We studied ventilatory, hemodynamic, and subjective responses to different plasma theophylline concentrations in 10 patients with "irreversible" airflow obstruction. Subjects received theophylline at doses that produced low (9.0 to 12.5 micrograms/mL) and high (17 to 22 micrograms/mL) peak plasma concentrations; subjects also received placebo. A significant (P less than 0.05) dose-related difference in pulmonary function was observed between each treatment. The mean maximal increase in forced expiratory volume at 1 second over placebo was 21.3% for high-dose theophylline and 6.0% for low dose. Both treatments were well tolerated with respect to hemodynamic changes and other adverse effects. Despite improved findings in pulmonary function tests patients were unable to distinguish either treatment from placebo in terms of improvement in breathlessness.