Potential causal effect of physical activity on reducing the risk of dementia: a 6-year cohort study from the Japan Gerontological Evaluation Study

BACKGROUND: The causal effect of physical activity on reducing dementia risk has been questioned due to the possibility of reverse causation. This study examined the potential causal effects of physical activity on reducing dementia risk using residency in a snowy area as an instrumental variable (IV) representing the physical activity of older adults. METHODS: We used cohort data from the Japan Gerontological Evaluation Study, a longitudinal cohort enrolling people aged 65 or older who were physically and cognitively independent in 2013; study participants were followed for an average of 5.7 years. Participants in the present study included 73,260 individuals living in 19 municipalities in Japan. Physical activity was measured by self-report questionnaires and the incidence of dementia was ascertained by linking participants to the public registries of long-term care insurance. IV estimation was obtained from a piecewise Cox proportional hazard model using a two-stage regression procedure. RESULTS: During the study period, we ascertained 8714 cases (11.9%) of dementia onset. In the IV analysis, we found that the frequency of physical activity per week was negatively associated with dementia risk, though the association weakened over time (Year 1: hazard ratio = 0.53, 95% confidence interval: 0.39–0.74; Year 4: 0.69, 0.53–0.90; Year 6: 0.85, 0.66–1.10). CONCLUSIONS: Our IV analysis indicated a potential causal effect of physical activity on reducing dementia risk that persisted for at least 4 years of follow-up. Thus, we conclude that physical activity should be recommended for older adults to reduce dementia risk. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12966-021-01212-w.