HEPATOBILIARY ANZJSurg.com Predictors of mortality in cirrhotic patients undergoing extrahepatic surgery: comparison of Child–Turcotte–Pugh and model for end-stage liver disease-based indices Dong Hyun Kim,* Sung Hoon Kim,* Kyung Sik Kim,† Woo Jung Lee,† Nam Kyu Kim,† Sung Hoon Noh† and Choong Bai Kim† *Department of Surgery, Yonsei University Wonju College of Medicine, Wonju Severance Christian Hospital, Wonju, Korea and †Department of Surgery, Yonsei University College of Medicine, Yonsei University Health System, Seoul, Korea Key words extrahepatic surgery, liver cirrhosis, mortality. Correspondence Dr Sung Hoon Kim, Department of Surgery, Yonsei University Wonju College of Medicine, Wonju Christian Hospital, 162 Ilsan-dong, Wonju, Kangwon-do 220-701, Korea. Email: drgs01@yonsei.ac.kr D. H. Kim MD; S. H. Kim MD; K. S. Kim MD, PhD; W. J. Lee MD, PhD; N. K. Kim MD, PhD; S. H. Noh MD, PhD; C. B. Kim MD, PhD. Accepted for publication 25 March 2013. doi: 10.1111/ans.12198 Abstract Background: Underlying liver cirrhosis is associated with high morbidity and mortality after surgery. Previous studies have reported conflicting results about the value of Child–Turcotte–Pugh (CTP) and model for end-stage liver disease (MELD) scores as predictors of post-operative mortality. This study was designed to compare the capacities of CTP, MELD and MELD-based indices in predicting mortality for patients with liver cirrhosis who underwent elective extrahepatic surgery. Methods: The medical records of 79 patients with liver cirrhosis who underwent elective extrahepatic surgery under general anaesthesia from December 2000 to December 2009 were reviewed retrospectively. Results: The median follow-up period was 21 months, and the mortality rate was 24.1% (n = 19). Among the 19 mortalities, nine (11.4%) occurred while the patient was hospitalized after surgery. Intraoperative transfusion amount (ⱖ700 mL; odds ratio 6.294, P = 0.004) and the integrated MELD score (ⱖ34; odds ratio 6.654, P = 0.007) were significantly correlated with post-operative mortality. CTP score (hazard ratio 1.575, P = 0.012) was significantly correlated with overall mortality. Conclusions: Integrated MELD may be a more accurate predictor of operative mortality in cirrhotic patients undergoing extrahepatic surgery than CTP and other MELD-Na based indices. However, overall mortality may be reflected more accurately by CTP score. Further large-scale study will be needed to validate this result. Introduction It is well established that patients with underlying chronic liver disease (CLD) have higher rates of morbidity and mortality after surgery than those without.1,2 While the development of effective treatments has increased the average life expectancies of patients with CLD dramatically,3 surgeons have met another problem that patients with CLD may have benign or malignancy disease requiring surgery. Many with both CLD and benign or malignant disease are often recommended to avoid surgery because of perioperative risk.4,5 Consequently, these patients may eventually require surgery in emergency settings, increasing the perioperative risk even further.6–8 Preoperative assessment of liver function has been emphasized for patients with CLD. However, even with adequate preoperative assessment and perioperative management, operative mortality rates ANZ J Surg 84 (2014) 832–836 in patients with cirrhosis remain high. Mortality rates in patients with cirrhosis undergoing cholecystectomy and colectomy are 3.4fold and 3.7-fold higher, respectively, than in patients without cirrhosis.9 The Child–Turcotte–Pugh (CTP) score has been used to evaluate the risk of mortality in patients with CLD undergoing surgery. The model for end-stage liver disease (MELD) score is usually