HP

MPP*BU

BrtnahtMw4c3wpsoc(mwvldsthmenfrsbe

Httw

GASTROENTEROLOGY 2006;130:1135–1143

yponatremia Impairs Early Posttransplantation Outcome inatients With Cirrhosis Undergoing Liver Transplantation

ARIA–CARLOTA LONDOÑO,*,‡ MÓNICA GUEVARA,*,‡ ANTONI RIMOLA,*,‡ MIQUEL NAVASA,*,‡

ILAR TAURÀ,‡,§ ANTONI MAS,*,‡ JUAN-CARLOS GARCÍA–VALDECASAS,‡,� VICENTE ARROYO,*,‡ andERE GINÈS*,‡

Liver Unit, Hospital Clínic, Universitat de Barcelona, Barcelona; ‡Institut d’Investigacions Biomèdiques August Pi-Sunyer (IDIBAPS),arcelona; §Department of Anesthesiology, Hospital Clínic, Universitat de Barcelona, Barcelona; and �Department of Surgery, Hospital Clínic,

niversitat de Barcelona, Barcelona, Spain

rlrttpthtsdlnr

stpsptdvoedheopstt

H

ackground & Aims: Hyponatremia is associated witheduced survival in patients with cirrhosis awaiting liverransplantation. However, it is not known whether hypo-atremia also represents a risk factor of poor outcomefter transplantation. We aimed to assess the effects ofyponatremia at the time of transplantation on post-ransplantation outcome in patients with cirrhosis.ethods: Two-hundred forty-one consecutive patientsith cirrhosis submitted to liver transplantation during a-year period (January 2000–December 2003) were in-luded in the study. The main end point was survival atmonths after transplantation. Secondary end points

ere complications within the first month after trans-lantation. Results: Patients with hyponatremia (serumodium lower than 130 mEq/L) had a greater incidencef neurologic disorders, renal failure, and infectiousomplications than patients without hyponatremiaodds ratio; 4.6, 3.4 and 2.7, respectively) within the first

onth after transplantation. By contrast, hyponatremiaas not associated with an increased incidence of se-ere intra-abdominal bleeding, acute rejection, or vascu-ar and biliary complications. Hyponatremia was an in-ependent predictive factor of early posttransplantationurvival. Three-month survival of patients with hypona-remia was 84% compared with 95% of patients withoutyponatremia (P < .05). Survival was similar after 3onths. Conclusions: In patients with cirrhosis, the pres-

nce of hyponatremia is associated with a high rate ofeurologic disorders, infectious complications, and renalailure during the first month after transplantation andeduced 3-month survival. In cirrhosis, hyponatremiahould be considered not only a risk factor of deathefore transplantation but also a risk factor of impairedarly posttransplantation outcome.

yponatremia is a common feature of patients withadvanced cirrhosis.1 The development of hypona-

remia is related to an impairment of the renal capacityo excrete solute-free water, which causes retention of

ater in an amount disproportionate to that of sodium

etained. This results in a decrease in serum sodiumevels despite the existence of increased renal sodiumeabsorption and high total body sodium content.1,2 Forhis reason, this condition is usually referred to as dilu-ional hyponatremia. Several mechanisms are known toarticipate in the impairment of solute-free water excre-ion in cirrhosis and the subsequent development ofyponatremia, including a reduced delivery of filtrate tohe distal nephron, impaired renal prostaglandin synthe-is, and hypersecretion of arginine vasopressin (the anti-iuretic hormone). Among them, the increased plasmaevels of arginine vasopressin, which are secondary to aonosmotic hypersecretion of the hormone from the neu-ohypophysis, appear to be the most important.1,2

It has been known for years that the impairment inolute-free water excretion and hyponatremia are impor-ant prognostic markers in the general population ofatients with cirrhosis.3–6 In addition, several recenttudies in patients with cirrhosis awaiting liver trans-lantation have extended these observations by showinghat patients with hyponatremia are at high risk of earlyeath before transplantation and that the prognosticalue of low serum sodium concentration is independentf the severity of cirrhosis, as assessed by model fornd-stage liver disease (MELD) score.7–9 Nevertheless,espite this growing interest in the relationship betweenyponatremia and liver transplantation, to our knowl-dge there is no information as to whether the presencef hyponatremia before transplantation may influenceosttransplantation outcome. Therefore, the currenttudy was designed to investigate whether pretransplan-ation hyponatremia may affect morbidity and/or mor-ality after transplantation.

Abbreviations used in this paper: CPM, central pontine myelinolysis;RS, hepatorenal syndrome.© 2006 by the American Gastroenterological Association Institute

0016-5085/06/$32.00

doi:10.1053/j.gastro.2006.02.017

tro2asoa[ht

mMimirhtttaaastpshc1tma

tpstswipptpdc

twmctw

tiwepbepbltpfpitddiiatdoec

agmtssm0boottib1tg

1136 LONDOÑO ET AL GASTROENTEROLOGY VOL. 130, NO. 4

Materials and Methods

Study Population

The current investigation is a single-center retrospec-ive cohort study of 241 consecutive adult patients with cir-hosis submitted to liver transplantation at the Hospital Clínicf Barcelona during a 4-year period (January 2000–December003). Seventy other patients undergoing transplantation alsot our center during this period of time were excluded from thetudy because of retransplantation (n � 24), liver diseasesther than cirrhosis (acute liver failure [n � 19], familialmyloidotic polyneuropathy [n � 13], polycystic liver diseasen � 4], and hemangioendothelioma and idiopathic portalypertension [1 patient each]), and combined liver-kidneyransplantation (n � 8).

