We studied 16 patients with stable COPD in a double blind, double dummy, placebo-controlled, within patient study to see if formoterol could be used as a rescue drug. We compared the of onset of bronchodilation obtained with formoterol 12 microg (metered dose corresponding to 9 microg delivered dose) and formoterol 24 microg (metered dose corresponding to 18 microg delivered dose), both delivered via Turbuhaler, with that of salbutamol 400 microg and salbutamol 800 microg delivered via pressurized metered-dose inhaler (pMDI). Patients inhaled single doses of placebo, formoterol and salbutamol on five separate days. FEV1 was measured in baseline condition and 3, 6, 9, 12, 15, 18, 21, 24, 30, 40, 50, and 60 min after inhalation of each treatment. We examined two separate criteria for deciding if a response was greater than that expected by a random variation of the measurement: (1) a rise in FEV1 of at least 15% from the baseline value; (2) an absolute increase in FEV1 of at least 200 ml. Formoterol 12 microg (15.2 min; 95% CI 9.5-21.0) and formoterol 24 microg (15.1 min; 95% CI 8.9-21.2) caused a rise in FEV1 of at least 15% from the baseline value almost rapidly as salbutamol 400 microg (13.6 min; 95% CI 7.1-20.1) and salbutamol 800 microg (14.5 min; 95% CI 7.1-21.9). No significant difference (P=0.982) in onset of action was seen between the four active treatments. According to Criterion 2, the mean time to 200 ml increase in FEV1 was 11.1 min (95% CI: 7.0-15.2) after salbutamol 400 microg, 13.0 min (95% CI: 7.9-18.1) after salbutamol 800 microg, 14.7 min (95% CI: 7.1-22.4) after formoterol 12 microg, and 12.7 min (95% CI: 7.4-18.0) after formoterol 24 microg. Again, there was no significant difference (P= 0.817) between the four active treatments. Formoterol Turbuhaler 12 microg and 24 microg caused bronchodilation as rapidly as salbutamol 400 microg and 800 microg given via pMDI.