Retrospective Pará cohort links tuberculosis death to age, HIV and mixed disease

Tuberculosis-attributed mortality in Pará, in the Brazilian Amazon, was associated with older age, male sex, TB–HIV/AIDS, alcohol use, smoking, and combined pulmonary and extrapulmonary disease in a population-based retrospective cohort study. The analysis included 52,066 tuberculosis cases notified between January 1, 2012, and December 31, 2022, and identified 1,032 deaths attributed to tuberculosis. Researchers used cause-specific Cox proportional hazards regression and a complementary Fine–Gray competing-risk model to assess factors associated with mortality. The similar results from the two models support the robustness of the reported associations, but the retrospective observational design cannot establish that any identified factor caused an individual death.

Older age had one of the strongest associations with mortality. Each additional decade of life was associated with a 50% higher hazard of tuberculosis-attributed death in the Cox model, corresponding to an adjusted hazard ratio of 1.50 and a 95% confidence interval of 1.45–1.55. The Fine–Gray model produced a similar estimate of 1.49, with a 95% confidence interval of 1.44–1.54. Both models reported p-values below 0.001 for the age association. Male sex was also associated with a higher mortality hazard, with an adjusted hazard ratio of 1.16 (95% CI 1.01–1.32; p = 0.035). These estimates describe the mortality hazard within the analyzed cohort rather than population-wide death rates.

TB–HIV/AIDS positivity was associated with a 64% higher hazard of tuberculosis-attributed death in the Cox analysis, with an adjusted hazard ratio of 1.64 (95% CI 1.32–2.03; p < 0.001). Unknown or unrecorded TB–HIV/AIDS status was likewise associated with mortality, with an adjusted hazard ratio of 1.38 (95% CI 1.20–1.58; p < 0.001). The source evidence does not identify the reference category used for the unknown-status estimate, so the comparison should not be interpreted beyond the reported adjusted association. The finding indicates that incomplete HIV-related information was present in the surveillance records and was associated with the outcome measured by the models.

Alcohol use and smoking were also associated with tuberculosis-attributed death. Alcohol use was linked to a 53% higher hazard in the Cox model, with an adjusted hazard ratio of 1.53 (95% CI 1.26–1.86; p < 0.001), while smoking was associated with a 30% higher hazard, with an adjusted hazard ratio of 1.30 (95% CI 1.06–1.61; p = 0.014). Patients with both pulmonary and extrapulmonary tuberculosis had the largest reported clinical association in the results: their adjusted hazard ratio was 3.00 (95% CI 2.29–3.92; p < 0.001). Because the cohort was retrospective and based on notified cases, these values characterize associations in the source population rather than causal effects or the prospectively measured risks for individual patients.

Incarceration showed an inverse association with tuberculosis-attributed mortality, with an adjusted hazard ratio of 0.42 (95% CI 0.27–0.65; p < 0.001). The authors advised caution, noting that the result could reflect structured follow-up or differential surveillance rather than a protective effect of imprisonment. The main survival cohort excluded cases without a valid treatment-closure date, post-mortem notifications, and records with inconsistent follow-up time, so the estimates apply specifically to the 52,066 retained cases. The competing-risk analysis treated cure and death from other causes as competing events, providing a complementary assessment of the time to tuberculosis-attributed death. Overall, the study identified clinical and information-related factors associated with mortality in the analyzed population while supporting restraint about causal interpretation.

Subject of Research: Clinical and surveillance factors associated with tuberculosis-attributed mortality in Pará, Brazil

Article Title: Factors associated with tuberculosis-attributed mortality in the Brazilian Amazon: a population-based retrospective cohort study with survival and competing-risk analyses

Article References: Marcos, W., Mota, N. C. V., Carvalho, F. I. C., da Silva Sousa Júnior, A., de Souza Melo, M. I., Sardinha, D. M., & Batista Lima, K. V. (2026). Factors associated with tuberculosis-attributed mortality in the Brazilian Amazon: a population-based retrospective cohort study with survival and competing-risk analyses. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14431-y

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14431-y

Keywords: tuberculosis, mortality, Pará, Brazilian Amazon, HIV coinfection, survival analysis, competing risks, population cohort, health surveillance, Factors, associated, tuberculosis-attributed