Abstract
Introduction
Lack of physical activity and disturbed sleep have been linked to older adult’s poor cognitive outcomes; however, little is unknown how they interact to affect cognition long-term. The purpose of this study was to examine the association of baseline sleep duration and physical activity (PA) with change in cognition independently and interactively over four years.
Methods
The sample included 1126 community-dwelling older adults aged 60+ (mean age 67.1±5.9 years, 51% female) from the 2011 baseline and 2015 follow-up data of the China Health and Retirement Longitudinal Study (CHARLS). All variables were assessed through interviews. Sleep duration was measured with hours per 30-minute interval and categorized as very-short (<5h), short (5-6.5h), normal (7-8.5h), and long (≥9h). PA was calculated based on PA intensity, duration, and number of days. Cognition was a composite score of mental capacity, episodic memory, and visuospatial abilities. Data were analyzed using multiple regression (primary outcome: change in cognition; main independent variables: baseline sleep, PA, and sleep PA interaction).
Results
At baseline, 19% of participants had very-short sleep duration, 34.4% had short sleep, 39.2% had normal sleep, and 7.2% had long sleep. At follow-up, 57.5% of participants experienced cognitive decline (-3.5±2.5). After controlling for age, gender, education, region, body mass index, smoking, drinking, number of chronic conditions, pain, depression, and cognition at baseline, compared to participants reporting 7-8.5h sleep, those with ≥9h sleep had significantly greater decline in cognition [β=-1.4, 95% CI=2.4, -0.4], while those with <5h sleep [β=-0.5, 95% CI=-1.2, 0.2] and 5-6.5h sleep did not [β=-0.1, 95% CI=-0.7, 0.5]. PA was neither associated with cognitive decline, nor moderated the relationship between sleep duration and cognitive decline.
Conclusion
Long sleep might be a marker of cognitive decline in older adults. Prospective analysis, using objectively measured PA and sleep should be conducted to further examine these associations.
Support
National Institute of Nursing Research R00NR016484