Effect of prior TB preventive therapy on all-cause mortality during TB treatment among people with HIV/TB in rural eastern Uganda: an observational causal analysis

Tuberculosis preventive therapy (TPT) is the cornerstone for preventing TB disease. However, it is uncertain whether prior TPT completion improves survival once TB disease develops in people with human immunodeficiency virus (PWH) while on anti-retroviral therapy (ART). We evaluated the effect of prior TPT completion on all-cause mortality during TB treatment among PWH who developed TB disease while on ART in rural eastern Uganda. Prior TPT completion served as a marker of sustained engagement in HIV care.

We applied inverse probability of treatment weighting using propensity scores, a causal inference analysis approach, to mimic a randomized controlled trial using real-world retrospective cohort data. Inverse probability of treatment weighting addresses design flaws in observational studies, such as selection and confounding biases, providing less biased causal effect estimates. Individuals who had completed a full course of TPT before the start of the index TB treatment episode formed the intervention group, while the control group comprised those with no history of TPT before the index TB treatment initiation. Inverse probability of treatment weighting ensured balanced baseline covariates between the exposure groups. We estimated all-cause mortality rates using person-time methods and Kaplan-Meier curves and performed propensity-score weighted Cox proportional hazards analysis for cause-effect estimation. We reported adjusted hazard ratios (aHR) with 95% confidence intervals (CI).

Of 719 participants, 296 (41.2%) had completed a full course of TPT before the start of the index TB treatment episode, and 83 (11.5%) had died. The mortality rate was 7.13 per 10,000 person-days, higher among participants without prior TPT than those with prior TPT completion (18.0% vs. 2.4%, log-rank χ²₍₁₎ = 35.3, p < 0.001). Median survival was 277 days (95% CI: 213–689). Prior TPT completion was associated with an 87% lower hazard of all-cause mortality (aHR 0.13, 95% CI: 0.06–0.29).

Prior TPT completion substantially reduced all-cause mortality among PWH who developed drug-susceptible TB disease while on ART. Therefore, there is a need to strengthen TPT completion and engagement in HIV care across HIV programs to lower all-cause mortality among PWH.