Abstract MP48: Cross-Sectional and Prospective Associations Between Slow-Wave Sleep and Type 2 Diabetes: the Multi-Ethnic Study of Atherosclerosis

Introduction: Slow-wave sleep (also known as stage N3 sleep), a marker of deep restorative sleep, is implicated in the regulation of insulin resistance and glucose homeostasis and may impact cardiometabolic health. Hypothesis: We conducted both cross-sectional and prospective analyses to test the hypothesis that higher proportion and longer duration of N3 sleep are associated with reduced diabetes risk in older adults. Methods: A sub-sample of participants from the Multi-Ethnic Study of Atherosclerosis completed one-night polysomnography at Exam 5 (2010-2013) and were prospectively followed until Exam 6 (2016-2018). Continuous measures of N3 proportion (%) and duration (minutes) and their corresponding quartiles were used as primary exposures. Diabetes status was assessed in both Exams and was defined as fasting blood glucose≥126 mg/dl, HbA1c≥6.5%, antidiabetic medications use, or self-reported doctor diagnosis of diabetes. We used Modified Poisson regression to examine the cross-sectional associations of N3 proportion and duration with prevalent diabetes in Exam 5. Cox proportional hazards models were used to estimate risk of incident diabetes according to N3 measures. We sequentially adjusted for sociodemographic factors, lifestyles, adiposity measures, and multiple sleep-related factors, such as average sleep duration, apnea-hypopnea index, and insomnia symptoms. Results: In cross-sectional analyses (N=2,026, mean age: 69 years), diabetes prevalence was 28% (n=572). Compared to participants in the first quartile of N3 proportion (<2.0%), participants in the fourth quartile (≥15.4%) were 29% less likely to have prevalent diabetes (Prevalence ratio: 0.71; 95% CI: 0.58,0.87; P -trend: 0.0016]. The association was attenuated and not significant after adjusting for confounders, particularly BMI. In prospective analyses (N=1,251), there were 129 incident diabetes cases over 6,346 person-years of follow-up (mean follow-up time: 5.1 years). A curvilinear relationship was observed between N3 proportion and incident diabetes risk. The hazard ratio (95% CI) of developing diabetes versus the lowest quartile was 0.45 (0.25,0.81) for quartile 2, 0.31 (0.14,0.70) for quartile 3, and 0.30 (0.10,0.90) for quartile 4 ( P -nonlinearity=0.0169, based on restricted cubic spline model). The association remained significant in the fully adjusted model ( P -nonlinearity=0.0319). The results were similar for N3 duration. Conclusions: Higher N3 proportion and longer N3 duration were prospectively associated with lower diabetes risk in a nonlinear fashion among older American adults.