ORIGINAL ARTICLES LIVER, PANCREAS, AND BILIARY TRACT

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2012;10:65 71 ORIGINAL ARTICLES LIVER, PANCREAS, AND BILIARY TRACT Acute Kidney Injury Is an Early Predictor of Mortality for Patients With Alcoholic Hepatitis JOSÉ ALTAMIRANO,*,, CLAUDIA FAGUNDES,*,, MARLENE DOMINGUEZ,*, ELISABET GARCÍA,*,, JAVIER MICHELENA,*,, ANDRÉS CÁRDENAS,*, MONICA GUEVARA,*, GUSTAVO PEREIRA,*,, KARINA TORRES VIGIL,*,, VICENTE ARROYO,*,, JUAN CABALLERÍA,*,, PERE GINÈS,*,, and RAMÓN BATALLER*,,, *Liver Unit, Institut d Investigacions Biomèdiques August-Pi-Sunyer, Hospital Clínic, University of Barcelona, Barcelona; Centre Esther Koplowitz, Barcelona; and Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas, Barcelona, Spain BACKGROUND & AIMS: Alcoholic hepatitis (AH) is a severe condition with high mortality. To improve therapeutic strategies, it is important to identify factors that affect survival times. The age, bilirubin, international normalized ratio, and creatinine scoring system (also known as the ABIC scoring system) was developed previously to determine the prognosis of patients with AH. We studied effects of acute kidney injury (AKI) on survival of patients with AH. METHODS: We retrospectively analyzed data from 103 patients with biopsyproven AH. AKI was defined as an abrupt reduction (within 48 h) in kidney function that resulted in an absolute increase of at least 0.3 mg/dl (or a 50% increase) in serum levels of creatinine from baseline (the AKI network [AKIN] criteria). RE- SULTS: Twenty-nine patients (28%) developed AKI during hospitalization, with a median time to diagnosis of 3 days. Overall 90-day mortality was 23%, which was significantly higher among patients with AKI than those without (65% vs 7%; P.0001). The age, bilirubin, international normalized ratio, and creatinine score (P.0001) and development of AKI (P.0001) were the most accurate independent predictors of 90-day mortality. The presence of systemic inflammatory response syndrome (P.0001), serum bilirubin (P.01), and international normalized ratio at admission (P.03) were the most accurate predictors of AKI. Importantly, the AKIN criteria were more accurate than traditional criteria for renal failure (serum creatinine 1.5 mg/dl) in predicting 90-day mortality (area under the receiver operating characteristic, 0.83 vs 0.70, respectively; P.02). CONCLUSIONS: Development of AKI reduces survival of patients with AH, in the short term. The AKIN criteria are useful and more accurate than traditional criteria in predicting mortality. Strategies to prevent AKI therefore should be considered in the management of patients with AH. Keywords: Liver Disease; Acute Kidney Injury; Alcoholic Liver Disease; Prognosis; Alcoholic Hepatitis. Alcoholic hepatitis (AH) is a severe clinical condition in patients with alcoholic liver disease. 1 Severe forms carry a high mortality rate as a result of liver failure and severe portal hypertension, and increased susceptibility to renal failure and bacterial infections. 2 Current therapies fail in many patients and there is a clear need to develop new pathophysiologically oriented therapies. 3 To design clinical trials, a prognostic classification of patients into different risk subgroups is required. Recently, we developed a new scoring system based on the combination of age, bilirubin, international normalized ratio (INR), and creatinine (ABIC score) at admission that allows the prognostic stratification of patients with AH (available from: abic). The ABIC score identifies patients with low ( 6.71 points), intermediate ( points), and high risk ( 9 points) of death at 90 days and 1 year. 4 Acute renal failure occurs in a variety of settings with manifestations ranging from minimal increase of serum creatinine (SCr) level to advanced renal failure. 5 Recently, the Acute Kidney Injury Network (AKIN) proposed a new term for acute renal failure, named acute kidney injury (AKI), and developed a diagnostic definition and staging system for AKI (Supplementary Table 1). 6 The negative impact on short- and long-term mortality of AKI development has been evaluated and validated in critically ill patients with several conditions, identifying AKI as a potential therapeutic target. 7,8 However, there are no studies evaluating the development of AKI among patients with AH. Our primary objectives were to retrospectively evaluate the clinical impact of AKI in a prospectively gathered dataset of patients with biopsy-proven AH. Specifically, we studied the occurrence of AKI during hospitalization, the influence on 90-day mortality and the factors associated with AKI development in these patients. Abbreviations used in this paper: ABIC, age, bilirubin, international normalized ratio, and creatinine; AH, alcoholic hepatitis; AKI, acute kidney injury; AKIN, Acute Kidney Injury Network; AUROC, area under the receiver operating characteristic curve; HE, hepatic encephalopathy; HRS, hepatorenal syndrome; INR, international normalized ratio; SCr, serum creatinine ratio; SIRS, systemic inflammatory response syndrome by the AGA Institute /$36.00 doi: /j.cgh

