Nasal CPAP is presently accepted as first-line therapy for obstructive sleep apnea, but a significant minority of patients do not tolerate nasal CPAP. The purpose of this study was to compare the benefits of nasal CPAP, nasal oxygen (O2), and placebo (air) using patients as their own controls. We studied eight men, aged 33 to 72 (mean 57 years), who had mild obstructive sleep apnea. To be eligible for study, patients had to have an apnea plus hypopnea index greater than or equal to 5, plus one or more of the following: blood pressure greater than 150/95 mm Hg, multiple sleep latency test mean score less than or equal to 10 minutes, or significant nocturnal cardiac ectopy. After a baseline study, patients received a month each of nocturnal O2 at 4 LPM and air at 4 LPM, presented in random order. The third month of treatment consisted of nasal CPAP (range 2.5 to 12.5 cm H2O). Patients underwent evaluation at baseline and after each month of treatment. It was concluded that oxygen was more effective in improving oxygenation and hypopneas than is nasal CPAP. However, oxygen did not reduce apneas or improve daytime hypersomnolence as well as nasal CPAP in patients with mild OSA. Oxygen might be considered as an alternate form of treatment for patients who are not hypersomnolent, or as an adjunct to nasal CPAP.