Duration of effectiveness of tuberculosis preventive treatment across various tuberculosis burden settings: a systematic review and individual-participant meta-analysis.

Abstract

Millions of individuals at risk of tuberculosis are given tuberculosis preventive treatment (TPT) annually. However, questions remain regarding the duration of protection from TPT and potential variability across populations and settings with disparate tuberculosis risk. We aimed to assess the effect of TPT on time to tuberculosis among close contacts of individuals with tuberculosis and with evidence of Mycobacterium tuberculosis infection across high-burden, medium-burden, and low-burden settings. In this systematic review and individual-participant meta-analysis, we searched MEDLINE, PubMed, Web of Science, BIOSIS, and Embase for cohort studies and randomised controlled trials published between Jan 1, 1998, and Aug 28, 2023. Studies were eligible if they included individuals (of any age) with household or close exposure to a person with microbiologically or clinically diagnosed tuberculosis (contacts); had at least 6 months of contact follow-up from baseline; and reported start and end of follow-up dates or days of follow-up. We pooled individual-participant data from included cohorts, including contact and index patient characteristics and setting. Analyses were restricted to participants who tested positive for M tuberculosis infection at baseline and stratified by background tuberculosis burden (low burden was defined as a country-level annual incidence of <20 per 100 000 population, medium as 20-100 per 100 000 population, and high as >100 per 100 000 population). The primary outcome was incident tuberculosis, which was defined as tuberculosis diagnosed more than 90 days after baseline. We calculated incidence rates by TPT status, generated Nelson-Aalen cumulative hazard plots, and estimated adjusted hazard ratios (HRs) using mixed-effects Cox models with a time interaction. This study is registered with PROSPERO, CRD42018087022. Of 466 523 participants from 47 studies, we included 84 128 individuals from 44 cohort studies across 24 countries. TPT was highly protective within the first year but declined over the course of follow-up in all settings, with greater effects observed in low-burden and medium-burden settings. At 1 year, the adjusted HR was 0·57 (95% CI 0·31-1·07) in high-burden settings, 0·14 (0·09-0·21) in medium-burden settings, and 0·21 (0·17-0·26) in low-burden settings. The protective effect persisted up to the end of follow-up in medium-burden settings (adjusted HR at 8 years 0·15 [95% CI 0·07-0·33]) and up to 13 years in low-burden settings (0·65 [0·43-1·00]). In medium-burden and low-burden settings, participants who completed TPT initially had lower risk of tuberculosis than those who did not complete TPT (adjusted HR at 1 year 0·59 [95% CI 0·36-0·95]); within 2 years, these groups were not statistically different (adjusted HR at 2 years 0·72 [0·41-1·26]) and both showed strong protection for at least 8 years. Among close contacts with evidence of M tuberculosis infection, we found that the initial protection from TPT eroded relatively quickly in high-burden settings, but was sustained long-term in medium-burden and low-burden settings. In medium-burden and low-burden settings, completing TPT provided greater initial protection, but TPT non-completers had similar long-term tuberculosis risk. None.