Objective: Worsening of sleep quality during menopause is well recognized. However, the underlying hormonal regulation is insufficiently described. In this study, we evaluated associations between sleep and cortisol levels.

Study design: Seventeen perimenopausal and 18 postmenopausal women were enrolled in a three-night sleep study. Diurnal blood sampling was performed during the third night and the following day.

Main outcome measures: Self-reported insomnia and sleepiness were evaluated with the Basic Nordic Sleep Questionnaire and sleep architecture with all-night polysomnography. Diurnal cortisol samples were collected at 20-min intervals. Correlation analyses and generalized linear models adjusted by age, body mass index, vasomotor symptoms and depressive symptoms were conducted.

Results: In correlation analyses, self-reported insomnia and sleepiness were not associated with cortisol levels. Lower sleep efficiency, slow-wave sleep and stage 1 percentages, number of slow-wave sleep and of rapid-eye-movement (REM) periods, longer slow-wave sleep latency and higher wake after sleep onset percentage were associated with higher cortisol levels (all p < 0.05). Further, lower slow-wave sleep percentage and longer slow-wave sleep latency correlated with steeper daytime cortisol slope (i.e. day cortisol decrease, both p < 0.05). In adjusted generalized linear models, lower sleep efficiency and number of rapid-eye-movement periods as well as higher wake after sleep onset percentage correlated with higher cortisol levels; lower slow-wave sleep percentage correlated with higher cortisol awakening response.

Conclusions: Worse sleep architecture but not worse self-reported insomnia and sleepiness was associated with higher cortisol levels. This is important for understanding sleep in women, especially during the menopausal period.