Sleep duration, insomnia and cognitive performance in the Elsa-Brasil cohort: a cross-sectional analysis.

Objective: To investigate the single and combined associations between sleep disturbances (sleep duration, insomnia symptoms in the last 30 nights, and daytime tiredness) and performance in cognitive tests.

Methods: Cross-sectional analysis of data from visit 2 (2012-2014) of the Longitudinal Study of Adult Health from a cohort of active and retired civil servants from six Brazilian capitals. Polynomial regression with quadratic term and multiple linear regression models were performed to assess single and combined associations between sleep disturbances and memory performance, fluency, executive functions, and global cognition.

Results: A total of 7,248 participants were included, with a mean age of 62.7 years (standard deviation [SD]=5.9), and 55.2% were women. Inverted U-shaped associations were observed between sleep duration and performance on all cognitive abilities, suggesting that durations shorter or longer than seven hours are associated with worse performance, regardless of age. Reported insomnia was associated with worse executive function (β: -0.08; 95% confidence interval [CI]: -0.15 to -0.01), and the magnitudes of associations were higher for individuals with insomnia at two or more moments (β: -0.12; 95%CI -0.19 to -0.05) or, especially, insomnia combined with short sleep (β: -0.18; 95%CI -0.24 to -0.11). Insomnia in two or more periods was also associated with lower memory and global cognition. There was no association between any sleep disturbance tested and verbal fluency. Isolated daytime tiredness was not associated with performance in the evaluated tests.

Conclusion: The results suggest that extreme sleep durations are detrimental to almost all cognitive abilities investigated, whereas insomnia appears to affect more severely the executive function.