Background: A randomized double-blind placebo-controlled trial was conducted to estimate the efficacy of preventive therapy for tuberculosis (TB) in HIV-infected adults in Lusaka, Zambia. The main outcome measures were the incidence of TB, mortality and adverse drug reactions.
Methods: During a 2 year period, 1053 HIV-positive individuals without evidence of clinical TB were randomly assigned to receive 6 months of isoniazid twice a week (H), or 3 months of rifampicin twice a week (R) plus pyrazinamide (Z), or a placebo. Therapy was taken twice a week and was self administered. Subjects presenting with symptoms during the follow-up period were investigated for TB.
Results: The 1053 subjects in the study were followed up for a total of 1631 person-years (median = 1.8 years). Twenty-nine subjects were taken off treatment as a result of adverse drug reactions. A total of 96 cases of TB/probable TB (59 TB and 37 probable TB) were diagnosed during the study period and 185 deaths were reported. One hundred and fifteen subjects (11%) did not return to the study clinic at any time after enrolment. The incidence of TB was lower in those subjects on preventive therapy (H and RZ groups combined) compared with those on placebo (rate ratio = 0.60, 95% CI: 0.36-1.01, P = 0.057), as was the incidence of TB/probable TB (rate ratio = 0.60, 95% CI: 0.40-0.89, P = 0.013). The effect of preventive therapy was greater in those with a tuberculin skin test (TST) of 5 mm or greater, in those with a lymphocyte count of 2x10(9)/l or higher, and in those with haemoglobin of 10 g/dl or higher. There was no difference in mortality rates between the preventive therapy and placebo groups. The effect of preventive therapy declined after the first year of the study so that by 18 months the rates of TB in the treated groups were similar to that in the placebo group.
Conclusion: This study has demonstrated that preventive therapy with either twice weekly isoniazid for 6 months or a combination of rifampicin and pyrazinamide for 3 months reduced the incidence of TB in HIV-infected persons in Zambia. No effect was observed on mortality. The effect was greatest in persons who had a positive TST or a lymphocyte count of 2x10(9)/l or greater, indicating that preventive therapy may be more effective in people with less advanced immunosuppression. The limited duration of the protective effect reported in this study raises the question of the need for lifelong preventive therapy or re-prophylaxis.