Hierarchical analysis of unsuccessful tuberculosis treatment among people living with <scp>HIV</scp> in Brazil using nationwide surveillance data

Abstract Introduction People with tuberculosis‐HIV coinfection face multiple barriers to effective treatment, including social vulnerability, stigma and limited access to healthcare. This study examined factors associated with loss to follow‐up and death among individuals with tuberculosis‐HIV in Brazil. Methods We conducted a longitudinal study using a nationally linked database from surveillance systems. Poisson regression models with robust variance were applied to identify factors associated with unfavourable outcomes, guided by a theoretical‐conceptual hierarchical framework. Results We analysed data from 54 516 individuals. The median time to loss to treatment follow‐up was 115 days, with a cumulative proportion of 29.56%. Among the most consistent predictors of loss to follow‐up were homelessness (relative risk, RR 1.18; 95% confidence interval, 95% CI 1.16–1.19), tuberculosis retreatment (RR 1.16; 95% CI 1.15–1.17) and drug use (RR 1.15; 95% CI 1.14–1.16), whereas antiretroviral therapy use (RR 0.95; 95% CI 0.95–0.96) showed a negative association. The median time to death during tuberculosis treatment was 27 days, with a cumulative proportion of 27.54%. Higher risk of death was observed among individuals with CD4 counts &lt;350 cells/mm 3 (RR 1.09; 95% CI 1.08–1.10), those experiencing homelessness (RR 1.08; 95% CI 1.06–1.10) and those with rifampicin resistance (RR 1.11; 95% CI 1.07–1.15). Conclusion Key social, clinical and programmatic factors were associated with loss to follow‐up and death during tuberculosis treatment among people with HIV. Addressing these vulnerabilities is essential to improving treatment outcomes and advancing progress towards the 2030 targets.