In contrast to methacholine, a stimulus that induces airway constriction mainly by "direct" stimulation of airway smooth muscle cells, AMP airway responsiveness reflects "indirectly" induced airway narrowing via inflammatory or neural reflex mechanisms. In order to determine inflammatory contribution to airway narrowing in COPD, we performed AMP and methacholine inhalation provocation tests in nonatopic subjects with COPD and compared the results with those obtained from atopic nonsmoking asthmatics and from healthy smoking volunteers. AMP caused airway narrowing in all but two subjects with COPD and in only three of the 12 healthy smoking subjects. Patients with COPD were significantly more responsive to AMP and methacholine than were healthy smoking volunteers. Geometric mean PC20 AMP was significantly lower in the smokers with COPD (7.2 mg/ml) than in the nonsmokers with COPD (58.5 mg/ml), whereas PC20 methacholine values and baseline FEV1 were comparable. In the nonatopic nonsmoking subjects with COPD, PC20 AMP was significantly higher than in the atopic nonsmoking asthmatics (3.8 mg/ml), whereas they responded similar to methacholine provocation. These results indicate that most subjects with COPD respond to AMP provocation and that smoking determines the degree of airway responsiveness to AMP in COPD. We suggest that increased susceptibility to mediator release by mast cells or neural reflex mechanisms are involved in AMP-induced airway constriction in asthma and in COPD.