Associations of actigraph sleep characteristics with blood pressure among older adults

Abstract

Objectives

Many studies have examined links between sleep and blood pressure, with mixed findings, mostly using self-reported sleep data and cross-sectional designs. We examined whether actigraph-estimated sleep characteristics are associated with concurrent blood pressure or 5-year blood pressure change in a national cohort of older adults (National Social Life, Health and Aging Project), and whether these associations differ by hypertension medication use.

Methods

Subjects were 669 older adults (62-90 years), 471 with 5-year follow-up data. Sleep characteristics were duration (linear plus quadratic terms); sleep percentage; and categorical onset, midpoint, and waking times. Multivariable linear models adjusted for age, race, gender, obesity, smoking, daytime napping, and hypertension medication use. Interactions between sleep characteristics and hypertension medication were tested among the 401 subjects with consistent hypertension medication status over time.

Results

We found U-shaped cross-sectional and longitudinal relationships between duration and blood pressure, with shorter and longer sleep times associated with higher blood pressure. Later onset times were cross-sectionally associated with higher systolic blood pressure, while earlier onset times were longitudinally associated with systolic blood pressure increase. Midpoint, wake time, and sleep percentage were not significantly associated with blood pressure. Significant interaction terms suggested hypertension medications attenuated associations of sleep onset and wake time with diastolic blood pressure.

Conclusions

These results with actigraph-estimated parameters confirm some, but not all, associations reported from research based on self-reported sleep data. Our findings are consistent with recommended intermediate sleep durations for cardiovascular health and suggest hypertension medication use may attenuate some associations between sleep timing and blood pressure.