysates and/or replacementsolutions with a high chloride load [22]. Many patients treated by CRRT inone study developed metabolic alkalosis and hypomagesemia after 7 days whenusing commercial bicarbonate based fluids [26]. To overcome some of theseproblems, some centers have used citrate dextrose-A rather than 4% trisodiumcitrate, or reduced the citrate concentration, in combination with a reducedbicarbonate dialysate concentration to 25 mmol/l [26]. Another approach whenusing a twin bag bicarbonate system was not to add the bicarbonate compart-ment when patients developed alkalosis.

Despite the apparent complexity of using citrate, and the potential array ofmetabolic disturbances, citrate use is growing in popularity for CRRT, as it is ahighly effective anticoagulant, with much longer average circuit lives of 29–82 hcompared to standard heparin [22–26]. Indeed as citrate is such an efficientregional anticoagulant, access problems are a more common cause of circuit clot-ting, especially if the calcium infusion is returned into the venous dialysis line.

ProstanoidsProstacyclin (PGI2) is a natural anticoagulant produced by endothelial cells,

by the breakdown of arachidonic acid. PGI2, and its analogue epoprostenol, arepotent antiplatelet agents blocking cAMP, and have been shown to reduceplatelet micro thrombi during hemodialysis and CRRT compared to standardheparin and LMWH [4]. Although both agents are potent arterial vasodilators,most patients do not develop symptomatic hypotension at the doses used (PGI2

5 ng/kg/min, range 2.5–10 ng/kg/min), as some 40% of the dose is lost duringpassage through the dialyzer [28]. As the half-life is in minutes, any hypoten-sive episode can readily be reversed by stopping the infusion.

Other prostanoids, such as PGE1 (alprostadil), PGE2 and PGD also haveantiplatelet effects and can be used as extracorporeal anticoagulants. As PGE1

is metabolised in the lung it has less systemic vasodilatory properties comparedto prostacyclin, and does not cause hypotension. These prostanoids are not aspotent as PGI2, and thus the dose of alprostadil required is 5–20 ng/kg/min.

PGI2 does not have any direct effect on the plasma coagulation pathways,so its anticoagulant activity can not be readily assessed, even by thrombo-elastography [4]. Thrombin generation does occur during dialysis with PGI2,


and some authors have advocated a combination of reduced doses of bothheparin and PGI2 [29]. Some groups have found PGI2 and prostanoids to beequally effective as heparin in maintaining CRRT circuit life, with better filterpatency [30].

When used as the sole extracorporeal anticoagulant, PGI2 has been shownto significantly reduce the incidence of haemorrhage in patients at risk of bleed-ing [30], as PGI2 and its analogues are essentially regional rather than systemicanticoagulants. However, PGI2 and epoprostenol are some 25� more expensivethan standard heparin.

Nafamostat MaleateNafamostat maleate is a serine protease inhibitor, which inhibits broad

enzymatic systems including coagulation cascades, platelet and complementactivation, kinin and fibrinolysis cascades. Nafamostat mesilate has a shorthalf-life of 5–8 min, and some 40% is cleared during passage through the dia-lyzer [31]. By inhibiting thrombin, factor Xa and XIIa, nafamostat prolongs theWBCT, ACT and aPTT, thus allowing bedside monitoring. Most experiencecomes from Japan, where the circuit is primed with 20 mg of nafamostat in1.0 l normal saline, and then nafamatostat initially infused at 40 mg/h duringCRRT, to maintain a target aPTT ratio of 2.0 � normal [32]. As nafamostat isa small molecule which is partially cleared by the extracorporeal circuit, thedose required for CRRT using hemodiafiltration is greater than that forhemofiltration (0.3 vs. 0.1 mg/kg/h) [4]. Nafamostat contains a cationic portionwhich binds to negatively charged polyacrylonitrile membranes, and also tosome extent to polymethylmethacrylate membranes, if these membranes areused then the nafamostat dose has to be increased.

Circuit Design for CRRT

The key with any extracorporeal circuit design is to minimise activation ofleukocytes and platelets, as platelets preferentially move to the periphery of theflow, so contact the wall of the plastic tubing and dialyzer fibers. Clotting startsin the outer fibers of the dialyzer, where blood flow is slower.

Membrane design, in terms of composition, surface and geometry hasbeen shown to affect thrombus formation [33]. Similarly, a dialyzer with a largesurface area will be more likely to result in thrombus deposition, than a moreefficient, or porous membrane of smaller surface area. If ultrafiltration rates arehigh then this may additionally result in increased protein deposition or foulingof the hemofilter membrane and the combination with hemoconcentrationincrease the risk of filter clotting. Thus, administration of the hemofiltration

Davenport 236

substitution fluid pre-filter minimises hemoconcentration within the fiber bun-dles and prolongs circuit life [34].

Mechanical activation due to turbulent blood flow occurs both with theaccess catheter and the blood pump. Thus venous access catheter insertion site,catheter design, composition and coating all affect the risk of thrombosis [4],for example heparin coated venous access catheters have been reported to causeless circuit clotting [35]. Surprisingly, the change from spontaneous to pumpedCRRT circuits was associated with a reduction in circuit lives [8], due to theintroduction of faster blood flows, which generated wide pressure swings [4].A redesign of the blood pump, to improve laminar flow may well help to reducethe propensity to clot the circuit.

Conclusions

When CRRT is truly continuous, it provides very effective solute control,with acid-base balance, electrolyte and water balance. However, the Achilles’heel of CRRT is circuit clotting and down time. The advent of pumped CRRTwith increasing ultrafiltration volumes to achieve greater treatment efficiency,has paradoxically led to shortened circuit lives with standard heparin regimes.Increased systemic heparinisation is associated with an increased risk of hemor-rhage. This has led to an increased interest in regional anticoagulants such as cit-rate. When citrate was first introduced into CRRT, it was shown to be a highlyeffective regional anticoagulant with a prolonged circuit life, but required an on-site sterile pharmacy to produce specialised dialysate and replacement fluids.Today, units are now experimenting with dialysates and fluid replacement usingcommercially available bicarbonate twin bag systems.

References

1 Ronco C, Bellomo R, Hamel P, Brendolan A, Dan M, Piccinni P, La Greca G: Effects of differentdoses in continuous veno-venous hemofiltration on outcomes of acute renal failure. Lancet 2000;356:26–30.

2 Mehta RL, McDonald B, Gabbai FB, Aphl M, Pascual MTA, Karkas A, Kaplan RM: A ran-domised controlled trial of continuous versus intermittent dialysis for acute renal failure. Kid Int2001;60:1154–1163.

3 Uchino S, Fealy N, Baldwin I, Morimatsu H, Bellomo R: Continuous is not continuous: The inci-dence and impact of ‘circuit down time’ on uremic control during continuous veno-venoushemofiltration. Intens Care Med 2003;29:575–578.

4 Davenport A: Anticoagulation in patients with acute renal failure treated with continuous renalreplacement therapies. Home Hemodial Int 1998;2:41–60.

5 Holt AW, Bierer P, Berstein AD, Bury LJ, Vedig AE: Continuous renal replacement therapy in crit-ically ill patients: Monitoring circuit function. Anaesth Intens Care 1996;24:423–424.


6 Shulman RI, Singer M, Rock J: Keeping the circuit open: Lessons from the lab. Blood Purif 2002;20:275–281.

7 Davenport A: Problems with anticoagulation; in Lameire N, Mehta R (eds): Complications ofDialysis: Recognition and Management. Boston, Marcel Dekker, 1999, pp 215–240.

8 Favre H, Martin Y, Stoermann C: Anticoagulation in continuous extracorporeal renal replacementtherapy. Semin Dial 1996;9:112–118.

9 van der Wetering J, Westendorp RGJ, van der Hoeven JG, Stolk B, Feuth JDM, Chang PC: Heparinuse in continuous renal replacement therapies: The struggle between filter coagulation and patienthemorrhage. J Am Soc Nephrol 1996;7:145–150.

10 Bellomo R, Teede H, Boyce N: Anticoagulant regimens in acute continuous hemodiafiltration:A comparative study. Intens Care Med 1993;19:329–332.

11 Davenport A: Heparin induced thrombocytopenia during renal replacement therapy. Hemodial Int2004;in press.

12 Leitienne P, Fouque D, Rigal D, Adeleine P, Trzeciak MC, Laville M: Heparins and bloodpolymorphonuclear stimulation in hemodialysis: An expansion of the biocompatibility concept.Nephrol Dial Transplant 2000;15:1631–1637.

13 Wynckel A, Bernieh B, Toupance O, N’Guyen PH, Wong T, Lavaud S, Chanard J: Guidelines tothe use of enoxparin in slow continuous dialysis. Contrib Nephrol. Basel, Karger, 1991, vol 93,pp 221–224.

14 Schepens D, De Keulenaer B: Efficacy and safety of nadropine as anticoagulant therapy in con-tinuous venovenous hemofiltration. Blood Purif 2002;20:314.

15 Reeves JH, Cumming AR, Gallagher L, O’Brien JL, Santamaria JD: A controlled trial of lowmolecular weight heparin (dalteparin) versus unfractionated heparin as anticoagulant during con-tinuous venovenous hemodialysis with filtration. Crit Care Med 1999;27:2224–2228.

16 Lindhoff-Last E, Betz C, Bauersachs R: Use of a low-molecular-weight heparinoid (danaparoidsodium) for continuous renal replacement therapy in intensive care patients. Clin Appl ThrombHemost 2001;7:300–304.

17 Fischer KG: Hirudin in renal insufficiency. Semin Thromb Hemostat 2002;28:467–482.18 Schneider T, Heuer B, Deller A, Boesken WH: Continuous haemofiltration with r-hirudin

(lepirudin) as anticoagulant in a patient with heparin induced thrombocytopenia (HIT II). WienKlin Wochenschr 2000;112:552–555.

19 Frank RD, Farber H, Lanzmich R, Floege J, Kierdorf HP: In vitro studies on hirudin eliminationby haemofiltration: Comparison of three high flux membranes. Nephrol Dial Transplant 2002;17:1957–1963.

20 Hursting MJ, Alford KL, Becker JC, Brooks RL, Joffrion JL, Knappenberger GD, Kogan PW,Kogan TP, McKinney AA, Schwarz RP Jr: Novostatin (brand of argatroban): A small molecule,direct thrombin inhibitor. Semin Thromb Hemostat 1997;23:503–516.

21 Reddy BV, Nahlik L, Trevino S, Murray PT: Argatroban anticoagulation during renal replacementtherapy. Blood Purif 2002;20:313–314.

22 Mehta RL, McDonald BR, Aguilar MM, Ward DM: Regional citrate anticoagulation for continu-ous arteriovenous hemodialysis in critically ill patients. Kid Int 1990;38:976–981.

23 Palsson R, Niles JR: Regional citrate anticoagulation in continuous venovenous hemofiltration incritically ill patients with a high risk of bleeding. Kid Int 1999;55:1991–1997.

24 Tolwani AJ, Campbell RC, Schenk MB, Allon M, Warnock DG: Simplified citrate anticoagulationfor continuous renal replacement therapy. Kid Int 2001;60:370–374.

25 Kutsogiannis DJ, Mayers I, Chi WD, Gibney RT: Regional citrate anticoagulation in continuousvenovenous hemodiafiltration. Am J Kid Dis 2000;35:802–811.

26 Bunchman T, Maxvold NJ, Barnett J, Hutchings A, Benfield MR: Pediatric hemofiltration:Normocarb dialysate solution with citrate anticoagulation. Pediatr Nephrol 2002;17:150–154.

27 Meier-Kriesche H-U, Gitomer J, Finkel K, DuBose T: Increased total to ionized calcium ratioduring continuous venovenous hemodialysis with regional citrate anticoagulation. Crit Care Med2001;29:748–752.

28 Zobel G, Ring E, Kuttnig M, Grubbauer HM: Continuous arteriovenous hemofiltration versuscontinuous venovenous hemofiltration in critically ill pediatric patients. Contrib Nephrol. Basel,Karger, 1991, vol 93, pp 257–260.

Davenport 238

29 Langenecker SA, Felfernig M, Werba A, Mueller CM, Chiari A, Zimpfer M: Anticoagulation withprostacyclin and heparin during continuous venovenous hemofiltration. Crit Care Med 1994;22:1774–1781.

30 Davenport A, Will EJ, Davison AM: Comparison of the use of standard heparin and prostacyclinanticoagulation in spontaneous and pump driven extracorporeal circuits in patients with combinedacute renal and hepatic failure. Nephron 1994;66:431–437.

31 Akizawa T: Beneficial characteristics of protease inhibitor as an anticoagulant for extracorporealcirculation. Rinsho Ketsueki 1990;31:782–786.

32 Matsuo T, Kario K, Nakao K, Yamada T, Matsuo M: Anticoagulation with nafamostat mesilate, asynthetic protease inhibitor, in hemodialysis patients with a bleeding risk. Haemostatasis 1993;23:135–141.

33 Sperschneider H, Deppisch R, Beck W, Wolf H, Stein G: Impact of membrane choice and bloodflow pattern on coagulation and heparin requirement potential consequences on lipid concentra-tions. Nephrol Dial Transplant 1997;12:2638–2646.

34 Davenport A: Pre vs. postdilution for continuous veno-venous hemofiltration? That is the ques-tion. Nephron Clin Pract. 2003;94:c83–c84.

35 Davenport A: Central Venous Catheters for Hemodialysis: How to overcome the problems. HomeHemodial Int 2000;2:43–45.

Dr. A. DavenportCenter for Nephrology, The Royal Free HospitalPond Street, London NW4 2QG (UK)Tel. �44 20 783 022 91, Fax �44 20 783 021 25, E-Mail [email protected]


Peritoneal Dialysis in Acute RenalFailure of Adults:The Under-UtilizedModality

S.R. Ash

Greater Lafayette Health Systems, Arnett Clinic, and HemoCleanse, Inc., Lafayette; Adjunct Associate Professor, Purdue University, West Lafayette, Ind., USA

Among various dialysis therapies for acute renal failure in the adult, today’snephrologists think of CVVHD or intermittent hemodialysis long before consid-ering peritoneal dialysis, making acute peritoneal dialysis a considerably under-used therapy [1]. This is in spite of the numerous problems encountered with theextracorporeal therapies that are eliminated with use of the peritoneal membrane.In fact, the peritoneum can be considered the ‘perfect’ membrane for continuousarteriovenous hemofiltration and dialysis (CAVHD) providing: (a) natural mem-branes located within the body; (b) permeability to uremic toxins and limited pas-sage of albumin and tightly-bound toxins; (c) limited passage of antibodies andthat cause kidney failure in some patients; (d) infallible blood access with bloodflow rate of about 200 ml/min; (e) controllable ultrafiltration rate; (f) biocompat-ibility of blood pathways, obviating need for anticoagulants; (g) impermeabilityto bacteria in dialysate, preventing septicemia after dialysate contamination;(h) permeability to white cells into dialysate if there is bacterial contamination,to limit proliferation and provide a visible sign of the contamination; (i) passageof effluent blood from the membranes directly to the liver, allowing metabolicconversion of lactate, glucose or various nutrients, and (j) ease of use, allowingcontinuous 24-hour dialysis by merely intermittently infusing and drainingmodest volumes of sterile dialysate through a permanent access.

As with all dialysis procedures, peritoneal dialysis (PD) was first used intherapy of acute renal failure (ARF) [2]. Now CAPD or cycler therapy supportsabout 12% of patients with ESRD in the United States [3]. The success ofCAPD in support of patients with ESRD has reminded physicians that PD can

Which Treatment for ARF in ICU?

Ash 240

also be used for treatment of ARF in adults. In some countries such as Japan,PD was a more common choice for treatment of ARF in adults than CAVH in1987 [4]. In ARF, PD is appropriate for the same types of patients for whichCAVH is chosen: those with heart failure and low cardiac index who cannottolerate the rapid fluid removal rate of standard hemodialysis (HD) [4]. PD isstill the mainstay for treatment of ARF in infants and children [5, 6] and has arisk of complications in the range or lower than CAVHD (as discussed below).In infants and children in one study, and using a surgically placed Tenckhoffcatheter, incidence of all types of complications during the course of renal fail-ure were only 9%, versus 49% for patients with a transcutaneously placed acutecatheters [7]. With chronic peritoneal dialysis, complications are measured innumber per year, rather than number per week.

The first 6 studies in table 1 have shown that patients with ARF treated byPD have a mortality and incidence of renal recovery at least equal to similarpatients treated by HD, and possibly better. In these studies the percentage ofpatients dying from ARF and complications was less than or equal for patientstreated with PD treatments versus those treated with HD. The article by Firmatreviewed literature reports including over 1,100 patients, and in summation themortality rate was identical for ARF patients receiving PD and HD. Most ofthese studies were performed in the 1970s and 1980s. However, Dr. Struijkcontinued analyzing patients treated by each modality at the Academic FreeHospital in Amsterdam from 1986 through 1999. Although the mortality washigh for ARF patients treated by PD, the mortality was identical to that ofpatients treated with HD. In our own study of 100 patients with ARF in twocommunity hospitals reported in 1983, over half were treated by PD [2]. Therewas a higher rate of recovery of renal function (and survival) following ARF forpatients treated by PD vs. HD (fig. 1). Whether the course of ARF and dialysis

Table 1. Comparison of mortality in ARF�PD vs. HD

Group (first author) Year Patients Mortality, %

PD HD

Orofino [12] 1976 82 52 62Firmat [13] 1979 1,101 50 50Ash [2] 1983 97 38 48Swartz [11] 1980 77 44 60Struijk [14] 1980 45 45 (same)Struijk [15] 1986–1999 50 78 (same)Phu [12] 2002 70 47 15

PD in Acute Renal Failure 241

was short (less than 7 days) or long (more than 7 days) there was a higher rateof recovery for the PD patients versus HD patients (though results were not sig-nificant). We repeated this study 10 years later and in study of 145 patients wefound the same result, a 10% higher patient survival for patients treated by PDversus HD. In a more recent study by Phu et al. [16], 70 adult patients withacute renal failure, 48 due to severe falciparum malaria and 22 due to sepsis,were randomized to treatment by peritoneal dialysis or by CVVH. The mortal-ity was significantly higher in the group treated by peritoneal dialysis versushemofiltration (47 vs. 15%) and there was a diminished rate of renal recovery.The peritoneal dialysis schedule was very aggressive (70 liters of fluid per day),and delivered a dialysis dose for urea probably equal to that of CVVH but a cre-atinine dose probably half of CVVH [49]. This study stands in opposition toprior studies of table 1 above, and what is most unusual is the exceedingly lowmortality of the group treated with CVVH, rather than an unusually high mor-tality in the group treated with PD. Also the study dealt with a high proportionof renal failure following falciparum malaria. It is possible that the heparin anti-coagulation of CVVH was of benefit to these patients. Also, hyperglycemiamay have stimulated malarial growth in the liver or red cells, or high osmolal-ity may have diminished white cell function [49]. Failure to correct acidosismay have been due to use of acetate rather than lactate or bicarbonate as buffer

Fig. 1. Comparison of patient recovery (survival) after acute renal failure, in patientstreated by continuous peritoneal dialysis (left) versus those treated by intermittent hemodial-ysis (right) in two hospitals in Lafayette, Indiana between 1979 and 1982. For both types ofdialysis, one bar indicates outcome for patients treated less than 7 days before recovery ordeath, and the other bar indicates outcome for patients treated for more than 7 days.

Mode and length of treatment*Statistics by chi-squared

�7 days �7 days �7 days �7 days

Per

cent

rec

over

yPD

70

60

50

40

30

20

10

045 00 13

1

24

HDp�0.3

p�0.3

p�0.17

Ash 242

in the PD solution. Finally, the authors used rigid acute catheters for access andopen drainage systems; there was a high incidence of cloudy fluid though avery low incidence of proven peritonitis [49].

Except for the study by Phu, none of these studies were randomized orprospectively controlled, and there was bias in patient selection in the studies.However, this bias was both for and against patients treated by PD. ARFpatients with abdominal trauma, abdominal surgery and remaining drains, andsevere ileus cannot be treated by PD, and these patients may have a higher mor-tality rate. However, in general, patients with ARF following surgery have ahigher rate of recovery from ARF than do patients with other causes of ARFsuch as sepsis or shock. PD was often chosen in these studies for patients withhypotension or cardiovascular instability that would make HD dangerous, alsoselecting a group with a potentially worse outcome. The techniques used for PDin most of these studies were antiquated by today’s standards. In many of thestudies, semi-rigid acute PD catheters were used. These catheters had irregularoutflow characteristics, and had to be removed and re-inserted every three days.Each insertion carried an increased risk of bowel puncture and outflow failure.PD fluid was infused from bottles in many of the studies, and there were few Y-sets to allow drainage and infusion of PD fluid though a single catheter con-nection. In spite of all of these disadvantages, in these studies PD patientsrecovered renal function and survived at least as frequently as patients treatedwith HD, with the notable exception of the study by Phu.

In many of the studies of PD versus HD for ARF, the reason for improvedsurvival in the PD group was related to an increased rate of renal recovery. Veryfew patients with ARF recover general health and leave the hospital withoutrecovering renal function, thus to be supported by dialysis as ESRD patients. Inpatients with ESRD, treatment by CAPD results in better preservation of intrin-sic renal function than treatment by intermittent HD [17–20]. This preservationof renal function is important in ESRD because it maintains endocrine functionof the kidneys, diminishes the clearance requirements for dialysis, minimizesrequired ultrafiltration during dialysis and therefore diminishes physiologicstress during dialysis [17]. Intermittent HD has several known nephrotoxiceffects: (a) generation of inflammatory mediators by the extracorporeal circuit[17]; (b) concomitant and rapid decrease in osmolality and vascular volume,diminishing renal perfusion, and (c) hypotensive episodes resulting in freshischemic lesions in the kidneys [21].

By contrast, CAPD therapy has effects which help to maintain renalperfusion: (a) smaller daily variation in body weight; (b) more constant bloodpressure; (c) continued mild overhydration, with higher mean pulmonary arterialpressure [17]; (d) persistent high blood osmolality, partly due to glucose [17], and(e) continued removal of proteins from the blood, including �2-microglobulin,


albumin, plasminogen-activator inhibitor type 1 (PAI-1) and immunoglobulins[12, 22, 23].

Given the beneficial effects of PD, it is not surprising that some patientswith ESRD and recent initiation of CAPD actually recover intrinsic renal func-tion and no longer need dialysis (3.3%). Given the negative physiologic effectsof HD, it is not surprising that very few ESRD patients treated by intermittentHD recover renal function (0.8%) [17]. Recovery of renal function is mostfrequent in patients whose renal failure was caused by: uncontrolled hyperten-sion, cardiac failure, nephrotic syndrome, rapidly progressive renal failure,analgesic nephropathy, urinary obstruction, and cholesterol emboli [17]. Manyof these underlying conditions are better corrected by CAPD than HD, due toits continuous chemical removal, better preservation of renal perfusion andglomerular filtration rate [10], and slow removal of immunoglobulins. Thesesame physiologic and chemical benefits may account for the higher recovery ofrenal function in most studies, in patients with ARF treated by PD than with HDdiscussed above [8–14].

There is general consensus that continuous dialysis therapies such asCVVH or CVVHD are the most chemically effective therapies for ARF, withthe fewest adverse physiologic effects. Approximately 1/4 of all ARF patientsin the US are treated with CVVH and CVVHD. These ‘gentle’ forms of therapyremove fluid at a slow rate and therefore do not decrease cardiac output. Theyalso do not adversely affect pulmonary function and do not activate the com-plement cascade [6]. CAVH was first described in 1967 [24]. Pump-assistedCVVH and CVVHD were developed to make blood flow rate through thehemofilter more consistent, improve clearances and make the therapy morenearly continuous. Continuous blood therapies require considerable attentionby nurses to assure adequate blood flow, monitor anticoagulation status, adjustthe ultrafiltration rate, and calculate the fluid balance of the patient. The patientis generally immobilized during the treatment. Continuous heparin administra-tion increases risk of bleeding. Vascular access catheters often provide insuffi-cient blood flow and have a risk of infection leading to sepsis [4–6]. In spite ofthe name and intent, the average duration of individual treatments with‘continuous’ blood therapies is 20 h before clotting of the system or need fordiscontinuation to transport the patient for diagnostic or therapeutic procedures.By contrast, PD is a truly continuous dialysis therapy, with less risk and lessnursing effort than CVVH or CVVHD, more mobility during therapy. Somecenters previously performing CVVH or CVVHD for acute dialysis have begunusing slow low efficiency hemodialysis (SLED) with extended duration (6–8 hper day) to avoid many of the problems of CVVH or CVVHD, limiting heparinuse and immobility of the patient while keeping the advantages of vascularstability and improved clearances [25].

Ash 244

The major criticism of PD, of course, is a low clearance of uremic toxins,and for low-molecular-weight toxins the clearance for small toxins is in factlower than other therapies. For continuous therapies, the time-averaged clear-ance is the same as the immediate clearance. For intermittent therapies, thetime-averaged clearance is diminished in proportion to the time between dialy-sis therapies. Table 2 compares the time averaged clearance of various modali-ties used in treatment of ARF, over 24-hour periods. The assumptions used inthis analysis are: (a) CAVH is performed with ultrafiltrate rate of 1 liter/h(15 ml/min); (b) CVVHD is performed with ultrafiltrate rate of 15 ml/minand an additional 2 liters/h of dialysis flow; (c) HD is performed for 4 h, everyother day, with blood flow of 250 ml/min and urea clearance of 150 ml/min, and(d) PD is performed with 2 liter exchanges every 2 h.

From table 2, it is apparent that PD with a modest dialysate use of 1 liter/his less efficient than other modalities for urea and creatinine, but is similarlyefficient in removal of larger molecules such as vitamin B12. The real questionis what are the real uremic toxins, and what are their molecular weights? Ureaand creatinine are not toxic. The very success of peritoneal dialysis in ESRDand SLED in acute dialysis would indicate that it is likely that larger molecularweight toxins are the real causes of uremic illness. Peritoneal dialysis is quiteeffective in removing various anionic organic compounds that function as mid-dle molecules (fig. 2). In ARF, clinical experience confirms that PD results inequal or greater rate of resolution of uremic symptoms and patient survival thesame as the other therapies (described above).

Besides removal of uremic toxins, of course, dialysis must also remove fluidand salt from the patients. With a properly functioning PD catheter, exchangesof 2 liters of dialysate with 2.5 or 4.25% glucose concentration provides dailyfluid removal at the same or greater rate than other regimens, without causinghypotension in most patients. In patients with refractory congestive heart failure(CHF) fluid removal is the main goal, and with PD therapy and hypertonicdialysate, improvement in clinical symptomatology and left ventricular function

Table 2. Comparison of time-averaged clearance offour modalities in treatment of ARF (average clearance perday, ml/min)

CAVH CVVHD Daily HD PD

Urea 15 35 25 12Creatinine 15 30 25 8Phosphate 10 12 10 6Vitamin B12 5 6 6 5


is routine. In one study of 20 patients with resistant CHF, all improved with onlytwelve in/out cycles [26].

Just as with CVVHD, the small molecular clearance of PD can be greatlyincreased by increasing the flow rate of dialysate to 1.5–2 liters/h or more. Tidalperitoneal dialysis (TPD) can easily deliver 2 liters/h into and out of the peri-toneum, and using a cycler automated TPD is only a little more complicatedthan manual in/out exchanges. In a recent study of ARF patients, a peritonealdialysate flow rate of 2 liters/h produced an average normalized creatinine clear-ance of 68.5 liters/week/1.73 m2 BSA and a urea Kt/V of 2.43, versus averagevalues for ‘equilibrium’ PD of 58.9 and 1.80, respectively. Both TPD andmanual exchange PD were adequate for treatment of ARF patients with mild-to-moderate hypercatabolism [27].

The future of peritoneal dialysis for ARF may rest with an old but goodidea, continuous flow PD (CFPD). CFPD utilizes two access points and CFPD,one for inflow of dialysate and the other for outflow. Since there is no inter-ruption of inflow to allow outflow, flow rates are determined only by the rate atwhich the draining catheter can reproducibly drain the abdomen. With CFPDdialysate flow rates of up to 300 ml/min can be maintained through the peri-toneum [28]. With use of an external dialyzer to ‘regenerate’ the dialysate, urea,

Fig. 2. Increase in anionic organic compounds within peritoneal dialysis fluid, overperiods of 2, 4 and 8 h in a patient with ESRD. Chromatograms generated by direct anion-exchange chromatography without protein removal [22].

Gradient

Gradient: 0�100% in 20minSensitivity: 0.01 A.U.F.S.Flow rate: 1ml/min

Anion exchangechromatograms of 2Lof peritoneal fluidremoved from onepatient after dwells of2, 6, and 8 hours

Ab

sorb

ance

at

254

nm U

V (%

)100

0

0%

0%

4 h2 h 8 h

0%

100%

100%

100%

�

�

��

�

Dwell: 2 hDwell: 4 h

Dwell: 8 h

Ash 246

creatinine, and urate clearances average 57, 35, and 39 ml/min, respectively, inadult patients [29]. At 170 ml/min of dialysate flow, urea and creatinine clear-ances have averaged 31 and 23 ml/min, respectively [30]. Recent studies withdialyzer-regenerated PD fluid and using dual Tenckhoff catheters have con-firmed urea clearances of 50 ml/min or more in several ARF patients [31].CFPD has also been used principally for fluid overload. In 6 pediatric patientswith ARDS due to sepsis or SIRS, CFPD at 10–30 ml/kg/h with two Tenckhoffcatheters resulted in a decrease of body weight by an average of 33% and animprovement in alveolar-arterial oxygen gradient [33]. With CFPD the time-averaged clearance of PD can theoretically exceed that of daily 4-hour HDand come close to those of CVVH or CVVHD, and approach the KoA or max-imal clearance theoretically obtainable from the peritoneum [32]. These highdialysate high flow rates seem unrealistic today only because we are accus-tomed to using expensive, pre-packaged dialysate and gravity flow. However,if PD machines reappear which proportion fluid on site, or if sorbent-basedregenerative systems are commercialized [34–36], peritoneal dialysate will beavailable at just about any flow rate desired.

Another requirement for CFPD to be successful is to have effectivedrainage of the peritoneum at relatively high flow rates. Theoretically this is notdifficult because the standard Tenckhoff can drain the abdomen at 300 ml/minor more under gravity flow in CAPD during the early part of outflow. However,flow from Tenckhoff catheters is somewhat variable, and during in CAPDexchanges if there is a diminution in flow then this merely represents a sloweroutflow and some inconvenience. In CFPD at 300 ml/min, it is possible to buildup an extra liter of fluid in the peritoneum in only a few minutes. TheAdvantage™ ‘t-fluted’ peritoneal catheter is a new chronic peritoneal catheterwith grooves rather than holes on two limbs that lie against the parietal peri-toneum. The catheter provides higher flow rate and more complete drainage ofperitoneal fluid than the standard Tenckhoff catheter. The use of this cathetermay make performance of PD in patients easier in patients with ARF, even ifthey have mild ileus [37, 38]. The catheter may also provide faster and morereliable drainage for CFPD, using a second site for infusion of fluid into theperitoneum or one limb of the T-shaped catheter for infusion and one fordrainage [39].

The final challenge for CFPD to be successful is to develop methods forcontrolling the intraperitoneal volume. An animal study has demonstrated thatthere is an optimal peritoneal volume for highest efficiency of the peritoneumin CFPD, about 1.5 liters in the dog [40]. With in-and-out manual PD theabdomen is drained fairly well at the end of each outflow cycle. With CFPD, aswith TPD, the amount of fluid going into and out of the abdomen may beknown, but the intraperitoneal volume is unknown because of the variability of


UF rate. In TPD the abdomen is drained in the middle of an 8 hour treatment,to restart the treatment with a near zero intraperitoneal volume. With CFPD asimilar approach could be taken, but a simpler approach is to control theintraperitoneal pressure rather than the volume. Since the compliance of theperitoneum is relatively constant and symptoms relate to pressure rather thanvolume, controlling the pressure in the peritoneum provides a relatively con-stant volume and lack of symptoms [28]. To provide a constant peritoneal pres-sure using an Advantage catheter or other well-functioning catheter is fairlyeasy for a supine patient with ARF. Infusion of peritoneal fluid through a sec-ond site is performed at up to 1.5–3 liters/h, and the Advantage catheter isattached to a drain bag that lies flat upon a bedside stand, 6 cm above theumbilicus. This provides an intraperitoneal pressure of 12–15 cm, a comfortablelevel for the patient (fig. 3a). UF rate is determined by intermittently weighingthe PD drain bag and subtracting the inflow bag weight. This method should berelatively simple to perform manually, but even simpler with automated regen-erating PD equipment. Though pressure-control not yet been tested in CFPD inpatients, it has been used in one horse with non-oliguric ARF and a creatinine

Fig. 3 a, b. Simplified method for controlling intraperitoneal pressure and volume dur-ing CFPD by placing the drain bag above the umbilicus. a Proposed method in patient.b Method used in adult horse with ARF.

Gravityinfusion at1.5–3 liters/h

Drain bagplaced flat at6 cm aboveumbilicus:Advantagecatheter

Viscera

a b

Ash 248

level of 16 [41]. The draining catheter was a large chest tube in the lowerabdomen, and for pressure control the drain bags were attached to the witherssimilar to saddle bags (fig. 3b). Fluid infusion rate was 72 liters per day andwith CFPD chemical efficiency was approximately twice that of in/outexchanges (D/P near 1 for creatinine and urea). The horse recovered from renalfailure and is doing well today.

The fact that PD results in significant protein loss (5–20 g/day) is gener-ally considered a nutritional problem. However, this loss of protein contributesto the chemical effectiveness of the procedure. In patients with hemolytic ure-mic syndrome, PD significantly reduces plasinogen-activator inhibitor type 1(PAI-1) which inhibits fibrinolysis in hemolytic uremic syndrome [42]. Most ofthe organic anions removed by PD in uremic patients are in fact strongly boundto protein, so protein output or ‘loss’ increases their clearance [22]. Theseprotein-bound organic anions act like middle molecules because the proteinbinding restricts their passage across dialysis membranes; they are still accu-mulating in peritoneal dialysate at 8 h dwell (fig. 2). The presence of proteinwithin the dialysate facilitates the transfer of these compounds into the peri-toneum. The peritoneal transfer of protein can be increased by application ofhypertonicity and pharmacotherapy, the globulin removal by PD on a dailybasis could equal or exceed daily therapeutic plasmapheresis [23].

Every therapy for ARF has some risks. In ARF the success and risks of var-ious dialysis modalities relate in part to the access devices needed to providethem. If acute catheters are used in peritoneal dialysis, each catheter can beused for only three days without high risk of peritonitis or bowel perforation,and each successive catheter has a higher risk of these complications. For acutePD to be performed effectively and safely, a chronic tunneled peritoneal dialy-sis catheter must be the access device. Chronic PD catheters were used in thestudies by Ash and Struijk but rigid acute catheters were used in the study byPhu [2, 14, 16].

Chronic PD catheters can be placed at the bedside in the ICU, procedurerooms, or surgery under local anesthesia. Using the peritoneoscopic techniqueand local anesthesia, it takes only 15–30 minutes to place a chronic 2-cuffTenckhoff catheter for by peritoneoscopy with the Y-TEC procedure, just as isdone for patients with ESRD [43]. Peritoneoscopically placed catheters are usu-ally placed at the lateral border of the rectus and can be directed to lie againstthe parietal peritoneum in a direction avoiding adhesions and bowel loops. Theparietal peritoneal surface provides a more consistent flow of fluid from theabdomen during outflow. Peritoneoscopically placed catheters have the highestrate of successful hydraulic function in the first few weeks of use and over yearsof use [44]. With a properly functioning chronic peritoneal access, the effec-tiveness of PD is increased and risks are considerably diminished, since one


catheter is used for the duration of ARF therapy. Tenckhoff catheters allow peri-toneal access for years rather than the days of safe use for acute PD catheters.Advantage catheters can also be placed peritoneoscopically by expanding theinitial 2 mm puncture site to 9 mm diameter [45, 46]. During the same proce-dure a second small infusion catheter can also be placed with an entry to theperitoneum several inches or more from the Advantage catheter and directedtowards the opposite quadrant if CFPD is planned.

Among therapies for ARF, PD has the unique risk of causing peritonitis.However, in patients in whom infection is suspected as a cause of ARF, per-forming PD can be helpful in assuring that peritonitis is not present. In otherpatients, if peritonitis is detected, diagnostic tests can be implemented to deter-mine the source and antibiotic therapy begun to treat the infection. When nursesare properly trained they soak the connectors with pvp-iodine, put on a maskand nonsterile gloves and use care in performing connections. The risk of con-tamination of PD fluid is minimal during each exchange. The incidence of peri-tonitis in PD therapy of ARF is much different than in CAPD therapy. Ifperitonitis is detected during therapy of ARF with PD it usually occurs within2 or 3 days of starting therapy [11, 47]. This indicates that PD may detect con-tamination of the peritoneum that predates the implementation of PD. Theorganisms causing peritonitis in ARF patients are much different from theorganisms causing peritonitis in CAPD patients. There is predominance ofStaphylococcus epidermidis and Candida species not usually seen in peritoni-tis in CAPD patients (table 3) and mixed infection is frequent [48]. If peritoni-tis occurs during PD therapy, it causes cloudy dialysate and sometimes localsymptoms, but does not usually result in septicemia in the ARF patient. This isa much different outcome than catheter infection during hemodialysis orCVVH, which always results in septicemia.

Table 3. Bacteriology of peritonitis in PD of ARF

Organism n %

Staphylococcus aureus 2 7Staphylococcus epidermidis 7 26Multiple gram-negative organisms 4 15Escherichia coli 1 4Other gram-negative organisms 2 7Multiple (gram-positive and -negative) 2 7Candida sp. 7 26Culture-negative 2 7

Total 27 100

Ash 250

The complications of PD and hemodialysis for ARF have been comparedin one center providing both types of therapy [9]. In this study by Swartz, thepatients treated by HD had a high incidence of severe hypotension and severehemorrhage, acidosis and shunt clotting. PD patients had a high incidence ofhyperglycemia, poor catheter drainage, and asymptomatic peritonitis (table 4).The major causes of death of ARF patients were also different for patientstreated by HD and PD. Death from dialysis unrelated sepsis was higher for theHD group, while cardiac deaths were higher in the PD group due to the morefrequent implementation of this therapy in patients with underlying heartdisease (table 5).

When comparing the overall risks of each type of therapy for ARF, thereare marked differences between CVVH, CVVHD, HD and PD (table 6). Theblood treatment therapies have a significant risk of septicemia, low flow fromblood access, hypotension, membrane clotting, and bleeding. PD therapy

Table 4. Complications of dialysis in ARF1 [9]

HD PD

Number (dialyses, patients) 240, 34 65, 43Severe hypotension2 85/240 (35%) 8/65 (12%)Severe hemorrhage3 15/34 (44%) 2/43 (5%)

Metabolic complicationsHyperglycemia (250 mg/dl) 37/65 (57%)Hypernatremia (150 mEq/l) 2/65 (3%)Acidosis 9/34 (26%)

Neurologic complicationsSeizures 1/34 (3%) 3/43 (7%)Deterioration in state of consciousness 9/65 (14%)

Mechanical complicationsMild bleeding 17/65 (26%)Poor drainage, leaking 34/65 (52%)Shunt clotting 11/34 (32%)

InfectionShunt infection 2/34 (6%)Peritonitis 4/34 (12%)Asymptomatic positive peritoneal cultures 19/65 (29%)

1Expressed as a fraction of total dialysis or total patients.2Blood pressure �90 mm Hg systolic, requiring blood products or pressor

administration.3Requiring transfusion.


includes risks of PD catheter outflow failure, hyperglycemia, and asymptomaticperitonitis. Of these risks of PD, peritonitis is the only one with potential toadversely affect the patient. However, if the initiation of PD detects a pre-existingperitonitis, then antibiotic or surgical therapy may resolve the infection thatcaused ARF. In patients treated by PD during ARF, recognition and therapy ofpre-existing peritonitis contributes to the improved outcome of these patients.

There are no studies comparing the cost to perform of these four uremictherapies, or the cost of treatment of various complications (outside of the study

Table 6. Risks of various dialysis therapies for ARF

CAVH CVVHD HD PD

Septicemia � � � �Vascular occlusion � � � �Hypotension � � � �Membrane clotting � � � �Bleeding due to anticoagulant � � � �PD catheter outflow failure �Hyperglycemia �Asymptomatic peritonitis, �often pre-existing

Table 5. Cause of death from ARF1 [9]

HD (n � 34) PD (n � 43)

Dialysis-unrelated 16 (48%) 14 (33%)Sepsis 11 2Cardiac 2 72

Hemorrhage 3 3Hepatorenal 4 2Other 1 0

Dialysis-related 4 (12%) 5 (12%)Sepsis3 3 3Cardiac 3 22

Hemorrhage 2 0

1Several patients died from more than one cause.2Of the 9 cardiac deaths in the PD group, 7 occurred in

patients with underlying heart disease (p � 0.05).3Shunt sepsis in HD; peritoneal sepsis in PD.

Ash 252

by Phu et al. [16], where all therapies were relatively inexpensive). Nurses per-forming acute PD will confirm that performing this therapy is simple; every2–4 h, a clamp is opened to drain the peritoneum, a new bag is attached to theinflow line and the inflow clamp is opened. The cost of the therapy is only thecost of 6–12 bags of peritoneal dialysate each day, plus the labor of an ICUnurse to open a clamp to drain the peritoneum then attach and infuse the vol-ume of a new bag. Data collection is simple; the outflow volume is measuredand recorded, and the fluid is inspected to determine whether it is clear orcloudy. Much more nursing time is required for procedures and measurementsrelated to CVVH, CVVHD, and HD treatments.

When a properly functioning chronic peritoneal access device is placed,PD is a safe, effective, and inexpensive modality for treatment of ARF. Thismodality is greatly underutilized for treatment of ARF in the United States andwhen improvements in chemical efficiency such as through CFPD are provenand implemented PD will become a much more widely used therapy.

References

1 Rao P, Passadakis P, Oreopoulos DG: Peritoneal dialysis in acute renal failure. Perit Dial Int 2003;23:320–322.

2 Ash SR, Wimberly AL, Mertz SL: PD for acute and ESRD: An update. Hosp Pract 1983;2:179–210.

3 US Renal Data Systems: USRDS 2002 Annual Data Report. Bethesda, NIH, NIDDKD, 2002. 4 Sugino N, Kubo K, Nakazatwo S, Nihei H: Therapeutic modalities and outcome in ARF; in Solez K,

Racusen LC (eds): Acute Renal Failure. New York, Dekker, 1992, pp 443–454. 5 Hansen HE: Dialysis treatment of ARF in children; in Solez K, Racusen LC (eds): Acute Renal

Failure. New York, Dekker, 1992, pp 407–416.6 Fivush BA, Porter CC: Pediatric dialysis therapy; in Solez K, Racusen LC (eds): Acute Renal

Failure. New York, Dekker, 1992, pp 417–431.7 Chadha V, Warady BA, Blowey DL, Simckes AM, Alon US: Tenckhoff catheters prove superior to

Cook catheters in pediatric acute peritoneal dialysis. Am J Kidney Dis 2000;35:1111–1116. 8 Stott RB, Cameron JS, Ogg CS, Bewick M: Why the persistently high mortality in ARF? Lancet

1972;ii:598.9 Posen GA, Lutsello J: Continuous equilibration PD in the treatment of ARF. Dial Bull 1980;1:6.

10 Ng RCK, Suki WN: Treatment of ARF; in Brenner BN, Stein JH (eds): Acute Renal Failure. NewYork, Churchill-Livingstone, 1980, p 231.

11 Swartz RD, Valk TW, Brain AJW, Hsu CH: Complications of HD in ARF. ASAIO J 1980;3:98. 12 Orofino L, Lampreable I, Muniz R, De Sancho JL, Villar F, Gomez Ullate P, Montenegro J, Garcia

Damborenea R: Survival of acute renal failure (ARF) on dialysis: Review of 82 patients. Rev ClinEsp 1976;141:155–160.

13 Firmat J, Zucchini A: Peritoneal dialysis in acute renal failure. Contrib Nephrol. Basel, Karger,1979, vol 17, pp 33–38.

14 Struijk DG, Krediet RT, de Glas-Vos JW, Boeschoten EW, Arisz L: Experiences with acute peri-toneal dialysis in adults. Ned Tijdschr Geneeskd 1984;128:751–755.

15 Struijk DG: Pers. commun., 2000. 16 Phu NH, Hien TT, Mai NT, Chau TT, Chuong LV, Loc PP, Winearls C, Farrar J, White N, Day N:

Hemofiltration and peritoneal dialysis in infection-associated acute renal failure in Vietnam.N Engl J Med 2002;347:895–902.


17 Rottembourg J: Residual renal function and recovery of renal function in patients treated byCAPD. Kidney Int 1993;43:S106–S110.

18 Rottembourg J, Allouache M, Issad B, Baiab R, Benhmida M, Baumelous A, Jacobs C: Outcomeand follow-up on CAPD. Contrib Nephrol. Basel, Karger, 1991, vol 89, pp 16–27.

19 Slingeneyer A, Mion C: Five year follow-up of 155 patients treated by CAPD in European Frenchspeaking countries (abstract). Perit Dial Int 1989;9:176.

20 Lysaght MJ, Vonesh EF, Gotch F, Ibels L, Keen M, Lindholm B, Nolph KD, Pollock CA, Prowant B,Farrel PC: The influence of dialysis treatment modality on the decline of remaining renal func-tion. ASAIO Trans 1991;37:598–604.

21 Conger JD: Does HD delay recovery from ARF? Semin Dial 1990;3:146–148.22 Ash SR, Bungu ATJ, Regnier FE: Dependence of middle molecular clearance on protein concen-

tration of peritoneal fluid; in Maher JF, Winchester JF (eds): Frontiers in Peritoneal Dialysis.New York, Field, Rich & Associates, 1986, pp 56–63.

23 Popovich R, He Z, Moncrief J: Peritoneal Membrane Plasmapheresis. Trans ASAIO 1992;38:M668–M672.

24 Henderson L, Besarab A, Michaels A, Blumle LW: Blood purification by ultrafiltration and fluidreplacement (diafiltration). Trans ASAIO 1967;13:216–222.

25 Kumar VA, Craig M, Depner TA, Yeun JY: Extended daily dialysis: A new approach to renalreplacement for acute renal failure in the intensive care unit. Am J Kidney Dis 2000;36:294–300.

26 Aggarwal HK, Sumit N, Sumit N, Sen J, Singh M: Evaluation of role of acute intermittent peri-toneal dialysis in resistance congestive heart failure. J Assoc Phyns India 2002;50:1115–1119.

27 Chitalia VC, Almeida AF, Rai H, Bapat M, Chitalia KV, Acharya VN, Khanna R: Is peritoneal dial-ysis adequate for hypercatabolic acute renal failure in developing countries. Kindey Int 2002;61:747–757.

28 Roberts M, Ash SR, Lee DBN: Innovative peritoneal dialysis: Flow-through and dialysate regen-eration. ASAIO J 1999;45:372–378.

29 Shinaberger JH, Shear L, Barry KG: Increasing efficiency of peritoneal dialysis: Experience withperitoneal-extracorporeal recirculation dialysis. Trans ASAIO 1965;11:76–82.

30 Stephen RL, Atkin-Thor E, Kolff WJ: Recirculating PD with subcutaneous catheter. Trans ASAIO1976;22:575–585.

31 Amerling R, Glezerman I, Savransky E, Dubrow A, Ronco C: Continuous flow peritoneal dialy-sis: Principles and applications. Semin Dial 2003;16:335–340.

32 Cruz C, Melendez A, Gotch F, Folen T, Levin NW: Continuous flow peritoneal dialysis (CFPD):Preliminary clinical experience. Perit Dial Int 2000;20(suppl 1):S6.

33 Sagy M, Silver P: Continuous flow peritoneal dialysis as a method to treat severe anasarca in chil-dren with acute respiratory distress syndrome. Crit Care Med 1999;27:2532–2536.

34 Ash SR, Carr DJ, Blake DE, Thornhill JA: The sorbent suspension reciprocating dialyzer for usein peritoneal dialysis; in Maher JF, Winchester JF (eds): Frontiers in Peritoneal Dialysis. NewYork, Field, Rich & Associates, 1986, pp 148–156.

35 Gordon A, Lewin AJ, Maxwell MH, Morales ND: Augmentation of efficiency by continuous flowsorbent regeneration peritoneal dialysis. Trans ASAIO 1976;22:599–604.

