Multilaboratory evaluation of real-time PCR tests for hepatitis B virus DNA quantification

Abstract

The performance characteristics of four different assays for hepatitis B virus (HBV) quantification were assessed: the Abbott RealTime HBV IUO, the Roche Cobas AmpliPrep/Cobas TaqMan HBV test, the Roche Cobas TaqMan HBV test with HighPure system, and the Qiagen artus HBV TM ASR. Limit of detection (LOD), linear range, reproducibility, and agreement were determined using a serially diluted plasma sample from a single chronically infected subject. Each assay was tested by at least three laboratories. The LOD of the RealTime and two TaqMan assays was approximately 1.0 log(10) IU/ml; for artus HBV (which used the lowest volume of extracted DNA), it was approximately 1.5 log(10) IU/ml. The linear range spanned 1.0 to at least 7.0 log(10) IU/ml for all assays. Median values were consistently lowest for artus HBV and highest for Cobas AmpliPrep/Cobas TaqMan HBV. Assays incorporating automated nucleic acid extraction were the most reproducible; however, the overall variability was minor since the standard deviations for the means of all tested concentrations were ≤0.32 log(10) IU/ml for all assays. False-positive results were observed with all assays; the highest rates occurred with tests using manual nucleic acid extraction. The performance characteristics of these assays suggest that they are useful for management and therapeutic monitoring of chronic HBV infection.

Fig. 1.

Linear range of the viral load assays.

Fig. 2.

Assay bias of the four real-time PCR assays for HBV quantification. Bias is defined as measured viral load (log₁₀ IU/ml) minus the nominal concentration of the sample (log₁₀ IU/ml).