0296 The Relationship Between Sleep Metrics, Age and Sex in a Large Cohort Study

Abstract Introduction The aims of the study are to investigate the relationship between sleep metrics, age and sex in a large cohort study Methods Wrist acceleration data (GT9x, ActiGraph) was collected from 10,175 subjects in the National Health and Nutrition Examination Survey (NHANES) 2013-2014 for up to 8 days. Data from the tri-axial accelerometer was sampled at 80hz and was processed using the sleep analysis pipeline from ActiGraph. The sleep analysis pipeline first consisted of wear detection, followed by total sleep opportunity detection and then finally sleep / wake classification within the total sleep opportunity. The total sleep opportunity algorithm detects one sleep opportunity per 24-hour cycle (in noon-to-noon windows) and allows for up to 60 minutes of continuous sleep disruption within the sleep opportunity. The following sleep metrics were calculated in each total sleep opportunity using the DACNN sleep / wake classification algorithm: total sleep time (TST), wake after sleep onset (WASO) and sleep efficiency (SE). These sleep metrics were combined with age and gender data from the study to perform a Pearson product correlation (r). Results For females in the study (n=3790), TST had a Pearson r of -0.285, WASO had a Pearson r of 0.149 and SE had a Pearson r of 0.058. For males in the study (n=3529), TST had a Pearson r of -0.300 versus age, WASO had a Pearson r of 0.192 versus age and SE had a Pearson r of -0.005 versus age. Conclusion Sleep measures showed low correlation to age in both the male and female cohorts. This is contrary to the information that sleep quality declines as individual ages. Perhaps the participants in this cohort were healthy enough that decline in sleep metrics was not observed. Support (if any) This research was funded by ActiGraph L.L.C.