Twelve patients with documented obstructive sleep apnea were enrolled in a double-blind placebo controlled crossover trial of oral theophylline, (Uniphyllin) 800 mg, taken at night for four weeks. Overnight polysomnography, using standard techniques, was performed at the end of each treatment period. The total number of apneas (A) and hypopneas (H) decreased significantly while receiving theophylline compared to placebo, from 398 (69), mean (SEM), to 283 (72), p less than 0.01. Sleep quality was, however, significantly worse while receiving theophylline. Obstructive A and H were very much decreased with theophylline (p less than .001), and even when the data were adjusted for the more disturbed sleep with theophylline, this decrease remained significant; the obstructive A and H index fell from 49 (8.7) on placebo to 40 (9) while receiving theophylline, p = 0.02. There was no difference in the numbers of central or mixed A and H, and mean A and H duration was unchanged on the two study nights. Oxygen desaturations greater than 4 percent were less with theophylline treatment (p = 0.02), but mean overnight SaO2 was unchanged. We conclude that theophylline may be beneficial in patients with OSA, but part of the improvement is due to a deterioration in sleep quality.