UCLA

RATIONALE: Recent empirical research suggests that isoniazid may lead to a risk reduction of incident tuberculosis (TB) among close TB contacts of someone with multidrug-resistant TB (MDR-TB). OBJECTIVES: To evaluate the association between isoniazid TB preventive treatment (TPT), compared to no treatment, upon incident TB in household contacts of MDR-TB cases using a large global consortium of TB contact tracing studies. METHODS: We conducted a systematic review and individual-participant meta-analysis among observational studies of household contact tracing studies. Participants were included if they were exposed to someone with MDR-TB and were given either 6 months of isoniazid TPT or no TPT. Our primary outcome was incident TB in contacts exposed to TB. We derived adjusted hazard ratios (aHRs) using mixed-effects, multivariable survival regression models with study-level random effects. The effectiveness of isoniazid TPT against incident TB was estimated through propensity score matching. MEASUREMENTS AND MAIN RESULTS: We included participant-level data from 6668 contacts exposed to MDR-TB from 17 countries. The effectiveness of isoniazid TPT against incident TB in contacts of MDR-TB was 57% (aHR, 0.43 [95% CI, 0.26-0.71]) and did not appreciably change with adjustment for additional potential confounders. The reduction in incident TB was marginally greater among child (<20 years old) contacts (aHR, 0.51 [95% CI, 0.28-0.92) compared to adult contacts (aHR, 0.69 [95% CI, 0.22-2.20]). The reduction in incidence was 73% (aHR, 0.27 [95% CI, 0.11-0.70]) in the first year of follow-up; effectiveness dropped to 60% (aHR, 0.40 [95% CI, 0.15-1.06]) from 12 to 23 months of follow-up and was nonsignificant after 2 years (28% effectiveness; aHR, 0.72 [95% CI, 0.33-1.54]). CONCLUSIONS: Among >6500 contacts of MDR-TB, isoniazid TPT was highly effective in preventing incident TB. The reduction was greatest in high-burden countries and waned after 2 years of follow-up.