(C15) Use of A Wearable Activity Tracker to Assess Differences in REM Sleep Duration Among Adolescents With and Without Depression

Publication Number: 959

Background: An increased percentage of rapid eye movement (REM) sleep at night has been identified as a potential biomarker of depression. Recently, wearable activity trackers have developed features to detect REM sleep in healthy adolescents.

Objectives: To assess whether REM sleep measured using wearable activity trackers differs between adolescents with depression and healthy controls.

Methods: Prospective cohort study of a convenience sample of adolescents 12 to 17 years with or without depression, recruited in 2021-2022. Sociodemographic characteristics, depressive symptom scores (Children's Depression Rating Scale-Revised [CDRS-R]) and PROMIS self-reported sleep quality were assessed at baseline. Adolescents received a Fitbit Inspire HR® activity tracker at baseline and were instructed to wear them 24/7 to capture sleep activity (i.e., daily minutes spent in bed and minutes of REM sleep) over 28 days of follow-up. Depression status was determined based on a CDRS-R score greater than 28 and being on antidepressants. Nighttime sleep was defined as any sleep events which started and ended between 9pm and noon of the next day. Percentage REM sleep was defined as the proportion of total minutes in bed during nighttime which were identified by the tracker as REM sleep. T tests compared mean (SD) percentage REM sleep at baseline, and overall recorded follow-up between adolescents with and without depression. Generalized estimating equation (GEE) was used to compare percentage REM sleep between adolescents with and without depression, adjusted for age, sex, follow-up day, baseline CDRS-R score and PROMIS self-reported sleep quality.

Results: We identified 29 adolescents -19 with depression and 10 controls - with ≥2 nights of Fitbit-recorded total and REM sleep time. Mean age was 15 years, 59% females. On average, there were 13 nights with missing sleep measurements in the group with depression, compared to 8 nights among controls. Mean percentage of REM sleep (SD) at baseline was 17 (5) % among adolescents with depression, relative to controls [14 (5) %], p =0.15. Over an average 17 days of follow-up, mean percentage REM sleep (SD) was 17 (6) % among adolescents with depression, relative to controls [16 (5) %], p =0.34. GEE showed an adjusted 4 percentage-points increase in percentage of REM sleep time among adolescents with depression, relative to controls (4.3, 95% CI [0.03 – 8.61] %).

Conclusions: With expected improvements in the sensitivity of wearable trackers to REM sleep over time, these devices show promise as tools for population-scale surveillance of REM sleep alterations. Such surveillance could be applied in detecting depression or response to antidepressant treatment among adolescents.