Sleep patterns and impaired cognitive function in adults aged ≥ 55 years: a longitudinal study in China

Research on sleep duration and cognitive function has primarily relied on cross-sectional designs or single-time assessments, overlooking the longitudinal patterns of sleep behavior. This study aims to explore the synergistic effects of nighttime sleep and daytime napping on cognitive changes.

We identified distinct long-term sleep patterns using Dynamic Time Warping -based K-medoids clustering analysis of three sleep data sets from 4,385 participants aged 55 and older. We examined the association between these patterns and cognitive decline and AACD-defined cognitive impairment incidence via linear mixed-effects models and time-censored Cox proportional hazards regression analysis.

Three sleep pattern clusters were identified: Cluster 1 (moderate-stable sleep duration, increasing napping), Cluster 2 (long-stable sleep, frequent napping), and Cluster 3 (progressively shorter sleep, infrequent napping). Compared with Cluster 1, both Cluster 2 (β = -0.11, 95% CI: -0.16 to -0.05) and Cluster 3 (β = -0.09, 95% CI: -0.13 to -0.04) exhibited significantly greater cognitive decline. In fully adjusted Cox models, Cluster 3 was modestly associated with a higher likelihood of AACD-defined cognitive impairment (HR = 1.15, 95% CI: 1.02 to 1.23), while Cluster 2 showed no significant association.

Distinct long-term sleep-nap patterns are differentially associated with cognitive outcomes. Shortening sleep duration over time with limited napping may be associated with a higher likelihood of AACD-defined cognitive impairment. These findings highlight the potential importance of monitoring longitudinal sleep behaviors in older adults, although further studies are needed to clarify the underlying mechanisms.