2 66 ALTAMIRANO ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 10, No. 1 Patients and Methods Patients and Interventions A total of 144 consecutive patients admitted to the Liver Unit (Hospital Clinic, Barcelona, Spain) between 2000 and 2007 with clinical and analytic criteria of AH were identified prospectively. Inclusion criteria and cohort details have been described previously. 4 A total of 103 patients with biopsyproven AH were included. Severe AH was defined as a Maddrey s discriminant function greater than 32 and/or an ABIC score greater than 6.71 at admission. Patients with Maddrey s discriminant function of 32 or greater who fulfilled the histologic criteria for AH were treated with prednisone orally for 4 weeks followed by a 2-week taper period. Contraindications for corticosteroids were severe bacterial infections or diabetes mellitus with poor metabolic control. During treatment with prednisone all patients received gastric acid suppression with proton pump inhibitors. During hospitalization, patients with clinical complications such as ascites, spontaneous bacterial peritonitis, renal dysfunction, overt hepatic encephalopathy (HE), or gastrointestinal bleeding associated with portal hypertension were treated according to current international guidelines and the clinical protocol for AH of the Liver Unit in effect when the patients were admitted. 4,9 12 Portal pressure was estimated by the hepatic venous pressure gradient, as described in detail previously. 13 This study was approved by the Ethics Committee of the Hospital Clinic and all patients provided written informed consent. Data Collection and Definitions The demographic, clinical, and analytic parameters collected in these series have been described previously. 4 The presence of the systemic inflammatory response syndrome (SIRS) was evaluated at admission. The presence of SIRS was considered in patients with 2 or more of the following criteria: (1) temperature less than 36 C or greater than 38 C, (2) heart rate higher than 90 beats/min, (3) respiratory rate greater than 20 breaths/min, and (4) leukocyte count higher than 12,000/mm 3 or less than 4000/mm In addition to baseline parameters, liver and renal function tests were measured every 24 to 48 hours according to the patient s disease severity. Criteria to define renal impairment were assessed in a retrospective manner. AKI was defined as an abrupt (within 48 h) reduction in kidney function characterized by an absolute increase in SCr of 0.3 mg/dl or greater or a percentage increase in SCr of 50% or greater (1.5-fold from baseline). Patients were categorized according to AKIN criteria, based on SCr level at admission as baseline. Stage 1 was defined as an increase of greater than 0.3 mg/dl or a 50% increase from baseline, stage 2 was defined as a greater than 200% to 300% increase from baseline SCr level, and stage 3 was defined as an increase in SCr greater than 300% from baseline, an SCr greater than 4.0 mg/dl with an increase of at least 0.5 mg/dl, or the receipt of renal replacement therapy. The most severe degree of AKI was recorded as the final AKI stage. We also classified AKI evolution in 2 categories as follows: (1) transient, when the final value returned to baseline levels or if there was a decrease in SCr level greater than 25% compared with the value of diagnosis of AKI, and (2) persistent, when there was a decrease less than 25% or an increase in SCr level compared with the value at AKI diagnosis. Statistical Analysis Continuous variables were described as means (95% confidence interval) or medians (interquartile range). Categoric variables were described by means of counts and percentages. Comparisons between groups were performed using the Student t test or the Mann Whitney U test when appropriate. Differences between categoric variables were assessed by the chi-square test or the Fisher exact test. The main end points were the development of AKI during hospital admission and death at 90 days. A stepwise analysis approach was performed for this post hoc analysis. For the first aim, univariate and