36 Ash SR: The Allient™ Dialysis System. Semin Dial 2004;17:164–165.37 Ash SR, Janle EM: T-fluted peritoneal dialysis catheter. Adv Perit Dial 1993;9:223–226.38 Ash SR, Sutton JM, Mankus RA, Rossman J, de Ridder V, Nassvi MS, Ross J: Clinical trials of the

T-fluted (ash advantage) peritoneal dialysis catheter. Adv Renal Replacement Ther 2002;9:133–143.39 Diaz-Buxo JA: What is the role of automated peritoneal dialysis and continuous flow peritoneal

dialysis? Contrib Nephrol. Basel, Karger, 2003, vol 140, pp 264–271.40 Ash SR, Janle EM: Continuous flow-through peritoneal dialysis (CFPD): Comparison of effi-

ciency to IPD, TPD, and CAPD in an animal model. Periton Dialysis Int 1997;17:365–372.41 Gallatin L, Couetil L, Ash SR: Continuous flow peritoneal dialysis for the treatment of acute renal

failure in an adult horse. 2004;in press. 42 Bergstein JM, Riley M, Bank NU: Role of plasminogen-activator inhibitor type 1 in the patho-

genesis and outcome of the hemolytic uremic syndrome. N Engl J Med 1992;327:755–759.43 Ash SR: Peritoneal access devices and placement techniques; in Nissenson AR, Fine RN (eds):

Dialysis Therapy. Philadelphia, Hanley & Belfus, 1992, pp 23–30.

Ash 254

44 Ash SR, Daugirdas JT: Peritoneal access devices; in Daugirdas JT, Ing TS (eds): Handbook ofDialysis. Boston, Little, Brown, 1985, pp 195–218.

45 Ash SR: Chronic peritoneal dialysis catheters: Procedures for placement, maintenance, and removal.Semin Nephrol 2002;22:221–236.

46 Ash SR: Chronic peritoneal dialysis catheters: Overview of design, placement, and removal pro-cedures. Semin Dial 2003;16:323–334.

47 Dandecha P, Sangthawan P: Peritonitis in acute peritoneal dialysis in a university hospital. J MedAssoc Thai 2002;85:477–481.

48 Sharma RK, Kuma J, Gupta A, Gulati S: Peritoneal infection in acute intermittent peritoneal dial-ysis. Ren Fail 2003;25:975–980.

49 Daugirdas JT: Peritoneal dialysis in acute renal failure – why the bad outcome? N Engl J Med2003;347:933–935.

Stephen R. Ash, MD, FACP HemoCleanse, Inc., 3601 Sagamore Parkway N., Lafayette, IN 47904 (USA)Tel. �1 765 742 4813, Fax �1 765 742 4823, E-Mail [email protected]


Intermittent Hemodialysis for Acute Renal Failure Patients – An Update

Norbert Lameire, Wim Van Biesen, Raymond Vanholder, Eric Hoste

Department of Medicine, University Hospital, De Pintelaan, Gent, Belgium

Acute renal failure (ARF) is a frequent complication in the intensive careunit (ICU), and is associated with increased morbidity, mortality and healthcare costs [1]. According to Maher et al. [2], ARF requiring dialysis reducedsurvival by 50% and Morgera et al. [3] found an in-hospital mortality of 69%.Of the patients discharged from the hospital, however, 77 and 70% still survivedafter 6 and 12 months, respectively, but 10% remained dialysis-dependent.

Although considerable advances have been made in the treatment of ICU-related ARF, several important issues still remain a matter of debate.

It has repeatedly been suggested that continuous renal replacement thera-pies (CRRT) as compared to intermittent dialysis (IHD) were linked toimproved outcomes, but convincing evidence is lacking. Even for secondaryendpoints like hemodynamic stability or removal of water and solutes, data onthe superiority of CRRT are missing or at least contradictory. Studies on thesequestions are jeopardized by biases, leading to unsatisfactory randomizationand stratification of risk factors between the observed populations, and to theinclusion of insufficient numbers of patients [4].

Hard data remain absent or conflictive regarding the timing to start dialy-sis, the adequacy levels necessary to optimize survival as well as to the impactof biocompatible membranes [5].

CRRT vs. IHDTwo major types of renal replacement therapy (RRT) are currently avail-

able for the treatment of ARF: IHD and CRRT. A survey by the Acute Dialysis

Lameire/Van Biesen/Vanholder/Hoste 256

Quality Initiative (ADQI) group revealed that the choice between CRRT andIHD was based on local circumstances rather than on scientific evidence [6].Although the most obvious difference between the two treatments is the timespan over which they are applied, some other technical differences may in factbe more relevant. IHD is performed as a highly efficient technique, relying ondiffusion, and thus necessitating high dialysate flows to maintain high concen-tration gradients. CRRT is mostly performed as a low efficiency technique,mainly relies on convection, and implies the need for sterile substitution fluids.As consequence, a water treatment system and a dialysis monitor are manda-tory for IHD, whereas CRRT can principally be performed with more simplehardware, but necessitates the use of industrially prepared substitution fluids.These technical peculiarities imply also that an IHD machine can be pro-grammed to perform continuous therapy, but that a CRRT machine cannot beused to perform intermittent treatment, as the low-efficient nature of the set-upnecessitates prolonged treatment duration to achieve adequacy goals. To enhanceefficiency, CRRT has evolved from continuous arteriovenous hemofiltration(CAVH), without blood pump, to high volume continuous venovenoushemo(dia)filtration (CVVH(D)), applying sophisticated blood pumps and fluidbalance systems equilibrating hemofiltered and substituted fluids. In this way,the technical simplicity as a major advantage of CRRT is completely lost.Hence, the evolution to hybrid therapies, whereby IHD machines are used toperform extended treatments, is a logical next step.

Only a limited number of randomized controlled studies compared out-comes on CRRT vs. IHD in ARF [7–11]. A meta-analysis by Tonelli et al. [12]including over 600 patients, found no significant difference in mortality of crit-ically ill patients with ARF treated with IHD compared to CRRT. This resultremained present even after addition of non-randomized trials, and aftercontrolling for potential remaining differences in baseline severity of illness.Another meta-analysis by Kellum et al. [4] allows no definite conclusion, as itincluded mainly non-randomized trials, and omitted some available randomizedtrials. In addition, the weighing of the impact of the different studies and themethodology to correct for differences in severity of disease were not appro-priate. In a recent Canadian study, mortality in the CRRT group was higher thanin the IHD group (71.9 vs. 42.2%, p � 0.001, respectively) [11]. The well-performed randomized controlled trial by Mehta et al. [7] demonstrated thenear impossibility to stratify for all underlying risk factors. In view of the emer-gence of hybrid therapies, it can even be questioned whether there is a real needfor another large randomized trial in this area, since overall, both techniqueshave a similar outcome. In specific conditions, however, one of both is anabsolute preference, like, for example, CRRT in patients with cerebral edemaor liver failure, or IHD in patients with increased bleeding risk. Another recent

Intermittent Hemodialysis for Acute Renal Failure Patients 257

comparative multicenter study has been performed in France by Guérin et al.[13]. Among the 587 patients who were included in this study, 354 receivedCRRT and 233 intermittent renal replacement therapy as first choice. CRRTpatients had a higher number of organ dysfunctions on admission and at thetime of ARF and higher SAPS II at time of ARF. Mortality was 79% in theCRRT group and 59% in the IRRT group. Logistic regression analysis showeddecreased patient survival to be associated with SAPS II on admission, oliguria,admission from hospital or emergency room, number of days between admis-sion and ARF, cardiac dysfunction at time of ARF, and ischemic ARF. Nounderlying disease or nonfatal disease, and absence of hepatic dysfunction wereassociated with an increase in patient survival. The type of renal replacementtherapy was not significantly associated with outcome.

The most logical conclusion, as we pointed out before [14], is that all cen-ters taking care of critically ill patients with ARF should have the possibility toperform both techniques in a qualitatively acceptable way. The emergence ofSLEDD (slow low efficient daily dialysis) might have a positive influence inthis regard, as this allows performing both strategies with one single machine.

An important aspect might be the cost. CRRT is more expensive due tohigher labor costs, more expensive all-inclusive disposables, and the need forspecific ‘additional’ machines that can only be used for treatment of ARF inthe ICU. A cost-analysis in 32 Canadian hospitals [11] revealed that the meancost of CRRT per week ranged from USD 3,486 to 5,117, compared to onlyUSD 1,342 for IHD. In contrast, the length of dialysis treatment in survivorswas longer in the IHD group, but not statistically significant (14.8 vs. 19.1days, p � 0.9 � NS).

Dialysis DoseIn chronic renal failure (CRF), Kt/Vurea (urea clearance per minute, multi-

plied by dialysis time, normalized for distribution volume of urea) is currentlyused as a marker of small solute clearance. This is based on the assumption thatKt/Vurea not only reflects the removal of toxic waste products, but offers also anindication of dietary protein intake and catabolism.

In ARF patients, blood urea levels are often used as a marker to start RRTand to monitor dialysis adequacy. Blood urea levels in highly catabolic ICUpatients can however be the reflection of many factors besides dialysis adequacy.A recent Canadian survey [15] found that in more than 75% of the centers,no formal method of the monitoring of the dialysis prescription was applied.The use of conventional urea kinetic modeling might lead to erroneous conclu-sions, since the assumptions at the basis of these calculations are violated inmost ICU patients. The distribution volume, e.g., is frequently expanded, andthe urea generation rate is not in a steady state. Clark et al. [16] have developed


different kinetic models to capture the differences in kinetic behavior betweenchronic and acute renal failure patients. If Kt/V is used as an index of adequacyin ARF, the thresholds should undoubtedly exceed those aimed at in CRF.

An accurate estimate of volume of distribution of urea (Vurea) is criticallyimportant to guide the prescription of therapy and the quantification of deliv-ered dialysis dose in patients with chronic and acute renal failure (ARF). WhileVurea has been shown to be substantially the same as total body water (TBW) inother patient populations, this relationship is not automatically true in ARFpatients. A study by Himmelfarb et al. [17] showed that determination of TBWby anthropometric measurements yielded significantly lower measures com-pared to TBW determined by physiological formulae and by bioelectricalimpedance analysis and that all measures of Vurea by blood-based kineticsexceeded TBW measurements by any method (7–50% difference). It was thusconcluded that estimates of TBW cannot be used as a surrogate for Vurea indetermining dialysis adequacy in ARF patients.

It has been registered before that most nephrologists prescribe a dialysisdose below those recommended for CRF patients, and that actually delivereddoses are often still below those prescribed [18] due to clotting, insufficientvascular access, hemodynamic instability, or practical pitfalls.

Prospective, formal, blood-side, urea kinetic modeling (UKM) has yet tobe applied in intermittent hemodialysis for acute renal failure (ARF). Methodsfor prescribing a target, equilibrated Kt/V (eKt/V) are described for this settingby Kanagasundaram et al. [19]. eKt/V was derived using delayed posthemodial-ysis urea samples and formal, double-pool UKM (eKt/Vref), and by applying theDaugirdas-Schneditz venous rate equation to pre- and posthemodialysis sam-ples (eKt/Vrate). Individual components of prescribed and delivered dose werecompared. Prescribed eKt/V values were determined using in vivo dialyzerclearance estimates and anthropometric (Watson and adjusted Chertow) andmodeled urea volumes. eKt/Vref was well-approximated by eKt/Vrate. ModeledV exceeded Watson V by 25 � 29% and Adjusted Chertow V by 18 � 28%,although the degree of overestimation diminished over time. This differencewas influenced by access recirculation (AR) and use of saline flushes. Themedian % difference between Vdprate and Watson V was reduced to 1% afteradjusting for AR for the 22 sessions with � or � 1 saline flush. The mediancoefficients of variation for serial determinations of Adjusted Chertow V,modeled V, urea generation rate, and eKt/Vref were 2.7, 12.2, 30.1 and 16.4%,respectively. Because of comparatively higher modeled urea Vs, deliveredeKt/Vref was lower than prescribed eKt/V, based on Watson V or AdjustedChertow V, by 0.13 and 0.08 Kt/V units. The median absolute errors ofprescribed eKt/V vs. delivered therapy (eKt/Vref) were, however, not large andwere similar in prescriptions based on the Adjusted Chertow V (0.127) vs. those


based on various double-pool modeled urea volumes (approximately0.127). Equilibrated Kt/V can be thus be derived using formal, double-poolUKM in intensive care unit ARF patients, with the venous rate equation pro-viding a practical alternative. A target eKt/V can be prescribed to within amedian absolute error of less than 0.14 Kt/V units using practical prescriptionalgorithms.

A recent study by Liao et al. [20] compared the effective dose delivery bythree acute dialysis therapies: continuous venovenous hemofiltration (CVVH),daily HD, and sustained low-efficiency dialysis (SLED). A modified equivalentrenal clearance (EKR) approach to account for the initial unsteady-state stageduring dialysis was employed. Effective small solute clearance in CVVH wasfound to be 8 and 60% higher than in SLED and daily HD, respectively.Differences are more pronounced for middle and large solute categories, andEKR in CVVH is approximately 2- and 4-fold greater than the correspondingvalues in daily HD and SLED, respectively. The superior middle and largesolute removal for CVVH is due to the powerful combination of convection andcontinuous operation. In CVVH, a decrease in the initial BUN from 150 to50 mg/dl is predicted to decrease TAC and, therefore, increase EKR by approx-imately 35%. After clinical validation, the quantification method presented inthis article could be a useful tool to assist in the dialytic management of criti-cally ill ARF patients.

Only a few studies have tried to relate dialysis dose to outcome in ARFpatients. Ronco et al. [21] randomized 425 patients with ARF treated with CRRTto different doses of hemofiltration with substitution: 20, 35 or 45 ml/min/kg.Survival in the 20-ml/kg/min group was inferior to that observed in the 35- and45-ml/kg/min groups. No survival advantage was observed between the twohighest doses, implying that 35 ml/kg/min is probably the most optimal dose. Incontrast, a randomized controlled study by Bouman et al. [22] found that neitherthe 28th day of survival nor the recovery of renal function were improved byhigh ultrafiltrate volumes (up to a median of 48.2 ml/kg/min).

In the field of IHD, the Cleveland group pointed out that dose of dialysiswas inversely related to mortality, at least in the subgroup of patients with a‘moderate risk’ profile [23]. In patients with a low or high risk, dialysis dosedid not independently predict mortality, probably because other, non-RRT-related factors play a more important role in these patients. For patients with‘moderate’ risk, a urea reduction rate higher than 58% was associated with asignificant reduction in mortality.

Schiffl et al. [24] compared daily versus alternate day IHD treatment in160 ARF patients. An intention-to-treat analysis revealed a mortality of 28% inthe daily dialysis group, vs. 46% in the alternate day group. In a multipleregression analysis, alternate day dialysis was an independent risk factor for


mortality. The delivered cumulative weekly Kt/V, as measured by the Daugirdasformula, was twice as high in the daily treatment group compared to the alternateday group. The time averaged blood urea concentration was 60 vs. 104 mg/dl(p � 0.001) in the daily vs. alternate day group. Both delivered dose and result-ing ‘clearance’ of toxic waste products were thus superior in the more inten-sively treated group. In addition, hypotension was less prevalent in the dailytreated (5 vs. 25%, p � 0.001) group, which was most likely related to thelower ultrafiltration volume per session, whereas also time to recover renalfunction was shorter (9.2 vs. 16 days, p � 0.001).

Biocompatibility of the Dialysis MembranesConsiderable controversy exists whether in patients with ARF the use of

biocompatible dialysis membranes can positively influence patient survival andrecovery of renal function as compared to bio-incompatible membranes. Theproblem ‘biocompatibility’ in this discussion is limited to the property of somedialyzer membranes, e.g. cuprophane, to activate leukocytes and complement.Different trials yielded contradictory results. The discussion is hampered by thedivergence of definitions of ‘biocompatibility’. Some authors compared cellu-losic with synthetic membranes, whereby ‘cellulosic’ is considered bioincom-patible. However, modified cuprophane membranes (e.g. hemophan) have alower complement activating capacity than cuprophane, and should hence beconsidered ‘biocompatible’. It has also been argued that differences in flux areresponsible for the observed differences, whereby it has to be understood thatthe non-modified cellulosic membranes are always low-flux, whereas biocom-patible (either synthetic or modified cellulose) membranes may be both low orhigh flux. Increasing flux improves the removal of middle- and high-molecular-weight uremic retention products, which might be important in catabolic ICUpatients [25]. Whether the removal of cytokines by RRT is possible and/orbeneficial remains controversial [26–29].

In a recent meta-analysis, including a total of 867 patients, Subramanianet al. [30] observed a relative risk of mortality for cellulosic membranes of 1.37(CI: 1.02–1.83). A separate subanalysis of studies with only unmodified cellu-lose as control group vs. studies where both unmodified as modified cellulosemembranes served as control, revealed that the observed survival benefit forsynthetic membranes was largely due to the difference with unmodified cellu-lose. When only correctly randomized trials were included in the meta-analysis,the statistical significance of the survival advantage of biocompatible mem-branes was lost. Jaber et al. [31] performed another meta-analysis comparingunmodified cellulose (bioincompatible) vs. synthetic and modified cellulosemembranes (biocompatible), including a total of 722 patients. The survivaladvantage for the biocompatible membranes did not reach statistical significance.


In none of these two meta-analyses was an attempt made to dissect the impactof biocompatibility from that of flux.

Meta-analyses of the published literature are increasingly being used,allowing similar clinical trials to be combined quantitatively, thereby increasingthe precision of the estimation of treatment effect. However, several meta-analyses on the topic of membrane biocompatibility in ARF were discordantand in analysis of these meta-analyses, Teehan et al. [32] observed that thisdivergence was due in part to the differences among meta-analyses in inclusionand exclusion criteria, the paucity of randomized controlled trials, variation inindividual study quality, and heterogeneity in the study populations and settings.Understanding these issues is important to properly interpret results from thesemeta-analyses.

The use of low complement-activating membranes implies a potentialsurvival advantage, but this effect might in clinical practice be lower thanexpected. However, none of the published studies has ever demonstrated a neg-ative impact of low complement-activating membranes. Hence, the only reasonnot to prefer these membranes is their cost.

In conclusion, both CRRT and IHD have advantages and disadvantages,and any attempt to improve either of them makes them more and more similar.A RRT modality for ICU should be easy to use, convenient, and offer the pos-sibility to deliver a broad range of therapeutic modalities while reducing theworkload. The emergence of ‘hybrid techniques’ such as slow low efficientextended daily dialysis (SLEDD) is therefore a logical consequence [33, 34]. Inthese hybrid techniques, a classical dialysis monitor allowing online prepara-tion of dialysate, and, if necessary, substitution fluid is used to performextended dialysis treatments, whereby the intensity and the duration of thetreatment can be adapted to the needs of the patient. Regardless of whether thetreatment is diffusive, convective, or a combination of both, the technical set-upof the machine is always the same for the ICU nurse.

The advantages and results of SLEDD will be discussed elsewhere in thisvolume.

References

1 Lameire N, Van Biesen W, Vanholder R, Colardijn F: The place of intermittent hemodialysis in thetreatment of acute renal failure in the ICU patient. Kidney Int Suppl 1998;66:S110–S119.

2 Maher ER, Robinson KN, Scoble JE, Farrimond JG, Browne DR, Sweny P, Moorhead JF:Prognosis of critically-ill patients with acute renal failure: APACHE II score and other predictivefactors. Q J Med 1989;72:857–866.

3 Morgera S, Kraft AK, Siebert G, Luft FC, Neumayer HH: Long-term outcomes in acute renal fail-ure patients treated with continuous renal replacement therapies. Am J Kidney Dis 2002;40:275–279.


4 Kellum JA, Angus DC, Johnson JP, Leblanc M, Griffin M, Ramakrishnan N, Linde-Zwirble WT:Continuous versus intermittent renal replacement therapy: A meta-analysis. Intensive Care Med2002;28:29–37.

5 Abdeen O, Mehta RL: Dialysis modalities in the intensive care unit. Crit Care Clin 2002;18:223–247.6 Ronco C, Brendolan A, Bellomo R: Continuous renal replacement techniques. Contrib Nephrol.

Basel, Karger, 2001, vol 132, pp 236–251.7 Mehta RL, McDonald B, Gabbai FB, Pahl M, Pascual MT, Farkas A, Kaplan RM: A randomized

clinical trial of continuous versus intermittent dialysis for acute renal failure. Kidney Int 2001;60:1154–1163.

8 John S, Griesbach D, Baumgartel M, Weihprecht H, Schmieder RE, Geiger H: Effects of contin-uous haemofiltration vs. intermittent haemodialysis on systemic haemodynamics and splanchnicregional perfusion in septic shock patients: A prospective, randomized clinical trial. Nephrol DialTransplant 2001;16:320–327.

9 Uehlinger DE, Jacob S, Eichelberger M: A randomized, controlled single center study for thecomparison of continuous renal replacement therapy with intermittent hemodialysis in criticallyill patients with acute renal failure (abstract). J Am Soc Nephrol 2001;12:278A.

10 Sandy D, Moreno L, Paganini EP: A randomized stratified dose equivalent comparison of contin-uous veno-venous hemodialysis vs. intermittent hemodialysis support in ICU acute renal failurepatients (abstract). J Am Soc Nephrol 1998;9:225A.

11 Manns B, Doig CJ, Lee H, Dean S, Tonelli M, Johnson D, Donaldson C: Cost of acute renal fail-ure requiring dialysis in the intensive care unit: Clinical and resource implications of renal recov-ery. Crit Care Med 2003;31:449–455.

12 Tonelli M, Manns B, Feller-Kopman D: Acute renal failure in the intensive care unit: A system-atic review of the impact of dialytic modality on mortality and renal recovery. Am J Kidney Dis2002;40:875–885.

13 Guerin C, Girard R, Selli JM, Ayzac L: Intermittent versus continuous renal replacement therapyfor acute renal failure in intensive care units: Results from a multicenter prospective epidemio-logical survey. Intens Care Med 2002;28:1411–1418.

14 Dhondt A, Van Biesen W, Vanholder R, Lameire N: Selected practical aspects of intermittenthemodialysis in acute renal failure patients. Contrib Nephrol. Basel, Karger, 2001, vol 132,pp 222–235.

15 Hyman A, Mendelssohn DC: Current Canadian approaches to dialysis for acute renal failure inthe ICU. Am J Nephrol 2002;22:29–34.

16 Clark WR, Ronco C: Renal replacement therapy in acute renal failure: Solute removal mecha-nisms and dose quantification. Kidney Int Suppl 1998;66:S133–S137.

17 Himmelfarb J, Evanson J, Hakim RM, Freedman S, Shyr Y, Ikizler TA: Urea volume of distribu-tion exceeds total body water in patients with acute renal failure. Kidney Int 2002;61:317–323.

18 Evanson JA, Himmelfarb J, Wingard R, Knights S, Shyr Y, Schulman G, Ikizler TA, Hakim RM:Prescribed versus delivered dialysis in acute renal failure patients. Am J Kidney Dis 1998;32:731–738.

19 Kanagasundaram NS, Greene T, Larive AB, Daugirdas JT, Depner TA, Garcia M, Paganini EP:Prescribing an equilibrated intermittent hemodialysis dose in intensive care unit acute renal fail-ure. Kidney Int 2003;64:2298–2310.

20 Liao Z, Zhang W, Hardy PA, Poh CK, Huang Z, Kraus MA, Clark WR, Gao D: Kinetic compari-son of different acute dialysis therapies. Artif Organs 2003;27:802–807.

21 Ronco C, Bellomo R, Homel P, Brendolan A, Dan M, Piccinni P, La Greca G: Effects of differentdoses in continuous veno-venous haemofiltration on outcomes of acute renal failure: A prospec-tive randomised trial. Lancet 2000;356:26–30.


23 Paganini EP, Tapolyai M, Goormastic M, Halstenberg WK, Kozlowski L, Leblanc M, Lee J,Moreno L, Sakai K: Establishing a dialysis therapy/patient outcome link in intensive care unitacute dialysis for patients with acute renal failure. Am J Kid Dis 1996;28(suppl 3):S81–S89.



25 Vanholder R, De Vriese A, Lameire N: The role of dialyzer biocompatibility in acute renal fail-ure. Blood Purif 2000;18:1–12.

26 De Vriese AS, Colardijn F, Phillippé JJ, Vanholder R, De Sutter JH, Lameire NH: Cytokineremoval during continuous hemofiltration in septic patients. J Am Soc Nephrol 1999;10:846–853.

27 Tetta C, D’Intini V, Bellomo R, Bonello M, Bordoni V, Ricci Z, Ronco C: Extracorporeal treat-ments in sepsis: Are there new perspectives? Clin Nephrol 2003;60:299–304.

28 Venkataraman R, Subramanian S, Kellum JA: Clinical review: Extracorporeal blood purificationin severe sepsis. Crit Care 2003;7:139–145.

29 Ronco C, Inguaggiato P, D’Intini V, Cole L, Bellomo R, Poulin S, Bordoni V, Crepaldi C,Gastaldon F, Brendolan A, Trairak P, Khajohn T: The role of extracoporeal therapies in sepsis.J Nephrol 2003;16(suppl 7):S34–S47.

30 Subramanian S, Venkataraman R, Kellum JA: Influence of dialysis membranes on outcomes inacute renal failure: A meta-analysis. Kidney Int 2002;62:1819–1823.

31 Jaber BL, Lau J, Schmid CH, Karsou SA, Levey AS, Pereira BJ: Effect of biocompatibility ofhemodialysis membranes on mortality in acute renal failure: A meta-analysis. Clin Nephrol 2002;57:274–282.

32 Teehan GS, Liangos O, Lau J, Levey AS, Pereira BJ, Jaber BL: Dialysis membrane and modalityin acute renal failure: Understanding discordant meta-analyses. Semin Dial 2003;16:356–360.

33 Van Den Noortgate N, Verbeke F, Dhondt A, Colardijn F, Van Biesen W, Vanholder R, Lameire N:The dialytic management of acute renal failure in the elderly. Semin Dial 2002;15:127–132.

34 Van Biesen W, Vanholder R, Lameire N: Dialysis strategies in critically ill acute renal failurepatients. Curr Opin Crit Care 2003;9:491–495.

Norbert Lameire, MDRenal Division, Department of Medicine, University Hospital185, De Pintelaan, BE–9000 Gent (Belgium)Tel. �32 9 2404524, Fax �32 9 2404599, E-Mail [email protected]


Continuous Renal ReplacementTechniques

William R. Clarka,b, Claudio Roncoc

aNxStage Medical, Inc., Lawrence, Mass., and bIndiana University School of Medicine, Indianapolis, Ind., USA; cDepartment of Nephrology, Ospedale San Bortolo, Vicenza, Italy`

Acute renal failure (ARF) occurring in patients admitted to the intensivecare unit is often different from the syndrome observed in renal wards. Patientsare critically ill and several organ systems are involved in the syndrome.Frequently, sepsis and a multiple organ dysfunction syndrome complicate theclinical picture. Under such circumstances, an adequate renal replacement ther-apy must be instituted providing effective blood purification and correction ofthe various homeostatic disorders [1]. Standard hemodialysis and peritonealdialysis have displayed significant limitations in these patients while continu-ous renal replacement therapies are rapidly gaining consensus and displayinginteresting clinical advantages. The critically ill patient presents a severe hemo-dynamic instability, sepsis and septic shock, he or she may require mechanicalventilation, cardiac mechanical support and other types of vital supports. Inthese conditions, intermittent hemodialysis appears to produce further hemo-dynamic instability and only partial correction of the uremic syndrome [2].Peritoneal dialysis on the other hand may present mechanical and infectiouscomplications and it is limited by the low clearances and ultrafiltration rates.Under all these circumstances the logical approach is the one of providing atherapy with the following characteristics:

(a) Excellent clinical tolerance.(b) Excellent capacity of blood purification for different molecules.(c) Optimal correction of electrolyte disorders.(d) Optimal correction of acid-base disorders.(e) Excellent biocompatibility with minimal or no proinflammatory effects.(f) Minimal or absent side effects including a negative impact on recovery

of the native organ function.

Continuous Renal Replacement Techniques 265

(g) Possibly improved outcomes.(h) Easy institution and easy monitoring of treatment.Most of these targets have been achieved with the use of continuous renal

replacement therapy (CRRT). This is especially true considering the mostrecent techniques in which the treatment dose and the applied technology havebeen optimized.

In this paper, the major modalities comprising CRRT are discussed. Priorto doing this, we feel however that it is important to review the basic conceptsunderlying each single technique to better interpret the possible advantages andto better target the prescription of the selected therapy. Finally, we will also pre-sent emerging applications of CRRT for disorders beyond ARF.

Mechanisms of Solute and Water Transport

The most important mechanisms of solute and water transport acrosssemipermeable membranes can be considered: diffusion, convection andultrafiltration [3].

DiffusionDiffusion is a process of transport in which the molecules that are present

in a solvent and can freely cross the membrane tend to move from the region athigher concentration into the region at lower concentration (fig. 1). In reality,molecules present a random movement. However, since they tend to reach thesame concentration in the available space occupied by the solvent, the numberof particles crossing the membrane towards the region at lower concentrationwill be statistically higher. This is therefore a transport mechanism that occursin the presence of a concentration gradient for solutes that are not restricted indiffusion by the porosity of the membrane. Besides the concentration gradient(dc), the diffusion flux is influenced by the characteristics of the membraneincluding surface area (A) and thickness (dx), the temperature of the solution(T), and the diffusion coefficient of the solute (D). The diffusion flux of a givensolute (Jx) will therefore result from equation (1):

Jx � D T A (dc/dx) (1)

Based on the above-mentioned concepts, one could predict with reasonableaccuracy the clearance value in the presence of given solute, solvent, membraneand operational conditions. However, several factors influencing the final clear-ance values may lead to a certain discrepancy between the theoretically expectedvalue and the empirically obtained value. As an example, protein binding orelectrical charges in the solute may negatively affect the final clearance value.

Clark/Ronco 266

On the other hand, an increased amount of convection may contribute to agreater transport of solutes especially in the higher-molecular-weight range.

ConvectionConvection is a form of transport that requires a movement of fluid across

the membrane as a consequence of a transmembrane pressure gradient (TMP). Inconjunction with this fluid transport, the crystalloids present in the solution (butnot the cells or the colloids that are retained by the membrane) are transported tothe other side of the membrane with a mechanism defined solvent drag. The fluidtransport is defined ultrafiltration (fig. 2) and it can be described by equation (2):

Jf � Kf � TMP (2)

where Kf is the coefficient of hydraulic permeability of the membrane andTMP � (Pb � Puf) � �. In the TMP expression, Pb is the hydrostatic pressureof blood, Puf the hydrostatic pressure of ultrafiltrate or dialysate, and � theoncotic pressure generated by plasma proteins in the blood.

Fig. 1. Mechanisms of solute removal in different blood purification techniques.

N P

C

Diffusion Convection

Fig. 2. The mechanism of ultra-filtration in response to a transmembranepressure gradient.

Pressure

Uf

Uf Membrane


The convective flux of a given solute x will, therefore, depend on theamount of ultrafiltration (Jf), on the concentration of the solute in plasma water(Cb) and by the sieving characteristics of the membrane for the solute (S):

Jx � Jf Cb (1 � �) � Jf CbS (3)

The sieving coefficient (S) is, under theoretical conditions, regulated bythe reflection coefficient of the membrane (Staverman reflection coefficient: �)according to equation (4) [4]:

S � 1 � �. (4)

In clinical practice, however, because plasma proteins and other factors modifythe original reflection coefficient of the membrane, the final observed sievingcoefficient is smaller than that expected from a simple theoretical calculation[5–7].

Dialysis membranes are classified according to their ultrafiltration coeffi-cient [8] and solute sieving profile [9] as high flux and low flux (fig. 3). In clin-ical practice, membranes are incorporated into specific devices designed tooptimize the performance of the membrane itself. These devices may either bedesigned as dialyzers, working prevalently in diffusion with a countercurrentflux of blood and dialysate, or as hemofilters, working prevalently in convec-tion. Improvements in membrane design have allowed diffusive and convectivemass transport to be combined, leading to therapies (high flux dialysis andhemodiafiltration) in which the advantages of both mechanisms are signifi-cantly enhanced.

CRRT Techniques

Several techniques are today available in the spectrum of CRRT.Techniques may differ in terms of vascular access and extracorporeal circuit

Fig. 3. Characteristics of dialysis membranes.

Ultr

afilt

ratio

n (m

l/min

)

Transmembrane pressure (mmHg)

Low-flux membraneKf�6ml/h/mmHg�m2

High-flux membraneKf�40ml/h/mmHg�m2

Sie

ving

coe

ffic

ient

0.00

Molecular weight (daltons)

10 102 103 104 105

1.00

0.80

0.60

0.40

0.20

High flux

Low flux

Clark/Ronco 268

design, frequency and intensity of treatment, predominant mechanism of trans-port utilized and type of membrane [10]. The following description is based pri-marily on the operational parameters normally employed and the targetefficiency with respect to solute and fluid control.

Slow Continuous Ultrafiltration (SCUF)In the intensive care unit, SCUF is typically employed for 24 h/day, but

may also be applied during some portion of the day. The treatment is carried outwith high flux membranes and the objective is to achieve volume control inpatients with severe, diuretic-resistant volume overload [11, 12]. Relative tohemofiltration, low filtration rates (approximately 2–5 ml/min) are required. Assuch, filters of relatively small surface area and low blood flow rates can beemployed. Machines used for this therapy require an ultrafiltration control sys-tem to prevent excessive ultrafiltration. Although very effective for volumereduction, the low filtration rates used and lack of substitution fluids (seebelow) render this therapy ineffective as a blood purification modality. Extra-corporeal ultrafiltration is now being carefully assessed as an adjunctive ther-apy for patients with refractory congestive heart failure (see below).

Continuous Venovenous Hemofiltration (CVVH)Continuous hemofiltration is normally applied for an extended period of

time up to several weeks [13]. The technique utilizes high flux membranes andthe prevalent mechanism of solute transport is convection. Ultrafiltration ratesin excess of the amount required for volume control are prescribed, requiringpartial or total replacement of ultrafiltrate losses with fresh substitution fluid.Blood flow is regulated by a pump and systems for ultrafiltration and reinfu-sion control are generally utilized. Different machines use either volumetriccontrol systems or volumetric pumps regulated by one or multiple scales.Heparin is infused in the arterial line to prevent clotting of the circuit. Thereplacement solution can either be infused before the filter (pre-dilution) orafter the filter (post-dilution) [14]. In the first case, ultrafiltration must be rel-atively increased to maintain the same efficiency observed in post-dilutionmode [15, 16]. Since the ultrafiltrate is replaced by toxin-free substitution fluid,the treatment is appropriate for both blood purification and volume control.

As noted above, the primary solute removal mechanism in HF is convec-tion. In post-dilution HF, the relationship between solute clearance and ultrafil-tration rate is relatively straightforward. In this situation, solute clearance isdetermined primarily by and related directly to the solute’s sieving coefficientand the ultrafiltration rate. Consequently, as discussed below, the concept thatultrafiltrate volume can be assumed to be a surrogate for treatment dose is rea-sonable, particularly in the context of small solute clearance [17]. However,


post-dilution HF is limited inherently by the attainable blood flow rate and thepatient’s hematocrit (Hct). More specifically, the ratio of the ultrafiltration rateto the plasma flow rate delivered to the filter, termed the filtration fraction (FF),is the limiting factor:

(5)

Previous studies suggest FF values �35% in post-dilution may be undesirabledue to hemoconcentration-related effects on filter performance [18, 19].

Pre-dilution HF avoids the post-dilution hemoconcentration-related effectson hemofilter performance and is being increasingly used for CVVH therapy.However, the above mass transfer benefits must be weighed against the pre-dictable dilution-induced reduction in plasma solute concentrations, one of thedriving forces for convective solute removal. The extent to which this reductionoccurs is determined mainly by the ratio of the replacement fluid rate to theblood flow rate [20]. However, the ultrafiltration rate afforded by such a highreplacement fluid rate allows the dilution-related loss of efficiency to be over-come. For reasons described above, the relationship between clearance andultrafiltration rate may not be as predictable in pre-dilution, relative to the caseof post-dilution.

Until recently, a standard ultrafiltration rate employed in CVVH has been1–2 liters/h. However, based on recent clinical outcome data indicating survivalis improved by the use of higher ultrafiltration rates, ‘high-volume’ hemofiltra-tion is increasing being applied in many intensive care units (see below).

Continuous Venovenous Hemodialysis (CVVHD)Continuous hemodialysis is a treatment carried out over an extended

period of time with a pump-driven circuit. Because of the nature of the mem-brane and the gradient provided by the dialysate, the prevalent mechanism ofsolute transport in this technique is diffusion. As such, either a low-flux or high-flux filter can be used [21, 22], although the latter is typically prescribed.Because of the nature of the membrane and the gradient provided by thedialysate, the prevalent mechanism of solute transport in this technique is dif-fusion. Ultrafiltration is obtained exactly in the range of values adequate tomaintain patient’s fluid control without requirement of fluid reinfusion. For thisreason, dialyzers with higher surface area and modified cellulosic membranessuch as triacetate could be effectively used [23]. Even when CVVHD is per-formed with a relatively small surface area filter (�0.5 m2), the use of relativelylow dialysate flow rates (�25 ml/min) results in saturation of the effluentdialysate with respect to small solutes [24]. This saturation phenomenon can bepreserved at even higher dialysate flow rates (�35 ml/min) by use of higher

FFQ

Q (1 Hct)uf

b

��

Clark/Ronco 270

surface area filters [25]. As dialysate flow rate is increased, although saturationof the effluent dialysate is not complete, an increase in small-molecular-weightsolute clearances is nevertheless achieved [26]. In most cases, designatedmachines must be used to control inlet and outlet dialysate flows and to achievethe desired volume of ultrafiltration.

For a given set of operating parameters (including filter effluent rate), theclearance of large-molecular-weight solutes in CVVHD is less than thatachieved with CVVH [27, 28]. The primarily diffusive nature of CVVHDexplains this finding. However, for typical blood and dialysate flow rates inCVVHD, a pressure profile that promotes enhanced removal of large toxins iscreated when a high flux filter is used [29]. Due to the high water permeabilityof such a filter, the spontaneous filtration would be much greater than thepatient’s desired net volume removal. However, the ultrafiltration control mech-anism present in most contemporary CRRT devices prevents such ‘runaway’filtration by applying positive pressure in the dialysate compartment [30]. In acertain segment of the filter at the venous end, the dialysate compartment pres-sure is actually greater than blood compartment pressure. Thus, an ‘internalfiltration’ circuit, characterized by a high rate of blood-to-dialysate filtration inthe arterial end of the device and a nearly equal rate of dialysate-to-bloodfiltration in the venous end, routinely occurs during CVVHD. This phenome-non, which has also been recognized in high flux dialysis in the ESRD setting[31], promotes convective solute removal. Thus, despite its diffusive founda-tion, convective forces also are operative in CVVHD.

Continuous Venovenous Hemodiafiltration (CVVHDF)Continuous hemodiafiltration requires a high flux hemodiafilter and oper-

ates combining the principles of hemodialysis and hemofiltration [32]. As such,this therapy may allow for an optimal combination of diffusion and convectionto provide clearances over a very broad range of solutes. Dialysate is circulatedin countercurrent mode to blood and at the same time ultrafiltration is obtainedin excess of the desired fluid loss from the patient. This is totally or partiallyreplaced with substitution fluid either in pre-dilution or in post-dilution mode.Recent machines allow a combination of pre- and post-dilution aiming at com-bining the advantages of both modalities. Information from the chronic hemodi-afiltration literature suggests that a combination of pre- and post-dilution may beoptimal [33]. This may also be the case for CVVHDF although this has not beenassessed carefully. The optimal balance is most likely dictated by the specific setof CVVHDF operating conditions, namely blood flow rate, dialysate flow rate,ultrafiltration rate, and filter type.

The specific manner in which diffusion and convection interact in con-tinuous hemodiafiltration differs significantly from the situation when this


treatment is applied in the ESRD setting. In the latter situation, diffusion andconvection interact in such a manner that total solute removal is significantlyless than what is expected if the individual components are simply addedtogether. This phenomenon is explained in the following way. Diffusiveremoval results in a decrease in solute concentration in the blood compartmentalong the axial length (i.e. from blood inlet to blood outlet) of the hemodia-lyzer/hemofilter. As convective solute removal is directly proportional to theblood compartment concentration, convective solute removal decreases as afunction of this axial concentration gradient. On the other hand, hemoconcen-tration resulting from ultrafiltration of plasma water causes a progressiveincrease in plasma protein concentration and hematocrit along the axial lengthof the filter. This hemoconcentration and resultant hyperviscosity causes anincrease in diffusive mass transfer resistance and a decrease in solute transportby this mechanism.

Due to the markedly lower flow rates used and clearances obtained inCVVHDF, the effect of simultaneous diffusion and convection on overallsolute removal is quite different. Therefore, the small solute concentration gra-dient along the axial length of the filter (i.e. extraction) is minimal comparedto that which is seen in chronic hemodiafiltration, in which extraction ratios of50% or more are the norm. Thus, the minimal diffusion-related change insmall solute concentrations along the filter allows any additional clearancerelated to convection to be simply additive to the diffusive component. Thishas been demonstrated clearly both in continuous hemodialysis [22] andhemodiafiltration [25].


The evolution of CRRT has been accompanied by a parallel evolution inthe related technology. A series of double lumen catheters has been developedin order to achieve higher blood flows with lower flow resistance and reducedrisk of recirculation [34]. Double lumen catheters are in some cases substitutedby twin separate catheters in order to maximize blood flow and preventunwanted recirculation. Several machines have incorporated the heparin pumpor other systems for regional heparinization and citrate anticoagulation. Themost common anticoagulant remains heparin, although high blood flows andpre-dilution techniques allow for a smooth conduction of CRRT without anyanticoagulant in patients at risk. Regional heparinization and the use of citrateis mostly reserved for special cases as it is for low-molecular-weight heparinand prostacyclin. In recent years, catheters, blood lines and filters with heparinbound on the inner surface have been developed. Their use, however, is still

Clark/Ronco 272

experimental and requires further evaluation. Dialyzers with different mem-branes have been created making possible to choose among a variety of mem-brane materials. Membranes with different porosity and ultrafiltrationcoefficients are available. There is a tendency to increase the filter surface areasince the pumped circulation can operate at higher blood flows compared toarteriovenous circuits. A series of on-line monitoring techniques are todayunder evaluation including blood volume monitoring and blood temperaturemonitoring [35–36]. Finally, a great deal of development has taken place in theoperator interface of the CRRT machines [37]. Most of these machines areequipped with large color screens and step-by-step guidelines to prime the cir-cuit and run the treatment smoothly and effectively.

Clinical Indications for CRRT

The different techniques described in this chapter have been used in dif-ferent settings and different clinical conditions. In those departments in whichdialysis machines or CRRT equipment are not available, arteriovenous treat-ments still represent an important resource. In most cases however one or moremachines or simply different types or adaptive technologies are available andvenovenous pump-driven therapies can be carried out. In these cases, the indi-cation for one or another technique is based on the knowledge of the capabili-ties of each technique and the clinical objectives that the clinician seeks toachieve. If small molecule clearance is the main target, there is no point in usingexpensive high-flux membranes. On the other hand, if a wider spectrum of mol-ecules is to be removed, convective therapies or combined diffusive-convectivetherapies should be implemented. Combination therapies are today more andmore utilized especially in the setting of multiple organ failure and acute renalfailure complicated by sepsis.

When blood purification is the main target, the efficiency of continuousrenal replacement therapies seems to be unparalleled. While intermittenthemodialysis in fact presents the typical limitations imposed by the doublepool kinetics of most molecules [38], CRRT, in spite of a lower clearance,presents an improved removal due to the continuous action and the steady con-centration of the solutes in blood [39]. Recent studies have demonstrated thatCRRT can improve survival in acute renal failure patients if the dose of treat-ment is increased up to 35 ml/h/mm Hg [17]. This observation has been elabo-rated by Gotch [40] who described this level of efficiency as the only oneapproaching the function of the native kidney. Comparing the efficiency ofdifferent therapies is complicated and it required special parameters of calcu-lation such as the standard Kt/V [40]. When comparing hemodialysis and


peritoneal dialysis, weekly standard Kt/V as high as 2–2.5 can be obtained.In CRRT, standard Kt/V can be four or five times higher displaying the enor-mous superiority of continuous therapies. The same concept can be describedfor the correction of acid-base and electrolyte derangement. By performing acontinuous correction, the electrolyte pools can be normalized and so can bedone with the serum concentrations, leading to a stable maintenance of anoptimal homeostatic equilibrium [2]. The effect of slow and continuous waterremoval is the other important advantage of these therapies since a continuousrefilling from the interstitial space can be obtained. Under such circumstances,overhydrated patients or patients with congestive heart failure can be treatedachieving the normalization or the improvement of cardiac filling pressures,preload and afterload, in the absence of dangerous reductions of circulatingblood volume [35].

New Horizons

In the clinical scenario of acute renal failure and the application of CRRT,it is worth underscoring not only the continuous evolution of the techniques butalso the continuous evolution of the possible clinical indications. As an exam-ple, there is increasing evidence that a beneficial effect from the hemodynamicpoint of view can be obtained by CRRT in patients with multiple organ failureand septic shock. Since the explanation for some of the possible benefits seemsto lie on the capacity of these therapies to remove chemical mediators from thepatient’s circulation, new studies and research have been directed towards themechanism of humoral response to sepsis, and the possibility to attenuatethe immunological disequilibrium that seems to characterize the septic patients[41]. Since some effects induced by CRRT could be related to the removal ofproinflammatory mediators, this hypothesis has spurred new interest in theapplication of therapies with increased amount of convection such as high vol-ume hemofiltration, or with membranes characterized by increased sievingcoefficients [42].

High volume hemofiltration is a purely convective therapy which can beperformed with two basic schedules: (a) continuous hemofiltration with a fluidexchange rate �3 liters/h, and (b) hemofiltration performed for some hours[3–6] during the day exchanging 6–8 liters/h while the patient continues a stan-dard CVVH for the rest of the day. In the first case, if the therapy is performedfor 24 h, clearances in the range of 80 liters/day can be obtained. Technicalrequirements for this technique consist especially on the increased blood flowrates and the availability of large volumes of substitution fluid. These therapieshave been shown to produce a beneficial effect on patient’s hemodynamics,

Clark/Ronco 274

with a significant reduction of vasopressor drugs requirement [42]. The tech-nology involved is in most cases borrowed from the chronic hemodialysis set-ting. The large volumes of fluid exchanged may render the treatment somehowimpractical. The new methods for on-line production of substitution fluid mayhowever contribute in the coming future to reduce the costs and the problemsof fluid supply.

Congestive heart failure (CHF) represents another disease state whichmay be benefit from the application of extracorporeal therapy. Currentpharmacotherapy of CHF involves agents which generally have very narrowtherapeutic ranges and are associated with a wide range of adverse effects.Extracorporeal ultrafiltration is increasingly being studied as a CHF adjunc-tive therapy which may decrease the morbidity and possibly mortality associ-ated with CHF and its medical management. Numerous studies [43–50] havecharacterized UF’s specific clinical benefits, which include decreases in car-diac filling pressures and improvements in diuretic responsiveness, hypo-natremia, edema, renal function, and dyspnea. A common element of many ofthese studies has been the ability of UF to ‘reset’ the neurohormonal axis, asevidenced by decreases in plasma norepinephrine, aldosterone, and reninactivity. A relatively common prescription in these studies was an ultrafiltra-tion rate of 300–600 ml/h, administered over a several-hour treatment periodfor consecutive days.

The clinical benefits reported for UF in CHF patients may relate to the dif-ference in the composition of the volume removed by UF vs. diuretics. In UF,the fluid removed is an ultrafiltrate of plasma and, as such, has electrolyte con-centrations that are isotonic with respect to plasma water. On the other hand,urine inherently is hypotonic with respect to plasma water. Therefore, sodiumremoval is significantly greater in ultrafiltrate relative to the same volume ofurine. Moreover, due to the isotonicity of the ultrafiltrate, UF induces no acutechanges in electrolyte concentrations.

Conclusions

Different techniques are today available for the therapy of acute renalfailure in the critically ill patient. Continuous therapies seem to displayimportant advantages in terms of clinical tolerance and blood purificationcapacity [51]. The field is in continuous evolution and prospective random-ized controlled trials will soon prove if there is any significant rationale forthe application of CRRT and derived techniques in patients with sepsis andmultiple organ dysfunction syndromes, beyond the simple indication of acuterenal failure.


