The effectiveness of isoniazid preventive treatment against tuberculosis among contacts of multidrug-resistant tuberculosis: A systematic review and individual-participant meta-analysis

ObjectiveTo evaluate the impact of isoniazid on incident tuberculosis in household contacts of MDR tuberculosis cases. DesignSystematic review and individual-participant meta-analysis. Data sourcesMEDLINE, Web of Science, BIOSIS, and Embase without language restrictions for case-contact cohort studies of tuberculosis contacts. Eligibility criteria and data analysisHousehold contact tracing studies that investigated the development of tuberculosis in persons closely exposed to individuals with tuberculosis and followed for incident disease. Both retrospective and prospective cohort studies were included. Participants were included if they were exposed to someone with multidrug-resistant tuberculosis and were given either 6 months of isoniazid TPT or no TPT. Two reviewers independently assessed quality using a modified quality assessment of tool. We derived adjusted hazard ratios (aHRs) for incident tuberculosis using mixed-effects, multivariable Cox regression models with study-level random effects. The effectiveness of isoniazid TPT against incident tuberculosis was estimated through propensity score matching. We stratified our results by contact age, HIV, and Mycobacterium tuberculosis infection status. Main outcome measuresOur primary outcome was incident tuberculosis in contacts exposed to tuberculosis (defined as a diagnosis >90 days after baseline). We derived adjusted hazard ratios (aHRs) for incident tuberculosis using mixed-effects, multivariable Cox regression models with study-level random effects. ResultsWe included participant-level data from 4,945 contacts exposed to multidrug-resistant tuberculosis from eight countries. The effectiveness of 6 months of isoniazid TPT against tuberculosis in contacts of multidrug-resistant tuberculosis was 70% (aHR, 0.30; 95% CI, 0.16-0.56) and did not appreciably change with adjustment for additional potential confounders. Effectiveness was higher among contacts <18 years of age (aHR, 0.39; 95% CI, 0.18-0.87) compared to adult contacts (aHR, 0.44; 95% CI, 0.14-1.41). Effectiveness was 93% (aHR, 0.07; 95% CI, 0.02-0.52) in the first year of follow-up; effectiveness dropped to 80% (aHR, 0.20; 95% CI, 0.05-0.89) in the second year and was non-significant after two years (26% effectiveness; aHR, 0.74; 95% CI, 0.34-1.59). ConclusionsAmong almost 5,000 contacts of multidrug-resistant tuberculosis cases, isoniazid IPT was 70% effective against incident tuberculosis. Protection waned after 2 years of follow-up. These results have important implications for the clinical management of individuals exposed to multidrug-resistant tuberculosis and future clinical trials.