Résumé
Background: The decline in hepatitis B virus surface antigen (HBsAg) may be an early predictor of the viral efficacy of Hepatitis B virus (HBV) therapy. The HBsAg levels obtained by different immunoassays now need comparing and the relationships between levels of HBsAg and HBV DNA alongside HBsAg and genotype must be evaluated.
Methodology/Principal Findings: HBsAg levels were compared among 80 patients using the Abbott Architect assay, a commercial immunoassay approved for HBsAg detection and quantitation, and three other assays derived from immunoassays approved for HBsAg detection (manufactured by Diasorin, Bio-Rad and Roche). Good correlation was found between the Abbot vs. Diasorin, Bio-Rad and Roche assays with narrow 95% limits of agreement and small mean differences: -0.06 to 0.11, -0.09 log(10) IU/mL; -0.57 to 0.64, -0.04 log(10) IU/mL; -0.09 to 0.45, -0.27 log(10) IU/mL, respectively. These agreements were not affected by genotypes A or D. HBsAg was weakly correlated with HBV DNA, whatever the HBsAg assay used: Abbott, rho = 0.36 p = 0.001, Diasorin rho = 0.34, p = 0.002; Bio-Rad rho = 0.37, p<0.001; or Roche rho = 0.41, p<0.001. This relationship between levels of HBsAg and HBV DNA seemed to depend on genotypes. Whereas HBsAg (Abbott assay) tended to correlate with HBV DNA for genotype A (rho = 0.44, p = 0.02), no such correlation was significant for genotypes D (rho = 0.29, p = 0.15).
Conclusion/Significance: The quantitation of HBsAg in routine clinical samples is comparable between the reference assay and the adapted assays with acceptable accuracy limits, low levels of variability and minimum discrepancy. While HBsAg quantitation is not affected by HBV genotype, the observed association between levels of HBsAg and HBV DNA seems genotype dependent.