Short-Term Prognosis in Critically Ill Patients With Cirrhosis Assessed by Prognostic Scoring Systems

Short-Term Prognosis in Critically Ill Patients With Cirrhosis Assessed by Prognostic Scoring Systems MARKUS WEHLER, 1 JUDITH KOKOSKA, 1 UDO REULBACH, 2 ECKHART GEORG HAHN, 1 AND RICHARD STRAUSS 1 The short-term prognosis of acutely ill patients with cirrhosis is influenced by the degree of hepatic insufficiency and by dysfunction of extrahepatic organ systems. The purpose of this study was to assess and compare the prognostic accuracy of the Child-Pugh classification, the Acute Physiology and Chronic Health Evaluation (APACHE) II system and the Sequential Organ Failure Assessment (SOFA) for predicting hospital mortality in patients with cirrhosis when used 24 hours after admission to a medical intensive care unit (ICU). Prospective data were recorded on 143 patients. Cumulative mortality rates were 36% in the ICU, 46% in the hospital, and 56% at 6-month follow-up. By using the area under receiver operating characteristic (AUROC) curves, the SOFA showed an excellent discriminative power (AU- ROC 0.94), which was clearly superior to the APACHE II (AUROC 0.79) and the Child-Pugh system (AUROC 0.74). Hospital mortality rates below and above a cutoff of 8 SOFA points were 4% and 88%, respectively (P <.0005). The SOFA score also reflected resource use during the ICU treatment as measured by daily workload and length of stay. The SOFA is an easily applied tool with excellent prognostic abilities and can be used to enhance clinical judgment of prognosis as well as providing patients and families with objective information. (HEPATOLOGY 2001;34:255-261.) Chronic liver disease, including cirrhosis, was the tenth most frequent cause of death in the United States in 1996. 1 Because end-stage liver disease is characterized by episodes of liver decompensation and gastrointestinal hemorrhage, treatment in the intensive care unit (ICU) is often needed in the management of these complications. In patients with cirrhosis who develop extrahepatic organ failure, hospital mortality rates of 63% to 100% have been reported. 2-7 For ethical reasons but also because of limited resources and growing questions about the efficacy of intensive care treatment, physicians need early and reliable Abbreviations: ICU, intensive care unit; APACHE, Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment; TISS, therapeutic intervention scoring system; AUROC, area under receiver operating characteristic; 95% CI, 95% confidence interval. From the 1 Departments of Medicine I and 2 Medical Informatics, Biometry, and Epidemiology of the University of Erlangen-Nuremberg, Germany. Received February 26, 2001; accepted May 29, 2001. Address reprint requests to: Markus Wehler, M.D., Department of Medicine I of the University of Erlangen-Nuremberg, Krankenhausstrasse 12, 91054 Erlangen, Germany. E-mail: markus.wehler@med1.imed.uni-erlangen.de; fax: (49) 9131-853 6012. Copyright 2001 by the American Association for the Study of Liver Diseases. 0270-9139/01/3402-0007$35.00/0 doi:10.1053/jhep.2001.26522 255 outcome predictors to identify cases in which aggressive treatment for cure or potential liver transplantation is merited, as well as those in which such care is likely to be futile. 8,9 The Child-Turcotte classification, and its subsequent modification by Pugh, 10 is one of the most commonly used clinical instruments to risk-stratify cirrhotic patients. 11,12 The Acute Physiology and Chronic Health Evaluation (APACHE) II, 13 a physiologically based prognostic system, is a widely used severity-of-illness scale to predict hospital mortality in all ICU patients. More recently the Sequential Organ Failure Assessment (SOFA) 14 has been developed and validated. Although originally an instrument to classify various degrees of organ dysfunction it has been shown that high SOFA scores for any individual organ are associated with increased mortality. 