Twenty-one of 40 patients with chronic non-A, non-B hepatitis (37 anti-HCV positive) were randomised to receive interferon alpha 2b (3 million units subcutaneously thrice weekly for 24 weeks) and then to be observed for six months. Among the other 19 patients (controls) randomised to be observed without treatment for 12 months, eight have subsequently been treated with interferon for six months. One treated patient and three controls were lost to follow-up. A return to normal serum alanine aminotransferase levels which lasted until the end of the treatment period occurred in 18 (64%) of the 28 patients given interferon (and in 13 of 21 (62%) randomised to treatment), but only in one of the 16 untreated controls (p less than 0.001). Multivariant analysis indicated that, compared with the ten nonresponders, the 18 patients who responded to interferon were more likely to have acquired infection by intravenous drug abuse than by blood transfusion (p less than 0.05), and were more likely to have histologically less severe chronic liver disease (p less than 0.01). Thus, all 13 patients with less severe liver disease histologically responded to interferon, but only five of 15 patients with cirrhosis or bridging fibrosis responded. Among 17 responders followed for more than four months, five (28%) are still in remission a median of 13 months (range four months to 24 months) after stopping interferon. The characteristics which favoured a response during treatment also appeared to distinguish those who experienced sustained post-treatment remission.(ABSTRACT TRUNCATED AT 250 WORDS)