Predicting harmful alcohol relapse after liver transplant: The HALT score

Background Alcohol-associated liver disease (AALD) is a rapidly growing indication for liver transplantation (LT). We aimed to examine various clinical, demographic, and behavioral factors to predict post-LT alcohol relapse and graft survival. Methods Retrospective analysis was performed on 241 LT recipients with AALD as either a primary or secondary indication for LT (2006-2015). Results Patients with 6 months of abstinence (P = .0041, Log-Rank). We identified four variables to predict harmful alcohol relapse post-LT: age at LT, non-alcohol-related criminal history, pre-LT abstinence period (Ref >6 months of alcohol abstinence), and drinks per day (Ref <10 drinks/day). Area under the curve (AUC) for the final model was 0.79 (95% CI: 0.68-0.91). Our multivariable model was evaluated with internal cross-validation; random sampling of the study subjects 100 times yielded a median C statistic of 75 (+/- SD 0.097) and accuracy of 91 (+/- SD 0.026). The four-variable model served to form the harmful alcohol use post-LT (HALT) score. Graft survival remained significantly lower in patients with <6 months of pre-LT alcohol abstinence and those with blue-collar jobs. Conclusion The HALT score identifies LT candidates with AALD at significant risk for alcohol relapse, potentially guiding transplant centers for pre- and post-LT interventions for improved patient outcomes.