American Society of Critical Care Anesthesiologists

Section Editor: Michael J. Murray

Early Determinants of Death Due to Multiple OrganFailure After Noncardiac Surgery in High-Risk PatientsSuzana M. Lobo, MD, PhD,* Ederlon Rezende, MD,† Marcos F. Knibel, MD, PhD,‡Nilton B. Silva, MD, PhD,§ Jose A. Paramo, MD,� Flavio E. Nacul, MD,¶ Ciro L. Mendes, MD,#Murilo Assuncao, MD,** Rubens C. Costa, MD,†† Cíntia C. Grion, MD, PhD,‡‡Sergio F. Pinto, MD,§§ Patricia M. Mello, MD,�� Marcelo O. Maia, MD,¶¶ Pericles A. Duarte, MD,##Fernando Gutierrez, MD, PhD,†† Joao M. Silva Junior, MD,*** Marcel R. Lopes, MD,†††Jose A. Cordeiro, PhD,‡‡‡ and Charles Mellot, MD, PhD§§§

BACKGROUND: Prediction of perioperative cardiac complications is important in the medicalmanagement of patients undergoing noncardiac surgery. However, these patients frequently die asa consequence of primary or secondary multiple organ failure (MOF), often as a result of sepsis. Weinvestigated the early perioperative risk factors for in-hospital death due to MOF in surgical patients.METHODS: This was a prospective, multicenter, observational cohort study performed in 21Brazilian intensive care units (ICUs). Adult patients undergoing noncardiac surgery who wereadmitted to the ICU within 24 hours after operation were evaluated. MOF was characterized bythe presence of at least 2 organ failures. To determine the relative risk (RR) of in-hospital deathdue to MOF, we performed a logistic regression multivariate analysis.RESULTS: A total of 587 patients were included (mean age, 62.4 � 17 years). ICU and hospitalmortality rates were 15% and 20.6%, respectively. The main cause of death was MOF (53%).Peritonitis (RR 4.17, 95% confidence interval [CI] 1.38–12.6), diabetes (RR 3.63, 95% CI1.17–11.2), unplanned surgery (RR 3.62, 95% CI 1.18–11.0), age (RR 1.04, 95% CI 10.01–1.08), and elevated serum lactate concentrations (RR 1.52, 95% CI 1.14–2.02), a highcentral venous pressure (RR 1.12, 95% CI 1.04–1.22), a fast heart rate (RR 3.63, 95% CI1.17–11.2) and pH (RR 0.04, 95% CI 0.0005–0.38) on the day of admission were independentpredictors of death due to MOF.CONCLUSIONS: MOF is the main cause of death after surgery in high-risk patients. Awarenessof the risk factors for death due to MOF may be important in risk stratification and can suggestroutes for therapy. (Anesth Analg 2011;112:877–83)

Although millions of operations are performedannually worldwide, relatively few patients un-dergoing major surgery are deemed high risk

and at an increased risk of postoperative complicationsand death.1 A large observational study reported thatthis population accounts for only 12.5% of all surgicalprocedures, but for �80% of related deaths.1 Despite thehigh mortality rates, fewer than 15% of these patientswere admitted to the intensive care unit (ICU),1 showingthat individual risk can easily be underestimated andhigh-risk patients not recognized.

In patients with major operative trauma, multiple organfailure (MOF) may be induced by various mechanismsincluding an aggressive inflammatory response, although

the mechanisms by which this occurs are not clear.2 In-creasing evidence suggests that oxygen requirements in-crease significantly as a result of the injury and metabolicresponse to trauma. However, very frequently, high-riskpatients are unable to spontaneously increase their cardiacoutput to match this increased oxygen demand.3 Suchpatients are therefore more likely to develop oxygen debtand, as a consequence, severe systemic inflammation withdeath due to ongoing organ dysfunction and nosocomialsepsis.4 In addition, not infrequently, these patients un-dergo surgery for peritonitis and therefore already havesepsis when submitted to surgery.

Several attempts have been made to help identify pa-tients at risk of complications and death after surgery.Studies have identified predictors of morbidity and mor-tality after colon, esophageal, and gastric surgery, and afterpulmonary resection.5–7 Other studies have evaluated pre-dictors of postoperative cardiac complications in noncar-diac surgical patients.8–11 High-risk surgical patientsadmitted to the ICU frequently die as a consequence ofprimary or secondary MOF, the latter of which is fre-quently a result of sepsis.12,13 Predictors of death due toMOF have never been investigated in high-risk surgicalpatients. Therefore, we investigated the early perioperativerisk factors for in-hospital death due to MOF in a popula-tion of surgical patients admitted to the ICU.

Author affiliations are provided at the end of the article.

Accepted for publication February 25, 2010.

Supported by Centro de Estudos e Pesquisa em Medicina Intensiva, Hospitalde Base, Sao Jose do Rio Preto, Brazil.

The authors report no conflicts of interest.

Address correspondence and reprint requests to Dr. Suzana MargarethLobo, Faculdade de Medicina de Sao Jose do Rio Preto, Servico de TerapiaIntensiva do Hospital de Base e Laboratorio de Sepse, Avenida BrigadeiroFaria Lima, 5544 CEP 15090-000 Sao Jose do Rio Preto, SP, Brazil. Addresse-mail to suzanalobo@yahoo.com.