References

1 Bellomo R, Ronco C: Acute renal failure in the ICU: Adequacy of dialysis and the case for con-tinuous therapies. Nephrol Dial Transplant 1996;11:424–428.

2 Bellomo R, Ronco C: Continuous versus intermittent renal replacement therapy in the intensivecare unit. Kidney Int 1998;53(suppl 66):S125–S128.

3 Ronco C, Ghezzi P, Bellomo R: New perspective in the treatment of acute renal failure. BloodPurif 1999;17:166–172.

4 Henderson LW: Biophysics of ultrafiltration and hemofiltration; in Jacobs C (ed): Replacement ofRenal Function by Dialysis, ed 4. Dortdrecht, Kluwer Academic Publishers, 1996, pp 114–118.

5 Rockel A, Hertel J, Fiegel P, Abdelhamid S, Panitz N, Walb D: Permeability and secondary mem-brane formation of a high flux polysufone hemofilter. Kidney Int 1986;30:429–432.

6 Langsdorf LJ, Zydney AL: Effect of blood contact on the transport properties of hemodialysismembranes: A two-layer model. Blood Purif 1994;12:292–307.

7 Morti SM, Zydney AL: Protein-membrane interactions during hemodialysis: Effects on solutetransport. ASAIO J 1998;44:319–326.

8 Ronco C: Continuous renal replacement therapies in the treatment of acute renal failure in inten-sive care patients. 1. Theoretical aspects and techniques. Nephrol Dial, Transplant 1994;9(suppl 4):191–200.

9 Clark WR, Ronco C: Determinants of hemodialyzer performance and the effect on clinical out-come. Nephrol Dial Transplant 2001;16(suppl 3):56–60.

10 Clark WR, Ronco C: Renal replacement therapy in acute renal failure: Solute removal mechanismand dose quantification. Kidney Int 1998;53(suppl 66):S133–S137.

11 Silverstein ME, Ford CA, Lysaght MJ, Henderson LW: Treatment of severe fluid overload by ultra-filtration. N Engl J Med 1974;291:747–751.

12 Paganini EP, Nakamoto S: Continuous slow ultrafiltration in oliguric acute renal failure. Trans AmSoc Artif Intern Organs 1980;26:201–204.

13 Macias WL, Mueller BA, Scarim SK, Robinson M, Rudy D: Continuous venovenous hemofiltra-tion: An alternative to continuous arteriovenous hemofiltration and hemodiafiltration in acuterenal failure. Am J Kidney Dis 1991;18:451–458.

14 Henderson LW: Pre- vs. post-dilution hemofiltration. Clin Nephrol 1979;11:120–124.15 Clark WR, Turk JE, Kraus MA, Gao D: Dose determinants in continuous renal replacement ther-

apy. Artif Organs 2003;27:815–820.16 Liao Z, Zhang W, Poh CK, Huang Z, Hardy PA, Kraus MA, Clark WR, Gao D: Kinetic compari-

son of different acute dialysis therapies. Artif Organs 2003;27:802–807.17 Ronco C, Bellomo R, Homel P, Brendolan A, Dan M, Piccinni P, La Greca G: Effects of different

doses in continuous veno-venous haemofiltration on outcomes of acute renal failure: A prospec-tive randomised trial. Lancet 2000;356:26–30.

18 Grootendorst F, Van Bommel E, Van De Hoven B: Light-volume hemofiltration improves hemo-dynamics of endotoxin-induced shock in the pig. Int Care Med 1992;18:235–240.

19 Grootendorst F, Van Bommel E, Van Leenooed L, Van De Hoven B: Infusion of ultrafiltrate fromendotoxemic pigs depresses myocardial performance in normal pigs. J Crit Care 1993;8:161–169.

20 Ahuja A, Rodby R, Huang Z, Gao D, Zhang W, Clark WR: Effect of pre-dilution replacement fluidadministration on solute clearance in high-volume hemofiltration (abstract). J Am Soc Nephrol2003;14:734A.

21 Sigler MH, Teehan BP: Solute transport in continous hemodialysis: A new treatment for acuterenal failure. Kidney Int 1987;32:562–571.

22 Relton S, Greenberg A, Palevsky P: Dialysate and blood flow dependence of diffusive solute clear-ance during CVVHD. ASAIO J 1992;38:691–696.

23 Ronco C, Ghezzi PM, Hoenich N, Delfino PG: Membranes and filters for hemodialysis. ContrNephrol. Basel, Karger, 2000.

24 Bonnardeaux A, Pichette V, Ouimet D, Geadah D, Habel F, Cardinal J: Solute clearances with highdialysate flow rates and glucose absorption from the dialysate in continuous arteriovenoushemodialysis. Am J Kidney Dis 1992;19:31–38.

Clark/Ronco 276

25 Brunet S, Leblanc M, Geadah D, Parent D, Corteau S, Cardinal J: Diffusive and convective soluteclearance during continuous renal replacement therapy at various dialysate and ultrafiltration flowrates. Am J Kidney Dis 1999;34:486–492.

26 Schlaeper C, Amerling R, Manns M, Levin NW: High clearance continuous renal replacementtherapy with a modified dialysis machine. Kidney Int Suppl 1999;72:S20–S23.

27 Clark WR, Ronco C: CRRT efficiency and efficacy in relation to solute size. Kidney Int 1999;56(suppl 72):S3–S7.

28 Troyanov S, Cardinal J, Geadah D, Parent D, Courteau S, Caron S, Leblanc M: Solute clearancesduring continuous venvenous haemofiltration at various ultrafiltration flow rates using Multiflow-100 and HF1000 filters. Nephrol Dial Transplant 2003;18:961–966.

29 Ronco C: Continuous renal replacement therapies for the treatment of acute renal failure in inten-sive care patients. Clin Nephrol 1993;4:187–198.

30 Ronco C, Ghezzi PM, Hoenich N, Delfino PG: Membranes and filters for hemodialysis. ContrNephrol. Basel, Karger, 2000.

31 Clark WR, Gao D: Low-molecular weight proteins in end-stage renal disease: Potential toxicityand dialytic removal mechanisms. J Am Soc Nephrol 2002;13:S41–S47.

32 Mehta RL: Therapeutic alternatives to renal replacement for critically ill patients in acute renalfailure. Semin Nephrol 1994;14:64–82.

33 Pedrini LA, de Christofaro V, Pagliari B, Sama F: Mixed predilution and post-dilution onlinehemodiafiltration compared with traditional infusion modes. Kidney Int 2000;58:2155–2165.

34 Ronco C, Bellomo R: Critical Care Nephrology. Dortrecht, Kluwer Academic Publishers, 1998.35 Ronco C, Brendolan A, Bellomo R: On-Line monitoring in continuous renal replacement thera-

pies. Kidney Int 1999;56(suppl 72):S8–S14.36 Rahmati S, Ronco F, Spittle M, Morris AT, Schlaeper C, Rosales L, Kaufman A, Amerling R,

Ronco C, Levin NW: Validation of the blood temperature monitor for extracorporeal thermalenergy balance during in vitro continuous hemodialysis. Blood Purif 2001;19:245–250.

37 Ronco C, Brendolan A, Bellomo R: Current technology for continuous renal replacement therapies;in Ronco C, Bellomo R (eds): Critical Care Nephrology. Dortrecht, Kluwer Academic Publishers,1998, pp 1269–1308.

38 Clark WR, Leypoldt JK, Henderson LW, Mueller BA, Scott MK, Vonesh EF: Quantifying theeffect of changes in the hemodialysis prescription on effective solute removal with a mathemati-cal model. J Am Soc Nephrol 1999;10:601–610.

39 Clark WR, Mueller BA, Kraus MA, Macias WL: Extracorporeal therapy requirements for patientswith acute renal failure. J Am Soc Nephrol 1997;8:804–812.

40 Gotch F: The current place of urea kinetic modeling with respect to different dialysis schedules.Nephrol Dial Transplant 1998;13(suppl 6):10–14.

41 Tetta C, Mariano F, Ronco C, Bellomo R: Removal and generation of inflammatory mediators dur-ing continuous renal replacement therapies; in Ronco C, Bellomo R (eds): Critical Care Nephrology.Dortrecht, Kluwer Academic Publishers, 1998, pp 1239–1248.

42 Bellomo R, Baldwin I, Cole L, Ronco C: Preliminary experience with high volume hemofiltra-tion in human septic shock. Kidney Int 1998;53(suppl 66):S182–S185.

43 Marenzi G, Grazi S, Giraldi F, et al: Interrelation of humoral factors, hemodynamics, and fluidand salt metabolism in congestive heart failure: Effect of extracorporeal ultrafiltration. Am J Med1993;94:49–56.

44 Agostoni P, Marenzi G, Pepi M, et al: Isolated ultrafiltration in moderate congestive heart failure.J Am Coll Cardiol 1993;21:424–431.

45 Agostoni P, Marenzi G, Lauri G, et al: Sustained improvement in functional capacity after removalof body fluid with isolated ultrafiltration in chronic cardiac insufficiency: Failure of furosemideto provide the same result. Am J Med 1994;96:191–199.

46 Blake P, Paganini EP: Refractory congestive heart failure: Overview and application of extracor-poreal ultrafiltration. Adv Ren Replace Ther 1996;3:166–173.

47 Canaud B, Leblanc M, Leray-Moragues H, Delmas S, Klouche K, Beraud JJ: Slow continuous anddaily ultrafiltration for refractory congestive heart failure. Nephrol Dial Transplant 1998;13(suppl 4):51–55.


48 Ronco C, Ricci Z, Bellomo R, Bedogni F: Extracorporeal ultrafiltration for the treatment of over-hydration and congestive heart failure. Cardiology 2001;96:155–168.

49 Marenzi G, Lauri G, Grazi M, Assanelli E, Campodonico J, Agostoni P: Circulatory response tofluid overload removal by extracorporeal ultrafiltration in refractory congestive heart failure. J AmColl Cardiol 2001;38:963–968.

50 Ronco C, Bellomo R, Ricci Z: Hemodynamic response to fluid withdrawal in overhydratedpatients treated with intermittent ultrafiltration and slow continuous ultrafiltration: Role of bloodvolume monitoring. Cardiology 2001;96:196–201.

51 Ronco C, Bellomo R: Continuous versus intermittent renal replacement therapy in the treatmentof acute renal failure. Nephrol Dial Transplant 1998;13:79–85.

William R. Clark, MDWishard Hospital/Myers D7111001 West 10th St., Indianapolis, IN 46202 (USA)Tel. 1 317 613 2315 (ext 327), Fax 1 317 613 2317, E-Mail [email protected]


Hybrid Renal Replacement Therapies for Critically Ill Patients

Thomas A. Golper

Vanderbilt University Medical Center, Nashville, Tenn., USA

Despite the fact that in prospective randomized controlled trials continuousrenal replacement therapy (CRRT) was not shown to deliver superior outcomesto intermittent hemodialysis (IHD) [1, 2], many clinicians, including me, con-tinue to prefer CRRTs over IHD for critically ill patients with hemodynamicinstability. The reason for the preference is that our collective clinical experienceshows that fluid removal under continuous operating conditions is less compli-cated than during intermittent therapy. Most patients needing renal replacementtherapy have already undergone fluid resuscitation and are often fluid over-loaded, compromising ventilation, despite being hypotensive. Outcomes havebeen shown to be superior when IHD is applied daily vs. thrice weekly [3] andwhen the ultrafiltration rate (UFR) is �35 ml/kg/h [4].

As currently practiced, many CRRTs require specially prepared sterilesolutions to serve as either filtration substitution fluid (hemofiltration) ordialysate (hemodialysis). Specialized equipment has been developed by severalmanufacturers, but in each case the CRRT machine is more expensive to purchasethan are machines used predominantly for IHD (table 1). Thus, in the absenceof clear outcome superiority for CRRT over IHD, and because there is an enor-mous cost difference, hybrid therapies have emerged [5–7]. This recent work isbased on the work of clinicians who attempted this approach with traditionalhemodialysis equipment [8, 9]. One such name for this hybrid of continuousand intermittent therapy is called slow low efficiency dialysis (SLED).

Cost Analysis

At Vanderbilt University, we utilize Gambro’s PRISMA™ machine toperform hemodiafiltration with a routine delivered clearance of approximately

Hybrid Renal Replacement Therapies 279

Table 1. Cost comparison of SLED with Fresenius K™ machine and CRRT withGambro PRISMA™ in USD

Fresenius K 20,000.00 PRISMA 25,000.00

Portable R.O. 5,000.00

F50 dialyzer (each) 17.00 Prisma circuit 170.00includes dialyzer

Medisystem™ bloodlines 2.90

Liquid bicarbonate (4 gal) 12.32 Baxter pre-mixed dialysate 324.00(USD 27.00 per 5 liters)@ 2 liters/h use over 24 h, or

Citrasate™ 4 gal 27.00 Gambro Prismasate™ with 420.00or bicarbonate (USD 35.00 8 gallons 54.00 per 5 liters) @ 2 liters/h

use over 24 h

NS prime (2 liters) 1.26 Saline prime (4 liters) 2.52Heparin (per day) 15.00

Diasafe™ ultrafilter for USD 4/dayFresenius K USD 50.00/filter over wholereplacement every 7 weeks � 4 yearmachines USD 1,485.72 per year � Biomed labor for installation

Monthly AAMI standard USD 6/dayculture and endotoxin for over wholeeach Fresenius K (4) and yeareach portable R.O. (4) plus bi-annual AAMI standard chemical analysis per R.O. USD 2,144.00 per year � follow-up samples as needed

Supply cost 1 SLED 97.48 Supply cost 1 CRRT setup 511.52setup and projected bicarbonate pre-mixed orrun 14 h (8 gallons dialysate and average 607.52of Citrasate™) 24-hour run per circuit

Labor description

Dialysis RN setup and monitor Dialysis RN setup and monitorover first hour then check over first hour then checkminimally three times/day; minimally three times/day;provide 24/7 coverage for provide 24/7 coverage fortroubleshooting and troubleshooting and newnew setups when needed setups due to clotting

Golper 280

Table 1 continued

ICU RN monitor and ICU RN monitors andcalculate UF each hour; calculates UF each hour;basic troubleshooting change dialysate bagsalarms and terminate every 2.5 h and emptyif needed; notify dialysis staff. effluent bag every 2.5 h;

basic troubleshooting alarms and terminates if needed; notify dialysis staff.

2 liters/h (half by dialysis diffusion, half by hemofiltration convective).Alternatively, SLED is employed using the Fresenius K™ hemodialysismachine, with dialysis operating for 6–12 h on a daily basis. Nursing intensityvaries by therapy with the greatest consumption of Intensive Care Unit (ICU)nurses’ time being changing the substitution fluid bags, dialysate bags andeffluent bags on the PRISMA™. For SLED the ICU nurses simply adjust theUFR on an hourly basis, which takes 15–30 s. A more detailed cost analysis isshown in table 1.

Clearances and Dose Versatility

Ronco’s landmark study demonstrating the survival advantage of clearance�35 ml/kg/h of convective clearance has set the standard certainly for purelyconvective therapies [4]. This is approximately a urea clearance of 40 ml/min.The lowest operating conditions of SLED with a dialysate flow rate (Qd) of100 ml/min generate a urea clearance of �60 ml/min [10]. CRRTs run approx-imately 22 of 24 h and the hybrid therapies such as SLED are generally appliedfor �6 h [11]. Depending on the dose needed, SLED and hybrid therapies canrun longer and even continuously. Since many critically ill patients require pro-cedures, often outside of the ICU, the starting and stopping of renal replace-ment therapy is common. The hybrids lend themselves better to this disruptionbecause the disposable components are considerably cheaper. If SLED is per-formed intermittently, then daily application can be considered more effica-cious, extrapolating from the IHD experience [3].

At a Qd of 100 ml/min, a jug of dialysate concentrate lasts �16 h. Obviouslyas Qd is increased the concentrate is depleted more rapidly. We utilize high fluxpolysulfone modest surface area hemodialyzer (e.g. Fresenius F-50™) forSLED treatments. The K™ machine has an internal filter which filters the


dialysate water prior to its reaching the hemodialyzer blood-dialysate interface.Prior to that stage a portable reverse-osmosis device is utilized to process thehospital water. So the dialysate is essentially ultrapure for SLED treatments, arecommendation with European origins [12]. By adjusting blood flow rate,dialysate flow rate and treatment length, a large operational diversity and poten-tial for dose delivery are created. The dialysate can vary in concentration ofelectrolytes which is particularly relevant as K�, HCO3

� and Ca2� need adjust-ment. As citrate is utilized more often, alkalosis can be easily rectified byreducing the dialysate HCO3

� and, depending on the need for Ca2� infusion,dialysate can contain low levels or be absent of Ca2�.

Nursing Time

Vanderbilt’s ICU nurses spend at least an hourly average of 15 min whenperforming PRISMA™ CRRT. This time is spent handling fluid bags. ForSLED the time commitment is to perform any calculations of intakes and out-puts (which are done regardless of the renal replacement therapy) and then toset the hourly UFR on the K machine.

Dialysis nurses set up both SLED and PRISMA™ machines and dispos-ables. Dialysis nurses trouble-shoot both therapies during their hours of opera-tion. There is limited use of SLED during hours that dialysis nurses are out ofthe hospital, but upon completion of additional in-service training, this willchange, such that ICU nurses can discontinue an inoperable SLED treatment inthe absence of a dialysis nurse.

Future Issues in Hybrid Therapies

Hybrid therapies as currently used are predominantly diffusion based. Aswe develop cleaner water systems, there is no reason that on-line hemodiafil-tration substitution fluid and dialysate are not utilized to improve the convec-tive component [13]. The use of the high flux dialyzers already greatlyenhances the convective contribution to total clearance. If a greater delivereddialysis dose does improve survival [3, 4, 14], hybrids will be the preferreddelivery system.

Regional anticoagulation regimens are constantly improving. The bloodpump of dialysis machines can generate much higher flow rates can CRRTmachines. This may lead to less aggressive anticoagulation, especiallybecause the filtration fraction is minimal (e.g. �4%) in the hybrid therapies.Furthermore, the greater blood flow may allow for Ca2� reinfusion via stop cock

Golper 282

specifically at the blood access return site, obviating the need for a central Ca2�

infusion line. As mentioned above, regional citrate regimens are evolving that uti-lize the enormous small solute clearance capacity of SLED to correct potentialside effects of citrate anticoagulation, such as hypernatremia, hyperglycemia(if Anticoagulant Citrate Dextrose-A™ is used), alkalosis, or hypocalcemia.

Conclusion

Not only are the hybrid therapies here to stay, but their advantages andversatility will lead them to be the mainstay renal replacement therapy for thecritically ill before the end of this decade.

References

1 Mehta RL, McDonald B, Gabbai FB, Pahl M, Pascual MTA, Farkas A, Kaplan RMl: A randomizedclinical trial of continuous versus intermittent dialysis for acute renal failure. Kidney Int 2001;60:1154–1163.

2 Misset B, Timsit J-F, Chevret S, Renaud B, Tamion F, Carlet J: A randomized cross-over compar-ison of the hemodynamic response to intermittent hemodialysis and continuous hemofiltration inICU patients with acute renal failure. Intens Care Med 1996;22:742–746.

3 Schiffl H, Lang S, Fischer R: Daily hemodialysis and the outcomes of acute renal failure. N EnglJ Med 2002;346:305–310.

4 Ronco C, Bellomo R, Homel P, Brendolan A, Dan M, Piccinni P, La Greca GI: Effects of differ-ent doses in continuous veno-venous haemofiltration on outcomes of acute renal failure:A prospective randomised trial. Lancet 2000;356:26–30.

5 Schlaeper C, Amerling R, Manns M, Levin NW: High clearance continuous renal replacementtherapy with a modified dialysis machine. Kidney Int 1999;56(suppl 72):S20–S23.

6 Kumar V, Craig M, Depner T, Yeun J: Extended daily dialysis: A new approach to renal replace-ment therapy for acute renal failure in the intensive care unit. Am J Kidney Dis 2000;36:294–300.

7 Marshall M, Golper TA, Shaver MJ, Chatoth DK: Sustained low efficiency dialysis for criticallyill patients requiring renal replacement therapy: Clinical experience. Kidney Int 2001;60:777–785.

8 Hombrouckx R, Bogaert AM, Leroy F, De Vos JY, Larno L: Go-slow dialysis instead of continu-ous arteriovenous hemofiltration. Contrib Nephrol. Basel, Karger, 1991, vol 93, pp 149–151.

9 Kihara M, Ikeda Y, Shibata K, Masumori S, Fujita H, Ebira H, Toya Y, Takagi N, Shionoiri H,Umemura S, Ishii M: Slow hemodialysis performed during the day in managing renal failure incritically ill patients. Nephron 1994;67:36–41.

10 Marshall MR, Golper TA, Shaver MJ, Chatoth DK: Urea kinetics during sustained low-efficiencydialysis. Am J Kidney Dis 2002;39:556–570.

11 Frankenfield DC, Reynolds HN, Wiles CE 3rd, Badellino MM, Siegel JH: Urea removal duringcontinuous hemodiafiltration. Crit Care Med 1994;22:407–412.

12 Lonnemann G, Floege J, Kliem V, Brunkhorst R, Koch K: Extended daily veno-venous high fluxhaemodialysis in patients with acute renal failure and multiple organ dysfunction syndrome usinga single path batch dialysis system. Nephrol Dial Transplant 2000;15:1189–1193.

13 Marshall MR, Ma T, Galler D, Rankin APN, Williams AB: Sustained low-efficiency daily diafil-tration (SLEDD-f) for critically ill patients requiring renal replacement therapy: Towards an ade-quate therapy. Nephrol Dial Transplant 2004;in press.


14 Paganini EP, Tapolyai M, Goormastic M, Halstenberg W, Kozlowski L, Leblanc M, Lee JC,Moreno L, Sakal K: Establishing a dialysis therapy/patient outcome link in intensive care unitacute dialysis for patients with acute renal failure. Am J Kidney Dis 1996;28(suppl 3):S81–S89.

Thomas A. Golper, MDS3301 MCN, Vanderbilt University Medical Center21st Avenue South, Nashville, TN, 37232 (USA)Tel. �1 615 343 2220, Fax �1 615 322 8653, E-Mail [email protected]


Pediatric Acute Renal Failure:Demographics and Treatment

Stuart L. Goldstein

Baylor College of Medicine, Medical Director, Renal Dialysis Unit, Texas Children’s Hospital, Houston, Tex., USA

Provision of appropriate renal replacement therapy for pediatric patientswith acute renal failure (ARF) requires special considerations not commonlyencountered in the care of adult patients. Pediatric patients with acute renalfailure may range in weight from a 1.5-kg neonate to a 200-kg young adult. Inaddition, disease states that may require acute renal replacement therapy in theabsence of significant renal dysfunction, such as inborn errors of metabolismor post-operative care of an infant with congenital cardiac defects, are moreprevalent in the pediatric setting. Optimal care for the pediatric patient requir-ing renal replacement therapy demands an understanding of the causes andpatterns of pediatric ARF and multi-organ dysfunction syndrome (MODS) andrecognition of the local expertise with respect to the personnel and equipmentresources. The aim of this paper is to review the pediatric specific causes neces-sitating renal replacement therapy provision with an emphasis on emergingpractice patterns with respect to modality and the timing of treatment.

Pediatric ARF Epidemiology

Advancements and improvements in care for critically ill neonates, infantswith congenital cardiac disease, and children with bone marrow and solid organtransplantation has lead to a dramatic broadening of pediatric ARF epidemiol-ogy. While multicenter epidemiological pediatric ARF data do not exist, singlecenter data from the 1980s report hemolytic uremic syndrome, other primaryrenal causes, sepsis and burns [1, 2] as the most prevalent causes leading topediatric ARF. More recent single center data detail the underlying causes ofpediatric ARF in large cohorts of children and demonstrate an epidemiological

Pediatric Acute Renal Failure 285

shift where ARF is more often a co-morbidity of another underlying disease orsystemic process. Bunchman et al. [3] reported data from 226 children withARF treated with renal replacement therapy with the most common causesbeing congenital heart disease, ATN and sepsis.

Transition from the use of adaptive continuous renal replacement therapy(CRRT) equipment [4, 5] to production of hemofiltration machines withvolumetric control allowing for accurate ultrafiltration flows has likewise lead toa change in pediatric renal replacement therapy modality prevalence patterns.Accurate ultrafiltration (UF) and blood flow rates are crucial for pediatric CRRTsince the extracorporeal circuit volume can comprise more than 15% of a smallpediatric patients’ total blood volume and small UF inaccuracies may represent alarge percentage of a small pediatric patient’s total body water. Polls of US pedi-atric nephrologists demonstrate increased CRRT use over peritoneal dialysis asthe preferred modality for treating pediatric ARF [6, 7]. In 1995, 45% of pediatriccenters ranked PD and 18% ranked CRRT as the most common modality used forinitial ARF treatment. In 1999, 31% of centers chose PD versus 36% of centersreported CRRT as their primary initial modality for ARF treatment [7].

Until very recently, most pediatric ARF treatment studies were limited toreview articles [8–10]. Such dearth of pediatric data may result, in part, fromthe only recent changes noted above in the clinical spectrum of pediatric ARFand the increased use of CRRT to treat more critically ill children with ARF.The remainder of this article will focus upon the application of the differentrenal replacement therapy modalities and assessment of the outcome of childrenwith ARF who receive renal replacement therapy.

Specific Pediatric Patient Populations

The Critically Ill Pediatric PatientSurvival rates for critically children with ARF receiving renal replacement

therapy have been fairly consistent from 1978 through 2001. Overall reportedpatient survival ranges from 52 to 58% [2, 3, 11, 12]. In the last decade, survivalrates stratified by renal replacement therapy modality have also been stable;survival rates for patients receiving hemodialysis (73–89%) are higher thanthose receiving PD (49–64%) or CRRT (34–42%) [3, 12, 13].

Understanding the pattern of pediatric multi-organ dysfunction syn-drome (MODS) lends insight into some of the shortcomings and strengths ofpediatric ARF outcome data presented later in the article. As opposed to adultpatients, Proulx et al. [14] demonstrated that children develop severe and life-threatening MODS very early in their ICU course; 87% of children developedthe maximum number of organ failures occurs within 72 h of ICU admission

Goldstein 286

and that children die with MODS very early in ICU course as 88.4% of deathsoccur within 7 days of MODS diagnosis. Thus, methods to quickly identify chil-dren at risk of developing MODS would lead to early and aggressive initiationof supportive measures, including renal replacement therapy to treat or preventARF sequelae, which could conceivably improve pediatric patient outcome.

Unfortunately, many issues plague the pediatric ARF outcome literature,which include a relative lack of prospective study, the mixture RRT modalitywith a lack of modality stratification in subject populations studied and theinconsistent use of methods to control for patient illness severity in outcomeanalysis. A few studies have considered the effect of a clinical variable on out-come. Smoyer et al. [15] reviewed the outcome of 98 infants and children withARF who received either arteriovenous or venovenous CRRT modalities andfound that higher mortality in patients on pressors. Subsequent work byBunchman et al. [3] upheld this finding by showing patient survival was only35% for those requiring pressors versus 89% for those without a pressorrequirement in the course of their ARF treatment. As a result, pressor use hasbeen construed as a surrogate for worse patient illness severity in the pediatricpopulation requiring CRRT.

Few pediatric outcome studies use a standardized scoring system controlfor patient illness severity, which may result from the fact that published dataprovide contradictory conclusions with respect to the utility of various illnessseverity scoring systems in predicting death in pediatric patients with ARF.Faragson et al. [16] demonstrated significant overlap in Pediatric Risk ofMortality (PRISM) scores [17] between surviving and non-surviving childrenwho received intermittent hemodialysis and therefore concluded that PRISMscores should not be used to differentiate patients who would likely or not ben-efit from dialysis initiation. Zobel et al. [18] demonstrated that children whoreceived CRRT with worse illness severity by PRISM score had increased mor-tality, but this study included patients who received both arterio-venous andvenovenous CRRT therapies, and did not stratify by modality.

A more recent study performed at our center [12] examined the outcomeof 22 critically ill children who received only venovenous CRRT modalities andused the PRISM 2 score to control for illness severity at ICU admission andCRRT initiation. Neither mean PRISM scores at the time of PICU admissionnor time of CRRT initiation differed between survivors and non-survivors. Ofthe clinical variables studied (GFR, pressor number, mean airway pressure,patient size or % fluid overload), only the degree of % fluid overload at the timeof CRRT initiation differed between survivors (16.4 � 13.8%) and non-survivors (34.0 � 21.0%, p � 0.03), even when controlled for severity of illnessby PRISM score using a multiple regression model. In addition, we found that75% of non-survivors died within 21 days of ICU admission. We hypothesize


that initiation of CRRT a lesser patient % fluid overload may allow for moreoptimal nutrition and blood product provision without the further accumulationof additional fluid or catabolic waste products.

Our data coupled with the predilection for early multi-organ system failureand death in critically ill children with acute renal failure argues for early andaggressive initiation of CRRT. Although mean PRISM scores were no differentbetween survivors and non-survivors, controlling for patient illness severityusing PRISM scores was essential to mitigating concerns that the patients whoreceived more fluid prior to CRRT initiation were more ill, and therefore had ahigher risk of mortality. Further study of a much larger cohort, using PRISM orother illness severity scoring systems, is clearly warranted to substantiate thefindings of our relatively small study.

InfantsInfants and neonates with ARF present unique problems for renal replace-

ment therapy provision. As noted earlier, delivery of hemodialysis or CRRT tothese small patients entails a significant portion of their blood volume to bepumped through the extracorporeal circuit. Therefore, extracorporeal circuitvolumes that comprise more than 10–15% of patient blood volume should beprimed with whole blood to prevent hypotension and anemia. Since the primevolume is not discarded, it is important to not re-infuse the blood into thepatient at the end of the treatment in order to prevent volume overload andhypertension.

Acute peritoneal dialysis requires much less technical expertise, expense andequipment compared to intermittent hemodialysis and CRRT. PD catheters canbe placed quickly and easily. Initial dwell volumes should be limited to 10 cm3/kgof patient body weight in order to minimize intra-abdominal pressure and poten-tial for fluid leakage along the catheter tunnel. Although PD may deliver lessefficient solute removal than hemodialysis or CRRT, its relative simplicity andminimal associated side effects allow for renal replacement therapy provision insettings lacking pediatric dialysis specific support and personnel.

CRRT has been prescribed since the mid-1980s for treatment of ARF incritically ill infants [19, 20]. The first CRRT modalities were arteriovenous inconfiguration, since the extracorporeal volumes were small in these circuits andultrafiltration was driven by patient perfusion pressure, thereby reducing the riskof hypotension from too much ultrafiltration. As mentioned before, introductionof more accurate machines with volumetric control has increased venovenousmodality CRRT in pediatric patients, including neonates and infants. Zobel et al.[20] noted that technical problems occurred only with CVVH in an early neona-tal outcome 1991 study reporting patients who received either CAVH or CVVH.Symons et al. [21] reported data from a more recent retrospective multi-center

Goldstein 288

study evaluating the CVVH course for 90 infants less than 10 kg from 1993through 2001, which demonstrate very few technical complications using newerCVVH machinery. Infant for survival for patients receiving CRRT has also beenconsistent over the past decade at 35–38%, which is similar to survival ratesnoted above for older pediatric patients [20, 21], although patients less than 3 kgexhibited a trend toward worse survival (24%) when compared to infants largerthan 3 kg (41%) [21].

Congenital Heart DiseaseInfants with acute renal failure after corrective congenital heart surgery

comprise a well-studied cohort [22–28]. These patients represent a nearlyunique group of pediatric patients in that the timing of the event leading toacute renal failure, namely cardiopulmonary bypass (CPB), is precisely known.In this sense, children undergoing cardiopulmonary bypass are akin to adultsreceiving nephrotoxic radiologic contrast or emergent surgery for aorticaneurysms; they all provide an opportunity to follow the time course of ARFfrom beginning to end in patients without significant underlying renal disease.

The incidence of infant ARF after CPB ranges from 2.7 to 5.3% withsurvival rates ranging from 21 to 70% [22, 23, 26]. Risk factors for mortalityinclude increasing underlying complexity of the congenital heart disease andpoor cardiac function [22, 26]. A recent trend toward providing PD therapy ear-lier in the post-CPB course has been reported, with one study of 20 patientsdemonstrating 80% patient survival [27]. While improved survival with earlyPD initiation may result from prevention of fluid overload, some posit [28]improved survival with early PD initiation results from increased clearance ofCPB-induced pro-inflammatory cytokines, although further study is required tosupport this hypothesis. In our center, patients with underlying an underlyingdiagnosis of hypoplastic left heart syndrome, transposition of the great arteriesor anomalous pulmonary venous return receive PD immediately postopera-tively in order to prevent fluid accumulation.

Future Studies

The recent epidemiological pediatric ARF data presented in this paper hasdemonstrated the need and laid the groundwork for future pediatric prospectivestudy. Since pediatric ARF is relatively rare, multicenter study will be requiredto enroll a sufficient patient number for appropriate statistical analysis. Currently,a group of 10 US pediatric centers, the Prospective Pediatric Continuous RenalReplacement Therapy Registry Group (ppCRRT Registry) is prospectivelygathering data with respect to critically ill children who receive CRRT.


The ppCRRT Registry aims to provide insight into the potential clinical factorsthat affect pediatric patient outcome, compare the efficacy of different anti-coagulation protocols, assess specialized patient populations including patientswith metabolic disorders and bone marrow transplantation, and evaluate theefficacy and effects of CRRT cytokine removal. At the current time, ppCRRTRegistry data comprise information from 180 patients representing over20,000 h of CRRT time. Future endeavors will include assessment of CRRTpharmacokinetics, broadening the scope to include all pediatric patients withARF and performing prospective randomized trials to evaluate the effect ofCRRT dose and modality upon patient outcome.

References

1 Lattouf OM, Ricketts RR: Peritoneal dialysis in infants and children. Am Surg 1986;52:66–69.2 Williams DM, Sreedhar SS, Mickell JJ, Chan JC: Acute kidney failure: A pediatric experience

over 20 years. Arch Pediatr Adolesc Med 2002;156:893–900.3 Bunchman TE, McBryde KD, Mottes TE, Gardner JJ, Maxvold NJ, Brophy PD: Pediatric acute

renal failure: Outcome by modality and disease. Pediatr Nephrol 2001;16:1067–1071.4 Bunchman TE, Donckerwolcke RA: Continuous arterial-venous diahemofiltration and continuous

veno-venous diahemofiltration in infants and children. Pediatr Nephrol 1994;8:96–102.5 Bunchman TE, Maxvold NJ, Kershaw DB, Sedman AB, Custer JR: Continuous venovenous

hemodiafiltration in infants and children. Am J Kidney Dis 1995;25:17–21.6 Belsha CW, Kohaut EC, Warady BA: Dialytic management of childhood acute renal failure: A sur-

vey of North American pediatric nephrologists. Pediatr Nephrol 1995;9:361–363.7 Warady BA, Bunchman T: Dialysis therapy for children with acute renal failure: Survey results.

Pediatr Nephrol 2000;15:11–13.8 Andreoli SP: Acute renal failure. Curr Opin Pediatr 2002;14:183–188.9 Flynn JT: Choice of dialysis modality for management of pediatric acute renal failure. Pediatr

Nephrol 2002;17:61–69.10 Parekh RS, Bunchman TE: Dialysis support in the pediatric intensive care unit. Adv Ren Replace

Ther 1996;3:326–336.11 Gallego N, Gallego A, Pascual J, Liano F, Estepa R, Ortuno J: Prognosis of children with acute

renal failure: A study of 138 cases. Nephron 1993;64:399–404.12 Goldstein SL, Currier H, Graf C, Cosio CC, Brewer ED, Sachdeva R: Outcome in children receiv-

ing continuous venovenous hemofiltration. Pediatrics 2001;107:1309–1312.13 Maxvold NJ, Smoyer WE, Gardner JJ, Bunchman TE: Management of acute renal failure in the

pediatric patient: Hemofiltration versus hemodialysis. Am J Kidney Dis 1997;30:S84–S88.14 Proulx F, Gauthier M, Nadeau D, Lacroix J, Farrell CA: Timing and predictors of death in pediatric

patients with multiple organ system failure. Crit Care Med 1994;22:1025–1031.15 Smoyer WE, McAdams C, Kaplan BS, Sherbotie JR: Determinants of survival in pediatric

continuous hemofiltration. J Am Soc Nephrol 1995;6:1401–1409.16 Fargason CA, Langman CB: Limitations of the pediatric risk of mortality score in assessing

children with acute renal failure. Pediatr Nephrol 1993;7:703–707.17 Pollack MM, Ruttimann UE, Getson PR: Pediatric risk of mortality (PRISM) score. Crit Care Med

1988;16:1110–1116.18 Zobel G, Kuttnig M, Ring E, Grubbauer HM: Clinical scoring systems in children with continu-

ous extracorporeal renal support. Child Nephrol Urol 1990;10:14–17.19 Ronco C, Brendolan A, Bragantini L, Chiaramonte S, Feriani M, Fabris A, Dell’Aquila R,

La Greca G: Treatment of acute renal failure in newborns by continuous arterio-venous hemo-filtration. Kidney Int 1986;29:908–915.

Goldstein 290

20 Zobel G, Ring E, Kuttnig M, Grubbauer HM: Five years experience with continuous extracorpo-real renal support in paediatric intensive care. Intensive Care Med 1991;17:315–319.

21 Symons JM, Brophy PD, Gregory MJ, McAfee N, Somers MJ, Bunchman TE, Goldstein SL:Continuous renal replacement therapy in children up to 10 kg. Am J Kidney Dis 2003;41:984–989.

22 Book K, Ohqvist G, Bjork VO, Lundberg S, Settergren G: Peritoneal dialysis in infants and chil-dren after open heart surgery. Scand J Thorac Cardiovasc Surg 1982;16:229–233.

23 Rigden SP, Barratt TM, Dillon MJ, De Leval M, Stark J: Acute renal failure complicating car-diopulmonary bypass surgery. Arch Dis Child 1982;57:425–430.

24 Giuffre RM, Tam KH, Williams WW, Freedom RM: Acute renal failure complicating pediatriccardiac surgery: A comparison of survivors and nonsurvivors following acute peritoneal dialysis.Pediatr Cardiol 1992;13:208–213.

25 Fleming F, Bohn D, Edwards H, Cox P, Geary D, McCrindle BW, Williams WG: Renal replace-ment therapy after repair of congenital heart disease in children: A comparison of hemofiltrationand peritoneal dialysis. J Thorac Cardiovasc Surg 1995;109:322–331.

26 Picca S, Principato F, Mazzera E, Corona R, Ferrigno L, Marcelletti C, Rizzoni G: Risks of acuterenal failure after cardiopulmonary bypass surgery in children: A retrospective 10-year case-control study. Nephrol Dial Transplant 1995;10:630–636.

27 Sorof JM, Stromberg D, Brewer ED, Feltes TF, Fraser CD Jr: Early initiation of peritoneal dialy-sis after surgical repair of congenital heart disease. Pediatr Nephrol 1999;13:641–645.

28 Bokesch PM, Kapural MB, Mossad EB, Cavaglia M, Appachi E, Drummond-Webb JJ, Mee RB:Do peritoneal catheters remove pro-inflammatory cytokines after cardiopulmonary bypass inneonates? Ann Thorac Surg 2000;70:639–643.

Stuart L. GoldsteinAssistant Professor of PediatricsTexas Children’s Hospital, 6621 Fannin StreetMail Code 3–2482, Houston, TX 77030 (USA)Tel. �1 832 824 3800, Fax �1 832 825 3889, E-Mail [email protected]


Vascular Access for ExtracorporealRenal Replacement Therapy in theIntensive Care Unit

Bernard Canauda,c, Cédric Formeta, Nathalie Raynala,Laurent Amiguesb, Kada Kloucheb, Hélène Leray-Moraguesa, Jean-Jacques Béraudb

aIntensive Care Unit, Nephrology, bMetabolic Intensive Care, and cRenalResearch and Training Institute Lapeyronie, University Hospital Montpellier,Montpellier, France

Vascular access (VA) management in intensive care unit (ICU) is faced toa twofold problem: first, the installation of an angioaccess that must beadequate for extracorporeal renal replacement therapy (RRT) in the acutesetting; second, the prevention of angioaccess-related morbidity [1].

VA is a basic and common tool required to perform all modalities of extra-corporeal RRT [2]. Central venous dialysis catheters have become the preferredform of vascular access in acute renal failure (ARF) patients where peripheralarteriovenous shunts (e.g. Scribner, Busselmeier) have virtually disappeared.Catheters provide a rapid and easy way for blood access permitting to launchdialysis without delay in critically ill situation. Central venous catheters areused temporarily during the phase of acute renal failure and removed at the timeof renal recovery (usually 4–6 weeks later). Despite significant technicaladvances, temporary dialysis catheters remain a major cause of morbidity andmortality in intensive care units (ICU) and catheter care must be reinforced.

Renal replacement therapy in the ICU relies on two main therapeuticoptions defined by blood flow regimen, periodicity and length of sessions,namely continuous low-flow methods versus intermittent high-flow methods.Indication for continuous or intermittent modality relies on clinical judgmentaccounting for local technical expertise and patient clinical conditions. Venouscatheter design does not differ fundamentally either for intermittent or contin-uous RRT modalities.


Canaud/Formet/Raynal/Amigues/Klouche/Leray-Moragues/Béraud 292

In this chapter, we will address the following issues: (1) implantation andmanagement of venous catheters in the ICU setting, and (2) prevention ofvenous angioaccess-related morbidity.

Insertion and Management of Venous Catheters in the ICU Setting

Venous catheters are clearly the preferred option of blood access in acutepatient. Catheters are easily inserted at the bedside; they are immediatelyusable; they provide adequate blood flow for all forms of dialysis modality.

Catheters and MaterialSeveral types of hemodialysis catheters are available on the market. Catheters

are usually made of polymers (e.g. polyvinyl chloride, polytetrafluoroethylene,polyethylene, polyurethane, silicone elastomer, tecoflex) that translate intospecific catheter characteristics (resistance, softness, hemocompatibility) [3].Catheter design and engineering are important factors contributing to theirperformances (low or high-flow catheter) and complications [4]. Due to theirhemocompatibility properties, polyurethane and silicone material are now themost widely used.

Catheter stiffness is a physical characteristic that dictates the procedurefor percutaneous vein insertion. Catheter rigidity depends mainly on thepolymer nature, the plasticizer content and the extrusion mode. Rigid andsemirigid catheters, usually referred to acute catheters, are easily introducedpercutaneously using the Seldinger method over a metallic guidewire.However, this stiffness leads to a greater risk of vascular perforation andvascular wall lesion. In this respect, semirigid polyurethane catheters are likelyto cause more stenosis and thrombosis to host vein than that softer ones. Now,it must be noted that certain type of catheters made of polyurethane which aresemirigid at ambient temperature soften up at the body temperature, therebylosing their mechanical aggressiveness. Soft silicone catheters, usuallyreferred to chronic or permanent catheters, are more difficult to insertpercutaneously but nevertheless have a major advantage associated to theirtunneling capacity. Several modalities of percutaneous insertion, derived fromthe Seldinger’s method using introducer with sheath dilator (pealable or not)have been substituted to the surgical approach [5]. Most of these siliconerubber catheters are manufactured with a subcutaneous anchoring system(dacron cuff, purse ring suture). The subcutaneous anchoring system is advan-tageous: it stabilizes the catheter, it enlarges the distance from catheter skinemergence to vein insertion site, it promotes tissue scaring around the catheter,

Vascular Access for Extracorporeal RRT in the ICU 293

and finally it provides a physical barrier against bacterial intrusion alongcatheter pathway.

Single- and double-lumen dialysis catheters are available on the market.Single lumen catheters have a single port used alternatively for both inflow andoutflow. Although, two single-lumen catheters (inflow and outflow) inserted inthe same central vein are used in some units, catheters featuring a double-lumendesign (one arterial port and one venous port) are most popularly used (fig. 1).Layout of the lumen and distal ends of catheters may vary considerably fromone type to another. According to venous and arterial port locations, two maintypes of catheter may be designed: one with port sites attached in a double-barreled gun fashion (coaxial, double-D, double-O); the other with independentor separated port sites (Dual catheter, Split catheter) (fig. 2). From a flowperspective, it has been shown that independent catheter lines offered higherand more consistently adequate blood flow than attached ones.

To summarize, one may consider that semirigid double-lumen polyurethanecatheters (acute catheters) are a fairly reasonable choice for a short time use up to2 weeks while soft silicone double catheters or double-lumen catheters (chronicor permanent catheters) are a best indication for prolonged RRT up to 4–6 weeks.

Inflow

Outflow

Outflow

Fig. 1. Acute catheter: Double-lumen polyurethane catheter.

Inflow

Outflow

Double-lumen, PermCath

Double-catheter, DualCath

Double split catheter, SplitCath

Fig. 2. Chronic or permanent catheters.


Technical Aspects of Catheter InsertionOver the last decade, the methods of percutaneous catheter implantation

have gained in easiness and safety, contributing to reduce traumatic complica-tions and to increase catheter life expectancy.

Major advances in securing catheter insertion have been provided by ultra-sound based methods. These methods have several benefits over the conventionalanatomical landmark method, including a reduction in catheter failure rate, areduction in time spent to insert catheter, an increase in patient’s comfort and asignificant reduction of major traumatic complications (pneumothorax and/orhemothorax) [6]. It has also been observed that routine use of ultrasound explo-ration identifies anatomical variations of the jugular vein position (right or left)in about 15–20% of cases. Ultrasound method may also be used either to locatevein prior any attempt of venipuncture offering a precise skin external landmarkor to guide directly the vein cannulation at the time of catheter insertion.

Basic rules for catheter insertion are known [7]. Catheter insertion must beperformed in a clean room following strict aseptic rules including a meticulousskin preparation and disinfection of the patient, the use of sterile drape, gown,gloves, mask and hat being worn by the operator.

Sites of InsertionThe placement of the hemodialysis catheter is restricted to anatomical vein

locations offering access to a central high-flow system (superior and inferiorvena cava system). Selection of the anatomical venous site depends on the clin-ical context, the patient’s state and the physician’s expertise.

Femoral approach should preferred as first option with critically ill respi-ratory condition (pulmonary edema, respiratory failure), or when thoracicaccess appears risky or because patient’s condition dictates prolonged stay inbed (coma, ventilatory assistance, and multiple injuries). Internal jugular accessis preferable in the absence of thoracic life-threatening conditions. Subclavianaccess should be reserved for short-term treatment or when no other centralvein access option exists. In all cases the percutaneous vein cannulation shouldbe performed after skin preparation, under local anesthesia in very strict asep-tic condition. The detailed procedures of catheter insertion will not be describedhere, and the interested reader is referred to textbook and technical reports inthe field.

Femoral Vein

Femoral vein represents the first and original route for percutaneous veincatheterization described by Shaldon in the early 1960s. Nowadays, the femoral


vein remains an indication of choice in case of emergency when hemodialysishas to be launched without delay in a patient with critical cardiopulmonarycondition. Femoral catheterization is also useful in bedridden patients withsevere neurologic disorders, ventilatory assistance (tracheotomy) or multipleorgan failure (fig. 3).

Double-lumen polyurethane catheters represent the best option in this caseto start the dialysis within few minutes. The right or left femoral vein offers thesame facility of insertion. The length of femoral catheters that provide optimalflow performances is between 25 and 35 cm. The insertion site is locatedapproximately 1–2 cm below the crural arcade and 1 cm medially apart from thefemoral artery. Although such type of femoral catheters may be extended toseveral weeks, it is commonly accepted for safety reason to restrict their use upto 2 weeks. Two single-lumen polyurethane catheters inserted in the same ipsi-femoral vein may represent another useful option to increase flow perfor-mances. Tunnelized soft silicone catheters used in chronic patients [8] havebeen also successfully used in acute condition. These types of catheters repre-sent a new and quite interesting option in the acute setting providing high flowperformances with reduced complication rates that deserve further studies.

Subclavian Vein

Subclavian catheter cannulation for hemodialysis was popularized byUldall in the early 1980s. Subclavian catheters entailing a major risk of steno-sis and/or thrombosis of the subclavian vein or one of its affluent vein superiorvena cava, innominate vein, brachiocephalic troncus, have reduced dramatically

Femoral double-lumen

silicone catheter (tunnel)

Femoral double-lumen

polyurethane catheter

Fig. 3. Venous accesses via the inferior cava system.


the indications for subclavian cannulation [9]. It is however important toacknowledge that these complications were mainly observed with use ofsemirigid and poorly hemocompatible catheters. From these historical reports,no prospective study has been carried out to really assess the incidence ofcatheter-related complications with soft and hemocompatible material. Indeed,subclavian cannulation should be considered a second line venous angioaccessintent for very short-term use or in absence of other alternative (fig. 4).

