Hepatitis B Core Related Antigen - Does it meet our expectations? Evidence from Cohorts in the United Kingdom and South Africa

Abstract

Introduction: Hepatitis B virus (HBV) infection is a pressing global public health threat, responsible for increasing mortality with a disproportionate impact on populations in the WHO African region. There is a need for evaluation of new biomarkers that may be able to provide insights into risk stratification and disease pathogenesis. We therefore set out to evaluate Hepatitis B core related antigen (HBcrAg), in cohorts in the United Kingdom (UK) and South Africa (SA). Methods: We undertook a cross-sectional retrospective observational study, using serum samples obtained from adults living with chronic HBV infection enrolled at Oxford University Hospitals (OUH) NHS Foundation Trust in the UK (n=142), and from a clinical cohort in Cape Town and Bloemfontein, SA (n=211). We recorded routine clinical and laboratory parameters gathered at each site, and undertook quantification of HBcrAg and host immune biomarkers, IL-21, IP-10 and PD-1. We report on HBcrAg distribution, relationship with other clinical and immunological biomarkers, and performance as a risk stratification tool in each setting. Results: Sex and age were comparable between SA and UK cohorts (p>0.05). There were significant differences between cohorts in ethnicity (p <0.001), HIV coinfection (2.3% in UK vs 56% in SA, p<0.001), HBeAg-positivity (12% in the UK vs 29% in SA, p<0.001), and proportion with HBV DNA >200,000 IU/ml (6% in the UK vs. 22% in SA, p<0.001). Host immune markers were all significantly higher in the SA vs UK cohorts (p<0.001 for all); (i) HBcrAg was positively associated with HBV DNA in both cohorts (p<0.0001), and in the UK this relationship was mediated by HBeAg-positivity. (ii) HBcrAg was also positively associated with overall liver inflammation assessed by ALT (p<0.001 in UK, p<0.01 in SA), however there was no specific significant correlation with more precise host immune markers, IP-10, IL-21 or PD-1 (p>0.05 in all cases). (iii) In univariable and multivariable analysis, there was no association between HBcrAg and liver fibrosis using elastography, APRI or FIB-4 scores in both cohorts. Discussion: In spite of similar HBcrAg levels in both the UK and SA cohorts, prediction of HBV VL seems to differ depending on settings, suggesting a more tailored and personalised approach to HBcrAg testing is required. In our cohorts, HBcrAg did not correlate with any liver disease outcomes.