Objective: Research has suggested that a higher risk of type 2 diabetes associated with sleep characteristics exists. However, studies have not thoroughly assessed the potential confounding effects of mental health conditions associated with alterations in sleep.
Research design and methods: We prospectively assessed the association between sleep characteristics and self-reported incident diabetes among Millennium Cohort Study participants prospectively followed over a 6-year time period. Surveys are administered approximately every 3 years and collect self-reported data on demographics, height, weight, lifestyle, features of military service, sleep, clinician-diagnosed diabetes, and mental health conditions assessed by the PRIME-MD Patient Health Questionnaire and the PTSD Checklist-Civilian Version. Statistical methods for longitudinal data were used for data analysis.
Results: We studied 47,093 participants (mean 34.9 years of age; mean BMI 26.0 kg/m2; 25.6% female). During 6 years of follow-up, 871 incident diabetes cases occurred (annual incidence 3.6/1,000 person-years). In univariate analyses, incident diabetes was significantly more likely among participants with self-reported trouble sleeping, sleep duration<6 h, and sleep apnea. Participants reporting incident diabetes were also significantly older, of nonwhite race, of higher BMI, less likely to have been deployed, and more likely to have reported baseline symptoms of panic, anxiety, posttraumatic stress disorder, and depression. After adjusting for covariates, trouble sleeping (odds ratio 1.21 [95% CI 1.03-1.42]) and sleep apnea (1.78 [1.39-2.28]) were significantly and independently related to incident diabetes.
Conclusions: Trouble sleeping and sleep apnea predict diabetes risk independent of mental health conditions and other diabetes risk factors.