Abstract 4144299: Association of Sleep and Exercise with Chronic Disease: Insights from Long-Term Wearable Data Among All of Us Participants

Introduction: Sleep and physical activity (PA) are essential components of ideal health. However, epidemiological data assessing PA and sleep have been limited by short measurement periods, small samples, or recall bias. We assess the relationship between objective measures of long-term PA and sleep across the human phenome. Methods: We included 9399 individuals in the All of Us cohort with >6 months of wearable (Fitbit) data and linked electronic health records. Fitbit data included daily total sleep duration (TSD) and moderate and vigorous physical activity (MVPA) over a median time of 1103 days. We used Cox regression to model the association of MVPA and TSD across 1711 phecodes (Cox pheWAS), adjusting for age, sex, and race. Top associations were investigated individually using restricted spline curves to quantify the change in hazard ratio (HR) of disease per minute of sleep or MVPA. Individual phenotype associations were further adjusted for smoking status, alcohol intake, education level, and BMI. We performed interaction analyses to assess whether TSD modifies the effect of MVPA on disease. We accounted for multiple hypothesis testing with Bonferroni correction. Results: TSD and MVPA levels were uncorrelated (r2: 0.004) and independently associated with several cardiometabolic and non-cardiometabolic diseases, including obesity (MVPA: HR 0.97, p 4e-14; TSD: HR 0.99, p 2e-16), hypertension (TSD: HR 0.99, p 4e-5), major depressive disorder (MVPA: HR 0.98, p 2e-7; TSD: HR 0.99, p 3e-3), and chronic migraine (MVPA: HR 0.97, p 3e-5) (Fig.1). The association between MVPA and chronic disease often demonstrated a linear or curvilinear dose-response curve. By contrast, the association of TSD and chronic disease was often sigmoidal, with increasing benefit of longer sleep between ~5hr20min and 6hr40min of sleep/day, but relatively constant health benefits/risks for longer/shorter sleep durations. Conclusion: In All of Us participants, TSD and MVPA are independently and inversely associated with chronic disease based on objective, long-term measures. These data highlight the importance of both sleep and MVPA in maintaining ideal health, and will help inform personalized, disease-specific sleep and PA recommendations.