Annual Incidence of Latent Tuberculosis Infection among Newly Employed Nurses at a Tertiary Care University Hospital

OBJECTIVE. To investigate the annual rate of tuberculosis (TB) infection among newly employed nurses using both tuberculin skin test (TST) and QuantiFERON-TB Gold (QFT-G; Cellestis Limited) assay. DESIGN. A prospective cohort study involving newly employed nurses. SETTING. A tertiary care university hospital in South Korea. METHODS. All participants (n = 196) were tested with the TST and QFT-G assay at baseline. After 1 year, the TST and QFT-G assay were reperformed for subjects who had negative TST results at baseline and for all subjects, respectively. RESULTS. The baseline TST and QFT-G assays were positive for 101 subjects (51.5%) and 28 subjects (14.3%), respectively; 22 subjects (11.2%) had positive results of both tests. Although the overall between-test agreement was 54.9% (kappa = 0.151 [95% confidence interval, 0.047-0.245]), agreement improved to 78.5% (kappa = 0.462 [95% confidence interval, 0.007-0.917]) for subjects who had not received bacille Calmette-Guerin vaccination. After 1 year, the TST yielded positive results for 16 (21.3%) of 75 nurses with negative baseline results, and the QFT-G assay yielded positive results for 21 (14.4%) of 146 subjects with negative baseline results. Collectively, 5 subjects (3.0%) experienced conversion to positive results with both tests, and 32 subjects (18.9%) experienced conversion to positive results with one of the tests. Neither the employing hospital department nor exposure to patients with TB affected test conversion status. CONCLUSIONS. The poor overall agreement between TST and QFT-G results may have been caused by the confounding effect of bacille Calmette-Guerin vaccination. The annual risk of TB infection among newly employed nurses was at least 3% on the basis of results of both the TST and QFT-G test. Stricter preventive strategies against TB spread should be implemented in our hospital.