In the present study we evaluated the performance of the new LCx HIV RNA quantitative kit (Abbott Laboratories, Delkenheim, Germany) for the quantitative detection of HIV-1 RNA in human plasma in comparison to the Cobas Amplicor HIV-1 Monitor assay (Roche Diagnostics, Branchburg, N.J.), including samples containing a variety of HIV-1 subtypes. LCx and Cobas were compared using archived EDTA plasma samples collected from HIV-infected patients. Considering the lower limit of the linear range of 50 copies/ml, the detection rate of the LCx was 139 out of 174 (79.9%) versus 131 out of 174 (75.3%) of the Cobas. Overall agreement was 95.4% (166/174) at a cut-off of 50 copies. LCx and Cobas results on clinical samples were found linearly associated (r2 = 0.900) and strongly correlated (r = 0.949). The mean viral load in the 174 frozen patient samples was 3.25 log10 copies/ml by LCx compared to 2.71 log10 copies/ml by Cobas. Considering only samples with a viral load > or =50 copies/ml, the average difference was -0.132 log copy/ml. Using a panel consisting of 9 plasma samples spiked with 9 different HIV-1 cultured isolates (A-H, and O) LCx detected the 9 subtypes with a high degree of precision, i.e., 9-33% coefficient of variation. As expected, the Cobas failed to detect the group O isolates. The results of the remaining samples showed a higher degree of variation (when testing four replicates of the subtype panel) than the LCx of 14.2-40.3%. Nevertheless, the results were comparable with the LCx data.