Randomised controlled trial of isoniazid preventive therapy in South African adults with advanced HIV disease
BACKGROUND: Tuberculosis (TB) preventive therapy is not effective in human immunodeficiency virus (HIV) infected adults with negative tuberculin skin tests (TSTs). However, there are insufficient data on patients with advanced HIV disease from high TB incidence areas. METHODS: We conducted a randomised, double-blind, controlled trial comparing isoniazid (INH) with placebo among TST-negative adults with World Health Organization Stage 3 or 4 HIV disease. INH/placebo was admin-istered for 12 months by patient-nominated supervisors. TST-positive participants were given open-label INH. Participants, who did not have access to antiretroviral therapy (ART), were followed up for 24 months with 6-monthly sputum culture and chest radiography. RESULTS: A total of 118 participants were enrolled: TST was negative in 98. In the randomised arms, the incidence of TB was 18/100 person-years (py) in the INH arm and 11.6/100 py in the placebo arm (hazard ratio 1.59, 95%CI 0.57-4.49). There were no significant differences in mortality, hospitalisation rate or CD4+ lymphocyte decline. INH/placebo adherence was 85%, and was significantly higher among participants with work-based treatment supervisors. CONCLUSIONS: We did not find any association between INH preventive therapy and reductions in incident TB among TST-negative adults with advanced HIV disease, but the study had limited statistical power. High levels of adherence were observed with patient-nominated supervisors. © 2007 The Union.