Prognostic factors and development of a predictive model for hepatitis B virus-related acute-on-chronic liver failure

WU Xi-rong; XIAO Zhi-yang; GU Lin

doi:10.13820/j.cnki.gdyx.20242933

Abstract：Objective To identify clinical indicators that influence the prognosis of hepatitis B virus-related acute-on-chronic liver failure(HBV-ACLF)patients and to establish a novel predictive model for HBV-ACLF prog-nosis.Methods We conducted a retrospective analysis of data from 415 HBV-ACLF patients admitted to The Third Af-filiated Hospital of Sun Yat-sen University between 2017 and 2020.Patients were divided into a survival group(321 ca-ses)and a mortality group(94 cases).Univariate and multivariate logistic regression analyses were used to determine prognostic factors,and a new predictive model was developed.The model's discrimination ability was assessed with the ar-ea under the receiver operating characteristic(ROC)curve(AUC),and calibration was evaluated using a calibration curve.Results A total of 415 patients were analyzed,with a 90-day mortality rate of 22.7％(94/415).Stepwise lo-gistic regression identified age,neutrophil-to-lymphocyte ratio(NLR),prothrombin activity(PTA),procalcitonin(PCT),and hepatic encephalopathy(HE)as independent prognostic factors for 90-day mortality.The predictive index(PI)for the new model was:Logit(P)=In(P/(1-P))=-3.316+0.057 × age+0.076 × NLR-7.095 × PTA+0.693 × PCT+0.759 × hepatic encephalopathy(1 for those with hepatic encephalopathy and 0 for those without hepatic encephalopathy).Compared with the MELD score,the new model showed a significantly higher AUC of 0.796(95％CI:0.741-0.851)compared to MELD's AUC of 0.701(95％CI:0.640-0.762;Z=3.084,P=0.002).Conclusion Age,NLR,PTA,PCT,and hepatic encephalopathy are independent prognostic factors for short-term outcomes in HBV-ACLF patients.The new predictive model provides a more accurate 90-day prognosis for HBV-ACLF patients compared to MELD,offering a valuable tool for clinical risk stratification.