Management of Patients BeforeTransplantation

Complications of cirrhosis before transplantation wereanaged according to preestablished protocols as follows.oderate ascites was treated with low-sodium diet and diuret-

cs, and large ascites was treated with paracentesis plus albu-in, followed by low-sodium diet and diuretics, as described

n detail elsewhere.10 Hyponatremia was managed with fluidestriction (1000 mL/day), except for those patients with severeyponatremia (serum sodium �120 mEq/L) (n � 3), who werereated with hypertonic saline. The administration of hyper-onic saline is not a standard therapy for dilutional hypona-remia in cirrhosis because its effects on sodium concentrationre inconsistent and is associated with rapid accumulation ofscites and edema.11,12 Nevertheless, in the absence of effectivelternative therapies, this treatment was used in patients withevere hyponatremia in an attempt to improve serum concen-ration before transplantation and reduce the risk of centralontine myelinolysis (CPM) related to rapid increases in serumodium after transplantation.13 Of the 3 patients with severeyponatremia treated with hypertonic saline, serum sodiumoncentration did not change significantly in 2 (from 116 and15 mEq/L before treatment to 118 and 114 mEq/L afterreatment, respectively) and increased markedly in 1 (from 112Eq/L to 131 mEq/L). All of these 3 patients were alive 1 year

fter transplantation.Hepatorenal syndrome (HRS) in patients awaiting liver

ransplantation was treated according to following criteria:atients with type 1 HRS and those with type 2 HRS witherum creatinine greater than 2 mg/dL (174 �mol/L) receivederlipressin and albumin.14 Patients with HRS type 2 witherum creatinine equal to or lower than 2 mg/dL (174 �mol/L)ere not treated with terlipressin and albumin because the

nformation on the efficacy of treatment in this group ofatients was very limited at the time of the study.14 Of the 241atients included in the study, 14 patients had HRS beforeransplantation, and 2 of them met criteria for treatment. Oneatient had type 1 HRS and underwent transplantation after 3ays of therapy while renal function was improving (serum

reatinine decreased from 6.7 mg/dL to 3.6 mg/dL), whereas a

he other patient had type 2 HRS and recovered from HRSith serum creatinine decreasing from 2.1 mg/dL to 1.0g/dL. Spontaneous bacterial peritonitis was treated with

eftriaxone plus intravenous albumin followed by norfloxacino prevent recurrence, and the other complications of cirrhosisere managed with conventional therapy.10,15

Management of Patients During and AfterTransplantation

According to the transplantation protocol in our cen-er, all patients were seen by an experienced hepatologistmmediately before transplantation. Preoperative routineorkout included a complete medical history, with particular

mphasis on the possible events occurring the preceding days,hysical examination; electrocardiogram; chest x-ray; andlood samples measuring liver and renal function tests, serumlectrolytes, blood cell count, urine sediment, and coagulationarameters. Biochemical parameters measured immediatelyefore transplantation were those used to evaluate the effect ofiver and renal function on patient’s outcome after transplan-ation. In patients with ascites, a diagnostic paracentesis waserformed for cell count and culture. Patients were then trans-erred to the surgical room in which liver transplantation waserformed. In all patients, transplantation was performed us-ng the piggyback technique.16 After surgery, patients wereransferred to the intensive care unit for continued care. Stan-ard liver and renal function tests were measured twice a dayuring the first 2 days and daily thereafter, unless otherwisendicated. Stable patients were transferred to a standard wardn which blood samples were routinely obtained at least twiceweek to measure blood cell count, liver and renal function

ests, serum electrolytes, and coagulation parameters. Afterischarge from the hospital, patients were followed in theutpatient clinic every month during the first 3 months andvery 2–3 months during the first year, or more frequently iflinically indicated.

Standard immunosuppression in our center consisted of thedministration of prednisone and tacrolimus. Prednisone wasiven at decreasing doses from 200 mg/day the first day to 20g/day at day 6 after transplantation. This dose was main-

ained until 1 month after transplantation. If there were noigns of rejection, prednisone was tapered progressively andtopped within 6–12 months after transplantation. Tacroli-us was started the first day after transplantation at a dose of

.1 mg/kg/day and then adjusted to maintain blood levelsetween 8 and 15 ng/mL within the first 3 months. The dosef tacrolimus was then adjusted to achieve target blood levelsf 7–12 ng/mL during months 4 to 12 and 5 to 10 ng/mLhereafter. Patients with diabetes mellitus before transplanta-ion were treated with cyclosporine instead of tacrolimus, at annitial dose of 10 mg/kg per day to achieve trough blood levelsetween 150 and 300 ng/mL during the first 3 months,00–200 ng/mL from months 4 to 12, and 50–150 ng/mLhereafter. In patients with serum creatinine levels equal to orreater than 1.5 mg/dL before transplantation or immediately

fter transplantation, the initiation of tacrolimus or cyclospor-

ictdgpdww

mdaict

oiofdgadtsstlcinwdwha(vcrm1A

scru

tMFfIc

atDta(gc

T

AS

E

AH

RSP

ASSMC

Na

5(b

L1pc

iI

April 2006 HYPONATREMIA AND LIVER TRANSPLANTATION 1137

ne administration was delayed until a reduction in serumreatinine concentration was observed. In the meantime, pa-ients were treated with prednisone, at identical doses asescribed previously, and mycophenolate mofetil at a dose of 1/12 hours. In patients developing renal failure after trans-lantation, the doses of tacrolimus or cyclosporine were re-uced or the drugs temporarily withdrawn. In patients inhom these drugs were withdrawn, mycophenolate mofetilas given until renal function recovered.

