For patients with cirrhosis, increased operative risk relative

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2010;8: ORIGINAL ARTICLES LIVER, PANCREAS, AND BILIARY TRACT Factors That Predict Outcome of Abdominal Operations in Patients With Advanced Cirrhosis DANA A. TELEM,* THOMAS SCHIANO, ROBERT GOLDSTONE,* DANIEL K. HAN,* KERRI E. BUCH,* EDWARD H. CHIN,* SCOTT Q. NGUYEN,* and CELIA M. DIVINO* *Division of General Surgery, Department of Surgery, and Recanati Miller Transplant Institute, Division of Liver Disease, The Mount Sinai Hospital, New York, New York This article has an accompanying continuing medical education activity on page e58. Learning Objectives At the end of this activity, the learner should be able to define the expected frequency of morbidity and mortality of abdominal operations in patients with advanced cirrhosis, and how best to predict such outcomes. See Editorial on page 399. BACKGROUND & AIMS: Patients with cirrhosis have an increased risk of complications during surgery that is relative to the severity of liver disease; it is a challenge to determine which patients are the best candidates for surgery. We performed a hospital-based study to identify factors that might facilitate selection of operative candidates and guide their management. METHODS: A retrospective review was performed of 100 cirrhotic patients (50 classified as Child Turcotte Pugh [CTP] A, 33 as CTP B, and 17 as CTP C) who underwent abdominal surgery at an institution specializing in liver medicine and transplant from Significant univariate variables were evaluated by multivariate logistic regression models to identify factors that correlate with outcome. RESULTS: The overall, 30-day postoperative mortality rate was 7%. The mortality for patients who were CTP A was 2%, CTP B was 12%, and CTP C was 12%; 33 patients had a Model for End-Stage Liver Disease (MELD) score 15, with 29% mortality. On the basis of multivariate analyses, risk factors for adverse outcome were American Society of Anesthesiologists (ASA) score 3; procedures being emergent; intraoperative blood transfusion; intraoperative blood loss 150 ml; presence of ascites; total bilirubin level 1.5 mg/dl; and albumin level 3 mg/dl. Addition of serum albumin to MELD score showed that patients with MELD score 15 and albumin 2.5 mg/dl (vs 2.5 mg/dl) had significantly increased mortality (60% vs 14%, P.01) and independently increased probability of adverse outcome (odds ratio, 8.4; P.015). CONCLUSIONS: For patients with MELD scores >15, the preoperative albumin level correlates with outcome and could guide operative decisions. Intraoperative packed red blood cell transfusion correlates with adverse outcome and should be limited. Keywords: Cirrhosis; Operative Outcome; CTP; MELD; General Surgery. For patients with cirrhosis, increased operative risk relative to the severity of liver disease is well-established. 1 3 Determination of appropriate operative candidates is difficult, particularly for patients with advanced disease. Child Turcotte Pugh (CTP) and Model for End-Stage Liver Disease (MELD) scoring systems are often used to guide operative decisions. 1 8 Studies demonstrate that patients classified as CTP A, B, and C have postoperative mortality rates of up to 10%, 30%, and 80%, respectively. 4,5 Assessment by MELD criteria demonstrates scores of 0 11 correlate with 5% 10% mortality rates, with 25% 54% mortality rates, and scores greater than 26 with a 90% postoperative mortality rate A limitation of both the CTP and MELD scoring systems is that neither accurately predicts operative outcome. 3 8 This limitation might be attributable to the exclusion of other influential outcome factors including acuity of patient presentation and operative course. 4 6,8 In addition, cardiac literature recently identified preoperative platelet count, a surrogate marker for degree of portal hypertension, as an independent predictor of patient outcome after coronary artery bypass grafting. 11 No literature specifically addressing the impact of platelet count on outcome after abdominal procedures currently exists. 11 The importance of patient selection based on assessment of operative risk is reaffirmed by a recent population-based study of 22,569 cirrhotic patients who underwent elective operative procedures. 7 This study demonstrated significantly worse operative outcomes for cirrhotic patients, particularly those with portal hypertension. 7 A major study limitation, however, was the inability to classify patients by CTP or MELD score or identify variables associated with adverse outcome. 7 Thus, we undertook this hospital-based study to assess the impact of Abbreviations used in this paper: ASA, American Society of Anesthesiologists; CI, confidence interval; CTP, Child Turcotte Pugh; ICU, intensive care unit; LOS, length of stay; MELD, Model for End-Stage Liver Disease; OR, odds ratio by the AGA Institute /10/$36.00 doi: /j.cgh

