Sleep Architecture, Glucose Regulation and Subclinical Cardiovascular Risk in Adults With Type 1 Diabetes Mellitus: A Wearable Sleep Tracking and Continuous Glucose Monitoring Study

Abstract

ABSTRACT Objective To investigate the associations of wearable‐derived sleep architecture with CGM‐derived glucose metrics, HbA1c and markers of subclinical atherosclerosis in adults with Type 1 diabetes mellitus (T1DM) in the CORDIA study. Study Design and Methods One hundred and seventy‐nine adults (79 women) with long‐standing T1DM wore an Apple Watch for 14 nights; glycaemia was assessed from routine 14‐day CGM reports (FreeStyle Libre 2, Dexcom G6/G7 or Guardian 4) and HbA1c. Carotid plaques were assessed by ultrasound and coronary artery calcium (CAC) by computed tomography. Participants also completed the Munich Chronotype Questionnaire (MCTQ). Spearman correlations with false discovery rate (FDR) correction and multivariable linear regression were used. Results Shorter REM sleep correlated with higher HbA1c ( ρ = −0.24; p adj = 0.011) and, with borderline significance after FDR correction, with higher GRI, TAR2, mean sensor glucose and glucose SD and lower TIR (| ρ | 0.17–0.21; p adj = 0.052). REM sleep expressed as a proportion of total sleep correlated with GRI, TIR, mean glucose and glucose SD (| ρ | 0.18–0.20; p adj = 0.044). In adjusted models, REM duration was associated with all glycaemic metrics; the association with glucose SD persisted after additional adjustment for HbA1c and after exclusion of nights with less than 4 h of recorded sleep. After adjustment for age and sex, participants with two or more carotid plaques had 9.4 min shorter REM sleep ( p = 0.034); associations with CAC and deep sleep were attenuated. Older age was associated with lower social jetlag ( ρ = −0.20; p = 0.010). Conclusions In adults with T1DM, shorter REM sleep was modestly associated with worse glycaemia, most consistently with glucose variability and HbA1c, and with carotid plaque burden. These cross‐sectional findings generate hypotheses for longitudinal studies combining sleep tracking with CGM.