End Points and Definitions

The primary end point of this study was survival at 3onths after transplantation. Secondary end points were the

evelopment of complications, including early severe intra-bdominal bleeding and renal failure, neurologic disorders,nfectious complications, acute rejection, and biliary and vas-ular complications during the first month after transplanta-ion.

Severe intra-abdominal bleeding was defined as the presencef abdominal hemorrhage requiring reintervention for bleed-ng control within the first 48 hours after surgery, eliminationf intra-abdominal hematomas, and/or removal of gauzes usedor packing during the surgical procedure. Renal failure wasefined as an increase in serum creatinine concentration ofreater than 50% of the immediate pretransplantation value tofinal value greater than 2 mg/dL (174 �mol/L). Neurologicisorders were defined as the development of signs or symp-oms of abnormal neurologic function and were classified asevere, which included coma, seizures, and/or CPM, and non-evere (all other abnormalities of neurologic function).17 Infec-ious complications were defined by the presence of systemic orocal signs of infection together with positive cultures and/orompatible radiologic findings requiring intravenous admin-stration of antimicrobial agents. Acute rejection was diag-osed on the basis of Banff criteria.18 Biliary complicationsere defined as the presence of biliary strictures or bile leaksiagnosed by imaging techniques. Vascular complicationsere defined as the development of thrombosis or stenosis ofepatic artery or portal vein diagnosed by doppler ultrasoundnd angiography. Patients receiving a liver from a living donorn � 35) were excluded from the analysis of biliary andascular complications because this population has an in-reased risk of surgical complications compared with patientseceiving a liver from a cadaveric donor.19 Finally, hyponatre-ia was defined as serum sodium concentration lower than

30 mEq/L according to the definition of the Internationalscites Club.1

Statistical Analysis

Clinical and analytical variables were analyzed as pos-ible predictors of survival in a univariate analysis, and survivalurves (Kaplan–Meier method) were compared with the log-ank test. A multivariate analysis of survival was performed

sing a Cox regression method. Comparisons of variables be- i

ween groups of patients were made using the nonparametricann–Whitney test for continuous data and the �2 test or

isher test for categorical data. Statistical analysis was per-ormed using the SPSS 10 for Windows (SPSS Inc., Chicago,L). Results are expressed as mean � SD. P � .05 wasonsidered statistically significant.

Results

Characteristics of the Patients

The study population includes 241 consecutivedult patients with cirrhosis undergoing liver transplan-ation during a 4-year period, from January 2000 toecember 2003. Two hundred six out of the 241 pa-

ients included in the study (85%) received a liver fromcadaveric donor, whereas the remaining 35 patients

15%) received a graft from a living donor. Demo-raphic, clinical, and biochemical data of patients in-luded in the study at time of transplantation are shown

able 1. Demographic and Clinical Data and Liver andRenal Function Tests at Time of Transplantation inthe 241 Patients Included in Study

ge (y) 55 � 9 (24–69)exMale 152 (63%)Female 89 (37%)

tiology of cirrhosis, No. patientsHepatitis C 135 (56%)Alcohol 46 (19%)Othera 60 (25%)

scites 162 (67%)epatic encephalopathy, No.

patientsPast history 82 (34%)At transplantation 15 (6%)

enal failure,b No. patients 19 (8%)erum bilirubin (mg/dL) 2.8 � 3.5 (0.3–30)rothrombin timeRatio (%) 65 (26–100)INRc 1.5 � 0.3 (1–3.3)

lbumin (g/L) 32 � 6 (18–47)erum creatinine (mg/dL) 1.0 � 0.3 (0.4–3.6)erum sodium (mEq/L) 136 � 5 (114–146)ELD score 17 � 6 (6–40)hild-Pugh score 8 � 2 (5–13)

OTE. Values are mean � SD (ranges).Hepatitis B (18 patients; 7%), hepatitis C and alcohol (13 patients;%), chronic cholestatic liver diseases (11 patients; 5%), and other18 patients; 7%).As defined by serum creatinine greater than 1.5 mg/dL (133 �mol/). Causes of renal failure were hepatorenal syndrome in 14 (type I inand type II in 13) and chronic, nonfunctional renal diseases in 5

atients.International normalized ratio was calculated according to the follow-ng formula: (prothrombin time of patient/control prothrombin time)ISI,SI being the International Sensitivity Index for thromboplastins.

n Table 1.

thf2awhctpfh

(tm(adpctiaitcttiaNrtamtwTspnTmwl

mCctmreh3lsndtmp(np2dbtfcc

tcvmb3dtlicfdpttwitdcdty

1138 LONDOÑO ET AL GASTROENTEROLOGY VOL. 130, NO. 4

Prevalence of Hyponatremia atTransplantation

Nineteen of the 241 patients (8%) had hypona-remia at the time of transplantation. Patients withyponatremia at transplantation had been hyponatremicor a mean period of 106 days (range, 12–237days). Tableshows the comparison of demographic and clinical data

nd liver and renal function tests in patients with andithout hyponatremia at transplantation. Patients withyponatremia had greater frequency of ascites and en-ephalopathy, more marked impairment of liver functionests, and higher MELD and Child–Pugh scores, com-ared with values in patients without hyponatremia. Therequency of HRS was similar in patients with and withoutyponatremia (5% and 6%, respectively; P � ns).