2 452 TELEM ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 5 perioperative factors on patient outcome after abdominal procedures at an institution specializing in liver disease and transplant medicine. We aimed to identify factors that might facilitate selection of appropriate operative candidates and guide perioperative patient management. Methods Patient Acquisition After approval by the Mount Sinai School of Medicine Institutional Review Board, a retrospective chart review was performed of 100 cirrhotic patients who underwent general surgical procedures at The Mount Sinai Medical Center from January 1, 2002 to December 31, Patients were identified from an administrative database by cross-matching International Classification of Diseases, 9th revision codes for cirrhosis (571.0, 571.2, 571.5, 571.6) with Current Procedural Terminology procedure codes for cholecystectomy (47600, 47605, 47610, 47562, 47564), ventral or incisional hernia repair (49560, 49561, 49565, 49566, 49568), umbilical hernia repair (49585, 49587), enterectomy (44120, 44121, 44125, 44130, 44202, 44203), colectomy (44140, 44141, 44143, 44144, 44145, 44150, 44151, 44153, 44155, 44156, 44160, 44204, 44205, 44206, 44207, 44208, 44210, 44211, 44212), and appendectomy (44950, 44955, 44960, 44979). Minors, patients who underwent prior liver transplantation, and those undergoing hepatic resection or vascular, cardiovascular, or thoracic procedures were excluded. Etiology and diagnosis of cirrhosis were confirmed by patient medical history and liver biopsy. Data Collection Patient electronic medical records were reviewed for demographics and medical, surgical, and social history. Etiology of cirrhosis and accompanying disease complications including history of spontaneous bacterial peritonitis, presence of esophageal or gastric varices, portal hypertension, and refractory ascites requiring shunt placement were recorded. Presence of esophageal or gastric varices was confirmed by esophagogastroduodenoscopy and portal hypertension by radiographic imaging. CTP and MELD scores were calculated on the basis of preoperative laboratory values, grade of encephalopathy, and ascites. Grade of encephalopathy was determined according to the West Haven classification system. 12 Ascites was graded on a scale of escalating severity based on the International Ascites Club proposed guidelines. 13 Operative and anesthesia records were reviewed for ASA score, type of procedure, intraoperative time, intraoperative blood loss, transfusion requirement, presence of ascites at time of operation, gross description of the liver, and intraoperative complications. Acuity of operation was determined from anesthesia records. Emergent procedures were defined according to institutional guidelines. Our institution has 2 categories classifying emergent operations: category 1, which is defined as a threat to life or limb without intervention within 1 hour, or category 2, which is defined as a threat to life or limb without intervention within 6 hours. All patients within this study classified as either a category 1 or 2. Hospital course including intensive care unit (ICU) admission and length of stay (LOS), hospital LOS, and 30-day postoperative morbidity and mortality were assessed. Patient outcome at a mean of 36 months assessing hospital readmission or reoperation, necessity of liver transplantation, and mortality was recorded. Patient mortality was determined by medical record review and by cross-referencing patient social security number with the online social security death index database. 