Efficacy of endoscopic diagnostics in smear-negative TB

Background. Diagnostics of smear-negative TB often is a challenging situation for both clinician and endoscopist. Data regarding usefulness of different endoscopic biopsies are scarce. Aim. To assess the comparative yield of different endoscopic biopsies in diagnostics of smear-negative pulmonary TB. Methods. 203 pts with smear-negative pulmonary TB, referred to bronchoscopy were retrospectively analyzed. Diagnostic yield for brushing and|or washing/BAL in terms of diagnostic yield by luminescence, automated system for PCR detection, Lowenstein-Jensen culture and Bactec MGIT 960 were calculated. Results. 82 pts (40,4%) underwent brushing combined with washing/BAL (Group I), 121 pts (59,6%) had a liquid-based biopsy: BAL/washing (Group II). Diagnostic yield for Group I was 17/82 (20,7%), 35 (42,6%), 4 (4,9%), and 23 (28,1%) for luminescence, PCR, LY, and BACTEC, respectively. Diagnostic yield for Group II was 21/121 (17,4%), 55 (45,5%), 6 (4,9%), and 51 (42,1%) for the same biopsy types, respectively. No significant difference was observed between two approaches of biopsies. Diagnostic yield for all types of biopsy combined for Group I reached 44/82 (53,7%), and for Group II reached 68/121 (56,1%), and also did not differ between groups. Conclusions. Total efficacy of endoscopic diagnostics of smear-negative pulmonary TB varies around 55% for all types of biopsies combined, and does not significantly differ between liquid and smear approaches. The most sensitive technique to detect TB is PCR, which can allow to detect three-thirds of all detectable TB cases by all methods combined.