Objective: To verify the analytical performance of cobas® HBV PCR and cobas® HCV PCR assays with Abbott m2000 RealTime System as the reference method.
Design: De-identified residual, archived patient specimens, and College of American Pathologists (CAP) proficiency testing samples were used. Analytical parameters verified were accuracy, precision, limit of detection (LOD), linear range, and cross-contamination. Experiments were designed in accordance with Clinical Laboratories Standards Institute (CLSI) guidelines and CAP standards. Analysis of accuracy was done through regression plots and Bland Altman analyses. Precision was analyzed through coefficient of variation and ANOVA; LOD through probit analysis; and linear range through polynomial fit analysis.
Results: The regression plots for accuracy showed a slope nearing 1, with a y-intercept close to zero, while Bland Altman analyses also showed no systematic evidence of bias, though concordance of results was not perfect near the lower limit of quantification. Coefficients of variation were all below 15%, while ANOVA returned p-values above 0.99, indicating no statistically significant imprecision. The LOD verified were an order of magnitude higher than the manufacturer reported ones for both assays, while the linear range verified was more limited. Within the verified range, polynomial fit analysis showed line to be the best fit for the data.
Conclusions: cobas® HBV PCR and cobas® HCV PCR assays showed acceptable accuracy, acceptable precision, as well as no evidence of cross-contamination. The LOD verified were higher, and linear ranges more limited than those reported by the manufacturer. Verifications of these may be limited by availability of appropriate testing specimens.
Keywords: Analytical; HBV; HCV; Roche; Verification; cobas®.
© 2023 The Authors.