Association of nucleos(t)ide analogue treatment history with mortality in HBV-DNA-positive HBV-ACLF: a multicenter cohort study

Background and aims Hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF) is associated with high short- and long-term mortality. Prior nucleos(t)ide analogue (NA) treatment history, particularly treatment withdrawal and virological resistance, may contribute to ACLF development and influence prognosis after onset. However, the clinical impact of different NA treatment histories in patients with HBV-ACLF remains insufficiently characterized. This study evaluated the associations of baseline HBV-DNA status and prior NA treatment history with mortality in patients with HBV-ACLF. Methods This multicenter retrospective cohort study included hospitalized patients with HBV-ACLF from three tertiary hospitals in Chongqing, China, between January 2018 and December 2024. Patients were stratified according to baseline HBV-DNA status. Among HBV-DNA-positive patients, prior NA treatment history was categorized as NA naïve, NA unsupervised withdrawal, or NA suboptimal response(NA-SR). Cox proportional hazards models were used to evaluate associations with 28-, 90-, 180-, and 360-day mortality. Propensity score matching was applied to reduce baseline confounding. Because of the small size of the NA-SR group, these analyses were considered exploratory. Results Among 702 patients with HBV-ACLF, 625 were HBV-DNA-positive and 77 were HBV-DNA-negative at admission. Baseline HBV-DNA positivity showde a consistend trendtoward increased 28-day mortality (HR 2.19, 95% CI 0.83-5.83; P = 0.115). Among HBV-DNA-positive patients, compared with NA-naïve patients (n=480; 136 events), prior NA unsupervised withdrawal (n=127; 47 events) was associated with increased 360-day mortality (adjusted HR 1.41, 95% CI 1.00-2.00; P = 0.051), while NA suboptimal response (n=18; 11 events) showed a higher point estimate for 360-day mortality (adjusted HR 2.68, 95% CI 1.43-5.03; P = 0.002). Propensity score-matched analyses yielded directionally consistent results. Conclusion Among HBV-DNA-positive patients with HBV-ACLF, prior NA unsupervised withdrawal showde a borderline associated with worse long-term survival, and NA suboptimal response showed a stronger and statistically significant association with worse long-term survival that requires confirmation in larger cohorts. These findings highlight NA treatment history as a clinically relevant disease-background factor in HBV-ACLF and support careful documentation of antiviral exposure, avoidance of unsupervised treatment discontinuation, and timely monitoring for virological breakthrough in chronic HBV care.