Validation of MELD 3.0 scoring system in East Asian patients with cirrhosis awaiting liver transplantation

Abstract

Recently, a new predictive model that jointly considers the Model of End-stage Liver Disease (MELD) 3.0 and albumin has been proposed. This study investigated the performance of the MELD 3.0 score in predicting the 3-month survival of East Asian patients with cirrhosis compared with the other MELD-based scores. Validation was performed with the retrospective data of 2153 patients in South Korea who were listed for liver transplantation (LT). Discrimination and calibration analyses were performed using the MELD-based scores as an independent variable. On average, patients had the original MELD score of 18.70 ± 9.65. Alcohol (39.99%) and chronic HBV (38.55%) were the 2 main etiologies. The MELD 3.0 with albumin showed slightly better discrimination [c-index = 0.738, incremental AUC (iAUC) = 0.719] compared with the MELD 3.0 without albumin (c-index = 0.737, iAUC = 0.715), MELD-Na (c-index = 0.730, iAUC = 0.707), or the original MELD (c-index = 0.718, iAUC = 0.687) for predicting 3-month survival but not significantly different compared with prior models. Likewise, in the stratified analysis according to the strata of MELD, although the performance of MELD 3.0 was better throughout all the MELD strata than MELD original, there was no statistical difference in performance. The MELD 3.0 with albumin reclassified 22.61% of cases classified by the original MELD to higher MELD score categories, and there was no significant difference in the reclassification rate between males and females. The predictive power of the MELD-based system is lower in Asian populations than in western countries. Nonetheless, the MELD 3.0 score with albumin was significantly better in predicting the short-term prognosis of East Asian patients on the LT waitlist than the current allocation system, original MELD.