Although orally administered theophylline has been prescribed widely in patients with nonreversible airway obstruction, symptomatic benefit has not been established. To assess the effects of orally administered theophylline on dyspnea, we performed a randomized, double-bind, crossover, placebo-theophylline clinical trial in 12 ambulatory male patients with moderate to severe nonreversible airway obstruction. Dyspnea was rated using 2 clinical indexes based on 3 components affecting breathlessness: functional impairment, magnitude of task that evokes dyspnea, and the associated magnitude of effort. Dyspnea and physiologic parameters were measured on 4 occasions: at baseline, 4 wk after initial treatment, a second baseline after a 2-wk washout period, and 4 wk after the second medication. For the 12 patients, mean age (+/- SD) was 60 +/- 7 yr, forced expiratory volume in one second was 1.36 +/- 0.67 L (mean, 40% of predicted), and arterial oxygen tension was 71 +/- 10 mmHg. During the treatment phase, all patients had a therapeutic theophylline blood level (range, 12 to 19 micrograms/ml). Theophylline significantly decreased the components of functional impairment (p = 0.02) and magnitude of task (p = 0.02) relating to dyspnea, as well as the overall dyspnea rating (p = 0.01). There were no significant differences between placebo and theophylline therapy for spirometry, arterial blood gas tensions, and the 12-min walking distance. Thus, sustained-release theophylline significantly reduced dyspnea in these ambulatory patients with moderate to severe nonreversible airway obstruction without altering lung function, gas exchange, or exercise performance.