Background:  HIV/TB co-infection, with its associated high mortality, is one of the biggest challenges facing AIDS and TB treatment programs in Africa. Barriers to initiation of ART in co-infected patients include concerns about drug interactions, high pill burden, drug toxicities, and immune reconstitution. There are no clinical trial data to guide timing of ART initiation in TB patients.

Methods:  We conducted a prospective, open-label randomized controlled trial to determine optimal timing of ART initiation in relation to TB treatment. From June 2005 to July 2008, 645 newly diagnosed sputum smear-positive TB patients with HIV infection and CD4 counts <500 cells/mm³ were enrolled at the CAPRISA TB/HIV clinic in Durban, South Africa, and randomized to the integrated TB/HIV treatment arm (ART initiation during TB treatment) or the sequential treatment arm (ART initiation upon TB treatment completion). All participants received standard TB therapy and cotrimoxazole prophylaxis. ART comprised once-daily didanosine, lamivudine, and efavirenz in both arms. The intent-to-treat interim analysis of the primary endpoint of all-cause mortality using Kaplan-Meier survival and proportional hazards regression analysis includes all data until September 1, 2008, when the Drug Safety Monitoring Board (DSMB) recommended stopping the sequential arm.

Results:  Both arms were similar at baseline with respect to age, gender, CD4 count, viral load, WHO stage, multi-drug resistant TB prevalence, and prior TB. On average, participants in the integrated and sequential arms started ART 67 days and 261 days after starting TB treatment, respectively. Mortality was 55% lower (HR 0.451, 95%CI 0.26 to 0.79; p = 0.0049) in the integrated treatment arm (5.1/100 person-years [24 deaths; n = 431]) compared to the sequential treatment arm (11.6/100 person-years [26 deaths; n = 214]). Mortality rates were significantly lower in the integrated treatment arm regardless of CD4 count; for participants with CD4 counts <201 cells/mm³, the HR was 0.54 (p = 0.044), while for CD4 counts from 201 to 500 cells/mm³, the HR was 0.08 (p = 0.023).

Conclusions:  Initiating ART in HIV/TB-co-infected patients, with CD4 counts <500 cells/mm³, during TB treatment significantly improves survival. The findings offer the first clinical trial evidence to inform clinical management of TB/HIV co-infection and provide further impetus for closer linkages between TB and AIDS care services.