Bidirectional Relationships of Daily Physical Activity and Sleep Among Patients with Heart Failure and Insomnia

Abstract

Physical activity is associated with better sleep quality across a wide range of populations, but the causal direction of this relationship is unclear due to lack of longitudinal and experimental studies. Patients with heart failure (HF) bear particular risk of poor sleep and low physical activity. PURPOSE: Examine the relationships between within-person day-to-day fluctuations in physical activity and sleep quality among patients with stable NYHA Class I-IV HF and insomnia (insomnia severity index > 7). METHODS: Patients wore a uniaxial wrist accelerometer (Actiwatch) for 14 consecutive days and nights to measure total daytime activity counts, total sleep time, sleep onset latency and sleep continuity (sleep efficiency and wake time after sleep onset (WASO)). Two-level multilevel models with daily and individual variation predicted daytime activity outcomes and sleep outcomes, separately. We adjusted for covariates within (day of the week) and between subjects (age, Charlson Comorbidity Index (CCI), NYHA HF Class, and body mass index (BMI)). Significance p < 0.05. RESULTS: Participants (n=114, M = 62.4 ± 12.1years, female 43%, black 22%, white 75%, NYHA Class 1.9 ± 0.8) on average obese (BMI 31.4 ± 7.6 kg/m2) with multiple comorbidities (CCI 3.2 ± 2.0). Daytime activity (177 ± 82 x 103 counts/day) was associated with younger age (β = -1.32 ± 0.50), fewer comorbidities (β = -10.57 ± 3.26), lower NYHA class (β = -17.99 ± 7.80), and a tendency for lower BMI (β = -1.58 ± 0.83, p = 0.06). Comorbidity was associated with poorer sleep efficiency (β = -0.98 ± 0.48) and more WASO (β = 5.94 ± 2.05). After adjustment for all significant covariates, daytime activity was not associated with sleep characteristics the next night, but every minute less total sleep time (β = -0.075 ± 0.015) or WASO (β = -0.114 ± 0.042) was associated with ~70 - 100 more activity counts the next day. CONCLUSIONS: Similar to studies in other populations, less WASO and less total sleep time both was associated with more activity the next day, but these were not bidirectional relationships since activity the previous day was not associated with sleep characteristics. Future research should confirm these results by polysomnography and hip accelerometry and evaluate mechanisms. Support: NIH- R01NR016191; Yale Center for Sleep Disturbance in Acute and Chronic Conditions.