Improving Matching Between Recipient and Donor in Liver Transplantation: Are Scores Able to Predict Long-term Survival?

Background: The primary goal was to calculate the value of delta Model for End-Stage Liver Disease (D-MELD) and Balance of Risk (BAR) scores in patients who underwent liver transplant. The secondary objective was to evaluate D-MELD and BAR scores' ability to predict patient and graft survival.
Methods: We retrospectively evaluated 336 patients who underwent liver transplant in a tertiary medical center between January 2010 and December 2020. The D-MELD and the BAR scores were evaluated through a receiver operating characteristic curve with the calculation of area under the curve (AUC) to evaluate the predictive score power for 3-month, 6-month, 1-year, and 5-year patient and graft survivals.
Results: The AUCs of D-MELD score in predicting 5-year patient and graft survival were 0.506 (95% CI, 0.43-0.57) and 0.49 (95% CI, 0.42-0.56), respectively. The AUCs of BAR score in predicting 5-year patient and graft survival were 0.50 (95% CI, 0.33-0.66) and 0.49 (95% CI, 0.30-0.67), respectively.
Conclusions: We could not confirm the ability to predict long-term survival by using D-MELD and BAR scores in our sample; however, there is a statistically significant trend in receiver operating characteristic curves of 5-year patient and graft survivals. We encourage the use of new scoring systems with a greater external validation to improve allograft allocation.
(Copyright © 2021 Elsevier Inc. All rights reserved.)