0780 WEARABLE SLEEP EPIDEMIOLOGY IN THE FRAMINGHAM HEART STUDY

• Published in: Sleep (New York, N.Y.) N.Y.), 2017-04, Vol.40 (suppl_1), p.A289-A289
• Main Authors: Heckman, EJ, Salazar, R, Hardy, S, Manders, E, Liu, Y, Au, R, O’Connor, G, Thomas, R
• Format: Article
• Language: English
• Subjects: Electrocardiography; Epidemiology; Oxygen saturation; Sleep
• ISSN: 0161-8105; 1550-9109
• DOI: 10.1093/sleepj/zsx050.779

Abstract Introduction: Wearable devices for sleep assessments offer a cost-effective and convenient alternative to traditional measures of sleep. Devices are now available to measure oxygenation, respiration electrocardiogram, and electroencephalogram in the home environment. This study assessed standard (oximetry) and novel (cardiopulmonary coupling) measures of sleep state in a well-established epidemiology cohort. Methods: Data were collected from 846 participants of the Framingham Heart Study’s second generation and Omni cohort (mean age: 67.9; 484 female). Sleep studies were mailed to each participant, with a single-lead ECG device manufactured by MyCardio, LLC (www.sleepimage.com), an oximetry device manufactured by Nonin, and a brochure to direct the application of the devices. The FDA approved M1 measures electrocardiogram, body position, trunk activity, and snoring. The analysis uses cardiopulmonary coupling to generate the following measures: high and low frequency coupling (HFC and LFC, stable and unstable NREM, respectively), and a biomarker of high loop gain (narrow-band elevated low frequency coupling). The mean, standard deviation, and intra-class correlation coefficient (ICC) were calculated for HFC, LFC, oxygen desaturation index (ODI), and time with oxygen saturation below 90%. Results: A total of 972 participants agreed to participate. 126 participants were unable or refused to complete the study. 830 and 836 participants obtained at least 4 hours of data with the M1 and oximetry device for at least one night, respectively. 574 participants wore both devices for 2 consecutive nights (803 wore M1, 695 wore Ox for 2 consecutive nights). The mean (SD) were as follows: HFC 43.5%(18.8), LFC 37.28%(17.03), ODI 8.3(8.5), oxygen saturation below 90% 48.1(77.24) minutes, and 52.5% of the sample had narrow band coupling. The ICC for these variables ranged from 74.5%-99.9%, suggesting high night to night data and physiological signal stability. Associations with common medical co-morbidities will be presented. Conclusion: The results suggest that home/wearable assessment of sleep is 1) feasible, cost-effective, and yields reliable results; 2) inter-individual differences are stable; 3) measures can be readily repeated; 4) in-person visits are not required, markedly simplifying data collection. Both standard and novel measures can be collected. Support (If Any): NHLBI N01 HC 25195, BIDMC Chief Academic Officer’s Innovation Grant.