Abstract
Background/Objectives: Alcohol-related acute-on-chronic liver failure (ACLF) carries high early mortality, but prognostic evidence varies by definition, endpoint, and adjustment. We estimated short-term mortality and synthesized clinically useful predictors. Methods: PubMed/MEDLINE, Scopus, and Web of Science were searched to 25 June 2026 and, with a revised strategy, to 28 September 2026. Direct alcohol-related ACLF cohorts formed the core evidence base. Mortality was pooled using REML random-effects models with Hartung–Knapp confidence intervals and 95% prediction intervals (PIs); predictors were synthesized qualitatively. Risk of bias and certainty were assessed with QUIPS and GRADE. Results: Forty reports (35 independent cohorts) were included. Pooled mortality was 36.0% at 28/30 days (11 cohorts; 95% CI 28.9–43.8; PI 17.3–60.3%) and 50.2% at 90 days (11 cohorts; 95% CI 39.5–60.8; PI 19.2–81.0%), without significant differences between APASL/AARC- and EASL-CLIF-defined cohorts. Qualitatively, organ-failure burden, infection or sepsis, hepatic encephalopathy, renal dysfunction, shock, and higher MELD-, CLIF-, or AARC-based severity were the most consistent adverse signals; pooled predictor effects could not be estimated. Conclusions: About one-third of patients die by 28/30 days and half by 90 days, but mortality varies widely across settings. Predominantly very low-certainty predictor evidence supports framework-appropriate scores and repeated assessment rather than a universal cutoff.