Summary Background Tuberculosis is a common complication of HIV-1 infection, especially in developing countries. Practical and effective chemoprophylaxis regimens for HIV-1-related tuberculosis are needed. Our aim was to test the efficacy of isoniazid versus rifampicin with pyrazinamide for prevention of tuberculosis in HIV-1-positive individuals. Methods We compared the efficacy of 6 months of isoniazid with 2 months of rifampicin and pyrazinamide for prevention of tuberculosis in HIV-1-seropositive individuals. Eligible participants were aged 16–77 years, HIV-1 seropositive, had a positive purified-protein derivative (PPD) skin test reaction of at least 5 mm, and had a normal chest radiograph. Participants were randomly assigned partially supervised twice weekly isoniazid for 24 weeks or twice weekly rifampicin and pyrazinamide for 8 weeks. Participants were followed up for up to 4 years for the development of tuberculosis and survival. Findings Tuberculosis developed in 14 (3·8%) of 370 participants assigned isoniazid and 19 (5·0%) of 380 participants assigned rifampicin and pyrazinamide (Cox model rate ratio 1·3 [95% CI 0·7–2·7]). The Kaplan-Meier estimate of the risk of tuberculosis during the first 10 months after entry was 3·7% among participants who received rifampicin and pyrazinamide compared with 1·0% (p=0·03) among participants who received isoniazid, and 5·4% versus 5·1%, respectively (p=0·9) at 36 months after entry. Higher rates of tuberculosis were observed in people with baseline CD4 percentages (of total lymphocytes) of less than 20 (rate ratio 4·0 [95% CI 1·8–9·0]). There were no significant differences in total mortality at any time. Interpretation Twice-weekly isoniazid preventive therapy for 6 months or rifampicin and pyrazinamide for 2 months provided similar overall protection against tuberculosis in HIV-1-infected, PPD-positive adults. The better protection among recipients of isoniazid during the first 10 months was most likely secondary to the longer duration of chemoprophylaxis. Preventive therapy for HIV-1-seropositive, PPD-positive individuals could be practical in developing countries with a once weekly clinic visit, but optimum duration of chemoprophylaxis has not been determined.