Yoav Nygate, MSc1 • Jan Wodnicki, BSc1 • Sam Rusk, BSc1 • Chris Fernandez, MSc1 • Dennis Hwang2 • Nathaniel F. Watson, MD, MS3
Introduction
Obstructive sleep apnea (OSA) is a prevalent sleep-related breathing disorder. The majority of adults with OSA remain undiagnosed and untreated, highlighting the importance of accurate diagnosis to ensure timely and effective treatment. While single-night polysomnography remains the gold standard for diagnosing OSA, the lack of access coupled with night-to-night variability contributes to underdiagnosis and inconsistent results, compromising diagnostic accuracy and increasing patient management cost.
Methods
A non-randomized all-comers retrospective dataset of 4,527 patients was collected throughout 2024 from clinical home sleep apnea testing using FDA-cleared wearable PPG devices. Each patient had at least three nights of recordings, and only nights with a minimum of four hours of technically adequate PPG signals and at least one hour of sleep were included. The mean number of nights was 3.29 and the maximum number of nights was 16. Furthermore, the dataset had an overall mean AHI of 15 events/hour and a median AHI of 9 events/hour. Night-to-night variability was assessed by determining whether the first night’s results would have resulted in misdiagnosis. Additionally, maximum difference in OSA severity (None, Mild, Moderate, Severe) across all nights was evaluated. The overall pooled standard deviation of multi-night AHI was also calculated to provide a comprehensive assessment of variability in OSA severity.
Results
When comparing the first night to subsequent nights, 11.8% of OSA patients would have been missed, and 26.4% of patients experienced an increase in OSA severity by at least one severity category. When assessing overall variability across all nights, 50% of patients showed no change in OSA severity, 46% experienced an increase of one severity category, and 4% showed an increase of two severity categories. Additionally, the dataset revealed a pooled AHI standard deviation of 5.8 events per hour.
Conclusion
Multi-night sleep testing revealed a level of variability that could negatively affect diagnostic accuracy when conducting single-night sleep studies. Multi-night sleep testing can mitigate this issue by aggregating the sleep quality and respiratory disturbance information across multiple nights, providing a more comprehensive assessment of a patient’s condition, and reducing the risk of misdiagnosis by ensuring that treatment decisions are based on a fuller understanding of OSA severity.
1 EnsoData Research, Ensodata, Madison, WI, USA | 2 Kaiser Permanente, Pleasanton, CA 94588 | 3 Department of Neurology, University of Washington School of Medicine, Seattle, WA