14 Approach to the Care of Cirrhotic Patients The Mount Sinai Medical Center performs the largest volume of hepatic resections nationwide and has performed more than 3100 liver transplants. For this reason, our institution has developed specific pathways pertaining to perioperative care of cirrhotic patients. Protocols for cirrhotic patients with advanced disease include strict guidelines regarding fluid management. Crystalloid administration is restricted both intraoperatively and postoperatively. For example, patients with advanced cirrhosis are often placed on an albumin drip as maintenance fluid postoperatively until an oral diet is tolerated. Specific protocols regarding medication administration and dietary management have also been developed. Ketorolac and epidural anesthesia are not used, and patients are often started on lactulose the first postoperative day to prevent encephalopathy. Hospital nutritionists have developed low sodium, low fat diets with adequate protein balance adapted specifically for cirrhotic patients. In addition, cirrhotic patients requiring general surgical procedures who are candidates for liver transplantation are evaluated by our multidisciplinary liver transplant service. At our institution, patients with MELD score 15 are considered candidates for liver transplant. For these patients, general surgical operations are performed by liver transplant surgeons, and patients are subsequently admitted to a multidisciplinary liver service where postoperative care is undertaken by physicians specializing in liver transplant medicine and transplant surgery. Statistical Analysis Univariate statistical analysis was performed by using unpaired Student t test with two-tail distribution for quantitative variables and 2 test for categorical values. Multivariate logistic regression models were used to estimate odds ratios (ORs) and associated 95% confidence interval (CI). Final multivariate models were created by elimination of nonsignificant variables from univariate analysis. P values of less than.05 for associations were considered to indicate statistical significance. Table 1. Major Postoperative (n) Wound complication (12) Infection (6) Hematoma (4) Leakage of ascites (2) Liver decompensation (12) Altered mental status (5) Worsening ascites (5) Shock liver (2) Postoperative ileus or obstruction (4) Respiratory failure (10) Tracheostomy (4) Decannulated (4) Sepsis (7) Postoperative variceal bleed (3) Anastomotic leak (1)

3 May 2010 ABDOMINAL OPERATIONS IN ADVANCED CIRRHOSIS 453 and mortality by MELD score. MELD scores 15 had a calculated OR of 5.0 with 95% CI of (P.001) for postoperative morbidity. Scores 17 had a 6.9 OR with 95% CI of (P.01) for postoperative mortality. Figure 1. and mortality by CTP class. Prism 4.0 statistical software (April 2003, San Diego, CA) was used for all analyses. Results Population Characteristics and Outcome One hundred cirrhotic patients qualified for the study. Mean patient age was 58.1 years, and 58% of patients were male. Twenty-eight patients had documented history of esophageal varices, 8 of spontaneous bacterial peritonitis, and 8 patients had a portosytemic shunt placed before operative procedure. Operative procedures consisted of 47 herniorrhaphies (35 umbilical and 12 ventral), 26 cholecystectomies, 17 colectomies, 3 appendectomies, 2 pancreaticoduodenectomies, and 5 other abdominal procedures. Mean ASA score was 3.0. Sixty-eight procedures were performed electively and 32 on an emergent basis. The overall 30-day morbidity and mortality rates were 43% and 7%, respectively. Patient morbidity data are shown in Table 1. Fourteen patients required ICU stay, with mean ICU stay of days (range, 2 58 days). Mean LOS was 6.3 days, 1.8 days for patients without and 12.1 days for patients with postoperative complications (P.001). Five patients underwent liver transplantation within 1 year of operative procedure. Follow-up at a mean of 36 months demonstrated an overall 9% mortality rate. Outcome by Child Turcotte Pugh Class and Model for End-Stage Liver Disease Score Fifty patients (50%) were classified as CTP A, 33% as CTP B, and 17% as CTP C. Figure 1 demonstrates operative morbidity and mortality by CTP class. Of the 17 CTP C patients, 5 required liver transplantation within 1 year of operative procedure; 3 30 days, 1 45 days, and one 240 days after operation. The 3 liver transplants that occurred within 30 days of operative procedure were a result of postoperative liver decompensation. Follow-up at a mean of 36 months demonstrated a 20% 1-year mortality rate. Six CTP C patients are currently alive and did not require liver transplantation, 5 who underwent umbilical herniorrhaphy and 1 who had cholecystectomy. Mean MELD score was 12.4; 33 patients had scores 15, and 3 patients had a score 26. Figure 2 demonstrates morbidity Outcome by Operative Procedure and Advanced Cirrhosis Table 2 shows intraoperative variables and LOS