Several methods have been described for the percutaneous insertion ofsubclavian catheters. Insofar as subclavian cannulation is indicated, softcatheters made preferably of silicone rubber should be considered for insertion.The right subclavian approach is preferable to reduce the distance from inser-tion site to right atrium. The left subclavian vein is more difficult to cannulateand exposed to a higher risk of left brachiocephalic thrombosis. In case ofpolyurethane catheters short cannula (20–25 cm) are indicated to preventcardiac trauma (atrium and ventricule perforation). This fatal risk is clearlyminimized with the use of soft silicone rubber catheters.

Internal Jugular Vein

Over the last 15 years, internal jugular access has gained popularity inacute conditions. Irrespective of the catheter used, jugular vein catheterization

Fig. 4. Acute catheters, Double-lumen polyurethane.


(particularly the right vein) entails a reduced incidence of venous thrombosesrelative to the subclavian access [10, 11]. The jugular route is prone to infec-tious complications. Straight catheters exiting close to the ear and/or in patientswith tracheotomy bear a higher risk of infection (fig. 4). Kinked double-lumenpolyurethane catheters have recently proposed to reduce this risk. Such type ofacute catheter does not provide a safe and comfortable venous access site eitherfor patient of nurses. Interestingly, soft polyurethane and silicone rubbercatheters (double-lumen, double-catheter, split-catheter) (PermCath, QuintonInstrument Co, Seattle, Wash., USA; DualCath, Hemotec, Ramonville, France& Medcomp, Harleysville, Pa., USA) implanted percutaneously into the inter-nal jugular with a tunnel have been employed as a first intent in acute patientswith excellent results (fig. 5).

The percutaneous insertion of the catheter into the internal jugular vein ispreferably performed in the low position of the Sedillot’s triangle. The rightinternal jugular vein represents the best anatomical option offering a straightand short distance to reach the right atrium. The left internal jugular vein maybe indicated as second intent since it is technically more difficult to cannulateand it bears higher risk of vein thrombosis. Soft silicone rubber catheters offersthe possibility of creating a 10- to 12-cm subcutaneous tunnel on the chestwall below the clavicle. According to the type of catheters (single part, doubleparts) the tunnelization will be performed either upwards or downwards.

Fig. 5. Chronic tunnelized dual-catheter silicone inserted in the internal jugular vein(DualCath).


Subcutaneous anchoring system (dacron cuff, purse ring suture) will secure andprotect catheter from infection.

From this section, it must be acknowledged that each vein location andeach type of catheters has specific advantages but bears the risks of foreignmaterial inserted into a vein. To ensure continuity of renal replacement therapyin acute patient it is often necessary to change the venous site (catheterdysfunction, catheter infection). The femoral vein is the first site of insertion incritical conditions for commodity. The right internal jugular and subclavianapproaches are in most cases the preferred sites of insertion for mid-term use.They facilitate the mobilization of patients (care, physiotherapy, imaging) andshould be preferred in conscious or mobile patients.

Catheter PositioningCorrect positioning of the catheter’s tip is essential to prevent mechanical

complications and catheter dysfunction.Thoracic catheters inserted in the superior vena cava system should have

their tips located between the superior vena cava and the right atrium to opti-mize flow rate. However, such position is associated with specific risks ofcardiac trauma or atrial thrombosis. Fluoroscopic or chest X-ray checking cor-rect catheter position and tip location must performed prior catheter use. Toensure safety, semirigid catheters should have their tips located in the superiorvena cava system (1 or 2 cm above the right atrium) and not in the heart.Although reduced, this risk still persists with soft catheter and a high level ofvigilance must be maintained. Twenty centimeters is the optimal length for athoracic catheter inserted on the right site in an adult. Three to four additionalcentimeters are required when catheters are inserted on the left side.

Femoral catheters accessing the inferior vena cava system should havetheir tips positioned in the central lumen of the inferior vena cava. This positionsupports a regular and correct functioning of the catheter and reduces the riskof vein perforation. Twenty-five to thirty centimeters length is needed for afemoral catheter to reach the desired location in an adult.

Catheter Care and Catheter Maintenance

In a recent meta-analysis study, it has been shown that appropriate andstrict hygienic rules and good medical practices are the best warranty of suc-cessful catheter handling [12]. In order to achieve highest quality hemodialysiscatheter should be used exclusively for renal replacement therapy avoiding theiruse for blood samples, parenteral feeding and intravenous injections. These actsincrease the risk of catheter contamination. Nurse training is essential in the


prevention of infection. Strict aseptic conditions in catheter handling must beadopted at any time [13, 14]. Interestingly, the use of a betadine ointment and/orantiseptic protective box on catheter hubs significantly reduces the incidence ofbacteremias in hemodialyzed patients [15, 16]. Catheter dressing is highlydesirable to protect the emergence of catheters. Indeed, tight occlusivepolyurethane occlusive dressings promoting moisture and proliferation ofcutaneous bacteria are not suitable since they have been associated with anincrease risk of catheter infection [17, 18]. Preliminary studies comparingbioactive material (silver or antibiotic impregnated material) to regular catheterhave led to encouraging results in reducing infection incidence. Catheter lock-ing solutions based on antithrombotic/antiseptic or antibiotic mixture have alsoproved to be efficient in preventing endoluminal bacteria contamination. Realbenefice of these approaches deserve further studies before being proposed aspreemptive therapeutic option.

Catheter Performances

The nature of the material used and the inner diameter of the catheterdirectly condition its performances. Thus, compliance and elastance of thematerial, catheter’s length and diameter and the presence of distal holes, condi-tion the resistance to the blood circulation. An internal port lumen diameter of2.0–2.2 mm, such as found in silicone rubber catheters, is able to provide anoptimal flow/resistance regimen [19]. Intermittent RRT modalities requiringhigh blood flows (250–400 ml/min) are mostly dependent upon catheter’sphysical characteristics. Continuous RRT modalities requiring fairly low bloodflows (150–200 ml/min) are much less blood flow dependent. Patency of theinternal lumen and its stability in the venous bulk flow are key factors forproper operation of a catheter. Two independent catheter ports (or split catheter)bearing circular distal holes are essential to reduce probability of catheter dys-function (parietal suction, partial lumen obstruction, fibrin sleeve formation).

Extracorporeal blood pressure (arterial and venous side) recorded by thedialysis monitoring module provide indirect evidence of catheter’s patency. Thenegative pressure recorded on the inflow side and/or the positive pressurerecorded on the outflow side, reflect blood resistance in the extracorporealcircuit. These values may be used as indirect bed-side index of partial catheterobstruction [20]. For example, it is convenient to observe that the venouspressure is half the blood flow (e.g. for a blood flow rate of 300 ml/min, venouspressure is close to 150 mm Hg) in the silicone rubber DualCath device.

Early catheter mechanical dysfunction is more frequent with double-lumencatheters inserted via the subclavian way than with those inserted in other sites.


In these cases, causes of catheter kinking and striction (subclavian pathway orsuture) must be ruled out.

Catheter recirculation, a factor reducing dialysis efficacy, depends both onsite of catheter insertion and flow prescribed [21]. The deleterious role of recir-culation is much less important in continuous modalities than in intermittentones. Femoral catheters, particularly short ones, exhibit a high recirculation rateaveraging 20% (5–38%). Internal jugular and subclavian catheters have muchlower recirculation rate averaging 10% (5–15%). It is also interesting to notethat the inversion of the connecting lines (by using the arterial line for venousreturn and the venous line for blood aspiration) on double-lumen catheters, thatis sometimes performed to increase flow when a pressure problem occur,significantly increases blood recirculation rates. In this context recirculation upto 20 or 30% have been observed [22, 23].

Catheter Complications

The causes of hemodialysis catheter dysfunction are usually related to thedelay after insertion [24]. Immediate or early dysfunction usually result frommechanical problems, among which are a malpositioning of the catheter tip(sucking the wall of the vein), a kinking of the catheter and a striction causedby ligatures or aponevrosis [25]. Late dysfunctioning (more than 2 weeks) ismore often caused by thrombotic problems: partial or total obstructive throm-bosis of the catheter lumen, thrombosis or stenosis of the cannulated vein,external sheath formation on the catheter distal end (external fibrin sleeve),internal coating of the catheter (endoluminal fibrin sleeve) [26]. In the lattercase, the partial or total occlusion of the lumen and distal and/or lateralperforations greatly increase the extracorporeal resistances and reduce accord-ingly the effective blood flow delivered.

Thrombotic complications can occur in several ways. Endoluminalthrombosis of the catheter is the most common. This is revealed by intermit-tent or permanent catheter dysfunction. The catheter may be reopened inthese cases by mechanical (brush) or chemical methods (fibrinolytic).External thrombosis of the catheters caused by a fibrin sheath covering the tipof the catheters requires either fibrinolysis, catheter stripping through apercutaneous femoral route, or removing and replacement of catheters.Thrombosis of the host cannulated veins is a severe situation (potential sourceof pulmonary embolism). The incidence may vary from 20 to 70% of reportsaccording to the sites and diagnostic modalities used [27]. Thrombosis of theright atrium is the most serious and potentially lethal complication.Symptoms of thrombotic complications are dysfunction of the catheter, rarely


onset of edema of the ipsilateral limb or unexplained febricula are the appeal-ing symptom. Several factors contribute to the thrombogenicity of thecatheter including the catheter (material and composition of the catheter, soft-ness, aspect and surface treatment), the mode of insertion, the type of vein(diameter, local hemodynamics), the duration of cannulation, the clotting andinflammatory state of the patient (hyperfibrinemia, inflammatory syndrome,hyperthrombocytemia, previous venous thrombosis), and the contaminationof the catheter.

Infections represent a major burden for hemodialysis catheters [28–30].Nontunneled polyurethane catheters used as a short-term therapy entail abacteremia risk estimated to a median value of 5.9 episodes per 1,000 patient-days in intensive care units [31]. The incidence of bacteremias varies greatlyaccording units and catheter practice [32]. Non-tunneled internal jugularaccess bears a higher risk of infections, notably in patients with a tracheotomy[33, 34].

Early infection may be related to catheter placement problems or skinand catheter track infection. Placement of a percutaneous catheter disruptsthe continuous protecting solution of the skin. Bacterial colonization of theskin is most often incriminated in these catheter infections. The skin acts asa bacterial reservoir and contributes to the subcutaneous penetration of germsalong the catheter pathway. Hence, the need to carefully disinfect the skinprior to any catheter insertion but equally to prevent the onset of cutaneouslesions and ensure a particular care in the patients with catheters is alwayssuitable.

Late infections are most often associated with endoluminal cathetercontamination and may be the expression of microbial biofilm seeding thecatheter lumen. The prolonged use of a dialysis catheter in acute patiententails an infectious risk increasing with time. This risk is known but unfor-tunately unavoidable. It must be alleviated through suitable nursing care andhandling. Two types of infections are observed with catheters: local infection(skin exit, track infection) and systemic infection (bacteremia, septicemia,infected thrombosis). Skin exit and bacteremia are the most frequent forms ofinfection that may be treated by local and systemic antibiotic therapy whilekeeping the catheter in situ. Catheter track infection, septicemia and infectedvenous thrombosis are most severe form of infections requiring bothcatheters withdrawal and systemic antibiotic therapy. Endoluminal cathetercontamination from hubs may be the source of microbial biofilm. In this case,bacteria entering into the lumen, adhere onto the catheter surface, growth,produce glycocalyx (slyme) and become resistant to antibiotic. Occasionally,bacteria may be released from this biofilm (e.g. higher stress conditions dueto blood pump speed) being the source of bacteremia and fever episode.


In the event of an unexplained septic condition, it is reasonable to refer firstto the role of the catheter as the infection carrier. Several authors haveproposed a catheter replacement over a guidewire through the same subcuta-neous track [35]. This unsafe microbiological approach appears undesirablein our opinion and should be abandoned. In another approach, the cathetersare systematically changed every 4 days and inserted in different venoussites. However, whether this change is pertinent remains highly debatable. Inall cases it is essential to cultivate the withdrawn catheter. The insertion ofsoft tunnelized catheters (with or without anchoring system) for long-termuse appears more suitable to prevent catheter hazards. Strict aseptic rules(gloves, mask, drapes, antiseptic) shall be followed at any time to handlecatheters and particularly at the time of blood lines connection to preventcontamination of catheter hubs.

Stenosis of the host vein is a common risk of catheter in the long term. Itis more common with semirigid than with soft catheters. This troublesomecomplication may compromise the creation of arteriovenous fistula in a chronicrenal failure patient.

Technical Catheter Advances

Biomaterial research and technical development have suggested newsolutions to render safer the use of catheter.

New polymer or surface treatment of the catheter may reduce the hemore-activity and the prothrombotic activity of the catheter lines substantially [36, 37].Introduction of bioactive material with ion bombardment, silver impregnationand anticoagulant or antibiotic/antiseptic are intended to prevent platelet adhe-sion and subsequent clotting and to prevent bacterial adhesion [38]. Recentstudies comparing the incidence of infection using antibiotic impregnatedcatheters to regular ones have confirmed the efficacy in reducing streamlineinfection over short period of time [39, 40]. However, the true protective effectof this bioactive material is effective on longer period of time (�6 months)remains to be proved [41].

Catheter locking solution during the interdialytic period, based onantithrombotic/antiseptic mixture, appears more appealing in this context.Locking solutions (Neutrolin, Citralock) are already available on the market[42, 43]. They have been proved highly effective both in reducing significantlycatheter-related infection and in preventing bacterial biofilm formation inchronic patients [44].

It remains to be proved that these solutions are applicable to acute settingwithout risks and this concern deserve further studies [45].


Prevention of Venous Catheter-Related Morbidity

Reducing venous catheter-related complications implies for an ICU man-aging team to establish and follow very stringent rules of use. These rules maybe summarized in a ‘good medical practice act’ that we call humorously ‘TheTen Commandments for Catheter Users in the ICU’.

Thou shalt select the best venous access site according to the clinicalcondition of your patient.

It is important to choose the venous access site taking into account theclinical condition and particularly the critical illness of the patient. Femoralapproach should preferred as first option with critically ill respiratory conditionor because patient’s condition dictates prolonged stay in bed. Internal jugularaccess is preferable in the absence of thoracic life-threatening conditions.Subclavian access should be reserved for short-term treatment or when no othercentral vein access option exists.

Thou shalt discuss the type of catheter according to the modality of renalreplacement therapy that you indicate.

The type of catheter to be inserted will be defined according to theexpected duration of use and to the critical condition of the patient. Acutecatheters (e.g. polyurethane double lumen) are indicated as emergency and sal-vaging angioaccess method to launch immediate hemodialysis in very criticallyill patient. Chronic tunnelized catheters (e.g. dual catheters, split catheters) areindicated for prolonged treatment period when life-threatening conditions arecontrolled.

Thou shalt select an experienced and trained operator.The insertion of venous catheter should be reserved to experienced opera-

tor. This technical act should not be trivialized and performed by the youngestinexperienced fellow. Ultrasound-based methods provide a major advance forsafety and easiness of catheter insertion particularly for trainees. They shouldbe used as much as possible. These methods may be used either to locate veinprior any attempt of venipuncture offering a precise skin external landmark orto guide directly the vein cannulation at the time of catheter insertion.

Thou shalt establish a precise protocol in your unit for the catheter use.A precise procedure for catheter use and handling must be established in

each ICU. This is quite important to respect the working constraints of eachunit. All details must be mentioned and periodic checking should be performedto verify that the procedure is applied and fit ICU needs. Very strict aseptic con-ditions (catheter and local disinfection, use of sterile gown, mask and gloves)applied any time when catheter is manipulated are the first protective line [46].

Thou shalt educate your nursing staff and collaborators to carefullyrespect the defined handling procedures.


Training of the nursing personnel is essential to prevent catheter dysfunc-tion and infection. This aspect should be part of a continuous training programin the ICU. It is recommended to have a referent experienced nurse for thisaspect that will coordinate actions and report results of the ICU.

Thou shalt evaluate the particular risk of your patient according to theclinical situation.

ICU treated patients have multiple risk factors for developing venouscatheter complications. Among them, one can identify some very risky factorssuch as sepsis, ventilation, stomized (colostomy, ureterostomy…), severeinflammation, malnutrition, and postoperative patients with multiple drainage.All these factors should be considered for venous access choice and for cathetercare in order to enhance catheter protection barriers.

Thou shalt watch carefully and regularly catheter to detect earlier anycomplication.

Catheter performances should be checked regularly in order to achieve dial-ysis delivery dose. That could be performed by evaluating dialysis efficacy andblood flow and flow resistance. Catheter aspect and skin exit must be assessedregularly to detect a local inflammation or infection. This is an important part ofthe nursing task that should be reported and recorded in the patient flowchart.

Thou shalt restrict exclusively the use of catheter for renal replacementtherapy.

Venous catheter should be used exclusively for dialysis. Parenteral nutri-tion, intravenous fluid infusion, blood samplings and drug administrationshould be prohibited via dialysis catheters except when dialysis is running.Such manipulations increase significantly the risk of catheter dysfunction(partial thrombosis or precipitation) and bacterial contamination.

Thou shalt correct individual patient risk factors in order to preventcomplications.

When specific patient’s risks are identified, they should be corrected.Chronic bearers of bacteria (e.g. Staphylococcus aureus) should be treated bytopical antiseptics or antibiotics to eradicate the source. Skin cleaning anddisinfection of the surrounding area should be performed with an appropriatedisinfectant. Dressing covering the catheter must be proof and adequately adhe-sive to skin surface. Stomies should be adequately isolated. The use of catheterlocking solution with dual action, antithrombotic and antiseptic, are stronglyrecommended within the interdialytic period to prevent catheter clotting andbacterial contamination.

Thou shalt institute with your nursing staff a continuous quality improve-ment assurance program.

All these measures are virtually ineffective if they are not part of a con-tinuous quality improvement program. In other words, a specific quality


assurance program for catheter care must be implemented in each unit underthe expertise of a dedicated staff member. Catheter care protocols will beadjusted to the results of the ICU and nursing staff will be associated to thisthought [47].

Conclusion

Venous catheters are the preferred form of angioaccess in the acute setting.Catheter insertion is facilitated by the use of ultrasound devices. Catheter careand handling require complying with strict protocols in order to preventcatheter-related complications. Catheter performances should be evaluated inthe acute setting in order to deliver adequate dialysis dose in ARF patients. Bestclinical results with venous catheters will be guaranteed when applying ‘TheTen Commandments for Catheter Users in the ICU’ every day.

References

1 Canaud B, Martin K, Nguessan C, Klouche K, Leray-Moragues H, Beraud JJ: Vascular access forextracorporeal renal replacement therapies in the intensive care unit in clinical practice. ContribNephrol. Basel, Karger, 2001, vol 132, pp 266–282.

2 Uldall R: Hemodialysis access. Part A: Temporary; in Jacobs C, Kjellstrand CM, Koch KM,Winchester JFW (eds): Replacement of Renal Function by Dialysis. Dordrecht, Kluwer AcademicPublishers, 1996, pp 277–292.

3 Maclean-Ross AH, Griffith CDM, Anderson JR, Grieve DC: Elastomer subclavian catheters.J Parenter Enteral Nutr 1982;6:61–63.

4 Curelalu I, Gustavsson B, Hansson AH: Material thrombogenicity in central venous catheteriza-tion. II. A comparison between plain silicone elastomer and plain polyethylene, long, antebrachialcatheters. Acta Anaesthesiol Scand 1983;27:158–164.

5 Canaud B, Leray-Moragues H, Garred LJ, Turc-Baron C, Mion C: Permanent central vein access.Semin Dial 1996;9:397–400.

6 Defalque RJ, Campbell C: Cardiac tamponade from central venous catheter. Anesthesiology1979;50:249–252.

7 Food and Drug Administration: Precautions Necessary with Central Venous Catheters.Washington, US Government Printing Office, 1989, p 15.

8 Montagnac R, Bernard CI, Guillaumie J, Hanhart P, Clavel P, Yazji J, Martinez LM, Schillinger F:Indwelling silicone femoral catheters: Experience of three haemodialysis centres. Nephrol DialTransplant 1997;12:772–775.

9 De Moor B, Vanholder R, Ringoir S: Subclavian vein hemodialysis catheters: Advantages and dis-advantages. Artif Organs 1994;18:293–297.

10 Cimochowski GE, Worley E, Rutherford WE, Sartain J, Blondin J, Harter H: Superiority of theinternal jugular over the subclavian access for temporary dialysis. Nephron 1990;54:154–161.

11 Schillinger F, Schillinger D, Montagnac R, Milcent T: Post-catheterisation vein stenosis inhemodialysis: Comparative angiographic study of 50 subclavian and 50 internal jugular access.Nephrol Dial Transplant 1991;6:722–724.

12 Mcgee WT, Ackerman BL, Rouben LR, Prasad VP, Bandi V, Mallory DL: Accurate placement ofcentral venous catheters: A prospective, randomized, multicenter trial. Crit Care Med 1993;21:1118–1123.


13 Vanherweghem JL, Dhaene M, Goldman M, Stolear JC, Sabot JP, Waterlot Y, Serruys E, Thayse C:Infections associated with subclavian dialysis catheters: The key role of nurse training. Nephron1986;42:116–119.

14 Alonso-Echanove J, Edwards JR, Richards MJ, Brennan P, Venezia RA, Keen J, Ashline V,Kirkland K, Chou E, Hupert M, Veeder AV, Speas J, Kaye J, Sharma K, Martin A, Moroz VD, GaynesRP: Effect of nurse staffing and antimicrobial-impregnated central venous catheters on the risk forbloodstream infections in intensive care units. Infect Control Hosp Epidemiol 2003;24:916–925.

15 Levin A, Mason AJ, Jindal KK, Fong IW, Goldstein MB: Prevention of hemodialysis subclavianvein catheter infections by topical providone-iodine. Kidney Int 1991;40:934–938.

16 Maki DG, Alvarado CJ, Ringer M: A prospective, randomized trial of providone-iodine, alcoholand chlorhexidine for prevention of infection with central venous and arterial catheters. Lancet1991;338:339–344.

17 Conly JM, Grieves K, Peters B: A prospective, randomized study comparing transparent and drygauze dressings for central venous catheters. J Infect Dis 1989;159:310–315.

18 Hoffman KK, Weber DJ, Samsa GP, Rutala WA: Transparent polyurethane film as an intravenouscatheter dressing: A meta-analysis of infection rates. JAMA 1992;267:2072–2076.

19 Athirakul K, Conlon P, Schwab S: Cuffed central venous hemodialysis catheters and adequacy ofdialysis. J Am Soc Nephrol 1996;7:1402–1403.

20 Stroud CC, Meyer SL, Bawkon MC, Smith HG, Klein MD: Vascular access for extracorporealcirculation: Resistance in double lumen cannulas. ASAIO J 1991;37:M418–M419.

21 Sherman RA, Matera JJ, Novik L, Cody RP: Recirculation reassessed: The impact of blood flowrate and the low-flow method reevaluated. Am J Kidney Dis 1994;23:846–848.

22 Leblanc M, Fedak S, Mokris G, Paganini EP: Blood recirculation in temporary central cathetersfor acute hemodialysis. Clin Nephrol 1996;45:315–319.

23 Level C, Lasseur C, Chauveau P, Bonarek H, Perrault L, Combe C: Performance of twin centralvenous catheters: Influence of the inversion of inlet and outlet on recirculation. Blood Purif2002;20:182–188.

24 Leblanc M, Bosc JY, Paganini EP, Canaud B: Central venous dialysis catheter dysfunction. AdvRenal Replacement Ther 1997;4:377–389.

25 Kelber J, Delmez JA, Windus DW: Factors affecting delivery of high-efficiency dialysis usingtemporary vascular access. Am J Kidney Dis 1993;22:24–29.

26 Hoshal VL, Ause RG, Hoskins PA: Fibrin sleeve formation on indwelling subclavian centralvenous catheters. Arch Surg 1971;102:353–358.

27 Trottier SJ, Veremakis C, Oíbrien J, Auer AI: Femoral deep vein thrombosis associated with centralvenous catheterization: Results from a prospective, randomized trial. Crit Care Med 1995;23:52–59.

28 Gil RT, Kruse JA, Thill-Baharozian MC, Carlson RW: Triple- vs single-lumen central venouscatheters: A prospective study in a critically ill population. Arch Intern Med 1989;149:1139–1145.

29 Collignon P, Soni N, Pearson I: Sepsis associated with central vein catheters in critically illpatients. Intens Care Med 1988;14:227–232.

30 Richet H, Hubert B, Nitemberg G: Prospective multicenter study of vascular-catheter-related com-plications and risk factors for positive central-catheter cultures in intensive care unit patients.J Clin Microbiol 1990;28:2520–2526.

31 Maki DG: Nosocomial infection in the intensive care unit; in Parillo JE, Bone RC (eds): CriticalCare Medicine Principles of Diagnosis and Management. St-Louis, Mosby, 1995, pp 893–954.

32 Eyer S, Brummitt C, Crossley K: Catheter-related sepsis: Prospective, randomized study of threemethods of long-term catheter maintenance. Crit Care Med 1990;18:1073–1078.

33 Pezzarossi HE, Ponce De Léon S, Calva JJ: High incidence of subclavian dialysis catheter-relatedbacteremias. Infect Control 1986;7:596–602.

34 Cheesbrough JS, Finch RG, Burden RP: A prospective study of the mechanisms of infectionassociated with hemodialysis catheters. J Infect Dis 1986;154:579–586.

35 Cobb DK, High KP, Sawyer RG: A controlled trial of scheduled replacement of central venous andpulmonary-artery catheters. N Engl J Med 1992;327:1062–1068.

36 Baumann M, Witzke O, Dietrich R, Haug U, Deppisch R, Lutz J, Philipp T, Heemann U:Prolonged catheter survival in intermittent hemodialysis using a less thrombogenic micropat-terned polymer modification. ASAIO J 2003;49:708–712.


37 Eberhart RC, Clagett CP: Catheter coatings, blood flow, and biocompatibility. Semin Hematol1991;28:42–48.

38 Tobin EJ, Bambauer R: Silver coating of dialysis catheters to reduce bacterial colonization andinfection. Ther Apher 2003;7:504–509.

39 Darouiche RO, Raad II, Heard SO, Thornby JI, Wenker OC, Gabrielli A, Berg J, Khardori N,Hanna H, Hachem R, Harris RL, Mayhall G: A comparison of two antimicrobial-impregnated cen-tral venous catheters. Catheter Study Group. N Engl J Med 1999;340:1–8.

40 Chatzinikolaou I, Finkel K, Hanna H, Boktour M, Foringer J, Ho T, Raad I: Antibiotic-coatedhemodialysis catheters for the prevention of vascular catheter-related infections: A prospective,randomized study. Am J Med 2003;115:352–357.

41 Bambauer R, Mestres P, Schiel R, Bambauer S, Sioshansi P, Latza R: Long-term catheters forapheresis and dialysis with surface treatment with infection resistance and low thrombogenicity.Ther Apher Dial 2003;7:225–231.

42 Sodemann K, Polaschegg HD, Feldmer B: Two years’ experience with Dialock and CLS (a newantimicrobial lock solution). Blood Purif 2001;19:251–254.

43 Quarello F, Forneris G: Prevention of hemodialysis catheter-related bloodstream infection usingan antimicrobial lock. Blood Purif 2002;20:87–92.

44 Weijmer MC, Debets-Ossenkopp YJ, Van De Vondervoort FJ, ter Wee PM: Superior antimicrobialactivity of trisodium citrate over heparin for catheter locking. Nephrol Dial Transplant2002;17:2189–2195.

45 Polaschegg HD, Sodemann K: Risks related to catheter locking solutions containing concentratedcitrate. Nephrol Dial Transplant 2003;18:2688–2690.

46 Raad II, Gilbreath J: Prevention of central venous catheter-related infections by using maximalsterile barrier precautions during insertion. Infect Control Hosp Epidemiol 1994;15:231–236.

47 Coopersmith CM, Zack JE, Ward MR, Sona CS, Schallom ME, Everett SJ, Huey WY,Garrison TM, McDonald J, Buchman TG, Boyle WA, Fraser VJ, Polish LB: The impact of bedsidebehavior on catheter-related bacteremia in the intensive care unit. Arch Surg 2004;139:131–136.

Professor Bernard CanaudDepartment of Nephrology, Lapeyronie University HospitalCHU Montpellier, FR–34295 Montpellier (France)Tel. �33 4 67 33 84 95, Fax �33 4 67 60 37 83, E-Mail [email protected]


Anticoagulation in Continuous Renal Replacement Therapy

O. Vargas Hein, W.J. Kox, C. Spies

Department of Anesthesiology and Intensive Care, University Hospital Charité, Campus Mitte, Berlin, Germany

The incidence of acute renal failure (ARF) varies in critically ill patientsbut it can be as high as 25% [1]. If ARF is part of the multiple organ dysfunctionsyndrome the mortality increases to over 50% [2]. Continuous renal replace-ment therapy (CRRT), occurring in up to 20% of the ICU patients with ARFhas become the treatment of choice for ARF in the intensive care unit (ICU) [3].During CRRT the blood-foreign membrane interactions lead to an activation ofdifferent cascades involving the plasmatic and cellular coagulation system, thecomplement system and the inflammatory cascade. In the last years, a closerelationship of interactions between the inflammatory and the coagulation sys-tem have been investigated [4, 5]. The result is clotting of the hemofiltrationsystem, mainly the filter [4, 6]. To counteract these interactions anticoagulationis necessary [4, 7]. However, the critically ill patients under CRRT are mostlyhighly compromised in their hemostatic system. Coagulopathies and underlyingdiseases with a high risk for bleeding are present as well as hypercoagulablestates [8, 9]. The equilibrium between avoiding bleeding complications andkeeping the system open considering patients’ multiple diseases is difficult toachieve with the anticoagulation regimes used at present.

High-Molecular-Weight (HMW) Heparin

HMW heparin is used as the standard for anticoagulation in CRRT.Heparin, a mucopolysaccharide composed of equal amounts of D-glucosamineand uronic acid, has a molecular weight of 3,000–30,000 daltons and a short half-life of 1–1.5 h. The mechanism of action of HMW heparin is antithrombin III

Anticoagulation in CRRT 309

(AT III) dependent, potentiating the inhibitory action of AT III mainly on factors IIa,IXa, Xa, XIa and XIIa [10]. The main metabolism and elimination is renal func-tion independent so that no accumulation has to be expected. However, as HMWheparin metabolism is dependent on the interaction of various proteins andendothelial cells the effective time of action can be very variable [7, 10].Antagonizing heparin effectively can be achieved by protamine. Nevertheless,bleeding complications are as high as 20–30% [1, 6, 11]. Especially in criticallyill patients with high risk of bleeding the preexisting coagulation state has to betaken into consideration in order to choose the right dose. The hemofiltrationsystem is rinsed with saline containing 5,000–20,000 IU of HMW heparin. Atinitiation of CRRT a bolus of 500–1,000 IU of HMW heparin is given followedby a continuous infusion of 300–1,000 IU/h or 5–10 IU/kg/h [6, 12]. Monitoringaccurately the dosing regime is essential to avoid bleeding complications. Thepartial thromboplastin time (aPTT) has shown a good correlation with filtersurvival, heparin dose has not. An aPTT of 35–45 s seems to keep the bestbalance between bleeding complications and keeping the system free of clotting[6, 12]. Besides bleeding events heparin induced thrombocytopenia type II(HIT II) is a further complication with an incidence of up to 10% in critically illpatients. After its diagnosis further anticoagulation with heparin is strictly con-traindicated [1]. Because of pathological thrombocyte adhesion and agglutina-tion in HIT II life-treathening thromboembolisms as well as augmented filterclotting can occur. Alternative anticoagulation is mandatory.

Low-Molecular-Weight (LMW) Heparin

LMW heparins are depolymerized HMW heparin and have a molecularweight of 4,500–9,000 daltons. As HMW heparin LMW heparin binds to AT IIIbut has a major action on factor Xa [6]. The half-life is 3–4 h and the elimina-tion is to 50% renal. LMW heparin is not reliably filtered and antagonized byprotamine. The dose correlation with aPTT is insufficient. Dosage monitoringhas to be performed measuring the anti-Xa-activity, which is not a routineparameter. In renal failure LMW heparin can accumulate, reinforced by theproblematic dosage monitoring [6]. The studies performed with LMW heparinfor anticoagulation in CRRT have described either comparable results or lessbleeding complications than with HMW heparin with comparable filter effec-tivity [13, 14]. The optimum dose lies in the range of 15–20 IU/kg bolus fol-lowed by 5–10 IU/kg/h with a target anti-Xa-activity between 0.3–0.6 IU/ml [12,14, 15]. Even though LMW heparin has a weaker action on thrombocytes thanHMW heparin, 90% cross-reactivity with HIT II antibodies have been described.Therefore, LMW heparin is contraindicated in patients with HIT II [6, 12].

Vargas Hein/Kox/Spies 310

Danaparoid-Sodium

Danaparoid is a low-molecular glucosaminoglykan (heparinoid) with amolecular weight of 5,500 daltons. The mechanism of action is AT III depen-dent inhibiting factor Xa. The half-life is 24 h. Renal elimination is 50%.Therefore, in renal failure half-life is prolonged to 31 h with the risk of drugaccumulation. Danaparoid cannot be filtered or antagonized [6]. The monitoringof anticoagulation is performed with the anti-Xa-activity. Danaparoid has beendescribed to be effective and safe in intermittent hemodialysis [6, 16]. A studyperformed in CRRT with danaparoid as anticoagulant with a dose of70–225 IU/h with an approximate level of 0.3–0.8 IU/ml showed a satisfactoryfilter patency. However, in all patients with coagulopathies bleeding complica-tions occurred [16]. In patients with HIT II 10% in vitro cross-reactivity hasbeen observed for danaparoid sodium. The clinical relevance (in vivo cross-reactivity) is described to be negligible and danaparoid has been used successfullyin patients having HIT II [16–18]. However, case reports have described life-threatening thromboembolic complications in patients who have beenswitched from HMW heparin to danaparoid because of HIT II [18]. Therefore,under close monitoring of thrombocytes count and when no other alternativeanticoagulation is feasible danaparoid can be seen as an alternative to heparinin HIT II patients.

Hirudin

Hirudin, a polypeptide made by recombinant technology with a molecularweight of 7,000 daltons, acts independently of cofactors and directly inhibitsbound and unbound thrombin. Its elimination half-life is 1–3 h and the elimi-nation pathway is �90% unmetabolized through the kidneys. In the presence ofrenal failure the half-life of hirudin is considerably prolonged up to 100 timesthe normal half-life [1]. Hirudin is removed through hemofiltration membranesat a rate dependent on the sieving coefficient of the used membrane [19]. Incase of ARF, hirudin could persist at low levels even when it is not given anymore because of redistribution. In interaction with coagulation disorders due toother underlying diseases it could possibly lead to bleeding complications [20].aPTT has been considered an unreliable monitoring parameter during hirudinanticoagulation, because it does not correlate with high hirudin levels [1].Ecarin clotting time (ECT), a bedside whole blood plasma clotting time assaybased on thrombin activation through the snake venom ecarin, has shown ade-quate dose-response curves [20]. However, ECT is no routine parameter. Inthis case aPTT can be used with a target of 50–60 s. Hirudin has been applied


continuously (12–500 �g/kg/day) and as bolus (4–128 �g/kg/day) applicationin CRRT showing adequate filter patency in comparison to HMW heparin[1, 20, 21, 22]. However, continuous application led to bleeding complications inpatients with plasmatic or cellular coagulopathies even though hirudin wasunder or in the therapeutic range [1, 20, 21]. Beside the coagulation status ofthe patient, an important issue is also the difference in risk of bleeding betweensurgical and medical patients. Comparable doses of continuous hirudin led tobleeding complications in surgical patients as medical patients did not bleed[1, 20–22]. Regarding HIT II patients, hirudin does not show cross-reactivitywith HIT II antibodies and is the first-line choice in HIT II patients.

Prostaglandins

Prostaglandins are products of the arachidonic acid metabolism and areproduced in the vascular endothelial tissue. The mechanism of action is an inhi-bition of thrombocyte aggregation without influence on the plasmatic coagula-tion pathway [6]. The half-life is �5 min. The synthetic prostaglandinsprostacyclin (PGI2) and alprostadil (PGE1) have been successfully used asanticoagulants in CRRT. Because of their potent vasodilatory action withpossible hemodynamic compromise and because of higher rates of bleedingepisodes using higher doses of prostaglandins, they have been used in combi-nation with HMW heparin in lower doses of both anticoagulants [23, 24].Systemic vasodilatory action can also be reduced in applying prostaglandins pre-filter [5, 7]. The studies performed with prostaglandins in a dose range of1–20 ng/kg/min showed a better or comparable filter patency in comparison toheparin with lower incidence of bleeding complications [23–25]. The idea is toreduce the incidence of bleeding complications through reduced dose whileblocking effectively not only the thrombin production but also the thrombocyteactivation and consecutively the inhibition of HMW heparin through plateletfactor 4 [7]. One study using low-dose prostacyclin alone in CRRT showedcomparable filter patency to HMW heparin without higher incidence of bleed-ing complications. However, 15% of the patients had to be treated because ofhypotension [26].

Sodium-Citrate

Systemic effective anticoagulation for CRRT is crucial if the patient is atrisk of bleeding. Sodium-citrate anticoagulation is a regional anticoagulationregime [11, 27–29]. Citrate binds ionized calcium which is required as co-factor


Tabl

e 1.

Ant

icoa

gula

tion

in C

RR

T –

sum

mar

y

Ant

icoa

gula

ntD

ose

(bol

us)

Dos

e (c

ont.)

Mon

itori

ngTa

rget

val

ues

Lim

itatio

nsA

ntic

oagu

latio

nco

sts/

24h

(€)

HM

W h

epar

in50

0–1,

000

IU/k

g30

0–1,

000

IU/k

g/h

aPT

T35

–60

sA

T I

II d

epen

denc

y,2

[6, 1

1, 1

2]co

ntra

indi

cate

d in

HIT

II

patie

nts

Dal

tepa

rin

15–2

0IU

/kg

5–10

IU/k

g/h

aXa

0.3–

0.6

aXa

IU/m

lno

rou

tine

mon

itori

ng, r

enal

62(L

MW

hep

arin

)el

imin

atio

n, n

ot f

ilter

ed, A

T I

II

[6, 1

2, 1

4, 1

5]de

pend

ency

, con

trai

ndic

ated

in

HIT

II

patie

nts

Dan

apar

oid

750

IU50

–150

IU/h

aXa

0.4

aXa

IU/m

lno

rou

tine

mon

itori

ng, l

ong

209

[6, 1

6]ha

lf-l

ife,

ren

al e

limin

atio

n,no

rou

tine

mon

itori

ng, n

otfi

ltere

d, A

T I

II d

epen

denc

y,

in v

itro

cros

s-re

activ

ity w

ith

HIT

II

antib

odie

s

Hir

udin

5–10

�g/

kg2–

5�

g/kg

/haP

TT

50

–60

sso

le r

enal

elim

inat

ion

with

25

[1, 6

, 20]

EC

T80

–100

sac

cum

ulat

ion

risk

, the

refo

re

hiru

din

300–

600

ng/m

lpr

imar

ily b

olus

dos

age

conc

entr

atio

n

Ilop

rost

2–10

ng/k

g/m

inno

neno

nehe

mod

ynam

ic in

stab

ility

1,

500

(pro

stag

land

in)

in h

ighe

r do

se, t

oget

her

with

[6

, 23,

24]

HM

W h

epar

in c

ontr

aind

icat

ed

in H

IT I

I pa

tient

s

Na-

citr

ate

Na-

citr

ate

is d

osed

acc

ordi

ng to

ioni

zed

calc

ium

0.

25m

mol

/lhy

pern

atre

mia

, met

abol

ic

20[6

, 11,

27]

post

-filt

er; C

aCl 2

is d

osed

acc

ordi

ng to

ioni

zed

1.0–

1.3

mm

ol/l

alka

losi

s an

d ac

idos

is,

calc

ium

in th

e bl

ood

hypo

calc

emia

, in

HIT

II

patie

nts

alte

rnat

ive

antic

oagu

latio

n is

man

dato

ry

Con

t.�

Con

tinuo

us; H

MW

�hi

gh m

olec

ular

wei

ght;

aPT

T�

activ

ated

par

tial t

hrom

bopl

astin

tim

e; A

T II

I�

antit

hrom

bin

III;

LM

W�

low

mol

ecul

ar w

eigh

t; aX

a�

anti-

fact

or X

a-ac

tivity

; HIT

II

�he

pari

n-in

duce

d th

rom

bocy

tope

nia;

EC

T�

ecar

in c

lotti

ng ti

me.


in multiple steps of the coagulation cascade. 4% tri-sodium-citrate solution isapplied pre-filter in a rate adapted to the measured ionized calcium post-filterwhich should be 0.25 mmol/l ideally. Calcium free, low-sodium and bicarbonatelow or free dialysate and replacement solution are used. Calcium replacement isdone through a separate venous line directly into the patient according to themeasured ionized calcium in the patients’ blood which should be hold at normallevels. In this way, just the dialysis circuit is anticoagulated. Approximately 60%of the sodium-citrate applied is extracted through the filter. Citrate is metabo-lized to bicarbonate in the liver [11, 27–29]. The filter run time achieved in theperformed studies were extensively longer than with other anticoagulationregimes without causing bleeding complications in high risk patients. However,in performing citrate anticoagulation some problems can arise regarding tometabolic disorders which must be accounted for [12]. Possible sodium overloadis taken into account by using hypotonic dialysate and replacement solution.Metabolic alkalosis, caused by the metabolism of citrate to bicarbonate in theliver, can be avoided in most cases by using bicarbonate low or free dialysate andreplacement solution. If the patient develops liver insufficiency citrate is notmetabolized to bicarbonate which can lead to metabolic acidosis and elevatedtotal calcium content due to an accumulation of calcium-citrate complexes [30].Therefore, accurate monitoring of the pH and electrolytes of the patient ismandatory when performing sodium-citrate anticoagulation [7, 12]. In the lastyears, many different protocols have been studied in order to reduce the meta-bolic derangement and ease the procedure without remarkable better results[11, 31, 32]. In terms of HIT II it is important to mention that citrate anticoagu-lation is no alternative for heparin anticoagulation. The patient must be system-ically anticoagulated to counteract HIT II pathophysiology.

In conclusion, considering the characteristics of an ideal anticoagulant(rapid onset, short time of action, rapid elimination independent of renal andliver function, effective antagonization, routine monitoring, minimal sideeffects) HMW heparin remains as first-line anticoagulant. Also, taking theimportant point of cost-effectiveness into account HMW heparin remains thecheapest anticoagulation regime for CRRT. In terms of costs citrate anticoag-ulation seems to be cheaper than HMW heparin anticoagulation if overallcosts, including the hemofiltration systems needed, are compared [unpubl.data]. The other alternatives are clearly more expensive but must be taken intoconsideration in patients with HIT II, with the exclusion of LMW heparin.Kozek-Langenecker et al. [24] could show, that the overall costs of CRRT withprostaglandins were more intensive than with HMW heparin. Prostaglandinsseem to offer some advantages just in combination with HMW heparin,excluding their use for patients with HIT II. Danaparoid is used effectivelyin HIT II patients, although a risk for cross-reactivity remains. Hirudin is


effective but extreme caution must be taken because of accumulation espe-cially in patients with risk of bleeding. A promising alternative could be thedirect thrombin inhibitor argatroban. This anticoagulant is eliminated throughthe liver with a half-life of 21–61 min and no cross-reactivity with HIT II anti-bodies could be found [33, 34]. A few case reports have shown preliminarysatisfactory results regarding effectiveness and safety [33, 34]. Experiencewith heparin-coated or other regional anticoagulation systems involvingimmobilized heparinase or protamine remain experimental at this time point[35, 36]. Also experimental, but potentially promising, are nitric oxide (NO)-releasing silicon rubbers. The potent inhibition of platelet activation causedby NO did show an improved thromboresistivity in an extracorporeal circuitanimal study [37].

References

1 Vargas Hein O, von Heymann C, Lipps M, Ziemer S, Ronco C, Neumayer HH, Morgera S, Welte M,Kox WJ, Spies C: Hirudin versus heparin for anticoagulation in continuous renal replacementtherapy. Intens Care Med 2001;27:673–679.

2 Vargas Hein O, Spies C, Kox WJ: Renal dysfunction in the Perioperative Periode; in Gullo A (ed):Anaesthesia, Pain, Intensive Care and Emergency Medicine (A.P.I.C.E.). Milano, Springer, 2001,pp 309–315.

3 Kellum JA, Angus DC, Johnson JP, Leblanc M, Griffin M, Ramakrishnan N, Linde-Zwirble WT:Continuous versus intermittent renal replacement therapy: A meta-analysis. Intens Care Med2002;1:29–37.

4 Van de Wetering J, Westendorp RGJ, van der Hoeven JG, Stolk B, Feuth JD, Chang PC: Heparin usein continuous renal replacement procedures: The struggle between filter coagulation and patienthemorrhage. J Am Soc Nephrol 1996;1:145–150.

5 Suhlman RI, Singer M, Rock J: Keeping the Circuit Open: Lessons from the Lab. Blood Purif2002;20:275–281.

6 Vargas-Hein O-R: Antikoagulation für die CVVHF. J Anästh Intensivbehandl 2001;3:136–138.7 Schetz M: Anticoagulation in continuous renal replacement therapy. Contrib Nephrol. Basel,

Karger, 2001, vol 132, pp 283–303.8 Davenport A: The coagulation system in the critically ill patient with acute renal failure and the

effect of an extracorporeal circuit. Am J Kidney Dis 1997;5(suppl 4):20–27.9 Metha RL: Anticoagulation strategies for continuous renal replacement therapies: What works?

Am J Kidney Dis 1996;5(suppl 3):8–14.10 Hirsh J, Anand SS, Halperin JL, Fuster V: Mechanism of action and pharmacology of unfractionated

heparin. Arterioscler Thromb Vasc Biol 2001;7:1094–1096.11 Palsson R, Niles JL: Regional citrate anticoagulation in continuous venovenous hemofiltration in

critically ill patients with a high risk of bleeding. Kidney Int 1999;5:1991–1997.12 Abramson S, Niles JL: Anticoagulation in continuous renal replacement therapy. Curr Opin

Nephrol Hypertens 1999;6:701–707.13 Jeffrey RF, Khan AA, Douglas JT, Will EJ, Davison AM: Anticoagulation with low molecular

weight heparin (Fragmin) during continuous hemodialysis in the intensive care unit. Artif Organs1993;8:717–720.

14 Reeves JH, Cumming AR, Gallagher L, O’Brien JL, Santamaria JD: A controlled trial of low-molecular-weight heparin (dalteparin) versus unfractionated heparin as anticoagulant duringcontinuous venovenous hemodialysis with filtration. Crit Care Med 1999;10:2224–2228.


15 de Pont AC, Oudemans-van Straaten HM, Roozendaal KJ, Zandstra DF: Nadroparin versus dal-teparin anticoagulation in high-volume, continuous venovenous hemofiltration: A double-blind,randomized, crossover study. Crit Care Med 2000;2:421–425.

16 Lindhoff-Last E, Betz C, Bauersachs R: Use of a low-molecular-weight heparinoid (danaparoidsodium) for continuous renal replacement therapy in intensive care patients. Clin Appl ThrombHemost 2001;4:300–304.

17 Schenk JF, Pindur G, Stephan B, Morsdorf S, Mertzlufft F, Kroll H, Wenzel E, Seyfert UT: On theprophylactic and therapeutic use of danaparoid sodium (Orgaran) in patients with heparin-inducedthrombocytopenia. Clin Appl Thromb Hemost 2003;1:25–32.

18 Plassat R, Cognet F, Ternisien C, Menoret N, Dubus-Bausiere V, Brunel P, Bontoux L, Bernat C,Richard I: Heparin induced thrombocytopenia: Case report with acute thrombotic complicationsand literature review. Ann Readapt Med Phys 2002;5:216–223.

19 Frank RD, Farber H, Stefanidis I, Lanzmich R, Kierdorf HP: Hirudin elimination by hemofiltration:A comparative in vitro study of different membranes. Kidney Int 1999;56(suppl 72):41–45.

20 Vargas Hein O, von Heymann C, Diehl T, Ziemer S, Ronco C, Morgera S, Siebert G, Kox WJ,Neumayer HH, Spies C: Intermittent hirudin versus continuous heparin for anticoagulation in con-tinuous renal replacement therapy; in press.

