1018 Daily Associations Between Sleep and Glucose: Evidence from Actigraphy and Continuous Glucose Monitoring

Abstract Introduction Sleep duration and sleep continuity disturbance have been linked to impaired glucose regulation, but most evidence comes from laboratory-based or cross-sectional studies, limiting the ability to evaluate day-to-day associations between sleep and glycemic outcomes. This study tested whether nightly fluctuations in total sleep time (TST) and total wake time (TWT) predict next-day glycemic outcomes (i.e., glucose immediately after waking and average glucose during waking hours). We hypothesized that shorter TST and greater TWT would predict higher glucose levels the following day. Methods Seventy-seven adult participants with insomnia (Mage = 39.2 years, SD = 11.9; 61.8% women; 73.7% White) wore an actigraphy watch and continuous glucose monitor for 10 consecutive days. Sleep variables included nightly TST and TWT derived from actigraphy. Glucose variables included fasting glucose (defined as the first glucose value recorded upon awakening) and average daytime glucose (from wake time to bedtime) derived from continuous glucose monitoring data sampled at 5-minute increments. Linear mixed-effects models accounted for repeated daily observations nested within individual participants. Separate models tested whether TST and TWT predicted the following day’s glucose outcomes. Results TST significantly predicted next-day waking glucose, such that nights with shorter sleep duration were associated with higher next-morning blood glucose (F(1, 552) = 4.89, p = .028). By contrast, TST did not significantly predict next-day average glucose. TWT was also examined as a predictor in additional models but did not significantly predict either glucose outcome. Conclusion Night-to-night variations in sleep duration were associated with next-morning glycemic levels, such that shorter sleep predicted elevated glucose immediately upon waking. In contrast, sleep duration did not predict average glucose across the subsequent day. These findings suggest that glucose levels upon awakening may be particularly sensitive to the prior night’s sleep. Overall, the results highlight sleep duration as a modifiable behavioral factor with implications for metabolic functioning and underscore the value of multimodal, real-world monitoring for capturing dynamic sleep–glycemia relationships. Additional analyses are underway to evaluate within-subject effects, including whether person-centered fluctuations in sleep predict corresponding fluctuations in glucose. Support (if any) SRS Foundation and AASM Foundation Small Research Grant