Prevalence and Mortality due to

Background: HIV-associated tuberculosis is a leading cause of mortality. Methods: We enrolled inpatients from medical wards irrespective of tuberculosis symptoms and outpatients with tuberculosis symptoms. All participants had HIV and were ≥18 years old, and all were recruited from 7 countries: Malawi, South Africa, Tanzania, Thailand, Uganda, Vietnam, and Zambia. Participants also had <3 doses of antituberculosis treatment in the past 60 days and no isoniazid preventive therapy in the past 6 months. We estimated the prevalence of MTB-BSI. In Bayesian survival models, we estimated the hazards of mortality for inpatients with MTB-BSI vs those without. Results: Between 2019 and 2021, 1703 participants were included: 44% were hospitalized, the median CD4 count was 361 cells/μL, and 77% reported using antiretroviral therapy. Crude prevalence of MTB-BSI varied by country but was 4.4% (32/723) among all inpatients, 22.5% (32/142) among inpatients with microbiologically confirmed TB, and 0.2% (2/913) among all outpatients. Among all inpatients, those with MTB-BSI had 2.65-times (95% credible interval, 1.06-5.01) increased hazard of death by 30 days and 1.79-times (95% credible interval, .81-3.11) increased hazard by 70 days. Lower CD4 and older age were strongly associated with mortality. Conclusions: MTB-BSI is far more common for inpatients than outpatients with HIV. Across multiple settings with high tuberculosis prevalence, MTB-BSI was strongly associated with heightened risk of early mortality in PWH. Novel optimized diagnostic and treatment strategies are still needed for HIV-associated MTB-BSI.