Early versus delayed initiation of highly active antiretroviral therapy for HIV-positive adults with newly diagnosed pulmonary tuberculosis (TB-HAART): a prospective, international, randomised, placebo-controlled trial

Summary Background WHO guidelines recommend early initiation of antiretroviral therapy (ART) irrespective of CD4 cell count for all patients with tuberculosis who also have HIV, but evidence supporting this approach is poor quality. We assessed the effect of timing of ART initiation on tuberculosis treatment outcomes for HIV-positive patients with CD4 counts of 220 cells per μL or more. Methods We did this randomised, placebo-controlled trial between Jan 1, 2008, and April 31, 2013 at 26 treatment centres in South Africa, Tanzania, Uganda, and Zambia. We enrolled HIV-positive patients with culture-confirmed tuberculosis who had tolerated 2 weeks of tuberculosis short course chemotherapy. Participants were randomly allocated (1:1) to early ART (starting after 2 weeks of tuberculosis treatment) or delayed ART (placebo, then starting ART at the end of 6 months of tuberculosis treatment). Randomisation was computer generated, with permuted blocks of size eight, and stratified by CD4 count (220–349 cells per μL vs ≥350 cells per μL). Patients and investigators were masked to treatment allocation until completion of 6-months' tuberculosis treatment, after which the study was open label. The primary endpoint was a composite of failure of tuberculosis treatment, tuberculosis recurrence, and death within 12 months of starting tuberculosis treatment in the modified intention-to-treat population. Secondary endpoints included mortality. The study is registered with controlled-trials.com (ISRCTN77861053). Findings We screened 13 588 patients and enrolled 1675: 834 assigned early ART, 841 delayed ART. The primary endpoint was reached by 65 (8·5%) of 767 patients in the early ART group versus 71 (9·2%) of 771 in the delayed ART group (relative risk [RR] 0·91, 95% CI 0·64–1·30; p=0·9). Of patients with a CD4 cell count of 220–349 cells per μL, 26 (7·9%) of 331 patients versus 33 (9·6%) of 342 reached the primary endpoint (RR 0·80, 95% CI 0·46–1·39; p=0·6). For those with 350 cells per μL or more, 39 (8·9%) of 436 versus 38 (8·9%) of 429 reached the primary endpoint (RR 1·01, 95% CI 0·63–1·62; p=0·4). Mortality did not differ significantly between treatment groups (RR 1·4, 95% CI 0·8–2·3; p=0·23). Grade 3 and 4 adverse events occurred in 149 (18%) of 834 patients assigned early ART versus 174 (21%) of 841 assigned delayed ART (p=0·37). 87 (10%) of 834 versus 84 (10%) of 841 had immune reconstitution inflammatory syndrome (p=0·56). Interpretation ART can be delayed until af