Objectives: To better characterize viral load profiles through the course of HIV-1 disease and in response to treatment, and to further evaluate quantitative competitive polymerase chain reaction for measurement of viral load, we extended our comparative evaluation of this and other viral load measurements to a total of 118 patients, representing all stages of HIV-1 disease. Design: For cross-sectional analysis across the spectrum of HIV-1 disease, plasma viral load was evaluated in 112 HIV-1-infected patients by quantitative competitive polymerase chain reaction analysis, plasma p24 antigen assay, plasma immune complex-dissociated p24 antigen assay and an endpoint dilution viral culture. Longitudinal specimens from six additional patients were analyzed, extending from the time of presentation with symptomatic acute HIV-1 infection through up to more than 2 years of follow-up. Longitudinal specimens were also studied for three patients over the period of initiation of zidovudine treatment, for 6 weeks of treatment and following temporary withdrawal of the treatment. Methods: All measurement techniques were assessed in replicate aliquots of plasma. Results: Quantitative competitive polymerase chain reaction was the most sensitive measure of viral load, and was best correlated with CD4+ T-cell counts. In longitudinally studied patients, this technique also allowed measurement of plasma virus levels throughout the period of follow-up, even when culture and p24 assays became negative following resolution of acute HIV-1 infection. The quantitative competitive polymerase chain reaction was also able to detect rapid and substantial changes in viral load associated with initiation and temporary withdrawal of antiviral treatment. Conclusions: The quantitative competitive polymerase chain reaction is promising as a sensitive and accurate method for measuring plasma viral load in HIV-1-infected patients, and is useful for following changes in viral load over the natural history of infection and following treatment intervention. The technique is particularly useful for patients with > 200 x 106 CD4+ T cells/l, in whom other viral markers are typically negative.