Background:  The dual epidemic of HIV/TB has resulted in significant childhood morbidity and mortality in Africa. P1041 was a prospective, phase II/III, randomized, double-blind placebo-controlled study to determine the efficacy of primary isoniazid (INH) prophylaxis to prevent tuberculosis (TB) and latent TB infection among HIV-1 perinatally exposed infants (both infected [HIV⁺] and uninfected [HIV^–]) in South Africa. We now report the initial results for the HIV^– group.  

Methods:  We randomized 3-month-old Bacille Calmette Guerin (BCG) vaccinated, HIV^– infants without prior TB exposure to daily INH (10 to 20 mg/kg/day) or placebo for 96 weeks, and then followed for another 96 weeks off INH/placebo. Main outcome measures were latent TB infection, TB disease, and death. A log rank test was used to compare treatment groups in an intent-to-treat analysis. The Kaplan-Meier estimate was used to estimate the rate of loss to follow-up.

Results:  From December 1, 2004 to June 7, 2006, 804 HIV^– infants (targeted 800) were enrolled and followed at 3 sites in South Africa. As of March 1, 2008 (data cutoff), the median length follow-up was104 weeks, the loss to follow-up rate was 17.2% (95%CI 13.0 to 20.5%), and 10% had reached a study endpoint as shown below. As of June 26, 2008 when the Data Safety Monitoring Board (DSMB) recommended discontinuing the study drug, all of the uninfected infant had competed the 96-week treatment phase. There was no significant difference in adverse event rates or compliance with study drug (range 73 to 89%) between groups.   

Conclusions:  Over 96 weeks, primary INH prophylaxis failed to prevent the high burden of TB disease and TB infection among young HIV perinatally exposed, uninfected infants in South Africa. Similar results reported earlier were observed in the HIV⁺ group.