Abstract
Standard diagnostics for Mycobacterium tuberculosis (MTB) including acid-fast bacilli (AFB) smear and culture, and Xpert™ MTB/RIF real-time Polymerase Chain Reaction (RT-PCR; Xpert) have variable sensitivity and/or long turnaround times. We describe the clinical performance of a laboratory-developed tissue-based MTB PCR compared with AFB culture and Xpert using a composite reference standard (CRS). Over an 8-year period, MTB PCR was performed on pulmonary, pleural, or lymph node specimens for 36 patients. Of these, 11 met criteria for confirmed/probable MTB using CRS. MTB PCR was positive in 100% (11/11), AFB cultures were positive in 73% (8/11), and Xpert in 0% (0/4). MTB PCR was negative in 25 cases of "No MTB" (100% specific). The MTB PCR assay resulted faster than positive AFB culture (mean time 4.3 versus 21.2 days). Tissue-based MTB PCR was associated with increased and rapid detection of MTB, improving clinical sensitivity in strongly suspected MTB cases.