Performance of Xpert MTB/RIF RUO assay and IS6110 real-time PCR for Mycobacterium tuberculosis detection in clinical samples.

The Cepheid Xpert MTB/RIF research-use-only (RUO) assay and a laboratory-developed test (LDT) targeting IS6110 were evaluated and compared to mycobacterial culture as the gold standard. The performance characteristics of both molecular assays were determined by using 112 specimens from 90 patients, including 89 pulmonary specimens and 23 extrapulmonary specimens. Of the specimens tested, 37 (33%) were culture positive for the Mycobacterium tuberculosis complex; 29 were pulmonary, and 8 were extrapulmonary. Of these culture-positive specimens, 83% of the pulmonary specimens and 50% of the extrapulmonary specimens were smear positive. There was complete concordance between the smear-positive culture-positive specimens, independent of the anatomical site (100% sensitivity). The sensitivity of the MTB/RIF RUO assay for smear-negative specimens was 60% for pulmonary and 75% for extrapulmonary specimens, while the IS6110 LDT sensitivities were 40% and 0%, respectively. There was also complete concordance among the culture-negative specimens tested. Both assays showed 95% specificity, with four culture-negative specimens testing as positive. A review of patient records indicated that there was a high likelihood of the presence of M. tuberculosis complex DNA in the false-positive specimens. Biosafety analysis was performed and showed an acceptable reduction in organism viability using the processing methods described above. Both molecular assays are suitable for the detection of M. tuberculosis isolates in smear-positive pulmonary and extrapulmonary specimens, while the sensitivity of the detection of M. tuberculosis isolates in smear-negative specimens was variable.