Complications After Transplantation

One hundred twenty-seven of the 241 patients53%) developed at least 1 major complication duringhe 30-day postoperative period. Overall, the most com-on complication was acute rejection in 52 patients

21%). In all patients, acute rejection resolved after thedministration of steroid boluses and/or adjustment ofoses of baseline immunosuppressive agents. Forty-sixatients (18% of the whole series) developed infectiousomplications: 15 (33%) had pneumonia, 9 (20%) biliaryract infection, 7 (15%) urinary tract infection, 7 (15%)ntraabominal infection, 5 (11%) catheter-related sepsis,nd 3 (7%) sepsis of unknown origin. Resolution of thenfection was obtained in 34 of the 46 patients (74%). Inhe remaining 12 patients, infection did not resolve andontributed to the death of the patients. Thirty-nine ofhe 241 patients (16%) developed renal failure withinhe 30-day postoperative period. Peak serum creatininen these patients was 3.7 mg/dL (range, 2.1–9.1 mg/dL),nd the median time to reach the peak value was 8 days.ine of the 39 patients (23%) with renal failure required

enal replacement therapy. No patient required long-erm hemodyalisis. Patients who developed renal failurefter transplantation did not have higher levels of im-unosuppressive agents (tacrolimus or cyclosporine)

hroughout the 30-day postoperative period comparedith those patients who did not develop renal failure.hirty-two of the 241 (13%) patients included in the

tudy developed neurologic disorders within the 30-dayostoperative period. Median time to the development ofeurologic disorders was 10 days (range, 2–29 days).wenty-nine of the 32 patients (91%) developed alteredental status without focal motor disorders, associatedith seizures in 4 patients, in the absence of significant

esions in cerebral computerized tomography (CT) or o

agnetic resonance imaging (MRI); 2 patients developedPM and 1 patient cerebral hemorrhage. Serum sodiumoncentrations before and immediately after transplanta-ion in the 2 patients who developed CPM were 146Eq/L and 124 mEq/L vs 142 mEq/L and 140 mEq/L,

espectively. In all patients, neurologic function recov-red completely, except for the patient with cerebralemorrhage and 1 of the 2 patients with CPM, who died8 and 103 days after transplantation, respectively. Thisatter patient died from hepatic artery thrombosis whenymptoms because of CPM were improving. There wereo significant differences between patients who did andid not develop neurologic complications with respect tohe blood levels of immunosuppressive agents (tacroli-us or cyclosporine) obtained throughout the 30-day

ostoperative period. Twenty-five of the 241 patients10%) had severe intra-abdominal bleeding. The averageumber of packed red blood cells transfused in these 25atients was 19, compared with only 5 in the remaining16 patients (P � .0001). Resolution of the intra-ab-ominal bleeding was obtained in 23 patients. Persistentleeding contributed to death in the remaining 2 pa-ients. Finally, out of the 206 patients receiving a liverrom a cadaveric donor, 7 patients (3%) developed biliaryomplications, and 11 patients (5%) developed vascularomplications during the 30-day postoperative period.

To assess whether the presence of hyponatremia beforeransplantation was associated with an increased risk ofomplications after transplantation, patients were di-ided according to the presence or absence of hyponatre-ia at transplantation, and the rate of complications

etween the 2 groups was compared. As shown in Table, patients with hyponatremia had a greater risk ofeveloping complications after transplantation than pa-ients without hyponatremia, the difference being at theevel of significance. When complications were analyzedndividually, patients with hyponatremia had a signifi-antly greater frequency of neurologic disorders, renalailure, and infectious complications. Severe neurologicisorders were also more frequent in patients with hy-onatremia as compared with patients without hypona-remia (21% and 5%, respectively; P � .003). By con-rast, the presence of hyponatremia was not associatedith an increased risk of severe intra-abdominal bleed-

ng, acute rejection, and biliary or vascular complica-ions. Figure 1 shows the cumulative probability ofevelopment of neurologic disorders, infectious compli-ations, and renal failure after transplantation in patientsivided according to the presence or absence of hypona-remia at time of transplantation. Using regression anal-sis, which included demographic and clinical and lab-

ratory variables obtained at the time of transplantation,

tpaiptpH

fii

cbhi

apn3h9l

T

AS

E

AD

HH

SP

ASSMC

Na

T

ANRISVBA

N

April 2006 HYPONATREMIA AND LIVER TRANSPLANTATION 1139

he study found hyponatremia to be an independentredictive factor of development of neurologic disordersnd renal failure. HRS before transplantation was also anndependent predictive factor of renal failure after trans-lantation. Six of the 14 patients with HRS beforeransplantation (43%) developed renal failure after trans-lantation compared with 33 of the 227 patients withoutRS (19%) (P � .04).

