Sleep apnea detection: accuracy of using automated ECG analysis compared to manually scored polysomnography (apnea hypopnea index)

• Published in: Sleep & breathing Vol. 23; no. 1; pp. 125 - 133
• Main Authors: Hilmisson, Hugi, Lange, Neale, Duntley, Stephen P.
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.03.2019; Springer Nature B.V
• Subjects: Adult; Aged; Algorithms; Apnea; Automation; Dentistry; EKG; Electrocardiography; Electrocardiography - methods; Female; Humans; Internal Medicine; Male; Medicine; Medicine & Public Health; Middle Aged; Neurology; Otorhinolaryngology; Pediatrics; Pneumology/Respiratory System; Polysomnography - methods; Retrospective Studies; Sensitivity and Specificity; Sleep; Sleep apnea; Sleep Apnea, Obstructive - diagnosis; Sleep Breathing Physiology and Disorders • Original Article; Sleep disorders; Sleep Stages; Statistical analysis; Apnea Hypopnea Index; Sleep apnea; Cardiopulmonary coupling; Cyclic variation of heart rate
• ISSN: 1520-9512; 1522-1709; 1522-1709
• DOI: 10.1007/s11325-018-1672-0

Introduction Adequate sleep is fundamental to wellness and recovery from illnesses and lack thereof is associated with disease onset and progression resulting in adverse health outcomes. Measuring sleep quality and sleep apnea (SA) at the point of care utilizing data that is already collected is feasible and cost effective, using validated methods to unlock sleep information embedded in the data. The objective of this study is to determine the utility of automated analysis of a stored, robust signal widely collected in hospital and outpatient settings, a single lead electrocardiogram (ECG), using clinically validated algorithms, cardiopulmonary coupling (CPC), to objectively and accurately identify SA. Methods Retrospective analysis of de-identified PSG data with expert level scoring of Apnea Hypopnea Index (AHI) dividing the cohort into severe OSA (AHI > 30), moderate (AHI 15–30), mild (AHI 5–15), and no disease (AHI < 5) was compared with automated CPC analysis of a single lead ECG collected during sleep for each subject. Statistical analysis was used to compare the two methods. Results Sixty-eight ECG recordings were analyzed. CPC identified patients with moderate to severe SA with sensitivity of 100%, specificity of 81%, and agreement of 93%, LR+ (positive likelihood ratio) 5.20, LR− (negative likelihood ratio) 0.00 and kappa 0.85 compared with manual scoring of AHI. Conclusion The automated CPC analysis of stored single lead ECG data often collected during sleep in the clinical setting can accurately identify sleep apnea, providing medically actionable information that can aid clinical decisions.