RESUMEN
PURPOSE: This study determined which obesity measurement correlates the best with diabetes and prediabetes. MATERIALS AND METHODS: This cross-sectional study enrolled 1603 subjects (611 men, 992 women; age 30-64 years) at the Cardiovascular and Metabolic Diseases Etiology Research Center. Body mass index, waist circumference, waist-height ratio, waist-hip ratio, waist-thigh ratio, and visceral fat were used as measures of obesity. Visceral fat was acquired using dual-energy X-ray absorptiometry (DXA). The prevalences of diabetes and prediabetes were defined using the criteria in the American Diabetes Association 2015 guidelines. RESULTS: After adjusting for age and other potential confounding factors, participants with a visceral fat mass in the upper 10th percentile had a higher odds ratio (OR) for diabetes and prediabetes than the upper 10th percentile of other adiposity indices [men, OR=15.9, 95% confidence interval (CI)=6.4-39.2; women, OR=6.9, 95% CI=3.5-13.7]. Visceral fat mass also had the highest area under the curve with diabetes and prediabetes in both men (0.69, 95% CI=0.64-0.73) and women (0.70, 95% CI=0.67-0.74) compared to other anthropometric measurements of obesity. CONCLUSION: Visceral fat mass measured using DXA is an indicator of diabetes or prediabetes, due to its ability to differentiate between abdominal visceral and subcutaneous fat.