used to prioritize organ allocation in waiting lists for liver transplantation, and is used to assess risks after surgery. Several studies have published conflicting results comparing the predicting capabilities of these two scores. However, recent review article ultimately recommends using both models to evaluate risk.10 Recently, MELD-based indices have been proposed as additional risk assessment models.11 We compared the predictive value of mortality of CTP, MELD and MELD-based indices in patients with cirrhosis who underwent elective extrahepatic surgery. © 2013 The Authors ANZ Journal of Surgery © 2013 Royal Australasian College of Surgeons Extrahepatic surgery in cirrhosis Methods Patients Seventy-nine patients with liver cirrhosis underwent elective extrahepatic surgery between January 2000 and December 2009 at Severance Hospital, Yonsei University Health System in Seoul, Korea. We reviewed the medical records of these patients retrospectively. Definitions The MELD-based indices utilized in our study include sodium incorporated into MELD (MELD-Na), integrated MELD (iMELD) and MELD to sodium ratio (MESO). CTP, MELD and MELD-based indices were calculated by using preoperative data. Liver cirrhosis diagnoses were made based on clinical, laboratory and radiological findings as reported previously.7 Operative mortality was defined as death within 1 month of surgery or during hospitalization after surgery. Outcomes Operative mortality and overall mortality were assessed. Clinical and operative characteristics were analysed, and along with CTP, MELD and MELD-based indices, were evaluated for correlation with mortality. Statistical analysis Statistical analysis was performed using SPSS v15.0 (SPSS Inc., Chicago, IL, USA). All continuous variables are presented as a median (range) and all categorical variables as a number (percentage). Statistical analysis using a receiver operating characteristic (ROC) curve was performed to determine cut-off values for CTP, MELD and MELD-based scores. Cut-off values were validated by determining areas under the ROC curves. Univariate and multivariate analysis for prognostic factors of operative mortality was performed using a forward stepwise analysis of binary logistic regression. Univariate and multivariate analysis for prognostic factors of overall mortality was performed using a forward stepwise analysis of Cox’s proportional hazard model. Statistical significance was defined as P value < 0.05. Results Baseline characteristics The median patient age was 59 years old. Sixty-three (79.7%) patients were male. The most common causes of underlying liver disease were hepatitis B (n = 36, 45.6%) and alcohol (n = 27, 34.2%). Forty-one (51.9%) patients were classified as having CTP class A cirrhosis. Only one patient (1.3%) was classified as having CTP class C cirrhosis. CTP, MELD and MELD-based scores are presented in Table 1. Operative data Almost all patients underwent major organ resection, except for four (5.1%) who underwent wedge resection of the stomach. Twentyeight (35.4%) patients underwent stomach resection, and 23 (29.1%) patients underwent colon resection. Two patients underwent pancreatoduodenectomy. Operative data are listed in Table 2. Most surgical procedures except splenectomy were performed due to cancer treatment. © 2013 The Authors ANZ Journal of Surgery © 2013 Royal Australasian College of Surgeons 833 Table 1 Clinical characteristics (n = 79) Gender Female Male Age Cause of cirrhosis Hepatitis B virus Hepatitis C virus Alcohol Non-B, non-C virus Platelet, 103/mL Albumin, g/dL Total bilirubin, mg/dL International normalized ratio Creatinine, mg/dL Sodium, mmol/L CTP score CTP class A B C MELD score MELD-Na score iMELD sore MESO score 16 (20.3%) 63 (79.7%) 59 (20–84) 36 (45.6%) 9 (11.4%) 27 (34.2%) 7 (8.9%) 123 (19–293) 3.6 (2.3–4.9) 1.1 (0.2–8.2) 1.09 (0.75–2.12) 0.9 (0.6–11.9) 138.8 (116.1–144.0) 6 (5–12) 41 (51.9%) 34 (43.0%) 1 (1.3%) 10 (6–23) 11 (6–27) 30 (19–48) 7 (0–18) CTP, Child–Turcotte–Pugh; iMELD, integrated MELD; MELD, model for endstage liver disease; MELD-Na, model for end-stage liver disease-sodium; MESO, MELD to serum sodium ratio. ROC curve analysis ROC curve analysis was performed to determine cut-off values for operative and overall mortality for CTP, MELD and MELD-based indices. The cut-off values for CTP, MELD, MELD-Na, iMELD and MESO scores for operative mortality were 7, 11, 12, 34 and 8, respectively, and the transfusion amount was 700 mL. The cut-off values for overall mortality for these scores were 7, 10, 11, 32 and 7, respectively, and the transfusion amount was 400 mL. Surgical outcomes and prognostic factors There were nine (11.4%) post-operative mortality cases. Three patients underwent resection of the stomach, one patient underwent resection of the small bowel, and one patient underwent splenectomy. Two cases of mortality followed Hartmann’s operations. The other cases of mortality occurred after distal pancreatectomy and pancreatoduodenectomy. During the median 21-month follow-up periods, 19 (24.1%) patients died. Only one patient expired due to carcinomatosis. Prognostic factors for operative mortality are presented in Table 3. Intraoperative transfusions greater than 700 mL (odds ratio (OR) 6.294; 95% confidence interval (CI) 1.266–31.302; P = 0.025) and iMELD scores greater than 34 (OR 6.654; 95% CI 1.174–37.727; P = 0.032) were significantly correlated with operative mortality in both univariate and multivariate analyses. Prognostic factors for overall mortality by univariate analysis included estimated blood loss of more than 1200 mL, high CTP score, high CTP class, high MELD scores and high iMELD scores. However, by multivariate analysis, only CTP score (hazard ratio 1.575; 95% CI 1.107–2.241; P = 0.012) was significantly correlated with overall mortality (Table 4). Kim et al. 834 Table 2 Operation data Organ Oesophagus Stomach Small bowel Colon Spleen Pancreas Disease (n: cases) Operation Oesophageal cancer (3) Stomach cancer (2) Peptic ulcer (2) Stomach cancer (24) n (laparoscopic cases) Esophagectomy Wedge resection Hernia (3) Peritoneal tumour (2) AOV cancer (1) GI bleeding (1) Ascending colon cancer (6) Transverse colon cancer (1) Intestinal obstruction (1) Intra-abdominal abscess (1) Enterocutaneous fistula (1) Krukenberg tumour (1) Sigmoid colon cancer (8) Rectal cancer (1) Hepatic flexure, descending colon cancer (1) Sigmoid colon cancer (1) Enterocutaneous fistula (1) Hypersplenism (11) ITP (1) Variceal bleeding (2) Schwannoma (1) Pancreas pseudocyst (1) IPMT (1) AOV cancer (1) 3 4 (1) RTG† RSTG Segmental resection 5 (2) 19 (4) 7 (1) RHC‡ 9 (3) Segmental resection 2 AR or LAR 9 (3) Subtotal colectomy Hartmann 1 2 Splenectomy 12 (1) Splenectomy Kobayashi operation Splenectomy, wedge resection of stomach Distal pancreatectomy PD 2 1 1 2 †Two patients underwent radical total gastrectomy and splenectomy. ‡Two patients underwent right hemicolectomy and splenectomy. AOV, ampulla of Vater; AR, anterior resection; GI, gastrointestinal; IPMT, intraductal papillary mucinous tumour; ITP, idiopathic thrombocytic purpura; LAR, low anterior resection; PD, pancreatoduodenectomy; RHC, right hemicolectomy; RSTG, radical subtotal gastrectomy; RTG, radical total gastrectomy. Table 3 Univariate and multivariate analyses of operative mortality Variable Age Gender Underlying disease Laparoscopy Operation time Estimated blood loss Transfusion amount CTP score CTP class MELD MELD-Na iMELD MESO Univariate analysis Female HCV Alcohol Non-B, non-C