Diagnostic Accuracy of Positive Airway Pressure Device for Sleep Apnea Detection in Acute Stroke Patients

Background and Purpose— Sleep apnea has been associated with a poor outcome in acute stroke patients. Polysomnography is the gold standard diagnostic method for sleep apnea, but it is not feasible as a routine in the acute stroke setting. The current generation of positive airway pressure (PAP) devices can detect the different types of respiratory events. This study aimed to compare the algorithms used in PAP device to manually scored events on polysomnography in patients with acute stroke. Methods— A sleep study was performed with standard polysomnography and PAP device, simultaneously, within the first 48 hours after acute stroke onset. Results— We prospectively evaluated 29 patients with acute stroke (59.5±12.1 years). The area under the receiver operating characteristic curve for each apnea-hypopnea index value was above 0.90 by PAP device. There was a good correlation of apnea-hypopnea index (r s =0.92;P <0.001), hypopnea index (r s =0.89;P <0.001), and apnea index (r s =0.70;P <0.001) between device-detected events and manually scored polysomnography.Conclusions— Given the high frequency of sleep apnea during the acute phase of stroke and the complexity of a full polysomnography study in this setting, PAP device on diagnostic mode can be used as an alternative tool for sleep apnea detection in acute stroke patients.