21 Kern H, Ziemer S, Kox WJ: Bleeding after intermittent or continuous r-hirudin during CVVH.Intens Care Med 1999;25:1311–1314.

22 Fischer KG, Van de Loo A, Böhler J: Recombinant hirudin (lepirudin) as anticoagulant inintensive care patients treated with continuous hemodialysis. Kidney Int 1999;56(suppl 72):46–50.

23 Kozek-Langenecker SA, Spiss CK, Gamsjager T, Domenig C, Zimpfer M: Anticoagulation withprostaglandins and unfractionated heparin during continuous venovenous haemofiltration: A ran-domized controlled trial. Wien Klin Wochenschr 2002;3:96–101.

24 Kozek-Langenecker SA, Kettner SC, Oismueller C, Gonano C, Speiser W, Zimpfer M:Anticoagulation with prostaglandin E1 and unfractionated heparin during continuous venovenoushemofiltration. Crit Care Med 1998;7:1208–1212.

25 Greuel MS, Birnbaum J, Klotz E, Saalmann R, Spies C, Lehmann C, Kox WJ: Der Einfluss vonIloprost auf die Filterlaufzeit, die Thrombozytenzahl und die Retentionsparameter im Rahmeneiner CVVH-Behandlung. J Anästh Intensivmed 2003;2:17.

26 Fiaccadori E, Maggiore U, Rotelli C, Minari M, Melfa L, Cappe G, Cabassi A: Continuoushaemofiltration in acute renal failure with prostacyclin as the sole anti-haemostatic agent. IntensCare Med 2002;5:586–593.

27 Mehta R, McDonald B, Aguilar M, Ward D: Regional citrate anticoagulation for continuous arte-riovenous hemodialysis in critically ill patients. Kidney Int 1990;38:976–981.

28 Gabutti L, Marone C, Colucci G, Duchini F, Schönholzer C: Citrate anticoagulation in continuousvenovenous hemodiafiltration: A metabolic challenge. Intens Care Med 2002;28:1419–1425.

29 Pinnick E, Wiegmanson TE, Diederich DA: Regional citrate anticoagulation for hemodialysis ina patient at high risk for bleeding. N Engl J Med 1983;308:258–261.

30 Meier-Kriesche HU, Gitomer J, Finkel K, DuBose T: Increased total to ionized calcium ratio dur-ing continuous venovenous hemodialysis with regional citrate anticoagulation. Crit Care Med.2001;4:748–752.

31 Tobe SW, Aujla P, Walele AA, Oliver MJ, Naimark DM, Perkins NJ, Beardsall M: A novel regionalcitrate anticoagulation protocol for CRRT using only commercially available solutions. J Crit Care2003;2:121–129.

32 Mitchell A, Daul AE, Beiderlinden M, Schafers RF, Heemann U, Kribben A, Peters J, Philipp T,Wenzel RR: A new system for regional citrate anticoagulation in continuous venovenous hemodialy-sis (CVVHD). Clin Nephrol. 2003;2:106–114.

33 Dager WE, White RH: Argatroban for heparin-induced thrombocytopenia in hepato-renal failureand CVVHD. Ann Pharmacother 2003;9:1232–1236.

34 Ohteki H, Furukawa K, Ohnishi H, Narita Y, Sakai M, Doi K: Clinical experience of Argatrobanfor anticoagulation in cardiovascular surgery. Jpn J Thorac Cardiovasc Surg 2000;1:39–46.

35 Zhang Y, Singh VK, Yang VC: Novel approach for optimizing the capacity and efficacy of a pro-tamine filter for clinical extracorporeal heparin removal. ASAIO 1998;5:M368–M373.


36 Ameer GA, Barabino G, Sasisekharan R, Harmon W, Cooney CL, Langer R: Ex vivo evaluationof a Taylor-Couette flow, immobilized heparinase I device for clinical application. Proc Natl AcadSci USA 1999;5:2350–2355.

37 Zhang H, Annich GM, Miskulin J, Stankiewicz K, Osterholzer K, Merz SI, Bartlett RH,Meyerhoff ME: Nitric oxide-releasing fumed silica particles: Synthesis, characterization, and bio-medical application. J Am Chem Soc 2003;17:5015–5024.

Dr. med. Ortrud Vargas HeinDepartment of Anesthesiology and Intensive Care, University Hospital CharitéCampus Mitte, Schumannstrasse 20/21, DE–10117 Berlin (Germany)Tel. �49 30 450 631043, Fax �49 30 450 531911, E-Mail [email protected]


Replacement and Dialysate Fluids for Patients with Acute Renal FailureTreated by Continuous Veno-VenousHaemofiltration and/orHaemodiafiltration

Andrew Davenport

Centre for Nephrology, Royal Free Hospital Campus, Royal Free and University College Hospital Medical School, London, UK

When Peter Kramer first described continuous arteriovenous haemofiltra-tion, it was a technique designed to provide ultrafiltration and was not asubstitute for intermittent haemodialysis, so fluid replacement was not required[1]. With the advances in CRRT over the last three decades, CRRT has becomean efficient treatment for acute renal failure in the ICU. The ideal replacementor dialysate solution would bear resemblance to normal plasma and extracellu-lar fluid (table 1). Originally peritoneal dialysates were used, as these were aready source of sterile fluid, and became the basis for substitution anddialysates. Recent studies have shown that survival can potentially be improvedwith larger exchanges of fluid and/or dialysate flows [2], and is supported by invivo animal experimental data [3]. This trend for high volume CRRT, coupledwith the resurgence in citrate based anticoagulation, has necessitated thedevelopment of specialised dialysate and substitution fluids.

Sterility

During haemofiltration, substitution fluids are directly infused, andsimilarly during continuous dialysis, due to back filtration, again there is directinfusion of dialysate. Thus sterile fluids are required. Commercially availablefluids are heat sterilised. Heat sterilisation of glucose containing solutions leads

Davenport 318

to the formation of glucose degradation products (GDPs). In vitro these lead toincreased carbonyl stress, and have been reported to have cytotoxic effects andimpair macrophage and other inflammatory cell function [4]. There is no dataon the effect of exposure to GDPs in the critically ill patient with acute renalfailure (ARF). Although peritoneal dialysates contain hypertonic glucose, andare therefore hyperosmolar, the relative blood flow through the haemofilter/dialyzer to replacement solution/dialysate flow, means that there is only�1 mosm/kg increase in plasma osmolality. The newer dialysates, and substitu-tion fluids are isotonic and glucose free, so do not contain GDPs.

Standard CRRT is more expensive than daily intermittent haemodialysis,primarily due to the cost of the commercially available sterile dialysates/substi-tution fluids. These costs increase with high volume CRRT. To reduce costs,some centres prepare their own fluids to the same standards as ultra-puredialysates, used in haemodiafiltration [5]. Quality control of these on-site fluidsis paramount, not only in terms of bacterial and endotoxin contamination, butalso chemical contamination, so that trace elements, nitrates and organiccompounds, such as chloramines have been removed.

Although the AAMMI and European Dialysis and Transplant Associationhave issued guidelines for the quality of dialysate [5], there is no consensus onthe purity of ultrapure dialysate, in terms of how to quantify bacterial and endo-toxin contamination [6]. Although contamination of dialysate and/or substitu-tion fluids with environmental microorganisms and small amounts of endotoxinhas in vivo effects [7], it is unknown whether this has significant clinicalimpact. However, in the absence of specific legislation [6], it is recommendedthat in Europe, the European Community Directive 93/42/EEC on medicaldevices is followed for on-line fluid production.

Table 1. Differences in the electrolyte composition of peritonealdialysate (Baxter Health Care, Deerfield, USA), plasma, extracellular fluid(ECF) and intracellular fluid (ICF): all units expressed in mmol/l

Fluid Dianeal® Plasma ECF ICF

Sodium 132 143 144 10Potassium 0 4.0 4.0 155Calcium 1.75 2.5 1.25 2.0Magnesium 0.75 1.0 0.75 15Chloride 102 100 114 2.0Bicarbonate 0 27 30 10Lactate 45 1.0 1.5 variableGlucose 72 4.0 4.0 4.0

Dialysates and Substitution Fluids for CRRT 319

Electrolyte Composition

The first dialysates and substitution fluids were based on peritonealdialysate (table 1), which is hyponatraemic. As critically ill patients with ARFare often enterally or parenterally fed with hyponatraemic feeds, receivemultiple drug infusions, often in dextrose, then patients could become hypona-traemic. Thus the newer specialised fluids usually have a sodium concentrationwithin the normal plasma range (tables 2, 3).

Diffusive sodium transport depends upon the concentration gradientbetween the plasma and the dialysate. However, the plasma sodium activity – the

Table 2. Standard lactate based fluids for haemofiltration and/orhaemodiafiltration mmol/l

Fluid Baxter® Filtrasol® Lactasol® Baxter UK

Sodium 140 140 140 140Potassium 1 0 0 0Calcium 1.60 1.60 1.45 1.5Magnesium 0.80 0.75 0.75 0.75Chloride 100 100 105 110Bicarbonate 0 0 0 0Lactate 46 45 40 35Glucose 10.8 0 0 0

Baxter® � Baxter Health Care, Sydney, Australia; Filtrasol® � PharmaciaAB, Uppsala, Sweden; Lactasol® � Hospal, Gottenberg, Sweden.

Table 3. Differences in electrolyte composition of bicarbonate based fluids used forCRRT (mmol/l)

Solution Normocarb® Bicarbonate Bicarbonate Bicarbonate Lactate free*

Sodium 140 155 140 140 110Calcium 0 1.8 1.5 1.75 1.75Magnesium 0.75 0.77 0.5 0.5 0.75Chloride 106.5 120 109 110 110Bicarbonate 35 40 35 34.5 0Lactate 0 3.0 0 3.0 0Reference 22 20 21 24 23

*Lactate-free solution – sodium bicarbonate added separately to achieve zero base excess.

Davenport 320

amount of sodium available for potential diffusion, is not the same as the plasmasodium concentration. In healthy patients plasma sodium activity is similar toplasma concentration. This is the result of two opposing effects, firstly plasmawater sodium is about 7% greater than that of plasma sodium concentration, dueto the volume occupied by plasma proteins, but then plasma sodium activity isreduced by plasma sodium binding to proteins, and proteins deposited on thedialyzer membrane. As proteins are generally negatively charged, sodium is heldback to maintain electrical neutrality, in accordance with the Gibbs-Donnaneffect. In the septic ICU patient, the plasma sodium is usually low or at the lowerend of the normal range, due to increased capillary permeability, coupled withhypoalbuminaemia. If the patient is also on a low sodium diet, and in receipt ofno other sodium containing fluids, then CRRT performed as dialysis alone,could provide a modest daily positive sodium balance.

Most ICUs offer haemofiltration and/or haemodiafiltration. The convec-tive losses of haemofiltration result in sodium losses due to solvent drag.However, again due to the Gibbs-Donnan effect, the sieving coefficient ofsodium will be �1.0. Thus if the replacement fluids have a sodium concentra-tion greater than plasma, patients will have a positive sodium balance [8]. Thesieving coefficient for sodium is greater for pre- than post-dilutional fluidreplacement (table 4), due to the effect of diluting proteins, so reducing theGibbs-Donnan effect. This is exemplified by comparing high volume haemofil-tration (6 litres/h), when the sodium sieving coefficient fell from 0.96 with pre-dilution to 0.94 with post-dilution. Although these changes in sievingcoefficient appear small, when one considers the volumes exchanged, patientswould be in a daily positive sodium balance of almost 250 mmol/day. Thus, if

Table 4. Differences in sieving coefficients of pre- andpost-dilutional continuous venovenous haemofiltration (2 litreexchanges using Baxter UK substitution fluid and 1.2 m2

polysulfone hemofilter, blood pump speed 200 ml/min)

Sieving coefficient Pre-filter Post-filter

Urea 0.74 1.06Creatinine 0.72 1.04Sodium 0.97 1.00Potassium 0.918 0.923Calcium 0.709 0.706Phosphate 0.71 1.04Chloride 1.01 1.06Protein 0.001 0.003


patients are going to be treated by high volume CRRT for prolonged periods,then lower sodium solutions will be required.

All currently commercially available fluids contain calcium and magne-sium, designed to maintain the plasma levels within the normal range,although the concentrations vary from manufacturer to manufacturer(calcium 1.65–2.0 and magnesium 0.75–0.77). Whereas the magnesiumconcentrations are set towards the lower limit of the normal range, the cal-cium concentrations are higher than normal plasma ionised calcium, andstudies from chronic dialysis have suggested a zero calcium balance with adialysate calcium concentration of 1.25 mmol/l. Although more recent studieswith daily extended hours haemodialysis have reported that higher calcium of1.5 mmol/l was required to maintain a zero calcium balance [10]. In ARF, cal-cium binding can be affected by albumin concentration and acidosis. Duringdialysis the fraction of diffusible calcium decreases, due to both ultrafiltrationleading to increased plasma albumin concentration, and the rise in plasmabicarbonate, which increases plasma protein binding. Thus during CRRTusing dialysis, calcium balance will differ according to whether lactate orbicarbonate dialysates are used. However, using these high calcium dialysates(tables 2, 3) will lead to a net positive calcium influx into the patient.Similarly during convective forms of haemofiltration, the calcium sievingcoefficient (table 4), would suggest that patients develop a positive dailycalcium balance. During high volume haemofiltration, the calcium sievingcoefficient fell from 0.78 to 0.6 when the site of fluid replacement wasswitched from the pre- to post-dilution mode [9]. As there is no apparentincrease in plasma calcium concentration, this would suggest a net increasein intracellular calcium.

Magnesium concentrations in the dialysates/replacement fluids are similarto free diffusible plasma magnesium (normal plasma 0.8–1.2 mmol/l of which60–65% is free). During dialysis mode CRRT, as with calcium the concentra-tion of diffusible magnesium will decrease during passage through the dialyzer,and this may allow a positive magnesium balance. As with calcium duringhaemofiltration, the sieving coefficient of magnesium was greater for pre-dilution compared to post-dilution (table 4). Similarly, during high volumehaemofiltration, the magnesium sieving coefficient changed from around 0.84with pre-dilution, to 0.70 with post-dilution. Although this may suggest a pos-itive magnesium balance with high volume post-dilution mode, there were nochanges in plasma magnesium. However, magnesium is primarily an intracel-lular cation, so that changes in magnesium flux may occur without discernablechanges in plasma concentration.

Typically, the commercial dialysates/replacement solutions are potassium andphosphate free, as patients with ARF require potassium and phosphate removal.

Davenport 322

Some patients may then require potassium and/or phosphate supplementation, inaddition to that in administered in parenteral/enteral nutrition [11, 12].

Acid-Base Balance

One of the main purposes of CRRT is to correct the metabolic acidosis,which develops in patients with ARF. Although bicarbonate is the naturalbuffer, bicarbonate based solutions have not been available until recent times,as bicarbonate is not stable in the presence of calcium and magnesium ions, andthere is an increased risk of bacterial contamination. Initially peritonealdialysate fluids were used for CRRT, as these were sterile fluids with aprolonged shelf storage life, containing acetate or a racemic mixture of lactate,as the anionic base. Acetate and l-lactate are metabolised, indirectly in liver andmuscle, to bicarbonate on an equimolar basis. D-Lactate, is not the naturallyoccurring isomer of lactate, and has a different and slower metabolic pathwaythan l-lactate, mainly by skeletal muscle. In ARF D-lactate metabolism isreduced and D-lactate can accumulate [13]. Specialised solutions were thendeveloped for CRRT based on l-lactate and acetate. However, studies in patientsundergoing intermittent haemodialysis using acetate based dialysate, noted thatacetate accumulated during dialysis, resulting in hyperacetataemia, and cardio-vascular instability, typically in patients with left ventricular dysfunction [14].Replacement of acetate by bicarbonate was reported to improve both patientintra-dialytic symptoms and cardiovascular tolerance. Endogenous lactatemetabolism has been estimated at 1,500 mmol/day in healthy subjects. Duringhigh volume CRRT using lactate based solutions, the rate of lactate adminis-tration may well exceeded hepatic metabolism (normal �100 mmol/h inhealthy subjects and �0.6 mmol/kg/h in patients with ARF) [15] resulted inhyperlactataemia. In cases of severe liver failure, the obligatory losses of bicar-bonate and lactate during intermittent high volume haemofiltration can exceedthe rate of bicarbonate generation, so resulting in an increase in arterialhydrogen ion concentration, and worsening acid-base status [16].

During CRRT the rate of administration of these substitution fluids/dialysates is substantially less than that during intermittent haemodialysis and/orfiltration, and thus the accumulation of acetate or lactate with a worsening acid-base status has rarely been reported [17]. Even so, the currently available com-mercial lactate based solutions can lead to acid-base disturbances. Currentlylactate concentrations vary from 45 mmol/l (chloride 100 mmol/l) down to35 mmol/l (chloride 110 mmol/l) (table 2). These differences in solution compo-sition can lead to clinical consequences, with the development of a hypochlo-raemic metabolic alkalosis on one hand after several days of CRRT using a high


lactate low chloride solution, and a hyperchloraemic metabolic acidosis on theother, when a low lactate high chloride fluid is used [19]. In short term studiesof high volume haemofiltration, although patients became hyperlactataemic,with a fall in pH and base excess, this was limited by hypochloraemia by usinga replacement solution with a lower chloride concentration, and increased lossof strong acids [19]. Chloride, being negatively charged has a sieving coefficient�1.0, and studies have shown that pre-dilution mode has a higher sieving coef-ficient than post dilution (table 4) [9]. Thus pre-dilution mode for high volumeexchanges will have least effect on acid-base balance, by increasing losses ofchloride, CO2, and other acids.

Whereas bicarbonate based dialysate is universally used for treatingpatients with ARF by intermittent haemodialysis, the use of commercially avail-able bicarbonate based fluids for CRRT has been delayed due to problems inmanufacture and storage of fluids containing both bicarbonate and the divalentcations, calcium and magnesium. Such fluids are at greater risk of bacterialcontamination, due to the higher pH, and precipitation of calcium and magne-sium carbonate. Commercially available bicarbonate based fluids are nowavailable; as either solutions containing bicarbonate and calcium, which have alimited shelf life [20]; or as twin bag systems, one containing the bicarbonate,the other the electrolytes and the two compartments being mixed immediatelyprior to use [21, 22]; or as ‘buffer free’ or ‘lactate free’ solutions, with bicar-bonate being added distally [23].

The question arises as to whether these more expensive bicarbonatesolutions provide any advantages over standard acetate and lactate based fluids.Heering et al. [21] reported that both bicarbonate and lactate based solutionsresulted in better control of acid-base balance, and improved cardiovasculartolerance compared to acetate. Similarly, others have reported superior controlof acidosis and cardiovascular stability with bicarbonate when compared tolactate [23, 24]. These latter studies were biased by their patient recruitment;ARF post cardiac surgery and liver failure, but both showed better correction ofacidosis with less total anionic base administered, compared to lactate. Not allstudies have reported such a positive effect for bicarbonate based solutions[20, 21], and this may be due to the differences in the amount of bicarbonateand the other components of the solutions (table 3), as the best results wereobtained when a variable amount of bicarbonate was titrated to meet individualpatient requirements [23].

During CRRT bicarbonate is given as a net iso-osmolar solution, at a rateof 60–80 mmol/h, which markedly differs from that during acute bolus adminis-tration (60 mmol/min). Although in vitro experiments suggested that the acuterapid bicarbonate administration could result in increased intracellular acidosis,due to the generation and passage of CO2 into cells. More recent studies,

Davenport 324

designed specifically to emulate the clinical situation, have reported an increasein intracellular pH [25]. Similarly, although the rapid administration of bicar-bonate causes an increase in end-tidal CO2 and PaCO2, bicarbonate based CRRT,has not been reported to result in a significant increase in PaCO2, such thatchanges in ventilatory settings were required [21, 22]. The advantage of using abuffer free replacement solution, and then titrating the dose of bicarbonate to atarget base excess, means that acid-base control can be achieved in ventilatedpatients treated with permissive hypercapnia, whereas this could be moreproblematical in patients given a fixed bicarbonate concentrate.

CRRT Using Citrate Anticoagulation

Citrate is increasingly being used as a regional anticoagulant for CRRT,and as each citrate molecule is indirectly metabolised through to threebicarbonates, then the need for additional anionic base varies from circuitdesign to design. When citrate anticoagulation is used, then the predilutionalfluid traditionally has been calcium and bicarbonate free [26]. A specialhyponatraemic dialysate (117 mmol/l) is required containing no calcium oralkali, with a relatively high chloride content of 122.5 mmol/l [26]. The hypona-traemic dialysate is required because of the high sodium load pre-filter due tothe combination of trisodium citrate and the pre-dilution infusion of normalsaline. Dextrose was then added to the dialysate to maintain osmolality. Othershave used citrate anticoagulation for continuous haemodiafiltration, and whilstusing a similar dialysate, found that they had to add bicarbonate to the post-filter replacement fluid (table 5) [27]. When citrate is used to anticoagulatehaemofiltration circuits, then customised post-dilutional fluids are required[28] (table 5). As calcium is complexed with citrate and lost into the dialysate,

Table 5. Differences in electrolyte composition of replacement and dialysate fluidsused for citrate based CRRT (units mmol/l, or in case of dextrose %)

Solution Replacement Dialysate Replacement Dialysate Replacement

Sodium 150 117 140 117 140Calcium 0 0 0 0 0Magnesium 0.70 0.75 0 0.75 0.75Chloride 121 121.5 140 117 101.5Bicarbonate 33.3 0 0 0 0Dextrose 0 0 0 2.5 0.2Reference 27 27 26 26 28


or ultra-filtrate, calcium must be infused as calcium chloride centrally, tomaintain a normal ionised calcium returning to the patient [26].

With the increasing interest in citrate anticoagulation, centers without thefacility to produce their own specialised dialysates and replacement fluids, haveused commercially available bicarbonate based calcium containing solutions asdialysates and replacement fluids [28]. As each citrate molecule is eventuallymetabolised, to three bicarbonate molecules, up to 26% of patients will developa mild metabolic alkalosis, more common in those given additional bicarbon-ate, patients with hepatic dysfunction, and those given blood products(containing acid citrate dextrose anticoagulant). Patients given commerciallyavailable bicarbonate based fluids have the highest incidence of alkalosis, butthis can often be simply managed by not opening the bicarbonate compartmentof the twin bag systems, and reducing the citrate load by using acid-citrate dex-trose [29]. Alkalosis also occurs more often with haemofiltration circuits, as theclearance of citrate is less than that during dialysis [28].Modest alkalosis can bemanaged by either reducing the citrate infusion rate, or increasing bicarbonatelosses and chloride gain. This can be done most effectively by increasing thedialysate flow, or less effectively by increasing the pre-dilution infusion. Inextreme cases, 0.2 M HCl can be infused through a major vein [26].

Other metabolic complications include hypercitrataemia, which has beenreported in 10–15% of patients, usually those with some degree of hepaticdysfunction. This can usually readily be corrected by slowing the citrate infu-sion to 1.5–2.5% of the blood flow. Typically, hypercitrataemia occurs in thesetting of alkalosis. Hypernatraemia occurs in less than 10% of patients [30],more commonly with trisodium citrate and haemofiltration, as the prefiltersodium load can not be removed as effectively as with dialysis, or when aninsufficiently hyponatraemic dialysate/substitution fluid has been used.

Occasionally, patients may develop hypercitrataemia, hypercalcaemia andmetabolic acidosis [30]. This is usually in the setting of ARF with rhabdomyol-ysis and initial hypocalcaemia. Despite a normal ionised plasma calcium, totalplasma calciums in excess of 4.0 mmol/l have been observed, in association withhypercitrataemia (plasma citrate �20 mmol/l, normal range 0.07–0.14 mmol/l)[31]. Such hypercitrataemia suggests hepatic dysfunction, and if the citrate cannot be metabolised through to bicarbonate, then the patient will become acidotic,due to the loss of plasma bicarbonate through the dialyzer.

Temperature

Centres differ in their approach to heating replacement solutions and/ordialysates [32]. During CRRT significant thermal energy is lost, which may

Davenport 326

have advantages in improving cardiovascular stability, but could also potentiallyadversely effect nutritional balance, immune function, and increase the risk ofclotting of the CRRT circuit. In addition cooling could theoretically reducesolute clearances both during diffusive techniques, as diffusion coefficients aretemperature dependent, and also convective therapies by increasing membraneprotein deposition.

Even warming both the dialysate and replacement solutions to 37�C stillleads to thermal energy losses, which are increased with larger exchange/dialysate volumes.

Conclusions

The advance of CRRT from simple arteriovenous ultrafiltration to a treat-ment using 50 plus litres a day has spurred on the creation of specialiseddialysate/replacement solutions. As the volumes used are large, quality controlis vitally important not only in terms of electrolyte composition, but also pre-venting bacterial and endotoxin contamination. There are small differences insolute flux between dialysis and both pre- and post-dilution haemofiltration. Ifhigh volume haemofiltration is to be used for any extended period, then pre-dilution mode will cause less electrolyte imbalance than post-dilution.

Lactate remains the standard anionic base, as this increases fluid storagetime and reduces the risk of bacterial contamination. The composition varyingbetween manufacturers, such that those patients given high lactate, low chloridefluid, will be prone to hypochloraemic alkalosis, whereas those give a lowerlactate but higher chloride can develop hyperchloraemic acidosis. There is anincreasing usage of citrate anticoagulation, and the citrate load provides asupply of anionic base. Varying protocols have been developed to cope with thesodium and citrate loads, but now several centers are using citrate based CRRTwith commercial fluids.

Hypotensive patients with severe tissue acidosis and/or liver failure, andthose receiving high volume CRRT may be unable to convert the lactate, orcitrate load effectively, resulting in hyperlactataemia or hypercitrataemia withacidosis. In these circumstances bicarbonate based fluids are advantageous, butare currently somewhat more expensive that standard lactate fluids. The use ofbicarbonate based dialysate/replacement fluid during CRRT does not masklactate overproduction, and lactate remains a reliable marker of tissueoxygenation in patients treated by CRRT. Bicarbonate CRRT alone does nottreat the underlying cause of metabolic acidosis, but allows better control ofacidosis, so allowing time for the institution of other therapies designed toreverse the underlying cause.


References

1 Kramer P, Wigger W, Rieger J, Mathaei D, Scheler F: Arteriovenous hemofiltration: A new andsimple method for the treatment of over hydrated patients resistant to diuretics. Klin Wochenschr1977;5:1121–1122.

2 Ronco C, Bellomo R, Hamel P, Brendolan A, Dan M, Piccinni P, La Greca G: Effects of differentdoses in continuous veno-venous hemofiltration on outcomes of acute renal failure. Lancet 2000;356:26–30.

3 Grootendorst AF, van Saase JLCM: Blood purification by hemofiltration in septic shock andmultiple organ dysfunction syndrome patients; in Bellomo R, Ronco C (eds): Update in IntensiveCare and Emergency Medicine. Heidelberg, Springer-Verlag, 1995, vol 20, pp 311–325.

4 Davenport A: Peritoneal host defence in patients with end-stage renal failure treated by peritonealdialysis; in Sweny P, Rubin R, Tolkoff-Rubin N (eds): The Infectious Complications of RenalDisease. Oxford, Oxford University Press, 2003, pp 41–68.

5 The EBPG Expert Group on Haemodialysis: European Best Practice Guidelines for haemodialy-sis. Part 1; section IV. Dialysis fluid purity. Nephrol Dial Transplant 2002;17(suppl 7):45–62.

6 Ministère de l’Emploi et de la Solidarieté: Circulaire relative aux specifications techniques de lasecuritié de pratique de l’hémofiltration et de l’hémodiafiltration en ligne dans établissements desanté. Republique Française, Circulaire DGS/DH/AFSSAPS no 311, June 7, 2000.

7 Schindler R: Passage of LPS and non-LPS like cytokine inducing substances across syntheticdialyzer membranes. Blood Purif 2003;21:343–344.

8 Morimatsu H, Uchino S, Bellomo R, Ronco C: Continuous renal replacement therapy: Doestechnique influence electrolyte and bicarbonate control? Int J Artif Organs 2003;26:289–296.

9 Uchino S, Cole L, Morimatsu H, Goldsmith D, Ronco C, Bellomo R: Solute mass balance duringisovolemic high volume hemofiltration. Intens Care Med 2003;9:1541–1546.

10 Lindsay RM, Leitch R, Heidenheim AP, Kortas C: Daily short hours and long slow nocturnalhemodialysis: The London, Ontario study. Blood Purif 2003;21:335–340.

11 Davenport A: Hemofiltration in patients with fulminant hepatic failure. Lancet 1991;338:1604.12 Davenport A, Kirby SA: Hemofiltration/dialysis treatment in patients with acute renal failure.

Care Crit Ill 1996;12:54–58.13 Veech RL: The untoward effects on anions of dialysis fluids. Kid Int 1991;34:587–589. 14 Katzarski KS: Fluid state and blood pressure control in patients on maintenance hemodialysis,

MD thesis, Stockholm, 1999.15 Wright DA, Forni LG, Carr LG, Treacher DF, Hilton PJ: Use of continuous hemofiltration assess

the rate of lactate metabolism in acute renal failure. Clin Sci 1996;90:507–510.16 Davenport A, Will EJ, Davison AM: Hyperlactatemia and metabolic acidosis during hemofiltra-

tion using lactate buffered fluids. Nephron 1991;59:461–465.17 Nimmo GR, Grant IS, Mackenzie SJ: Lactate and acid base changes in the critically ill. Post-Grad

Med J 1991;67(suppl 1):S56–S61.18 Davenport A, Worth DP, Will EJ: Hypochloremic alkalosis after high flux continuous hemofiltra-

tion and continuous arterio-venous hemofiltration with dialysis. Lancet 1988;i:658.19 Cole L, Bellomo R, Baldwin I, Hayhoe M, Ronco C: The impact of lactate buffered high-volume

hemofiltration on acid-base balance. Intens Care Med 2003;7:1113–1120.20 Thomas AN, Guy JM, Kishen R, Bowles BMJ, Vadgama P: Comparison of lactate and bicarbon-

ate buffered hemofiltration fluids: Use in critically ill patients. Nephrol Dial Transplant 1997;12:1212–1217.

21 Heering P, Ivens K, Thümer O, Moergera S, Heintzen M, Passlick-Deetjen J, Willers R, Strauber BE,Grabenensee B: The use of different buffers during continuous hemofiltration in critically ill patientswith acute renal failure. Intens Care Med 1999;25:1244–1251.

22 Zimmerman D, Cotman P, Ting R, Karanicolas S, Tobe SW: Continuous veno-venous haemodial-ysis with a novel bicarbonate dialysis solution: Prospective cross-over comparison with a lactatebuffered solution. Nephrol Dial Transplant 1999;14:2387–2391.

23 MacLean AG, Davenport A, Cox D, Sweny P: Effects of continuous haemodiafiltration againstlactate buffered and lactate-free dialysate in patients with, and without liver dysfunction. Kid Int2000;58:1765–1772.

Davenport 328

24 Kierdorf HP, Leue C, Arns S: Lactate or bicarbonate buffered solutions in continuous extracorpo-real renal replacement therapies. Kid Int 1999;56(suppl 72):S32–S36.

25 Goldsmith DJA, Forni LG, Hilton PJ: Bicarbonate therapy and intracellular acidosis. Clin Sci1997;93:593–598.

26 Mehta RL, McDonald BR, Aguilar MM, Ward DM: Regional citrate anticoagulation for continu-ous arterio-venous hemodialysis in critically ill patients. Kid Int 1990;38:976–981.

27 Kutsogiannis DJ, Mayers I, Chi WD, Gibney RT: Regional citrate anticoagulation in continuousvenovenous hemodiafiltration. Am J Kid Dis 2000;35:802–811.

28 Palson R, Niles JL: Regional citrate anticoagulation in continuous veno-venous hemofiltration incritically ill patients with a high risk of bleeding. Kid Int 1999;55:1991–1997.

29 Bunchman T, Maxvold NJ, Barnett J, Hutchings A, Benfield MR: Pediatric hemofiltration:Normocarb dialysate solution with citrate anticoagulation. Pediatr Nephrol 2002;17:150–154.

30 Davenport A: Dialysate and substitution fluids for patients treated by continuous forms of renalreplacement therapy (CCRT). Contrib Nephrol. Basel, Karger, 2001, vol 132, pp 313–322.

31 Nowak MA, Campbell TE: Profound hypercalcaemia in continuous veno-venous hemofiltrationdialysis with trisodium citrate anticoagulation and hepatic failure. Clin Chem 1997;43:412–413.

32 Yagi N, LeBlanc M, Sakai K, Wright EJ, Paganini EP: Cooling effect of continuous renal replace-ment therapies in critically ill patients. Am J Kid Dis 1999;32:1023–1030.

Dr. A. DavenportCentre for Nephrology, Royal Free HospitalPond Street, London NW3 2QG (UK)Tel. �44 20 783 022 91, Fax �44 20 783 021 25, E-Mail [email protected]


A Practical Tool for Determiningthe Adequacy of Renal ReplacementTherapy in Acute Renal Failure Patients

Trairak Pisitkuna, Khajohn Tiranathanagula, Sonya Poulinb, Monika Bonellob, Gabriella Salvatorib, Vincenzo D’Intinib, Zaccoria Riccib, Rinaldo Bellomoc, Claudio Roncob

aDivision of Nephrology, Department of Medicine, Faculty of Medicine,Chulalongkorn University, Bangkok, Thailand; bDivision of Nephrology, St. Bortolo Hospital, Vicenza, Italy; cDepartment of Intensive Care, Austin and Repatriation Hospital, Melbourne, Australia

Acute renal failure (ARF) is one of the components of multiple organ fail-ure, a devastating syndrome in intensive care unit (ICU) patients. Although sev-eral aspects of medical care in ARF have been improved, including renalreplacement therapy (RRT), morbidity and mortality of the patients is still high[1–3]. Nevertheless, a great deal of development in RRT is ongoing concerningboth quality and quantity of therapy in the attempt to ameliorate the outcome inARF patients. Improvement in quality involves new techniques, membranes,and integrated equipment. The quantity aspect is also important, but still manyquestions remain unanswered. This article will focus on the main controversialissues on the quantity of RRT:

– How would you define an adequate renal replacement therapy?– Does treatment dose correlate with outcome?– How can you measure treatment dose in different treatments?

How Would You Define an Adequate Renal Replacement Therapy?

There is still no consensus about the definition of adequacy for RRT in acuterenal failure patients [4]. Indeed, the true meaning of adequacy itselfcould already express the authentic definition. In Latin, adequacy, ‘ad aequatum’

CRRT Information Technology

Pisitkun/Tiranathanagul/Poulin/Bonello/Salvatori/D’Intini/Ricci/Bellomo/Ronco 330

means ‘equal to…’. Thus, adequacy should be the state of treatment that closelymimics the features of the native kidney [5]. Therefore, adequacy should coverthree aspects of kidney function, including blood purification, homeostasis regu-lation, and biosynthesis. As far as blood purification is concerned, RRT shouldeliminate the wide range, small-to-larger sizes, of metabolic end products andother molecules normally handled by the human kidney. To regulate homeostasis,RRT should restore and maintain the ‘milieu enterieur’ regarding acid-base, elec-trolytes, and water balance in a continuous fashion as performed by the humankidney. For the biosynthesis aspect, RRT should provide an adequate stimulus toerythropoiesis and participate in the control of mineral balance and bone physi-ology. Unfortunately, no such RRT currently exists. Thus, the more realistic def-inition of adequacy for RRT should be the treatment that could sufficientlydominate the uremic syndrome and improve the survival rate. ‘How much treat-ment is sufficient?’ is a question that all of us should define. Beyond this ques-tion, a second question will rise: ‘How much treatment is optimal?’ The optimaldose should in fact be the one that provides best possible outcomes given a cer-tain degree of risk. In this condition, in fact, no further benefits can be observedfrom an increased dose and outcome becomes dependent on other variables.

Does Treatment Dose Correlate with Outcome?

While the treatment dose of RRT has been shown to affect outcome inchronic hemodialysis patients, it is still a matter of debate in acute renal failurepatients. Gillum et al. [6] found that there is no advantage to intensive dialysisin the management of acute renal failure. However, this study should be inter-preted with caution because there was no effective control of prescription anddelivery; no random allocation; no control of urea generation rate, proteincatabolic rate, and hydration status; and the results were only based on bloodlevels of urea nitrogen and creatinine. Bouman et al. [7] showed that survival at28 days was not improved by using high ultrafiltration volume when comparedwith low ultrafiltration volume in critically ill patients with oliguric acute renalfailure. Nevertheless, all patients enrolled in this study were postsurgical andsurvival rate was relatively high compared with previous reports.

On the other hand, recent evidence supported the concept of dose-outcomelinkage in acute renal failure patients. Storck et al. [8] reported that in acute renalfailure after surgery, the survival rate was significantly higher with pump-drivenhemofiltration (PDHF) than with spontaneous continuous arteriovenoushemofiltration (CAVH). The main difference between these two groups wasdaily ultrafiltration volume that was significantly higher with PDHF than withCAVH, 15.7 vs. 7.0 liters/day, respectively. Paganini et al. [9] showed the

The Adequacy Calculator 331

survival benefit of a higher delivered dose in patients with intermediate severity,making their assessment with the Cleveland Clinical Foundation ARF score.However, the significance was not seen in the two tales of the severity curveincluding mild and severe patients. The explanation was that the outcome in suchpatients might be predominated by the underlying comorbidity of patients ratherthan the effects of treatment. Schiffl et al. [10] reported the superiority of sur-vival in daily intermittent hemodialysis as compared with alternate-day intermit-tent hemodialysis among patients with acute renal failure. The authors stated thatthis partially resulted from the higher dose in the daily treatment group. Roncoet al. [11] demonstrated that the higher ultrafiltration rate, 35 vs. 20 ml/h/kg, ofcontinuous venovenous hemofiltration (CVVH) improved survival significantly.A further increase in ultrafiltration rate did not result in additional benefit interms of survival although in septic patients an improvement was observed. Theauthors recommended that ultrafiltration should be prescribed according topatient’s bodyweight and should reach at least 35 ml/h/kg. At this point, it seemsthat a dose and outcome relationship of RRT in acute renal failure patients exists,especially in specific groups of patients such as those with moderate levels ofseverity and patients with sepsis. Furthermore, it seems that such a correlation isonly linear within specific ranges of severity and doses. Beyond a given dose, thecorrelation becomes less significant and survival tends to steady. However, it isstill not clear how much an adequate or optimal dose is and therefore we aregoing to scrutinize where the curve reaches the plateau and which other factorsaffect the level of the plateau. To define the dose effect on outcome, we needlarge-scale multicenter studies using a quantitative measuring tool that is effec-tive, practical, and equivalent among the study groups. Furthermore, we need toselect some marker molecules to be used to measure dose. Finally, we need toestablish the effect of different modalities (convection, diffusion, and adsorption)at the same dose for a particular marker molecule such as urea.

How Can You Measure Treatment Dose in Different Treatments?

Renal replacement therapy consists of various modalities which differ inmany features including continuity of treatment (intermittent vs. continuous),vascular access (arteriovenous vs. venovenous), and mechanism of solute removal(diffusion vs. convection) [12, 13]. Accordingly, it is difficult to find an idealmarker employed for comparing the doses of these different treatments. As such,it needs the comprehensive view to search out the best solution for this matter.Using urea as a marker molecule, as is done in chronic hemodialysis, the treat-ment dose of RRT can be defined by various aspects such as efficiency, intensity,frequency, and clinical efficacy. In spite of its moderate toxicity, urea is often used


as a marker. Urea in fact has the advantage of being easily measurable, and rep-resenting the end of protein metabolism, its accumulation defines the requirementfor dialysis while its elimination defines the ability of treatment to remove toxins.Of course, urea just represents a surrogate of the low-molecular-weight toxin.

Efficiency

The efficiency of RRT can be represented by clearance (K). Technically, Kdepends on blood flow rate (Qb), dialysate flow rate (Qd), ultrafiltration rate(Qf), reference molecules, and hemodialyzer type and size. K can be normallyused to compare the treatment dose within each modality. Between differentmodalities, however, K is typically higher in intermittent hemodialysis (IHD)than in continuous renal replacement therapy (CRRT) and sustained low effi-ciency dialysis (SLED) (fig. 1a) even though IHD does not really remove thesolutes better than the others. This is not surprising since K represents only theamount of treatment per unit of time. Therefore, K cannot be employed tocompare various modalities which differ in treatment duration. Furthermore,

Dai

ly c

lear

ance

(ml/d

ay)

CVVHD24h

43,200

SLED8h

38,400

IHD3h

36,000

Wee

kly

clea

ranc

e (m

l/wee

k)

CVVHD

24hcontinuous

302,400

SLED

8halt. days

134,400

SLED

8h�7 days

268,800

IHD

3h�3 days

108,000

IHD

252,000

3h �7 days

Cle

aran

ce (m

l/min

)

CVVHD

30

SLED

80

IHD

200

a

c

b

Fig. 1. Comparisons of doses between different RRT modalities: (a) comparison byclearance (K); (b) comparison by daily clearance (Kt), and (c) comparison by weekly clear-ance (Kt � treatment days/week). CVVHD � Continuous venovenous hemodialysis;IHD � intermittent hemodialysis; SLED � sustained low efficiency dialysis.


K represents an instantaneous measurement and it correlates with the amountof solute removal at the time point of the measurement. Although K mightremain stable over time, if blood levels of the reference molecule change, theremoval rate will also change. In its original conception, in fact, K was signedto confer renal function among disparate individuals where, however, functionwas operating 24 h per day and blood levels were at a steady state.

Intensity

The intensity of RRT can be described by the product clearance � time (Kt).Because the time is accounted, Kt is more effective than K in the comparison ofvarious RRT modalities. As depicted in figure 1b, Kt in daily clearance is higherin CVVHD than in SLED and IHD even though CVVHD has the lowest K com-pared with the other treatments (fig. 1b).

Frequency

Frequency is an essential factor to further describe treatment dose in differ-ent modalities. Thus, weekly clearance, intensity � frequency (Kt � treatmentdays/week), is superior to Kt since it offers the comparison of different modali-ties in the more extensive view (fig. 1c).

Efficacy

Efficacy of RRT represents the effective clinical outcome resulting fromthe implication of a given treatment. It can be described by a fractional clear-ance (Kt/V), where V is the volume of distribution of the marker molecule.Kt/V is an established marker of adequacy correlating with survival in chronichemodialysis patients [14]. Kt/V, however, has not been proved as a marker ofadequacy in acute renal failure patients [15].

Of interest, all of the above methods have one common problem impairingtheir ability to compare between different RRT modalities. This problem resultsfrom the limitation of using K as a representation of solute removal capacity.Because solute removal is a product of K and serum levels of solute (C) (soluteremoval � K � C). The same K in different modalities does not mean the samesolute removal. In CRRT, for example, serum levels of solutes are nearly con-stant during the treatment while these decline during a dialysis session in IHD[13]. As a consequence, solute removal in CRRT is stable while it decreases


along the dialysis time in IHD. Furthermore, the rapid removal of solutes inIHD results in unfinished equilibration of serum levels of solutes between thebody pools as evidenced by the rebound of solutes after dialysis (fig. 2) [16, 17].Thus, CRRT would remove more solutes from the total body pool and accom-plish much greater blood purification than IHD even though K and Kt/Vof CRRT are lower than IHD (both daily short HD and SLED) (table 1) [18].After a correction of the equilibration problem by measuring the equilibratedKt/V (Eq Kt/V) [19], this contention is still the same since Eq Kt/V of CRRTremains lower than IHD (table 1). There is an attempt to make the different RRT

Table 1. Comparison of treatment dose between different modalities

Daily short HD SLED CVVHD

K, ml/min 200 80 30Predialysis BUN, mg/dl 110 110 70Postdialysis BUN, mg/dl 30 30 65Treatment time, min 180 480 1440Kt/V 1.34 1.32 0.9Eq Kt/V 1.12 1.24 0.9Daily clearance, liters 36 38.4 43.2Daily urea nitrogen removal, g 18 27 33.6Rebound 22% 6% 0%

Eq Kt/V � Equilibrated Kt/V.

120

100

80

60

40

20

0

BU

N (m

g/d

l)

Minutes of treatment0 60 120 180 240 300 360 420 480 540

D short HDKt/V�1.34Rebound�22%Eq Kt/V�1.12

CVVHDKt/V�0.9No reboundEq Kt/V�0.9

SLEDKt/V�1.32Rebound�6%Eq Kt/V�1.24

Fig. 2. Postdialytic rebound in different modalities. CVVHD provided a gradualremoval of BUN without rebound as compared to daily short hemodialysis which speedyremoved BUN with significant rebound.


modalities comparable by transforming the intermittent clearance to continuousclearance. Equivalent renal clearance (EKR) [20] and standard Kt/V (stdKt/V)[21] are established as the tools for this purpose in chronic hemodialysispatients. However, these methods cannot apply to calculations in the setting ofacute renal failure because the key assumption, urea generation rate is equal tourea removal rate, is not true in this hypercatabolic state.

The solute removal index (SRI) is another representative of efficacy and ithas been proposed as the best method for comparing the dose between differ-ent RRT modalities [15, 22]. SRI reflexes fraction urea removal during RRT byusing the dialysate-side measuring method. Therefore, SRI is not influenced byblood-side kinetics, including intercompartmental distribution of the solutes, ofvarious RRT [22]. When the frequency of RRT is included in the consideration,weekly SRI (SRI � treatment days/week) might become the gold standard forcomparing between disparate RRT modalities in acute renal failure patients.

Essentially, it needs the scientific proof for the merit of these various metho-dologies formerly described. In practice, it is difficult to manually calculate thedose of RRT by all of these various methods. To simplify the complicated calcu-lations, computer software is introduced to circumvent this obstacle. Computersoftware can assist in further studies to clarify the best dose measuring methodand it can also be utilized in current RRT practice. Accordingly, we have devel-oped a program on Microsoft Excel entitled ‘Adequacy Calculator for ARF’which is easily used and can calculate most of the described methods in a realtime basis.

In the program, it consists of calculators for the various RRT modalitiesincluding continuous mode (SCUF, CVVH, CVVHD, and CVVHDF) and inter-mittent mode (IHD and SLED) as shown in the main menu (fig. 3).

Continuous Mode

SCUFSCUF is a technique where blood is driven through a highly permeable fil-

ter via an extracorporeal circuit in the veno-venous mode. The ultrafiltrate pro-duced during membrane transit is not replaced and it corresponds exactly to thepatient’s weight loss. It is used only for fluid control in overhydration [12].Therefore, it is not necessary to measure the treatment dose in terms of soluteremoval and the program calculates only fluid balance per day (fig. 4). Therequired input data are SCUF prescription, including treatment time per dayand ultrafiltration rate, and non-CRRT fluid input-output per day. The outputwill be shown in both descriptive (‘Removal’, ‘Zero’, or ‘Repletion’) andnumerical data (ml/day).


Fig. 3. Main menu of Adequacy Calculator for ARF program.

Fig. 4. Slow continuous ultrafiltration (SCUF) calculator sheet.


CVVHCVVH is a technique whereby blood is driven through a highly permeable

filter via an extracorporeal circuit in the venovenous mode. The ultrafiltrateproduced during membrane transit is replaced in part or completely to achieveblood purification and volume control. Ultrafiltration is in excess of patientweight loss and replacement is needed. Because it is purely a convective ther-apy, the clearance for most solutes equals ultrafiltration rate [12].

At the start of the CVVH program, the methods of fluid replacement com-prising predilution, postdilution, or pre- � postdilution are requested (fig. 5).The input data consist of body weight, height, hematocrit, blood flow rate, treat-ment time per day, fluid replacement rate, and ultrafiltration rate.

The estimated urea clearance will be derived directly from ultrafiltrationrate in the postdilution setting as seen by the following equations:

Equation 1:

,

where, Cuf � ultrafiltrate urea nitrogen level; Quf � ultrafiltration rate, andCbi � prefilter BUN level.

Since the ultrafiltrate urea nitrogen level will be equivalent to the prefilterBUN level in a usual CVVH setting then:

Equation 2:

Clearance � Quf.