15,16 We analyzed and compared the prognostic accuracy of these 3 scoring systems in predicting hospital mortality of cirrhotic patients when scores were measured 24 hours after admission to a medical ICU. PATIENTS AND METHODS This study was performed in a 12-bed noncoronary medical ICU of a university hospital between October 1996 and December 1998. A total of 143 consecutive patients with hepatic cirrhosis requiring intensive monitoring and/or treatment that could not be provided outside of the ICU were enrolled. Eight cirrhotic patients with known end-stage cancer unresponsive to therapy were not admitted to the ICU. Readmissions were also excluded. Prospectively collected data included demographics, reason for ICU admission, acute diagnosis, presence of comorbid disease, severity of illness, organ function, therapeutic activity on all ICU days, and length of ICU and hospital stay. Comorbid illnesses were classified using the Charlson comorbidity index, which assigns weights for 19 major comorbid conditions affecting the patient s prognosis. 17 The severity of liver disease on ICU admission was graded by the Child-Pugh system. 10 The severity of illness was assessed by the APACHE II system after the first 24 hours of ICU admission. The APACHE II system is based on age, chronic health status, and acute physiology score; the variables included in the acute physiology scoring are mean arterial pressure, temperature, respiratory rate, heart rate, Glasgow Coma Score, hematocrit, white blood cell count, serum sodium, serum potassium, serum creatinine, serum bicarbonate or arterial ph, and arterial oxygen tension or alveolar-arterial oxygen tension gradient (Appendix Table). 13 Organ function was evaluated by the SOFA. 14 The SOFA is composed of scores from 6 organ systems (respiratory, cardiovascular, hepatic, renal, coagulation, and neurologic) graded from 0 to 4 points according to normal function or the degree of dysfunction (Table 1). In the first 24-hour period of ICU admission the most abnormal value for each of the 6 organ systems was recorded, and points were assigned for each organ system. The SOFA score presents the sum of all 6 organ system points. Organ failure was defined by a SOFA score of 3 or more

256 WEHLER ET AL. HEPATOLOGY August 2001 TABLE 1. The Sequential Organ Failure Assessment (SOFA) Score SOFA Score 0 1 2 3 4 Respiration 400 301-400 201-300 101-200 100 PaO 2 /FiO 2 with respiratory support with respiratory support Coagulation Platelets, 10 3 /mm 3 150 101-150 51-100 21-50 20 Liver Bilirubin, mg/dl 1.2 1.2-1.9 2.0-5.9 6.0-11.9 12.0 ( mol/l) ( 20) (20-32) (33-101) (102-204) ( 204) Cardiovascular Hypotension MAP 70 mm Hg MAP 70 mm Hg Dopamine 5or dobutamine (any dose)* Dopamine 5or epi 0.1 or norepi 0.1* Dopamine 15 or epi 0.1 or norepi 0.1* CNS Glasgow Coma Score 15 13-14 10-12 6-9 6 Renal Creatinine, mg/dl 1.2 1.2-1.9 2.0-3.4 3.5-4.9 5.0 ( mol/l) ( 110) (110-170) (171-299) (300-440) ( 440) or urine output or 500 ml/d or 200 ml/d Abbreviations: PaO 2, arterial oxygen tension; FiO 2, fractional inspired oxygen; MAP, mean arterial pressure; CNS, central nervous system; epi, epinephrine; norepi, norepinephrine. *Adrenergic agents administered for at least 1 hour (doses are given in g/kg/min). Reprinted with permission. 14 points for the respective organ system. 15 Therapeutic activity was determined by the Therapeutic Intervention Scoring System (TISS), a weighted score that quantitates the use of 76 diagnostic, therapeutic, and monitoring procedures commonly performed in the ICU. 18 The principal study outcome was hospital mortality rate. Follow-up 6 months after hospital discharge was performed by telephone interview. If necessary, the registry office provided information as to patient survival or date of death. The diagnosis of hepatic cirrhosis was made histologically (n 37) or by clinical evaluation (n 106). The clincal diagnosis of cirrhosis was made by a history of portal hypertension excluding other etiologies, impaired liver function tests and clotting profile, and ultrasound or computer tomography criteria. 19 The Glasgow Coma Scale as the neurologic component of the APACHE II and SOFA was scored conservatively (for sedated patients normal function is assumed unless there is evidence for intrinsically altered neurologic function). The criteria for diagnosing severe sepsis were as previously defined. 