Performance evaluation of the Cobas TaqMan MTB assay on respiratory specimens according to clinical application

Objective To evaluate the performance of the Cobas TaqMan MTB assay (Cobas assay) with respect to its clinical application. Methods This was a retrospective analysis of 1154 results from 1034 patients for whom mycobacterial cultures and the Cobas assay were performed simultaneously. Based on the patient medical records, two categories of clinical application were defined: (1) the diagnosis of patients with a high probability of pulmonary tuberculosis according to clinical and radiological features (n = 128), and (2) the exclusion of tuberculosis in clinically indeterminate patients (n = 1026). Standard culture was used as the reference method. Results The sensitivity of the Cobas assay for the detection of Mycobacterium tuberculosis was 70.4% (95% confidence interval (CI) 49.7–85.5%) for category 1, but only 25.0% (95% CI 4.5–64.4%) for category 2. The specificity was ≥95.0% for both categories. The positive predictive value was 79.2% (95% CI 57.3–92.1%) for category 1 and 33.3% (95% CI 6.0–75.9%) for category 2, while the negative predictive value was 92.3% (95% CI 85.0–96.4%) for category 1 and 99.4% (95% CI 98.7–99.8%) for category 2. Conclusions The results of this study indicate that Cobas assay results must be interpreted carefully according to the clinical purpose of the assay.