0325 Snoring and Obstructive Sleep Apnea Associations Through the Lens of a Smart Bed Platform

Abstract Introduction Snoring and obstructive sleep apnea (OSA) are bidirectionally associated. The majority of OSA patients snore and a substantial percentage of those who snore have OSA. The intensity and the acoustic properties of snoring in apnea patients have been shown to correlate with OSA presence. The goal of this research is to use objective sleep data collected by the Sleep Number platform and survey responses to examine the relationship between snoring, sleep metrics, demographics, and OSA. Methods An IRB-approved survey, which included demographics and questions about sleep and its associated disorders, was presented to a cohort of Sleep Number (SN) customers during the period between June 12 to 26, 2023. Objective data collected by the SN platform included sleep duration, restful sleep, sleep latency, sleep quality, mean respiration rate, mean heart rate, and mean heart rate variability. Data were collected May 1st - June 30-2023, and were augmented by survey responses to examine associations between apnea presence, snoring properties, and objective data. Results Out of 22048 (12476F/9537M) respondents with mean age 55.8 (SD: 14.1) years, 3163 reported a diagnosis of apnea and 10558 reported no sleep disorder. Snoring was reported by 89.9% and 65.2% of the apnea and healthy sleepers, respectively. Among the respondents who snored, 29.2% had an apnea diagnosis (apnea and snoring are associated with R2=0.02). A mixed model considering age, gender, reported breathing interruption, snore intensity and frequency as predictors and apnea as the dependent variable resulted in a higher R2 = 0.19. The addition of objective data collected by the smart bed increased the R2 = 0.53. A logistic regression model using demographics, snoring properties, and objective metrics as independent and apnea presence as dependent variables, was trained and tested with the data from 80% and 20% of the individuals respectively. This model resulted in 0.91 area under the receiving operator curve, and 84% sensitivity and 90% specificity to detect apnea. Conclusion Objective data from a smart bed combined with demographic and snore properties can be used to screen apnea risk. This research is limited by the absence of information about apnea severity. Support (if any)