Copyright © 2011 International Anesthesia Research SocietyDOI: 10.1213/ANE.0b013e3181e2bf8e

April 2011 • Volume 112 • Number 4 www.anesthesia-analgesia.org 877

METHODSData were extracted from a prospective multicenterobservational cohort study, the SCORIS study, per-formed from April 1 to June 31, 2006, to which 21Brazilian ICUs from 18 institutions (8 public and 10private hospitals) contributed. The SCORIS study wasdesigned to evaluate the epidemiology and clinical out-comes of surgical ICU patients and to develop a model topredict the outcome of such patients in Brazilian ICUs.The IRB from the coordinator center waived the need forinformed consent because of the observational nature ofthe study. Institutional recruitment for participation wasby open invitation from the study steering committee.

Adult patients (834 cases) undergoing noncardiac sur-gery and admitted to a participating ICU within 24 hoursafter operation were evaluated for inclusion. Patients un-dergoing trauma, neurological, gynecologic, obstetric, orpalliative surgery were excluded.

Data were collected on age, gender, smoking habits,alcohol abuse, nutritional status, diabetes, renal function,chronic obstructive pulmonary disease, and presence ofmalignant disease.

Patients taking oral antidiabetic medications or insulinwere considered to have a diagnosis of diabetes. Cardiopa-thy was defined as the presence of moderate or severecardiomegaly, turgescent jugular veins, and use of digital,diuretics, antiangina, and antihypertensive drugs. Lowfunctional capacity was defined as inability to climb 2flights of stairs on subjective evaluation. Electrocardio-graphic abnormalities included nonsinus rhythms, frequentventricular extrasystoles (�5/min), Q waves, or ST-T seg-ment abnormalities. For the diagnosis of angina, the Cana-dian Cardiovascular Society classification system was used.A diagnosis of acute myocardial infarction required thepresence of typical electrocardiographic alterations to-gether with elevated cardiac enzymes and/or segmentalwall motion abnormalities on echocardiography. Otherclinical predictors of increased perioperative cardiovascu-lar risk were defined according to the American College ofCardiology/American Heart Association guidelines. All datawere entered on an electronic case report file (Comunicare®)and the variables were cross-checked by 2 of the authors.

The following procedures were considered major surgery:laparotomy, enterectomy, cholecystectomy with choledocho-stomy, major amputation, vascular and aortic procedures,rectal abdominoperineal resection, pancreatectomy, esopha-gectomy, and hepatectomy. Unplanned surgery comprisedadmissions after urgent (within 48 hours of referral) oremergent (immediately after referral/consultation) surgery.The Physiological and Operative Severity Score for the Enu-meration of Mortality and Morbidity (POSSUM), the AcutePhysiology and Chronic Health Disease Classification SystemII (APACHE II) score, the Multiple Organ Dysfunction Systemscore, and the Sequential Organ Failure Assessment scorewere calculated.14–17 Peritonitis was classified according tothe findings noted at laparotomy as either serum fluid asso-ciated with an infection site or the presence of collected ordiffuse purulent discharge in the abdominal cavity.14 Severebleeding was defined as an estimated blood loss �500 mLduring surgery. For vital signs, laboratory tests, and centralvenous pressure (CVP) measurements, the most abnormal

values collected over the first 24 hours of ICU stay wereregistered.

Sepsis syndromes were defined according to the consen-sus conference definitions.18 Early- and late-onset sepsiswere defined as a diagnosis of sepsis made within 72 hoursof ICU admission or thereafter, respectively.

Causes of death were classified as follows: MOF, char-acterized by the presence of at least 2 organ failurescontributing to death; refractory cardiovascular failure,characterized as uncontrollable hypotension despite high-dose vasopressors determining death; coagulation failure,characterized as need for massive transfusion, and hemor-rhagic shock after surgery; sudden death (unexpectedcardiac arrest); and unknown cause.

Data were analyzed using SPSS 13.0 for Windows (SPSS,Inc., Chicago, IL). Continuous variables are presented asmean � SD and/or median (range), and categorical vari-ables are reported as absolute numbers (percentages). Non-parametric tests of comparison were used for variablesevaluated as not normally distributed. Difference testingbetween groups was performed using the 2-tailed t test, �2

test, Wilcoxon test, analysis of variance, and Fisher exacttest as appropriate. Bonferroni adjustment was used formultiple comparisons. We considered P � 0.05 as statisti-cally significant.

We performed a logistic regression multivariate analysiswith in-hospital death due to MOF as the dependent factor.Variables considered for the regression analysis includedage, gender, comorbid diseases, the type of admission(planned or unplanned), the type of surgery (nonmajor ormajor), malnutrition, alcoholism, maximal heart rate (HR),lowest hemoglobin concentration and highest urea concen-tration before operation, severe bleeding during surgery(estimated blood loss �500 mL), arrhythmia during opera-tion, peritonitis, postoperative measurements of minimaland maximal CVP, minimal axillary temperature, totalleukocyte count, hemoglobin concentration, serum lactateconcentration, pH, and total platelet count. Colinearitybetween variables was excluded before modeling. Covari-ates were selected and entered in the model if they attaineda P � 0.2 on a univariate basis.