multivariate logistic regression (odds ratio [OR]) analyses were used to identify those variables associated with AKI. For the second aim, univariate and multivariate Cox regression (hazard ratio) analyses were performed. The area under the receiver operating characteristic curve (AUROC) analysis was used to determine the best cut-off value of continuous variables and the accuracy (sensitivity and specificity) of different systems of renal impairment. Comparison between AUROC curves was performed by the method of Hanley and McNeil using MedCalc statistics software (Medisoftware, Mariakerke, Belgium). Finally, to evaluate the influence of the most significant variables identified for AKI development and 90-day survival, comparative risk analysis using the Kaplan Meier method compared by the log-rank test was performed. A P value of less than.05 was required for significance. The SPSS statistical package (version 15.0; SPSS, Inc, Chicago, IL) was used for all analyses. More details of the statistical approach are provided in the Supplementary Materials and Methods. Results Patient Characteristics The baseline clinical, demographic, and biochemical characteristics as well as other important information of the study cohort have been described previously. 4 The median alcohol consumption was 100 g/d (range, ). At admission, SIRS criteria were fulfilled by half of the patients. Seventynine patients (77%) developed at least one clinical complication during hospitalization. Forty-seven patients (46%) were treated with corticosteroids and no patient received pentoxifylline. Patients receiving corticosteroids were not more susceptible to develop infections during hospitalization. Development of Acute Kidney Injury During hospitalization, AKI occurred in 29 patients (28%). The median time to AKI diagnosis was 3 days (range, 2 5 d). According to the AKIN criteria, 15 patients (52%) were classified in stage 1, 7 (24%) in stage 2, and 7 (24%) in stage 3. Two of the 7 patients in AKIN stage 3 received renal replacement therapy during the course of renal failure. The mean values of baseline SCr level (at admission) were , , and mg/dl in patients with stage 1, stage 2, and stage 3, respectively. At the time of AKI diagnosis, the median increases in absolute values of creatinine were 0.3 mg/dl (range, mg/dl) in stage 1, 1.7 mg/dl (range, mg/dl) in stage 2, and 2.9 mg/dl (range, mg/dl) in stage 3. Regarding the recovery of renal function, transient kidney injury was observed in 8 of 15 patients (53%) in stage 1 and in 2

3 January 2012 ACUTE KIDNEY INJURY IN ALCOHOLIC HEPATITIS 67 Table 1. Characteristics of Patients With AH According to the Presence of AKI AKI, % No AKI, % P n(%) 29 (28) 74 (72) Age, y a 49 (46 53) 49 (47 51).76 Male, % 19 (65) 49 (66) Alcohol intake, g/d 100 ( ) 100 ( ).39 Cirrhosis, % 21 (72) 41 (55).14 Corticosteroids, % 15 (52) 32 (43).57 At admission Ascites, % 25 (86) 44 (59).009 Variceal bleeding, % 1 (3) 11 (15).17 Encephalopathy, % 13 (45) 10 (13).001 Infection, % 7 (24) 7 (9).06 SIRS, % b 25 (89) 26 (38).001 SIRS criteria Temperature 38 C/ 36 C, % 8 (25) 5 (7).02 Heart rate 90 bpm, % 21 (75) 32 (47).01 Respiratory rate 20 bpm, % 23 (82) 37 (5).01 Leukocytes 12,000 mm 3 / 4000 mm 3, % 16 (57) 24 (35).04 Laboratory parameters at admission.31 Hemoglobin level, g/dl a 10.3 ( ) 10.8 ( ) Leukocyte count 10 9 /L 10.6 ( ) 8.5 ( ).39 Platelet count 10 9 /L 134 (54 168) 110 (70 183).97 AST level, U/L 125 (95 206) 128 (85 182).46 ALT level, U/L 58 (40 79) 48 (36 63).15 Serum Na level, mmol/l 131 ( ) 133 ( ).14 Serum albumin level, g/dl 2.5 ( ) 2.7 ( ).19 SCr level, mg/dl 1.1 ( ) 0.7 ( ).01 Serum bilirubin level, mg/dl a 18.4(14.3, 22.4) 10.0(8.3, 11.6).001 International normalized ratio 1.9 ( ) 1.5 ( ).004 GFR (MDRD) ml/min/1.73 m ( ) ( ).005 HVPG, mm Hg 20 (18 25) 20 (15 23).23 Clinical scores at admission Maddrey s discriminant function 65.5 ( ) 42.6 ( ).001 MELD score a 27 (24 31) 20 (18 2).001 ABIC score a 8.63 ( ) 7.25 ( ).001 ABIC class, % A( 6.71) 3 (10) 23 (30).001 B ( ) 13 (45) 45 (62) C( 9) 13 (45) 6 (8) During hospitalization.52 Ascites, % 5 (17) 17 (23) Variceal bleeding, % 5 (17) 4 (5).11 Encephalopathy, % 11 (38) 11 (15).01 Infection, % 22 (76) 24 (32).001 AST, aspartate aminotransferase; ALT, alanine aminotransferase; GFR, glomerular filtration rate; HVPG, hepatic venous pressure gradient; MDRD, Modification of Diet in Renal Disease Study; MELD, model for end-stage liver disease. a Expressed as median (95% confidence interval). b SIRS criteria were evaluated in 96 patients from the total cohort. of 7 patients (28%) in stage 2. All patients in stage 