*Purpose: ACCELERATE-AH is a multicenter consortium studying early liver transplant (LT) for alcoholic hepatitis (AH). To inform surveillance and intervention strategies for post-LT alcohol use, we sought to identify pre-LT factors to predict early vs. later post-LT alcohol use, and if these patterns were associated with post-LT survival.

*Methods: In this multi-center longitudinal analysis, 11 US sites provided detailed pre-LT psychosocial, clinical, post-LT alcohol use, and survival data. Consecutive patients with clinically-diagnosed severe AH, no prior diagnosis of liver disease or AH, who received LT from 2006-2018, were included. Alcohol use post-LT was any evidence of alcohol use post-LT: by clinical interview, biochemical testing, including ethyl glucuronide (ETG) or phosphotidylethanol (PEth). Alcohol use was categorized by date of first drink post-LT: none, early (≤1 year post-LT), later (>1 year post-LT). To evaluate factors predicting early vs. later post-LT alcohol use, Cox regression was performed, with LT recipients with no post-LT alcohol use as reference group, and center clustering.

*Results: 140 LT recipients for AH survived to home discharge (69% male, median pre-LT abstinence 55 days, MELD-Na 39, Lille 0.79, 49% overt encephalopathy), with median post-LT follow-up of 2.5 years (IQR 1.5-4.4). Post-LT alcohol use was as follows: 91 (65%) none, 32 (23%) early use, 17 (12%) later use. The proportion with sustained alcohol use among early (11/32; 34%) vs. later (4/17; 24%) alcohol use was similar (p=0.43). Probability of any alcohol use post-LT at 1-, 3-, 5- years was 24% (95%CI: 18-32), 37% (95%CI: 29-46), 42% (95%CI: 33-53). In MV analysis, predictors of early alcohol use post-LT were younger age (HR 1.06, p<0.001) and overt encephalopathy at LT (HR 1.75, p=0.03), and of later alcohol use post-LT were female sex (HR 1.96, p=0.001), >10 drinks/day pre-hospitalization (HR 2.45, p=0.04) and prior failed rehab attempt (HR 2.12, p=0.007). After adjusting for MELD in separate bivariate models, both early (HR 6.36, p=0.001) and later (HR 2.27, p=0.09) alcohol use were associated with increased risk of post-LT death, though the association with later alcohol use did not reach statistical significance.

*Conclusions: Pre-LT factors associated with early versus later post-LT alcohol use are different, which may inform surveillance strategies for post-LT alcohol use. Early (vs. later) post-LT alcohol use appears to be more harmful as evidenced by the higher mortality risk. This highlights the first year post-LT as an especially important period to target interventions to prevent and treat alcohol use.

« Back to 2019 American Transplant Congress