Predicting early allograft failure and mortality after liver transplantation: the role of the postoperative model for end-stage liver disease score.

Early allograft dysfunction (EAD) is a serious complication after liver transplantation (LT). There is no uniform definition of EAD, and most definitions are based on arbitrary laboratory values. The aim of this study was to devise a definition of EAD that maximizes the predictive power for early death and graft failure. In this single-center, retrospective study, the ability of the international normalized ratio (INR), total bilirubin, aspartate aminotransferase (AST), physiological Model for End-Stage Liver Disease (MELD) score, and serum albumin levels within 7 days after LT to predict 90-day mortality or graft loss was compared with 2 previously used definitions of EAD: (1) peak total bilirubin level >10 mg/dL on days 2 to 7 and (2) either a total bilirubin level >10 mg/dL or an INR >1.6 on day 7 or an AST or alanine aminotransferase level >2000 IU/L within the first 7 days. Of 572 enrolled LT patients 38 died or required retransplantation within 90 days. Peak INR, total bilirubin level, AST levels, and MELD scores were predictors of 90-day graft failure. MELD score on postoperative day 5 was the best predictor with an area under the curve of the receiver operating characteristic curve of 0.812 (95% CI: 0.739-0.886, P < 0.001). The best cutoff of MELD score on day 5 for predicting 90-day mortality or graft loss was 18.9. A MELD score >18.9 on postoperative day 5 was a better predictor than any other laboratory value or definition of EAD. This study has demonstrated that the MELD score can be a useful tool not only for pretransplant graft allocation but also for postoperative risk stratification.