We have developed a sensitive and reproducible one-step competitive reverse transcriptase (RT) PCR assay, which allows hepatitis C virus (HCV) RNA quantitation in plasma over a broad range of values. The RNA samples and a constant amount of an internal standard were reverse transcribed and coamplified with the same primers in the same tube. A standard curve was obtained from an additional series of tubes containing both the internal standard and known amounts of a wild-type HCV RNA transcript, thus eliminating the need for titrating samples with the competitor. Eighty-eight anti-HCV-positive samples were tested by RT-PCR and a branched-DNA (bDNA) assay which has a detection limit of 3.5 x 10(5) copies per ml. Fifty-five samples were quantifiable by both methods (correlation coefficient, 0.72), the ranges of values found by the RT-PCR and bDNA assays being, respectively, 0.127 x 10(6) to 18.4 x 10(6) and 0.44 x10(6) to 38 x 10(6) copies per ml. Six samples that had indeterminate values by the bDNA assay had RT-PCR values between 0.37 x 10(5) and 9.6 x 10(5) copies per ml. Twenty-two samples that had values below the cutoff value by the bDNA assay had RT-PCR values between 2.5 x 10(3) and 10.4 x 10(5) (18 less than and 4 more than the limit of 3.5 x 10(5) copies per ml). The remaining five samples were negative by both assays. The level of RT-PCR interassay reproducibility was high (correlation coefficient between duplicate values, 0.94). Our method, with a detection limit of 2,500 copies per ml, was markedly more sensitive than the bDNA assay. This method is convenient for following up patients with low viremia, a common situation with alpha interferon treatment.