POSSUM scores with respective estimated mortalityrates were calculated. Discrimination for the severity scoreswas assessed by area under receiver operating characteris-tic (ROC) curves.

RESULTSCharacteristics of the Study GroupsA total of 587 consecutive patients were admitted to theparticipating ICUs during the study period (55% male; meanage � 62.4 � 17 years). A total of 247 patients were excluded(127 for neurosurgery; 35 palliative surgery; 6 gynecologicsurgery; 32 trauma; 34 lost to follow-up; 8 were younger than18 years; and 5 had no indication for ICU admission).

The characteristics of the study groups are shown inTable 1. Cardiopathy (35.4%), cancer (32%), and diabetes(20%) were the most prevalent comorbid conditions. Sixty-six percent of the patients were admitted after majorsurgery and 32% after unplanned surgery. The most com-mon type of surgery was gastrointestinal surgery (44%)followed by vascular procedures (23%).

Multiple Organ Failure After Noncardiac Surgery

878 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA

A total of 135 patients (23%) had sepsis during their ICUstay; 64% of these died in the hospital. One hundred forty-onepatients (24.0%) had peritonitis of whom 66 (46.8%) devel-oped sepsis, with 50 cases of early-onset sepsis (37%).

The ICU mortality rate was 15%. Overall hospital mor-tality rates were 16.7% at 30 days, 20.6% at 60 days, and20.6% at 90 days. The causes of death in the ICU were MOF(53%), sudden death (14.9%), refractory shock (6.8%),bleeding (2.5%), and unknown (22.8%).

Of the patients who died, 94% had significant medicalcomorbidities at the time of surgery (3.4 � 2.2), 66% hadundergone urgent surgery, 70% were older than 60years, and 46% older than 70 years. In addition, 35% hadlow functional capacity, 28% malnutrition, and 26%hemodynamic instability before surgery. Serum lactateconcentrations, HR, and CVP values were higher and pHlower on the first day of ICU admission in nonsurvivors(Table 2).

Table 1. Characteristics of the Study Group on Admission to the ICU, Occurrence of Sepsis, and ICU andHospital Lengths of Stay in Survivors and Nonsurvivors

Survivors Death due to MOF Death due to other causesNo. of patients 466 64 57Age (y), mean � SD 61.1 � 17.2 66.7 � 16.4* 67.5 � 12.0*Gender (male %) 255 (54.7) 36 (56.2) 31 (54.4)Comorbidities (%)

Cardiopathy 164 (35.2) 25 (39.0) 19 (33.3)Cancer 154 (33.0) 20 (31.2) 14 (24.5)Diabetes 84 (18.0) 24 (37.5)† 12 (21.0)Smoking (active in the last year) 92 (19.7) 13 (20.3) 13 (22.8)Low functional capacity 70 (15.0) 26 (40.6)†‡ 16 (21)Electrocardiographic abnormalities 12 (21.0) 12 (18.7) 12 (21.0)COPD 68 (14.6) 9 (14.1) 10 (17.5)Malnutrition 52 (11.1) 20 (31.2)† 14 (24.5)*Uncontrolled arterial hypertension 41 (8.8) 1 (1.5) 5 (8.8)Previous acute myocardial infarction 45 (9.6) 3 (4.7) 4 (7.0)Heart failure 25 (5.3) 11 (17.2)* 7 (12.2)*Alcoholism 24 (5.1) 5 (7.8) 10 (17.5)†Hemodynamic instability before surgery 8 (1.7) 21 (32.8)† 10 (17.5)†Previous cerebral vascular accident 24 (5.1) 6 (9.4) 7 (12.2)Liver failure 22 (4.7) 8 (12.5) 2 (3.5)Angina 11 (2.4) 2 (3.2) 5 (8.8)Chronic renal failure (need for RRT) 6 (1.3) 4 (6.3) 1 (6.1)

Surgery within 24 h (%)Unplanned surgery 107 (22.9) 45 (70.3)† 33 (57.9)†Major surgery 287 (61.6) 55 (85.9) 47 (82.4)Gastrointestinal surgery 82 (17.6) 16 (25.0)‡ 2 (3.5)Peritonitis 214 (45.9) 24 (53.1) 23 (40.3)Vascular surgery 116 (24.9) 13 (20.3) 8 (14.0)Exploratory laparotomy 17 (3.6) 25 (39.0)† 24 (42.1)†Orthopedic surgery 37 (7.9) 0 (0) 1 (1.7)Others 37 (7.9) 10 (15.6) 22 (38.6)†

Severity scores on admission, mean � SDMODS 3.1 � 2.7 6.0 � 3.1* 6.0 � 3.7*SOFA 4.2 � 3.5 8.1 � 4.5* 7.5 � 4.3*APACHE II 13.3 � 5.8 19.9 � 6.9*‡ 17.1 � 6.0*POSSUM 33.4 � 8.8 47.3 � 11* 44.2 � 12*

