查看更多>>摘要:Background and Aims: Aspartate aminotransferase-to-platelet ratio index (APRI) and fibrosis-4 index (FIB-4) are widely used to assess liver fibrosis in chronic hepatitis B virus (HBV) infection. Currently, the definition of normal alanine aminotransferase (ALT) is controversial. We aimed to exam-ine the diagnostic value of APRI and FIB-4 in chronic HBV carriers with different upper limits of normal (ULNs) for ALT. Methods: 581 chronic HBV carriers were divided into the fol-lowing four groups based on different ULNs for ALT: chronic HBV carriers I,Ⅱ,Ⅲ, and IV. Furthermore, 106 chronic HBV carriers formed an external validation group. Predictive val-ues of APRI and FIB-4 were elucidated using the area un-der the curve (AUC). A liver fibrosis-predictive model-GPSA (named for its measure of gamma glutamyl transpeptidase, platelet count, HBsAg and albumin) was developed using multivariate logistic regression analysis. Results: In chronic HBV carriers I, the AUCs of APRI and FIB-4 were 0.680 and 0.609 for significant fibrosis and 0.678 and 0.661 for cir-rhosis, respectively. The AUCs of GPSA for significant fibrosis in the training group, internal group, and external valida-tion group were 0.877, 0.837, and 0.871, respectively. The diagnostic value of GPSA differed among chronic HBV carri-ers I,Ⅱ,Ⅲ, and IV, with AUCs for significant fibrosis being 0.857, 0.853, 0.868, and 0.905 and AUCs for cirrhosis being 0.901, 0.905, 0.886, and 0.913, respectively. GPSA showed a higher diagnostic value than APRI and FIB-4 for predict-ing significant fibrosis in the four groups. Conclusions: The GPSA model allows for accurate diagnosis of liver fibrosis in chronic HBV carriers with different ULN for ALT.