LAUSR

BACKGROUND The prognosis of hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF) is extremely poor due to multiple organ dysfunction. OBJECTIVES To investigate the prognostic risk factors and create a 90-day prognostic predictive model for the patients with HBV-ACLF. METHODS Demographic information, clinical examination, and laboratory test results of the enrolled patients were collected to study the prognostic risk factors. Univariate and multivariate analysis and stepwise Logistic regression were performed to develop the predictive model. External validation was performed to verify the model. RESULTS A total of 333 HBV-ACLF patients and 86 HBV-non-ACLF patients were included in this study. Age, alpha-fetoprotein (AFP), total bilirubin (TBIL), platelet (PLT), and international normalized ratio (INR) were found to be independent risk factors for poor outcomes of HBV-ACLF patients. The formula identified for the linear predictor (LP) of the prognosis of HBV-ACLF patients is thus: LPACLF=-5.04-0.056×age-0.002×AFP-0.010×PLT+0.002×TBIL+0.877×INR. The area under curve (AUC) of the receiver operating characteristic curve (ROC) was 0.7835 (95% CI 0.7248-0.8423). CONCLUSIONS A predictive model with good calibration and discrimination for 90-day survival of HBV-ACLF patients, including 5 variables, namely age, AFP, PLT, TBIL, and INR was established. Platelet count was a sensitive and dynamic variable for the prognosis of HBV-ACLF.