Study objectives: Sleep duration and difficulties have been shown to associate with cognitive function. This study examined how changes in sleep and in cognitive function are associated during retirement transition.

Methods: The study population consisted of 2,980 Whitehall II study participants, who retired during the follow-up, whose sleep was queried, and cognitive function measured (inductive reasoning and verbal memory) before and after retirement (follow-up 16 years). Using the last information on sleep before and the first after retirement, participants were categorized into constantly without (59%), increasing (13%), decreasing (11%), and constantly with (18%) sleep difficulties; and constantly short (26%), increasing (19%), decreasing (8.5%), and constantly mid-range (47%) sleep duration. Change in cognitive function during retirement transition was examined by sleep change groups using linear regression analyses with generalized estimating equations.

Results: More pronounced decline in inductive reasoning during retirement transition was observed among participants with increasing sleep difficulties (-1.96, 95%CI -2.52 to -1.41) compared to those constantly without sleep difficulties (-1.25, 95%CI -1.52 to -0.98) and constantly with sleep difficulties (-1.26, 95%CI -1.75 to -0.92). Decreasing sleep difficulties (-0.64, 95%CI -0.86 to -0.43) were associated with a more pronounced decline in verbal memory when compared to constantly without sleep difficulties (-0.42, 95%CI -0.52 to -0.32) in post-retirement period. No statistically significant differences across sleep duration groups in cognitive function were observed.

Conclusion: Increasing and decreasing sleep difficulties may be associated with accelerated decline in cognitive function during retirement transition and post-retirement.