To investigate the effect of theophylline on sleep and sleep-disordered breathing in patients with chronic obstructive pulmonary disease (COPD), we studied 12 male nonhypercapnic subjects with a mean +/- SEM age of 62.8 +/- 2.5 yr and a FEV1 of 1.36 +/- 0.11 L using a randomized double-blind crossover protocol. Sustained-action theophylline (250 mg three times or four times a day) or placebo was administered for 2 days, and the alternate drug was administered on the following 2 days. Sleep studies were performed on Nights 2 and 4 with spirometry at 9:00P.M. and 7:00A.M. Two puffs of metaproterenol or albuterol were administered at 10:00P.M. on both study nights. A theophylline level, drawn at bedtime (10:00 to 11:00P.M.), was 14.2 +/- 0.78 micrograms/ml on the theophylline nights and less than 2 on placebo nights. The morning FEV1 was significantly better during theophylline administration (1.27 +/- 0.12 versus 1.00 +/- 0.11 L, p less than 0.001). The mean arterial oxygen saturation (SaO2) and transcutaneous carbon dioxide pressure (PCO2) were also better during NREM sleep on theophylline nights. Neither the mean SaO2 and transcutaneous PCO2 during REM sleep nor the apnea plus hypopnea index (events per hour of sleep) differed between placebo and theophylline nights. Theophylline administration did not impair the amount or architecture of sleep as neither total sleep time nor the fraction of time spent in Stages 1, 2, and 3/4 and REM differed between the two regimens. The number of arousals per hour of sleep was slightly less on theophylline nights (19.9 +/- 1.7 versus 24.9 +/- 2.7, p less than 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)