Summary. Treatment of chronic hepatitis C in human immunodeficiency virus (HIV)-infected patients is associated with low response rates and high incidence of side effects. One hundred twenty-one hepatitis C virus (HCV)–HIV-coinfected patients were randomized to receive interferon alpha-2b (3 MU thrice weekly; n = 61) or peginterferon alpha-2b (1.5 μg/kg/week; n = 60), plus ribavirin (800 mg daily), for 24 (genotype 2 or 3) or 48 weeks (genotype 1 or 4). We assessed early virological response at 4, 8 and 12 weeks to predict sustained virological response (SVR). Safety assessment included frequent blood lactate measurement and relative quantitation of mitochondrial DNA (mtDNA) content in peripheral blood mononuclear cells. In intention-to-treat analysis, the SVR rate was higher in the peginterferon group (55%vs 26%; P = 0.002). The difference for HCV genotypes 1 and 4 was 45%vs 14% (P = 0.009) and 50%vs 27% (P = 0.387), respectively, and for genotype 2 or 3, 71%vs 43% (P = 0.12) Viral response at 4, 8 and 12 weeks of treatment was highly predictive of SVR. Among genotype 3 patients, 17 of 20 (85%) whose HCV RNA was already undetectable at 4 weeks had an SVR after 24 weeks of treatment. Hyperlactataemia occurred in 22 patients and was clinically significant in six, two of whom died. mtDNA decreased significantly 4–12 weeks after the start of treatment in patients developing clinically significant hyperlactataemia. Peginterferon alpha-2b plus ribavirin was more effective than interferon alpha-2b plus ribavirin in HIV-coinfected patients. Frequent monitoring of virological response may be very helpful to optimize treatment compliance, to tailor treatment duration and to minimize side effects.