20 The Institutional Ethics Committee approved the study and because of the observational nature of the study waived the need for informed consent. Statistical Methods. Descriptive statistics are expressed as mean SD unless otherwise stated. The primary analysis compared hospital survivors with nonsurvivors. All variables were tested for normal distribution by the Kolmogorov-Smirnov test. Student s t test was used for comparison of the means of continuous variables and normally distributed data. Mann-Whitney U test was used otherwise. Categorical data were tested using the 2 statistic. Sensitivity, specificity, overall correctness, and positive and negative predictive values were determined for the Child- Pugh, APACHE II, and SOFA scores. Cutoff points giving the best Youden index (sensitivity specificity 1) were calculated. 21 Survival percentages of patients dichotomized by the SOFA score with the best Youden index were described by the Kaplan-Meier method. 22 The ability of the scoring systems to discriminate between hospital survivors and nonsurvivors was assessed by using the area under the receiver operating characteristic (AUROC) curve. 23 The effect of the various organ systems as assessed by SOFA after the first 24 hours on the risk of hospital death was evaluated using a Cox proportional hazards nonstepwise regression analysis. 24 Resource use among patients with different numbers of organ system failures was studied by examining the median TISS score of all ICU days and ICU length of stay with the Kruskal-Wallis test. All statistical tests were 2-tailed, and a significance level of P.05 or less was used. Data were analyzed using SPSS for Windows 10 (SPSS Inc., Chicago, IL). RESULTS A total of 143 patients with hepatic cirrhosis were enrolled in the study. Fifty-two percent of the patients were admitted from the ward, 38% from the emergency room, and 10% from other hospitals. Demographic and clinical characteristics of the study patients are shown in Table 2. The median age was 53 years, 62% were men, and all patients were white. Liver disease was most commonly attributed to alcohol abuse, and most patients had 2 or more other comorbid conditions. The most common reason for ICU admission was upper gastrointestinal bleeding (42%) (Table 3). Management of the patients was time consuming and resource intensive (Table 2). Seventy-four (52%) patients needed mechanical ventilation (25% hospital survivors vs. 83% nonsurvivors, P.0005), 70 (49%) were treated with vasopressors (22% vs. 80%, P.0005), and 32 (22%) required renal replacement therapy (8% vs. 40%, P.0005). During the ICU stay 187 endoscopic procedures (66% gastrointestinal) were performed and transfusion requirements were high (total 2,450 units of blood products). Follow-up to 180 days or time of death was complete for the entire cohort. The cumulative incidence of death was 36% in the ICU, 46% in the hospital, and 56% at 6 months. During the first 24 hours of ICU admission 67% of the patients presented with 1 or more organ system failures. The number of organ systems failing was significantly related to hospital mortality, with mortality rates ranging from 6.4% in patients without any organ system failure to 96.7% in patients with 3 or more organ system failures (Fig. 1). To evaluate to what extent the applied scoring systems were valid for prediction of hospital mortality; the sensitivity, spec-

HEPATOLOGY Vol. 34, No. 2, 2001 WEHLER ET AL. 257 TABLE 2. Demographic and Clinical Characteristics of 143 Patients With Cirrhosis Admitted to a Medical ICU All Patients (n 143) Hospital Survivors (n 77) Hospital Nonsurvivors (n 66) P Gender (F/M) 54/89 29/48 25/41 NS Age (years SD) 53 11.6 52 11.8 54 11.3 NS Length of ICU stay (days SD) 8.8 13.2 Median 5 7.6 15.6 Median 4 10.3 9.6 Median 8.0005 Length of hospital stay (days SD) 26.6 28.5 Median 21 31.7 34 Median 24 20.6 19 Median 15.02 Charlson Comorbidity Index 3.9 1.2 3.8 1.1 4.1 1.3 NS Child-Pugh category (A/B/C; n) 6/40/97 6/30/41 0/10/56.0005 Child-Pugh points (mean SD) 10.1 2 9.35 1.9 10.9 1.8.0005 APACHE II (mean SD), first 24 h 20.6 10.7 15.5 8 26.6 10.5.0005 SOFA (mean SD), first 24 h 8.6 4.7 5.2 2.7 12.6 3.3.0005 Median TISS score ( SD)* 31 13 24 9 39 12.0005 Mechanical ventilation (days SD) 5.4 11.9 3.4 13.5 7.8 9.2.025 Renal replacement (days SD) 1.5 4.6 0.7 4.4 2.5 4.6.023 Vasopressor therapy (days SD) 3.9 7.2 1.6 6.6 6.5 7.0.0005 Packed red blood cells 6.3 8.2 3.9 5.4 9 9.8.0005 Single donor apheresis platelets 1.4 2.8 0.6 1.8 2.4 3.4.0005 Fresh frozen plasma infused during ICU stay (units SD) 9.45 16.5 6.7 3.3 16.7 21.1.0005 Abbreviation: NS, not significant. *Median TISS score of all days the patient was treated in the