Survival After TransplantationFourteen of the 241 patients (5%) died during the

rst 3 months after transplantation. Causes of death werenfections in 7 patients (bacterial in 5 and fungal in 2);

able 2. Demographic and Clinical Data and Liver and RenalPresence or Absence of Hyponatremia

Variable Hyponatremia (n

ge (y) 52 � 13 (24–6exMale 9 (47Female 10 (53

tiology of cirrhosis, No. patientsHepatitis C 11 (58Alcohol 4 (21Other 4 (21

scites, No. patients 19 (10iureticsa

Number of patients 12 (63Furosemide (mg/day) 80 � 69 (40–1Spironolactone (mg/day) 167 � 121 (10

epatorenal syndrome, No. patients 1 (5%epatic encephalopathy, No. patientsPast history 9 (47At transplantation 4 (21

erum bilirubin (mg/dL) 7.5 � 7.5 (0.6rothrombin timeRatio (%) 55 � 22 (26–9INR 1.7 � 0.5 (1–2

lbumin (g/L) 28 � 5 (22–41erum creatinine (mg/dL) 1.2 � 0.6 (0.4erum sodium (mEq/L) 125 � 4 (114–1ELD score 20 � 7 (9–40)hild-Pugh score 10 � 2 (6–13)

OTE. Values are mean � SD (ranges).Patients being treated with diuretics at the time of transplantation.

able 3. Frequency of Major Complications of TransplantatioPresence or Absence of Hyponatremia at Time of Tr

ComplicationsHyponatremia, n (%)

(n � 19)

ny complication 14 (74)eurologic disorders 7 (37)enal failure 7 (37)

nfectious complications 7 (37)evere intraabdominal bleeding 4 (21)ascular complications 0iliary complications 1 (7)cute rejection 2 (10)

OTE. Hyponatremia � serum sodium �130 mE/L. OR, odds ratio; CI, co

ardiovascular complications in 2; intra-abdominalleeding in 2; and gastrointestinal bleeding, cerebralemorrhage, and recurrence of hepatocellular carcinoman 1 patient each.

The presence of hyponatremia at transplantation wasssociated with a reduced short-term survival after trans-lantation. As shown in Figure 2, patients with hypo-atremia at the time of transplantation had a lower-month survival compared with that of patients withoutyponatremia at the time of transplantation (84% vs5%, respectively; P � .05). One-year survival was alsoower in patients with hyponatremia compared with that

tion Tests at Time of Transplantation According to

9) No hyponatremia (n � 222) P value

55 � 9 (24–69) ns

143 (64%) ns79 (36%) ns

ns124 (56%)42 (19%)56 (25%)

143 (64%) �.01

119 (54%) ns45 � 23 (20–160) ns

0) 105 � 42 (25–300) ns13 (6%) ns

73 (33%) ns11 (5%) .01

2.4 � 2.6 (0.3–29) �.01

66 � 16 (32–100) .021.5 � 0.3 (1–3.3) ns33 � 6 (18–47) �.011.0 � 0.3 (0.4–2.7) ns138 � 3 (130–146) �.0117 � 5 (6–30) .048 � 2 (5–13) �.01

the 241 Patients Included in Study Divided According tolantation

No hyponatremia, n (%)(n � 222) OR (95% CI) P value

113 (51) 2.7 (1–8) .0525 (11) 4.6 (1.6–13) .00632 (14) 3.4 (1.3–9.4) .0239 (17) 2.7 (1–7.3) .0421 (10) 2.5 (0.7–8.3) .111 (6) .66 (13) 1.6 (0.2–14) .7

50 (23) 0.4 (0.1–1.8) .2

Func

� 1

5)

%)%)

%)%)%)0%)

%)60)

0–40)

%)%)–30)

8).6))

–3.6)29)

n inansp

nfidence interval.

otwedpt.

fMttliw

va

3v3dLsacddi

pisgptctii3uishhF

Fv

Fcdo

1140 LONDOÑO ET AL GASTROENTEROLOGY VOL. 130, NO. 4

f patients without hyponatremia (74% vs 91%, respec-ively; P � .017). Nevertheless, when patients who diedithin the first 3 months after transplantation were

xcluded from the analysis, there were no significantifferences in 1-year survival between patients with hy-onatremia and patients without hyponatremia at theime of transplantation (88% vs 95%, respectively; P �16).

In univariate analysis, neither the severity of liverailure, as assessed by individual liver function tests or

ELD or Child–Pugh scores, serum creatinine concen-ration, renal failure at transplantation, nor the charac-eristics of the donor (cadaveric vs living related, age,iver tests, presence and degree of liver steatosis, time inntensive care unit, and ischemia time) were associated

igure 1. Probability of developing neurologic disorders, infectiousomplications, and renal failure in all patients included in the studyivided according to the presence (dashed line) or absence (solid line)f hyponatremia at transplantation.

ith prognosis after transplantation. The prognostic o

alue of hyponatremia was maintained in multivariatenalysis (odds ratio, 3.6).

Finally, the effect of postoperative complications on-month survival was also evaluated. Patients who de-eloped renal failure after transplantation had a reduced-month survival compared with that of patients whoid not develop renal failure (79% vs 97%, P � .001).ikewise, the development of infectious complications orevere intra-abdominal bleeding was also associated withreduced 3-month survival (85% and 84%, respectively,ompared with 96% and 95% in patients who did notevelop these complications). By contrast, neither theevelopment of neurologic disorders nor vascular or bil-ary complications were associated with a poor outcome.

