Abstract TMP88: High Accuracy of Auto-CPAP for Obstructive Sleep Apnea in Acute Stroke Patients

Abstract

Background: Obstructive sleep apnea (OSA) is very common in acute stroke patients and has been associated with a poor short-term and long-term outcome. Polysomnography (PSG) is the gold standard diagnostic method for OSA, but it is not feasible as a routine for all acute stroke patients. Therefore, it is essential to validade alternative methods to accurately diagnose OSA in acute stroke patients. The newest generation of automatic continuous positive airway pressure (auto-CPAP) devices innovates by the possibility of detecting the different types of respiratory events. However, the accuracy of auto-CPAP devices for OSA diagnosis has never been tested in comparison to PSG in the acute stroke setting. The main objective of this study was to evaluate the accuracy of an auto-CPAP device to diagnose OSA and to validate its algorithm of respiratory events detection to diagnose OSA in patients during the acute phase of stroke. Methods: A sleep study was performed with PSG and auto-CPAP device, simultaneously, within the first 48 hours after the acute stroke onset. The Receiver Operating Characteristic Curve (ROC), C-statistics, Spearman correlation coefficient, and intraclass correlation coefficient were analyzed. Results: We prospectively evaluated 31 adult patients with acute stroke. The mean age was 59.7 ± 12 years and 60% were males. All patients used auto-CPAP for longer than 4 hours. The PSG revealed an apnea-hypopnea index (AHI) mean of 34 ± 41 events/h, and the auto-CPAP showed an AHI of 18 ± 16 events/h. The area under the ROC curve for OSA diagnosis by the auto-CPAP was above 0.90, with sensitivity and specificity above 80% for each AHI value. The Spearman correlation coefficients (rs) of the AHI, of the hypopnea index, of the obstructive apnea index and of the central apnea index were 0.92; 0.89; 0.63 and 0.62, respectively. The intraclass correlation coefficients between device-detected and PSG manually scored events were 0.60 for AHI, 0.64 for hypopnea index, and 0.45 for apnea index. Conclusion: The auto-CPAP showed an excellent accuracy for the diagnosis of OSA and it was well tolerated by acute stroke patients. Our results suggest that auto-CPAP should be considered as the preferred diagnostic tool for OSA diagnosis during the acute phase of stroke.