Abstract MP90: Association of Sleep Duration and Quality with Incident Diabetes: Coronary Artery Risk Development in Young Adults Sleep Study
Background: Most prior longitudinal studies investigating the association of sleep characteristics with incident diabetes are based on self-reported sleep duration and quality. However, self-reported sleep is prone to differential error that is based on the amount of actual sleep that a person gets.
Objective: To test whether objectively-measured sleep duration and quality or self-reported sleepiness were associated with the development of diabetes in a population sample.
Methods: Participants from the Chicago, IL center of the CARDIA study who wore wrist actigraphs (ActiwatchTM) twice between 2003-2005 and who were free from diabetes in 2000-2001 were included in the analysis (n=594). The following sleep parameters were averaged over 6 days: sleep duration, sleep maintenance (percentage of sleep period actually sleeping) and minutes of wake after sleep onset (WASO). The Epworth Sleepiness Scale (ESS) assessed daytime sleepiness; higher scores reflect more sleepiness. Diabetes was determined based on fasting glucose >126 mg/dL, medication use, post-load glucose >200 mg/dL or Hemoglobin A1c >6.5% during examinations in 2005-2006 and 2010-2011.
Results: Participants were 40.1 years old (SD=3.6), average sleep duration was 365.8 min (SD=64.3), sleep maintenance was 88.9% (SD=5.4), WASO was 46.0 min (SD=24.0) and the ESS was 7.5 (SD=4.3). Over ~10 years, 68 participants (11.4%) developed diabetes. Each sleep measure was associated with diabetes incidence in unadjusted models; however, only the association with daytime sleepiness remained significant following adjustment for demographic characteristics and diabetes risk factors.
Conclusions: Self-reported daytime sleepiness is a straightforward measurement of sleep quality that can identify adults at risk for diabetes. Although objective measures of sleep duration and quality are associated with incident diabetes, the association is entirely explained by demographic characteristics.
- © 2013 by American Heart Association, Inc.