Discussion

The main finding of the current study is thatretransplantation hyponatremia was associated with anmpaired short-term posttransplantation outcome in aeries of 241 consecutive patients with cirrhosis under-oing transplantation in a single institution over a 4-yeareriod. Patients with hyponatremia at time of transplan-ation had a greater risk of major posttransplantationomplications within the first month after transplanta-ion, including neurologic disorders, renal failure, andnfectious complications. The presence of hyponatremiancreased the risk of developing these complications 4.6,.4, and 2.7 times, respectively, compared with patientsndergoing transplantation without hyponatremia. Moremportantly, 3-month survival after transplantation wasignificantly lower in patients with pretransplantationyponatremia compared with that of patients withoutyponatremia (84% vs 95%, respectively; P � .05;igure 2).

igure 2. Three-month survival after transplantation of patients, di-ided according to the presence (dashed line) or absence (solid line)

f hyponatremia at the time of transplantation.

nclcicotesdgtctsWtpdtc

owtpttstttvtnipcwim

twfsscea

TtaasvtpaispeTpmctf

efawnohratitlpcdcm�shanrncdtr2

April 2006 HYPONATREMIA AND LIVER TRANSPLANTATION 1141

It is well established that the occurrence of hypo-atremia entails a poor prognosis regardless of thelinical condition causing hyponatremia. Epidemio-ogic studies in hospitalized patients with a variety oflinical conditions have shown an increased risk ofn-hospital mortality in patients with hyponatremia asompared with nonhyponatremic controls.20,21 More-ver, studies in patients with heart failure have shownhat hyponatremia is one of the best parameters tostimate short-term mortality.22 Finally, a number oftudies in patients with cirrhosis have consistentlyemonstrated that hyponatremia is one of the stron-est predictive factors of poor survival in several pa-ient populations, including patients with stable as-ites, patients with bacterial infections, patientsreated with transjugular intrahepatic portosystemichunts, and patients listed for transplantation.5–9,23,24

hether the prognostic value of hyponatremia inhese different conditions is due to the fact that it is arecise marker of the disease stage or to negativeirect effects of low serum sodium levels on the func-ion of several organs (ie, central nervous system,irculatory system) or both is not known.

The results of the current study extend these previousbservations and show that hyponatremia is associatedith a reduced survival early (3-month) after transplan-

ation. However, if patients with hyponatremia pretrans-lantation survive 3 months, then the outcome is similaro that of patients without hyponatremia pretransplan-ation. The lower early posttransplantation survival ob-erved in patients with hyponatremia compared withhat of patients without hyponatremia in our transplan-ation population was probably related, at least in part,o the fact that hyponatremia increased the risk of de-elopment of major complications early after transplan-ation, such as renal failure, infectious complications, andeurologic disorders, which are known to have a negativempact on survival of transplantation patients.25–27 Theossible mechanism(s) by which hyponatremia may in-rease the risk of these latter complications in patientsith cirrhosis undergoing liver transplantation was not

nvestigated in the current study but deserves a com-ent.It is well established that the presence of hypona-

remia identifies a group of patients with cirrhosisith a particularly severe impairment in circulatory

unction.1,2 Clinical and experimental studies havehown convincingly that hyponatremia in cirrhosis isecondary to a marked impairment of the arterialirculation, which is characterized by a reduction inffective arterial blood volume related to an intense

rterial vasodilatation of the splanchnic circulation. b

his reduction in effective arterial blood volume leadso a marked stimulation of the renin-angiotensin-ldosterone system and sympathetic nervous systemnd nonosmotic hypersecretion of arginine vasopres-in.1,2 One of the main consequences of the increasedasopressin levels is an impairment of solute-free wa-er excretion and the subsequent development of hy-onatremia. Therefore, considering that changes inrterial pressure occur frequently in patients undergo-ng liver transplantation, both during or after theurgical procedure, the existence of a severely im-aired circulatory function before transplantation maynhance the risk of renal failure after transplantation.his situation of increased risk of renal failure inatients undergoing transplantation with hyponatre-ia is reminiscent of what occurs in patients with

irrhosis not submitted to transplantation in whomhe presence of hyponatremia is one of the main riskactors of the development of hepatorenal syndrome.28

Several previous studies have suggested that the pres-nce of hyponatremia before liver transplantation is a riskactor of neurologic complications after transplantation,lthough a study specifically comparing hyponatremicith nonhyponatremic patients has, to our knowledge,ot been reported.13,29 The most severe neurologic dis-rder described in hyponatremic patients is CPM, whichas been reported to occur in 1%–2% of liver transplantecipients and is thought to be mainly related to a rapidnd marked increase in serum sodium concentration afterransplantation.13,29,30 The results of the current studyndicate that patients with hyponatremia at transplanta-ion have a 4.6 times higher risk of developing neuro-ogic complications compared with patients without hy-onatremia. Patients who developed neurologicomplications did not differ from those who did notevelop neurologic complications with respect tohanges in serum sodium concentration before and im-ediately after transplantation (5.4 � 6.8 mEq/L vs 2.73.8 mEq/L, respectively; P � .12). Changes in serum

odium concentration before transplantation and at 48ours after transplantation between the 2 groups werelso not significantly different. These data suggest thateurologic complications in this setting are probablyelated to hyponatremia and its effects on the centralervous system rather than to changes in serum sodiumoncentration occurring after transplantation. The inci-ence of CPM in patients with hyponatremia included inhe current study (2 out of 19, 10%) is lower than thateported in a previous study (3 out of 12 patients,5%).29

