Patient-controlled analgesia (PCA) has become a standard modality for the management of postoperatively pain, although anecdotal reports of excessive sedation and respiratory depression impugn its safety. To study the prevalence and severity of nocturnal hypoxemia, we measured arterial oxygen saturation (Spo(2)) continuously overnight in 32 postoperative patients who were receiving morphine via PCA. To evaluate the potential benefit of providing concurrent supplemental oxygen, the patients breathed oxygen-enriched air the night of surgery and room air the next night. Patients experienced more pain and consumed twice as much morphine the first night. However, breathing supplemental oxygen that night, the nocturnal mean Spo(2) was 99% +/- 1%, 94% +/- 4% (P < 0.001), and only four patients had periods of hemoglobin desaturation <90%. In contrast, breathing room air the subsequent night, the mean Spo(2) was lower (94% +/- 4%; P < 0.001), and hypoxemia occurred more frequently and was more severe: 18 patients experienced episodes of Spo(2) <90%, 7 patients experienced episodes of Spo(2) <80%, and 3 patients experienced episodes of Spo(2) <70%. One patient required resuscitation for profound bradypnea and cyanosis, but none suffered permanent sequelae. We conclude that when postoperative patients use PCA at night, hypoxemia can be substantial and oxygenation can be improved by providing: supplemental oxygen. Implications: Oxygen saturation was measured postoperatively in patients using morphine patient-controlled analgesia. Substantial nocturnal hypoxemia occurred in half of the patients while they breathed room air. The severity of the hypoxemia was reduced when patients received supplemental oxygen.