[Induced expectoration or endoscopic samples in case of bronchopulmonary infection?].

The aim of this study was to compare induced sputum versus bronchoscopy in a non selected population. An induced sputum and a bronchoscopy with aspiration of secretions and bronchoalveolar lavage (BAL) were proposed to 30 patients with suspected lower respiratory tract infection, including 14 subjects with AIDS. Induced sputum failed in 3 patients, while endoscopy could not be performed (contra-indication, refusal or failure due to lack of cooperation) in 7 patients; a technical failure for BAL was observed in 3/23 cases. Contamination by oral germs was significantly less frequent with BAL (4/20) than with aspiration (15/23) or induced sputum (17/27). A relevant pathogen was cultured from induced sputum in 7/27 cases (3 mycobacteria, 4 usual pathogens) from aspiration in 3/23 cases (one mycobacterium, 2 usual pathogens) and from BAL in 4/20 cases (one mycobacterium, 3 usual pathogens). These results suggest that induced sputum is a promising method when difficulties in performing bronchoscopy are expected (severe respiratory insufficiency, psychiatric disease, lack of cooperation).