2024.10.23-HBV-ACLF.rar

We performed a retrospective analysis of clinical data in two cohorts of subjects diagnosed with HBV-ACLF. Seven indices were identified as independent risk factors associated with HBV-ACLF in the derivation cohort. Then the logistic regression and nomogram models were used to build three novel predictive models. These models’ AUROC, sensitivity, and specificity were evaluated and compared with classic prognostic scoring systems, respectively. Finally, we further verified the predictive value of the new models for HBV-ACLF in the validation cohort. The model based on seven predictors, including four laboratory indicators and three clinical signs, could accurately predict short-term mortality in HBV-ACLF patients, which was promising for guiding clinical management.