Risk factors for false-negative results of T-SPOT.TB and tuberculin skin test in extrapulmonary tuberculosis

T-SPOT.TB, a recently developed T cell-based assay, has shown promise in diagnosing extrapulmonary tuberculosis (EPTB). However, a limited number of reports have compared the risk factors for false-negative results of tuberculin skin tests (TSTs) and T-SPOT.TB assays in patients with EPTB. We, thus, conducted a prospective, blinded, observational study to evaluate the risk factors for false-negative T-SPOT.TB and TST results in patients with EPTB. Between April 2008 and November 2011, all adult patients with suspected EPTB were prospectively enrolled at Asan Medical Center, Seoul, South Korea (an intermediate TB-burden country). Only patients with confirmed and probable EPTB who underwent TST and T-SPOT.TB were included in the final analysis. Of the 324 patients who underwent both TST and T-SPOT.TB testing, 128 patients with 96 (75 %) culture- or polymerase chain reaction (PCR)-confirmed and 32 (25 %) probable EPTB were finally analyzed. T-SPOT.TB assays were less likely to yield false-negative results than TSTs [17 % (22/128) vs. 54 % (69/128), p < 0.001]. In a multivariate analysis, miliary TB was associated with false-negative TSTs [odds ratio (OR) = 5.3; 95 % confidence interval (CI) 1.7-16.1], while immunosuppression showed a trend toward false-negative TSTs (OR = 2.5; 95 % CI 0.9-6.8). Conversely, lymph node TB (OR = 0.2; 95 % CI 0.1-0.5) and skeletal TB (OR = 0.2; 95 % CI 0.1-0.5) were associated with true-positive TST results. The only risk factor for false-negative T-SPOT.TB results was TB meningitis (OR = 2.6; 95 % CI 1.0-6.6). Our findings suggest that T-SPOT.TB has a better sensitivity to diagnose EPTB than TST, especially in patients with immunosuppression or miliary TB.