“…3 By χ 2 test. In logistic regression analysis, each parameter needed to be categorized into the low and high groups with cutoff point that was as following: 4 <60 years old, 5 <25 kg/m 2 BMI, 6 <126 mL/dL fasting serum glucose or medication for diabetes, 7 <6.5% HbA1c or medication for diabetes, 8 <130/90 mm Hg for blood pressure, 9 <college, formal and current smoking, 10 <USD 2,000/month, 11 nonsmoking, 12 <20 g/day alcohol drinking, 13 <3 g/day coffee drinking, 14 <estimated energy intake, 15 <65 energy % carbohydrate intake, 16 <15 energy % fat, 17 <13 energy % protein, 18 <0.75 serving/day of dairy products adding milk, yogurt, and cheese, 19 <500 mg/day Ca intake, and 20 <100 g/day kimchi intake. The Best Model for the Association with ARC Assessed by Gene-Gene Interaction Using GMDR Since each SNP selected for ARC was weakly associated with ARC, a gene-gene interaction study was conducted to better explain the ARC genetic effect by combining multiple genetic variants.…”