Authors
Summary
Background: Recent studies have shown sleep of a poor quality and short duration to be a risk factor in Type 2 diabetes. The current study of the nationally representative Understanding Society (USoc) cohort aims to distinguish an association between four discrete sleep variables and self-reported diabetes. Methods: Four self-reported sleep measures (sleep duration; sleep onset latency; sleep quality; night time awakening) and self-reported diagnosis of diabetes were analysed from the first wave of USoc using multivariate logistic regression. A directed acyclic graph identified ten covariates (age, sex, ethnicity, BMI, alcohol intake, nutrition, exercise, smoking, highest qualification, profession and technology usage) that were the minimum amount of variables required to adjust for confounding. Results: Comprehensive data was available for 23,214 participants (mean age=46.4; ♂:♀=45.1:54.9). Following adjustment, no significant association between sleep duration or sleep onset latency and diabetes was identified. However, participants reporting “fairly bad” sleep quality had 1.39 (95%CI:1.13,1.70) the odds of having reported diabetes and those reporting night-time awakening as “more than once most nights” had 1.37 (95%CI:1.12,1.68) the odds of having reported diabetes. Discussion: Analysis suggests sleep quality and night-time awakening are a more important risk factor for diabetes than sleep duration. The main limitation of the study is that results are based on cross-sectional analysis of self-reported diagnostic data. Further longitudinal analyses need to be conducted using blood glucose and HbA1c data once these have been released in the next wave of data. Conclusion: Self-reported sleep quality and night time awakening may be better indicators of diabetes risk than sleep duration.
Subject
Link
http://www.cambridgemedicine.org/article/doi/10.7244/cmj-1394708353