There is randomized controlled evidence that patients with sleep apnea-hypopnea syndrome (SAHS) treated with continuous positive airway pressure (CPAP) have improved daytime function and quality of life. "Before-and-after" data indicate that CPAP improves sleep quality, but there is no randomized controlled evidence of this. We tested the hypothesis that CPAP improved sleep quality in patients with SAHS. We also sought correlations between polysomnographic (PSG) indices and benefit from use of CPAP. Twenty-two patients with SAHS (median, [interquartile range] apnea-hypopnea index, 40 [25-65] events/h) spent 1 mo receiving CPAP and 1 mo receiving placebo capsules, in random order, and home sleep studies were performed at the end of each month. CPAP resulted in a lower arousal index, less Stage 1, and more Stage 3+4 sleep, but there was no increase in Stage REM (rapid eye movement) sleep. Improvement in Epworth score after 6-12 mo of treatment correlated significantly with PSG indices on CPAP: sleep efficiency index (r = 0.78, p < 0.001), Stage REM (r = 0.55, p = 0.001), Stage 3+4 (r = 0.5, p = 0.02), and arousal index (r = -0.43, p = 0.05). This study provides randomized controlled trial evidence that patients with SAHS sleep better when receiving CPAP. Patients with good sleep quality on CPAP at 1 mo are likely to gain later benefit in subjective daytime sleepiness.