Poor agreement between diagnostic tests for latent tuberculosis infection among HIV-infected persons in Hong Kong

Background and objective The tuberculin skin test (TST), T-Spot.TB (T-Spot) and QuantiFERON-TB Gold-In Tube (QFT) were compared in diagnosing latent tuberculosis infection (LTBI) among human immunodeficiency virus (HIV)-infected persons. Methods Human immunodeficiency virus-infected persons without previous history of tuberculosis or LTBI were simultaneously tested by TST, T-Spot and QFT annually and followed up for tuberculosis. Results Among 110 HIV-infected subjects with 85% previous TST screening coverage, 75% on anti-retroviral therapy, well-preserved median CD4 count (414/μL) and low median viral load (<75/μL), baseline TST, T-Spot and QFT were positive in 5.5%, 5.6% and 4.9%, respectively, with almost complete discordance of positive results. Among 91 (83%), 66 (60%) and 26 (24%) subjects successfully undergoing the first, second and third annual retesting, TST, T-Spot and QFT were, respectively, positive in 11/123 (8.9%), 13/173 (7.5%) and 21/182 (11.5%) on retesting, with similar discordance of positive results. There was no significant association with the concurrent CD4 count or viral load. Conversion occurred in 11/123 (8.9%), 8/160 (5.0%) and 18/168 (10.7%) of TST, T-Spot and QFT, respectively, and none was associated with changes in CD4 count or viral load. More than half of the positive T-SPOT and QFT results reverted to negative on follow-up. None of these tests picked up the single case of culture-confirmed tuberculosis observed after 798 person-years of follow-up. Conclusion Major discordance in positive results, high reversion rates and low tuberculosis incidence among test-positive subjects cast serious doubt on the utility of the currently available LTBI tests in the annual screening of HIV-infected persons in an intermediate tuberculosis burden area.