Abstract
OBJECTIVES: Reaching the moderate-to-vigorous physical activity (MVPA) recommendations of 150 min/wk is difficult for older adults, particularly among those living with frailty and its associated risk of dementia. We examined the dose-response relationship between MVPA and dementia risk among at-risk persons living with and without frailty enrolled in the UK Biobank study. DESIGN: Survival analysis within a prospective cohort study. SETTINGS AND PARTICIPANTS: Participants at risk for all-cause dementia who wore an Axivity AX3 triaxial wrist-worn accelerometer between February 2013 and December 2015. METHODS: MVPA was estimated from wrist-worn accelerometry in a subpopulation of the UK Biobank study. A modified version of the physical frailty phenotype was used to define frailty. Associations between MVPA dose (including interactions with frailty) and first-time incident dementia were analyzed using Cox regression models. MVPA was treated continuously and categorically across 5 levels to estimate the dose-response curve. Models were adjusted for demographics, frailty status, and comorbidities. RESULTS: This study included 89,667 adults (median age, 63 years; 56% women), with 735 participants developing dementia over an average of 4.4 years. Average weekly MVPA was 126 minutes. Each 30 minutes higher MVPA was associated with a 4% reduction in the risk of all-cause dementia (hazard ratio, 0.96; 95% CI, 0.93-0.99). The hazard ratios for engaging in 0-34.9, 35-69.9, 70-139.9, and ≥140 MVPA minutes per week were 0.59, 0.40, 0.37, and 0.31, respectively (P < .05 for all) compared with 0 MVPA minutes per week. All associations were similar across frailty status (interaction P for all models > .21). CONCLUSIONS AND IMPLICATIONS: Our results suggest engaging in any additional amount of MVPA reduces dementia risk, with the highest benefit appearing among individuals with no MVPA. These associations are not substantially modified by frailty status.