Water turnover estimated by prediction equation and all-cause and cause-specific mortality in Japan: A cross-sectional prefecture-level ecological study

Objective

Although water is essential for maintaining life and health, the association of water turnover (WT) with mortality is, to our knowledge, unknown. We aimed to 1) determine the annual mean and variance of WT and 2) evaluate the association between WT and mortality in all 47 prefectures in Japan.

Methods

This cross-sectional, prefecture-level ecological study used data from government surveys that provided prefecture-level data. WT was calculated considering lifestyle and environmental factors using an equation previously developed by the international doubly labeled water database group. Each prefecture was classified into a WT quartile. We obtained data on age-standardized mortality from the 2015 Vital Statistics. The mortality rate ratio (RR) was calculated using a multivariable Poisson regression model.

Results

The annual mean and variance of WT were 3771 and 990 mL/d in men and 3084 and 990 mL/d in women, respectively. After adjusting for confounders, compared to the prefecture in the highest WT quartile, those in the lowest WT quartile were associated with a higher mortality RR for all-cause (Men: RR = 1.07 [95% confidence interval (CI) = 1.02–1.11]; Women: RR = 1.05 [95% CI = 1.00–1.11]) and stroke (Men: RR = 1.16 [95% CI = 1.00–1.33]; Women: RR = 1.18 [95% CI = 1.01–1.42]). This association was moderately weaker when adjusted for exercise habits, especially among women.

Conclusions

Our findings indicate large regional and seasonal variations in WT in Japan and that the risk of excess death due to lower WT may be partially explained by physical activity.