Background: Cross-sectional and prospective studies have demonstrated that self-reported short sleep is a predictor of cardiometabolic conditions including obesity, Type 2 Diabetes, and cardiovascular events. The relationship may be non-linear, with short and long sleep related to markers of cardiometabolic risk. Research on sleep duration and cardiometabolic health is limited by use of single item self-report sleep measures and homogeneous populations. The current study tested the hypothesis that accelerometer-measured sleep duration would be significantly associated with objective markers of cardiometabolic risk in older adult women.
Methods: Cross-sectional data were analyzed in 2662 women (mean age: 79.05, 52.8% white, 29.8% black, 17.4% Hispanic), from the Objective Physical Activity and Cardiovascular Health Study, ancillary study to the Women’s Health Initiative. Women wore accelerometers on the hip for 24 hours over 7 days and completed a daily sleep log. To be included in the present analysis women must have validated night wear for at least 3 nights. Sleep data were scored according to a standard protocol using sleep logs and visual inspection of the accelerometer data. Body mass index (BMI) and fasting blood samples were obtained at home visits prior to accelerometer monitoring. Adjusting for age and race/ethnicity, linear regression models estimated the relationship between sleep duration and BMI (kg/m2), HDL cholesterol, triglycerides, and glucose (mg/dL each), and C-reactive protein (CRP; mg/L). A quadratic term for sleep was included in the models to evaluate nonlinearity.
Results: The mean nightly sleep duration in the sample was 489.6 mins per night (8.16 hours) with 14.8% of the sample sleeping less than 7 hours per night and 21% sleeping more than 9 hours per night. After adjusting for age and race/ethnicity, sleep duration was significantly related to BMI (regression coefficient [B]:-0.034, p<.01), and the relationship appeared to be non-linear (quadratic, p=.02). The estimated quadratic function indicated a decreasing BMI as sleep duration increased up to 500 minutes/night, and then an increasing BMI as sleep duration continued to increase beyond 500 minutes/night. Additionally, sleep duration was significantly related to CRP (B: -0.005, p=.03), triglycerides (B: 0.002, p=.03) and glucose (B: -0.125, p=.05), with both the CRP (quadratic, p =.02) and glucose (quadratic, p =.03) relationships appearing to be u-shaped and the triglyceride relationship linear. Sleep duration was not significantly related to HDL cholesterol.
Conclusions: In older women, there is a significant relationship between sleep duration and cardiometabolic risk factors, independent of age and race/ethnicity. These cross-sectional relationships should be further explored in prospective studies to inform sleep guidelines for better cardiovascular health in older adult women.