Detection of Mycobacterium tuberculosis infection in United States Navy recruits using the tuberculin skin test or whole-blood interferon-gamma release assays

Abstract

Background: Military personnel are at risk for acquiring Mycobacterium tuberculosis infection because of activities in close quarters and in regions with a high prevalence of tuberculosis (TB). Accurate tests are needed to avoid unnecessary treatment because of false-positive results and to avoid TB because of false-negative results and failure to diagnose and treat M. tuberculosis infection. We sought to estimate the specificity of the tuberculin skin test (TST) and 2 whole-blood interferon-gamma release assays (QuantiFERON-TB assay [QFT] and QuantiFERON-TB Gold assay [QFT-G]) and to identify factors associated with test discordance.

Methods: A cross-sectional comparison study was performed in which 856 US Navy recruits were tested for M. tuberculosis infection using the TST, QFT, and QFT-G.

Results: Among the study subjects, 5.1% of TSTs resulted in an induration > or = 10 mm, and 2.9% of TSTs resulted in an induration > or = 15 mm. Eleven percent of QFT results and 0.6% of QFT-G results were positive. Assuming recruits at low risk for M. tuberculosis exposure were not infected, estimates of TST specificity were 99.1% (95% confidence interval [CI], 98.3%-99.9%) when a 15-mm cutoff value was used and 98.4% (95% CI, 97.3%-99.4%) when a 10-mm cutoff value was used. The estimated QFT specificity was 92.3% (95% CI, 90.0%-94.5%), and the estimated QFT-G specificity was 99.8% (95% CI, 99.5%-100%). Recruits who were born in countries with a high prevalence of TB were 26-40 times more likely to have discordant results involving a positive TST result and a negative QFT-G result than were recruits born in countries with a low prevalence of TB. Nineteen (50%) of 38 recruits with this type of discordant results had a TST induration > or = 15 mm.

Conclusions: The QFT-G and TST are more specific than the QFT. No statistically significant difference in specificity between the QFT-G and TST was found using a 15-mm induration cutoff value. The discordant results observed among recruits with increased risk of M. tuberculosis infection may have been because of lower TST specificity or lower QFT-G sensitivity. Negative QFT-G results for recruits born in countries where TB is highly prevalent and whose TST induration was > or = 15 mm suggest that the QFT-G may be less sensitive than the TST. Additional studies are needed to determine the risk of TB when TST and QFT-G results are discordant.