Unknown No Male HBV Yes ⱖ1200 mL ⱖ700 mL B and C ⱖ11 ⱖ12 ⱖ34 ⱖ8 A Multivariate analysis OR 95% CI P value OR 95% CI P value 1.01 1.14 4.00 0.64 0.00 0.00 0.00 1.00 5.90 9.17 2.01 2.18 3.15 3.36 9.72 3.04 0.950–1.082 0.214–6.114 0.707–22.616 0.108–3.781 0 0 0 0.998–1.009 1.306–26.653 2.005–41.916 1.174–3.455 0.482–9.893 0.725–13.723 0.771–14.638 1.846–51.196 0.699–13.190 0.676 0.876 0.117 0.622 0.999 0.999 0.998 0.182 0.021 0.004 0.011 0.311 0.126 0.107 0.007 0.138 — — — — — — — — — 6.294 — — — — 6.654 — — — — — — — — — — 1.2665–31.302 — — — — 1.174–37.727 — — — — — — — — — — 0.025 — — — — 0.032 — CI, confidence interval; CTP, Child–Turcotte–Pugh; HBV, hepatitis B virus; HCV, hepatitis C virus; iMELD, integrated MELD; MELD, model for end-stage liver disease; MELD-Na, model for end-stage liver disease-sodium; MESO, MELD to serum sodium ratio; OR, odds ratio. Discussion We analysed the outcomes of 79 cirrhotic patients who underwent extrahepatic surgery. Previous studies have reported conflicting results about the capacities of CTP and MELD score as prognostic markers. Our results suggest that iMELD and intraoperative transfusion scores are the prognostic factors for operative mortality, and CTP class is the prognostic factor for overall mortality. The CTP score was originally introduced to predict prognoses for patients with liver disease undergoing surgery for portosystemic shunts; however, it has since been used to predict risks for extrahepatic surgeries. Indeed, Neal Garrison et al.12 and Mansour et al.8 have reported a 10% perioperative mortality risk following abdominal surgery for patients in CTP class A, 30–31% for patients in CTP class B and 76%–82% for patients in CTP class C. However, two major problems with CTP are that scoring involves subjective © 2013 The Authors ANZ Journal of Surgery © 2013 Royal Australasian College of Surgeons Extrahepatic surgery in cirrhosis 835 Table 4 Univariate and multivariate analyses of overall survival Variable Age Gender Underlying disease Laparoscopy Operation time Estimated blood loss Transfusion amount CTP score CTP class MELD MELD-Na iMELD MESO Univariate analysis Female HCV Alcohol Non-B, non-C Unknown No Male HBV Yes ⱖ600 mL ⱖ400 mL B and C ⱖ10 ⱖ11 ⱖ32 ⱖ8 A Multivariate analysis OR 95% CI P value OR 95% CI P value 1.01 1.26 2.28 0.78 0.00 1.73 0.18 1.00 1.99 3.337 1.58 3.63 3.24 3.05 3.25 2.11 0.962–1.049 0.415–3.812 0.7–7.454 0.26–2.331 0 0.215–13.946 0.024–1.326 0.998–1.006 0.766–5.156 1.289–8.636 1.206–2.064 1.279–10.368 1.065–9.873 1.002–9.287 1.211–8.689 0.813–5.465 0.823 0.685 0.171 0.654 0.985 0.606 0.092 0.319 0.158 0.013 0.001 0.016 0.038 0.05 0.019 0.125 — — — — — — — — — — 1.575 — — — — — — — — — — — — — — — 1.107–2.241 — — — — — — — — — — — — — — — 0.012 — — — — — CI, confidence interval; CTP, Child–Turcotte–Pugh; HBV, hepatitis B virus; HCV, hepatitis C virus; iMELD, integrated MELD; MELD, model for end-stage liver disease; MELD-Na, model for end-stage liver disease-sodium; MESO, MELD to serum sodium ratio; OR, odds ratio. parameters, including grading ascites and encephalopathy, and that patients are classified broadly into three categories. Thus, many have looked to MELD as a new and more useful prognostic tool for evaluating patients with CLD. Several groups have compared the predictive capacities of these two scores. Farnsworth et al. analysed the outcomes of 40 cirrhotic patients who underwent elective and emergent surgery. In this study, CTP classes A, B and C were comparable to MELD scores of less than 8, 9 to 16, and more than 17, respectively, in predicting 1- and 3-month mortality rates.12 Additionally, Hoteit et al. compared the ability of CTP and MELD scores to predict outcomes in 195 cirrhotic patients who underwent surgery. The