Three procedures for the quantification of human immunodeficiency virus type 1 (HIV-1) RNA from plasma were compared at three laboratories. The comparison involved the Quantiplex branched DNA assay (version 1.0) by Chiron Diagnostics, the NASBA-QT assay by Organon Teknika, and the Amplicor Monitor assay by Roche Molecular Systems. The laboratories performed each of the three assays with the same sets of reconstructed HIV-1-infected human plasma samples, cross-sectionally collected clinical plasma samples and longitudinally collected plasma samples from patients starting zidovudine therapy. Analysis of the reconstruction panel results for interlaboratory variation demonstrated that no laboratory differences in results were detected for any of the assays. A comparison of the reproducibilities of duplicate samples analyzed by batch and in separate assay runs demonstrated that the reproducibilities of the test results were similar within one assay and appeared to be independent of the HIV-1 concentration. The best reproducibility was obtained with the Quantiplex assay, but all three assays demonstrated equal reliability, which was independent of batched or unbatched analysis of replicate samples. Differences in the absolute concentrations calculated were observed for the assays, in particular in the analysis of reconstructed samples. In all assays, similar changes in plasma HIV-1 RNA concentrations were determined for longitudinally collected clinical samples.