by operative procedure. Cholecystectomy. All procedures were performed by laparoscopic approach, with 2 open conversions as a result of technical difficulty. No emergent procedure was performed. Mean ASA score was 2.8. One death and 6 morbidities occurred. Five patients were classified as CTP B and 1 as CTP C. Two morbidities (40%) and 1 (20%) death occurred in CTP B patients and 1 morbidity (100%) in CTP C patients. Five patients had MELD 15, with 3 (60%) postoperative morbidities and 1 (20%) death. Umbilical herniorrhaphy. Of the 35 umbilical herniorrhaphies, 20 were elective and 15 emergent. Indications for emergent operation included acute incarceration with inability to reduce hernia (n 13) and umbilical perforation with leakage of ascites (n 2). Seven (35%) were performed by laparoscopic approach, and there were no open conversions. All laparoscopic procedures occurred in the elective setting. For those patients undergoing the laparoscopic approach, 6 (85%) had an uneventful postoperative course, and 1 (14%) developed postoperative morbidity (P.001). Mean ASA score was 3.0. One death (3%) and 20 (57%) morbidities occurred. Twelve patients were CTP B, and 13 were CTP C. Twenty-one patients had ascites at time of operation. and mortality rates for CTP B were 50% and 0% and for CTP C 77% and 8%, respectively. Eleven patients had MELD scores 15, with postoperative morbidity rate of 64% and 11% mortality rate. Of the 21 patients who had ascites, 15 patients presented with incarceration and 6 with spontaneous umbilical rupture. The overall mortality rate was 5%, and morbidity rate was 71%. Two patients required perioperative liver transplantation, and 5 developed ascites-related wound complications. Colectomy. Nine procedures were elective, and 8 were emergent. All elective procedures were attempted by laparoscopic approach, with 2 (22%) open conversions as a result of adhesions. Indications for emergent colectomy included colonic Figure 2. and mortality by MELD score.

4 454 TELEM ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 5 Table 2. Univariate Comparison of Intraoperative Variables and Postoperative LOS by Operative Procedure and Postoperative Complication Operative procedure Overall ( ) ( ) P value a Cholecystectomy (n 26) Mean operative time (min) NS Mean EBL (ml) Transfusion requirement PRBC 8% 0% 8%.003 Mean LOS (days) Umbilical hernia (n 35) Mean operative time (min) NS Mean EBL (ml) NS Transfusion requirement PRBC 9% 6% 3% NS Mean LOS (days) Colectomy (n 17) b Mean operative time (min) Mean EBL (ml) NS Transfusion requirement PRBC 35% 0% 35%.05 Mean LOS (days) NS EBL, estimated blood loss; PRBC, packed red blood cells. a P value represents univariate comparison of patients with uncomplicated versus complicated operative course. b Indications for colectomy: colorectal cancer (8), inflammatory bowel disease (4), diverticular disease (2), fulminant Clostridium difficile colitis (1), gastrointestinal bleed due to arteriovenous malformation (1), and colopleural fistula (1). perforation (n 4), gastrointestinal bleed (n 3), and fulminant colitis (n 1). Mean ASA score was 3.1. Four deaths (24%) and 6 morbidities (35%) occurred. Seven patients were classified as CTP B, with a total of 5 (71%) morbidities and 2 deaths (29%). One patient who was classified as CTP C died (100%). Eight patients had MELD 15, with 3 (38%) morbidities and 4 (50%) deaths. Risk Factors for Adverse Outcome Tables 3 and 4 demonstrate the results of univariate analysis of potential risk factors for adverse operative outcome. Mean platelet count was not significant; however, only 11 patients had preoperative platelet count / L. Of these patients, 1 was CTP A, 7 were CTP B, and 3 were CTP C. Analysis revealed a mortality rate of 18% (2/11) and morbidity rate of 54% for these patients. Of the 2 mortalities, 1 patient was CTP A who underwent right hemicolectomy, and the other was CTP C who underwent low anterior resection. Table 5 demonstrates the associated OR with 95% CI for significant predictors of adverse operative outcome after multivariate analysis of significant univariate variables. Albumin and Model for End-Stage Liver Disease Serum albumin in combination with MELD score significantly correlated with postoperative outcome. Addition of serum albumin to MELD score demonstrated that patients with MELD score 15 and albumin 2.5 mg/dl versus 2.5 mg/dl had significantly increased mortality or transplant rate (60% vs 14%, P.01) and independently