Associations Between Sleep Habits and Dysglycemia in Adults in the United States: A Cross-Sectional Analysis.

Objectives: To examine the independent and joint associations between sleep duration and quality with glycated hemoglobin (A1C) levels and dysglycemia in non-institutionalized adults living in the United States.

Methods: Data from the United States National Health and Nutrition Examination Survey (2005-2008) were used (N=9478; ≥20 years). Information on sleep quantity and quality were derived from the Sleep Disorders Questionnaire and used to classify sleep quality as good, fair, poor, or very poor.

Results: Overall, sleep quantity and quality were related to A1C levels in our unadjusted models. In general, a U-shaped relationship between sleep quantity and A1C levels was observed. Compared to those who slept for 7 to 8 hours per night, sleeping for 4 hours or fewer was associated with higher A1C levels (mean, 95% CI; 5.49%, 5.45 to 5.53 vs. 5.69%, 5.60 to 5.77; p<0.05), whereas only those reporting good and very poor sleep quality had higher A1C levels than poor sleepers (mean, 95% CI: 5.63%, 5.57 to 5.69; 5.56%, 5.52 to 5.60 vs. 5.46%, 5.42 to 5.50; p<0.05). The relationships among sleep duration and quality and the joint effects of sleep quality and quantity and dysglycemia were not significant after multivariable adjustment.

Conclusions: Between 7 and 8 hours of sleep and fair/poor sleep quality were associated with optimal A1C levels, while sleeping for fewer or more hours appeared to increase dysglycemia, without adjustment for covariates. These relationships were attenuated following multivariable adjustment. Future research is necessary to refine our understanding of the sleep/glycemic-control relationship to provide a context for the clinical significance of these findings for longer-term A1C control in adults with diabetes.