Background: Tuberculosis Preventive Treatment (TPT) can reduce TB incidence and mortality in people living with HIV. However, low levels of screening and uptake, poor adherence, and loss to follow-up considerably reduce its effectiveness. We aimed, therefore, to assess the losses within all steps of the screening and treatment cascade.
Methods: To enhance data generalizibility we included articles which reported the proportion of people living with HIV completing any step of the TPT cascade in low- and high-TB burden countries published before March 2024. Random effects meta-analysis produced pooled estimates of the proportion proceeding to the next step along the cascade. Results were explored through subgroup analyses and meta-regression. PROSPERO registration: CRD42020162396.
Findings: Data from 368 cohorts containing 2.7 million participants were included. High levels of heterogeneity in outcomes were seen. Most participants were from Africa (80.6%). Isoniazid monotherapy was used for TPT in 92.6% of cohorts, usually for six months. Substantial loss to follow-up was found throughout the treatment cascade with more than one in six patients lost at the following steps: initial screening, immunological testing, treatment start and completion. Treatment regimens lasting four months or less were more likely to be completed than longer ones - 88.4% compared to 61.6%.
Interpretation: Our analysis highlights substantial loss to follow-up at multiple steps during the care cascade. This may significantly lower the reported effectiveness of TPT in real-world settings. Research and policy should focus on simplified care pathways and novel, shorter treatment regimens that optimise retention in care.
Copyright ©The authors 2025. For reproduction rights and permissions contact permissions@ersnet.org.