Systematic review with meta‐analysis: hepatitis B surface antigen decline and seroclearance in chronic hepatitis B patients on nucleos(t)ide analogues or pegylated interferon therapy

Background Surface antigen (HBsAg) seroclearance in chronic hepatitis B (CHB) is associated with improved outcomes and permits safe treatment discontinuation. Current treatments (nucleos(t)ide analogues [NA], pegylated interferon [pegIFN]) have low HBsAg seroclearance rates. New therapies aim to improve these rates. Aim To use historical data to quantify the mean change from baseline in HBsAg at week 24 and the seroclearance rate at week 48 under standard-of-care treatment. Methods We searched PubMed, ClinicalTrials.gov, and conference presentations for CHB studies with NAs (tenofovir, entecavir) or pegIFN. HBsAg decline and seroclearance were evaluated by a random effects meta-analysis. Treatment type, HBsAg baseline values, and hepatitis B virus (HBV) genotype, recruitment region were explored as heterogeneity sources. Results One hundred and fifty-eight cohorts (over 10 000 patients; 71 studies [randomised controlled; prospective cohorts; retrospective cohorts]) were analysed. With NAs, 24-week HBsAg decline was minimal for treatment-naïve HBeAg-negative, and virologically suppressed patients. Considerable heterogeneity was observed in NA-treated HBeAg-positive patients, impacted by HBV genotype, NA type and HBsAg baseline levels. 48-week HBsAg seroclearance rates were low in all NA-treated groups (<1%/group). HBsAg declines and seroclearance rates were greater with pegIFN (also greater intra-subject variation within studies) vs NAs. Conclusions This meta-analysis demonstrates good consistency between studies and limited within-study variation for the endpoint of HBsAg seroclearance at week 48 on NA treatment. For novel agents, these results can be applied to quantify control arm performance in study designs with historical borrowing.