The assessment of reversibility is recognized as being an essential part of the management of airways obstruction, but testing for reversibility of airways obstruction may not be useful for identifying patients with chronic obstructive pulmonary disease (COPD) who are likely to benefit from bronchodilator treatment. We studied 100 patients with stable COPD. Early reversibility of airways obstruction to 200 mg salbutamol and peak bronchodilation to 50 micrograms salmeterol or 200 micrograms salbutamol were evaluated in four different sessions (two for salbutamol and two for salmeterol), using a double-blind, cross-over, randomized study. Fifteen minutes after inhalation of salbutamol, 47 patients presented an increase in FEV1 greater than 15% of baseline, whereas 48 showed an increase of FEV1 of at least 160 ml from basal value and 33 an increase in FEV1 greater than both 15% of baseline and 200 ml. On the contrary, 69 patients presented an increase in FEV1 greater than 15% of baseline, 65 an increase of FEV1 of at least 160 ml from basal value and 60 an increase in FEV1 greater than both 15% of baseline and 200 ml as maximum increase over baseline usually 2-4 h after inhalation of salmeterol. Twenty-seven other subjects defined as irreversible by the lack of acute improvement in FEV1 after salbutamol, showed an increase in FEV1 greater than 15% of baseline, 20 showed an increase of FEV1 of at least 160 ml from basal value, and 18 showed an increase in FEV1 greater than both 15% of baseline and 200 ml; all had the maximum degree of bronchodilation usually 1-2 h after salbutamol administration. These findings clearly demonstrate that many patients suffering from stable COPD show early reversibility to a short-acting beta 2-agonist and patients who do not manifest early reversibility to salbutamol can still benefit from salbutamol or salmeterol. Therefore, treatments with beta 2-agonists must always be tried.