We audited our practice of bronchoscoping patients admitted with community acquired pneumonia to detect latent underlying abnormalities. Fifty-five bronchoscopies were performed immediately after clinical recovery in 64 sequential patients without obvious underlying carcinoma. Five (9.1%) showed abnormalities (four carcinomas and one mild tracheal stenosis). The detection rate of abnormalities and the number of unnecessary bronchoscopies would be improved if only those patients aged 50 years or over who were current or ex-smokers were bronchoscoped. In this group 13.9% (5/36) had an underlying abnormality. The cost implications of such a policy of early bronchoscopy are discussed and compared with traditional follow-up.