ICU. Continuous veno-venous hemofiltration. Intravenous epinephrine or norepinephrine. ificity, overall correctness of prediction, and positive and negative predictive values were all determined. Table 4 shows these data calculated at the cutoff point giving the best Youden index. The best Youden index and highest overall correctness of prediction was found for the SOFA score. ICU and hospital mortality rates below and above a cutoff of 8 SOFA points were 3% and 71%, and 4% and 88%, respectively (P.0005 for both). Figure 2 shows the cumulative rates of survival during hospitalization for the study group dichotomized by a cutoff of 8 SOFA points. TABLE 3. Causes of Cirrhosis and Reasons for Admission of 143 Patients Admitted to the Medical ICU All Patients n Hospital Survivors n * Hospital Nonsurvivors n * P Receiver operating characteristic curves were used to evaluate the discriminative power of the scores. The SOFA score calculated 24 hours after ICU admission was found to be the most reliable scoring system to discriminate between hospital survivors and nonsurvivors (AUROC 0.947, SE 0.02, 95% confidence interval [95% CI] 0.91-0.987) (Fig. 3). No differences were seen between the APACHE II (AU- ROC 0.79, SE 0.04, 95% CI 0.72-0.87) and the Child-Pugh score (AUROC 0.74, SE 0.04, 95% CI 0.66-0.82). When calculated for ICU outcome AUROC data did not change significantly: SOFA (AUROC 0.92, SE 0.02, 95% CI 0.87-0.96), APACHE II (AUROC 0.79, SE 0.04, 95% CI 0.72- Cause of cirrhosis Alcoholic 108 (75) 52 (48) 56 (52).019 Posthepatitic 20 (14) 15 (75) 5 (25) NS Cryptogenic 5 (4) 4 (80) 1 (20) NS Other causes 10 (7) 6 (60) 4 (40) NS Primary ICU admission cause UGI bleeding 60 (42) 41 (68) 19 (32).001 Severe sepsis 20 (14) 3 (15) 17 (85).0005 Acute renal failure 19 (13) 11 (58) 8 (42) NS Hepatic coma 18 (13) 7 (39) 11 (61) NS Pneumonia with respiratory failure 17 (12) 7 (41) 10 (59) NS Drug/alcohol 6 (4) 6 (100) 0 (0).03 overdose Acute pancreatitis 3 (2) 2 (67) 1 (33) NS Abbreviations: UGI, upper gastrointestinal; NS, not significant. *Number of patients with the condition who survived or died. Primary biliary cirrhosis, secondary biliary cirrhosis, autoimmune hepatitis. FIG. 1. Numbers of organ systems failing on the first day of ICU admission vs. hospital mortality in 143 patients with cirrhosis ( 2 test for trend 72, P.0005). Organ failure defined by a SOFA score of 3 or more points for the respective organ system.

258 WEHLER ET AL. HEPATOLOGY August 2001 TABLE 4. Prediction of Subsequent Hospital Mortality in 143 Cirrhotic Patients After the First Day of ICU Admission Scoring System Cutoff Point* Youden Index Sensitivity Specificity Overall Correctness PPV NPV Child-Pugh 10 0.47 66 80 74 74 74 APACHE II 22 0.51 66 84 76 78 75 SOFA 8 0.84 95 88 91 87 96 Abbreviations: PPV, positive predictive value; NPV, negative predictive value. *Value giving the best Youden index. 0.87) and Child-Pugh score (AUROC 0.73, SE 0.04, 95% CI 0.64-0.82). To evaluate the relative contribution to hospital outcome of each of the 6 organ systems comprising the SOFA, a Cox proportional hazards analysis was performed. The exponent of the estimated coefficient for each organ represents the factor by which the relative risk of hospital death changes when the score for that particular organ increases 1 point. This analysis showed that SOFA scores for the cardiovascular, neurologic, renal, and hepatic system were significantly related to the risk of hospital death (Table 5). Results remained unchanged when ICU mortality was used as the dependent variable. Resource use, as measured by median TISS of all ICU days and median length of ICU stay, was directly proportional to the number of organ failures assessed by the SOFA (Table 6). The total SOFA after the first 24 hours showed a good correlation with median TISS of all ICU days (Pearson s r 0.7, P.01). Moreover, because of the declining survival with the increasing number of organ failures the mean cumulative TISS points for patients discharged alive increased 94-fold from 136 TISS points for patients with no organ failure to 12,775 TISS points for patients with 3 or more organ failures. ically ill cirrhotic patients. The overall predictive accuracy of the SOFA was 15% and 17% greater than that of the APACHE II and Child-Pugh systems, respectively. Several past studies analyzed the predictive abilities of prognostic systems on the short-term mortality of cirrhotic patients. Zimmerman et al. 5 showed that the APACHE III system accurately risk stratifies critically ill cirrhotic patients requiring mechanical ventilation. Zauner et al. 25 retrospectively compared the prognostic abilities of liver-specific and general scoring systems in 198 ICU patients with cirrhosis. They found that the APACHE III system was the most accurate prognostic system (AUROC at admission 0.78, after 48 hours 0.8) and that a cutoff point of 80 had an overall correct prediction of 70%. 