3 showed persistent kidney injury and none of them recovered normal renal function. The evolution of SCr among patients with transient and persistent AKI is depicted in Supplementary Figure 1. The causes of renal failure in patients with AKI were as follows: 10 patients (35%) with hepatorenal syndrome (HRS) (9 patients with HRS type 1 and 1 patient with HRS type 2), 15 patients (62%) with renal failure associated with infections, and 4 patients (14%) with mixed causes (2 patients with parenchymal nephropathy associated with bacterial infection and hypovolemia and 2 patients with nephrotoxicity secondary to nonsteroidal anti-inflammatory drugs). Development of AKI was more prevalent among infected patients compared with noninfected patients (47% vs 8%; P.0001). Patients developing AKI were more likely to have SIRS at admission (89% vs 38%; P.0001) and higher model for end-stage liver disease and ABIC scores (Table 1). In addition, patients with AKI developed more infections (76% vs 32%; P.0001) and HE (38% vs 15%; P.01) episodes during hospitalization compared with patients without AKI. Variables associated with AKI development in the univariate analysis were serum bilirubin levels, INR, serum creatinine levels, presence of SIRS, ascites, and HE at admission. In the multivariate analysis, the presence of SIRS (OR, 18.41; P.0001), serum bilirubin levels (OR, 1.08; P.01), and INR (OR, 2.89; P.03) were associated independently with AKI development (Table 2). The best cut-off level for bilirubin was 16 mg/dl

4 68 ALTAMIRANO ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 10, No. 1 Table 2. Factors Influencing the Development of AKI on Univariate and Multivariate Logistic Regression in Patients With AH Univariate logistic regression OR 95% confidence interval At admission Serum bilirubin INR Serum creatinine SIRS, y/n Encephalopathy, y/n Cirrhosis, y/n Infection, y/n Ascites, y/n During hospitalization Encephalopathy, y/n Infection, y/n Multivariate logistic regression SIRS, y/n Serum bilirubin INR (AUROC, 0.72; P.0001; sensitivity, 59%; specificity, 82%) and 1.7 for INR (AUROC, 0.68; P.003; sensitivity, 69%; specificity, 67%). Among patients with SIRS and poor liver function (bilirubin level, 16 mg/dl; INR, 1.7), AKI developed in 71%. Factors Influencing Patient Survival: Role of Acute Kidney Injury The overall in-hospital and 90-day mortality rates were 15% and 23%, respectively. The main causes of death at 90 days were sepsis (29%) and multiorgan failure (52%). The univariate analysis identified age, serum bilirubin levels, INR, serum creatinine levels, presence of SIRS, ABIC score, and HE at admission as well as infection, AKI development, and HE during hospitalization associated with 90-day mortality. In the multivariate analysis, the ABIC score (hazard ratio, 2.08; P.001) and development of AKI (hazard ratio, 8.07; P.0001) were the best predictors of 90-day mortality (Table 3). Importantly, we found that the development of AKI markedly influenced 3-month survival in patients with AH (35% vs 93% in patients with and without AKI, respectively; P.001; Figure 1). Regarding the influence of AKIN stages, patients with stage 1 had a better prognosis than patients with either stages 2 or 3 (47% vs 21% survival at 90-d; P.03; Supplementary Figure 2). Moreover, the development of AKI equally influenced survival both in patients receiving and those not receiving corticosteroids (Supplementary Figure 3). Impact of Systemic Inflammatory Response Syndrome on Acute Kidney Injury Development and Patient Survival At admission, 51 patients (49%) met the diagnostic criteria of SIRS (56% fulfilled 2 criteria and 44% fulfilled 2 criteria). Interestingly, SIRS was present in all infected patients at admission and in 61% of patients who further developed an episode of infection during hospitalization. The median time for infection diagnosis in patients with and without SIRS was 6 P Table 3. Predictive Factors of 90-Day Mortality in the Univariate and Multivariate Cox Regression in Patients With AH Univariate Cox regression Hazard ratio 95% confidence interval At admission Age Serum bilirubin INR Serum creatinine SIRS, y/n Encephalopathy, y/n Cirrhosis Infection, y/n ABIC a During hospitalization AKI, y/n Encephalopathy, y/n Infection, y/n Multivariate Cox regression AKI, y/n ABIC a a As a nondichotomous variable, the hazard ratio applies to each increment of 1 U of the variable. days (range, 4 14 d) and 4 days (range, 3 7 d), respectively. Of note, 49% of patients with SIRS at admission developed AKI during hospitalization compared with only 7% of patients without SIRS (P.0001; Figure 2). Importantly, the presence of SIRS was associated with a reduced survival (60% survival rate at 3 mo in patients with SIRS vs 93% in patients without SIRS; P.001; Figure 3). Figure 1. Three-month survival probability since the time of admission of patients with AH according to the presence of AKI. P