increased probability of adverse outcome, OR of 8.4 and 95% CI of (P.015). Two patients (3%) with MELD score 15 had albumin 2.5 mg/dl. Both patients underwent uncomplicated umbilical herniorrhaphy. Discussion Operative decisions concerning cirrhotic patients are challenging. Identifying preoperative factors that might help determine appropriate operative candidates, the optimal timing of operative intervention, is that potentially reduce postoperative complications are of great importance. Our study has identified multiple factors, in addition to those already reported by current literature, that strongly influence postoperative course. Perhaps the most significant finding our data demonstrated was the association between preoperative serum albumin and MELD score. Preoperative serum albumin, in addition to MELD score, strongly correlated with postoperative outcome in patients with advanced cirrhosis. Patients with MELD scores 15 and albumin levels 2.5 mg/dl had a postoperative mortality or transplantation rate of 60% versus 14% in patients with albumin 2.5 mg/dl (P.01). In addition, albumin 2.5 mg/dl in association with MELD score 15 was an independent predictor (OR, 8.4; P.015) of adverse operative outcome. On the basis of this result, we recommend preoperative albumin be considered a criterion for operative decisions in cirrhotic patients with MELD 15. Because serum albumin is not a component of MELD, it might not fully be taken into account preoperatively at institutions that use MELD score to guide operative decisions. The basis for the low serum albumin remains unclear. Low albumin is often a result of malnutrition, a known risk factor for postoperative morbidity and mortality in cirrhotic patients. 15 This study, however, was not able to objectively assess nutritional status. The lower serum albumin could also be a marker for muscle wasting and decreased muscle stores or representative of difficulties with management of hyponatremia and ascites. Regardless, the subset of patients with MELD 15 and albumin 2.5 mg/dl appear to be poor operative candidates, and operative intervention will likely result in adverse outcome. Operative blood loss and intraoperative transfusion of packed red blood cells were also strong predictors of adverse outcome. Blood loss greater than 150 ml had an OR of 3.9, and intraoperative transfusion requirement had an associated OR of 16.8 for postoperative morbidity and mortality. Although mul-

5 May 2010 ABDOMINAL OPERATIONS IN ADVANCED CIRRHOSIS 455 Table 3. Univariate Analysis of Potential Preoperative Variables Influencing Operative Risk Preoperative parameter (n 57) % (n 43) % (n 12) % P value Mean age (y) 56.9 n/a 59.6 n/a 61.8 n/a NS Gender Male NS Female NS Mean body mass index 28.1 n/a 26.9 n/a 25.7 n/a NS Comorbidity NS Hypertension NS Diabetes NS Coronary artery disease NS Cardiac arrhythmia NS Cerebrovascular event NS End-stage renal disease NS COPD NS Hypothyroidism NS Etiology of cirrhosis a Hepatitis C NS Alcohol NS Hepatitis B NS Biliary NS Cryptogenic NS Autoimmune NS Other NS Prior SBP NS ( ) Esophageal/gastric varices NS ( ) Portal hypertension NS Preoperative TIPS Encephalopathy Grade Grade NS Grade NS Ascites Grade Grade NS Grade Mean laboratory values (n 57) 95% CI (n 43) 95% CI (n 12) 95% CI P value Total bilirubin (mg/dl) Albumin (mg/dl) Blood urea nitrogen (U/L) Prothrombin time (s) Partial thromboplastin time (s) Leukocyte count ( 10 3 / L) Hematocrit (%) Aspartate aminotransferase (U/L) NS Alanine aminotransferase (U/L) NS Alkaline phosphatase (U/L) NS Direct bilirubin (mg/dl) NS Sodium (meq/l) NS Creatinine (mg/dl) NS International normalization ratio NS Platelets ( 10 3 / L) NS COPD, chronic obstructive pulmonary disease; SBP, spontaneous bacterial peritonitis; TIPS, transjugular intrahepatic portosystemic shunt. a n 100, 4 patients with cirrhosis due to hepatitis C and alcohol. tivariate analysis demonstrated blood loss and transfusion requirement to be independent predictors of poor outcome, differentiating the effect of each variable is difficult. Increased blood loss might represent technical difficulty, whereas adverse events associated with transfusion might be due to operative difficulty or effects of transfusion on the immune system impeding recovery. 16 Future studies designed to distinguish the effects of transfusion and blood loss on outcome are necessary. At this time, however, the authors recommend limiting packed red blood cell transfusion when possible.

