Limited added value of T-SPOT.TB blood test in diagnosing active TB: a prospective bayesian analysis.

OBJECTIVES

To determine the diagnostic value of a blood interferon-gamma release assay in suspected active tuberculosis (TB).

METHODS

136 subjects with suspected pulmonary TB (pTB) at a single London centre with intermediate TB incidence, were clinically graded into low (<25%), medium, or high (>75%) likelihood of active pTB and then tested by T-SPOT®.TB assay. The diagnosis was confirmed by culture (n = 33), treatment response (n = 13) or a firm alternative diagnosis (n = 90).

RESULTS

Overall, the T-SPOT.TB sensitivity was 74% (95% confidence intervals 60-84%), positive predictive value (PPV) 56% (43-68%), negative predictive value (NPV) 83% (71-90%), positive likelihood ratio (PLR) 1.75 and negative likelihood ratio (NLR) 0.45. Results for high pTB likelihood subjects: PPV 100%, NPV 25% (7-60%), PLR >69, NLR 0.31. Results for intermediate pTB likelihood subjects: PPV 67% (41-85%), NPV 88% (65-96%), PLR 2.39, NLR 0.26. Results for low pTB likelihood subjects: PPV 15% (6-34%), NPV 92% (79-97%), PLR 1.23, NLR 0.80. False negatives occurred in 24% of cases of active tuberculosis (4 smear and culture-positive, 3 smear negative and culture-positive, and 4 culture negative).

CONCLUSIONS

The predictive values and likelihood ratios show the T-SPOT.TB test does not assist in confidently confirming or excluding active TB, regardless of the pre-test probability of disease.