Model for End-Stage Liver Disease, Model for Liver Transplantation Survival and Donor Risk Index as predictive models of survival after liver transplantation in 1,006 patients

Abstract

Objectives: Liver transplantation has not increased with the number of patients requiring this treatment, increasing deaths among those on the waiting list. Models predicting post-transplantation survival, including the Model for Liver Transplantation Survival and the Donor Risk Index, have been created. Our aim was to compare the performance of the Model for End-Stage Liver Disease, the Model for Liver Transplantation Survival and the Donor Risk Index as prognostic models for survival after liver transplantation.

Method: We retrospectively analyzed the data from 1,270 patients who received a liver transplant from a deceased donor in the state of São Paulo, Brazil, between July 2006 and July 2009. All data obtained from the Health Department of the State of São Paulo at the 15 registered transplant centers were analyzed. Patients younger than 13 years of age or with acute liver failure were excluded.

Results: The majority of the recipients had Child-Pugh class B or C cirrhosis (63.5%). Among the 1,006 patients included, 274 (27%) died. Univariate survival analysis using a Cox proportional hazards model showed hazard ratios of 1.02 and 1.43 for the Model for End-Stage Liver Disease and the Model for Liver Transplantation Survival, respectively (p<0.001). The areas under the ROC curve for the Donor Risk Index were always less than 0.5, whereas those for the Model for End-Stage Liver Disease and the Model for Liver Transplantation Survival were significantly greater than 0.5 (p<0.001). The cutoff values for the Model for End-Stage Liver Disease (≥29.5; sensitivity: 39.1%; specificity: 75.4%) and the Model for Liver Transplantation Survival (≥1.9; sensitivity 63.9%, specificity 54.5%), which were calculated using data available before liver transplantation, were good predictors of survival after liver transplantation (p<0.001).

Conclusions: The Model for Liver Transplantation Survival displayed similar death prediction performance to that of the Model for End-Stage Liver Disease. A simpler model involving fewer variables, such as the Model for End-Stage Liver Disease, is preferred over a complex model involving more variables, such as the Model for Liver Transplantation Survival. The Donor Risk Index had no significance in post-transplantation survival in our patients.

Figure 1

A) The Kaplan-Meier curve for the survival of 1,006 patients who received a liver transplant in the state of São Paulo between 2006 and 2009. B) The ROC curves of the MELD, MLTS, and DRI scores for 2-year mortality for 1,006 liver transplants performed in the state of São Paulo between 2006 and 2009. C) The Kaplan-Meier curves for mortality according to the MELD scores for 1,006 liver transplants performed in the state of São Paulo between 2006 and 2009. D) The Kaplan-Meier curves for mortality according to the MLTS scores for 1,006 liver transplants performed in the state of São Paulo between 2006 and 2009.