Noninvasive models to diagnose liver histology in HBeAg-negative chronic hepatitis B

Abstract

Abstract Introduction and objectives: Noninvasive, practical, and convenient means of assessment of liver histology in China are greatly needed.This research was to develop precise non-invasive tests to assess the histological activity grade and the degree of fibrosis. Patients and Methods: We included 141 patients with HBeAg-negative chronic hepatitis B patient who was treatment-naive and underwent liver biopsy, whose ALT was below twice the normal upper limit (ULN). Logistic regression analysis was performed, noninvasive prediction models were established to diagnose the degree of liver inflammatory and fibrosis. Results: The independent risk factors of moderate or severe inflammation were aspartate aminotransferase(AST) and Prealbumin. Based on the binary sequential logistic analysis,an API model consists of AST and PA is proposed, and the AUROC value was 0.789 (95%CI=0.682-0.896). AST, Plt, HBsAg(Log10) group were independent risk factors of significant fibrosis. Based on the binary sequential logistic analysis, a HAPI model consists of AST, Plt, HBsAg(Log10) group is proposed,with the AUROC of 0.825,and the optimum cut-off, the sensitivity and the specificity were 0.290, 66.67% and 91.49%., respectively. The AUROC of HAPI in predicting significant fibrosis were greater than that of AST(0.669, 95%CI=0.571-0.767, p=0.0014) , Plt(0.745, 95%CI=0.660-0.830, p=0.0216) and FIB-4(0.727, 95%CI=0.637-0.816, p=0.0112). Conclusions: The API and HAPI should become convenient and noninvasive markers to predict the degree of ncerosis, inflammation and fibrosis in HBsAg-negative chronic hepatitis B infection whose ALT was lower than 2 ULN.