OutcomesEarly-onset sepsis (%) 17 (3.6) 22 (34.3)* 11 (19.3)*Late-onset sepsis (%) 30 (6.4) 26 (40.6)* 26 (45.6)*ICU LOS (d), median (IQ) 2 (1–3) 9 (3–15)* 7 (2–16)*Hospital LOS (d), median (IQ) 8 (4–18) 16.5 (8–29) 19 (11.5–46.5)*

ICU � intensive care unit; MOF � multiple organ failure; COPD � chronic obstructive pulmonary disease; RRT � renal replacement therapy; MODS � MultipleOrgan Dysfunction System; SOFA � Sequential Organ Failure Assessment; APACHE II � Acute Physiology and Chronic Health Disease Classification System II;POSSUM � Physiological and Operative Severity Score for the Enumeration of Mortality and Morbidity; LOS � length of stay; IQ � interquartile range.* P � 0.05 vs survivors.† P � 0.001 vs survivors.‡ P � 0.05 vs death due to other causes.

Table 2. Indices of Cardiorespiratory Dysfunctionand Tissue Hypoperfusion in Survivorsand Nonsurvivors

SurvivorsDeath due

to MOFDeath due toother causes

No. of patients 466 64 57pH 7.31 � 0.08 7.20 � 0.13*† 7.27 � 0.09*Serum lactate (mEq/L) 2.4 � 2.1 4.0 � 2.4* 3.4 � 1.9*Heart rate (beats/min) 97 � 19 121 � 24* 118 � 24*Maximal CVP (mm Hg) 10.2 � 7.5 18.2 � 7.7*† 14.5 � 8.2Minimal CVP (mm Hg) 6.4 � 5.4 11.2 � 6.7* 9.8 � 5.7*

MOF � multiple organ failure; CVP � central venous pressure.Data are mean � SD.* P � 0.05 vs survivors.† P � 0.05 vs death due to other causes.

April 2011 • Volume 112 • Number 4 www.anesthesia-analgesia.org 879

Characteristics and Outcomes of the PatientsWho Died Due to MOFDeath was due to MOF in 64 patients (53%) (Table 1).Patients who died due to MOF were more severely ill onICU admission (APACHE II, 19.9 � 6.9 vs 17.1 � 6.0, P �0.05) and more frequently died in the ICU (98.4% vs 43.8%,P � 0.05) than patients who died due to other causes. Lowfunctional capacity (41% vs 21%) and gastrointestinal sur-gery (25% vs 3.5%) were significantly more frequent inpatients who died due to MOF than in patients who dieddue to other causes (P � 0.05 for both). In contrast,alcoholism and other types of surgery were more frequentlyassociated with deaths due to other causes (Table 1). Early

onset sepsis was more prevalent in patients who died dueto MOF (34.3%) compared with patients who died of othercauses (19.3%) but the difference was not statisticallysignificant (P � 0.09).

On the day of ICU admission, the maximal CVP wassignificantly higher (18.2 � 7.7 vs 14.5 � 8.2 mm Hg, P �0.01) and the pH significantly lower (7.20 � 0.13 vs 7.27 �0.09, P � 0.05) in patients who died due to MOF than inpatients who died of other causes (Table 2).

The variables identified as independent predictors ofdeath due to MOF were age, unplanned surgery, diabetes,peritonitis, and, on the first day of ICU admission, highCVP (mm Hg), increased HR, increased serum lactateconcentrations (mEq/L), and pH (Fig. 1).

Comparison Between the Outcomes of thePatients to That Predicted by theSeverity ScoresThe POSSUM score gave a standardized mortality ratio of1.2 (degrees of freedom � 0.02, P � 0.001). The area underthe ROC curve for hospital mortality was 0.80 (95% confi-dence interval 0.775–0.840) for the POSSUM score. Theareas under the ROC curves for hospital mortality forAPACHE II, Multiple Organ Dysfunction System, andSequential Organ Failure Assessment scores were 0.808,0.802, and 0.805, respectively (Fig. 2).

DISCUSSIONDespite the large amount of resources directed at evaluat-ing the risk of perioperative cardiovascular complications,our results indicate that MOF is the main cause of death inhigh-risk surgical patients, deemed to be the cause of death

Figure 1. Relative risk (95% confidence interval) of death due to multiple organ failure. MOF � multiple organ failure; RR � relative risk; CI �confidence interval.

Figure 2. Receiver operating characteristic curves for the severityscores. APACHE II � Acute Physiology and Chronic Health DiseaseClassification System II; MODS � Multiple Organ Dysfunction Sys-tem; POSSUM � Physiological and Operative Severity Score for theEnumeration of Mortality and Morbidity; SOFA � Sequential OrganFailure Assessment.

Multiple Organ Failure After Noncardiac Surgery

880 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA

in 53% of our patients. MOF has recently been shown to bethe main cause of morbidity and mortality in patientsadmitted to ICUs, and has been calculated to account for upto 80% of ICU deaths.19

In an Australasian study, MOF was the cause of death in20.3% of patients admitted to the ICU with severe nonin-fectious systemic inflammatory response syndrome and in69% of patients with severe sepsis.20 In a large prospectivestudy performed in a population of mainly surgical criti-cally ill patients, acute or chronic MOF prevailed as thecause of death in the ICU.13 In a study on macroscopicpostmortem findings in surgical intensive care patientswith sepsis, the main causes of death as reported in thepatient history were refractory MOF in 51.5%.21

The multivariable analysis confirmed that tachycardia,high CVP and serum lactate levels, and pH on the day ofadmission to the ICU are early predictors of death due toMOF. Moreover, we found that unplanned surgery, perito-nitis, older age, and the presence of diabetes significantlyincreased the risk of death due to MOF.

