Subjective and Objective Measures of Hypersomnolence Demonstrate Divergent Associations with Depression among Participants in the Wisconsin Sleep Cohort Study.

Study Objectives: To examine associations of depression with habitual sleep duration, daytime sleepiness, and objective sleep propensity in a nonclinical population.

Methods: Data from adults participating in the Wisconsin Sleep Cohort Study were utilized in analyses. There were 1,287 adults (3,324 observations) who were used in the analysis of subjective hypersomnolence measures; 1,155 adults (2,981 observations) were used in the analysis of objective sleep propensity assessed by the multiple sleep latency test (MSLT). Repeated-measures logistic regression estimated associations between presence of depression (defined as modified Zung Self-Rating Depression Scale ≥ 50 or use of antidepressant medications) and three primary hypersomnolence measures: subjective excessive daytime sleepiness (Epworth Sleepiness Scale [ESS] ≥ 11), self-reported sleep duration ≥ 9 h/d, and objective sleep propensity (MSLT mean sleep latency < 8 min).

Results: After adjusting for age, sex, body mass index, chronic medical conditions, sedative hypnotic medication use, caffeine, tobacco, and alcohol use, sleep disordered breathing, as well as insomnia and sleep duration when appropriate, estimated odd ratios (95% confidence interval) for depression were: 1.56 (1.31,1.86) for ESS ≥ 11; 2.01 (1.49, 2.72) for habitual sleep time ≥ 9 h; and 0.76 (0.63-0.92) for MSLT mean sleep latency < 8 min.

Conclusions: Our results demonstrate divergent associations between subjective and objective symptoms of hypersomnolence and depression, with subjective sleepiness and excessive sleep duration associated with increased odds of depression, but objective sleep propensity as measured by the MSLT associated with decreased odds of depression. Further research is indicated to explain this paradox and the impact of different hypersomnolence measures on the course of mood disorders.

Commentary: A commentary on this article appears in this issue on page 467.