Sleep, physical activity, sedentary behavior, and risk of incident dementia: A prospective cohort study of 431,924 UK Biobank participants

Abstract Background Sleep, physical activity (PA), and sedentary behavior have been individually associated with the risk of dementia, but findings are inconsistent. Their joint associations with dementia also remain unknown. We aimed to investigate the independent and joint associations of sleep duration, PA, and sedentary behavior with dementia risks and the underlying mechanisms. Method A total of 431,924 participants with a median follow‐up 9.0 years from UK Biobank without dementia were included. Cox proportional hazards regression was performed to test the independent and joint association of sleep, PA, and screen‐based sedentary behavior with dementia. We also test the non‐linearity of these associations. Non‐linear regression was performed to examine the associations of each behavior with brain morphometric measures. Result During a median observation follow‐up period of 9.0 years, 5390 participants developed an incident dementia among 431,924 individuals. Sleep duration, leisure‐time PA, and screen‐based sedentary behavior individually associated with dementia risks in different non‐linear association modes. Sleep duration associated with dementia risk in a U‐shape, and participants with a seven‐hour sleeping pattern showed the lowest dementia risk. Leisure‐time PA associated with dementia in a curvilinear relationship with diminishing returns, and sedentary behavior associated with dementia in a J‐shaped association. The dementia risk was 17% lower in the high leisure‐time PA group ((95%CIs), 0.83 (0.76‐0.91)) and 22% higher in the high sedentary behavior group (1.22 (1.10‐1.35)). A combination of seven‐hour sleep, moderate‐to‐high leisure‐time PA, and low‐to‐moderate sedentary behavior showed lower dementia risks (0.66 (0.56‐0.78)) compared to the referent group with atypical sleep, low leisure‐time PA, and high sedentary behavior. Notably, each of the three behaviors was non‐linearly associated with brain morphometric measures in the similar modes as their associations with dementia. Proper intensity/time‐spent on each behavior of interest was associated with higher brain volumes. The relationship of each behavior with dementia was partially mediated by cerebrovascular diseases. Conclusion A 7‐hour/day sleep duration, higher leisure‐time PA, and lower sedentary behavior are independently and jointly associated with lower dementia risks. Our findings underscore the potential of changing these three daily behaviors concomitantly to mitigate dementia risks in population‐wide strategies.