Tachycardia, lactic acidosis, or acidosis due to othercauses, such as hyperchloremic acidosis, are common oc-currences after major surgery. These factors are undoubt-edly related to traumatic and long operations, with anenhanced systemic inflammatory response, inadequate re-suscitation, and tissue hypoperfusion. Our findings areconsistent with numerous studies that have demonstratedan association between organ hypoperfusion and indices oftissue trauma and organ dysfunction.22,23 Likewise, studiesin surgical ICU patients, and in patients with infection,sepsis, and shock have reported worse outcomes related tohigher serum lactate concentrations.24–28 Prolonged lactateclearance is related to increased mortality after surgery,and lactate nonclearance during resuscitation was a strongindependent predictor of in-hospital death in patients withsevere sepsis.29–31

Importantly, we found a 12% increase in the risk ofdeath due to MOF for each unit increase in CVP. Tradition-ally, cardiac filling pressures have been used to assessvolume status in critically ill patients. The observationaldesign of our study does not allow us to conclude whetherelevated CVP levels represented excessive intravascularvolume, poor cardiovascular reserve, or both. Moreover, inmore complex cases, the presence of external or intrinsicpositive end-expiratory pressure, abdominal hypertension,compromised left ventricular compliance, which is fre-quently decreased in ICU patients with sepsis, ischemic orhypertrophic cardiopathy, can increase CVP measurementseven in the presence of hypovolemia. Nevertheless, severalstudies have associated a positive fluid balance with com-plications and death in ICU patients.32

MOF development in patients imposes a heavy burdenon staff and resources, and patients have long ICU lengthsof stay and high costs. Awareness of early risk factors forMOF would be valuable if changes in clinical managementcould be prompted by potentially avoidable predictors ofpoor outcome. Several important clinical trials have docu-mented that early aggressive resuscitation using well-defined protocols, such as goal-directed therapy (GDT),improves outcomes and is cost effective.33–40 These studies

have used therapeutic strategies aimed at boosting cardio-respiratory function and maintaining end-organ perfusionthrough a more individualized and targeted fluid therapyand reported reductions in the length of ICU and hospitalstay, a faster recovery of gastrointestinal function, and areduction in mortality when GDT was performed in higherrisk surgical patients.41

The mortality rate in our study was higher than thatpredicted by the POSSUM score. All scoring systems thatwe evaluated had the same capacity of predicting hospitalmortality in this population of surgical patients admitted tothe ICU postoperatively. Unfortunately, in our country,early GDT, best known in this set of patients as optimiza-tion of oxygen delivery, is still not widely used in clinicalpractice despite the growing body of evidence. As a conse-quence of our findings and the high mortality observed inthis population of surgical patients, we recommend se-quential measurements of serum lactate, pH, and CVP to beused in the context of well-defined protocols of optimiza-tion of oxygen delivery to guide adjustments of IV fluidadministration and use of dobutamine to maintain a maxi-mal stroke volume in the perioperative period.34–39

Age has been reported to be independently associatedwith MOF in medical patients and in heterogeneous popu-lations of critically ill patients, with death in patients withsystemic inflammatory response syndrome and MOF, andwith post-ICU mortality in surgical ICU patients.12,20,42,43

Serum glucose levels were tested in the logistic regressionmodel but were not retained; however, a history of diabeteswas a strong predictor increasing the odds of MOF by 3.63.Diabetes mellitus is a chronic, systemic debilitating condi-tion that affects many organ functions, increases the likeli-hood and extent of coronary artery disease, and increasesthe likelihood of infections. Other studies have shown thatglucose levels and glucose variability, rather than diabetes,were important factors associated with organ dysfunc-tion.44,45 Hyperglycemia was independently associatedwith organ failure and death in critically ill children.46

Unplanned surgery increased the odds of death due toMOF by almost 4-fold. Indeed, in a prospective, observa-tional, Australian study, performed in 1125 subjects under-going surgery in a tertiary teaching hospital, unplannedICU admission increased the risk of death by 4-fold.47

Other studies have shown emergency surgery to be animportant predictor of mortality in older patients.48

Peritonitis was an independent predictor of death due toMOF. A total of 141 patients (24.0%) had peritonitis, of whom18.4% developed sepsis within 48 hours of ICU admission.When sepsis progresses to sepsis-associated organ failure andhypotension, mortality increases from 27% to �50% in pa-tients with septic shock.49 Severe sepsis occurred in 23% of thepatients in our cohort of high-risk surgical patients, with ahospital mortality rate of 64%. The nosocomial sepsis that is sotypical of the later course of such patients in the ICU can occuras a consequence of organ dysfunction, prolonging their stayand increasing the risk of infection.

