Recognizing the limitations of CLIF-C ACLF in predicting outcomes following liver transplantation in acute-on-chronic liver failure

Acute-on-chronic liver failure (ACLF) is associated with a high short-term mortality. The Chronic Liver Failure-SOFA Consortium ACLF score (CLIF-C ACLF) predicts mortality in patients with ACLF, but its utility in predicting post-liver transplantation (LT) outcomes remains unclear. This study evaluated the predictive ability of CLIF-C ACLF score and its dynamic changes (ΔCLIF-C ACLF) for post-LT outcomes. We conducted a retrospective study of 521 patients with ACLF who underwent LT at Mayo Clinic Florida between January 2010 and December 2021. Prognostic scores were calculated at ICU admission and at transplantation, with ΔCLIF-C ACLF reflecting changes between these time points. Survival was analyzed using the Kaplan-Meier method, and predictive accuracy was assessed using the concordance index (C-index). The CLIF-C ACLF score at transplant demonstrated poor discrimination for overall survival (C-index: 30-day 0.55 [95% CI 0.49–0.61], 90-day 0.54 [0.48–0.60], 1-year 0.53 [0.47–0.58]). The ΔCLIF-C ACLF showed similarly limited discrimination (30-day 0.57 [0.52–0.63], 90-day 0.58 [0.52–0.63], 1-year 0.57 [0.52–0.62]), with no significant difference at any time point ( p > 0.05). For graft survival, the static score outperformed the dynamic score at all time points (30-day 0.71 [0.58–0.83], 90-day 0.71 [0.59–0.83], 1-year 0.72 [0.61–0.83] vs. 0.67 across all time points; p ≤ 0.008). The CLIF-C ACLF score and its dynamic change have limited utility in predicting post-LT overall survival. However, the static score at transplant demonstrated modest and significantly superior discrimination for graft survival compared to the dynamic score. High pre-transplant CLIF-C ACLF scores should not preclude transplantation, as favorable outcomes were observed even in the highest-risk patients. More robust post-LT prognostic models are needed.