Introduction and Objectives The Gene Xpert® MTB/RIF test has been validated in sputum samples, facilitating rapid mycobacterium tuberculosis (MTB) diagnosis with improved sensitivity compared to smear alone. Its utility in bronchoalveolar lavage (BAL) samples, in a low prevalence setting, is unclear. Our objective was to examine the sensitivity of the Xpert® test in BAL samples and evaluate its use as a rapid diagnostic test in non-productive or smear negative patients undergoing bronchoscopy.
Methods We conducted a retrospective analysis of all culture-proven MTB samples acquired by BAL between 01.08.2009 and 01.06.2013, which were also sent for the Xpert® test, n = 38. We assessed the proportion of Xpert® MTB positive samples and compared these results with smear status and time to culture positivity.
Results 26/38 of the culture-proven cases sent for analysis were MTB Xpert® positive, giving a sensitivity (if culture is taken as the 'gold standard') of 68%. 17/38 samples were smear positive. Of the 21 smear negative/culture positive samples, 43% were MTB Xpert® positive. All smear positive cases were Xpert® positive. The Xpert® positive samples had a lower mean time to positivity: 11.2 days vs. 17.2 days for Xpert® negative samples. There was one case of rifampicin resistance correctly identified by Gene Xpert®, giving a lead time of 28 days vs. culture sensitivity. Incidentally 4 additional Xpert® positive samples were identified which were BAL culture negative, all in patients with clinically likely MTB, one of whom went on to culture MTB from sputum, and one of whom had necrotising granulomas on subsequent endobronchial ultrasound-guided transbronchial needle aspiration.
Conclusions To our knowledge this is the largest series reported of MTB culture positive BAL samples analysed using the Xpert® MTB/RIF assay. This data supports the use of Xpert® in the diagnosis of pulmonary MTB in BAL fluid, with a sensitivity of 68% when compared to MTB culture, adding additional value to simple smear and early detection of rifampicin resistance. In fact, the true sensitivity may well be higher given the cases detected exclusively by Xpert®.