areas under the curve of CTP and MELD were 0.696 and 0.755, respectively, and were not significantly different (P = 0.3).13 Some studies have suggested that MELD scores are superior to CTP;14,15 however, none have demonstrated a definitive superiority. Our results show that both CTP and MELD scores are significantly higher in mortality cases than in non-mortality cases (seven versus six, eleven versus nine), but do not correlate with operative mortality. Several reports have suggested that CTP score and MELD should both be considered in assessing the operative risk of a patient with liver cirrhosis.9,10,16 Patients with decompensated liver cirrhosis may present with many complications. Indeed, variceal bleeding, ascites and hepatorenal syndrome are major causes of mortality in these patients.17 Because patients with low MELD scores were dying of complications relating to ascites and hyponatremia before they could undergo liver transplantation, it was thought that a new predictive model was needed. Thus, levels of serum sodium were incorporated into the MELD score to more accurately reflect waiting list mortality.18–20 This MELD-Na was previously reported as an independent prognostic factor after both surgical and non-surgical treatment in patients with hepatocellular carcinoma,21,22 and it showed superiority to MELD in cirrhotic patients undergoing extrahepatic surgery.23 In our study, the median MELD-Na scores in mortality cases and non-mortality cases were 12.5 and 10, respectively (P = 0.028). However, MELD-Na did not correlate with operative and overall mortality, nor did it demonstrate superiority to the other models. © 2013 The Authors ANZ Journal of Surgery © 2013 Royal Australasian College of Surgeons To improve the prognostic accuracy of several models, iMELD24 and MESO25 were developed, and both demonstrated improved results. Among these, iMELD scores were reported as useful prognostic markers for operative mortality11 and overall mortality.7 In our study, iMELD scores greater than 34 had a significant correlation with operative mortality (OR 6.654; P = 0.032). We found that iMELD scores may assess the operative mortality risk more accurately than CTP, MELD and other MELD-based indices. Physical status, as classified by the American Society of Anesthesiologists, and age have been found as independent prognostic factors in some studies.15,26 Although our study did not adjust for physical status as a variable, age and sodium were reflected in iMELD score. Thus, iMELD may be the more relevant model among CTP, MELD and MELD-based indices. Intraoperative transfusion has also been reported as an independent prognostic factor.27,28 Our study corroborates these findings, as intraoperative transfusions of greater than 700 mL were significantly correlated with operative mortality. Some have reported prognostic factors of long-term mortality cirrhotic patients undergoing extrahepatic surgery.26,29 In contrast, we found that only CTP score was a predictor of long-term mortality. These results were contrary findings compared with our previous reports7 that the iMELD score was a significant prognostic factor. Emergency status was known as an important prognostic factor, and the extent of the operation might affect post-operative results.2 In consideration of these difference (emergency versus elective surgery and differences in operation severity), two studies might show contrary results.Additionally, this study had just nine cases of operative mortality and ten cases of mortality that occurred outside of the operations, and among these cases, one occurred due to carcinomatosis. Small events may also produce these unexpected contrary results. Our analysis of mortality was, therefore, limited to providing high