5 January 2012 ACUTE KIDNEY INJURY IN ALCOHOLIC HEPATITIS 69 Figure 2. Probability of AKI development according to the presence of SIRS at admission. Performance of Acute Kidney Injury Network Criteria to Predict Short-Term Mortality Finally, we compared the performance of AKIN criteria versus the traditional definition of renal failure in advanced liver disease (SCr level, 1.5 mg/dl) in predicting short-term mortality. The AUROC analysis showed that the AKIN criteria had a stronger discriminatory capability than a SCr level greater than 1.5 mg/dl as criteria to predict 90-day mortality (AUROC, 0.83; 95% confidence interval, ; and 0.70; 95% confidence interval, , respectively; P.02, Supplementary Figure 4). The stratification of patients according to the AKIN criteria identified more patients with AKI (n 29) than with the traditional criteria (n 17). Finally, we fitted different logistic regression models for 90-day mortality including renal dysfunction diagnosis based on either AKIN or traditional criteria (Supplementary Table 2). We found that the models including AKIN criteria have better accuracy for predicting 90-day survival. Discussion The development of renal impairment in critically ill patients (including those with cirrhosis) is a key clinical event influencing survival. 15,16 Here, we investigate the prevalence and the significance of renal failure in AH. We provide evidence that AKI is a frequent and early complication in patients with AH. The development of AKI heavily impacts short-term prognosis in this clinical setting. Our study strongly indicates that renal impairment is a critical event in the course of AH. We also showed that the degree of systemic inflammatory response and liver failure are the main factors predicting AKI development. These results confirm previous studies suggesting that renal impairment negatively impacts the survival of patients with AH. 17 An important finding in the current study was that AKI is an early event (3 d; range, 2 5 d) after hospital admission. By using a sensitive system of renal injury (AKIN criteria) we were able to diagnose renal impairment at early stages during hospitalization. This finding is clinically relevant because it can help clinicians to initiate strategies aimed to prevent or reverse renal dysfunction in these patients. In this regard, AKIN criteria have been shown to be more accurate than the traditional criteria of renal failure in cirrhosis for predicting short-term mortality. Our results support the use of the AKIN criteria among hepatologists to diagnose renal impairment in patients with suspected AH. Whether AKIN criteria are also useful in patients with other types of advanced liver diseases is unknown and deserves further investigation. Based on our results, we propose that future clinical trials to prevent and/or treat renal failure in patients with AH use the AKIN criteria. We found that half of the patients fulfilled SIRS criteria at admission. Remarkably, the presence of SIRS at admission strongly predicted the development of AKI and was associated with poor survival. The high prevalence of SIRS in these patients can be the result of 2 main reasons. First, livers with AH actively secrete a number of proinflammatory chemokines to the systemic circulation including tumor necrosis factor- and CXC chemokines. 2,18 Second, most of the patients with SIRS had either ongoing bacterial infections or will develop an infectious episode during hospitalization. 19,20 Thus, the presence of occult infections in patients with SIRS cannot be ruled out. Future studies also should investigate whether the use of systemic antibiotics in patients with AH and SIRS prevent the development of renal failure and improve survival. This issue is relevant because patients with AH, similar to others with acuteon-chronic liver failure, are at a high risk for developing bacterial infections. 21,22 The increased susceptibility to bacterial infections could be owing to the presence of the so-called immune dissonance, which is characterized by an exaggerated proinflammatory and a defective anti-inflammatory response. 23 Recent studies at the critical care setting have indicated that Figure 3. Three-month probability of survival of patients with AH according to the presence of SIRS at admission.

