We quantified serum hepatitis C virus RNA titers and determined hepatitis C virus subtypes in chronic hepatitis C patients treated with interferon-beta to investigate relationships among serum ALT response, serum hepatitis C virus titer and hepatitis C virus subtype. Of 146 chronic hepatitis C patients who received interferon-beta therapy, 24 patients with sustained serum ALT normalization (complete responders) and 26 patients without serum ALT normalization (nonresponders) were randomly selected. Detection, typing and quantitation of hepatitis C virus were performed by means of the "single-tube" polymerase chain reaction method. Of the 24 complete responders, 21 (87.5%) became negative for hepatitis C virus RNA, whereas 21 (80.8%) of the 26 nonresponders remained positive. Hepatitis C virus infections with types I, II, III, IV, II + III and III + IV occurred in 0 (0%), 22 (51.2%), 10 (23.3%), 1 (2.3%), 7 (16.5%) and 3 (7.9%) patients, respectively. The mean pretreatment hepatitis C virus RNA titer of complete responders (0.4 +/- 2.0 x 10(4) CID50/ml) was significantly lower than that of nonresponders (3.8 +/- 4.5 x 10(4) CID50/ml) (p < 0.01). Regardless of HCV subtype, patients with more than 10(4) CID50/ml of HCV did not show serum ALT normalization, whereas complete serum ALT response was seen in most cases with less than 10(2) CID50/ml HCV. These results show that mixed infections with different hepatitis C virus subtypes appear to be more common than previously reported and that the pretreatment serum level of hepatitis C virus RNA is a more important predictor of outcome of interferon therapy than is virus genotype.