Associations Of 24-hour Compositional Behaviors With Cardiorespiratory Fitness: Study Of Muscle, Mobility And Aging (SOMMA)

Physical activity (PA) regression models do not adequately capture the time co-dependency of behaviors in a 24-hour period. For example, a regression modeling the association of higher amounts of moderate-to-vigorous PA (MVPA) with cardiorespiratory fitness (CRF) fails to account for the reallocation of time from another behavior (e.g., sleep, sedentary behavior (SB), or light intensity PA (LPA)) to attain the higher amounts of MVPA, thus simultaneously modeling the effects of higher MVPA and lower amounts of another behavior. PURPOSE: Utilize CoDA methods to address the time co-dependency and quantify the interrelation between 24-hour behaviors (sleep duration (hereinafter sleep), SB, LPA, and MVPA) and CRF. METHODS: At baseline, SOMMA participants wore an ActiGraph GT9X on their non-dominant wrist for at least three days and up to seven days. Intensities were categorized using Montoye et al. 2021 cutpoints for activity and the Cole-Kripke algorithm for sleep. CRF (VO2peak, mL/kg/min) was measured using treadmill cardiopulmonary exercise testing. Compositional isotemporal substitution models examined the predicted change in VO2peak when 30 minutes was theoretically reallocated from one behavior to another, after adjusting for age, sex, race, site, smoking status, and a multimorbidity index (0 - 11 conditions). RESULTS: Participants (N = 769; mean age 76.2 ± 4.9 years; 57.3% women) had a 24-hour compositional mean of 30.5% sleep (439 min/d), 49.9% SB (718 min/d), 7.0% LPA (102 min/d), and 12.6% MVPA (181 min/d). Mean VO2peak was 20.2 ± 4.9 mL/kg/min. Reallocating 30 minutes of SB to MVPA resulted in 0.55 mL/kg/min higher predicted VO2peak when compared to values at the mean composition (P < 0.05). Additionally, reallocating 30 minutes of LPA to MVPA resulted in the strongest observed association with 1.22 mL/kg/min higher predicted VO2peak (P < 0.05). Reallocating time from sleep to MVPA was not associated with predicted VO2peak. CONCLUSION: As VO2peak is a hallmark biomarker of aging, understanding its interrelations with 24-hour behaviors may inform targeted interventions—likely aimed at increasing MVPA at the expense of LPA—to promote the healthier, independent aging of older adults. SOMMA funded by: NIH R01 AG059416, P30 AG024827, P30 AG021332, and UL1 TR001420. BTS funded by NIA T32AG000181. SOMMA funded by: NIH R01 AG059416, P30 AG024827, P30 AG021332, and UL1 TR001420. BTS funded by NIA T32AG000181.