BACKGROUND Diagnosis of latent tuberculosis infection is a problem in children because of lack of a diagnostic standard and potential impact of previous Bacille Calmette-Guerin vaccination and exposure to environmental mycobacteria. Effectiveness and usefulness of interferon-gamma release assays in infants and younger children have yet to be clearly demonstrated. METHODS Prospective cohort study including 109 children (4 months to 16 years) seen in an international adoption clinic at Nationwide Children's Hospital, Columbus, OH. Children were adopted from 14 countries, mostly (72.5%) from China, Russia and Ethiopia. Correspondence between tuberculin skin test (TST) and the T-SPOT.TB assay was evaluated. Factors associated with positive results on the TST and T-SPOT.TB were determined, and the impact of age on test performance was specifically addressed. RESULTS TST was positive in 23.4% (25 of 107). T-SPOT.TB was positive in 4.6% (5 of 109). Overall agreement between TST and T-SPOT.TB was 71%, with prevalence-adjusted, bias-adjusted Kappa of 0.68. History of Mycobacterium tuberculosis exposure was associated with positive results on TST (odds ratio: 25.4, 95% confidence interval: 4.8-261.6, exact logistic regression) and T-SPOT.TB (odds ratio: 78.9, 95% confidence interval: 9.7-∞). All 5 children with positive T-SPOT.TB had TST induration ≥15 mm. No patient less than 1 year of age (n = 17) had positive TST or T-SPOT.TB. Positive TST was not associated with Bacille Calmette-Guerin vaccination or scar. CONCLUSIONS TST was positive in a significant percentage of international adoptees. T-SPOT.TB was rarely positive and discordant results reflected negative T-SPOT.TB with positive TST. In this population latent tuberculosis infection may be over-estimated by TST. Regardless, in our context at the time of the study, treatment decisions were based upon TST results, not results of the T-SPOT.TB assay. Age was consistently associated with findings on TST and T-SPOT.TB with no positive result on either test for any child
