OBJECTIVE To establish clinical determinants affecting the diagnostic yield of bronchial washing.

SETTING We performed bronchial washing in 241 consecutive patients with bronchoscopically invisible lung tumours. Of these, 150 patients known to have lung cancer were enrolled for the final analysis.

DESIGN A multi-centre study.

RESULTS Bronchial washing provided a diagnosis of lung cancer in 30 of the 150 patients (20%). Tumour size > or = 3 cm (P = 0.005), the location of the tumour within 8 cm of the carina (P = 0.003), and exposed type bronchus sign of tumour (P < 0.001) were factors affecting diagnostic bronchial washing for bronchoscopically invisible lung cancers. However, multivariate logistic regression revealed that exposed type bronchus sign was the sole determinant (OR 19.22, 95% CI 4.23-87.46, P < 0.001).

CONCLUSION Bronchial washing is a useful procedure for the diagnosis of bronchoscopically invisible lung cancers. As the tumour-bronchus relationship is the most important determinant of a diagnostic yield, the routine use of bronchial washing should be considered for tumours with exposed type bronchus sign.