Examination of Longitudinal Associations Between Actigraphic Sleep and Ambulatory Blood Pressure in a Community Sample: Three Analysis Comparisons

Abstract

Objective: Emerging evidence suggests that various dimensions of sleep health are uniquely associated with cardiovascular health, including through sleep effects on acute blood pressure (BP). Prior work, including work from our lab (Doyle et al., 2019) demonstrates that a poor night’s sleep is associated with higher concurrent nighttime BP and elevations the next day. If experienced chronically, this biobehavioral relationship may be a key pathway in the link between sleep and CV health. The current study sought to better establish this pathway to disease by clarifying and extending methodological approaches to conceptualizing BP, determining the reliability of this relationship through replication, and evaluating its chronicity over time. Methods: The current study sought to address these evidence gaps through analysis of data from in 300 men and women aged 21 to 70 years at baseline and 237 at a 2-year follow-up enrolled in The North Texas Heart Study (NTHS), which included ecologically valid assessment of sleep and blood pressure. Actigraphy-assessed sleep was monitored for two consecutive nights, while ambulatory BP was sampled randomly within 45 minute blocks on the first and second day as well as the second night. This protocol was repeated at the second time point, two years later. Results: The previously reported relationship between sleep and daytime BP largely replicated for sleep efficiency, but not for sleep duration, at Time 2 (unadjusted (systolic: (F(1,190)=7.44, p=.006, R2=0.033), diastolic: (F(1, 190)=5.88, p=0.016, R2=0.025); partial (systolic: (F(6, 180)=4.56, p=0.010, R2=0.10), diastolic: (F(6, 181)=2.83, p=0.025, R2=0.055), fully-adjusted (systolic: (F(8, 150)=8.46, p=0.021, R2=0.27)., diastolic: (F(8, 150)=5.13, p=0.049, R2=0.17). The temporal association between lower sleep efficiency one night and higher BP the next day was a particularly stable trend across time points, BP conceptualizations, and models. The results for sleep and nighttime BP at Time 1 (Doyle et al., 2019) largely did not replicate for sleep duration and sleep efficiency at Time 2, except for diastolic BP before full adjustment, indicating a small but meaningful difference between sleeping and nighttime BP readings. Longitudinally, Time 1 mean sleep duration and efficiency did not significantly predict change in mean BP from Time 1 to Time 2. This remained true across the three BP conceptualizations. Conclusion: The previously reported cross-sectional associations between sleep efficiency and mean/temporal daytime ABP were largely replicated at follow-up. There was no meaningful difference between “waking” and “daytime” BP values, but there were small but meaningful differences between “sleeping” BP and “nighttime” BP. Sleep did not predict change in BP over a two-year time span. These results support the hypothesis that sleep impacts BP, clarify conceptual use of nocturnal BP terms, and warrant further investigation of this potential pathway to disease.