LAUSR

Increasing trend of diabetes is significant public health threats. This is a consecutive cross-sectional study aimed to estimate the attribution of overweight for diabetes prevalence in Japan. We used individual data aged 20 years and older of National Health and Nutrition Survey in 1997 (2349 men and 3588 women), 2002 (2075 and 3087), 2007 (1614 and 2346), 2012 (5538 and 7981), and 2016 (4327 and 6270). Diabetes was defined as glycated haemoglobin (HbA1c) >6.5% or currently being treated. We estimated multivariable-adjusted odds ratio (OR) and 95% confidence interval (CI) for the association between diabetes and overweight (body mass index >25kg/m2) after adjusting for the number of steps per day, total energy intake (kcal/kg), fat energy ratio (%), smoking and drinking status by logistic regression. There was significant association between diabetes and overweight in all years: multivariable-adjusted OR (95%CI) were 1.78 (1.32-2.40) in 1997, 1.46 (1.10-1.93) in 2002, 1.77 (1.35-2.32) in 2007, 1.79 (1.55-2.07) in 2012, and 2.15 (1.82-2.54) in 2016 for men and 2.53 (1.92-3.32), 2.53 (1.87-3.42), 2.91 (2.15-3.95), 2.94 (2.51-3.45), and 2.72 (2.27-3.25), for women, respectively. The estimated population attribution on diabetes prevalence due to overweight was 14.8%, 11.9%, 18.1%, 18.0%, and 24.6% for men and 27.0%, 29.4%, 29.9%, 29.5%, and 26.7% for women. Overweight was a risk factor for diabetes consistently for 20 years. That attribution was increasing among men but decreasing after peaking in 2007 among women. The present study suggested a high burden of diabetes due to overweight in Japan.