Prevention of anti-tumor necrosis factor-associated tuberculosis: a 10-year longitudinal cohort study.

The extent to which anti-tumor necrosis factor (TNF)-associated tuberculosis can be prevented is unclear, and there is no established guidance on the optimal screening strategy for latent tuberculosis (LTBI) in patients about to start anti-TNF therapy. We aimed to determine the effectiveness of a comprehensive program for the prevention of anti-TNF-associated tuberculosis, and to evaluate 3 LTBI screening strategies and the need for retesting patients with negative results at baseline. In total, 726 patients were screened prior to anti-TNF therapy using 1 of 3 diagnostic strategies over 3 consecutive periods: first, a 2-step tuberculin skin test (TST); second, a 2-step TST plus QuantiFERON-TB Gold In-Tube test (QFT-GIT) (2-step TST/QFT); and third, a single-step TST plus QFT-GIT (TST/QFT). Infected patients were offered preventive therapy. We assessed differences in the incidence of tuberculosis between anti-TNF exposed and nonexposed patients, and between the 3 study periods. Tuberculosis developed during the first year in 2.85 per 1000 exposed patient-years (3/1052 patient-years) and 1.77 per 1000 nonexposed patient-years (1/566 patient-years). No cases occurred beyond the first year of treatment. LTBI diagnoses decreased with the single-step TST/QFT (26.5%) compared with the 2-step TST (42.5%; P < .001) and 2-step TST/QFT (38.5%; P = .02); the incidence of tuberculosis among exposed patients did not change significantly across the 3 periods (2.63/1000, 3.91/1000, and 2.4/1000 patient-years, respectively). Although anti-TNF-associated tuberculosis can be reduced, some risk remains during the first year of therapy. Neither the 2-step TST nor systematic retesting after negative baseline testing is justified.