6 456 TELEM ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 8, No. 5 Table 4. Univariate Analysis of Potential Intraoperative Risk Factors for Adverse Outcome (n 57) 95% CI (n 43) 95% CI (n 12) 95% CI P value ASA score EBL (ml) (n 57) % (n 43) % (n 12) % P value Acuity Emergent Elective Approach Laparoscopic Conversion NS Open NS PRBC transfusion EBL, estimated blood loss; PRBC, packed red blood cells. Overall, the laparoscopic approach was associated with significantly decreased postoperative complications. This result might be confounded by selection bias, because it is more likely used in healthier patients presenting for elective operative procedures. To control for bias, the laparoscopic approach was assessed by operative procedure and acuity of operation. When doing so, laparoscopy was associated with a significantly decreased rate of postoperative complication in patients undergoing umbilical herniorrhaphy. A trend toward decreased morbidity was also demonstrated by other operative procedures; however, sample size precluded significance. Although study power is limited, the overall reduction in morbidity supports the growing body of literature demonstrating that the laparoscopic approach is safe and feasible in cirrhotic patients without ascites Consistent with other studies, both emergent surgery and more invasive operative procedures conferred worse postoperative outcomes. 21,22 This is not unexpected because emergent operative procedures are typically performed in patients with more advanced liver disease. In addition, the increased morbidity and mortality rates associated with colon resection in this study, as opposed to other operative procedures, are corroborated by other studies within the literature. 22 Although the rate and severity of postoperative complication demonstrated by this study are similar to reported literature, they did not translate to comparable mortality rates. This study demonstrated an overall mortality rate for CTP A, B, and C Table 5. Independent Risk Factors for Adverse Outcome After Multivariate Analysis of Significant Univariate Variables Parameter OR 95% CI P value Intraoperative transfusion Albumin 3 mg/dl ASA score Total bilirubin 1.5 mg/dl Emergent procedure Presence of ascites Blood loss 150 ml cirrhotics of 2%, 12%, and 12%, respectively, and for MELD scores a mortality rate of 29%. We attribute this success to institutional volume, experience with intraoperative and postoperative management of cirrhotic patients, and most importantly our model of multidisciplinary care. Thus, despite the same frequency and severity of complications, the experience of the team and preoperative identification of the potential for morbidity allowed such management to prevent high mortality rates. Although futures studies are necessary, regionalization of care for patients with advanced cirrhosis to centers specializing in liver medicine with transplant capability might improve outcome. Regionalization of care has already been demonstrated to improve outcome for several operative procedures and disease processes The major strength of this study is that it represents one of the largest single institution experiences involving patients with advanced cirrhosis. In contrast to population studies, our study was able to account for patient characteristics and assess factors influencing operative outcome. A major limitation of this study was the retrospective study design. Definitive therapeutic recommendations based on retrospective studies are not ideal; however, the infrequent nature of operative intervention in this patient subset limits the applicability of prospective or randomized control studies. Nonetheless, methodologic limitations of our study limit definitive conclusions. Another limitation was the exclusion of cirrhotic patients requiring operative intervention where a decision was made not to intervene. Thus, sicker patients might have been excluded from analysis. One final limitation was the ability to assess the influence of preoperative platelet count on postoperative outcome. Although overall platelet count did not influence patient morbidity and mortality, only 11 patients presented with preoperative platelet count below / L. Although mortality rate (18%) was comparable, this small sample size precluded significant analysis. Future studies assessing the impact of platelet count on postoperative outcome after abdominal surgery are necessary. Conclusions In summary, preoperative albumin strongly correlated with outcome in patients with MELD 15 and should be