This study is limited by the relatively small number ofpatients and ICUs. However, although the size of the studywas relatively modest, its selected study group is more likelyto represent real surgical practice. Despite the large size of ourcountry, all regions were represented. Our cohort was a

April 2011 • Volume 112 • Number 4 www.anesthesia-analgesia.org 881

population of high-risk surgical patients as demonstrated bythe frequent prevalence of comorbid diseases and unplannedsurgery, high degrees of physiologic derangement as shownby the high POSSUM and APACHE II scores, and thecorrespondingly high mortality rates. Bed shortages in Brazil-ian hospital ICUs may be the cause of the admission of higherrisk cases than in some other countries.

In conclusion, the vast majority of deaths in this high-risk population of surgical patients were due to MOF. Weidentified several routinely available variables as strongpredictors of the development of a fatal outcome due toMOF in our high-risk surgical ICU patients.

AUTHOR AFFILIATIONSFrom the *Intensive care Unit, Hospital de Base and SaoJose do Rio Preto Medical School; †Intensive Care Unit,Hospital do Servidor Publico Estadual Francisco Morato deOliveira, Sao Paulo; ‡Intensive Care Unit, Hospital SaoLucas and Hospital Cardiotrauma Ipanema, Rio de Janeiro;§Intensive Care Unit, Hospital Moinhos de Vento, PortoAlegre; �Intensive Care Unit, Clínica Sorocaba, Rio deJaneiro; ¶Intensive Care Unit, Clínica Sao Vicente, Rio deJaneiro; #Adult Intensive Care Unit, Hospital Universitarioda Universidade Federal da Paraíba, Joao Pessoa; **Depart-ment of Anesthesiology, Pain and Intensive Care, Univer-sidade Federal de Sao Paulo, Sao Paulo; ††Intensive CareUnit, Hospital Pro-Cardíaco, Rio de Janeiro; ‡‡IntensiveCare Department, Universidade Estadual de Londrina,Londrina; §§Intensive Care Unit, Hospital Universitario daUniversidade Federal do Mato Grosso do Sul, CampoGrande; ��Universidade Estadual do Piauí, Teresina; ¶¶In-tensive Care Unit, Hospital Santa Luzia, Brasília; ##MedicalSchool, Universidade Estadual do Oeste do Parana, Cas-cavel; ***Hospital do Servidor Publico Estadual FranciscoMorato de Oliveira, Sao Paulo; †††Intensive Care Unit,Santa Casa de Misericordia, Passos; ‡‡‡Medical School ofSao Jose do Rio Preto, Sao Jose do Rio Preto, Brazil; and§§§Department of Intensive Care, Erasme Hospital, FreeUniversity of Brussels, Brussels, Belgium.

AUTHOR CONTRIBUTIONSSML and ER helped in study design and manuscriptpreparation; MFK, NBS, JAP, CLM, FEN, MA, RCC, CCG,SFP, PMM, MOM, PAD, FG, and MRL helped in conduct ofstudy; JMS helped in data analysis and conduct of study;and JAC and CM helped in data analysis.

REFERENCES1. Jhanji S, Thomas B, Ely A, Watson D, Hinds CJ, Pearse RM.

Mortality and utilisation of critical care resources amongsthigh-risk surgical patients in a large NHS trust. Anaesthesia2002;63:695–700

2. Sauaia A, Moore FA, Moore EE, Lezotte DC. Early risk factorsfor post injury multiple organ failure. World J Surg1996;20:392–400

3. Shoemaker WC, Appel PL, Kram HB, Waxman K, Lee TS.Prospective trial of supranormal values of survivors as therapeu-tic goals in high-risk surgical patients. Chest 1998;94:1176–86

4. Older P, Smith R, Courtney P, Hone R. Preoperative evaluationof cardiac failure and ischemia in elderly patients by cardio-pulmonary exercise testing. Chest 1993;104:701–4

5. Kirchhoff P, Dincler S, Buchmann P. A multivariate analysis ofpotential risk factors for intra- and postoperative complicationsin 1316 elective laparoscopic colorectal procedures. Ann Surg2008;248:259–65

6. McCulloch P, Ward J, Tekkis PP; ASCOT group of surgeons;British Oesophago-Gastric Cancer group. Mortality and mor-bidity in gastro-oesophageal cancer surgery: initial results ofASCOT multicentre prospective cohort study. BMJ 2003;327:1192–7

7. Bianchi RC, de Souza JN, Giaciani CA, Hoehr NF, Toro IF.Prognostic factors for complications following pulmonary re-section: pre-albumin analysis, time on mechanical ventilation,and other factors. J Bras Pneumol 2006;32:489–94

8. Lee TH, Marcantonio ER, Mangione CM, Thomas EJ, Polanc-zyk CA, Cook EF, Sugarbaker DJ, Donaldson MC, Poss R, HoKK, Ludwig LE, Pedan A, Goldman L. Derivation and pro-spective validation of a simple index for prediction of cardiacrisk of major noncardiac surgery. Circulation 1999;100:1043–9

