Impact of time spent walking on incident dementia in elderly Japanese.

Abstract

This study examined the hypothesis that time spent walking daily has a marked impact on incident dementia. First, we analyzed data from a Japanese cohort (n=13990 aged ≥65y) to obtain hazard ratios. Time spent walking per day (<0.5, 0.5-1, or ≥1h) was assessed using a self-reported questionnaire. Data on 5.7-year incident dementia were retrieved from the public long-term care insurance database. After estimating the multivariate-adjusted hazard ratios (HRs) of incident dementia using the Cox model, the population attributable fraction (PAF) was calculated using the prevalence in a representative Japanese survey, the "National Health and Nutrition Survey". The time spent walking per day showed an inverse association with incident dementia: the multiple-adjusted HRs (95% confidence intervals) were 1.00 (reference) for less than 0.5hour, 0.81 [0.71, 0.92] for 0.5 to 1hour, and 0.72 [0.62, 0.84] for more than or equal to 1hour. Our estimates indicate that 18.1% of dementia cases would be attributable to walking if all subjects walked more than or equal to 1h/d and 14.0% if subjects increased their daily time spent walking to one level above the present one (<0.5 to 0.5-1 or 0.5-1 to ≥1h). Our results suggest that the daily time spent walking has a considerable preventive impact on incident dementia in Japan.