Effect of Smoking on Longitudinal Interferon γ Release Assay Results Among Close Contacts of People with Pulmonary Tuberculosis

Smoking among close contacts of people with tuberculosis was associated with lower odds of IGRA reversion. Our findings highlight the importance of smoking on longitudinal IGRA results and clinical research in which IGRA status is used as an outcome. Diagnosis of Mycobacterium tuberculosis infection in close contacts is critical for tuberculosis control. Smoking is a risk factor for M. tuberculosis infection and tuberculosis disease but its effect on longitudinal interferon gamma release assay (IGRA) results remains unknown. We conducted a multisite prospective study in Brazil between 2015 and 2019, among close contacts of adults with culture-confirmed pulmonary tuberculosis. IGRA was performed at baseline, at month 6 if results were negative at baseline, and at months 24-30 after enrollment. IGRA results were categorized as IGRA positive (maintained from baseline to the last visit), IGRA conversion (from negative to positive at any time), IGRA reversion (from positive to negative at any time), and IGRA negative (maintained from baseline to the last visit). Associations between IGRA results and smoking status at baseline (current/former vs never) in contacts were evaluated using propensity score-adjusted logistic regression models. Estimated propensity score was used as a covariate in models, which regressed the outcome (IGRA positive, IGRA conversion, IGRA reversion) on smoking status. Of 430 close contacts, 89 (21%) were IGRA positive, 30 (7%) were converters, 30 (7%) were reverters and 22 were indeterminate. Smoking frequency was 26 (29%) among IGRA-positive contacts, 7 (23%) in converters, and 3 (10%) in reverters. Smoking in contacts was associated with lower odds of IGRA reversion (adjusted odds ratio, 0.16 [95% confidence interval, .03-.70]). We did not detect associations between smoking and IGRA positive or IGRA conversion. Our findings highlight the importance of smoking on longitudinal IGRA results. This has implications for clinical care and clinical trials in which IGRA status is monitored or used as an outcome.