levels of evidence due to retrospective design and small number of mortality cases. We endeavoured to identify the effects of the difference between operation types (open surgery versus laparoscopic surgery) or the extent of operation (minor operation versus major operation). However, we cannot find the difference between those variables 836 because there may be small cases in each group. We cannot find the indication of surgery in patients with CLD. However, laparoscopic surgery may decrease the amount of intraoperative bleeding, and laparoscopic surgery was suggested as a safe modality of surgery in patients with CLD. The role of laparoscopy should be needed to evaluate for identifying the indication of surgery. Additionally, our study had some limitations. First, it was a retrospective analysis and had a small sample size of 79 patients. Second, only one patient belonged to CTP class C. Previous studies reported the mortality rate as 76–82% in patients with CTP class C.8,12 Thereafter, many studies evaluated the prognosis of patients with CLD. Although a systematic review of those articles or review articles were analysed, there were few articles that included patients with CTP class C.9,30 In many cases, surgeons hesitated to perform elective operations on patients with CTP class C and preferred to wait. In our study, all surgeons hesitated to perform operations on patients with CTP class C, and then there was only one CTP class C patients in the study population. Therefore, our results had some degree of selection bias. In conclusion, iMELD scores may be more accurate markers of operative mortality than CTP and other MELD-Na based indices for cirrhotic patients undergoing extrahepatic surgery. Overall mortality may be reflected more accurately by CTP score. However, further study including more large patients must be needed. Acknowledgement This work was supported by a research grant from Yonsei University Wonju College of Medicine (YUWCM-2013-60). References 1. Csikesz NG, Nguyen LN, Tseng JF, Shah SA. Nationwide volume and mortality after elective surgery in cirrhotic patients. J. Am. Coll. Surg. 2009; 208: 96–103. 2. Ziser A, Plevak DJ, Wiesner RH, Rakela J, Offord KP, Brown DL. Morbidity and mortality in cirrhotic patients undergoing anesthesia and surgery. Anesthesiology 1999; 90: 42–53. 3. Jang J. Current status of liver diseases in Korea: liver cirrhosis. Clin. Mol. Hepatol. 2009; 15: 40–9. 4. Marrocco-Trischitta MM, Kahlberg A, Astore D, Tshiombo G, Mascia D, Chiesa R. Outcome in cirrhotic patients after elective surgical repair of infrarenal aortic aneurysm. J. Vasc. Surg. 2011; 53: 906–11. 5. Warnick P, Mai I, Klein F et al. Safety of pancreatic surgery in patients with simultaneous liver cirrhosis: a single center experience. Pancreatology 2011; 11: 24–9. 6. Del Olmo JA, Flor-Lorente B, Flor-Civera B et al. Risk factors for nonhepatic surgery in patients with cirrhosis. World J. Surg. 2003; 27: 647–52. 7. Kim SH, Han YD, Lee JG et al. MELD-based indices as predictors of mortality in chronic liver disease patients who undergo emergency surgery with general anesthesia. J. Gastrointest. Surg. 2011; 15: 2029–35. 8. Mansour A, Watson W, Shayani V, Pickleman J. Abdominal operations in patients with cirrhosis: still a major surgical challenge. Surgery 1997; 122: 730–6. 9. de Goede B, Klitsie PJ, Lange JF, Metselaar HJ, Kazemier G. Morbidity and mortality related to non-hepatic surgery in patients with liver cirrhosis: a systematic review. Best Pract. Res. Clin. Gastroenterol. 2012; 26: 47–59. 10. Bhangui P, Laurent A, Amathieu R, Azoulay D. Assessment of risk for nonhepatic surgery in cirrhotic patients. J. Hepatol. 