Predictability and implications of anthropometric indices for metabolic abnormalities in children: Nutrition and health survey in Taiwan elementary children, 2001-2002

Background: To determine whether separate anthropometric screening tools are needed for obesity and the metabolic syndrome in children, we compared the predictability of several anthropometric indices, including waist circumference (WC) and body mass index (BMI), with regard to metabolic disorders. Study design: The Nutrition and Health Survey in Taiwan Elementary School Children (2000-2001) collected data from 2,215 children. Logistic regression analysis was used to study the association between anthropometric indices and metabolic abnormalities, which was defined as two or more of the following conditions: high fasting triglycerides, high fasting glucose levels, high blood pressure and low high-density lipoprotein-cholesterol. The receiver operating characteristic curve was used to study the sensitivity and specificity of these anthropometric indices. Results: predictability was the ranked highest for WC (R² = 10.69%), followed by BMI (R² = 9.80%), arm girth (R² = 9.75%), hip circumference (R² = 9.43%), scapular skinfold thickness (R² = 9.28%) and waist-to-height ratio (R² = 9.25%). Waist circumference or BMI cut-offs for maximal balanced sensitivity and specificity were close to the 60^th percentile for each age and gender group. Values were greater in boys than in girls and increased with age. Conclusion: It is justifiable to use the WC criteria to define the metabolic syndrome in children. Due to its practicality, BMI remains the most suitable index for defining overweight/obesity. Only moderate levels of sensitivity and specificity were achieved with these two popular obesity indices with regard to metabolic abnormalities.