Abstract / Summary
Background: The Lille score is an established tool to guide corticosteroid (CS) response and treatment in severe alcohol-associated hepatitis (AH), MELD score >20. However, its use in clinical trials has remained inconsistent. We performed this meta-analysis to compare mortality and infection rates between studies using Lille-guided versus those using CS not guided by the Lille score. Methods: We performed a systematic review and meta-analysis of randomized controlled trials (2007–2025) evaluating treatment arms using CS in severe AH. Pooled mortality and infection rates at 28, 90, and 180 days were calculated for Lille-guided and non–Lille-guided arms. Results: Fourteen corticosteroid-treated arms from 13 trials (N=1,294) were included, of which four studies (n=437) used CS guided by the Lille score. Mortality was lower in Lille-guided arms compared with non–Lille-guided arms at 28 days: [8 (2-14) vs 21 (13-29)] %, P=0.01, and at 90 days [15 (10-20) vs 40 (25-56)] %, P<0.001, but similar at 180 days [35 (16-54) vs 48 (34-62)] %, P=0.27. Infection rates at corresponding time points were 14 (11-17), 26 (14-38), and 32 (19-44), without differences between Lille- and non-Lille-guided CS arms. Substantial heterogeneity existed across studies, mainly due to differences in the publication year of the 28-day mortality outcome. Conclusion: Lille-guided CS use is associated with lower mortality across follow-up periods without a significant difference in infection rates. These findings support the use of the Lille score to guide CS therapy in patients with severe AH.