6 70 ALTAMIRANO ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 10, No. 1 AKI development is not a mere bystander and could precede and favor the development of bacterial infections. 24 The presence of SIRS precedes and predicts the development of renal failure and is associated with poor survival. Different factors could explain this finding. First, increased systemic and intrahepatic chemokine levels could impair the circulatory dysfunction that commonly is found in these patients, leading to further activation of vasoconstrictor systems and favoring renal hypoperfusion and parenchymal injury. This hypothesis was not assessed in our study and deserves further investigation. Second, the existence of SIRS could promote endothelial dysfunction and hepatocellular damage and disturb the intrahepatic circulation, thus favoring the development of clinical complications. The fact that patients with SIRS have more complications related to portal hypertension supports this hypothesis. 20 Third, the presence of SIRS could be indicative of an ongoing bacterial infection, either occult or clinically evident, which is a known precipitating factor of renal failure in patients with chronic liver diseases. 25 Fourth, the cytokine overspill that characterized bacterial infections and SIRS could have a direct deleterious role in the genesis of renal dysfunction by stimulating apoptosis and necrosis of tubular cells. 26 It is possible that these events also occur in patients with cirrhosis, as suggested by previous studies showing an association between SIRS and acute renal failure in cirrhotic patients. 27 Another interesting finding is that the degree of liver failure (as assessed by serum bilirubin and INR) was associated independently with AKI development. This finding confirms previous studies showing that the severity of liver failure is an independent factor for the development of renal failure in cirrhotic patients with gastrointestinal bleeding, spontaneous bacterial peritonitis, and other bacterial infections. 25,28 30 Similarly, a recent study showed that the degree of liver insufficiency strongly associates with AKI development in nonparacetamol-induced acute liver failure. 8 The underlying mechanisms by which liver failure is associated with renal impairment include increased severity of circulatory dysfunction and susceptibility to develop bacterial infections. Whether intrarenal mechanisms also play a role in AKI in patients with AH is unknown and deserves future investigation. Our study had some limitations. First, the present study was a subanalysis of a previously reported cohort from a single center. Second, we did not measure the activation of vasoactive system (ie, plasma renin activity and the plasma concentrations of aldosterone, norepinephrine, arginine vasopressin and endothelin) in the systemic circulation. Third, renal parameters such as urine sodium content, as well as serum and urine markers of AKI (ie, Neutrophil gelatinase-associated lipocalin [NGAL], interleukin-18, and Kidney Injury Molecule-1 [KIM-1]), were not measured in many patients. Fourth, we could not evaluate patients with AKI at the time of admission because in the majority of patients previous values of SCr were not available. Fifth, we did not use timed urine output criteria for diagnosing AKI because the hourly urine output data for determining oliguria within a 6-, 12-, or a 24-hour window was not obtained. However, in advanced liver diseases, oliguria could be commonly related to sodium retention and does not necessarily reflect renal failure. In fact, it is a common misconception that oliguria is common in certain types of AKI (ie, HRS). 31 Furthermore, recent data have shown that urine output does not increase the accuracy of AKIN criteria when compared with SCr levels in the intensive care setting. 32 In conclusion, this study shows that early development of renal impairment in patients with AH during hospitalization negatively impacts their short-term survival. The AKIN criteria for AKI diagnosis are useful and more accurate than the traditional criteria to identify this complication. The presence of SIRS at admission and infection during hospitalization were strongly associated with the development of AKI and has a direct prognostic value in predicting the outcome of these patients. Further studies should be performed to confirm these results to identify maneuvers that effectively prevent the development of AKI in high-risk patients with AH. Supplementary