25 The prognostic value of the APACHE II system (AUROC at admission 0.69, after 48 hours 0.78) was as good as the liver-specific scoring systems (i.e. Mayo Risk Score AU- ROC at admission 0.72, after 48 hours 0.66, Composite Clinical and Laboratory Index Scoring AUROC at admission 0.68, after 48 hours 0.68). Cooper et al. 26 analyzed a cohort of 538 patients with decompensated end-stage liver disease enrolled in the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT). They developed a multivariable prognostic DISCUSSION In this study we have shown that the SOFA is an excellent prognostic system for predicting in-hospital mortality in crit- FIG. 2. Cumulative survival in 143 patients with cirrhosis according to their SOFA (solid curve, SOFA 8 points, n 71; dashed curve, SOFA 9 points, n 72) after the first day of admission to a medical intensive care unit. Day 0 is the day of intensive care unit admission. FIG. 3. Receiver operating characteristic curves for SOFA (, AUROC 0.94), APACHE II (, AUROC 0.79), and Child-Pugh points (Œ, AUROC 0.74) recorded after the first day of intensive care unit admission.

HEPATOLOGY Vol. 34, No. 2, 2001 WEHLER ET AL. 259 TABLE 5. Cox Proportional Hazards Regression Analysis Showing the Effect of Different Organ Systems on the Risk of Hospital Death Variable SE P Risk Ratio (95% CI) Cardiovascular 0.53 0.13.0005 1.69 (1.31-2.17) Neurologic 0.37 0.11.001 1.45 (1.16-1.82) Renal 0.32 0.11.004 1.38 (1.1-1.71) Hepatic 0.27 0.1.01 1.3 (1.06-1.6) Respiratory 0.08 0.11.45 1.09 (0.87-1.35) Coagulation 0.06 0.1.57 1.06 (0.87-1.29) NOTE. Organ systems assessed by SOFA 24 hours after ICU admission., coefficient. Risk ratios are presented for a 1-point change in the scores for each organ. model, which included renal insufficiency, Glasgow Coma Score, ventilatory insufficiency, age 65 years, and prolonged prothrombin time, that had a AUROC of 0.76 and 0.74 for death at 30 and 180 days, respectively. 26 Butt et al. 27 prospectively compared the prognostic accuracy of the Child-Pugh and the APACHE III systems in predicting hospital mortality in 282 general ward patients with cirrhosis. The overall correctness of prediction values were 75% and 67% for the APACHE III and Child-Pugh systems, respectively. Afessa and Kubilis 28 compared the prognostic performances of APACHE II and Child-Pugh score in 111 cirrhotic patients hospitalized for upper gastrointestinal bleeding and did not find significant differences between the 2 scoring systems (APACHE II AUROC at admission 0.78; Child-Pugh 0.76). Recently Zauner et al. 7 developed a new scoring system for predicting hospital mortality in critically ill cirrhotic patients. Their Intensive Care Cirrhosis Outcome score is based on a multiple logistic regression analysis including bilirubin, cholesterol, creatinine clearance, and lactate. In a study group of 70 cirrhotic patients the Intensive Care Cirrhosis Outcome score showed a high discriminative ability in predicting hospital mortality (AU- ROC 0.90) and had an overall correctness of prediction of 83%. 7 The predictive accuracies we found for the APACHE II and Child-Pugh systems in our sample compare well with data reported by Zauner et al. 25 and Afessa and Kubilis. 27 The discriminatory power of the SOFA score seems to be superior to previously evaluated prognostic systems in cirrhotic patients. The SOFA score also reflected resource use during the ICU treatment as measured by daily TISS and ICU length of stay. Resource use increased tremendously with growing organ dysfunction, and intense therapy was not associated with improved outcome. Similar findings were noted by Cooper et al. in their group of patients with decompensated end-stage liver disease. 