They conducted a prospective Cox Regression analysis using a cohort of 13,540 Japanese community residents. They found a significant inverse association
with a 35% lower risk of diabetes incidence in male, but not female, participants during a 10-year follow up.