In the predilution and pre- � postdilution setting, however, the BUN levelentering the filter will lower after fluid replacement, thus the ultrafiltration ratewill overestimate the exact clearance. Thus, the program will correct this errorby using the following equation:

Equation 3:

where Qr � predilution fluid replacement rate, and Qb � blood flow rate.In addition, if one needs to declare the actual in vivo urea clearance, the

instantaneous urea clearance can be measured by two options. Firstly, blood-side urea clearance is determined by collecting the simultaneous blood samplesof pre- and postfilter urea levels and calculating it with the following equation:

Equation 4:

ClearanceCbi Qbi Cbo Qbo

Cbi�

� � �,

ClearanceQuf

1 (Qr Qb),�

�

ClearanceCuf Quf

Cbi�

�


Fig. 5. Continuous venovenous hemofiltration (CVVH) calculator sheet.


where Cbi � prefilter BUN level, Cbo � postfilter BUN level, Qbi � pre-filter blood flow rate, and Qbo � postfilter blood flow rate.

Secondly, ultrafiltrate-side urea clearance is determined by collecting thesimultaneous blood and ultrafiltration samples of prefilter and ultrafiltrationurea levels, respectively, and calculating as for equation 1.

After the value of urea clearance (K) is obtained, the other values includingclearance per body surface area (K/BSA), daily urea clearance (Kt), and dailyKt/V are calculated simply. The BSA is derived from the Du Bois formula [23].The volume distribution of urea (V) is attained by estimating at 60% of the bodyweight since there is still no established method for V determination in acuterenal failure patients.

For the estimation of daily urea nitrogen removal, steady state prefilterBUN level (Cbi) is required for calculation by the following equation:

Equation 5:

Daily urea nitrogen removal � K � t � steady state Cbi.

However, if Cbi is not constant, daily average Cbi should be used insteadof steady state Cbi.

After daily urea nitrogen removal is obtained, the daily solute removalindex (SRI) of urea is simply estimated by dividing the daily urea nitrogenremoval with total body urea nitrogen level at the start of the same treatmentday as seen in the following equation [22]:

Equation 6:

Eventually, when the superior comparison between disparate RRT modal-ities is concerned, weekly urea clearance, weekly urea nitrogen removal, andweekly SRI are easily estimated by multiplying the daily urea clearance, thedaily urea nitrogen removal, and the daily SRI, respectively, with the number oftreatment day in a week.

Other than measuring the dose of RRT, the program can also calculate ureageneration rate (G), normalized protein equivalent of total nitrogen appearance(nPNA), and residual renal urea clearance (Kr). The urea generation rate wascalculated by the following equation:

Equation 7:

where UNA � urea nitrogen appearance; body UNA � (Cbit � Vt �Cbi0 � V0); ultrafiltrate UNA � (Quf � (Cuf0 � Cuft/2) � t); urine

Gbody UNA + ultrafiltrate UNA + urine UNA

t�

Daily SRIdaily urea nitrogen removal

start of the day Cbi�

�VV.


UNA � urine volume � average urine urea nitrogen; Cbi0 � prefilter BUN levelat time 0; Cbit � prefilter BUN level at time t; Cuf0 � ultrafiltrate urea nitrogenlevel at time 0; Cuft � ultrafiltrate urea nitrogen level at time t; V0 � volume dis-tribution of urea at time 0, V0 � BW0 � 0.6; Vt � volume distribution of urea attime t, Vt � BWt � 0.6; t � time between two points of sample collections.

These calculations required information on the change of BUN and theamount of urea nitrogen removal in ultrafiltrate and urine between two timepoints, preferably 24 h apart, which can be obtained by 3 methods. The bestmethod requires two time points of urea nitrogen levels in both blood and ultra-filtrate samples (option 1 in the part of urea generation rate calculation of theprogram). In option 2, it is only necessary to know the urea nitrogen levels inblood because the urea nitrogen levels in ultrafiltrate can be estimated fromthose in blood based on the assumption that, in the usual setting of CVVH, asmall molecule solute such as urea will be equivalent between blood and ultra-filtrate sides. On the other hand, option 3 requires only the urea nitrogen levelsin ultrafiltrate because the urea nitrogen levels in blood can be estimated basedon the same assumption. In the predilution setting, these estimations cannot beperformed directly since they require the involvement of the predilution fluidreplacement rate in the calculation. Nevertheless, the program will take thispoint into account and accomplish this complicated calculation. After the ureageneration rate is obtained, nPNA is approximated by the following equation:

Equation 8:

.

Residual renal urea clearance is simply calculated by dividing the urineUNA with time between two points of sample collections.

Moreover, the program will determine the fluid balance per day asdescribed above in the SCUF mode.

Continuous Venovenous HemodialysisCVVHD is a technique whereby blood is driven through a low permeable

dialyzer and a countercurrent flow of dialysis solution is delivered to thedialysate compartment. The ultrafiltrate produced during membrane transit cor-responds to patient weight loss. Solute clearance is mainly achieved by diffu-sion. A replacement solution is not needed. Because it is mainly a diffusivetherapy, its efficiency is limited to small molecules [12].

At the start of the CVVHD program, the input data consist of body weight,height, hematocrit, blood flow rate, treatment time per day, dialysate inflowrate, and dialysate outflow rate are requested (fig. 6).

nPNAG 6.25

average BW�

�


Fig. 6. Continuous venovenous hemodialysis (CVVHD) calculator sheet.


The estimated urea clearance will be derived directly from dialysate out-flow rate as the following equation:

Equation 9:

,

where Cdo � dialysate outflow urea nitrogen level, Qdo � dialysate outflow rate,and Cbi � prefilter BUN level.

Since the dialysate outflow urea nitrogen level will be equivalent to pre-filter BUN level in a usual CVVHD setting, then:

Equation 10:

Clearance � Qdo.

This assumption is not true when blood flow rate is too low and dialysateflow rate is too high because in this condition the dialysate outflow will not besaturated with urea nitrogen from the blood side [24]. However, this will not bethe problem for calculation if the instantaneous urea clearance is measured.

As in the CVVH mode, the instantaneous urea clearance can be deter-mined by two options. Firstly, blood-side urea clearance is determined by col-lecting the simultaneous blood samples of prefilter and postfilter urea levelsand calculating as in equation 4. Secondly, dialysate-side urea clearance isdetermined by collecting the simultaneous blood and dialysate samples ofprefilter and dialysate outflow urea levels, respectively, and calculating as inequation 9.

After the value of urea clearance (K) is obtained, the other values as in theCVVH mode including clearance per body surface area (K/BSA), daily ureaclearance (Kt), daily Kt/V, daily urea nitrogen removal, daily solute removalindex (SRI), weekly urea clearance, weekly urea nitrogen removal, and weeklySRI are determined by the program.

For the urea generation rate (G) determination, it can be done by using thedialysate outflow urea nitrogen level instead of ultrafiltrate urea nitrogen levelfor the calculation in equation 7. The normalized protein equivalent of totalnitrogen appearance (nPNA), residual renal urea clearance (Kr), and the fluidbalance per day are also determined by the program.

CVVHDFContinuous venovenous hemodiafiltration is a technique whereby blood is

driven through a highly permeable dialyzer and a countercurrent flow of dialysissolution is delivered on the dialysate compartment. The ultrafiltrate producedduring membrane transit is in excess to the patient weight loss. Solute clearanceis obtained both by diffusion and convection. Replacement solution is needed to

ClearanceCdo

Cbi�

� Qdo


obtain fluid balance. Since it is a convective plus diffusive therapy, its efficiencyis extended from small to larger molecules [12].

At the start of the CVVHDF program, the methods of fluid replacementcomprising pre- and postdilution are requested (fig. 7). The input data consistof body weight, height, hematocrit, blood flow rate, treatment time per day,fluid replacement rate, dialysate inflow rate, and dialysate outflow rate.

Based on the assumption that the dialysate outflow urea nitrogen level willbe equivalent to the prefilter BUN level in a usual CVVHDF setting, the esti-mated urea clearance will be derived directly from dialysate outflow rate in thepostdilution setting as in equations 9 and 10. In the predilution setting, however,dialysate outflow rate will overestimate the exact clearance. Thus, the programwill correct this error by using the following equation:

Equation 11:

This estimation cannot be used when the blood flow rate is too low and thedialysate flow rate is too high because in this condition the dialysate outflowwill not be saturated with urea nitrogen from the blood side [25]. However, theinstantaneous urea clearance, determined as in the CVVHD mode, can beemployed to circumvent this limitation.

After the value of urea clearance (K) is obtained, the other values as in theCVVH and CVVHD mode including clearance per body surface area (K/BSA),daily urea clearance (Kt), daily Kt/V, daily urea nitrogen removal, daily soluteremoval index (SRI), weekly urea clearance, weekly urea nitrogen removal, andweekly SRI are determined by the program.

For the urea generation rate (G) determination, it can be performed as inthe CVVHD mode; in addition, the program will also take account of the meth-ods of fluid replacement in the calculation. The normalized protein equivalentof total nitrogen appearance (nPNA), residual renal urea clearance (Kr), and thefluid balance per day are also determined by the program.

Intermittent Mode

This mode consists of standard intermittent hemodialysis (IHD) and sus-tained low efficiency dialysis (SLED). These therapies are mainly the diffu-sive therapy performed in intermittent fashion. The differences between thesetwo modalities are in the aspects of efficiency and treatment time. SLED isdesigned to operate at a lower efficiency (clearance) but longer treatment time

ClearanceQdo

1 (Qr/Qb)�

�.


Fig. 7. Continuous venovenous hemodiafiltration (CVVHDF) calculator sheet.


than IHD [26]. In the SLED mode, blood flow rate and dialysate flow rate areusually set at the range of 150–200 and 100 ml/min, respectively, while thetreatment time is extended to at least 8 h. In both IHD and SLED, however, thedoses of the treatments can be measured by using the same principle ofcalculation.

The program can determine dialyzer urea clearance of these therapies bytwo methods. Firstly, the estimated urea clearance can be obtained by multiplestep calculation, as described by Daugirdas [27], which requires input dataincluding in vitro urea clearance (K) at the specific in vitro blood flow rate (Qb)and in vitro dialysate flow rate (Qd), actual blood and dialysate flow rate, ultra-filtration rate, and hematocrit. In vitro data of K, Qb, and Qd can be obtainedautomatically from the program database by selecting the brand and the modelof dialyzer (as shown in the list boxes in fig. 8) or these data can be acquiredfrom the dialyzer manufacturer and entered manually into the program. Theultrafiltration rate is calculated from total body weight loss during dialysisdivided by treatment time. Secondly, the actual in vivo urea clearance can bemeasured by adding the pre- and postfilter levels of simultaneous BUN samplesinto the program which will perform the calculation by the equation 4.

After the value of dialyzer urea clearance (K) is obtained, the clearance perbody surface area (K/BSA), daily urea clearance (Kt), and Kt/V derived fromthe formal urea kinetic modeling (UKM) method are calculated. In the UKMmethod, the volume distribution of urea (V) is attained by a complicated math-ematical iteration of two formulae described by Gotch as the following [28]:

Equation 12:

Equation 13:

where V � the volume distribution of urea, Quf � ultrafiltration rate, t � treat-ment time, G � the urea generation rate, K � the dialyzer urea clearance,Kr � the residual renal urea clearance, C0 � the BUN level at the beginning of

G

(Kr ) C0 CtV

V

1V

V

Kr

Kr�

��

��

��

�

��

⎡⎣⎢

⎤⎦⎥

⎡

⎣

⎢⎢⎢

⎤

⎦

⎥⎥⎥

⎡⎣⎢

⎤⎦⎥

��⎡

⎣

⎢⎢⎢

⎤

⎦

⎥⎥⎥

V Quf t 1G Ct(K Kr Quf)

G C0(K Kr Quf)

Quf

K Kr Quf� � �

� � �

� � �

� �⎡

⎣⎢

⎤

⎦⎥

⎡

⎣

⎢⎢⎢⎢

⎤

⎦

⎥⎥⎥

⎡

⎣

⎢⎢⎢⎢

⎤

⎦

⎥⎥⎥⎥

�

�

1

1


a dialysis treatment, Ct � the BUN level at the end of a dialysis treatment,� � the rate of interdialytic volume expansion, � � the length of the interdia-lytic interval.

These formulas share common terms and solve for V and G. The programwill repeat these formulas until unique values for V and G are found that sat-isfy both equations. The requisite input for these calculations include bodyweights at before dialysis (BW pre1), after dialysis (BW post), and before thenext dialysis (BW pre2); treatment time; BUN level immediately before dialy-sis (BUN pre1), immediately after dialysis (BUN post, using slow flow/stop

Fig. 8. Intermittent hemodialysis (IHD) calculator sheet. Remark: Sustained low effi-ciency dialysis (SLED) calculator sheet is similar to IHD calculator sheet (figure not shown).


pump technique), 30 min after dialysis (BUN post 30), and immediately beforethe next dialysis (BUN pre2); duration between the beginning of a dialysis(pre1) and the next dialysis (pre2); and, for calculating the Kr, urine volume andaverage urine urea nitrogen level between the end of a dialysis (post) and thenext dialysis (pre2). For single-pool Kt/V (spKt/V) and equilibrate Kt/V(eqKt/V) determinations, the values of V are obtained by using ‘BUN post’ and‘BUN post 30’, respectively, as the Ct in the above UKM formulas.

In addition, the program also determines the spKt/V and eqKt/V by using anatural logarithm formula [29] and rate equation [19], respectively. The type ofvascular access between arteriovenous (AV) and venovenous (VV) is required toperform the rate equation.

For the estimation of daily urea nitrogen removal per and daily soluteremoval index (SRI), the dialysate samples for urea nitrogen levels (Cd) arerequested. Even though the dialysate samplings are tedious tasks, thesedialysate-side dose-measuring methods might provide the best solution forcomparing between disparate RRT modalities in acute renal failure patients. Formore practicality, spot sampling rather than total dialysate collection is pre-ferred. The program has designed a time table for dialysate sampling by usingthe mid point of time in the particular hours as references. After the urea nitro-gen levels from all dialysate samples are obtained, the daily urea nitrogenremoval is calculated using the following equation:

Equation 14:

Daily urea nitrogen removal � (Qd � Quf) � (Cd1 � t1 � Cd2 � t2 � Cd3 � t3 � …)

where Qd � dialysate flow rate, Quf � ultrafiltration rate, Cd1,2,3,… �dialysate urea nitrogen level at mid point of hours 1, 2, 3,…, and t1,2,3,… �duration of treatment (minute) at hours 1, 2, 3,…

As a consequence, SRI is simply estimated by dividing the daily urea nitro-gen removal with total body urea nitrogen level at predialysis as in the followingequation [22]:

Equation 15:

As in continuous mode, the weekly urea clearance, weekly urea nitrogenremoval, and weekly SRI are easily estimated by multiplying the daily ureaclearance, the daily urea nitrogen removal and the daily SRI, respectively, withthe number of treatment days in a week.

Because G is already obtained from the UKM calculation described above,the normalized protein equivalent of total nitrogen appearance (nPNA) is simplydetermined from G by the Borah formula [30].

SRIdaily urea nitrogen removal

predialysis BUN V�

�.


Conclusion

Even though the adequacy of RRT does not clearly influence the survivalrate of the critical acute renal failure patient at this moment, this issue is stilldrawing much concern from the intensive patient care team. To decipher theunresolved issues about the survival benefit and the sufficient quantity of thedose of RRT in acute renal failure, long-term, large-scaled multicenter studiesusing the same tool for measuring and comparing the dose of various treatmentmodalities are needed. Of interest, the Adequacy Calculator for ARF programis innovated to provide the practicality in this matter since this program isdesigned on the worldwide used software Microsoft Excel and it can performthe extensive calculations covering all dimensions of adequacy including effi-ciency, intensity, frequency and efficacy of RRT. Other than a tool for the study,this program can also be readily employed in clinical practice. Thus, this pro-gram is aspired to be a valuable tool for exploring the unforeseeable arena ofadequacy of RRT in acute renal failure patients.

References

1 Parker RA, Himmelfarb J, Tolkoff-Rubin N, Chandran P, Wingard RL, Hakim RM: Prognosis ofpatients with acute renal failure requiring dialysis: Results of a multicenter study. Am J KidneyDis 1998;32:432–443.

2 Metnitz PG, Krenn CG, Steltzer H, Lang T, Ploder J, Lenz K, Le Gall JR, Druml W: Effect ofacute renal failure requiring renal replacement therapy on outcome in critically ill patients. CritCare Med 2002;30:2051–2088.

3 McCarthy JT: Prognosis of patients with acute renal failure in the intensive-care unit: A tale oftwo eras. Mayo Clin Proc 1996;71:117–126.

4 Ronco C, Kellum JA, Mehta R: Acute dialysis quality initiative (ADQI). Nephrol Dial Transplant2001;16:1555–1558.

5 Bellomo R, Ronco C: Adequacy of dialysis in the acute renal failure of the critically ill: The casefor continuous therapies. Int J Artif Organs 1996;19:129–142.

6 Gillum DM, Dixon BS, Yanover MJ, Kelleher SP, Shapiro MD, Benedetti RG, Dillingham MA,Paller MS, Goldberg JP, Tomford RC, et al: The role of intensive dialysis in acute renal failure.Clin Nephrol 1986;25:249–255.


8 Storck M, Hartl WH, Zimmerer E, Inthorn D: Comparison of pump-driven and spontaneous con-tinuous haemofiltration in postoperative acute renal failure. Lancet 1991;337:452–455.

9 Paganini EP, Tapolyai M, Goormastic M, Halstenberg W, Kozlowski L, Leblanc M, Lee JC,Moreno L, Sakai K: Establishing a dialysis therapy/patient outcome link in intensive care unitacute dialysis for patients with acute renal failure. Am J Kidney Dis 1996;28(suppl 3):S81–S89.


11 Ronco C, Bellomo R, Homel P, Brendolan A, Dan M, Piccinni P, La Greca G: Effects of differentdoses in continuous veno-venous haemofiltration on outcomes of acute renal failure: A prospectiverandomised trial. Lancet 2000;356:26–30.


12 Ronco C, Bellomo R: Continuous renal replacement therapy: Evolution in technology and currentnomenclature. Kidney Int 1998;66(suppl):S160–S164.

13 Clark WR, Ronco C: Renal replacement therapy in acute renal failure: Solute removal mechanismsand dose quantification. Kidney Int 1998;66(suppl):S133–S137.

14 Gotch F, Sargent J: A mechanistic analysis of the National Cooperative Dialysis Study (NCDS).Kidney Int 1985;28:526–534.

15 Evanson JA, Ikizler TA, Wingard R, Knights S, Shyr Y, Schulman G, Himmelfarb J, Hakim RM:Measurement of the delivery of dialysis in acute renal failure. Kidney Int 1999;55:1501–1508.

16 Depner TA, Rizwan S, Cheer AY, Wagner JM, Eder LA: High venous urea concentrations in theopposite arm: A consequence of hemodialysis-induced compartment disequilibrium. ASAIOTrans 1991;37:M141–M143.

17 Schneditz D, Kaufman AM, Polaschegg HD, Levin NW, Daugirdas JT: Cardiopulmonary recircu-lation during hemodialysis. Kidney Int 1992;42:1450–1456.

18 Clark WR, Mueller BA, Alaka KJ, Macias WL: A comparison of metabolic control by continuousand intermittent therapies in acute renal failure. J Am Soc Nephrol 1994;4:1413–1420.

19 Daugirdas JT, Schneditz D: Overestimation of hemodialysis dose depends on dialysis efficiencyby regional blood flow but not by conventional two pool urea kinetic analysis. ASAIO J 1995;41:M719–M724.

20 Casino FG, Lopez T: The equivalent renal urea clearance: A new parameter to assess dialysis dose.Nephrol Dial Transplant 1996;11:1574–1581.

21 Gotch FA: Evolution of the single-pool urea kinetic model. Semin Dial 2001;14:252–256.22 Keshaviah P, Star RA: A new approach to dialysis quantification: An adequacy index based on

solute removal. Semin Dial 1994;7:85–90. 23 Du Bois D, Du Bois EF: A formula to estimate the approximate surface area if height and weight

be known. Nutrition 1989;5:303–311.24 Sigler MH, Teehan BP: Solute transport in continuous hemodialysis: A new treatment for acute

renal failure. Kidney Int 1987;32:562–571. 25 Brunet S, Leblanc M, Geadah D, Parent D, Courteau S, Cardinal J: Diffusive and convective solute

clearances during continuous renal replacement therapy at various dialysate and ultrafiltrationflow rates. Am J Kidney Dis 1999;34:486–492.

26 Marshall MR, Golper TA, Shaver MJ, Alam MG, Chatoth DK: Sustained low-efficiency dialysisfor critically ill patients requiring renal replacement therapy. Kidney Int 2001;60:777–785.

27 Daugirdas JT: Appendix A: Estimating dialyzer blood water clearance from KoA, QB, and QD

(Table A-1); in Daugirdas JT, Blake PG, Ing TS (eds): Handbook of Dialysis, ed 3. Philadelphia,Lippincott, Williams & Wilkins, 2001, p 674.

28 Gotch FA: Kinetic modeling in hemodialysis; in Nissenson AR, Fine RN, Gentile DE (eds):Clinical Dialysis, ed 3. Englewood Cliffs, Prentice Hall, 1995, pp 156–189.

29 Daugirdas JT: Second generation logarithmic estimates of single-pool variable volume Kt/V: Ananalysis of error. J Am Soc Nephrol 1993;4:1205–1213.

30 Borah MF, Schoenfeld PY, Gotch FA, Sargent JA, Wolfsen M, Humphreys MH: Nitrogen balanceduring intermittent dialysis therapy of uremia. Kidney Int 1978;14:491–500.

Claudio Ronco, MDDepartment of Nephrology, St. Bortolo HospitalViale Rodolfi, IT–36100, Vicenza (Italy)Tel. �39 0444 993 869, Fax �39 0444 993 949, E-Mail [email protected]


How to Approach Sepsis Today?

Jean-Louis Vincent

Department of Intensive Care, Erasme Hospital, Free University of Brussels,Brussels, Belgium

Over the last decade or so, vast amounts of time, energy, and money havegone into investigating the multiple facets of the pathophysiology and effects ofsepsis. These efforts have not been in vain, and while sepsis continues to be aleading cause of morbidity and mortality in intensive care units (ICUs) world-wide, the development and licensing of the first immunomodulatory therapy forpatients with severe sepsis, along with encouraging results from other clinicaltrials, has provided the intensivist with some strategies to curb the often tragicoutcomes of this disease process.

Diagnosis of Sepsis

Advances in understanding of the sepsis response have led to a need for anupdate of the widely used 1992 SCCM/ACCP definitions referring to the sys-temic inflammatory response syndrome (SIRS). At a Sepsis DefinitionsConference involving 29 physicians from Europe and North America held in2001, the participants agreed that sepsis should still be defined as infection plussigns of systemic inflammation, but that the SIRS criteria should be abandoned[6]. Unfortunately, as yet, no individual sign is specific for sepsis and clinicaldiagnosis relies on the combined presence of several signs and symptoms(the majority of which are listed in table 1). Severe sepsis is defined as sepsisplus organ dysfunction, and septic shock as severe sepsis with hypotensiondespite adequate fluid resuscitation, and evidence of perfusion abnormalities.This conference also introduced the idea of staging sepsis, as cancer is staged,to facilitate assessment of likely patient prognosis and potential to respond totherapy. The PIRO system (table 2) stratifies patients according to theirPredisposing conditions, the nature of the Infection, the extent of the hostResponse, and the degree of concomitant Organ dysfunction.

New Frontiers in the Management of ARF, MOF and Sepsis

How to Approach Sepsis Today? 351

Table 1. Signs of sepsis

• General signs and symptomsRigor-fever (sometimes hypothermia)Tachypnea/respiratory alkalosisPositive fluid balance – edema

• General hematologic/inflammatory reactionIncreased (sometimes decreased) WBC – increased immature formsIncreased CRP, IL-6 and procalcitonin concentrations

• Hemodynamic alterationsArterial hypotensionTachycardiaIncreased cardiac output/wide pulse pressure/low SVR/high SvO2

Altered skin perfusion (cold, mottled extremities, petechiae, …)Decreased urine outputHyperlactatemia – increased base deficit

• Signs of organ dysfunctionHypoxemia (ALI)Altered mental statusAlteration in renal functionThrombocytopenia, DICAlteration in liver tests (hyperbilirubinemia)Intolerance to feeding (altered GI motility)Hyperglycemia

Predisposing Conditions. This includes individual patient characteristicsthat make them more or less likely to develop sepsis and to succumb when theydo, e.g. age, gender, chronic disease processes such as cancer, history of alco-hol abuse, genetic polymorphisms, etc.

Infection. This includes characteristics of infection that may influence apatient’s response to that infection and/or response to treatment, e.g. the site ofinfection, the specific organism, the size of the inoculum, and the susceptibil-ity of the organism to antimicrobial agents.

Response. The degree of immune response to the infection can be charac-terized by the presence or absence, or the degree of elevation of various signsand symptoms, e.g. white cell count, C-reactive protein, procalcitonin, etc.(table 1). The response will vary between individual patients and in the samepatient at different times.

Organ Dysfunction. This can be measured and monitored using organdysfunction scores such as the sequential organ failure assessment (SOFA),which use simple, routinely available parameters to assess organ function invarious, usually six, organ systems.

Vincent 352

The PIRO system is in its infancy and requires further development andrefinement, but it is an important advance and, in addition to characterizingindividual patients, will help compare patient populations for clinical trial pur-poses, and help direct clinical research.

Basic Management of Sepsis

The essentials of the management of patients with severe sepsis remainunchanged and can be broadly divided into infection control and hemodynamicstabilization. For the purposes of this chapter, each aspect will be consideredseparately, although in practice many are started simultaneously, as ‘time is tis-sue’ and early, effective therapy saves lives.

Infection ControlInfection control involves an in depth search for the source of the sepsis,

with rapid institution of appropriate antibiotic therapy, after relevant cultureshave been taken, and surgical removal of any focus. Delayed or initially ineffec-tive antibiotic therapy has been shown to be associated with worse prognosis,and it is important that all likely microbial culprits are covered by the empiricantibiotic(s), which can then be altered as culture results become available.

Table 2. The PIRO system

P Predisposing factors AgeSexPast historyGenetic factorsAlcoholism, tobacco abuse, …

I Infection Source (lungs, abdomen, urine, cerebrospinal fluid, …)Degree of extension (e.g. for pneumonia, one lobe vs. one lung vs. bilateral)Type of microorganism (gram-positive, gram-negative, fungus, virus)

R Response Fever, tachycardia, tachypnea, …WBC, CRP, PCT, cytokines, …

O Organ dysfunction PaO2/FiO2

Platelet counte.g. SOFA score Bilirubin

CreatinineNeed for vasopressorsGCS


Hemodynamic StabilizationResuscitation of the patients with severe sepsis or septic shock can be

considered using the VIP mnemonic proposed by Weil and Shubin back in1969: V – ventilate; I – infuse; P – pump.

Ventilate

Adequate oxygenation is essential and, when needed, mechanical ventila-tion should be instituted early rather than late as, in addition to providing ade-quate oxygenation, mechanical ventilation can reduce the work of breathingand hence reduce oxygen demand. Specific indications for instituting mechan-ical ventilation include severe tachypnea, muscular respiratory failure (use ofaccessory muscles), altered mental status, and severe hypoxemia despite sup-plemental oxygen. However, if uncertain as to whether or not to intubate, thenintubate! While mechanical ventilation carries attendant risks, delaying venti-lation is equally harmful in that tissue oxygenation is further threatened.

The choice of ventilatory mode is large and studies continue to try anddetermine which method if any is superior to the next. In essence, mechanicalventilation should be aimed at providing hemoglobin oxygen saturations of atleast 90% (arterial oxygen pressure [PaO2] greater than 60 mm Hg). The use ofpositive end-expiratory pressure (PEEP) can increase mean airway pressure andreduce the concentrations of oxygen needed. Higher PEEP levels are not indi-cated, however, and the short-term use of higher inspired oxygen concentrationswhen necessary is preferred over the risks of barotrauma or hemodynamicimpairment seen with higher PEEPs.

Prone positioning can improve oxygenation in patients with acute respira-tory distress syndrome (ARDS), a common complication of sepsis. However, alarge randomized trial of this strategy failed to demonstrate any overallimprovement in survival in patients positioned prone, although mortality wasreduced in patients who responded to prone positioning by reducing theirPaCO2.

Considerable interest has focused on the optimal tidal volume since theNational Institutes of Health-sponsored ARDS Network study [13] showingimproved outcomes in patients ventilated with tidal volumes of 6 ml/kg versusthose ventilated with tidal volumes of 12 ml/kg predicted body weight. However,this study has been criticized in its choice of 12 ml/kg as the ‘traditional’ tidalvolume, and it is generally felt that rather than showing 6 ml/kg to be the idealtidal volume, the study rather demonstrated that tidal volumes of 12 ml/kg aretoo high. The optimum tidal volume is probably somewhere in between, round8–9 ml/kg [9].

Vincent 354

Other ventilatory techniques have also been explored for their potential ben-eficial effects in ARDS. These include: inhaled nitric oxide (NO), which has beenshown to improve oxygenation but not survival; aerosolized surfactant, which hadno effect on oxygenation or survival; and partial liquid ventilation, which mayimprove oxygenation and have anti-inflammatory effects. The optimum approachto ventilation of the patient with severe sepsis, and in particular, sepsis-relatedARDS is an active field of current research.

Infuse

Fluid administration is an essential aspect of resuscitation from septic shock,the aims being to preserve intravascular fluid volume, restore effective tissue per-fusion, and re-establish a balance between tissue oxygen demand and supply.However, while in most forms of circulatory shock, hypotension and reducedtissue perfusion are due to a fall in cardiac output, and endpoints of fluid resusci-tation are thus relatively easy to identify, in sepsis the situation is much morecomplex with cardiac output often normal or raised, and tissue perfusion beinghindered by microcirculatory alterations. The endpoints of fluid therapy in sepsisare thus much more difficult to define, largely because of problems in monitoringthe regional microcirculation and oxygenation, and changes may persist at a locallevel while systemic hemodynamic and oxygenation parameters appear to havestabilized. Several markers of perfusion and oxygenation have been proposed,including blood lactate levels, mixed venous oxygen saturation (SvO2), and thegastric tonometry-derived variables gastric intramucosal pH (pHi) or PCO2 gap,but none alone are reliable indicators of adequate resuscitation. The combinationof clinical parameters (mean arterial pressure, urine flow, skin perfusion, level ofconsciousness) with blood lactate levels is most useful.

Hyperlactatemia (blood lactate �2 mEq/l) is typically present in patientswith septic shock, and may be secondary to anaerobic metabolism due tohypoperfusion, although in septic patients, elevated lactate levels may alsoresult from cellular metabolic failure in sepsis and from decreased clearance bythe liver. Raised blood lactate levels nevertheless are associated with worse out-comes in septic shock patients, particularly if the high levels persist, and bloodlactate levels have been shown to have greater prognostic value than oxygen-derived variables.

The normal SvO2 is 70–75% in critically ill patients, but can be elevated inseptic patients due to maldistribution of blood flow. Admittedly, a normal orhigh SvO2 does not necessarily indicate adequate tissue oxygenation, but thereis no definitive evidence that further increasing oxygen supply to the tissues inthese circumstances will improve outcome. On the other hand, a low SvO2


should prompt rapid intervention to increase oxygen delivery to the tissues.This concept is supported by a recent study by Rivers et al. [10], who evaluatedthe effects of early goal-directed therapy in patients with severe sepsis and sep-tic shock versus standard therapy, using restoration of a central venous oxygensaturation (ScvO2) greater than 70% as one of its goals. This target wasachieved by 95% of the early goal-directed group, compared to just 60% of thestandard treatment group (p � 0.001), and the strategy was associated withimproved hospital mortality rates (30.5% in the goal-directed group versus46.5% in the standard therapy group, p � 0.009).

Gastric tonometry-derived parameters have been studied as possible toolsto detect regional hypoxia, but the use of traditional gastric tonometry is lim-ited due to potential methodological pitfalls, substantial measurement errors,and doubts regarding the validity of intramucosal pH (pHi) calculations.Automated air tonometry also has sources of error including buffering of gas-tric acid by bicarbonate secreted by the stomach. Several studies have shownthe prognostic value of pHi or PCO2 gap, but although some studies have sug-gested that the pHi may be a useful guide in surgical patients, treatment titratedagainst pHi was not shown to be beneficial in a heterogeneous population ofcritically ill patients. Newer techniques of monitoring the microcirculation arebeing developed including orthogonal polarization spectral (OPS) imaging, inwhich the microcirculation of mucosal surfaces, e.g. the sublingual area, isdirectly visualized. The microcirculatory changes seen in sepsis are related toprognosis and this technique may provide useful information regarding ade-quacy of resuscitation, although it remains experimental at present.

FluidsThe debate as to which fluid provides greatest benefit in sepsis continues

and we will not discuss it in great detail here. In fact, it is the quantity of fluidrather than the type of fluid per se that is of greatest importance. Certainly,more crystalloid is needed to achieve the same effect as colloid, thus potentiallyincreasing the risk of edema, but colloids are more expensive and carry theirown risks.

The most commonly used crystalloid solutions for resuscitation purposesare normal/isotonic (0.9%) saline and Ringer’s lactate (also called Hartmann’ssolution). Hypertonic saline rapidly improves hemodynamic status and with itspossible immunomodulatory and beneficial microcirculatory effects, may be avaluable initial resuscitation fluid in patients with septic shock [8]. However,there are few clinical data available at present to confirm or refute the benefitsof hypertonic saline in sepsis.

Colloid solutions include human albumin and artificial colloids such asdextrans, gelatins, and hydroxyethyl starch (HES) solutions. Albumin use has

Vincent 356

come under heavy fire in recent years since a meta-analysis suggested thathuman albumin administration was associated with an increased mortality rate,even though a subsequent meta-analysis failed to demonstrate any adverseeffect of albumin on mortality. Albumin may have a particular place in criticallyill patients with hypoalbuminemia. Of the artificial colloids, HES solutions arethe most frequently used. Experimental studies have suggested that HES solu-tions may have beneficial effects on the microcirculation in sepsis, but thesehave not been confirmed in clinical trials, and others have suggested that HESsolutions increase blood viscosity and may thus further compromise micro-circulatory flow. HES solutions can prolong prothrombin time and the totalamount infused thus needs to be limited, although its use has not been relatedto increased bleeding in septic shock patients. The molecular weight (MW) ofthese solutions may be important. In patients with severe sepsis, infusion of amedium MW HES solution (hexastarch) was associated with an increased riskof developing acute renal failure. Newer generation HES solutions with smallerMW may have beneficial effects, including on the microcirculation andendothelial cells, while limiting side effects [5].

Whatever fluid is decided on, assessment of adequacy of fluid infusion canbe facilitated by repeated fluid challenges, in which a predefined amount offluid (e.g. 250 or 500 ml) is infused over a set time. Preselected endpoints, e.g.mean arterial pressure 75 mm Hg, and pressure safety limits, e.g. central venouspressure (CVP) or pulmonary artery wedge pressure (PAWP), are then moni-tored, and fluids continued or stopped accordingly.

TransfusionWith several studies in recent years suggesting worse outcomes in critically

ill patients who received blood transfusions [16], this field has undergone some-thing of an upheaval with transfusion procedures and triggers being widely re-assessed. However, a more recent observational study, the SOAP study, showedno increased mortality in transfused patients, results that may be related to thewidespread introduction of leukodepletion in the preparation of blood for trans-fusion. The rationale behind transfusion of red blood cells is that they willimprove the blood’s oxygen carrying ability and hence improve tissue oxygena-tion; however, the resultant increase in hematocrit may increase viscosity, impair-ing blood flow, and limiting tissue oxygen delivery. This may be even more trueif red blood cells are stored for any length of time. Thus, transfusion policies needto be flexible, and while some patients will undoubtedly benefit from blood trans-fusion, each patient needs to be assessed individually. Hemoglobin solutions, withtheir beneficial effects on oxygen carriage and potential immunomodulatoryeffects, may find a place in the treatment of septic shock, but further clinical trialsare needed before these can be recommended [13].


Pump

Fluid administration alone is often not sufficient to restore hemodynamicstability, and vasoactive agents are required. As with all aspects of the manage-ment of sepsis, there has been keen debate in recent years as to which vasoac-tive agent(s) is superior to the others [4].

Vasopressor AgentsDopamine has often been used as the first-line vasopressor to raise blood

pressure. Dopamine combines �- and �-adrenergic properties, and at lowdoses, can also stimulate dopaminergic receptors, resulting in a selectiveincrease in splanchnic and renal blood flow. However, the routine use of low(renal) dose dopamine to prevent renal failure is not recommended, a majorargument being that a large randomized controlled trial failed to demonstrateany protective effect on renal function. Norepinephrine is also widely used inpatients with septic shock. Norepinephrine has predominantly �-adrenergicproperties, and hence is a stronger vasoconstrictor, but carries the risk thatperipheral blood flow will be reduced by an excessive increase in vasculartone. Dobutamine is therefore often added when norepinephrine is given.Some studies have suggested that norepinephrine may be a better first choicein patients with septic shock [7], but the jury is still out, and current guidelinesrecommend both drugs as appropriate first line agents [12, 15]. Epinephrinedecreases the splanchnic blood supply and is not recommended as a first linevasopressor in septic shock.

Patients with septic shock often have inappropriately low levels of serumvasopressin and low dose vasopressin (0.01–0.04 U/min) as replacement ther-apy has been demonstrated to produce a significant rise in mean arterial pres-sure in septic shock, and can reduce requirements for traditional vasopressors.Although this strategy has been shown to improve survival in animal models,no randomized, prospective clinical trials have assessed outcomes.

Inotropic AgentsWhile cardiac index is often normal or even high in the volume resusci-

tated septic shock patient, cardiac function is often impaired. The mechanismof the myocardial dysfunction is complex, likely involving alterations in intra-cellular calcium homeostasis and in �-adrenergic signal transduction, andmediated by various inflammatory mediators including cytokines and NO.Dobutamine, an adrenergic agonist that stimulates �1-, �2- and �1-adrenergicreceptors, is generally used as the first-line inotrope in septic shock.Dopexamine is an alternative inotrope, a dopamine analog, that initially seemed

Vincent 358

to have beneficial effects on splanchnic circulation, but these were not con-firmed. Studies have suggested that the phosphodiesterase inhibitor, enoxi-mone, may have beneficial anti-tumor necrosis factor (TNF) effects, andimprove the splanchnic circulation when compared to dobutamine, with similareffects on cardiac index and oxygen delivery, but these drugs are not widelyused in septic shock patients.

Other Management Strategies in Sepsis

Immunomodulatory TherapiesAfter many years of apparently fruitless research into immunomodulatory

therapies for patients with sepsis, there was a breakthrough in 2001 with thelicensing of drotrecogin alpha (activated) for the treatment of adult patientswith severe sepsis or septic shock. This drug, a recombinant version of a nat-ural anticoagulant protein, was shown in a multicenter randomized controlledtrial involving 1,690 patients to improve survival from 30.8% in the placebogroup to 24.7% in the drotrecogin alpha (activated) group, giving a 19.4% rel-ative reduction in mortality rate; i.e. only 16 patients need to be treated to saveone life [2]. Drotrecogin alpha (activated) (licensed under the name Xigris) alsoimproves organ function, and its effects on outcome are sustained beyond thetraditional 28-day endpoint. Obviously there are increased risks of bleedingwith drotrecogin alpha, but these are limited. In a study evaluating all reporteduse of this drug so far, serious bleeding events occurring during the infusionperiod in 2.8% of patients and during the 28-day study period in 5.3% ofpatients. Of the bleeding events during the infusion period, 43% were procedure-related. Patients at high risk of bleeding should not be given drotrecogin alpha(activated) and it is contraindicated in patients with active internal bleeding,recent hemorrhagic stroke, intracranial or intraspinal surgery, severe headtrauma, presence of an epidural catheter, intracranial neoplasm or evidence ofcerebral herniation. In addition, infusion should be interrupted for surgery orinvasive interventions. Importantly, early treatment with drotrecogin alpha(activated) seems to convey a survival advantage, and its use needs to be con-sidered early rather than as a last-resort option. Although expensive, its cost-effectiveness profile seems to be in keeping with other commonly usedinterventions in intensive care.

As our understanding of the sepsis response broadens, other immuno-modulatory strategies are continually being suggested, developed, and tested.Agents currently attracting interest include anti-oxidants, apoptosis manipulat-ing drugs, and anti-Toll-like receptor 4 (TLR-4) therapies. Hemofiltration strate-gies with large-pore hemofilters or coupled plasma filtration absorption [11]


have also been used in the hope that removal of inflammatory mediators maylimit the sepsis response. Further study is needed to evaluate the best filter andoptimum filtration rates, but this is a promising approach.

CorticosteroidsCorticosteroid therapy in septic patients has come in from the cold with

Annane et al. [1] demonstrating improved survival in patients with septic shockand relative adrenal insufficiency treated with a 50-mg intravenous bolus ofhydrocortisone every 6 h and fludrocortisone (50-�g tablet once daily). So far,this strategy has only been tested in patients with septic shock and whether italso applies to patients with less severe forms of sepsis remains to be seen. Inaddition, definitions of relative adrenal insufficiency are not clear and if corti-costeroid administration is to be guided by ACTH stimulation tests, which testshould be used, and when?

Glucose ControlIn a milestone study, Van den Berghe et al. [14] randomized more than 1,500

ICU patients to intensive management aimed at keeping blood sugar levels withintight limits of 80–110 mg/dl vs. conventional management of hyperglycemia;mortality rates were reduced from 8.0 to 4.6% (p � 0.04) in the intensive treat-ment group. In addition, intensive treatment was associated with shorter ICUstays, less requirement for renal replacement therapy, less hyperbilirubinemia,fewer bloodstream infections, fewer ICU neuropathies, and a reduced need fortransfusion. Further study has suggested that these results were due to the controlof glucose levels rather than to the insulin administered. While this strategy seemsan attractively cheap and simple means of improving outcomes in our septicpatients, additional blood sampling and nursing time can make it more complexand costly than would seem to be the case at first glance. Nevertheless, glucoselevels clearly need to be more carefully monitored and adjusted than has perhapspreviously been the case.

Conclusion

The last few years have seen exciting developments in the treatment ofsevere sepsis and septic shock. The standard, and still vitally important, man-agement of severe sepsis relies on adequate resuscitation with fluids andvasoactive agents, eradication of the causative infection using antibiotics andsurgical removal where necessary, and individual organ support including renaldialysis and mechanical ventilation. Now, immunomodulatory strategies have

Vincent 360

been shown to be effective at reducing mortality and can now be added to ourmanagement protocols.

These results are encouraging for all involved in the treatment of sepsis, butdo not provide all the answers. For example, as other immunomodulating drugs aredeveloped that also reduce mortality, the challenge will be to decide which agentshould be given to which patient? Or will combinations of drugs be needed, and ifso which? Genetic typing and improved markers of sepsis may help answer someof these questions. The recently suggested PIRO (predisposition, infection,immune response, organ dysfunction) system of ‘staging’ sepsis is still in the earlystages of development but will help to characterize patients, to target treatments,and to monitor response to therapy. Following the positive results from recenttrials, the future is certainly looking brighter for the patient with severe sepsis.

References

1 Annane D, Sebille V, Charpentier C, Bollaert PE, Francois B, Korach JM, Capellier G, Cohen Y,Azoulay E, Troché G, Chaumet-Riffaut P, Bellissant E: Effect of treatment with low doses ofhydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA 2002;288:862–871.

2 Bernard GR, Vincent JL, Laterre PF, LaRosa SP, Dhainaut JF, Lopez-Rodriguez A, Steingrub JS,Garber GE, Helterbrand JD, Ely EW, Fisher CJ Jr: Efficacy and safety of recombinant human acti-vated protein C for severe sepsis. N Engl J Med 2001;344:699–709.

3 Creteur J, Vincent JL: Hemoglobin solutions: Not just red blood cell substitutes. Crit Care Med2000;28:894–896.

4 De Backer D, Creteur J, Silva E, Vincent JL: Effects of dopamine, norepinephrine, and epineph-rine on the splanchnic circulation in septic shock: Which is best? Crit Care Med 2003;31:1659–1667.

5 Dieterich HJ: Recent developments in European colloid solutions. J Trauma 2003;54:S26–S30.6 Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D, Cohen J, Opal SM, Vincent JL,

Ramsay G: 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference.Crit Care Med 2003;31:1250–1256.

7 Martin C, Viviand X, Leone M, Thirion X: Effect of norepinephrine on the outcome of septicshock. Crit Care Med 2000;28:2758–2765.

8 Oliveira RP, Velasco I, Soriano F, Friedman G: Clinical review: Hypertonic saline resuscitation insepsis. Crit Care 2002;6:418–423.

9 Ricard JD, Dreyfuss D: ARDS controversies: Where do we stand now? in Vincent JL (ed):Yearbook of Intensive Care and Emergency Medicine. Heidelberg, Springer, 2004, pp 419–428.

10 Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, Peterson E, Tomlanovich M:Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001;345:1368–1377.

11 Ronco C, Brendolan A, Lonnemann G, Bellomo R, Piccinni P, Digito A, Dan M, Irone M,La Greca G, Inguaggiato P, Maggiore U, De Nitti C, Wratten ML, Ricci Z, Tetta C: A pilot studyof coupled plasma filtration with adsorption in septic shock. Crit Care Med 2002;30:1250–1255.

12 Task Force of the American College of Critical Care Medicine, Society of Critical Care Medicine:Practice parameters for hemodynamic support of sepsis in adult patients in sepsis. Crit Care Med1999;27:639–660.

13 The ARDS Network: Ventilation with lower tidal volumes as compared with traditional tidalvolumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000;342:1301–1308.


14 Van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, Vlasselaers D,Ferdinande P, Lauwers P, Bouillon R: Intensive insulin therapy in the critically ill patient. N EnglJ Med 2001;345:1359–1367.

15 Vincent JL: Hemodynamic support in septic shock: Guidelines for the management of severe sep-sis and septic shock. International Sepsis Forum. Intens Care Med 2001;27(suppl 1):S80–S92.

16 Vincent JL, Baron JF, Reinhart K, Gattinoni L, Thijs L, Webb A, Meier-Hellmann A, Nollet G,Peres-Bota D: Anemia and blood transfusion in critically ill patients. JAMA 2002;288:1499–1507.

Prof. Jean-Louis VincentDepartment of Intensive Care, Erasme University HospitalRoute de Lennik 808, BE–1070 Brussels (Belgium)Tel. �32 2 555 3380, Fax �32 2 555 4555, E-Mail [email protected]


High Volume Hemofiltration in Critically Ill Patients:Why,When and How?

Ciro Tettaa, Rinaldo Bellomob, John Kellumc, Zaccaria Riccid,Robert Pohlmeiera, Jutta Passlick-Deetjena, Claudio Roncod

aResearch Extracorporeal Therapies, Fresenius Medical Care Deutschland GmbH,Bad Homburg, Germany; bIntensive Care Research and Division of Surgery, AustinHospital and Melbourne University, Heidelberg, Australia; cUniversity of PittsburghSchool of Medicine, Department of Critical Care Medicine, and MANTRALaboratories, Pittsburgh, Pa., USA; dDepartment of Nephrology, Dialysis andTransplantation, St Bortolo Hospital, Vicenza, Italy

The systemic inflammatory response syndrome (SIRS) is characterized bywidespread endothelial damage caused by persistent inflammation from bothinfectious and non-infectious stimuli. The host uses hormonal and immunologicmechanisms to counteract SIRS. Hypoperfusion and shock result when homeo-static mechanisms are no longer able to keep the system in balance, leading toorgan dysfunction.

Septic shock can be defined as sepsis with hypotension, despite adequatefluid resuscitation, along with evidence of multiple organ dysfunction syndrome(MODS). Sepsis is the leading cause of acute renal failure (ARF). The mortalityrate of patients requiring dialysis for ARF in the ICU (50–60%) is nearly twiceas high as comparable patients without ARF [1, 2]. For example, one large study(17,000 patients admitted to Austrian ICUs) found that ARF was associated witha greater than fourfold increase in mortality [3]. Even after controlling for under-lying severity of illness, mortality was still significantly higher in people withARF (62.8 vs. 38.5%), suggesting that ARF is independently responsible forincreased mortality, even if renal replacement therapy is used.

The pathogenesis of sepsis is complex. Endotoxins or lipopolysaccharidesproduce both cellular and humoral effects. Increased levels of plasmacytokines such as tumor necrosis factor-� (TNF�), interleukins (IL-1, IL-6,IL-8 and IL-10), IL-1 receptor antagonist, soluble TNF receptors types I and II

The Rationale of High Volume Hemofiltration 363

and lipid mediators such as platelet-activating factor (PAF) are produced andare considered to be of pathogenetic relevance [4–6].