2012; 57: 874–84. Kim et al. 11. Costa BP, Castro Sousa F, Serôdio M, Carvalho C. Value of MELD and MELD-based indices in surgical risk evaluation of cirrhotic patients: retrospective analysis of 190 cases. World J. Surg. 2009; 33: 1711–9. 12. Neal Garrison R, Cryer HM, Howard DA, Polk HC. Clarification of risk factors for abdominal operations in patients with hepatic cirrhosis. Ann. Surg. 1984; 199: 648–55. 13. 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: 1774–80. 14. Northup PG, Wanamaker RC, Lee VD, Adams RB, Berg CL. Model for end-stage liver disease (MELD) predicts nontransplant surgical mortality in patients with cirrhosis. Ann. Surg. 2005; 242: 244–51. 15. Teh SH, Nagorney DM, Stevens SR et al. Risk factors for mortality after surgery in patients with cirrhosis. Gastroenterology 2007; 132: 1261–9. 16. Cholongitas E, Papatheodoridis GV, Vangeli M, Terreni N, Patch D, Burroughs AK. Systematic review: the model for end-stage liver disease – should it replace Child-Pugh’s classification for assessing prognosis in cirrhosis? Aliment. Pharmacol. Ther. 2005; 22: 1079–89. 17. Bosch J, García-Pagán JC. Complications of cirrhosis. I. Portal hypertension. J. Hepatol. 2000; 32: 141–56. 18. Biggins SW, Kim WR, Terrault NA et al. Evidence-based incorporation of serum sodium concentration into MELD. Gastroenterology 2006; 130: 1652–60. 19. Biggins SW, Rodriguez HJ, Bacchetti P, Bass NM, Roberts JP, Terrault NA. Serum sodium predicts mortality in patients listed for liver transplantation. Hepatology 2005; 41: 32–9. 20. Ruf AE, Kremers WK, Chavez LL, Descalzi VI, Podesta LG, Villamil FG. Addition of serum sodium into the MELD score predicts waiting list mortality better than MELD alone. Liver Transpl. 2005; 11: 336–43. 21. Huo TI, Lin HC, Hsia CY et al. The MELD-Na is an independent shortand long-term prognostic predictor for hepatocellular carcinoma: a prospective survey. Dig. Liver Dis. 2008; 40: 882–9. 22. Cucchetti A, Ercolani G, Vivarelli M et al. Impact of model for endstage liver disease (MELD) score on prognosis after hepatectomy for hepatocellular carcinoma on cirrhosis. Liver Transpl. 2006; 12: 966–71. 23. Cho HC, Jung HY, Sinn DH et al. Mortality after surgery in patients with liver cirrhosis: comparison of Child-Turcotte-Pugh, MELD and MELDNa score. Eur. J. Gastroenterol. Hepatol. 2011; 23: 51–9. 24. Luca A, Angermayr B, Bertolini G et al. An integrated MELD model including serum sodium and age improves the prediction of early mortality in patients with cirrhosis. Liver Transpl. 2007; 13: 1174–80. 25. Huo TI, Wang YW, Yang YY et al. Model for end-stage liver disease score to serum sodium ratio index as a prognostic predictor and its correlation with portal pressure in patients with liver cirrhosis. Liver Int. 2007; 27: 498–506. 26. Kim SY, Yim HJ, Park SM et al. Validation of a Mayo post-operative mortality risk prediction model in Korean cirrhotic patients. Liver Int. 2011; 31: 222–8. 27. Neeff H, Mariaskin D, Spangenberg HC, Hopt UT, Makowiec F. Perioperative mortality after non-hepatic general surgery in patients with liver cirrhosis: an analysis of 138 operations in the 2000s using child and MELD scores. J. Gastrointest. Surg. 2011; 15: 1–11. 28. Telem DA, Schiano T, Goldstone R et al. Factors that predict outcome of abdominal operations in patients with advanced cirrhosis. Clin. Gastroenterol. Hepatol. 2010; 8: 451–7. 29. Causey MW, Steele SR, Farris Z, Lyle DS, Beitler AL. An assessment of different scoring systems in cirrhotic patients undergoing nontransplant surgery. Am. J. Surg. 2012; 203: 589–93. 30. Nicoll A. Surgical risk in patients with cirrhosis. J. Gastroenterol. Hepatol. 2012; 27: 1569–75. © 2013 The Authors ANZ Journal of Surgery © 2013 Royal Australasian College of Surgeons Copyright of ANZ Journal of Surgery is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.