Formoterol Turbuhaler has been suggested for as-needed use in asthmatic patients. We investigated whether regular treatment with formoterol would modify the dose-response curves to formoterol in patients with partially reversible COPD. In this randomised, double-blind, cross-over study taking place over four non-consecutive days 16 outpatients with moderate to severe COPD, who were under regular treatment with formoterol Turbuhaler (18 microg in two daily doses) from at least 4 months, inhaled a conventional dose of formoterol Turbuhaler 9 microg or placebo. Two hours later, a FEV(1) value was established, following which a dose-response curve to formoterol (4.5 microg/inhalation) or placebo was constructed using four inhalations (1+1+2)--total cumulative delivered dose of 18 microg formoterol--with the following sequences: (1) formoterol pre-treatment + formoterol 18 microg, (2) formoterol pre-treatment + placebo, (3) placebo pre-treatment + formoterol 18 microg, (4) placebo pre-treatment + placebo. Formoterol 9 microg induced significant (P < 0.0001) bronchodilation at 2 h after inhalation (best mean increase in FEV(1): 0.170 L). Afterwards, dose-dependent increases in FEV(1) occurred with formoterol (maximum mean increase from 2-h value with formoterol: 0.072 after formoterol pre-treatment, and 0.201 L after placebo pre-treatment). Both maximum values of bronchodilation after the last inhalation of formoterol were statistically different (P < 0.001) from 2-h levels. These results show that dose-dependent bronchodilatation of formoterol is maintained despite regular treatment.