Materials Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at and at doi: / j.cgh References 1. Parés A, Caballería J, Bruguera M, et al. Histological course of alcoholic hepatitis. Influence of abstinence, sex and extent of hepatic damage. J Hepatol 1986;2: Dominguez M, Miquel R, Colmenero J, et al. Hepatic expression of CXC chemokines predicts portal hypertension and survival in patients with alcoholic hepatitis. Gastroenterology 2009;136: Louvet A, Diaz E, Dharancy S, et al. Early switch to pentoxifylline in patients with severe alcoholic hepatitis is inefficient in nonresponders to corticosteroids. J Hepatol 2008;48: Dominguez M, Rincón D, Abraldes JG, et al. A new scoring system for prognostic stratification of patients with alcoholic hepatitis. Am J Gastroenterol 2008;103: Palevsky PM. Epidemiology of acute renal failure: the tip of the iceberg. Clin J Am Soc Nephrol 2006;1: Mehta RL, Kellum JA, Shah SV, et al. Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007;11:R Ricci Z, Cruz D, Ronco C. The RIFLE criteria and mortality in acute kidney injury: a systematic review. Kidney Int 2008;73: Leithead JA, Ferguson JW, Bates CM, et al. The systemic inflammatory response syndrome is predictive of renal dysfunction in patients with non-paracetamol-induced acute liver failure. Gut 2009;58: Garcia-Tsao G, Sanyal AJ, Grace ND, et al. Prevention and management of gastroesophageal varices and variceal hemorrhage in cirrhosis. Hepatology 2007;46: Moore KP, Wong F, Gines P, et al. The management of ascites in cirrhosis: report on the consensus conference of the International Ascites Club. Hepatology 2003;38: Runyon BA. Practice Guidelines Committee, American Association for the Study of Liver Diseases (AASLD), management of adult patients with ascites due to cirrhosis. Hepatology 2004; 39: Ferenci P, Lockwood A, Mullen K, et al. Hepatic encephalopathy definition, nomenclature, diagnosis, and quantification: final report of the working party at the 11th World Congresses of Gastroenterology, Vienna, Hepatology 2002;35: Bosch J, Abraldes JG, Berzigotti A, et al. The clinical use of HVPG measurements in chronic liver disease. Nat Rev Gastroenterol Hepatol 2009;6: Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in

7 January 2012 ACUTE KIDNEY INJURY IN ALCOHOLIC HEPATITIS 71 sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992;101: D Amico G, Garcia-Tsao G, Pagliaro L. Natural history and prognostic indicators of survival in cirrhosis: a systematic review of 118 studies. J Hepatol 2006;44: Cholongitas E, Calvaruso V, Senzolo M, et al. RIFLE classification as predictive factor of mortality in patients with cirrhosis admitted to intensive care unit. J Gastroenterol Hepatol 2009;24: Louvet A, Naveau S, Abdelnour M, et al. The Lille model: a new tool for therapeutic strategy in patients with severe alcoholic hepatitis treated with steroids. Hepatology 2007;45: Fujimoto M, Uemura M, Nakatani Y, et al. Plasma endotoxin and serum cytokine levels in patients with alcoholic hepatitis: relation to severity of liver disturbance. Alcohol Clin Exp Res 2000; 24(Suppl 4):48S 54S. 19. Rangel-Frausto MS, Pittet D, Costigan M, et al. The natural history of the systemic inflammatory response syndrome (SIRS). A prospective study. JAMA 1995;11: Cazzaniga M, Dionigi E, Gobbo G, et al. The systemic inflammatory response syndrome in cirrhotic patients: relationship with their in-hospital outcome. J Hepatol 2009;51: Louvet A, Wartel F, Castel H, et al. Infection in patients with severe alcoholic hepatitis treated with steroids: early response to therapy is the key factor. Gastroenterology 2009;137: Katoonizadeh A, Laleman W, Verslype C, et al. Early features of acute-on-chronic alcoholic liver failure: a prospective cohort study. Gut 2010;59: Wasmuth HE, Kunz D, Yagmur E, et al. Patients with acute on chronic liver failure display sepsis-like immune paralysis. J Hepatol 2005;42: Mehta RL, Bouchard J, Soroko SB, et al. Sepsis as a cause and consequence of acute kidney injury: program to improve care in acute renal disease. Intensive Care Med 2011;37: Terra C, Guevara M, Torre A, et al. Renal failure in patients with cirrhosis and sepsis unrelated to spontaneous bacterial peritonitis: value of MELD score. Gastroenterology 2005;129: Simmons EM, Himmelfarb J, Sezer MT, et al. Plasma cytokine levels predict mortality in patients with acute renal failure. Kidney Int 2004;65: Thabut D, Massard J, Gangloff A, et al. Model for end-stage liver disease score and systemic inflammatory response are major prognostic factors in