26 For any predictive model to be clinically useful, it must show ease of use, accuracy, acceptance by the data collecting staff, and reproducibility. 29 The SOFA system has already been used sucessfully in large numbers of critically ill patients. 15,16,30,31 The variables needed to record the SOFA are derived from standard monitoring of critically ill patients and calculation at the bedside takes 3 minutes. 31 Although the SOFA was not developed to predict outcome but to describe the degree of organ dysfunction in critically ill patients, several recent studies showed its predictive ability not only in patients with sepsis but also in trauma patients and in medical cardiovascular patients. 14,15,30,31 However, there are some limitations of the SOFA that should be addressed. Serum bilirubin as the descriptor of liver function lacks specificity, it has limited ability to reflect the full spectrum of liver dysfunction in critical illness and cannot differentiate acute liver dysfunction from the effects of preexisting chronic disease. 32 The Glasgow Coma Scale as the variable reflecting neurologic dysfunction also has shortcomings. Its clinical evaluation is subject to expectation bias and random error and it is affected by sedative and analgesic drugs frequently used in critically ill patients. Nevertheless, besides the cardiovascular and renal organ systems, the hepatic and neurologic systems of the SOFA were also associated with an increased risk of death. These findings are in agreement with those of previous studies on the SOFA, which implicated the cardiovascular, the neurologic, and the renal systems in increasing the risk of death. 15,16,31 In these investigations no effect of the hepatic system was noted, but only a minority of their study patients had severe hyperbilirubinemia during the first ICU day. Interestingly, several recent studies analyzing the effect of different organ systems on outcome find no 15,31,33 or little effect (risk ratio 1.17) 16 associated with pulmonary dysfunction. Maybe this finding is related to improvements in respiratory support and the observation that refractory hypoxemia is an unusual cause of death. 33,34 Potential limitations of our study should also be mentioned. First, it was performed at an academic referral hospital, therefore our results may not be applicable to institutions with different patient populations. Second, it should be noted, that we tested the prognostic instruments in patients already admitted to intensive care and not as a preadmission screening tool. 8 Finally, we would like to stress that all prognostic systems are not readily applicable to individual patients because of the fact that scoring systems are developed to predict group outcomes. 35 In conclusion, our data show that the discriminatory power of the SOFA to predict short-term mortality in critically ill patients with cirrhosis is clearly superior to the APACHE II and Child-Pugh systems. SOFA also seems to be superior to previously evaluated prognostic systems. Prognostic scoring systems cannot replace the clinical evaluation of the patient, which includes an appreciation not only of the overall prognosis but also of the patient s views. However, we believe that the SOFA may improve the phy- TABLE 6. Resource Use for 143 Patients With Cirrhosis According to the Number of Organ Failures Assessed After the First Day of ICU Admission Number of organ failures 0 1 2 3 Number of patients 47 36 30 30 Patients alive at hospital discharge/ 6-month follow-up 94/72 67/67 27/13 3/3 Median SOFA score 4 7 11 15.5 Median TISS score* 22 28 38 45 Median length of ICU stay (days) 3 5.5 7.5 7.5 NOTE. Organ failure is defined as 3 or more SOFA points for the respective organ system. *Median TISS points of all days treated on the ICU. P.0005 by Kruskal-Wallis test.

