Validation of the five-variable Model for End-stage Liver Disease (5vMELD) for prediction of mortality on the liver transplant waiting list

Background: Modifications to the Model for End-Stage Liver Disease (MELD) have been proposed to improve prioritization of liver transplant (LT) candidates. Using a U. S. database, we derived a revised MELD including sodium and albumin [5-variable MELD (5vMELD)] that improved prediction of waiting list mortality. Our objectives were to confirm the association between hypoalbuminaemia and mortality and to externally validate 5vMELD in Canadian LT candidates. Methods: Among adults registered on the LT waiting list at the University of Alberta (01/2000-10/2009), Cox regression determined the association between albumin and 1-year waiting list mortality. The discrimination of MELD, MELDNa and 5vMELD for predicting 1-year mortality were compared using c-statistics. Results: Among 677 patients, 17% died and 51% underwent LT within 1 year of listing. Median serum albumin was 3.1 g/dl (IQR 2.6-3.6) and 70% of patients were hypoalbuminaemic (albumin <3.5 g/dl). One-year mortality in patients with normal serum albumin and hypoalbuminaemia were 14% and 29% respectively (P = 0.004). For patients with serum albumin between 2.0 and 4.0 g/dl, an approximately linear, inverse relationship was observed between albumin and 1-year mortality [adjusted hazard ratio (HR) 1.45; 95% CI 1.03-2.03; P = 0.03]. For this outcome, the c-statistic of 5vMELD (0.778) was superior to those of MELD (0.754) and MELDNa (0.765) (both P <= 0.05). Conclusions: Hypoalbuminaemia is an independent predictor of mortality on the LT waiting list. Compared with MELD and MELDNa, 5vMELD improves prediction of mortality suggesting that modification of these scores to include serum albumin should be considered as a means of prioritizing LT candidates.