Article Text
Abstract
Background Tackling tuberculosis (TB) requires testing and treatment of high-risk groups for latent tuberculosis infection. We estimated the predictive values of the tuberculin skin test (TST) and interferon gamma release assays (IGRAs) for development of active TB in migrants and contacts of active TB patients in the UK.
Methods Participants were prospectively recruited in clinics and the community and followed for a median of 2.9 years. We administered IGRAs (Quantiferon Gold In-Tube [QFT-GIT] and T-SPOT.TB) and TST (with 3 thresholds: 5 mm (TST5), 10 mm (TST10) and TST15 (5 mm in BCG-naïve, 15 mm in vaccinated). Potential incident TB cases were identified by telephone interview and national TB databases and confirmed by medical note review.
Results Ninety-seven (1.0%) of 9610 participants developed active TB (77 of 6386 who had Results for T-SPOT.TB, QFT-GIT and TST). All tests had very low incidence in test negatives (1.2–1.6 per 1000 per year). Incidence rates in test positives were highest for TSpot.TB (13.2 95% CI: (9.9–17.4)), TST15 (11.1 (8.3,14.6)) and QFT.GIT (10.1 (7.4,13.4)); positive test Results for these tests were significantly more predictive of progression than TST10 and TST5, TSpot.TB was also higher than QFT.GIT. TST5 predicted more at high risk (55%) than TST10 (45%), TSpot.TB (33%), TST15 (38%) and QFT.GIT (31%).
Conclusions IGRA-based or TST15 strategies are most suited for population screening in low-risk populations. Although TST5 and TST10 detect more TB cases this is at the cost of more individuals being classified at high risk with lower positive predictive values.