260 WEHLER ET AL. HEPATOLOGY August 2001 APPENDIX The APACHE II Severity of Disease Classification System High Abnormal Range Low Abnormal Range Physiologic Variable 4 3 2 1 0 1 2 3 4 Temperature rectal ( C) 41 39-40.9 38.5-38.9 36-38.4 34-35.9 32-33.9 30-31.9 29.9 Mean arterial pressure (mm Hg) 160 130-159 110-129 70-109 50-69 49 Heart rate (ventricular response) 180 140-179 110-139 70-109 55-69 40-54 39 Respiratory rate (nonventilated or ventilated) 50 35-49 25-34 12-24 10-11 6-9 5 Oxygenation AaDO 2 or PaO 2 (mm Hg) FiO 2 0.5 record 500 350-499 200-349 200 AaDO 2 FiO 2 0.5 record only PO 2 PO 2 PO 2 PO 2 PaO 2 70 61-70 55-60 55 Arterial ph 7.7 7.6-7.69 7.5-7.59 7.33-7.49 7.25-7.32 7.15-7.24 7.15 Serum sodium (mmol/l) 180 160-179 155-159 150-154 130-149 120-129 111-119 110 Serum potassium (mmol/l) 7 6-6.9 5.5-5.9 3.5-5.4 3-3.4 2.5-2.9 2.5 Serum creatinine (mg/100 ml) (Double point 3.5 2-3.4 1.5-1.9 0.6-1.4 0.6 score for acute renal failure) Hematocrit 60 50-59.9 46-49.9 30-45.9 20-29.9 20 White blood count (total/mm 3 ) (in 1,000s) 40 20-39.9 15-19.9 3-14.9 1-2.9 1 Glasgow Coma Score (GCS), score 15 minus actual GCS A Total Acute Physiology Score (APS) Sum of the 12 individual variable points Serum HCO 3 (venous mmol/l) (Not preferred, use if no ABGs) 52 41-51.9 32-40.9 22-31.9 18-21.9 15-17.9 15 B C Age points Assign points to age as follows: Age (years) Points 44 0 45-54 2 55-64 3 65-74 5 75 6 Chronic Health Points If the patient has a history of severe organ system insufficiency or is immunocompromised assign points as follows: (a) for nonoperative or emergency postoperative patients, 5 points; (b) for elective postoperative patients, 2 points. Definitions Organ insufficiency or an immunocompromised state must have been evident prior to this hospital admission and conform to the following criteria: Liver: Biopsy-proven and documented portal hypertension; episodes of past upper GI bleeding attributed to portal hypertension; or prior episodes of hepatic failure/encephalopathy/coma. Cardiovascular: New York Heart Association Class IV. Respiratory: Chronic restrictive, obstructive, or vascular disease resulting in severe exercise restriction, i.e., unable to climb stairs or perform household duties; or documented chronic hypoxia, hypercapnia, secondary polycythemia, severe pulmonary hypertension ( 40 mm Hg), respirator dependency. Renal: Receiving chronic dialysis Immunocompromised: The patient has received therapy that suppresses resistance to infection, e.g., immunosuppression, chemotherapy, radiation, longterm or recent high-dose steroids, or has a disease that is sufficiently advanced to suppress resistance to infection, e.g., leukemia, lymphoma, AIDS. APACHE II Score Sum of A B C A APS points B Age points C Chronic health points Abbreviations: AaDO 2, alveolar-arterial oxygen tension gradient; PaO 2, arterial oxygen tension; FiO 2, fractional inspired oxygen; APS, acute physiology score; ABG, arterial blood gas; GI, gastrointestinal; AIDS, acquired immunodeficiency syndrome. Reprinted with permission. 13

HEPATOLOGY Vol. 34, No. 2, 2001 WEHLER ET AL. 261 sician s estimate of prognosis and, therefore be useful in clinical decision making aimed at using medical resources appropriately as well as providing patients and families with objective information. REFERENCES 1. National Center for Health Statistics Centers for Disease Control, and Prevention: Monthly Vital Statistics Report; 1997:(Suppl 2). 2. Goldfarb G, Nouel O, Poynard T, Rueff B. Efficiency of respiratory assistance in cirrhotic patients with liver failure. Intensive Care Med 1983;9: 271-273. 3. Shellman RG, Fulkerson WJ, DeLong E, Piantadosi CA. Prognosis of patients with cirrhosis and chronic liver disease admitted to the medical intensive care unit. Crit Care Med 1988;16:671-678. 4. Moreau R, Hadengue A, Soupison T, Kirstetter P, Mamzer MF, Vanjak D, Vauquelin P, et al. Septic shock in patients with cirrhosis: hemodynamic and metabolic characteristics and intensive care unit outcome. Crit Care Med 1992;20:746-750. 5. Zimmerman JE, Wagner DP, Seneff MG, Becker RB, Sun X, Knaus WA. Intensive care unit admissions with cirrhosis: risk-stratifying patient groups and predicting individual survival. HEPATOLOGY 1996;23:1393-1401. 6. Singh N, Gayowski T, Wagener MM, Marino IR. Outcome of patients with cirrhosis requiring intensive care unit support: prospective assessment of predictors of mortality. J Gastroenterol 1998;33:73-79. 7. Zauner C, Schneeweiss B, Schneider B, Madl C, Klos H, Kranz A, Ratheiser K, et al. Short-term prognosis in critically ill patients with liver cirrhosis: an evaluation of a new scoring system. Eur J Gastroenterol Hepatol 2000;2:517-522. 8. Guidelines for intensive care unit admission, discharge, and triage. Task Force of the American College of Critical Care Medicine, Society of Critical Care Medicine. Crit Care Med 1999;27:633-638. 