Rates and Risk Factors for On‐Treatment Mortality Among a Cohort of Adults Treated for Drug‐Sensitive Tuberculosis: Analysis of Data From the Adherence Support Coalition to End Tuberculosis Consortium in Five Countries
Amare Worku Tadesse, Noriah Maraba, Jason Alacapa, Katya Gamazina, Tanyaradzwa Dube, Baraka Onjare, Norma Madden, Salome Charalambous, et al. (11 authors)
Tropical Medicine & International Health · 2025-10
Abstract
BACKGROUND: Tuberculosis remains a leading cause of death globally, particularly in countries with high tuberculosis and HIV burdens. Disruptions caused by the COVID-19 pandemic may have further impacted tuberculosis outcomes. This study examines on-treatment mortality and associated risk factors in five countries. METHOD: We conducted a secondary analysis of data from ASCENT cluster-randomised trials of digital adherence tools for improved adherence involving 23,799 adults with drug-sensitive tuberculosis in South Africa, Tanzania, Ethiopia, the Philippines, and Ukraine. Analyses were conducted separately by country. Mortality rates were measured from treatment initiation to the earliest of 6 months, death, or loss to follow-up. Cox regression models (with random effects or robust standard errors for clustering) assessed the associations between mortality and HIV status, ART use, tuberculosis diagnosis type, and calendar periods (COVID-19 pandemic and conflict in Ukraine). RESULTS: Mortality rates ranged from 7.6 (Ethiopia) to 23.2 (Tanzania) and 23.3 (Ukraine) per 100 person-years. Higher mortality was associated with: older age in all countries (age < 30 versus ≥ 60 years, adjusted rate ratio [aRR] ranging from 2.38 to 6.57 by country); HIV status (positive versus negative, aRR ranging from 1.44 to 2.98 by country); tuberculosis diagnosis type (clinical vs. bacteriological, aRR 1.5-1.6 in Ethiopia, Tanzania and South Africa); extrapulmonary tuberculosis (aRR 1.44 to 1.60 in Ukraine and Tanzania). ART versus HIV-positive not on ART was linked to lower mortality in South Africa and Ukraine but not in Tanzania. Analyses suggested possible mortality variations by calendar period. CONCLUSION: Our findings suggest variability in tuberculosis mortality across settings, influenced by HIV/ART and diagnosis type. The high mortality rates across countries may reflect underlying causes or potential misdiagnoses. Further investigation into these factors may be needed to improve tuberculosis outcomes globally.
MeSH terms
- Medicine
- Tuberculosis
- Cohort
- Environmental health
- Cohort study
- Mortality rate
- Developing country
- Demography
- Developed country
- MEDLINE
- Population
- Public health
- Epidemiology
- Young adult
- Cohort effect
- Risk assessment