A number of studies have shown that renal failure

efore transplantation is an important risk factor of poor

ssrpitcotTmlvmoob

tthttpppcloavattetoa

1

1

1

1

1

1

1

1

1

1

2

2

2

2

1142 LONDOÑO ET AL GASTROENTEROLOGY VOL. 130, NO. 4

urvival after liver transplantation.25,31,32 In the currenteries, neither serum creatinine levels nor the existence ofenal failure before transplantation was associated withoor survival after transplantation. Most published stud-es demonstrating the prognostic value of pretransplan-ation renal failure on postransplantation outcome in-luded patients undergoing transplantation in the 1980sr 1990s,25,31,32 whereas the current study includes pa-ients undergoing transplantation from 2000 to 2003.herefore, it is possible that recent improvement in theanagement of transplantation patients accounts for the

ack of effect of renal failure pretransplantation on sur-ival postransplantation. These may include an improve-ent in general postransplantation care, the routine use

f the surgical piggyback technique,16 the modificationsf immunosuppressive regimes, and treatment of HRSefore transplantation.14,33

In conclusion, the results of the current study indicatehat, in patients with cirrhosis, the presence of hypona-remia at the time of transplantation is associated with aigh rate of neurologic disorders, infectious complica-ions, and renal failure during the first month afterransplantation and an impaired 3-month survival com-ared with patients without hyponatremia. Whether therognostic value of hyponatremia is related directly to aossible deleterious effect of hyponatremia and the asso-iated hypoosmolality and/or indirectly to a more advancediver disease is not known and deserves investigation. Thebservation in patients awaiting transplantation,7–9 as wells the results of the current study, that the predictivealue of hyponatremia is independent of liver function, asssessed by MELD score, does not support the concepthat the poor outcome of hyponatremic patients is relatedo the coexistence of a more severe liver disease. What-ver the reason, the results of the current study indicatehat hyponatremia in cirrhosis should be considered notnly a risk factor of death before transplantation but alsorisk factor of impaired early posttransplantion outcome.

References1. Ginès P, Berl T, Bernardi M, Bichet DG, Hamon G, Jimenez W,

Liard JF, Martin PY, Schrier RW. Hyponatremia in cirrhosis: frompathogenesis to treatment. Hepatology 1998;28:851–864.

2. Ishikawa S, Schrier RW. Pathogenesis of hyponatremia: the roleof arginine vasopressin. In: Ginès P, Arroyo V, Rodés J, SchrierRW, eds. Ascites and renal dysfunction in liver disease. Oxford,UK: Blackwell Publishing, 2005:305.

3. Hecker R, Sherlock S. Electrolyte and circulatory changes interminal liver failure. Lancet 1956;271:1121–1125.

4. Arroyo V, Rodes J, Gutierrez-Lizarraga MA, Revert L. Prognosticvalue of spontaneous hyponatremia in cirrhosis with ascites.Am J Dig Dis 1976;21:249–256.

5. Llach J, Ginès P, Arroyo V, Rimola A, Tito L, Badalamenti S,Jimenez W, Gaya J, Rivera F, Rodés J. Prognostic value of arterial

pressure, endogenous vasoactive systems, and renal function in

cirrhotic patients admitted to the hospital for the treatment ofascites. Gastroenterology 1988;94:482–487.

6. Fernandez-Esparrach G, Sanchez-Fueyo A, Gines P, Uriz J, QuintoL, Ventura PJ, Cardenas A, Guevara M, Sort P, Jimenez W,Bataller R, Arroyo V, Rodés J. A prognostic model for predictingsurvival in cirrhosis with ascites. J Hepatol 2001;34:46–52.

7. Heuman DM, Abou-Assi SG, Habib A, Williams LM, Stravitz RT,Sanyal AJ, Fisher RA, Mihas AA. Persistent ascites and low serumsodium identify patients with cirrhosis and low MELD scores whoare at high risk for early death. Hepatology 2004;40:802–810.

8. Biggins SW, Rodriguez HJ, Bacchetti P, Bass NM, Roberts JP,Terrault NA. Serum sodium predicts mortality in patients listed forliver transplantation. Hepatology 2005;41:32–39.

9. Ruf AE, Kremers WK, Chavez LL, Descalzi VI, Podesta LG, VillamilFG. Addition of serum sodium into the MELD score predictswaiting list mortality better than MELD alone. Liver Transpl 2005;11:336–343.

0. Ginès P, Cardenas A, Arroyo V, Rodés J. Management of cirrhosisand ascites. N Engl J Med 2004;350:1646–1654.

1. Moore KP, Wong F, Ginès P, Bernardi M, Ochs A, Salerno F, AngeliP, Porayko M, Moreau R, Garcia-Tsao G, Jimenez W, Planas R,Arroyo V. The management of ascites in cirrhosis: report on theconsensus conference of the International Ascites Club. Hepa-tology 2003;38:258–266.

2. Ginès P, Arroyo V, Rodés J. Ascites, hepatorenal syndrome andspontaneous bacterial peritonitis. In: McDonald J, Burroughs A,Feagan BG, eds. Evidence based-gastroenterology and hepatol-ogy. Oxford, UK: Blackwell Publishing, 2005:487.

3. Wszolek ZK, McComb RD, Pfeiffer RF, Steg RE, Wood RP, ShawBW Jr, Markin RS. Pontine and extrapontine myelinolysis follow-ing liver transplantation. Relationship to serum sodium. Trans-plantation 1989;48:1006–1012.

4. Ginès P, Guevara M, Arroyo V, Rodés J. Hepatorenal syndrome.Lancet 2003;362:1819–1827.

5. Cardenas A, Ginès P. Management of complications of cirrhosisin patients awaiting liver transplantation. J Hepatol 2005;42(Suppl 1):S124–S133.