9. Schneiderman LJ, Gilmer T, Teetzel HD. Impact of ethics consultations in the intensive care setting: a randomized, controlled trial. Crit Care Med 2000;28:3920-3924. 10. Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973;60:646-649. 11. Conn HO. A peak at the Child-Turcotte classfication. HEPATOLOGY 1981; 1:673-676. 12. Infante-Rivard C, Esnaola S, Villeneuve JP. Clinical and statistical validity of conventional prognostic factors in predicting short-term survival among cirrhotics. HEPATOLOGY 1987;7:660-664. 13. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985;13:818-829. 14. Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A, Bruining H, Reinhart CK, et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. Working Group on sepsisrelated problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996;22:707-710. 15. Vincent JL, de Mendonca A, Cantraine F, Moreno R, Takala J, Suter PM, Sprung CL, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on sepsis-related problems of the European Society of Intensive Care Medicine. Crit Care Med 1998;26:1793-1800. 16. Moreno R, Vincent JL, Matos R, Mendonca A, Cantraine F, Thijs L, Takala J, et al. The use of maximum SOFA score to quantify organ dysfunction/failure in intensive care. Results of a prospective, multicentre study. Working Group on sepsis-related problems of the European Society of Intensive Care Medicine. Intensive Care Med 1999;25:686-696. 17. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373-383. 18. Keene AR, Cullen DJ. Therapeutic Intervention Scoring System: update 1983. Crit Care Med 1983;11:1-3. 19. Brown JJ, Naylor MJ, Yagan N. Imaging of hepatic cirrhosis. Radiology 1997;202:1-16. 20. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, Schein RM, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference Committee. Chest 1992;101:1644-1655. 21. Youden WJ. Index for rating diagnostic tests. Cancer 1950;3:32-35. 22. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457-481. 23. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29-36. 24. Christensen E. Multivariate survival analysis using Cox s regression model. HEPATOLOGY 1987;7:1346-1358. 25. Zauner CA, Apsner RC, Kranz A, Kramer L, Madl C, Schneider B, Schneeweiss B, et al. Outcome prediction for patients with cirrhosis of the liver in a medical ICU: a comparison of the APACHE scores and liver-specific scoring systems. Intensive Care Med 1996;22:559-563. 26. Cooper GS, Bellamy P, Dawson NV, Desbiens N, Fulkerson WJ, Jr., Goldman L, Quinn LM, et al. A prognostic model for patients with endstage liver disease. Gastroenterology 1997;113:1278-1288. 27. Butt AK, Khan AA, Alam A, Shah SW, Shafquat F, Naqvi AB. Predicting hospital mortality in cirrhotic patients: comparison of Child-Pugh and Acute Physiology, Age and Chronic Health Evaluation (APACHE III) scoring systems. Am J Gastroenterol 1998;93:2469-2475. 28. Afessa B, Kubilis PS. Upper gastrointestinal bleeding in patients with hepatic cirrhosis: clinical course and mortality prediction. Am J Gastroenterol 2000;95:484-489. 29. Holt AW, Bury LK, Bersten AD. Prospective evaluation of residents and nurses as severity score data collectors. Crit Care Med 1992;20:1688-1691. 30. Antonelli M, Moreno R, Vincent JL, Sprung CL, Mendoca A, Passariello M, Riccioni L, et al. Application of SOFA score to trauma patients. Sequential Organ Failure Assessment. Intensive Care Med 1999;25:389-394. 31. Janssens U, Graf C, Graf J, Radke PW, Königs B, Koch KC, Lepper W, et al. Evaluation of the SOFA score: single center experience of a medical intensive care unit in 303 consecutive patients with predominantly cardiovascular disorders. Intensive Care Med 2000;26:1037-1045. 32. Marshall JC, Cook DJ, Christou NV, Bernard GR, Sprung CL, Sibbald WJ. Multiple organ dysfunction score: a reliable descriptor of a complex clinical outcome. Crit Care Med 1995;23:1638-1652. 33. Russell JA, Singer J, Bernard GR, Wheeler A, Fulkerson W, Hudson L, Schein R, et al. Changing pattern of organ dysfunction in early human sepsis is related to mortality. Crit Care Med 2000;28:3405-3411. 34. Milberg JA, Davis DR, Steinberg KP, Hudson LD. Improved survival of patients with acute respiratory distress syndrome (ARDS): 1983-1993. JAMA 1995;273:306-309. 35. Teres D, Lemeshow S. Using severity measures to describe high performance intensive care units. Crit Care Clin 1993;9:543-554.