Combined HBsAg and anti-HBc testing is required to estimate hepatitis B virus seroprevalence in a low-endemic country: findings from a nationwide population-based serosurvey, Belgium, 2020

Abstract

Background: The World Health Organization aims to eliminate hepatitis B virus (HBV) by 2030 through reducing incidence and mortality. Accurate prevalence estimates are crucial to guide policies and monitor progress towards HBV elimination. However, HBV prevalence can be overestimated when relying solely on hepatitis B surface antigen (HBsAg) because of unconfirmed or false-positive results. Robust screening algorithms to improve diagnostic accuracy and minimise false positives are required. Aim: We conducted a nationwide, population-based serosurvey to estimate HBV prevalence in Belgium by using HBsAg alone or combined with hepatitis B core antibody (anti-HBc) positivity as infection criterion. Methods: We analysed HBsAg and anti-HBc in a total of 4,955 left-over serum samples from severe acute respiratory syndrome coronavirus 2 sero-epidemiology studies in 2020. Samples were stratified per region, 10-year age band and sex. A confirmatory anti-HBc neutralisation assay was performed in discordant samples. Results: We detected HBsAg in 0.75% (37/4,955) of samples, of which 62.2% (23/37) were anti-HBc-negative and showed no specific anti-HBc signal in the neutralisation assay. None of the samples from <= 5-year-olds (n = 87) were double-positive. Weighted analysis estimated HBsAg seroprevalence at 0.74% (95% confidence interval (CI): 0.50-1.04). However, considering double HBsAg and anti-HBc positivity, an HBV prevalence of 0.25% (95% CI: 0.13-0.42) was retained. The HBsAg/anti-HBc prevalence in <= 33-year-olds was lower than in older adults (0.079% vs 0.36%; p = 0.015), consistent with Belgium's vaccination policy. Conclusion: This serosurvey reinforces the importance of confirmatory anti-HBc testing in HBsAg-positive samples, particularly in low-endemic countries. Incorporating anti-HBc testing improves the correctness of prevalence estimates.