RESUMO
BACKGROUND: We evaluated liver fibrosis in patients with chronic hepatitis B (CHB) and mildly raised alanine transaminase (ALT) activities between 1-2 times the upper limit of normal (ULN) which was near the threshold for initiating treatment. METHODS: Nomogram-Fibrosis and Nomogram-Cirrhosis were elaborated with variables independently associated with significant fibrosis and cirrhosis determined by multivariate logistic regression. Calibration, receiver operator characteristic (ROC) and decision curves were applied to comparing nomograms with aspartate aminotransferase (AST) to platelet count (PLT) ratio index (APRI), age-AST-PLT-ALT index (FIB-4) and liver stiffness measurement (LSM). RESULTS: The Nomogram-Fibrosis was constructed with LSM, PLT, and gamma-glutamyl transpeptidase (GGT). Nomogram-Cirrhosis contained one more variable of age other than Nomogram-Fibrosis. The calibration demonstrated that the assessments of significant fibrosis or cirrhosis by nomograms were in line with liver biopsy. The AUROC of Nomogram-Fibrosis was 0.788, lager than APRI (0.586), FIB-4 (0.656) and LSM (0.735). The AUROC of Nomogram-Cirrhosis was 0.889, larger than APRI (0.642), FIB-4 (0.725) and LSM (0.837). Furthermore, the decision curve analysis suggested the most net benefits were provided by the nomograms. CONCLUSIONS: Nomogram-Fibrosis and Nomogram-Cirrhosis could be promising tools for recognizing significant fibrosis and cirrhosis for CHB patients with mild raised ALT activities.