7 May 2010 ABDOMINAL OPERATIONS IN ADVANCED CIRRHOSIS 457 considered a criterion guiding operative decisions. Patients with MELD 15 and serum albumin levels 2.5 mg/dl appear to be poor operative candidates, and adverse outcomes should be anticipated. Conversely, this study identified a subset of patients with advanced cirrhosis and preoperative serum albumin 2.5 mg/dl who might be suitable for operative procedures with better than anticipated outcome. Intraoperative packed red blood cell transfusion also independently correlated with adverse operative outcome (OR, 16.8) and should be limited when feasible. The lower than anticipated mortality rate for patients with advanced cirrhosis, demonstrated by this study, is attributed to institutional experience and multidisciplinary approach to patient care. Although future studies are necessary, regionalization of care for cirrhotic patients with advanced disease might improve mortality rates. References 1. Friedman SL. Hepatic fibrosis. In: Schiff ER, Sorrell MF, Maddrey WC, eds. Schiff s diseases of the liver. 8th ed. Philadelphia: Lippincott-Raven, 1999: Garrison RN, Cryer HM, Howard DA, et al. Clarification of risk factors for abdominal operations in patients with hepatic cirrhosis. Ann Surg 1984;199: Sleeman D, Namias N, Levi D, et al. Laparoscopic cholecystectomy in cirrhotic patients. J Am Coll Surg 1998;187: Befeler AS, Palmer DE, Hoffman M, et al. The safety of intraabdominal surgery in patients with cirrhosis: model for end-stage liver disease score is superior to Child-Turcotte-Pugh classification in predicting outcome. Arch Surg 2005;140: Mansour A, Watson W, Shayani V, et al. Abdominal operations in patients with cirrhosis: still a major surgical challenge. Surgery 1997;122: Wong R, Rappaport W, Witte C, et al. Risk of nonshunt abdominal operation in the patient with cirrhosis. J Am Coll Surg 1994;179: Csikesz NG, Nguyen LN, Tseng JF, et al. Nationwide volume and mortality after elective surgery in cirrhotic patients. J Am Coll Surg 2009;208: Teh SH, Nagorney DM, Stevens SR, et al. Risk factors for mortality after surgery in patients with cirrhosis. Gastroenterology 2007;132: Hoteit MA, Ghazale AH, Bain AJ, et al. Model for end-stage liver disease score versus Child score in predicting the outcome of surgical procedures in patients with cirrhosis. World J Gastroenterol 2008;14: Farnsworth N, Fagan SP, Berger DH, et al. Child-Turcotte-Pugh versus MELD score as a predictor of outcome after elective and emergent surgery in cirrhotic patients. Am J Surg 2004;188: Filsoufi F, Salzberg SP, Rahmanian PB, et al. Early and late outcome of cardiac surgery in patients with liver cirrhosis. Liver Transpl 2007;13: Blei AT, Córdoba J. Hepatic encephalopathy. Am J Gastroenterol 2001;96: 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: Social security death index [database online]. District of Columbia: United States Social Security Administration, Updated November 18, Merli M, Nicolini G, Angeloni S, et al. Malnutrition is a risk factor in cirrhotic patients undergoing surgery. Nutrition 2002;18: Corwin HL, Carson JL. Blood transfusion: when is more really less? N Engl J Med 2007;356: Cobb WS, Heniford BT, Burns JM, et al. Cirrhosis is not a contraindication to laparoscopic surgery. Surg Endosc 2005;19: Currò G, Iapichino G, Melita G, et al. Laparoscopic cholecystectomy in Child-Pugh class C cirrhotic patients. JSLS 2005;9: Cucinotta E, Lazzara S, Melita G. Laparoscopic cholecystectomy in cirrhotic patients. Surg Endosc 2003;17: Belli G, D Agostino A, Fantini C, et al. Laparoscopic incisional and umbilical hernia repair in cirrhotic patients. Surg Laparosc Endosc Percutan Tech 2006;16: Franzetta M, Raimondo D, Giammanco M, et al. Prognostic factors of cirrhotic patients in extra-hepatic surgery. Minerva Chir 2003;58: Meunier K, Mucci S, Quentin V, et al. Colorectal surgery in cirrhotic patients: assessment of operative morbidity and mortality. Dis Colon Rectum 2008;51: van Heek NT, Kuhlmann KF, Scholten RJ, et al. Hospital volume and mortality after pancreatic resection: a systematic review and an evaluation of intervention in the Netherlands. Ann Surg 2005; 242: Glasgow RE, Showstack JA, Katz PP, et al. The relationship between hospital volume and outcomes of hepatic resection for hepatocellular carcinoma. Arch Surg 1999;134: Harmon JW, Tang DG, Gordon TA, et al. Hospital volume can serve as a surrogate for surgeon volume for achieving excellent outcomes in colorectal resection. Ann Surg 1999;230: Reprint requests Address requests for reprints to: Celia M. Divino, MD, Department of Surgery, The Mount Sinai Medical Center, 5 East 98th Street, Box 1259, 15th Floor, New York, New York celia. divino@mountsinai.org; fax: (212) Conflicts of interest The authors disclose no conflicts.