Summary
Liver biopsy is the reference method for antiviral therapy decision‐making in chronic hepatitis B (CHB) when alanine aminotransferase (ALT) is less than two times of upper limit of normal (<2ULN). Our aim was to explore noninvasive markers for antiviral therapy decision in CHB with ALT <2ULN. A total of 452 treatment‐naïve CHB patients with ALT < 2ULN who had undergone liver biopsy were analysed in this prospective multi‐centre study. If liver biopsy showed moderate or severe inflammation (histology activity index ≥ 5) or significant fibrosis (Ishak fibrosis score ≥ 3), antiviral treatment was recommended. We analysed data using univariate and multivariate analyses and receiver operating characteristic curves (ROC). Two hundred and sixty‐nine (59.5%) of 452 cases with ALT < 2ULN had moderate, severe or significant inflammation. Aspartate aminotransferase (AST) (P = 0.03), anti‐hepatitis B virus core antibody (anti‐HBc) (P = 0.003) and liver stiffness measurement (LSM) (P = 0.000) were independent variables for antiviral therapy decision‐making, with area under the ROC curve (AUROC) of 0.718, 0.703 and 0.819, respectively. Our novel AAF index, which combined AST, anti‐HBc and LSM, showed better performance with AUROC of 0.876, 0.877 and 0.876 in estimation, validation and total set. Finally, 247 (54.6%) of 452 patients could avoid liver biopsy based on AAF index. Furthermore, performances of 23 noninvasive models were unsatisfactory for antiviral therapy decision with AUROC < 0.800, which were inferior to AAF index. In conclusion, AST, anti‐HBc and LSM were related to antiviral therapy decision‐making among CHB patients with ALT < 2ULN. Thus, the novel AAF index was a more reliable noninvasive model for antiviral therapy decision‐making.