Background: beta (2)-Agonists and corticosteroids or theophylline can interact to produce beneficial effects on airway function in asthma, but this has not been established in COPD. Methods: Eighty patients with well-controlled COPD were randomized to receive 3 months of treatment in one of four treatment groups: (1) salmeterol, 50 mug bid; (2) salmeterol, 50 mug, plus fluticasone propionate, 250 mug bid; (3) salmeterol, 50 mug, plus fluticasone propionate, 500 mug bid; and (4) salmeterol, 50 mug, plus titrated theophylline bid. At each visit, a dose-response curve to inhaled salbutamol was constructed using a total cumulative dose of 800 mug. Results: A gradual increase in FEV1 was observed with each of the four treatments. Maximum significant increases in FEV1 over baseline values that were observed after 3 months of treatment were as follows: salmeterol, 50 mug bid, 0.163 L (95% confidence interval [CI], 0.080 to 0.245 L); salmeterol, 50 mug, plus fluticasone propionate, 250 mug bid, 0.188 L (95% CI, 0.089 to 0.287 L); salmeterol, 50 mug, plus fluticasone propionate, 500 mug bid, 0.239 L (95% CI, 0.183 to 0.296 L); and salmeterol, 50 mug, plus titrated theophylline bid, 0.157 L (95% CI, 0.027 to 0.288 L). Salbutamol always caused a significant dose-dependent increase in FEV1 (p < 0.001), although the 800-<mu>g dose never induced further significant benefit when compared with the 400-mug close. The mean differences between the highest salbutamol FEV1 after salmeterol, 50 mug, plus fluticasone propionate, 500 mug bid, and that after salmeterol, 50 mug, plus titrated theophylline bid or salmeterol, 50 mug bid, were statistically significant (p < 0.05). Conclusion: These data show that both long-acting <beta>(2)-agonists and inhaled corticosteroids have a role in COPD. The data also show that fluticasone propionate and salmeterol given together are more effective than salmeterol alone. Moreover, it suggests that the addition of fluticasone propionate to salmeterol allows a greater improvement in lung function after salbutamol, although regular salmeterol is able to improve lung function in COPD patients without development of a true subsensitivity to its bronchodilator effect. In any case, patients must be treated for at least 3 months before a real improvement in lung function is achieved.