Performance of HBsAg quantification assays for detection of Hepatitis B virus genotypes and diagnostic...

The impact of hepatitis B virus (HBV) genomic variability on the measurement of HBsAg level has been poorly evaluated.

This study was designed to compare the performance of all the available assays measuring HBsAg level in this setting.

A large selection of wild type HBV genotypes (n=184) and HBsAg strains harboring mutations in the S gene (n=81) from clinical samples was studied with three HBsAg quantification assays: Architect HBsAg (Abbott), LiaisonXL Murex HBsAg Quant (DiaSorin) and the Elecsys HBsAgII (Roche).

The overall percentage of positive results was 99.2% for Abbott, 98.9% for DiaSorin and 98.1% for Roche. Abbott and Roche assays provided an excellent concordance in HBsAg quantification (global mean bias of -0.006 logIU/mL). By contrast, DiaSorin underestimated HBsAg level with values 0.112 logIU/ml and 0.103 logIU/ml lower than Abbott and Roche, respectively. By contrast, DiaSorin slightly over quantified gtC (2.5% over the expected value) while Abbott provided values 6.2% lower than expected and 16.2% lower than what observed for the other genotypes. HBsAg quantitative assays were influenced by HBs protein substitutions irrespective to the genotype but no specific protein pattern that would particularly impair the quantification by one technique has been identified. However, Roche seemed to be particularly impacted by substitutions at 145 residue: 75% of under quantified samples carried a substituted 145 residue.

This head-to-head comparison indicates a good correlation between all current systems used to quantify HBsAg but clearly shows an influence of both the genotype and the presence of "a" determinant variants in the absolute quantification of HBsAg. While these discrepancies may not translate into major clinical consequence, they may explain an absence of detection of weak concentration of HBsAg on some systems.