patients with cirrhosis and acute functional renal failure. Hepatology 2007;46: Cárdenas A, Ginès P, Uriz J, et al. Renal failure after upper gastrointestinal bleeding in cirrhosis: incidence, clinical course, predictive factors, and short-term prognosis. Hepatology 2001; 34: Fasolato S, Angeli P, Dallagnese L, et al. Renal failure and bacterial infections in patients with cirrhosis: epidemiology and clinical features. Hepatology 2007;45: Sort P, Navasa M, Arroyo V, et al. Effect of intravenous albumin on renal impairment and mortality in patients with cirrhosis and spontaneous bacterial peritonitis. N Engl J Med 1999; 341: Baldus WP, Feichter RN, Summerskill WH. The kidney in cirrhosis. I. Clinical and biochemical features of azotemia in hepatic failure. Ann Intern Med 1964;60: Lopes JA, Fernandes P, Jorge S, et al. Acute kidney injury in intensive care unit patients: a comparison between the RIFLE and the Acute Kidney Injury Network classifications. Crit Care 2008;12:R110. Reprint requests Address requests for reprints to: Ramón Bataller, MD, Liver Unit, Hospital Clinic, Villarroel 170, Barcelona 08036, Spain. bataller@clinic.ub.es; fax: (34) Conflicts of interest The authors disclose no conflicts. Funding This work was supported by grants from the Fondo de Investigación Sanitaria (FIS PI080237, FIS PS09/01164, and FIS PI to R.B., J.C., and P.G.); José Altamirano received a grant from the Fundación Banco Bilbao Vizcaya Argentaria and is enrolled in the Master on Research in Liver Diseases of the Universitat de Barcelona; Claudia Fagundes was supported by a grant from the Instituto Reina Sofía de Investigación Nefrológica; and the Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas is funded by the Instituto de Salud Carlos III.

8 January 2012 ACUTE KIDNEY INJURY IN ALCOHOLIC HEPATITIS 71.e1 Supplementary Materials and Methods Statistical Analysis To avoid co-linearity in multivariate modeling, those variables included in a prognostic score (eg, ABIC and MELD scores) were not included separately in the multivariate models. To avoid overfitting of the different models, a predefined ratio of candidate prognostic variables to the number of observed events (deaths or AKI) was set at 1:10. Reduction of the logarithm of the likelihood ratio (-2LL) was used as a descriptive measure of calibration (the closer -2LL gets to 0, the better the model adjusts to explain the outcome). For discrimination we used the AUROC of the predicted probability of mortality for each model. The final model was selected taking these parameters into account and were adjusted for other relevant clinical variables (ie, cirrhosis).

9 71.e2 ALTAMIRANO ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 10, No. 1 Supplementary Table 1. Classification/Staging System for AKI Based on AKIN Criteria Stage SCr criteria Urine output criteria 1 Increase in SCr or 0.3 mg/dl or increase 150% to 200% 0.5 ml/kg/h for 6 h (1.5- to 2-fold) from baseline 2 Increase in SCr to 200% to 300% ( 2- to 3-fold) from baseline 0.5 ml/kg/h for 12 h 3 a Increase in SCr 300% ( 3-fold) from baseline (or SCr 4.0 mg/dl with an acute increase of at least 0.5 mg/dl) 0.3 ml/kg/h for 24 h or anuria for 12 h a Individuals who received renal replacement therapy are considered to have met the criteria for stage 3 irrespective of the stage they are at the time of renal replacement therapy. Supplementary Figure 1. Mean ( standard error of the mean) values of SCr at different time points during hospitalization in patients with transient and persistent AKI. End, end of hospitalization or death during hospitalization. Supplementary Figure 2. Three-month survival in patients with AH and AKI according to the stage of AKIN. Figures under the curves are patients at risk at different time points.

10 January 2012 ACUTE KIDNEY INJURY IN ALCOHOLIC HEPATITIS 71.e3 Supplementary Figure 3. Three-month probability of survival of patients with AH according to the development of AKI and corticosteroid treatment. Figures under the curves are patients at risk at different time points. Supplementary Figure 4. Comparison of the ROC curves for 90-day survival according to the different criteria of renal failure in patients with AH. Supplementary Table 2. Multivariate Logistic Regression for 90-Day Mortality in Patients With AH Multivariate logistic regression OR 95% confidence interval P Model AUROC (95% confidence interval) AKIN, y/n ( ) ABIC score Traditional criteria, y/n ( ) ABIC score AKIN, y/n ( ) MELD score Traditional criteria, y/n ( ) MELD score