Nasal continuous positive airway pressure (CPAP) has become the nonsurgical treatment of choice for obstructive sleep apnea syndrome (OSAS). Recent evidence suggests that intermittent use of CPAP by patients is more common than nightly compliance. To determine the consequences of intermittent CPAP use, in terms of a return of sleep-disordered breathing and daytime hypersomnolence, 15 OSAS subjects were evaluated at three times: (1) before CPAP treatment (pretreatment), (2) after 30 to 237 days posttreatment during a night of CPAP use (on CPAP), and (3) during a night without CPAP (off CPAP). Evaluations of sleep-disordered breathing and three domains of hypersomnolence, physiologic sleep tendency, subjective sleepiness, and performance, were accomplished with the respiratory disturbance index (RDI), multiple sleep latency test (MSLT), Stanford sleepiness scale (SSS), and psychomotor vigilance task (PVT), respectively. CPAP use was encouraged and monitored from pretreatment to post-treatment by daily diaries for most subjects and an electronic device for a subset of subjects. As expected, CPAP eliminated apneas and hypopneas, and following the on CPAP night, there were statistically significant improvements in objective measures of sleepiness (MSLT and PVT). Subjective measures of sleepiness and fatigue also showed improvement. Sleeping without CPAP for one night reversed virtually all of the sleep and daytime alertness gains derived from sleeping with CPAP. This occurred despite a statistically significant reduction in the RDI on the night off CPAP (M = 36.8, SD = 28.0 events/h) relative to the pretreatment night (M = 56.6, SD = 24.8 events/h), which may be due to a lessening of the edema of the upper airway following CPAP use.(ABSTRACT TRUNCATED AT 250 WORDS)