Prediction model for sustained hepatitis B e antigen seroconversion to peginterferon alfa‐2a in chronic hepatitis B

Abstract

Abstract Background and Aim Clinically applicable models to predict hepatitis B e antigen (HBeAg)‐positive chronic hepatitis B (CHB) response to peginterferon (PEG‐IFN) are scarce. This study aimed to develop simple scoring systems, based on multiple parameters, for predicting sustained HBeAg seroconversion to PEG‐IFN. Methods Eighty‐five treatment‐naïve patients with HBeAg‐positive CHB underwent 52‐week PEG‐IFN treatment and 24‐week follow‐up. Logistic regression analysis assessed parameters at baseline and weeks 12, 24, and 52 to predict HBeAg seroconversion at week 24 off‐treatment. The best three predictors at each time point were included in prediction models of PEG‐IFN therapy efficacy. Results The three most meaningful predictors were alanine aminotransferase (ALT) > 5 × ULN, HBeAg ≤ 500 S/CO, and antibody to hepatitis B core antigen (anti‐HBc) > 10.7 S/CO at baseline; HBeAg ≤ 20 S/CO, anti‐HBc > 11.7 S/CO, and HBeAg decline > 1 log 10 S/CO at week 12; ALT > 2 × ULN, HBeAg ≤ 15 S/CO, and anti‐HBc > 10.4 S/CO at week 24; HBeAg ≤ 5 S/CO, anti‐HBc > 11.1 S/CO, and hepatitis B virus DNA decline > 2 log 10 copies/mL at week 52. Parameters meeting optimal cutoff thresholds were scored 1 or otherwise scored 0. For total scores of 0 versus 3 at baseline and weeks 12, 24, and 52, response rates were 6.3%, 12.5%, 0%, and 0% versus 90.0%, 83.3%, 76.9%, and 86.4%, respectively. Conclusions We successfully established prediction models for PEG‐IFN response in HBeAg‐positive CHB.