9. Eagle KA, Berger PB, Calkins H, Chaitman BR, Ewy GA,Fleischmann KE, Fleisher LA, Froehlich JB, Gusberg RJ, LeppoJA, Ryan T, Schlant RC, Winters WL Jr, Gibbons RJ, AntmanEM, Alpert JS, Faxon DP, Fuster V, Gregoratos G, Jacobs AK,Hiratzka LF, Russell RO, Smith SC Jr. ACC/AHA guidelineupdate for perioperative cardiovascular evaluation for noncar-diac surgery—executive summary. A report of the AmericanCollege of Cardiology/American Heart Association Task Forceon Practice Guidelines (Committee to Update the 1996 Guide-lines on Perioperative Cardiovascular Evaluation for Noncar-diac Surgery). Anesth Analg 2002;94:1052–64

10. Pinho C, Grandini PC, Gualandro DM, Calderaro D, Mona-chini M, Caramelli B. Multicenter study of perioperative evalu-ation for noncardiac surgeries in Brazil (EMAPO). Clinics2007;62:17–22

11. Davenport DL, Ferraris VA, Hosokawa P, Henderson WG,Khuri SF, Mentzer RM Jr. Multivariable predictors of postop-erative cardiac adverse events after general and vascularsurgery: results from the patient safety in surgery study. J AmColl Surg 2007;204:1199–210

12. Tran DD, Groeneveld AB, Van der Meulen J, Nauta JJ, Strackvan Schijndel RJ, Thijs LG. Age, chronic disease, sepsis, organsystem failure, and mortality in a medical intensive care unit.Crit Care Med 1990;118:474–9

13. Mayr VD, Dunser MW, Greil V, Jochberger S, Luckner G,Ulmer H, Friesenecker BE, Takala J, Hasibeder WR. Causes ofdeath and determinants of outcome in critically ill patients.Crit Care 2006;10:R154

14. Prytherch DR, Whiteley MS, Higgins B, Weaver PC, Prout WG,Powell SJ. POSSUM and Portsmouth POSSUM for predictingmortality. Physiological and Operative Severity Score for theenUmeration of Mortality and morbidity. Br J Surg 1998;85:1217–20

15. Knaus W, Draper E, Wagner D, Zimmerman JE. APACHE II: aseverity of disease classification system. Crit Care Med1985;13:818–29

16. Marshall JC, Cook DJ, Christou NV, Bernard GR, Sprung CL,Sibbald WJ. Multiple organ dysfunction score: a reliable de-scriptor of a complex clinical outcome. Crit Care Med1995;23:1638–52

17. Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A,Bruining H, Reinhart CK, Suter PM, Thijs LG. The SOFA(Sepsis-related Organ Failure Assessment) score to describeorgan dysfunction/failure. On behalf of the Working group onSepsis-Related Problems of the European Society of IntensiveCare Medicine. Intensive Care Med 1996;22:707–10

18. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, KnausWA, Schein RM, Sibbald WJ. Definitions for sepsis and organfailure and guidelines for the use of innovative therapies insepsis. The ACCP/SCCM Consensus Conference Committee.American College of Chest Physicians/Society of Critical CareMedicine. Chest 1992;101:1644–55

19. Baue AE, Durham R, Faist E. Systemic inflammatory responsesyndrome (SIRS), multiple organ dysfunction syndrome(MODS), multiple organ failure (MOF). Are we winning thebattle? Shock 1998;10:79–89

20. Dulhunty JM, Lipman J, Finfer S. Does severe non-infectiousSIRS differ from severe sepsis? Results from a multi-centreAustralian and New Zealand intensive care unit study. Inten-sive Care Med 2008;34:1654–61

Multiple Organ Failure After Noncardiac Surgery

882 www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA

21. Torgersen C, Moser P, Luckner G, Mayr V, Jochberger S,Hasibeder WR, Dunser MW. Macroscopic postmortem find-ings in 235 surgical intensive care patients with sepsis. AnesthAnalg 2009;108:1841–7

22. Bennett-Guerrero E, Welsby J, Dunn TJ, Young LR, Wahl TA,Diers TL, Phillips-Bute BG, Newman MF, Mythen MG. The useof a postoperative morbidity survey to evaluate patients withprolonged hospitalization after routine, moderate-risk, electivesurgery. Anesth Analg 1999;89:514–9

23. Bakker J, Gris P, Coffernils M, Kahn RJ, Vincent JL. Serial bloodlactate levels can predict the development of multiple organfailure following septic shock. Am J Surg 1996;171:221–6

24. Smith I, Kumar P, Molly S, Rhodes A, Newman PJ, GroundsRM, Bennett ED. Base excess and lactate as prognostic indica-tors for patients admitted to intensive care. Intensive Care Med2001;27:74–83

25. Meregalli A, Oliveira RP, Friedman G. Occult hypoperfusion isassociated with increased mortality in hemodynamicallystable, high-risk, surgical patients. Crit Care 2004;8:R60–5

26. Shapiro NI, Howell MD, Talmor D, Nathanson LA, Lisbon A,Wolfe RE, Weiss JW. Serum lactate as a predictor of mortalityin emergency department patients with infection. Ann EmergMed 2005;45:524–8

27. Bernardin G, Pradier C, Tiger F, Deloffre P, Mattei M. Bloodpressure and arterial lactate level are early indicators ofshort-term survival in human septic shock. Intensive Care Med1996;22:17–25

