Conventional manually adjusted continuous positive airway pressure (CPAP) is an effective therapy for sleep-disordered breathing. We prospectively investigated the efficacy of a self-titrating nasal CPAP system in the acute treatment of obstructive sleep apnea (OSA) syndrome. Twenty patients with moderately severe OSA [apnea hypopnea index (AHI) > 15/hour] were enrolled in a randomized, controlled, prospective clinical trial. An initial diagnostic sleep study was performed, followed by randomization to a manually adjusted CPAP titration on one night and self titrating CPAP on the other night. On the conventional CPAP night, the CPAP was manually adjusted until abolition of all apneas and electroencephalographic (EEG) arousals, whereas the self-titrating CPAP was set in automatic mode at lights out. The self-titrating CPAP system utilized an algorithm based on airway vibration patterns to detect airway stability. The AHI decreased from 50.8 +/- 28.8/hour [mean +/- standard deviation (SD)] at baseline to 3.8 +/- 3.1/hour (p < 0.005) during manually adjusted and 6.1 +/- 5.3/hour (p < 0.005) during self-titrating CPAP. The arousal index (Ar-I) decreased from 34.1 +/- 23.1/hour (baseline) to 11.2 +/- 5.0/hour on manual adjustment (p < 0.005) and 11.3 +/- 0.3/hour on self titration (p < 0.005), whereas total sleep time was unchanged. No significant differences in any measure of oxygenation or sleep architecture were observed between the manually adjusted and self-titrating CPAP nights except that the lowest arterial oxygen saturation (SaO2) was higher with manual titration (84.4 +/- 4.2% vs. 79.9 +/- 9.7%, p < 0.05). The maximum pressure required for abolition of apneas and arousals was significantly lower (p < 0.05) during the self-titrating study (10.1 +/- 3.8 cmH2O) as compared to manually adjusted CPAP (12.3 +/- 3.9 cmH2O). Failure to increase pressure and failure to maintain minimum pressure occurred in 7 of the 20 subjects during the self-titrating study. This required manual resetting of the system in five subjects, but the system self-corrected in two subjects. An unsupervised study would have resulted in undertreatment of OSA. Based on a single-night laboratory study, self-titrating CPAP was well tolerated and improved OSA and sleep architecture comparable to manually adjusted CPAP. The future modifications of this prototype will require further research to assess its efficacy and safety in the laboratory and home environments before its recommendation for general long-term use.