Clinical Non-invasive Model to Predict Liver Inflammation in Chronic Hepatitis B With Alanine Aminotransferase ≤ 2 Upper Limit of Normal

Background and Aim: Liver biopsy remains the gold standard for evaluating liver histology. However, it has certain limitations, and many patients refuse it. Non-invasive methods of liver evaluation are thus attracting considerable interest. In this study, we sought predictors of liver inflammation in chronic hepatitis B (CHB) patients with alanine aminotransferase (ALT) levels ≤ 2-fold the upper limit of normal (ULN); these may guide decisions on whether to commence antiviral therapy. Methods: We retrospectively analyzed 720 patients with CHB who underwent liver biopsy and whose ALT levels were ≤2 ULN. The patients were randomly divided into a training and validation set. We used univariate and multivariate regression analyses of data from the training set to construct a model that predicted significant (grade ≥2) liver inflammation, and validated the model employing the validation set. Results: Aspartate aminotransferase (AST) level, prothrombin time (PT), glutamyl transpeptidase (GGT) level, and anti-hepatitis B virus core antibody (anti-HBC) level were independent predictors of significant liver inflammation in CHB patients with ALT levels ≤ 2 ULN. A model featuring these four parameters afforded areas under the ROC curve of 0.767 and 0.714 for the training and validation sets. The model was more predictive than were the individual factors. Conclusion: AST, GGT, anti-HBC, and PT reflect significant liver inflammation among CHB patients with ALT levels ≤ 2 ULN. Their combination indicates whether antiviral therapy is required.