An Automated Sleep-Analysis System Operated through a Standard Hospital Monitor

Abstract

Sleep disordered breathing (SDB), a cause of clinically important cardiovascular comorbidity, is often not recognized and diagnosed. An automated system that detects SDB using signals from a standard hospital monitor might provide useful information about the presence and severity of SDB without the need to evaluate the patient in a sleep laboratory and without additional hardware. The aim of this study was to examine the feasibility and accuracy of routine overnight sleep testing for SDB detection by an automated analysis system that operates by analyzing signals derived from standard hospital monitors. Comparison of SDB detection by simultaneous "gold-standard" polysomnography and by Morpheus Hx (WideMed, Ltd., Herzliya, Israel), a bedside computerized analysis system (CAS) connected to a standard hospital monitor (ECG, respiratory impedance, end-tidal carbon dioxide (ETCO2), and SpO2). A total of 53 subjects were examined, 36 men and 17 women, all with suspected SDB. Each subject underwent an overnight sleep study, scored both by polysomnography and by CAS. The study was conducted in Brigham and Women's Hospital, Newton Center, MA. CAS-derived values for apnea-hypopnea index and total sleep time, were each found to be highly correlated with the corresponding polysomnography results, with linear regression values of r = 0.96 and r = 0.82, respectively. Mean apnea-hypopnea index values were also quite similar (CAS of 15.5 +/- 20.0 vs polysomnography of 15.4 +/- 24.0). An automated sleep-analysis system utilizing signals derived from a standard hospital monitor can be considered as a feasible and accurate method to detect and quantify SDB.