Multicenter evaluation of the cobas(R) HIV-1 quantitative nucleic acid test for use on the cobas(R) 4800 system for the quantification of HIV-1 plasma viral load.

BACKGROUND AND OBJECTIVES: Measurement of HIV-1 viral load (VL) is necessary to monitor treatment efficacy in patients receiving antiretroviral therapy. We evaluated the performance of the cobas(R) HIV-1 quantitative nucleic acid test for use on the cobas(R) 4800 system ("cobas 4800 HIV-1"). METHODS: Limit of detection, linearity, accuracy, precision, and specificity of cobas 4800 HIV-1, COBAS(R) AmpliPrep/COBAS(R) Taqman(R) HIV-1 version 2.0 (CAP/CTM HIV-1 v2) and Abbott RealTime HIV-1 were determined in one or two out of three sites. RESULTS: The limit of detection of the cobas 4800 HIV-1 for 400 muL and 200 muL input volumes was 14.2 copies/mL (95% CI: 12.5-16.6 copies/mL) and 43.9 copies/mL (37.7-52.7 copies/mL), respectively. Cobas 4800 HIV-1 demonstrated 100% specificity, and results were linear for all analyzed group M HIV-1 subtypes. Precision was high (SD < 0.19 log10) across all measured ranges, reagent lots and input volumes. Correlation between cobas 4800 HIV-1 and CAP/CTM HIV-1 v2 or RealTime HIV-1 was high (R(2) >/= 0.95). Agreement between cobas 4800 HIV-1 and CAP/CTM HIV-1 v2 was 96.5% (95.0%-97.7%) at a threshold of 50 copies/mL, and 97.2% (95.8%-98.3%) at 200 copies/mL. Agreement between cobas 4800 HIV-1 and RealTime HIV-1 was 96.6% (93.4%-98.5%) at 50 copies/mL, and 97.0% (94.0%-98.8%) at 200 copies/mL. The mean difference between cobas 4800 HIV-1 and CAP/CTM HIV-1 v2 or RealTime HIV-1 was -0.10 log10 or 0.01 log10, respectively. CONCLUSIONS: The cobas 4800 HIV-1 test is highly sensitive, accurate and correlated well with other assays, including agreement around clinically relevant thresholds, indicating minimal overall VL quantification differences between tested platforms.