We studied the effect of inhaled metaproterenol on exercise performance in ten patients with moderate to severe nonreversible chronic obstructive pulmonary disease (COPD), defined as a FEV1/FVC ratio below 60%, and less than 15% improvement in FEV1 and FVC after both one-time administration of inhaled metaproterenol and a 10-day therapeutic trial with oral theophylline, inhaled metaproterenol, and oral prednisone. After baseline evaluation, on two separate days all selected patients had spirometry and a self-paced 12-min treadmill walking test performed (in double-blind crossover fashion) before and after random administration of five puffs of inhaled metaproterenol or placebo. A few weeks later, testing was repeated single-blind before and after inhaled metaproterenol, but this time the 12-min walk was done in an enclosed hall. Pretreatment spirometry values were similar on all study days, and none of the subjects had a significant change (greater than 15% from baseline) in FEV1 or FVC after the administration of placebo or metaproterenol. The postmetaproterenol treadmill and hall walking distances improved by a mean (+/- SD) of 112 +/- 56 m (p less than 0.001) and 82 +/- 46 m (p less than 0.01), respectively. Individual improvements in exercise performance did not correlate with the corresponding changes in FEV1 or FVC, or with the baseline DLCO measurement. We conclude that an objective improvement in physical performance after administration of a relatively high dose of inhaled metaproterenol can be seen in some patients with moderate to severe "fixed" airway obstruction, despite a lack of significant response as judged by conventional spirometry criteria.