Abstract
Background
Approximately one-third of patients have durable responses after finite (48-week) treatment with peginterferon alfa-2a. The ability to identify patients likely to respond would be particularly useful in resource-limited settings.
Methods
Data from 1,363 peginterferon alfa-2a recipients (955 hepatitis B ‘e’ antigen [HBeAg]-positive and 408 -negative) in six studies were analysed. Baseline scoring systems were developed using generalized additive models and multiple logistic regression analysis to predict virological response (VR; HBV DNA <2,000 IU/ml), alone or combined with alanine aminotransferase (ALT) normalization (CR) at 24 weeks post-treatment.
Results
Based on the final models, points were assigned for age ≥45 (0) or <45 years (1); male (0) or female (1); hepatitis B surface antigen (HBsAg) >25,000 (0), >7,500-≤25,000 (1), >1,250-≤7,500 (2) or ≤1,250 IU/ml (4); HBV DNA >5 (0) or ≤5 log10 IU/ml (2) and ALT ratio >1-7 (0) or either ≤1 or >7 (1). Higher total scores (range 0-9) indicate higher likelihood of response. VR and CR rates were 28.5% (388/1,363) and 24.4% (332/1,363), respectively, and increased with increasing score: score 0-1 (n=257), VR 14.8%, CR 12.8%; score 2-3 (n=711), VR 23.1%, CR 20.1%; score ≥4 (n=395), VR 47.1%, CR 39.5%.
Conclusions
An easy-to-use baseline scoring system for use in settings where HBeAg status and HBV genotypes are unavailable would allow clinicians to identify patients with a low or high chance of achieving a durable post-treatment response to peginterferon alfa-2a. The tool can be used to inform treatment decisions in resource-limited settings.