Introduction Endobronchial ultrasound with transbronchial needle aspiration (EBUS-TBNA) provides a technique for safely sampling mediastinal and hilar lymph nodes. In the diagnosis of tuberculosis (TB) published data has shown excellent performance in identifying suggestive cytology but with disappointing culture-confirmation rates of <50%. We sought to investigate the proportion of TB cases referred to our tertiary referral service in whom culture and sensitivities were obtained.
Methods EBUS-TBNA was performed by 2 experienced consultants with on-site cytology review. Data was abstracted from our prospective database of all EBUS-TBNA cases between 01/2008 and 01/2013, our hospital electronic record and by contacting referring clinicians. A final diagnosis of active TB was made if treatment for active TB was commenced subsequent to EBUS-TBNA. Treated TB was defined as anyone who had received at least 2 months of anti-TB treatment prior to EBUS-TBNA.
Results A final diagnosis of active or treated TB was made in 142 of 2121 EBUS-TBNA cases (6.7%). Sampled nodes were: right para-tracheal 32%, left para-tracheal 9%, right hilar 16%, left hilar 8%, sub-carinal 35%. A median of 7 passes (range: 4–14) were performed per case. Granulomas with and without necrosis were identified in 91.5%, necrosis alone in 4.9% and lymphocytes alone in 3.5%. Of 118 (73%) diagnosed with active TB, culture-confirmation was obtained in 89 (75%) with a median time to positive liquid culture of 27 days (range: 4–42). Cytology from cases of treated TB and active TB with and without culture-confirmation are compared in the Table.
Discussion These data suggest that EBUS-TBNA can obtain higher proportions of culture-confirmed TB than has been previously reported. The similar cytology profile seen in active TB cases regardless of smear status, culture result or treatment history highlights the need for both adequate samples to be sent for culture and for improved TB diagnostics.