We studied the effect of 30 mg of prednisolone on 29 Japanese patients with chronic obstructive pulmonary disease (COPD). The mean value of the baseline forced expiratory volume in one second (FEV1; mean +/- SEM) was 1.14 +/- 0.12 l (46.9 +/- 3.9% pred) and the FEV1 following the steroid trial was 1.30 +/- 0.12 l (53.7 +/- 4.3% pred). Post-trial FEV1--baseline FEV1/predicted FEV1 was 6.8 +/- 1.9%. Five patients (17%) had more than a 15% increase in FEV1 as a percentage of predicted FEV1. Post-trial FEV1/baseline FEV1 was 117.3 +/- 4.3%, and 12 patients (41%) had more than a 20% increase in FEV1 after the trial. Acute bronchodilator response to beta-agonist correlated positively with the response to corticosteroid. Baseline spirometries, blood eosinophil counts, serum IgE levels, sputum eosinophil counts, family history of asthma, and history of paroxysmal dyspnea did not vary across responders and non-responders. Patients with severe COPD should be treated to achieve the best possible pulmonary functions indicated by a steroid trial within the limit of acceptable levels of adverse effects.