Background: Inhaled corticosteroids (ICS) markedly reduce bronchial mucosal inflammation in asthma but whether they have an anti-inflammatory effect in airway tissue in chronic obstructive pulmonary disease (COPD,) is unknown. Methods: A, study of endobronchial biopsy samples was conducted as part of a double blind, placebo controlled, randomised trial of parallel design. Patients had mild to moderately severe COPD (FEV(1) 25-80% of predicted) and were given 3 months treatment with ICS, fluticasone propionate (FP; 500 mug twice daily, n=14) or placebo (n=10). Biopsy tissue taken at baseline and after treatment was examined by transmission electron microscopy to count the numbers of all ultrastructurally distinct inflammatory cells. Results: Compared with their baseline values, FP resulted in a significant decrease (on average 65%) in the numbers of mucosal mast cells (median 7.8 (range 1-33) v 2.8 (1-14), p<0.05). The reductive effect of FP held true when the post-treatment values of the placebo and FP groups were compared: 8.8 (1-24) v 2.8 (1-14) (p<0.05). Unexpectedly, there were significantly more neutrophils in the FP than in the placebo group: 4.0 (0-23) v 1.7 (0-8), respectively (p<0.05). There were no alterations to other cell types including mononuclear cells. Symptoms markedly improved in the patients treated with FP for 3 months, Conclusion: Fluticasone propionate given for 3 months to patients with COPD has selective effects on the inflammatory cells in the bronchial mucosa: the reduction in mast cell numbers may account for the improvement in symptoms over this time.