Background: Optimal timing for initiation of antiretroviral therapy (ART) among HIV-infected patients with tuberculosis (TB) is not well established.
Methods: HIV/TB-coinfected patients were randomized to initiate tenofovir/lamivudine/efavirenz at 4 weeks (4-week group) or 12 weeks (12-week group) of TB treatment. The primary outcome was 1-year all-cause mortality.
Results: Of 156 patients, 79 were in 4-week group and 77 in 12-week group. Overall, median (interquartile range) CD4 was 43 (47-106) cells per cubic millimeter and median (interquartile range) HIV-1 RNA was 5.8 (5.4-6.3) log copies per milliliter. Eleven (7%) mortalities occurred in a total follow-up period of 137 patient-years. Seven percent (6/79, 8.76 per 100 patient-years) mortalities were in 4-week group, and 6% (5/77, 7.25 per 100 person-years) mortalities were in 12-week group [relative risk (RR) = 0.845, 95% confidence interval (CI) = 0.247 to 2.893]. Twenty-eight (35%) patients in 4-week group and 25 (32%) patients in 12-week group were hospitalized (RR = 1.142, 95% CI = 0.588 to 2.217). Grade 2-4 adverse events were 39% (31/79) in 4-week group and 34% (26/77) in 12-week group (RR = 1.267, 95% CI = 0.659 to 2.435). In multivariate analysis, "low albumin" (RR = 2.695, 95% CI = 1.353 to 5.475) and "low baseline CD4 count" (RR = 4.878, 95% CI = 1.019 to 23.256) were the independent predictors of mortality. Immune reconstitution inflammatory syndrome was more frequent in 4-week group with an incidence of 8.86 versus 5.02 per 100 person-months in 12-week group over the first 6 months of ART (P = 0.069).
Conclusions: In middle-income countries where ART is initiated at CD4 count of <350 cells per cubic millimeter, immediate initiation of ART in HIV-infected patients with active TB was not associated with survival advantage when compared to initiation of ART at 12 weeks.