Evaluation of the Cobas TaqMan MTB test for detection of Mycobacterium tuberculosis complex

Background: The Cobas TaqMan MTB assay is used for rapid detection of the Mycobacterium tuberculosis complex (MTC) in clinical samples. It is only validated for respiratory samples, but is often requested by physicians for non-respiratory specimens. The aim of this study was therefore to evaluate the performance of this assay in clinical praxis in a country with low prevalence of tuberculosis (TB). Methods: The results from 2388 respiratory and 1005 non-respiratory human specimens were analyzed by this real-time PCR technique. Using culture results as the gold standard, the sensitivity, specificity, positive predictive value (PPV), and negative predictive values (NPV) of the PCR results were calculated. Results: For smear-positive respiratory specimens all the four values investigated were 100%. The number of smear-positive non-respiratory specimens was only eight, and all of these eight specimens reacted positively. Sensitivity was 51% for smear-negative respiratory specimens, and 47% for smear-negative non-respiratory specimens. When all non-respiratory specimens were analyzed together the sensitivity was 51%. Specimens from 16 patients were PCR-positive but culture-negative and these cases are discussed. In one case, TB DNA was still present in sputum 2 years after a successful treatment. Conclusion: The study shows that the Cobas TaqMan MTB assay performs well in smear-positive samples, while the sensitivities are unsatisfactory for both respiratory and non-respiratory smear-negative specimens. Furthermore, the analyses emphasize that this assay cannot be used to evaluate treatment or contagiousness, or to detect relapses or screen for TB.