Additional role of second washing specimen obtained during single bronchoscopy session in diagnosis of pulmonary tuberculosis

BACKGROUND Flexible bronchoscopy with bronchial washing is a useful procedure for diagnosis of pulmonary tuberculosis (TB), when a patient cannot produce sputum spontaneously or when sputum smears are negative. However, the benefit of gaining serial bronchial washing specimens for diagnosis of TB has not yet been studied. Therefore, we conducted a retrospective study to determine the diagnostic utility of additional bronchial washing specimens for the diagnosis of pulmonary TB in suspected patients. METHODS A retrospective analysis was performed on 174 patients [sputum smear-negative, n = 95 (55%); lack of sputum specimen, n = 79 (45%)] who received flexible bronchoscopy with two bronchial washing specimens with microbiological confirmation of pulmonary TB in Samsung Medical Center, between January, 2010 and December, 2011. RESULTS Pulmonary TB was diagnosed by first bronchial washing specimen in 141 patients (81%) out of 174 enrolled patients, and an additional bronchial washing specimen established diagnosis exclusively in 22 (13%) patients. Smear for acid-fast bacilli (AFB) was positive in 46 patients (26%) for the first bronchial washing specimen. Thirteen patients (7%) were positive only on smear of an additional bronchial washing specimen. Combined smear positivity of the first and second bronchial washing specimens was significantly higher compared to first bronchial washing specimen alone [Total cases: 59 (34%) vs. 46 (26%), p < 0.001; cases for smear negative sputum: 25 (26%) vs. 18 (19%), p = 0.016; cases for poor expectoration: 34 (43%) vs. 28 (35%), p = 0.031]. The diagnostic yield determined by culture was also significantly higher in combination of the first and second bronchial washing specimens compared to the first bronchial washing. [Total cases: 163 (94%) vs. 141 (81%), p < 0.001; cases for smear negative sputum: 86 (91%) vs. 73 (77%), p < 0.001; cases for poor expectoration: 77 (98%) vs. 68 (86%), p = 0.004]. CONCLUSIONS Obtaining an additional bronchial washing specimen could be a beneficial and considerable option for diagnosis of TB in patients with smear-negative sputum or who cannot produce sputum samples.