6. Gonzalez FX, Garcia-Valdecasas JC, Grande L, Pacheco JL, CugatE, Fuster J, Lacy AM, Taura P, Lopez-Boado MA, Rimola A, Visa J.Vena cava vascular reconstruction during orthotopic liver trans-plantation: a comparative study. Liver Transpl Surg 1998;4:133–140.

7. Bronster DJ, Emre S, Boccagni P, Sheiner PA, Schwartz ME, MillerCM. Central nervous system complications in liver transplantrecipients—incidence, timing, and long-term follow-up. ClinTransplant 2000;14:1–7.

8. Ormonde DG, de Boer WB, Kierath A, Bell R, Shilkin KB, HouseAK, Jeffrey GP, Reed WD. Banff schema for grading liver allograftrejection: utility in clinical practice. Liver Transpl Surg 1999;5:261–268.

9. Olthoff KM, Merion RM, Ghobrial RM, Abecassis MM, Fair JH,Fisher RA, Freise CE, Kam I, Pruett TL, Everhart JE, Hulbert-Shearon TE, Gillespie BW, Emond JC; A2ALL Study Group. Out-comes of 385 adult-to-adult living donor liver transplant recipi-ents. A report from the A2ALL consortion. Ann Surg 2005;242:314–325.

0. Anderson RJ. Hospital-associated hyponatremia. Kidney Int1986;29:1237–1247.

1. Baran D, Hutchinson TA. The outcome of hyponatremia in ageneral hospital population. Clin Nephrol 1984;22:72–76.

2. Lee DS, Austin PC, Rouleau JL, Liu PP, Naimark D, Tu JV. Pre-dicting mortality among patients hospitalised for heart failure.Derivation and validation of a clinical model. JAMA 2003;290:2581–2587.

3. Follo A, Llovet JM, Navasa M, Planas R, Forns X, Francitorra A,Rimola A, Gassull MA, Arroyo V, Rodes J. Renal impairment after

spontaneous bacterial peritonitis in cirrhosis: incidence, clinical

2

2

2

2

2

2

3

3

3

3

Ce

0I

April 2006 HYPONATREMIA AND LIVER TRANSPLANTATION 1143

course, predictive factors and prognosis. Hepatology 1994;20:1495–1501.

4. Jalan R, Elton RA, Redhead DN, Finlayson NDC, Hayes PC. Anal-ysis of prognostic variables in the prediction of mortality, shuntfailure, variceal rebleeding and encephalopathy following thetransjugular intrahepatic portosystemic stent-shunt for varicealhemorrhage. J Hepatol 1995;23:123–128.

5. Nair S, Verma S, Thuluvath PJ. Pretransplant renal function pre-dicts survival in patients undergoing orthotopic liver transplanta-tion. Hepatology 2002;35:1179–1185.

6. Lewis MB, Howdle PD. Neurologic complications of liver trans-plantation in adults. Neurology 2003;61:1174–1178.

7. Paya CV, Hermans PE, Washington JA II, Smith TF, Anhalt JP,Wiesner RH, Krom RA. Incidence, distribution, and outcome ofepisodes of infection in 100 orthotopic liver transplantations.Mayo Clin Proc 1989;64:555–564.

8. Ginès A, Escorsell A, Ginès P, Saló J, Jiménez W, Inglada L,Navasa M, Clària J, Rimola A, Arroyo V, Rodés J. Incidence,predictive factors, and prognosis of the hepatorenal syndrome incirrhosis with ascites. Gastroenterology 1993;105:229–236.

9. Abbasoglu O, Goldstein RM, Vodapally MS, Jennings LW, LevyMF, Husberg BS, Klintmalm GB. Liver transplantation in hy-ponatremic patients with emphasis on central pontine myelinol-ysis. Clin Transplant 1998;12:263–269.

0. Pujol A, Graus F, Rimola A, Beltran J, Garcia-Valdecasas JC,

Navasa M, Grande L, Galofre J, Visa J, Rodés J. Predictive factors f

of in-hospital CNS complications following liver transplantation.Neurology 1994;44:1226–1230.

1. Rimola A, Gavaler JS, Schade RR, El-Lakany S, Starzl TE, VanThiel DH. Effects of renal impairment on liver transplantation.Gastroenterology 1987;93:148–156.

2. Ghobrial RM, Steadman R, Gornbein J, Lassman C, Holt CD,Chen P, Farmer DG, Yersiz H, Danino N, Collisson E, Baquarizo A,Han SS, Saab S, Goldstein LI, Donovan JA, Esrason K, BusuttilRW. A 10-year experience of liver transplantation for hepatitis C:analysis of factors determining outcome in over 500 patients.Ann Surg 2001;234:384–393.

3. Restuccia T, Ortega R, Guevara M, Ginès P, Alessandria C,Ozdogan O, Navasa M, Rimola A, Garcia-Valdecasas JC, Arroyo V,Rodes J. Effects of treatment of hepatorenal syndrome beforetransplantation on posttransplantation outcome. A case-controlstudy. J Hepatol 2004;40:140–146.

Received October 21, 2005. Accepted December 21, 2005.Address requests for reprints to: Pere Ginès, MD, Liver Unit, Hospital

línic de Barcelona, C/ Villarroel 170, 08036, Barcelona, Spain.-mail: pgines@clinic.ub.es; fax: (34) 93 4515522.Supported by grants from the Fondo de Investigación Sanitaria (FIS

5/0246), Ministerio de Educación y Ciencia (SAF 2005-01917), andnstituto de Salud Carlos III (CO3/2 and CO3/3) and a research grant

rom Fundación Carolina (to M-C.L).