Initially, several experimental and clinical approaches attempted to remove(or neutralize) bacterial components or cytokines produced early in the inflam-matory response. The major therapeutic approaches (reviewed in [4, 5, 7]) include:• Blocking of microbial products (from gram-negative or gram-positive

bacteria; anti-LPS antibodies, antibacterial proteins, circulating proteins).• Blocking of cellular stimulation (soluble CD14, anti-CD14 antibodies,

anti-lipopolysaccharide-binding protein).• Blocking of cytokine activity (monoclonal antibodies, soluble receptors).• Removal or neutralization of endotoxin by several strategies aimed specifi-

cally at endotoxin removal or neutralization included: blood passage throughpolymyxin immobilized cellulose filters, antibodies directed against theanti-O side chain antibody, anti-core or anti-lipid A antibodies, or by the useof binding peptides that neutralize or enhance endotoxin clearance.Although many in vitro and animal models provided encouraging results,

large-scale clinical trials provided little or no confirmation of clinical benefitderived from the removal of bacterial components. Therefore, there is wide-spread doubt on the efficacy of LPS removing strategies. In animal models ofendotoxic shock, intravenously injected endotoxin disappears rapidly from theplasma (usually in less than 2 h) [8]. It is common, even in these well-definedmodels, to observe a high degree of individual variability to endotoxin toler-ance, cytokine levels, mortality related to dose as well as genetic susceptibilityand cytokine polymorphism (fig. 1).

The concept of sepsis as a simply proinflammatory event has been subse-quently challenged [9–11]. Terms such as ‘monocyte hyporesponsiveness’ or‘monocyte deactivation’ have been introduced in order to indicate what is man-ifested as a ‘Compensatory Anti-inflammatory Response Syndrome’ (CARS).As the network acts like a cascade, early intervention would seem most benefi-cial. On the other hand, sepsis does not fit a one-hit-model but shows complexand multiple increases and decreases in mediator levels over time. Neithersingle-mediator-directed nor one-time interventions therefore seem appropri-ate. One of the major criticisms attributed to continuous renal replacement ther-apy (CRRT) in sepsis – its lack of specificity – could turn out to be a majorstrength. Non-specific removal of soluble mediators – be they pro- or anti-inflammatory – without completely eliminating their effect may be the mostlogical and adequate approach to a complex and long-running process like sep-sis. The concept of cutting peaks of soluble mediators, e.g. through continuoushemofiltration is a paradigm called ‘the peak concentration hypothesis’ [12].However, removal rates and clearances of pro-inflammatory cytokines by con-ventional hemodialysis are hindered by limited diffusive or convective rates.

Tetta/Bellomo/Kellum/Ricci/Pohlmeier/Passlick-Deetjen/Ronco 364

High Volume Hemofiltration (HVHF):Why?

For several years now, the real capability of hemofiltration in removinginflammatory mediators has remained controversial. Numerous ex vivo as wellas animal and human studies have shown that hemofilters can extract nearlyevery substance involved in sepsis to a certain degree [as reviewed in ref. 7].Prominent examples are complement factors [13, 14] TNF, IL-1, IL-6 [15, 16],IL-8, and PAF [17, 18]. Regarding plasma cytokine levels, their decreaseappeared nevertheless of minor degree. Other studies could not show any

HistamineProstaglandins ElastasePAF O2-Metabolites sCD14

C1,C3 activation

Macrophage monocytes

Mono/granulocytes

Complement activationCell and organ

damage Activation of coagulationEndothelial dysfunction

LBP BPI

MODS

Endotoxin

Exotoxin

TNF-�, IL-1, IL-6, IL-8

Genetic susceptibility

Cytokine polymorphism

SIRS

Fig. 1. Sepsis represents a complex series of cascade events involving the activation ofthe innate response, synthesis and release of pro- and anti-inflammatory mediators and theinitial functional derangement of vascular endothelium, and activation of the coagulation.These events together with the biological expression of complement activated products (viathe C1 classical or C3 alternate pathway) converge to inducing cell and organ functional andanatomical damage thus ensuing in the systemic inflammatory response syndrome (SIRS)and in the multiple organ dysfunction syndrome (MODS). Individual sensitivity to gram-negative endotoxin and gram-positive exotoxin is played by different mechanisms: at thecell-membrane receptor recognition pathway (Toll-like receptors), by the enhancing effect oflipopolysaccharide-binding protein (LBP) in complex with endotoxin and/or the competingclearance role of bactericidal/permeability-inducing protein. Furthermore, cytokine poly-morphism (for TNF and IL-10) is associated with increased susceptibility [for review, seerefs 52 and 53].


influence on cytokine plasma levels by CRRT [19–22]. Furthermore, even therelevance of plasma cytokines has been questioned (reviewed in [23]). Moreimportant is the proof that significant hemodynamic improvement has beenachieved even without measurable decreases in cytokine plasma levels [24].The removal of substances different to the measured cytokines might havebeen responsible for the achieved effect. Alternatively, bioactive substancesincluding some of the measured cytokines might be removed causing theobserved beneficial effect. Nowadays, the removal from plasma of substancesexerting measurable biologic effects rather than of a single or a single class ofmediators seems more important. Several explanations have been discussed inthe recent literature: the removal of the priming effect on polymorphonuclearneutrophils [18], of the myocardial depressant factor(s) [25, 26] and of theendotoxin-mediated pro-apoptogenic activity on myocytes [27]. When theresponse to sepsis is viewed in a network perspective, absolute values wouldbe less relevant than relative ones within an array of interdependent mediatorsas even small decreases could induce major balance changes. This makesmeasurement of cytokine plasma levels debatable while more local or tissuelevels should be preferably measured whenever possible. These issues are stillextremely controversial. However, these data emphasize that convection andpossibly also diffusion may achieve a certain degree of ‘blood purification’.

Consequently, it would be logical to try to improve the efficiency of theextracorporeal treatments by increasing the amount of plasma water exchange,i.e. increasing ultrafiltration rates.

Animal studies provide great support to this concept. Starting in the earlynineties, several studies using different septic animal models examined theeffect of high ultrafiltration rates (up to 300 ml/kg/h) on physiological para-meters and outcome. In a landmark study, a porcine model of septic shockinduced by endotoxin infusion was investigated [28]. The animals developedprofound arterial hypotension, a decrease in cardiac output, stroke volume,and right ventricular stroke work index. With HVHF (at 6 liters/h), rightventricular function, blood pressure and cardiac output showed a remark-able improvement compared to control and sham-filtered animals [28, 29].The same group extended their findings in the same model by i.v. administra-tion of ultrafiltrate from LPS-infused animals into healthy animals. The ani-mals receiving ultrafiltrate from endotoxemic animals rapidly developedhemodynamic features of septic shock while animals infused with ultrafiltratefrom healthy animals showed a moderate blood pressure rise [28]. In a furtherstudy by the same group, a bowel ischemia-reperfusion model in pigs wasinvestigated. HVHF started before clamping of the superior mesenteric arterysignificantly diminished bowel damage and prevented hemodynamic deterio-ration [30].


These studies established that a convection-based treatment can removesubstances with hemodynamic effects resembling septic shock, when sufficientlyhigh ultrafiltration rates are applied [31, 32].

Several animal studies confirmed and refined these results. In three ofthem [33–35], the correlation of survival with ultrafiltration rate was demon-strated. Significant improvements in cardiac function, systemic and pulmonaryvascular resistance, and hepatic perfusion were found. Another study in lambsshowed significant improvements in lung function [36].

A study in pigs made septic by induced pancreatitis compared low-volumehemofiltration with HVHF at 100 ml/kg/h of ultrafiltrate. In this study, the influ-ence of frequent filter changes on survival, changes in TNF levels as well asmonocyte and polymorphonuclear neutrophil function were analyzed [35]. Earlyfilter change was used to delineate the effect of cytokine removal by adsorptionon the filter since membrane capacity was saturated after a few hours. By chang-ing filters, adsorption continued to a certain extent. In this model, a hyper-dynamic septic state was induced through an intervention which approximates theunderlying conditions encountered in human sepsis. Additionally, the interventionstarted late to simulate real clinical conditions. Hemofiltration was commencedwhen the animals had already developed the clinical picture of hyperdynamicseptic shock. HVHF was superior for all measured endpoints. Of relevance,increasing ultrafiltration had more effect than frequency of filter change [35].

Closer to human sepsis has been the finding that the ultrafiltration dose iscorrelated to outcome in critically ill patients with ARF. In a large randomized,controlled study including 425 patients, an ultrafiltration dose of 35 ml/kg/hincreased survival rate from 41 to 57% compared to a dose of 20 ml/kg/h [37].Eleven to 14% (per randomization group) of the patients had sepsis. In thesesubgroups there was a trend of a direct correlation between treatment dose andsurvival even above 35 ml/kg/h in contrast to the whole group where a survivalplateau was reached. This lends support to the concept of a ‘sepsis dose’ ofhemofiltration in septic patients contrasting to a ‘renal dose’ in critically illpatients without systemic inflammation, the former being probably distinctlyhigher (without proven upper limit). Of note, there was no increase in adverseeffects even with the highest ultrafiltration dose [38].

Over the last few years several human studies have examined the clinicaleffects of HVHF. In 20 children undergoing cardiac surgery, zero-balancedHVHF was administered with UF rates equivalent to 7–9 liters/h for a 70-kgadult [39]. Cytokine plasma levels and clinical endpoints associated with thecardiopulmonary bypass were examined. There was a significant reduction inpostoperative blood loss, time to extubation and improvement in the alveolar toarterial oxygen gradient as well as reduction of cytokine plasma levels atdifferent time points.


In a prospective cohort study, observed and predicted mortality was com-pared in prospectively stratified prognostic groups (306 critically ill patientswith varying underlying diseases) treated with a mean ultrafiltration rate of3.8 liters/h [40]. Observed survival rates were significantly higher in the treatedpopulation compared to predicted survival by three well-validated scores.

In another trial in 11 septic patients with shock and MODS, a randomizedcross-over design of 8 vs. 1 liter/h ultrafiltration was applied [41]. The HVHFgroup displayed significantly greater reduction in vasopressor requirements.Both treatment groups showed a decrease in C3a and C5a plasma levels whichwas significantly greater in the HVHF group.

Impressive clinical results were obtained in an evaluation of short-termHVHF in 20 patients in catecholamine-refractory septic shock comprising apatient cohort with very poor expected survival [42]. A control group was notdefined. Only one 4-hour session of HVHF removing 35 liters of ultrafiltratereplaced by bicarbonate-containing fluid was applied as soon as mean bloodpressure could not be stabilized above 70 mm Hg with dopamine, norepineph-rine and epinephrine after appropriate volume resuscitation. HVHF wasfollowed by conventional continuous venovenous hemofiltration (CVVH).Endpoints included an increase in cardiac index, mixed venous oxygen satura-tion and arterial pH and decrease in norepinephrine requirements. Elevenpatients reached all predefined endpoints and showed impressively goodsurvival (9 of 11) at 28 days. Nine patients did not reach all endpoints and hada 100% mortality rate. Apart from responding to HVHF, only time from admis-sion to start of HVHF and body weight were survival-associated factors in theanalysis. Patients with higher body weight did worse possibly because theyreceived a smaller ultrafiltration dose per body weight as discussed by theauthors [42]. Furthermore, time to start treatment also appeared to be correlatedwith higher mortality [42].

These trials should be interpreted with caution with respect to their limiteddesign but they certainly deliver sound evidence of feasibility and efficacy toset the stage for a large-scale trial on HVHF in sepsis.

Studies comparing HVHF and standard CVVH often lack the statisticalpower. In their studies, Bouman et al. [43] showed no difference in 28-daymortality and recovery of renal function between early (within 12 h from ARFappearance) HVHF, early standard CVVH and late standard CVVH. In thispopulation, survival was very high compared to other reports, and recovery ofrenal function was 100% in all hospital survivors. Due to the high clearance,HVHF was at least not less effective in the study endpoints with the notewor-thy advantage that in case of need (for logistic reason, medical needs etc), itcould be interrupted. However, this study lacks the statistical power and wasperformed in a restricted number of closed format ICUs.


Despite the positive results of the Ronco trial [37], the practice of higherintensity CRRT has not been widely adopted in the common practice of ICUs.In Australia and New Zealand, almost 100% of treatments are continuous ther-apies (CVVH or continuous venovenous hemodiafiltration, CVVHDF). A sur-vey of several units active in the Australian and New Zealand Intensive CareSociety (ANZICS) Clinical Trials Group (CTG) found that very few (�3) unitshad adopted the intensive CRRT regimen proposed by Ronco’s study [37;Bellomo, unpubl. data, 2002]. Data from such Australian units shows insteadthat the vast majority (�90%) prescribes a ‘fixed’ standard therapy dose of 2liters/h which is not adjusted for body weight. Thus, a 100-kg man wouldreceive 20 ml/kg/h, the dose shown to have the worst outcome in Ronco’s study[37]. In another recent study that involved several Australian units [‘BESTkidney study’ – Bellomo et al., in press], the median body weight for Australianpatients was 81 kg, thus indicating that the vast majority receive a CRRTintensity of approximately 25 ml/kg/h of effluent. Finally, while in Ronco’sstudy [37], the technique of CRRT was uniform as CVVH with post-filter fluidreplacement, current practice in Australia includes a variety of techniques withCVVH in the pre-dilution mode representing perhaps 80% of techniques in useand CVVHDF representing most of the remainder.

There is some awareness of Ronco’s study [37] in the Intensive Carecommunity but specialists reported having doubts over the applicability of theresults to the wider Intensive Care population. Major concerns relate to thesomewhat unusual patient population studied (very low incidence of sepsiswhich, on the other hand, dominates as the major cause of ARF in Australia),the fact that the study was conducted over 5 years, the fact that the typical doseof CRRT in Australia is already on average 20% greater than the lowest doseprescribed in the study, the additional cost of intensifying therapy, and greatconcern about extrapolating the results of an unblinded, single-center study tothe wider Australian Intensive Care population. Thus, the findings of this studyhave not induced a change in practice despite the potential survival benefits ofmore intensive continuous therapies.

In essence, HVHF has been found to produce unexpected physiologicalbenefits in terms of ‘blood purification’ on different surrogate outcomemeasures in both animal and human studies. However, the real benefit (i.e. onsurvival) still needs to be proven.

HVHF:When?

The initial hemodynamic improvement with HVHF seems to be main-tained over the next 96 hours, with a significant decrease in norepinephrine


requirement (more than 70%) [42]. The reason for the reduction of norepi-nephrine requirement has not been clarified as this seems to be independentof the fluid balance and the decrease in body temperature. Removal of seda-tive drugs with vasodilatory effect such as midazolam and sufentanil has notextensively been studied. However, for morphine, only 1–3% of the infuseddose is removed by high efficiency, high flux hemodialysers [44]. Theremoval of mediators is likely to play a role. Filtration volume is an importantfactor influencing the efficacy of the technique. A sequence of adsorption andfiltration was shown with platelet-activating factor using high permeabilitypolysulfone membranes [17]. The question of when to start the treatment isobviously related to the major indication for initiating HVHF, i.e. restoringhemodynamic stability. Thus, HVHF could become the elective treatment inhemodynamically unstable patients. Treatment should be started as early asthis is possible as this was clearly shown to affect survival rates [45]. We haveproposed the concept of HVHF in the framework of a ‘pulse therapy’ (fig. 2).This concept meets also with practical and technical considerations. A dailyschedule of HVHF (85 ml/kg/h) associated with CVVH (at 35 ml/kg/h)leading to a cumulative dose of approximately 48 ml/kg/h (in a 70 kgpatient) would be practical and less costly. This schedule is now under clini-cal investigation.

HVHF: How?

To reach ultrafiltration rates such as 50–60 ml/kg/h, two major determi-nants of ultrafiltration need to be considered. A high blood flow rate (as com-pared to conventional CRRT) requires a vascular access (e.g. 14 F catheters) toattain a constant flow of 300 ml/min. A filtration fraction of 25% can then be

01,0002,0003,0004,0005,0006,0007,0008,0009,000

50 70Weight (kg)

ml/h

85ml/h/kg 35ml/h/kg

Fig. 2. The relevance of relating effluent volumes to body weight is shown here. In a70-kg patient, 6 liters/h.


Tabl

e 1.

Tech

nica

l fea

ture

s of

com

mer

cial

ly a

vaila

ble

CR

RT

mon

itors

Mon

itors

Com

pany

Pum

psB

lood

H

eate

rH

epar

in

Rei

nfec

tion

Scal

esA

vaila

ble

trea

tmen

tsfl

owpu

mp

site

Aqu

ariu

s*E

dwar

ds

40–

450

yes

yes

Pre-

Post

-Pre

/2

SCU

F/C

VV

H/H

VH

F/C

VV

HD

/Po

stC

VV

HD

F/M

PS(P

EX

)/H

P -

ped

Acc

ura*

Bax

ter

40–

450

yes

yes

Pre-

Post

-Pre

/2

SCU

F/C

VV

H/H

VH

F/C

VV

HD

/Po

stC

VV

HD

F/M

PS(P

EX

)/H

P -

ped

BM

25E

dwar

ds

330

–500

nono

Pre-

Post

2SC

UF/

CV

VH

/CV

VH

D/

PEX

- p

edB

M25

Bax

ter

330

–500

nono

Pre-

Post

2SC

UF/

CV

VH

/CV

VH

D/

PEX

- p

edD

iapa

ctB

raun

310

–500

yes

noPr

e-Po

st1

IHD

/IH

FD/I

HF/

Pex/

SCU

F/C

VV

H/C

VV

HD

/CV

VH

DF

Equ

a-sm

art

Med

ica

35–

400

yes

yes

Pre-

Post

3SC

UF/

CV

VH

/HV

HF/

CV

VH

D/

CV

VH

DF/

MPS

(PE

X)/

HP

- pe

dM

ultim

atB

ellc

o2

or 3

0–40

0no

yes

Pre-

Post

1SC

UF/

CV

VH

/HV

HF/

CV

VH

D/

CV

VH

DF/

CPF

AH

F 40

0*In

fom

ed4

0–45

0ye

sye

sPr

e-Po

st-P

re/

2SC

UF/

CV

VH

/CV

VH

D/C

VV

HD

F/Po

stH

VH

F/M

PS(P

EX

)/H

P -

ped

Hyg

eia

plus

Kim

al*

40–

500

yes

yes

Pre-

Post

-Pre

/vo

lum

etri

cSC

UF/

CV

VH

/CV

VH

D/

Post

CV

VH

DF/

MPS

(PE

X)/

HP

- pe

dPe

rfor

mer

Ran

d4

5–50

0ye

sye

sPr

e-Po

st1

SCU

F/C

VV

H/C

VV

HD

/CV

VH

DF/

MPS

(PE

X)/

HP

- pe

dPr

ism

a H

ospa

l4

0–18

0B

lood

yes

Pre-

Post

-Pre

/3

SCU

F/C

VV

H/C

VV

HD

/w

arm

Post

CV

VH

DF/

MPS

(PE

X)/

HP

- pe

dM

ultif

iltra

te*

FMC

410

–500

ye

sye

sPr

e-Po

st-P

re/

4SC

UF/

CV

VH

/HV

HF/

CV

VH

D/

Post

CV

VH

DF/

MPS

(PE

X)/

HP

- pe

d

SCU

F�

Slow

con

tinuo

us u

ltraf

iltra

tion;

CV

VH

�co

ntin

uous

ven

oven

ous

hem

ofilt

ratio

n; H

VH

F�

high

vol

ume

cont

inou

s ve

nove

nous

hem

ofil-

trat

ion;

CV

VH

D�

cont

inuo

us v

enov

enou

s he

mod

ialy

sis;

CV

VH

DF

�co

ntin

uous

ven

oven

ous

hem

odia

filtr

atio

n; M

PS�

mem

bran

e pl

asm

a se

pa-

ratio

n fo

r pla

smap

here

sis

(PE

X) o

r hem

oper

fusi

on; C

PFA

�co

uple

d pl

asm

a fi

ltrat

ion

adso

rptio

n; p

ed�

indi

cate

s w

hene

ver p

edia

tric

trea

tmen

ts c

anbe

per

form

ed. T

he a

ster

isk

indi

cate

s th

e m

onito

rs th

at m

ay p

erfo

rm H

VH

F. I

nfor

mat

ion

deriv

ed f

rom

Com

pany

bro

chur

es a

nd f

rom

ref

48.


set. Bicarbonate buffered hemofiltration fluid (35 mmol/l) should be adminis-tered 1/3 in predilution and 2/3 in post-dilution [46]. A highly biocompatible,synthetic membrane and a dialyzer surface of 1.8–2 m2 (in 70-kg patients)should be considered. Pre-dialyzer pressures can be informative in the earlystages of dialyzer clotting [47]. In this context, the membrane biocompatibilityand the dialyzer geometry are important. High pre-dialyzer pressures especiallyin HVHF will inevitably impact negatively on dialyzer life, making the tech-nique cumbersome, increasing patient’s risk and the staff workload. For severalyears, the major hindrance to the widespread use of HVHF has been the lack ofreliable monitors capable of insuring safety, precision and continuation of thistherapy. On the other hand, the controversy on the real advantages and theincreased costs of the therapy did not immediately persuade industry on such atherapy [48]. Nowadays, different monitors from different manufacturers existand are available. Table 1 lists the different features of commercially availablemonitors. Figure 3 provides an example of how continuous therapy monitorshave developed in order to make HVHF practical. Some important hardwareand software aspects must be considered such as high precision scales(equipped with software for on-line continuous testing and a high capacity) andpowerful heating systems for maintaining constant sufficiently high tempera-ture of the high volumes of infusion solution.

The practice of HVHF needs adequate tailoring of the therapy. Uchinoet al. [49] studied the effect of changing the amount of pre-dilution replacementfluid on the sieving coefficient (SC) and mass transfer of small solutes duringisovolemic HVHF. These authors interestingly concluded with the warning that

a

b

c

Fig. 3. New technology for HVHF. The example of MultiFiltrate (Fresenius MedicalCare) is shown. a Two integrated heating systems (35/39�C) with fast, effective temperaturecontrol. b 4 high precision scales with on-line continuous testing with a capacity up to 24liters HF solution; C � cassette for easy, quick line system set-up.


during isovolemic HVHF, small solute sieving coefficients and mass transfer(sodium, potassium, chloride, total calcium, total magnesium, phosphate, totalCO2, urea, creatinine and glucose) are significantly affected by the proportionof replacement fluid administered pre-filter. Isovolemic HVHF is neitherisonatremic nor isochloremic and requires close monitoring of sodium, glucose,and acid-base balance.

Conclusion

A vast array of mostly water-soluble mediators play a strategic role in theseptic syndrome. At variance with targeting single mediators, therapeutic inter-vention aiming at the non-selective removal of pro- and anti-inflammatorymediators seems a rational concept. Therefore, sequentially appearing peaks ofsystemic mediator overflow could be curbed as well as persistently high plasmalevels reduced. Antagonizing pro- and anti-inflammatory processes by reducingthe relative excess of active substances could be summarized under the term of‘peak concentration hypothesis’.

Recent animal and human trials have delivered much support to thisconcept. It has been conclusively shown that treatment dose in continuous ther-apies is a major factor influencing survival in the critically ill patient with ARF.Accumulating evidence suggest that HVHF compared to conventional CVVHimproves laboratory severity markers and clinical outcome, i.e. survival.Monitors to perform HVHF safely are now available.

Yet the evidence still is not strong enough to recommend HVHF outsideclinical studies taking into account possible adverse effects of the technique.A large-scale clinical trial is urgently needed to resolve the issue.

Other blood purification techniques using large pore membranes or plasmafiltration with adsorbent perfusion are in the early stages of clinical testing[50, 51]. They are conceptually promising and possibly constitute an importantrefinement.

Acknowledgment

The authors would like to thank Dr. S. Civardi and M. Brambilla for providing usefulinformation.

References

1 Mehta RL, Mc Donald B, Gabbai FB, et al: A randomized clinical trial for continuous versus inter-mittent dialysis for acute renal failure. Kidney Int 2001;60:1154–1163.


2 Liano F, Junco E, Pascual J, et al: The spectrum of acute renal failure in the intensive care unitcompared with that seen in other settings. The Madrid Acute Renal Failure Study Group. KidneyInt Suppl 1998;66:S16–S24.

3 Metnitz PG, Krenn CG, Steltzer H, et al: Effect of acute renal failure requiring renal replacementtherapy on outcome in critically ill patients. Crit Care Med 2002;30:2051–2058.

4 Cohen J: The immunopathogenesis of sepsis. Nature 2002;420:885–891.5 Hotchkiss RE, Karl IE: The pathophysiology and treatment of sepsis. N Engl J Med 2003;348:

138–150.6 Tetta C, De Nitti CM, Wratten ML, et al: Do circulating cytokines really matter in sepsis? Kidney

Int Suppl 2003;84:S69–S71.7 Tetta C, Bellomo R, Ronco C: Artificial organ treatment for multiple organ failure, acute renal

failure, and sepsis: Recent new trends. Artif Organs 2003;27:202–213.8 Taylor FB, Haddad PA, Hack E, et al: Two-stage response to endotoxin infusion into normal

human subjects: Correlation of blood phagocyte luminescence with clinical and laboratorymarkers of the inflammatory, hemostatic response. Crit Care Med 2001;29:326–334.

9 Pinsky MR: Sepsis: A pro- and anti-inflammatory disequilibrium syndrome. Contrib Nephrol.Basel, Karger, 2001, vol 132, pp 354–366.

10 Cavaillon JM, Adib-Conquy M, Cloez-Tayarani I, Fitting C: Immunodepression in sepsis andSIRS assessed by ex vivo cytokine production is not a generalized phenomenon: A review.J Endotoxin Res 2001;7:85–93.

11 Volk HD, Reinke P, Krausch D, et al: Monocyte deactivation: Rationale for a new therapeutic strat-egy in sepsis. Intens Care Med 1996;(suppl 4):S474–S481.

12 Ronco C, Ricci Z, Bellomo R: Importance of increased ultrafiltration volume and impact onmortality: Sepsis and cytokine story and the role of continuous veno-venous hemofiltration. CurOpin Nephrol Hypern 2001;10:755–761.

13 Goldfarb S, Golper TA: Proinflammatory cytokines and hemofiltration membranes. J Am SocNephrol 1994;5:228–232.

14 Hoffmann JN, Hartl WH, Deppisch R, et al: Hemofiltration in human sepsis: Evidence for elimi-nation of immunomodulatory substances. Kidney Int 1995;48:1563–1570.

15 Gasche Y, Pascual M, Suter PM, Favre H, Chevrolet JC, Schifferli JA: Complement depletionduring haemofiltration with polyacrylonitrile membranes. Nephrol Dial Transplant 1996;11:117–119.

16 Kellum JA, Johnson JP, Kramer D, Palevsky P, Brady JJ, Pinsky MR: Diffusive vs. convective ther-apy: Effects on mediators of inflammation in patients with severe systemic inflammatory responsesyndrome. Crit Care Med 1998;26:1995–2000.

17 Ronco C, Tetta C, Lupi A, Galloni E, Bettini MC, Sereni L, Mariano F, DeMartino A, Montrucchio G,Camussi G, LaGreca G: Removal of platelet-activating factor in experimental continuous arteri-ovenous hemofiltration. Crit Care Med 1995;23:99–107.

18 Mariano F, Tetta C, Guida GE, Triolo G, Camussi G: Hemofiltration reduces the priming activityon neutrophil chemiluminescence in septic patients. Kidney Int 2001;60:1598–1605.

19 Sander A, Armbruster W, Sander B, et al: Haemofiltration increases IL-6 clearance in earlysystemic inflammatory response syndrome but does not alter IL-6 and TNF alpha plasmaconcentrations. Intens Care Med 1997;23:878–884.

20 De Vriese AS, Colardyn FA, Philippe JJ, Vanholder RC, De Sutter JH, Lameire NH: Cytokineremoval during continuous hemofiltration in septic patients. J Am Soc Nephrol 1999;10:846–853.

21 De Vriese AS, Vanholder RC, Pascual M, Lameire NH, Colardyn FA: Can inflammatory cytokinesbe removed efficiently by continuous renal replacement therapies? Intens Care Med 1999;25:903–910.

22 Ronco C, Brendolan A, Lonnemann G, Bellomo R, Piccinni P, Digito A, Dan M, Irone M,LaGreca G, Ingaggiato P, Maggiore U, De Nitti C, Wratten ML, Tetta C: A pilot study on coupledplasma filtration with adsorption in septic shock. Crit Care Med 2002;30:1250–1255.

23 Tetta C, Bellomo R, D’Intini V, et al: Do circulating cytokines really matter in sepsis? Kidney IntSuppl 2003;84:S69–S71.

24 Heering P, Morgera S, Schmitz FJ, et al: Cytokine removal and cardiovascular hemodynamics inseptic patients with continuous veno-venous hemofiltration. Intens Care Med 1997;23:288–296.


25 Pathan N, Sandiford C, Harding SE, Levin M: Characterization of a myocardial depressant factorin meningococcal septicemia. Crit Care Med 2002;30:2191–2198.

26 Pathan N, Hemingway CA, Alizadeh AA, et al: Role of interleukin 6 in myocardial dysfunction inmeningococcal septic shock. Lancet 2004;363:203–209.

27 Carlson DL, Lightford E, Bryant DD, et al: Burn plasma mediates cardiac myocyte apoptosis viaendotoxin. Am J Physiol 2002;282:H1907–H1914.

28 Grootendorst AF, van Bommel EFH, van der Hoven B, van Leengoed LAM, van Osta ALM: Highvolume hemofiltration improves right ventricular function of endotoxin-induced shock in the pig.Intens Care Med 1992;18:235–240.

29 Grootendorst AF, van Bommel EFH, van der Hoven B, et al: High volume hemofiltration improveshemodynamics of endotoxin-induced shock in the pig. J Crit Care 1992;7:67–75.

30 Grootendorst AF, van Bommel EFH, van Leengoed LAM, et al: High volume hemofiltrationimproves hemodynamics and survival of pigs exposed to gut ischemia and reperfusion. Shock1994;2:72–78.

31 Grootendorst AF, van Bommel EF, van Leengoed LA, et al: Infusion of ultrafiltrate from endo-toxemic pigs depresses myocardial performance in normal pigs. J Crit Care 1993;8:161–169.

32 Nagashima M, Shin�oka T, Nollert G, et al: High-volume continuous hemofiltration during car-diopulmonary bypass attenuates pulmonary dysfunction in neonatal lambs after deep hypothermiccirculatory arrest. Circulation 1998;98(suppl 19):II378–II384.

33 Lee PA, Matson JR, Pryor RW, Hinshaw LB: Continuous arteriovenous hemofiltration therapy forStaphylococcus aureus-induced septicemia in immature swine. Crit Care Med 1993;21:914–924.

34 Rogiers P, Zhang H, Smail N, et al: Continuous venovenous hemofiltration improves cardiac per-formance by mechanisms other than tumor necrosis factor-alpha attenuation during endotoxicshock. Crit Care Med 1999;27:1848–1855.

35 Yekebas EF, Eisenberger CF, Ohnesorge H, et al: Attenuation of sepsis-related immunoparalysisby continuous veno-venous hemofiltration in experimental porcine pancreatitis. Crit Care Med2001;29:1423–1430.

36 Nagashima M, Shin�oka T, Nollert G, et al: High volume continuous hemofiltration during car-diopulmonary bypass attenuates pulmonary dysfunction in neonatal lambs after deep hypothermiccirculatory arrest. Circulation 1998;98(suppl 19):II378–II384.

37 Ronco C, Bellomo R, Homel P, et al: Effects of different doses in continuous veno-venoushaemofiltration on outcomes of acute renal failure: A prospective randomised trial. Lancet 2000;356:26–30.

38 Chertow GM, Lazarus JM, Paganini EP, et al: Predictors of mortality and the provision of dialy-sis in patients with acute tubular necrosis. The Auriculin Anaritide Acute Failure Study Group.J Am Soc Nephrol 1998;9:692–698.

39 Journois D, Israel Biet D, Pouard P, et al: High-volume, zero-balanced hemofiltration to reducedelayed inflammatory response to cardiopulmonary bypass in children. Anesthesiology 1996;85:965–976.

40 Oudemans-van Straaten HM, Bosman RJ, et al: Outcome of critically ill patients treated withintermittent high-volume hemofiltration: A prospective cohort analysis. Intens Care Med 1999;25:814–821.

41 Cole L, Bellomo R, Journois D, et al: High-volume hemofiltration in human septic shock. IntensCare Med 2001;27:978–986.

42 Honore PM, Jamez J, Wauthier M, et al: Prospective evaluation of short-term, high-volumeisovolemic hemofiltration on the hemodynamic course and outcome in patients with intractablecirculatory failure resulting from septic shock. Crit Care Med 2000;28:3581–3587.


44 Jamal JA, Hoh J, Bastani B: Removal of morphine with the high-efficiency and high-fluxmembranes uring hemofiltration and heamodiafiltration. Nephrol Dial Transpl 1998;13:1535–1537.

45 Rivers L, Nguyen B, Havstad S, et al: Early goal-directed therapy in the treatment of severe sep-sis and septic shock. N Engl J Med 2001;145:1368–1377.


46 Uchino S, Fealy N, Baldwin I, Morimatsu H, Bellomo R: Pre-dilution vs. post-dilution during con-tinuous veno-venous hemofiltration: Impact on filter life and azotemic control. Nephron ClinPract. 2003;94:c83–c84.

47 Pacitti A, Cantaluppi V, Fenoglio R, et al: Continuous hemofiltration with polyethersulfone mem-branes evaluated by tele-monitoring. Contrib Nephrol. Basel, Karger, 2003, vol 138, pp 126–143.

48 Bellomo R, Balwin I, Ronco C, Golper T: Atlas of Hemofiltration. London, Saunders, 2002.49 Uchino S, Cole L, Morimatsu H, et al: Solute mass balance during isovolaemic high volume

haemofiltration. Intens Care Med 2003;29:1541–1546. Epub 2003 Jul 10.50 Morgera S, Haase M, Rocktaschel J, et al: High permeability hemofiltration improves peripheral

blood mononuclear cell proliferation in septic patients with acute renal failure. Nephrol DialTransplant 2003;18:2570–2576.

51 Tetta C, D’Intini V, Bellomo R, et al: Extracorporeal treatments in sepsis: Are there new perspec-tives? Clin Nephrol 2003;60:299–304.

52 Stuber F, Petersen M, Bokelmann F, Shade U: A genomic polymorphism within the tumor necro-sis factor locus influences plasma tumor necrosis factor-alpha concentrations and outcome ofpatients with severe sepsis. Crit Care Med 1996;24:381–384.

53 Stuber F: Another definite candidate gene for genetic predisposition of sepsis: Interleukin-10. CritCare Med 2003;31:314–315.

Ciro Tetta, MDResearch Extracorporeal Therapies, Fresenius Medical Care Deutschland GmbH Else Kroener Strasse 1, DE–61352 Bad Homburg (Germany)Tel. �49 6172 609 2417, Fax �49 6172 609 2314, E-Mail [email protected]


Coupled Plasma Filtration Adsorption: Rationale,Technical Development and Early Clinical Experience

Alessandra Brendolana, Claudio Roncoa, Zaccaria Riccia, Valeria Bordonia, Monica Bonelloa, Vincenzo D’Intinia, Mary Lou Wrattenb, Rinaldo Bellomoc

aDepartment of Nephrology, St. Bortolo Hospital, Vicenza, and bScientific Research Department, Bellco SpA, Mirandola (MO), Italy; Department of Intensive Care, Austin and Repatriation Medical Centre, Melbourne, Australia

Treatment Rationale: Learning from Past Lessons

‘only by rethinking the assumptions underlying the failed treatments for sepsiscan we hope to make strides in saving patients from sepsis induced death’

R. Bone 1996 [1]

We still lack a good definition of sepsis. There have been many attemptsto better define clinical symptoms, categories and scoring systems, however astime passes the definitions always seem to grow more complex in spectrum.

Early attempts at treatment met with failure due to over-simplistic approachesto treatment. It was thought that, if an infectious agent was responsible forsepsis, early intervention by neutralizing the infectious agent or a component ofthe early phases of the inflammatory cascade would result in prevention of thesecondary symptoms. Often, in animal models that utilized targets towardsspecific mediators or endotoxin, such approaches had some success, due to thecontrolled nature of the experiment. These ‘magic bullets’ often stirred excitementwithin the medical community, only to be met with great disappointment aftertheir failure during human clinical trials.

Other attempts to define sepsis looked at more systemic approach to thepathology. Pinsky [2] suggested that sepsis was a ‘malignant form of intravascular

Coupled Plasma Filtration Adsorption 377

inflammation’. Others have suggested that it may be more of ‘a type of dis-seminated local inflammation in which the actual process occurs at the tissuelevel’ [3]. In this case, the increase in plasma cytokines is thought to originatefrom ‘spillage’ from the local sites into the circulation with consequent dis-seminated inflammation. What seems to be the accepted operating paradigm at thistime is that sepsis is an exaggerated immune response to infection that includesthe production and release of a wide array of both pro- and anti-inflammatorymolecules.

Coupled plasma filtration adsorption (CPFA) is an extracorporeal therapythat utilizes a plasma filter to separate plasma from the blood and then allowspassage of the separated plasma through a sorbent cartridge for the nonspecificremoval of various mediators [4–8]. After purification, the plasma is returnedto the blood. The blood can then pass through a hemodialyzer/hemofilter foradditional blood purification by conventional hemodialysis, hemofiltration orhemodiafiltration in patients who have acute renal failure (fig. 1). The treat-ment goal of CPFA is to target the excess in circulating mediators (both pro-and anti-inflammatory molecules) in order to restore normal immune function.

Plasma

Normalphysiologicalrange

Excessmediators

Resintarget

HemofiltrationHemodialysisHemodiafiltration

Unit for

Blood in

Blood out

Plasmafiltration unit

Fig. 1. Schematic diagram of CPFA. The plasmafilter separates the plasma. Thispasses through the resin cartridge and is returned to the patient after depurification. The resinadsorbs various inflammatory mediators by non-specific adsorption. The target moleculesfor adsorption are all of the excess pro- and anti-inflammatory mediators.

Brendolan/Ronco/Ricci/Bordoni/Bonello/D’Intini/Wratten/Bellomo 378

Figure 2 shows a diagrammatic scheme of inflammatory cell activation andhyporesponsiveness and how these may relate to the amplified response seen inseptic patients.

The development of CPFA was a complex process that underwent numerousevolutionary changes to find a suitable resin with good adsorption for a widevariety of mediators and at the same time suitable for extracorporeal therapy.Historically, the use of resins or sorbents appeared an attractive concept to achieveblood purification. Hemoperfusion was used in animals as early as 1948 to testthe removal of urea with an Amberlite ion exchange resin [9]. This was fol-lowed by the use of a lactated anion exchange resin by Schreiner in 1958 for apatient with pentobarbital poisoning and various other attempts in the mid-1960s [10–13]. These early trials were associated with severe side effects whichincluded hemolysis, electrolyte disturbances, pyrogenic reactions and thrombo-cytopenia. Attempts to increase the biocompatibility of such resins by externalcoating of the matrix often led to compromises in adsorption efficacy.

Today there is a renewed interest in the use of sorbent techniques to com-plement other types of extracorporeal blood purification therapies. There are awide range of commercial adsorption therapies ranging from specific adsorption

Activation of inflammation

Infectious or initiating agent

Chemo- attractants

Pro- inflammatory

cytokines

Anti- inflammatory

cytokines

Immuno- suppression

Activation of signal pathways

Inappropriate monocyte

hypo-responsiveness

Endothelial injury

Margination of neutrophils

Systemic activation of monocytes

Microcirculatory injury

Tissue/organ injury

Fig. 2. Different steps in the evolution of sepsis.


therapies such as a polymyxin B endotoxin adsorbing resin to more generalizedor combination systems for artificial liver support devices [14–18].

How Do Sorbents Work?

The resin used in the CPFA sorbent cartridge was chosen based on itsadsorption capacity for inflammatory mediators, low levels of extractable tox-ins/metals and good pressure flow performance. This section will describe ingreater detail some of the basic characteristics of the sorbent cartridge and themechanisms responsible for the inflammatory mediator removal.

The resin used in CPFA sorbent cartridge is a synthetic cross-linkedstyrenic divinylbenzene resin. This type of resin is used in diverse processesranging from laboratory chromatography to industrial purification of foods,beverages and pharmaceuticals. The resin is well suited for extracorporeal appli-cations because of its high homogeneity, good pressure-flow performance, andexcellent mechanical and chemical stability. In addition to its mechanical prop-erties, the choice of this particular resin was based on several different factorsrelated to its capacity to adsorb a wide variety of inflammatory mediators andits adsorption characteristics during different flow rates.

In this regard, it may be useful to compare the processes that occur duringthe adsorption of mediators with CPFA to similar processes that occur in tradi-tional chromatography (table 1). Chromatography is commonly used to sepa-rate molecules from a complex mixture. There are many factors that caninfluence this process and many of the important factors are summarized intable 2. Traditional column chromatography has two important phases: themobile phase (the fluid or solvent that moves through the column) and thestationary phase (the ‘resin’ or matrix). The active sites of the stationary phaseinteract with the functional groups of the compound to be separated by nonco-valent bonds, nonpolar interactions, Van der Waals forces and hydrophobicinteractions. The less tightly bound compounds are eluted out by the mobilephase at an earlier time allowing different classes of compounds or moleculesto be separated. The mobile phase can be easily changed by using a combina-tion of different solvents. In the case of CPFA, the patient’s plasma is a com-plex milieu that contains a wide variety of excess mediators – as well as lipids,proteins and other physiologically important molecules. The patient’s plasmacan be considered the mobile phase and the treatment goal is to maximize theinteraction of the mediators with the resin so that they will prefer the resinand be retained the cartridge. This is not an easy task as the excess inflamma-tory mediators have a wide range of sizes and different physicochemicalproperties.


Table 1. Comparison between chromatography and CPFA resin adsorption

Chromatography CPFA (resin adsorption)

• Separate and retain pro- and anti-inflammatory molecules that areproduced in excess of normalphysiological levels

• The primary treatment goal is to retainexcess molecules within the resin. Asecondary goal is the eventual restitutionof physiological molecules and theavoidance of large amounts of exogenousreinfusion fluid

• In this case – the mobile phase is actuallythe patient’s plasma

• This creates a unique milieu that variesfrom individual to individual and that canchange over the treatment period

• Separate molecules from a complexmixture

• Molecules of interest leave the column atdifferent times depending on the affinityfor either the resin or the mobile phase

• Eventually all molecules are usuallyeluted from the column/resin

• Works with well-defined mobile phasethat can be modified by the user(polarity, mixed solvents, pH, saltconcentration)

There are several factors that are important in determining the resin’sadsorption efficacy. These include: the type of resin, the size of the resin beadand the internal pores of the resin bead. Resins vary greatly in their chemicalproperties. The resin used for CPFA is a reverse phase type resin that interactswith hydrophobic sites on the molecule. The pore size can be important as poredimensions limit the size of the molecules that can cross them. An importantconcept related to the molecule’s size is the Stokes’ radius (fig. 3). In aqueoussolutions, molecules rotate around their center of gravity – describing a spherethat corresponds to their effective size. The radius of this ‘sphere’ governs themolecules’ access to the matrix pores. The capacity of the resin refers to thequantity of a given molecule that is adsorbed by a specific quantity of resin.

In vitro Experiments

The resin for CPFA was initially tested under static conditions to deter-mine the binding capacity for pro- and anti-inflammatory mediators. Individualmediators were first tested by adding a known quantity of the mediator to salineand observing the adsorption to a small quantity of resin in a test tube. Theadsorption could then be easily measured by monitoring the concentration inthe supernatant. This method allowed a general assessment of adsorption for awide variety of inflammatory mediators and different sorbents. Some of the


various sorbents that were tested included activated carbon, ion exchange resinsand several reverse phase resins with varying bead and pore dimensions.

The next step involved adding the inflammatory mediators to healthy con-trol plasma, incubating this with a small amount of resin and then monitoringthe supernatant. Not surprisingly, the adsorption can change dramatically whenincubated with saline versus plasma. An example of this difference was foundfor the high binding of �2-macroglobulin from plasma. This protein has a highaffinity for several cytokines. In some cases, cytokines that had only moderatebinding to the resin in saline showed dramatically improved binding after theywere added to plasma. It appeared that this was a consequence of the fact that�2-macroglobulin was able to bind to the resin and subsequently the cytokinebecame bound to the �2-macroglobulin. Thus, the resin became an effectiveremover of cytokines in plasma not through direct cytokine binding but throughan intermediary step. A final step for the completion of static measurements

Table 2. Factors affecting adsorption

What influences adsorption/elution?

The Resinparticle sizepore sizeresin typecharge

Columnresin quantitygeometrylinear velocityresin capacitypacking

The Molecule of Interestsizechargehydrophobicitycharge

The Mobile PhasesaltpHpolarity

0

50

100

150

200

250

300

50 100 150 200

Linear velocity

Ad

sorp

tion Fig. 3. Linear velocity is related to the

speed at which fluid – and thus sample com-ponents – moves along a flow path. The figureshows the binding of an inflammatory media-tor to three different resins at various linearvelocities. Some resins have good bindingover a wide range of linear velocities, whereothers have a sharp decrease in binding athigh linear velocities.


included incubating whole blood with endotoxin to produce a wide array ofinflammatory mediators, isolating the plasma and then incubating the plasmawith the resin in the test tube. All of these different static tests helped chooseresins that have a good capacity for the molecules of interest.

After several resins were chosen for their adsorptive capacity, they werethen evaluated for safety (leachable toxins or metals), ability to be sterilized andmechanical stability. The optimization of flow and column geometry is aparameter that also greatly influences adsorption efficacy. The linear velocity isa term that describes the relationship between fluid speed (and thus the samplecomponents) and a flow path. It is typically reported as cm/hr. Since the linearvelocity is related to the square of the column radius- it is directly influencedby sample flux through the column and the column radius.

Some resins have good adsorption of a particular molecule over awide range of linear velocities, while others may have a dramatic decrease inadsorption as the linear velocity increases (fig. 4). This translates into an impor-tant point related to plasma adsorption: there is a balance between the volumeof plasma being treated and the plasma contact time with the resin. Decreasedadsorption efficacy is often observed if plasma flow rate becomes too fast.

Figure 2 shows the influence of linear velocity on sample adsorption forthree different resins for an inflammatory mediator. In some cases the adsorp-tion is quite high at low linear velocities, but the adsorption is greatly dimin-ished at higher linear velocities. It is also possible to see the relationship ofcolumn diameter by comparing the diameters of a ‘thin’ column of 4 cm to a‘wide’ column of 8 cm. If there is a flux of 50 ml/min, this will give a linearvelocity of 238 cm/h for the ‘thin’ column and 60 cm/h for the ‘wide’ column.A small change in column diameter can greatly influence the linear velocity andadsorption, even if the flow rates are similar.

The next series of experiments evaluated resin adsorption under flow con-ditions in vitro. For these experiments small cartridges were constructed to

Linear velocityflux cm /

area cm�

3

2

h

Fig. 4. Stokes’ radius of proteins withsimilar molecular weights but differentstokes’ radii.

Linear structure Globular structure


enable testing of various resins under the same linear velocity as those expectedduring plasma filtration (about 20 ml/min). The parameters that were importantwith these experiments were overall adsorption of mediators in both plasma andsaline, good flow performance without any increases in pressure and theabsence of fines or small particles that could clog the external safety filter.After careful consideration of the static and dynamic in vitro results, the resinthat gave the best overall performance was a 100-�m reverse-phase resin withan average pore diameter of 30 nm.

Animal Studies

A rabbit model for endotoxic shock was used to determine whether the useof coupled plasma filtration and adsorption could reduce 72-hour mortality [19].Rabbits were subdivided into groups to receive endotoxin plus CPFA, endotoxinwith plasma filtration (no resin), CPFA only (no endotoxin) or CPFA plus endo-toxin. The rabbits were anesthetized, cannulated and underwent 3 hours of treat-ment. Plasma concentrations of endotoxin, bioactive tumor necrosis factor,resin-adsorbed platelet-activating factor, mean arterial pressure, base excess, andwhite cell count were assessed and a global severity score was established.

At 72 h, cumulative survival was significantly (p � 0.0041) improved inseptic rabbits treated with coupled plasma filtration-adsorption, and cumulativesurvival of the resin with LPS group was not significantly different (p � 0.05)from that of the control groups (plasma filtration or CPFA without any endotoxin).

