Development and validation of a scoring system to predict progression to acute-on-chronic liver failure in patients with acute exacerbation of chronic hepatitis B.

Abstract

The aim of this study was to develop and validate a scoring system to predict the progression to acute-on-chronic liver failure (ACLF) in patients with acute exacerbation (AE) of chronic hepatitis B (CHB). The baseline characteristics of 474 patients with AE of CHB were retrospectively reviewed; 280 and 194 patients were randomly assigned to the derivation and validation cohorts, respectively. Univariate risk factors associated with ACLF development were entered into a multivariate logistic regression. The score model was established, and its predictive value was evaluated by the receiver operating characteristic (ROC) curve and the area under the ROC curve (AUROC). Hepatitis B virus (HBV) DNA, international normalized ratio (INR) of prothrombin time, and patient age were identified as independent risk factors associated with progressing to ACLF. The prediction model was established as R = -13.323 + 0.553 × log HBV-DNA (copies/mL) + 3.631× INR + 0.053 × age. The AUROCs of our prediction model were higher than those of the Model for End-stage Liver Disease (MELD) and MELD-sodium (Na) for both cohorts. At the cut-off value of -2.43, our prediction model had higher sensitivity (87.5%), specificity (73.6%), positive predictive value (23.0%), positive likelihood ratio (3.30), and lower negative likelihood ratio (0.17) in the validation cohort than those of MELD and MELD-Na. The independent risk factors associated with progressing to ACLF in patients with AE of CHB are HBV-DNA, INR, and age. Our risk prediction model is useful for predicting the development of ACLF.