METHODS: A retrospective study performed in 4,818 non-diabetic participants (mean age 43 years, 78.3% men) in a health screening program in whom serial health examinations were performed in 5 consecutive years from 2010 to 2014. Average successive variability of weight (ASVW) was defined by the amount of body weight change in absolute value between the successive years over 5 years summed and divided by four. The subjects were divided into two groups according to body mass index (BMI), normal weight (<23 kg/m2) and overweight (≥23 kg/m2).
RESULTS: Over 4 years, 3.2% developed diabetes. When the subjects were divided into 3 groups according to tertile groups of ASVW, those in the highest tertile showed significantly increased risk for diabetes development compared to those with the lowest tertile {odds ratio (OR) 1.860; 95% CI 1.130-3.063}. When similar analyses were performed according to the 4 groups divided by baseline body weight and ASVW over four years, those who were more than overweight at baseline with high ASVW showed significantly increased risk of diabetes development compared to those had normal weight and low ASVW (OR 2.266; 95% 1.123-4.572). When the subjects were divided into six group according to weight change and ASVW, those with increased weight over 4 years and high ASVW showed the highest risk for diabetes development among the groups compared to those with stable weight and low ASVW over four years (OR 3.660; 95% CI 1.402-9.553).
CONCLUSIONS: Those with high ASVW showed significantly increased risk for diabetes development over four years compared with those who had low ASVW. Weight cycling was significantly associated with increased risk for diabetes.