One of the more surprising results of the study was that survival was notdirectly correlated with any single parameter (endotoxin, cytokine, MAP, etc.).Only the global severity score based on severity of several different parameterswas inversely correlated with survival. This supports the current trend to a moreglobal treatment of sepsis rather than restricting treatment to specific removalstrategies.

Human Clinical Studies

There have been two recent studies using CPFA for septic patients. Thefirst study by Ronco et al. [20] was a pilot cross-over study that comparedCPFA to hemodiafiltration by measuring hemodynamics and immuneresponsiveness in acute renal failure patients with septic shock. In this study,patients were randomly assigned to10 h of treatment using CPFA followed by10 h of hemodiafiltration (or vice versa). These authors observed a signifi-cant improvement in hemodynamics with the use of CPFA compared to


hemodiafiltration. They also observed a significant increase in leukocyteresponsiveness after CPFA treatment. For these experiments they monitoredspontaneous and endotoxin stimulated leukocyte TNF� production after 10 h oftreatment. At the beginning of the treatment there was a marked leukocytehyporesponsiveness to endotoxin stimulation (immunosuppression). As thetreatment progressed the responsiveness increased. Further support for the roleof CPFA in the restoration of immune responsiveness was observed by incu-bating pre- and post-resin plasma with monocytes obtained from healthydonors. The pre-resin plasma at the beginning of treatment had a strongimmunosuppressive effect – unless the plasma was first incubated with mono-clonal antibodies to IL10. In contrast, the post-resin plasma (at the beginning oftreatment) produced higher quantities of TNF� after endotoxin challenge, andnearly normal quantities after the 10-hour treatment.

One of the interesting observations of this study was that there were no sig-nificant changes in circulating plasma levels of IL10 or TNF�, even thoughthere was almost complete adsorption of these cytokines by the resin cartridge.This suggests that there may still be other factors that are adsorbed by the cartridgethat play a role in immunosuppression. For this reason, the results presented inthis study may be particularly relevant as the end point of the study was restora-tion of immune responsiveness, rather than a net increase or decrease in specificinflammatory mediators.

A second clinical study using CPFA was conducted by Formica et al. [21].This study examined the effect of repeated applications of CPFA on the hemo-dynamic response in septic shock patients. One of the unique aspects of this studywas that both ARF and non-ARF patients were eligible if they met the enrollmentcriteria. All patients had a high APACHE II score (24.8 � 5.6) and multi-organfailure. Six of the 10 patients had normal renal function. The authors performed10 consecutive sessions and observed a net decrease in vasopressor requirement,increased mean arterial pressure and decreased C reactive protein. The patientstreated with CPFA had a 70% survival.

Although the sample size was small for both clinical studies, they bothshowed improved hemodynamics and good treatment tolerance. Further studiesare needed to better identify patient groups that may benefit from early inter-vention in the complex interactions between various immunosuppressive andinflammatory mediators.

Conclusion

Sepsis is a leading cause of ARF in ICUs, however the use of extra-corporeal blood purification techniques in non-ARF septic patients remains


controversial [3, 22, 23]. Current guidelines, for example, state that ‘hemofil-tration should not be used in patients with sepsis without renal indicationsunless ongoing studies provide positive results’ [24]. The role of plasmaexchange remains equally controversial [25, 26]. In contrast, the initial findingswith CPFA are consistent and provide provocative observations, which will nodoubt encourage clinical investigators to further pursue the path of blood purifi-cation. It may be that CPFA will be the first non-renal blood purification tech-nique to be tested in a phase IIb multicentre randomized controlled trial for thetreatment of severe sepsis and multiorgan failure.

References

1 Bone R: Why sepsis trials fail. JAMA 1996;276:565–566.2 Pinsky M: Multiple systems organ failure: Malignant intravascular inflammation; in Pinsky M,

Matuschak G, (eds): Critical Care Clinics, Multiple Systems Organ Fialure, vol 5, 1989,pp 195–198.

3 Kellum J, Bellomo R: Hemofiltration in sepsis: Where do we go from here? Crit Care 2000;4:69–71.4 Bellomo R, Tetta C, Ronco C: Coupled plasma filtration adsorption. Intensive Care Med 2003;26:26.5 Bellomo R, Tetta C, Brendolan A, Ronco C: Coupled plasma filtration adsorption. Blood Purif

2002;20:289–292.6 Brendolan A, Bellomo R, Tetta C, Piccinni P, Digito A, Wratten ML, Dan M, Irone M, La Greca G,

Inguaggiato P, Ronco C: Coupled plasma filtration adsorption in the treatment of septic shock; inRonco C, Bellomo R, La Greca G, (eds). Blood Purification in Intensive Care. Contrib Nephrol.Basel, Karger, 2001, vol 132, pp 383–390.

7 Tetta C, Cavaillon JM, Schulze M, Ronco C, Ghezzi PM, Camussi G, Serra AM, Curti F,Lonnemann G: Removal of cytokines and activated complement components in an experimentalmodel of continuous plasma filtration coupled with sorbent adsorption. Nephrol Dial Transplantation1998;13:1458–1464.

8 Tetta C, Cavaillon JM, Camussi G, Lonnemann FG, Brendolan A, Ronco C: Continuous plasmafiltration coupled with sorbents. Kidney Int suppl 1998;66:S186–S189.

9 Muirhead E, Reid A: Resin artificial kidney. J Lab Clin Med 1948;33:841–844.10 Kissack A, Gliedman L, Karlson K: Studies with ion exchange resins. Trans Am Soc Artif Intern

Organs 1962;8:219.11 Yatzidis H: A convenient haemoperfusion micro-apparatus over charcoal for the treatment of

endogenous and exogenous intoxidations. Its use as an artificial kidney. Proc Eur Dial TransplantAssoc 1964;1:83.

12 Chang T: Semipermeable aqueous micocapsules (artificial cells): With emphasis on experimentsin an extracorporeal shunt. Trans Am Soc Artif Organs 1966;12:13.

13 Gordon A, Greenbaum M, Marantz L, McArthur M, Maxwell M: A sorbent based low volumerecirculating dialysate system. Trans Am Soc Artif Intern Organs 1969;15:347–352.

14 Greil J, Wyss PA, Ludwig K, Bonakdar S, Scharf J, Beck JD, Ruder H: Continuous plasma resinperfusion for detoxification of methotrexate. Eur J Pediatr 1997;156:533–536.

15 Takenaka Y: Bilirubin adsorbent column for plasma perfusion. Ther Apher 1998;2:129–133.16 Ryan CJ, Anilkumar T, Ben-Hamida AJ, Khorsandi SE, Aslam M, Pusey CD, Gaylor JD, Courtney

JM: Multisorbent plasma perfusion in fulminant hepatic failure: Effects of duration and frequencyof treatment in rats with grade III hepatic coma. Artif Organs 2001;25:109–118.

17 Sechser A, Osorio J, Freise C, Osorio RW: Artificial liver support devices for fulminant liver failure.Clin Liver Dis 2001;5:415–430.

18 Winchester J, Kellum J, Ronco C, Brady J, Quartaro P, Salsberg J, Levin N: Sorbents in acute renalfailure and the systemic inflammatory response syndrome. Blood Purif 2003;21:79–84.


19 Tetta C, Gianotti L, Cavaillon J, Wratten M, Fini M, Braga M, Bisagni P, Giavaresi G, Bolzani R,Giardino R: Continuous plasma filtration coupled with sorbent adsorption in a rabbit model ofendotoxic shock. Crit Care Med 2000;28:1526–1533.

20 Ronco C, Brendolan A, Lonnemann G, Bellomo R, Piccinni P, Digito A, Dan M, Irone M,La Greca G, Inguaggiato P, Maggiore U, De Nitti C, Wratten ML, Ricci Z, Tetta C: A pilot studyof coupled plasma filtration with adsorption in septic shock. Crit Care Med 2002;30:1250–1255.

21 Formica M, Olivieri C, Livigni S, Cesano G, Vallero A, Maio M, Tetta C: Hemodynamic response tocoupled plasmafiltration adsoprtion in human septic shock. Intensive Care Med 2003;29:703–708.

22 Kellum J, Bellomo R, Mehta R, Ronco C: Blood purification in non-renal critical illness. BloodPurif 2003;21:6–13.

23 Cuhaci B: Plasma exchange in multi organ failure: Changing gears in sepsis and organ failure. Crit Care Med 2003;31:1875–1877.

24 Carlet J: Immunological therpay in sepsis: Guidelines for the managment of severe sepsis andseptic shock. Intensive Care Med 2001;27:S93–S103.

25 Reeves JH, Butt W, Shann F, Layton JE, Stewart A, Waring P, Presneill JJ: Continuous plasmafil-tration in sepsis syndrome. Crit Care Med 1999;27:2096–2104.

26 Busund R, Koukline V, Utrobin U, Nedashkovsky E: Plasmapheresis in severe sepsis and septicshock: A prospective randomized controlled trial. Intensive Care Med 2002;28:1434–1439.

Claudio Ronco, MDDirector, Department of NephrologySt. Bortolo Hospital, Viale Rodolfi, IT–36100 Vicenza (Italy)Tel. �39 0444993869, Fax �39 0444993949, E-Mail [email protected]


Plasmapheresis in Sepsis

Giorgio Berlota, Gabriella Di Capuab, Paola Nosellaa, Sara Rocconia, Corrado Thomanna

aDepartment of Anaesthesia, Intensive Care and Pain Therapy, University of Trieste, Trieste, and bDepartment of Anaesthesia and Intensive Care, University of Naples ‘Federico II’, Naples, Italy

The successful treatment of sepsis is based on the surgical drainage of sep-tic foci, on the administration of appropriate antibiotics and on the correctionof the sepsis-induced cardiovascular, metabolic and respiratory derangements[1]. Thus, it is clear that all these approaches are implemented only once thenetwork of sepsis mediators with either pro- or anti-inflammatory actions havebeen activated [2].

In the recent past, two strategies have been developed to neutralize theactions of the inflammatory mediators in septic patients. The first one takesadvantage of the impressive advances in genetic engineering techniques, whichallowed the production of antagonists of the sepsis mediators, including antibod-ies directed against them and inhibitors of the mediator-cell interaction. However,despite the sound biological basis and the promising experimental results, manymulticenter, randomized, placebo-controlled trials (RCT) repeatedly failed todemonstrate any substantial positive effect on the survival [3, 4]. Only recently,the administration of the human recombinant activated protein C (rAPC) whichis a natural occurring anticoagulant with also a strong anti-inflammatory actionhas been demonstrated to exert a clear, albeit limited, positive action on the sur-vival of septic patients [5]. The second approach is based on the removal of thesepsis mediators from the bloodstream by means of extracorporeal techniques,including those commonly used in the treatment of the acute renal failure (ARF)occurring in critically ill patients and especially in those developing multipleorgan dysfunction syndrome (MODS) [6, 7]. Collectively, these techniques aregrouped under the term continuous renal replacement therapy (CRRT) as they arebasically used to provide renal support on a 24-hour basis in patients who cannottolerate hemodialysis (HD) [7].

Berlot/Di Capua/Nosella/Rocconi/Thomann 388

The rationale for their use with this indication is based on two assump-tions. The first is that either the size and molecular weight (MW) of the medi-ators remain within the cut-off value of the membrane used or, alternatively,that their chemicophysical properties allow their adsorption on the surface ofthe filter [8], and the second is that they are present in the bloodstream in rele-vant concentration when the procedure is running, since both too early or toolate treatments can be ineffective [4, 9]. Unfortunately, both conditions are hardto assess in the clinical settings, since the blood concentration of the sepsismediators are subjected to wide fluctuations [10]; furthermore, they can be con-sidered as the tip of the iceberg poorly reflecting what is going on at the tissuelevel [11]. Several studies have been performed aimed to evaluate either theefficacy of these techniques in the removal of septic mediators and the possibleimpact on the outcome. However, results from clinical investigations werelargely inferior to what was expectable from laboratory experiences, as theblood concentration of sepsis mediators decreased in some studies butremained stable or even increased in others [12]. Several factors can account forthis inconsistency, including the heterogeneity of the patients, the timeframe ofthe treatment, its duration, the different properties of the membrane used andthe volume of fluid exchanged [12, 13]. Then, it appears that the extracorporealremoval of the sepsis mediators can be accomplished via two differentapproaches. The first takes advantage of the characteristics of the membraneused in the CRRT, whose depurative efficacy can be enhanced either by increas-ing the volume of the ultrafiltrate or by maximizing their absorptive capabili-ties [8, 14]. The second strategy is based on the use membranes with largercut-off values commonly used during plasmapheresis (PP) or plasma exchange(PE), which allow the removal of substances whose MW exceeds the cut-off ofthe membranes commonly used in CRRT [15, 16].

Applications of Plasma-Separation Techniques in Sepsis

Description of the Technique and TerminologyThe process of separation of the blood into its main components is based on

two different approaches [15, 16], namely plasma-exchange (PE) and plasma-pheresis (PP). Albeit these two terms are often used as synonymous, the differ-ences between them are profound. The PE is basically a one-step procedure,consisting in the separation of the blood in its components, namely the cells andthe plasma; the former are returned to the patient, and the latter is discarded andsubstituted with other fluids, including plasma expanders, albumin or fresh-frozen plasma (FFP) in order to maintain an adequate volemia and/or to supplythe substances lost. Conversely, the PP is a double-step procedure, aimed at the

Plasmapheresis in Sepsis 389

selective removal of determined component(s) via single or double cascadeprocess, during which the plasma obtained by means of filtration or centrifuga-tion run flows through columns containing different types of beads or othermaterials able to adsorb more or less selectively the target substance(s); theprocessed plasma is subsequently reinfused. Both techniques have been used ina wide number of diseases characterized by the presence in the bloodstream ofsubstances supposed to be responsible for the disorders [15]. However, clinicalresults have often been inconclusive due to different reasons [16], including(a) the role played by the target substance(s) (causative agents vs. simple markerof the disease); (b) the low rate of occurrence of many disorders treated with PEor PP which prevents to collect a sufficient experience to draw definite conclu-sions about their efficacy, and (c) the bizarre clinical course of several diseasetreated with these techniques, which can undergo spontaneous relapses andremissions. For these reasons, the number of diseases in which PE and/or PP arecurrently used is far lower than a few years ago [17]; among them, the treatmentof sepsis could represent a promising field of application.

Laboratory ExperiencesAs stated above, the main theoretical indication for the use of PP or PE in

sepsis consists in the removal of mediators exerting a detrimental action.However, in different timeframes, one class of substances can prevail on theother, leading to a systemic inflammatory response syndrome (SIRS), associ-ated or not with an infection, or, conversely, to a status characterized by theprogressive blunting, till the exhaustion, of this response [10]. This lattercircumstance has been defined as a compensatory anti-inflammatory responsesyndrome (CARS). Although probably an oversimplification, this is a reason-able model to understand the bizarre clinical course of many septic patients inwhom a hypo- or anergic condition often follows, and sometimes concludes, acritical illness initiated with a typical inflammatory process (e.g. pneumonia orperitonitis) [18]. Recently, it has become clear that genetic factors are relevantin up- and downregulating these reactions [19]. Hence, as far as the extracor-poreal removal of sepsis mediators is concerned, the theoretical risk of remov-ing anti-inflammatory substances during the florid phase of sepsis and SIRSshould be taken into account, thus determining the potential harmful imbalancebetween the two classes of substances. Results from experimental studiescarried conflicting results. Busund et al. [20] studied three groups of dogs inseptic shock: the first and the second groups were treated with PE or the infu-sion of plasma (PI) whereas the third served as control. In the groups treatedwith PE and PI, the blood concentrations of endotoxin, TNF and interleukin 1(IL-1) decreased significantly as compared with the controls, and this decreasewas more marked in the PE group. Either the cardiovascular function or the


overall survival was better in both the treatment arms. However, these encour-aging results are in sharp contrast with those deriving from another similarstudy, in which a marked worsening of the hemodynamic variables and areduced survival was observed in the PE group as compared with the controland the sham PE groups, respectively [21]. A possible explanation could derivefrom another study, in which dogs were given i.v. endotoxin and subsequentlytreated with repeated PIs: in many cases an irreversible cardiovascular collapseoccurred, which was ascribed to a sudden decrease of the plasma ionized Ca2�;remarkably, this decrease was not observed in non-endotoxemic animals [22].

Clinical StudiesDespite the conflicting experimental results and the lack of conclusive clin-

ical trials clearly demonstrating a beneficial effect associated with their use, PEand PP, alone or in association with other blood depurative techniques, have beenused in sepsis patients. The available studies suffer from some of the limitationsencountered in those concerning the effects of PE and PI in diseases other thansepsis, namely the heterogeneity of patients and related clinical conditions.

Then, it appears more useful to consider separately the efficacy of PE andPP in terms of: (a) removal of mediators; (b) improvement of some variablescommonly deranged during sepsis and septic shock, and (c) effect on the out-come. As will appear, there is not a clear relationship among these variables.

As far as the first point is concerned, PE has been demonstrated to effectivelyreduce the plasma concentration of both endotoxin and sepsis mediators both inseptic patients [23] and in patients with acute liver failure [24]. Reeves et al. [25]treated septic patients with continuous PE, titrated to remove approximately 5 volof plasma during 30–36 h; the replacement regimen included FFP, colloids andcrystalloids. Some mediators, including C-reactive protein, haptoglobin, C3 frag-ment and �1-antitripsin, were removed with a sieving coefficient approaching 1.0,but other substances, e.g. IL-6, TBX2 and granulocyte-stimulating factors, werenot substantially affected. Although there was no significant decrease in mortality,a trend toward less-severe organ involvement was observed. A different pattern ofremoval of the sepsis mediators has also been demonstrated in other studies: as anexample, in septic patients treated with PE a decrease of the TNF-� concentrationbut not of other cytokines has been observed [26]. Possible mechanisms responsi-ble for the failed decrease of the blood levels of sepsis mediators include a volumeof distribution largely exceeding the intravascular space [11], a pulse production(as an example, following the administration of certain antibiotics with elevatedendotoxin-releasing properties) [27] or a positive feedback loop promoting anelevated production to cope with the increased elimination [2, 3].

Independent of the substances removed, it appears that PE/PP is associatedwith an improvement of several physiological variables. In a group of septic


patients undergoing PE aimed at the removal of a full volume of plasma whichwas replaced with crystalloids and FFP, Berlot et al. [28] observed a significantimprovement of several hemodynamic variables in the absence of any changeof the preload or afterload; interestingly, these changes were more marked inthose patients who, before the procedure, exhibited the lowest hemodynamicvalues. Despite these beneficial effects, the mortality rate exceeded 60%. Anoverall improvement of the cardiovascular variables in association with PE/PPhas also been demonstrated in other studies, involving both adult and pediatricseptic patients [29–30].

Although it is relatively easy to measure circulating septic mediators or toassess the variations of some cardiovascular variable, to draw definite conclu-sions on the effect on the outcome exerted by PE and/or PP is much lessstraightforward. As an example, a survival of 74% has been reported in a reviewof the available outcome data [31]; however, the analysis included a limitednumber of patients (total 76, ranging from 2 to 19 in 10 different studies) whohad been treated with 4 different plasma-processing techniques. More recently,the same authors [32] reported a survival of 82% in a group of septic patientstreated with PE as last option during MODS. These results are clearly encour-aging, but, likewise, other studies involving PE, PP and CRRT are biased by thelack of an objective indicator for its application [33].

Plasma-Processing Techniques

Description of the Technique and TerminologyThe availability of new molecules able to stick on their surface one or more

septic mediators lead to the development of newer extracorporeal techniques,based on the principle of coupled plasma removal and absorption (CPFA).Basically, this procedure consists of a PP in which the removed plasma isreturned to the bloodstream after flowing through a resin cartridge with ele-vated absorptive capabilities. If needed, the CPFA can be coupled with CRRT;in this case, the filter is located distally to the cartridge in order to avoid anexcessive hemoconcentration [34].

Laboratory ExperiencesBasically, the experimental investigations parallel what has been already

described with PE and PP. In a rabbit model of sepsis, Tetta et al. [35] were ableto demonstrate that a 3-hour CPFA significantly increased the survival at 72 hof the treated animals as compared with the control group, without any con-comitant decrease of blood endotoxin and TNF-� levels; however, endotoxin


concentrations were significantly lower in animals still surviving at 72 h.During the treatment time, significant amounts of PAF stuck onto the cartridge.The authors ascribed the failed decrease of the endotoxin either to differentkinetics of this molecule and/or to individual variability to its effects.

Clinical StudiesRonco et al. [36] used intermittent CPFA (each session lasting 10 h) in two

groups of patients with septic shock and ARF who also received HD or CRRTin a random order. Arterial pressure increased in both groups but moremarkedly in those treated with CRRT, in whom there was also a decrease in therequest for vasopressor support. Surprisingly enough, macrophages of patientsin the CPFA-CRRT group presented a time-dependent increased production ofTNF-� when challenged with endotoxin, and the authors ascribed this effect toa restoration of the leukocyte responsiveness induced by the removal of somesepsis mediators. Similar results have also been recently reported by Formicaet al. [37] who observed a marked hemodynamic improvement associated withan improved survival in patients with septic shock treated with CPFA.

Conclusions

The neutralization of sepsis mediators is a therapeutic option which under-went a profound evolutionary process. At the beginning of the 1990s, it wasthought that the cascade of mediators could be interrupted by some agent(s)directed against them and/or their cellular receptors. Several RCTs demonstratedthat this approach was not as effective as expected from the preliminary experi-mental studies. Then, it was supposed that these substances could be removed bymeans of CRRT, and some but not all studies demonstrated that the increase intheir depurative efficiency was associated with an improved survival of sepsis-induced MODS patients. The next step consisted of the use of PE and/or PP, buta firm conclusion on the efficacy of both techniques in this setting is difficult todraw because different procedures, different replacement regimens and differentstrategies (single vs. continuous treatments) have been used. The last innovationis CPFA, whose initial results appear extremely promising, and which can beused in association with the currently used CRTT devices.

References

1 Hotchkiss RS, Karl IE: Medical progress: The pathophysiology and treatment of sepsis. N Engl JMed 2003;348:138–150.


2 Adrie C, Pinsky MR: The inflammatory balance in human sepsis. Intens Care Med 2000;26:364–375.

3 Cain BS, Meldrum DR, Harken AH, McIntyre RC: The physiologic basis for anticytokine clinicaltrials in the treatment of sepsis. J Am Coll Surg 1998;186:337–350.

4 Abraham E: Why immunomodulatory therapies have not worked in sepsis. Intens Care Med 1999;25:556–566.

5 Bernard GR, Vincent JL, Laterre PF, et al: Efficacy and safety of recombinant human activatedprotein C for severe sepsis. N Engl J Med 2001;344:699–709.

6 van Bommel EFH: Should continuous renal replacement therapy be used for ‘non-renal’ indica-tions in critically ill patients with shock? Resuscitation 1997;33:257–270.

7 Bellomo R, Ronco C: Continuous renal replacement therapy in the intensive care unit. Intens CareMed 1999;25:781–789.

8 De Vriese AS, Colardyn FA, Philippè JJ, Vanholder RC, de Sutter JH, Lameire NH: Cytokineremoval during hemofiltration in septic patients. J Am Soc Nephrol 1999;10:846–853.

9 De Vriese AS, Vanholder RC, Pascual M, Lameire NH, Colardyn FA: Can inflammatory cytokinebe removed efficiently by continuous renal replacement therapies? Intens Care Med 1999;25:903–910.

10 van der Poll, van Deventer SJH: Cytokines and anticytokines in the pathogenesis of sepsis. InfectDis Clins N Am 1999;13:413–426.

11 Cavaillon JM, Munoz C, Fitting C, Misset B, Carlet J: Circulating cytokines: The tip of the iceberg?Circ Shock 1992;38:145–152.

12 Gotloib L: Hemofiltration in multiorgan failure syndrome secondary to sepsis: A critical analysisof heterogeneity. Nephron 1996;73:125–130.

13 Schetz M, Ferdinande P, Van Den Berghe Verwaest C, Lauvers P: Removal of proinflammatorycytokines with renal replacement therapy: Sense or non-sense? Intens Care Med 1995;21:169–176.

14 Ronco C, Bellomo R, Hormel P, Brendolan A, Dan M, Piccinni P, La Greca G: Effects of differ-ent doses in continuous veno-venous hemofiltration on outcomes of acute renal failure: Aprospective, randomised trial. Lancet 2000;355:26–30.

15 Madore F, Lazarus JM, Brady HR: Therapeutic plasma exchange in renal diseases. J Am SocNephr 1996;7:367–386.

16 Patten E: Therapeutic plasmapheresis and plasma exchange. Crit Rev Cin Lab Sci 1986;23:147–175.

17 Berlot G, Tomasini A, Silvestri L, Gullo A: Plasmapheresis in the critically ill. Kidney Int 1998;53(suppl 66):178–181.

18 Bone RC: Sir Isaac Newton, sepsis, SIRS and CARS. Crit Care Med 1996;24:1125–1136.19 Villar J, Siminovitch KA: Molecular intensive care medicine. Intens Care Med 1999;25:652–661.20 Busund R, Lindsetmo RO, Rasmussen LT, Rikke O, Revkig OP, Revhaug A: Tumour necrosis fac-

tor and interleukin–1 appearance in experimental gram-negative septic shock: The effect ofplasma exchange with albumin and plasma infusion. Arch Surg 1991;126:591–597.

21 Natanson C, Hoffman WD, Koev LA, et al: Plasma exchange does not improve survival in acanine model of human septic shock. Transfusion 1993;33:243–248.

22 Busund R, Lindsetmo RO, Balteskard L, Revkig OP, Revhaug A: Repeated plasma therapyinduces fatal shock in experimental septicemia. Circ Shock 1993;40:268–275.

23 Stegmayr B: Apheresis of plasma compounds as a therapeutic principle in severe sepsis and mul-tiorgan dysfunction syndrome. Clin Chem Lab Med 1999;37:327–332.

24 Iwai H, Nagaki M, Naito T, Ishiki Y, Murakami N, Sugihara J, Muto Y, Moriwaki H: Removal ofendotoxin and cytokines by plasma exchange in patients with acute hepatic failure. Crit Care Med1998;26:873–876.

25 Reeves JH, Butt WW, Shann F, Layton JE, Stewart A, Waring PM, Presneill JJ, Plasmafiltration inSepsis Study Group: Continuous plasmafiltration in sepsis syndrome. Crit Care Med 1999;27:2096–2104.

26 Gardlund B, Syolin J, Nillson A, Roll M, Wickerts CJ, Wretlind B: Plasma levels of cytokines inprimary septic shock in humans: Correlation with disease severity. J Infect Dis 1995;172:296–301.


27 Mock CN, Jurkovich GJ, Dries DJ, Maier RV: Clinical significance of antibiotic endotoxin-releasingproperties in trauma patients. Arch Surg 1995;130:1234–1241.

28 Berlot G, Gullo A, Fasiolo S, Serra S, Silvestri S: Hemodynamic effects of plasma exchange inseptic patients: Preliminary report. Blood Purif 1997;15:45–53.

29 Reeves JH, Butt WW: Blood filtration in children with severe sepsis: Safe adjunctive therapy.Intens Care Med 1995;21:500–504.

30 Mok Q, Butt WW: The outcome of children admitted to intensive care with meningococcal septi-caemia. Intens Care Med 1996;22:259–263.

31 Stegmayr BG: Plasmapheresis in severe sepsis or septic shock. Blood Purification 1996;14:94–101.

32 Stegmayr BG, Banga R, Berggren L, Norda R, Rydvall A, Vikerfors T: Plasma exchange as res-cue therapy in multiple organ failure including acute renal failure. Crit Care Med 2003;31:1730–1736.

33 Rogiers P: Hemofiltration treatment for sepsis. Is it time for controlled trials? Kidney Int 1999;56(suppl 72):S99–S103.

34 Bellomo R, Tetta C, Ronco C: Couplet plasma filtration adsorption. Intens Care Med 2003;29:1222–1228.

35 Tetta C, Gianotti L, Cavaillon JM, et al: Coupled plasma filtration-absorption in a rabbit model ofendotoxic shock. Crit Care Med 2003;28:1526–1533.

36 Ronco C, Brendolan A, Lonnemann G, et al: A pilot study of coupled plasma filtration withadsorption in septic shock. Crit Care Med 2002;31:1250–1255.

37 Formica M, Olivieri C, Livigni S, et al: Hemodynamic response to coupled plasmafiltration-adsorption in human septic shock. Intens Care Med 2003;29:703–708.

Prof. Giorgio BerlotDepartment of Anaesthesia and Intensive CareCattinara Hospital, Strada di Fiume 449, IT–34100 Trieste (Italy)E-Mail [email protected]

395

Adib-Conquy, M. 76Amigues, L. 291Ash, S.R. 239Ayus, J.C. 132

Balakrishnan, V.S. 63Baldwin, I. 105, 203Barsoum, R.S. 44Bellomo, R. IX, 105, 158,

203, 329, 362, 376Béraud, J.-J. 291Berlot, G. 387Bonello, M. 158, 191,

329Bonventre, J.V. 19Bordoni, V. 158Brendolan, A. IX, 376

Callegarin, L. 182Canaud, B. 291Carraro, R. 12Cavaillon, J.-M. 76Chien, C.-C. 53Clark, W.R. 264

Dan, M. 182Davenport, A. 228, 317Di Capua, G. 387D’Intini, V. 158, 191, 329,

376

Fitting, C. 76Formet, C. 291

Goldstein, S.L. 284Golper, T.A. 278

Hoste, E.A. 1, 255

Inguaggiato, P. 158

Jaber, B.L. 63

Kellum, J.A. 1, 362King, L.S. 53Klouche, K. 291Kox, W.J. 308

Lameire, N. 255Leblanc, M. 222Leray-Moragues, H. 291Liangos, O. 63

Moritz, M.L. 132

Naka, T. 105Nosella, P. 387

Palevsky, P.M. 214Passlick-Deetjen, J. 362Pereira, B.J.G. 63Piccinni, P. 12

Pinsky, M.R. 31, 94Pisitkun, T. 329Pohlmeier, R. 362Poulin, S. 329

Rabb, H. 53Ratanarat, R. 158Raynal, N. 291Ricci, Z. 12, 329, 362, 376Rocconi, S. 387Ronco, C. IX, 158, 182,

191, 264, 329, 362, 376Rossi, S. 182Rotonorot, R. 158

Salvatori, G. 158, 191, 329Schetz, M. 119, 166Spies, C. 308

Tetta, C. 362Thomann, C. 387Tiranathanagul, K. 329

Van Biesen, W. 255Van den Berghe, G. 119Vanholder, R. 255Vargas Hein, O. 308Vincent, J.-L. 350

Wratten, M.L. 376

Author Index

396

Subject Index

N-acetyl-�-D-glucosaminidase (NAG),acute renal failure marker 65, 72

Acid-base balance, see FluidsAcidosis, acute renal failure patients,

prognosis 3, 4Acute lung injury, see LungAdequacy, renal replacement therapy

continuous venovenous hemodiafiltration342–344

continuous venovenous hemodialysis340–342

continuous venovenous hemofiltration337–340

definition 329, 330dose correlation with outcome 330,

331dose measurement, different treatments

331, 332efficacy 333–335efficiency 332, 333frequency 333intensity 333intermittent hemodialysis 343,

345–347slow continuous ultrafiltration 335–337sustained low-efficiency dialysis 343,

345–347Anemia, acute renal failure patients,

prognosis 5Antibiotics, dosing, acute renal failure 6Anticoagulation, see Continuous renal

replacement therapy

Apoptosisacute renal failure mechanisms 159pathways 15

Argatroban, anticoagulation, extracorporeal circuits 232, 233

Association studiescandidate gene identification 66, 67cytokine polymorphisms, acute renal

failure 67, 68limitations 70

B cells, sepsis response 78Benzodiazepines, dosing, acute renal

failure 6, 7Blood transfusion, sepsis management

356

Canceracute renal failure association 158hyperuricemia, see Hyperuricemia

Cardiac surgery, acute renal failure risks 13

Cardiogenic shock, see ShockCatheters

care and maintenance 201, 298, 299complications

delayed complications 193, 195immediate complications 193–195infections 197–199, 301, 302prevention guidelines, intensive care

units 303–305stenosis, host vein 302

Subject Index 397

thrombosisfibrin sheath thrombus 196, 197,

300incidence 196prevention 197risk factors 196, 301treatment 197

double-lumen hemodialysis cathetermaterials 192

incidence, catheterization 191performance factors 299, 300performances and recirculation 199–201peritoneal dialysis 248, 249positioning 193, 298stiffness 292surface treatments 302temporary catheter ideal characteristics

191types 292, 293vascular access, see Vascular access

CD11b expression, sepsis 41CD14, inflammation signaling 35Central diabetes insipidus (CDI),

features 150Central venous catheters, see CathetersCerebral demyelination

features 145risk factors, hyponatremic patients 145

Cerebral salt wasting (CSW)diagnosis 142, 143management 143pathogenesis 143

Citrate, anticoagulation, extracorporealcircuits 233, 234, 311, 313, 314, 324, 325

Compensatory anti-inflammatory responsesyndrome (CARS)monocyte hyporesponsiveness

mechanismsdesensitizing agents, plasma 82, 83endotoxin-neutralizing molecules,

plasma 83nuclear factor-�B expression

84, 85toll-like receptor expression and

signaling pathwaydownregulation 83–87

overview 79, 80, 363reversal 87

overview 76Complement, ischemic acute renal

failure pathophysiology 22Continuous renal replacement therapy

(CRRT)adequacy calculation

continuous venovenoushemodiafiltration 342–344

continuous venovenous hemodialysis 340–342

continuous venovenous hemofiltration 337–340

anticoagulationargatroban 232, 233citrate 233, 234, 311, 313, 314,

324, 325comparison of agents 312danaparoid 231, 232, 310hirudin 232, 310, 311low-molecular-weight heparin 230,

231, 309nafamostat mesilate 235prostanoids 234, 235, 311unfractionated heparin 229, 230,

308, 309blood flow monitoring

circuit failure detection 211Doppler flow 206, 207, 210, 211ultrasound flow probe 210

catheters, see Catheters; Vascular accesscongestive heart failure patients 274cost analysis 278–280extracorporeal circuits, see

Extracorporeal circuitsfluids

buffers 223, 224, 322–324composition 223electrolytes 224, 225, 319–322glucose 225prescription mode influences 226, 227sterility 317, 318warming 325, 326

high-volume hemofiltration 273, 274hybrid renal replacement therapy, see

Slow low-efficiency dialysis

Subject Index 398

ideal characteristics 264, 265patient selection

acute renal failure 217–219, 272, 273non-renal indications 219

rationale 387, 388principles 243technical aspects 271, 272techniques

continuous venovenoushemodiafiltration 270, 271

continuous venovenous hemodialysis269, 270

continuous venovenous hemofiltration268, 269

high-volume hemofiltration, seeHigh-volume hemofiltration

slow continuous ultrafiltration 268transport mechanisms, solutes and water

convection 266, 267diffusion 265, 266

Corticosteroids, sepsis management 359Coupled plasma filtration adsorption (CPFA)

animal studies 383, 391, 392chromatography comparison 379, 380clinical trials, sepsis 383, 384, 392development 378principles 377, 378, 391prospects 384, 385, 392sorbents

mechanisms 379, 380testing 380–383

Danaparoid, anticoagulation, extracorporealcircuits 231, 232, 310

Delayed-type hypersensitivity, sepsisresponse 79

Distributive shock, see ShockDiuretic therapy

hyponatremia induction, thiazidediuretics 141

loop diuretics, acute renal failureabsence of protective effects 172–174,

176animal model studies 166, 167clinical trials 167–170

pathophysiology 170, 171rationale 172toxicity 174, 175

Dobutamine, sepsis management 357, 358Dopamine, dosing, acute renal failure 6Dopexamine, sepsis management 357, 358Double-lumen hemodialysis catheters, see

CathetersDrotrecogin alpha, sepsis management 358Drug dosing, acute renal failure

patients 5–7

Endothelial cells, ischemic acute renalfailure pathophysiology 20, 21

Enoximone, sepsis management 358Epinephrine, dosing, acute renal failure 6Extracorporeal circuits

blood flow assessment 206, 207,209–211

blood pumping 203–205design, continuous renal replacement

therapy 235, 236pumps 206, 207roller pumps, segment tubing 205, 206

Femoral vein, catheterization 294, 295Fluids

acid-base balance effectsacidosis consequences 108–110, 222,

223clinical consequences 111colloid versus crystalloid solutions

111, 112conventional view 107, 108‘strong ion difference’ calculation

106, 107, 222buffers

acetate 113bicarbonate 113citrate 113, 114high-volume hemofiltration 114–116lactate 113

colloid solutions 111continuous renal replacement therapy

buffers 223, 224, 322–324composition of fluids 223electrolytes 224, 225, 319–322

Continuous renal replacement therapy(CRRT) (continued)

glucose 225prescription mode influences 226, 227sterility 317, 318warming 325, 326

crystalloid solutions 106sepsis management 354–356shock resuscitation 98–100

Gene polymorphismscytokines, acute renal failure 67–71sites 66types 66

Glomerular filtration rate (GFR), acuterenal failure pathophysiology 170, 171

Glucose control, see Hyperglycemia

Heat shock proteins (HSPs)abundance, cells 38activation mechanism 38immunosuppression, sepsis 83mitochondrial oxidative stress

protection 39nuclear factor-�B interactions 39thermal preconditioning and survival 38

Heparin, see Low-molecular-weightheparin; Unfractionated heparin

High-volume hemofiltration (HVHF)fluid composition 114–116inflammatory cytokine removal 364, 365sepsis management

animal studies 365, 366clinical trials 366–368monitors 370, 371prospects 372rationale 364, 365technique 369, 371, 372timing 368, 369

Hirudin, anticoagulation, extracorporealcircuits 232, 310, 311

HLA-DR expression, sepsis 40, 77, 78Hybrid renal replacement therapy, see Slow

low-efficiency dialysisHyperglycemia

adverse effects 120, 121hypercoagulability 126, 127immune system impairment, infection

risk 126

induction, critical illness 119, 120insulin therapy

beneficial mechanisms 123–125dosing, intensive care units 121–123precautions 127, 128

neuropathy, association, critical illness125, 126

renal failure, association, critical illness 125

sepsis management 359Hyperkalemia, acute renal failure patients,

prognosis 5Hypernatremia

central diabetes insipidus 150clinical manifestations 147, 148definition 146diagnosis 146, 147dialysis therapy

continuous venovenous hemofiltration152

hemodialysis 152peritoneal dialysis 151, 152

edema, association 150, 151mortality 148, 149pathogenesis 146, 147treatment 149, 150

Hyperuricemiaacute renal failure mechanisms 159–161allopurinol therapy 161, 162pathogenesis, cancer 159rasburicase therapy 161, 162, 164

Hyponatremiaacute renal failure patients, prognosis 4, 5cerebral demyelination 145cerebral salt wasting 142, 143definition 133diagnosis 134–136disorders, impaired renal water excretion

133encephalopathy

brain cell volume regulation 138neurogenic pulmonary edema 138risk factors

age 138, 139hypoxia 139, 140sex 139

symptoms 137

Subject Index 399

incidence, hospital patients 136outpatients 140, 141pathogenesis 133prevention 136, 137syndrome of inappropriate antidiuretic

hormone 141, 142thiazide diuretic induction 141treatment 143, 144

Hypovolemic shock, see Shock

Incidence, acute renal failure, critically ill patients 1, 12

Infants, acute renal failure 287, 288Infection

acute renal failure patients, prognosis4, 13

tropical acute renal failure injurymechanismsdirect parenchymal invasion 45, 46hemodynamic disturbances 47, 48iatrogenic injury 48immune system mediated lesions

46, 47Inflammation

compartmentalization concept80–82

critical illness 126, 127intracellular events 34–36ischemic acute renal failure

pathophysiology 20–22, 27mechanical ventilation response 57mitochondria, role 39sepsis, pro- and anti-inflammatory

activities 39–41tubule contribution to inflammatory

injury 23Insulin therapy, see HyperglycemiaIntegrins, expression, sepsis 41Interferon gamma (IFN-�)

ischemic acute renal failurepathophysiology 21

sepsis response 79Interleukin-1 (IL-1)


sepsis response 79, 80

Interleukin-6 (IL-6)expression, sepsis 40, 41infection response 34prognostic marker, acute renal failure 65

Interleukin-10 (IL-10)gene

locus 69polymorphisms, acute renal failure 69

inducers, acute renal failure management 71

prognostic marker, acute renal failure 65sepsis response 79

Intermittent hemodialysisadequacy calculation 343, 345–347biocompatibility of membranes 260, 261continuous renal replacement therapy,

comparison studies, acute renal failure255–257, 261

cost analysis 278–280dosing, acute renal failure 257–260patient selection, acute extracorporeal

renal support 214–217Internal jugular vein, catheterization

296–298Ischemia

pathophysiology, acute renal failure 16,17, 19–28

preconditioning protection against renalinjury 23–25

Kidney injury molecule 1 (KIM-1), acutetubular damage marker 27, 65, 72

Linkage analysis, candidate geneidentification 66, 67

Lipopolysaccharide (LPS), binding proteinlevels, sepsis 83

Low-molecular-weight heparin (LMWH),anticoagulation, extracorporeal circuits230, 231, 309

Lungacute lung injury

animal model studiesacute lung injury effects on kidney

57, 58acute renal failure effects on lung

54, 55mortality, acute renal failure 53

Subject Index 400

Hyponatremia (continued)

acute renal failure effectscytokines 56edema 53, 54ion channel expression 55macrophage activation 55neutrophil activity 55, 56vascular permeability 53–55

kidney cross talk, ischemic acute renalfailure pathophysiology 25

mechanical ventilation effects on kidney 56–58

Mechanical ventilationeffects on kidney 56–58sepsis management 353, 354

Mitochondriacytotoxic mechanisms 39inflammation interactions 39

Mitogen-activated protein kinases(MAPKs), ischemic preconditioningresponse 23–25

Monocyte hyporesponsiveness, seeCompensatory anti-inflammatoryresponse syndrome

Mortalityacute renal failure, critically ill patients

1, 12, 13, 53, 214dialysis effects on acute renal failure

outcome statistics 255, 362hypernatremia 148, 149

Multiple organ failure, see Sepsis

Nafamostat maleate, anticoagulation,extracorporeal circuits 235

Natural killer cell, sepsis activity 78Necrosis, acute renal failure mechanisms

159Neutrophil, sepsis function 78, 79Nitric oxide (NO)

ischemic acute renal failurepathophysiology 20, 25, 26

synthase responses, ischemicpreonditioning 23, 24

Non-cardiogenic obstructive shock, seeShock

Norepinephrine, dosing, acute renal failure 6

Nuclear factor-�B (NF-�B)activation 36, 37endotoxin tolerance induced dysfunction

37heat shock protein interactions 39inflammation, role 35inhibitory subunit 35–37sepsis response 41, 42, 84, 85target genes 35, 36

Opioids, dosing, acute renal failure 6, 7Oxidative stress, see Reactive oxygen

species

Pediatric acute renal failurecongenital heart disease patients 288critically ill patients 285–287epidemiology 284, 285infants 287, 288prospects for study 288, 289

Peritoneal dialysis (PD)acute renal failure management

acute therapy 248catheters 248, 249complications compared with

hemodialysis 250–252continuous-flow peritoneal dialysis

245–248mortality versus hemodialysis

240–242peritonitis incidence 249popularity 239, 240renal function recovery 243tidal peritoneal dialysis 245, 247

clearance versus other dialysis modes244, 245

peritoneum advantages, dialysismembrane 239

protein loss 248renal perfusion maintenance 242, 243

Plasma exchangeanimal studies 389, 390clinical trials, sepsis 390, 391principles 388, 389

Plasmapheresis, see also Coupled plasmafiltration adsorptionanimal studies 389, 390

Subject Index 401

clinical trials, sepsis 390, 391principles 388, 389

Poly(ADP ribose) synthetase, ischemicacute renal failure pathophysiology 22

Prostaglandins, anticoagulation,extracorporeal circuits 234, 235, 311

Protein C, sepsis management 387

Rasburicase, hyperuricemia management161, 162, 164

Reactive oxygen species (ROS)ischemic acute renal failure

pathophysiology 22mitochondria generation 39

Renal blood flow (RBF)acute renal failure pathophysiology 170,

171regulation 183

Risk factors, acute renal failure, intensivecare units 12–15

Sepsiscirculatory dysfunction 183–187compartmentalization, inflammation

80–82cytokine and chemokine responses 34,

39–42, 79, 80, 362definition 376, 377delayed-type hypersensitivity

response 79diagnosis using PIRO system 350–352HLA-DR expression 40, 77, 78lymphocytes

proliferation 76subsets 77

managementblood transfusion 356corticosteroids 359coupled plasma filtration adsorption,

see Coupled plasma filtrationadsorption

fluids 354–356glucose control 359hemodynamic stabilization 353high-volume hemofiltration

animal studies 365, 366

clinical trials 366–368monitors 370, 371prospects 372rationale 364, 365technique 369, 371, 372timing 368, 369

immunomodulation 358, 359infection control 352inotropic agents 357, 358plasma exchange, see Plasma

exchangeplasmapheresis, see Plasmapheresisvasopressors 187–189, 357ventilation 353, 354

monocyte hyporesponsivenessmechanisms

desensitizing agents, plasma 82, 83endotoxin-neutralizing molecules,

plasma 83nuclear factor-�B expression 84, 85toll-like receptor expression and

signaling pathwaydownregulation 83–87

overview 79, 80, 363reversal 87

natural killer cell activity 78neutrophil function 78, 79pro- and anti-inflammatory activities

39–41septic shock definition 362therapeutic targets 363

Shockcardiogenic shock 95definition 182distributive shock 95hypovolemic shock 94, 95non-cardiogenic obstructive shock 95, 96resuscitation

cardiovascular resuscitation targets 101fluid resuscitation 98–100goals 96, 97hematocrit, goals 97, 98periods 96vasopressors 100, 101

septic shock, see Sepsisvasopressor therapy, septic shock

187–189

Subject Index 402

Plasmapheresis (continued)

Slow continuous ultrafiltration (SCUF)adequacy calculation 335–337technique 268

Slow low-efficiency dialysis (SLED)adequacy calculation 343, 345–347clearance rates 280, 281cost analysis 278–280dose versatility 280, 281nursing time 281prospects 281, 282

Sodium balance, see Hypernatremia;Hyponatremia

Stress hyperglycemia, see HyperglycemiaSubclavian vein, catheterization 295, 296Syndrome of inappropriate antidiuretic

hormone (SIADH)causes 141, 142management 142

Systemic inflammatory response syndrome(SIRS), see also Sepsiscytokine response, acute renal failure

34, 63, 64definition 31modifiers of immune response 76, 77

T cellsacute renal failure-acute lung injury

association 56ischemic acute renal failure

pathophysiology 21, 22Toll-like receptors (TLRs)

inflammation signaling 32sepsis, expression and signaling pathway

downregulation 83–87types 32, 83

Transforming growth factor beta (TGF-�),immunosuppression, sepsis 82

Tropical acute renal failureclinical considerations 50epidemiology 44infection and injury mechanisms

direct parenchymal invasion 45, 46hemodynamic disturbances 47, 48iatrogenic injury 48immune system mediated lesions

46, 47

management 50, 51toxin sources, etiology

animal toxins 48, 49industrial chemical toxins 49, 50plant toxins 49

Tubulesinflammatory injury contribution,


lesions, ischemic acute renal failure 16Tumor necrosis factor alpha (TNF-�)

genelocus 67polymorphisms, acute renal failure

67, 68infection response 33, 34inhibitors, acute renal failure

management 70, 71ischemic acute renal failure

pathophysiology 20, 21, 65systemic inflammatory response

syndrome, role 63, 64Tumor necrosis factor beta (TNF-�)

genelocus 67polymorphisms, acute renal failure

67, 68

Unfractionated heparin, anticoagulation,extracorporeal circuits 229, 230, 308, 309

Urate oxidase, see RasburicaseUric acid, see Hyperuricemia

Vascular access (VA), see also Catheterscontinuous renal replacement therapy

access sitesfemoral vein 294, 295internal jugular vein 296–298overview 207–209, 294subclavian vein 295, 296

insertion techniques 192, 193, 294positioning, catheters 193, 298

Vasopressorsseptic shock management 187–189, 357shock resuscitation 100, 101

Volume overload, acute renal failurepatients, prognosis 2, 3

Subject Index 403

tF_ipopf

Coupled Plasma Filtration Adsorption: Rationale, Technical Development and Early Clinical Experience

Documents

Transcript of Coupled Plasma Filtration Adsorption: Rationale, Technical Development and Early Clinical Experience