0513 Assessment of Sleep Fragmentation in a Large US Sample by Home Under-Mattress Devices

Clete Kushida,Andrew Cotton-Clay,Susan Baron,Laura Fava,Andrew Wray,Venkat Easwar,Arthur Kinsolving,Yue Leng,Jennifer Zitser,Philippe Kahn
DOI: https://doi.org/10.1093/sleep/zsae067.0513
IF: 6.313
2024-04-20
SLEEP
Abstract:Abstract Introduction Sleep fragmentation by arousals and sleep-disordered breathing (SDB) is known to increase with age and differ by gender, although the magnitude of these effects is largely unknown. Population studies allow large collection of continuous nightly data to explore these effects. Methods Sleep fragmentation index (SFI, arousals/hour) and SDB were analyzed by a commercially available home monitoring device (Sleeptracker-AI Monitor, Fullpower Technologies Inc., California, USA) that passively monitors sleep using piezo-electric sensors. Validated sleep/respiratory parameters were derived from device data. De-identified data were analyzed, following review and exemption of the study (#57681) from Stanford University IRB. SFI was assessed using an estimated arousal index, estimating individual arousal events from passive sensor data, with validation against PSG. Data 8/16/2023-12/15/2023 were reviewed. Individuals with birth year and gender provided and age 20-79 years in 2023 were included in the analytic dataset. Results 117,846 individuals (64,049 men, 48.2±12.6 years; 53,797 women, 48.3±12.6 years) with 10,090,778 recorded nights met the inclusion criteria. For AHI< 5, the SFI mean±SE for women and men were 13.6±0.16 and 12.6±0.13 at 20-24 years, respectively, and increased with age to 15.0±0.26 and 16.0±0.24 for 75-79 years. For AHI< 5, men showed a significantly greater SFI during REM sleep than women across all ages above 24 years, but not during NREM sleep. For AHI< 5, SFI significantly increased with age during NREM sleep (women 11.9±0.15 to 14.6±0.17, men 10.8±0.12 to 15.1±0.26), but for REM sleep increased up to 25-34 years (women 18.8±0.12, men 19.6±0.09), but then steadily decreased with age to 75-79 years (women 16.1±0.27, men 18.1±0.24), and a similar trend was found for AHI≥5. For AHI≥5 compared to AHI< 5, SFI was significantly higher by 6-13 arousals/hour across all ages. Conclusion Use of a noninvasive in-home monitoring device enabling continuous nightly collection/analysis of a large sample of sleep/respiratory data revealed differences in sleep fragmentation across age and by gender. Greater sleep fragmentation was observed for men across age during REM sleep, and fragmentation during REM sleep decreased with age after 34 years, which is unexpected and warrants further study given that overall quantities of NREM and REM sleep decrease with age. Support (if any) Fullpower
neurosciences,clinical neurology
What problem does this paper attempt to address?
This summary collection covers 4 different studies, each focusing on different aspects of sleep disorders. Below is a brief overview of the main objectives of these 4 studies: ### 1. Application of High-Frequency Radar Systems in Sleep-Related Breathing Disorders - **Objective**: To evaluate the performance of a non-contact high-frequency radar system in detecting sleep-related breathing disorder (SRBD) events and compare it with the standard polysomnography (PSG) method. - **Methods**: Data were collected using a radar system placed under the bed while participants underwent clinically indicated PSG. Algorithms were developed to quantify the reduction in breathing amplitude collected by the radar and compare it with the time-synchronized signals from respiratory impedance plethysmography belts. A convolutional neural network (CNN) was trained to identify apnea and hypopnea events. - **Conclusion**: The radar system performed excellently in detecting breathing events, showing high accuracy compared to PSG. It has the potential to reduce the number of physical contact sensors used in traditional sleep studies, thereby improving patient comfort and the accuracy of results. ### 2. Automatic Breathing Event Detection (ABED) - **Objective**: To propose a method named ABED for automatically locating and classifying obstructive apnea (OA), central apnea (CA), and hypopnea (HYPO) events. - **Methods**: ABED is based on an end-to-end deep learning architecture, including unified real-time object detection, ResNet feature extraction, and bidirectional long short-term memory layers for temporal processing. PSG data from 4 different cohorts were used for training and evaluation. - **Conclusion**: ABED can accurately predict the apnea-hypopnea index (AHI), showing high consistency with expert scoring and accurately diagnosing sleep apnea and its severity. ### 3. Evaluating Sleep Fragmentation in a Large US Sample Using a Home Under-Mattress Device - **Objective**: To passively monitor sleep fragmentation in a large sample using a commercial home monitoring device (Sleeptracker-AI Monitor) and explore the effects of factors such as age and gender. - **Methods**: Nightly sleep data from over 110,000 participants were analyzed. Sensor data from the under-mattress device were used to estimate the sleep fragmentation index (SFI) and sleep-disordered breathing (SDB). - **Conclusion**: Men experience more sleep fragmentation during REM sleep than women, and this difference increases with age. Sleep fragmentation during REM sleep peaks at ages 25-34 and then gradually decreases. ### 4. Can the Cardiac Component Envelope of Photoplethysmographic (PPG) Signals Reflect Respiratory Drive? - **Objective**: To investigate whether the cardiac component envelope (EC) in fingertip photoplethysmographic (PPG) signals can reflect respiratory drive during obstructive sleep apnea (OSA). - **Methods**: 76 participants suspected of OSA were recruited. PPG signals were recorded using the TipTraQ device while they underwent polysomnography analysis. Spectral analysis was used to extract mutations in EC and instantaneous heart rate (IHR). - **Conclusion**: The study results showed that EC typically exhibits mutations following OSA events, which may reflect sympathetic nervous stimulation induced by OSA preceding hemodynamic changes, indicating that EC encodes respiratory drive information.