28. Weil MH, Afifi AA. Experimental and clinical studies onlactate and pyruvate as indicators of the severity of acutecirculatory failure (shock). Circulation 1970;41:989–1001

29. McNelis J, Marini CP, Jurkiewicz A, Szomstein S, Simms HH,Ritter G, Nathan IM. Prolonged lactate clearance is associatedwith increased mortality in the surgical intensive care unit.Am J Surg 2001;182:481–5

30. Husain FA, Martin MJ, Mullanix PS, Steele SR, Elliot DC.Serum lactate and base deficit as predictors of mortality andmorbidity. Am J Surg 2003;185:485–91

31. Arnold RC, Shapiro NI, Jones AE, Schorr C, Pope J, Casner E,Parrillo JE, Dellinger RP, Trzeciak S. Multi-center study ofearly lactate clearance as a determinant of survival in patientswith presumed sepsis. Shock 2009;32:35–9

32. Payen D, de Pont AC, Sakr Y, Spies C, Reinhart K, Vincent JLA.Positive fluid balance is associated with a worse outcome inpatients with acute renal failure. Crit Care 2008;12:R74

33. Mythen MG, Webb AR. Perioperative plasma volume expan-sion reduces the incidence of gut mucosal hypoperfusionduring cardiac surgery. Arch Surg 1995;130:423–9

34. Gan TJ, Soppitt A, Maroof M, el-Moalem H, Robertson KM,Moretti E, Dwane P, Glass PS. Goal-directed intraoperativefluid administration reduces length of hospital stay after majorsurgery. Anesthesiology 2002;97:820–6

35. Venn R, Steele A, Richardson P, Poloniecki J, Grounds M,Newman P. Randomized controlled trial to investigate influ-ence of the fluid challenge on duration of hospital stay andperioperative morbidity in patients with hip fractures. Br JAnaesth 2002;88:65–71

36. McKendry M, McGloin H, Saberi D, Caudwell L, Brady AR,Singer M. Randomised controlled trial assessing the impact ofa nurse delivered, flow monitored protocol for optimisation ofcirculatory status after cardiac surgery. BMJ 2004;329:258

37. Polenen P, Ruokonen E, Hippelainen M, Poyhonen M, TakalaJ. A prospective randomized study of goal-oriented hemody-namic therapy in cardiac surgical patients. Anesth Analg2000;90:1052–9

38. Conway DH, Mayall R, Abdul-Latif MS, Gilligan S, TackaberryC. Randomised controlled trial investigating the influence ofintravenous fluid titration using oesophageal Doppler moni-toring during bowel surgery. Anaesthesia 2002;57:845–9

39. Lobo SM, Salgado PF, Castillo VG, Borim AA, Polachini CA,Palchetti JC, Brienzi SL, de Oliveira GG. Effects of maximizingoxygen delivery on morbidity and mortality in high-risksurgical patients. Crit Care Med 2000;28:3396–404

40. Fenwick E, Wilson J, Sculpher M, Claxton K. Pre-operativeoptimisation employing dopexamine or adrenaline for patientsundergoing major elective surgery: a cost-effectiveness analy-sis. Intensive Care Med 2002;28:599–608

41. Kern JW, Shoemaker WC. Meta-analysis of hemodynamicoptimization in high-risk patients. Crit Care Med 2002;30:1686–92

42. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, CarcilloJ, Pinsky MR. Epidemiology of severe sepsis in the UnitedStates: analysis of incidence, outcome, and associated costs ofcare. Crit Care Med 2001;29:1303–10

43. Sakr Y, Vincent JL, Ruokonen E, Pizzamiglio M, Installe E,Reinhart K, Moreno R. Sepsis and organ system failure aremajor determinants of post-intensive care unit mortality. J CritCare 2008;23:475–83

44. Sperry JL, Frankel HL, Vanek SL, Nathens AB, Moore EE,Maier RV, Minei JP. Early hyperglycemia predicts multipleorgan failure and mortality but not infection. J Trauma2007;63:487–93

45. Ali NA, O’Brien JM Jr, Dungan K, Phillips G, Marsh CB,Lemeshow S, Connors AF Jr, Preiser JC. Glucose variabilityand mortality in patients with sepsis. Crit Care Med2008;368:2316–21

46. Yung M, Wilkins B, Norton L, Slater A. Glucose control, organfailure, and mortality in pediatric intensive care. Pediatr CritCare Med 2008;9:147–52

47. Bellomo R, Goldsmith D, Russell S, Uchino S. Postoperativeserious adverse events in a teaching hospital: a prospectivestudy. Med J Aust 2002;176:216–8

48. Hamel MB, Henderson M, Khuri SF, Daley J. Surgical out-comes for patients aged 80 and older: morbidity and mortalityfrom major noncardiac surgery. J Am Geriatr Soc 2005;53:424–9

49. Martin GS, Mannino DM, Eaton SM, Moss M. The epidemiol-ogy of sepsis in the United States from 1979 through 2000.N Engl J Med 2003;348:1546–54

April 2011 • Volume 112 • Number 4 www.anesthesia-analgesia.org 883