Comparison of waist circumference percentiles versus body mass index percentiles for diagnosis of obesity in a large cohort of children

Context. Waist circumference may offer improved diagnosis of obesity in youth compared with body mass index (BMI), but empirical evidence is limited. Objective. To compare the ability of BMI percentile using UK reference data and waist circumference percentile using UK reference data to diagnose high fat mass in English children. Design and Methods. In 7 722 9-10-year-olds (3 809 boys, 3 913 girls) sensitivity and specificity were calculated and receiver operator characteristic (ROC) analyses undertaken to determine the diagnostic accuracy of BMI and waist circumference z-scores to define high fat mass measured by dual energy x-ray absorptiometry (DXA). High fat mass was defined as being in the top decile of fatness for each sex (359 boys and 367 girls). Results. The area under the ROC curve was slightly higher for BMI percentile (0.92 in boys, 95% CI: 0.91 -0.93; 0.94 in girls, 95% CI: 0.93-0.95) than waist circumference percentile (0.89 in boys, 95% CI: 0.86-0.91; 0.81 in girls, 95% CI: 0.73-0.90). Specificity of BMI percentile was slightly but significantly higher than that of waist circumference percentile for both sexes (p < 0.05 in each case). Conclusions. The present study suggests that waist circumference percentile has no advantage over BMI percentile for the diagnosis of high fat mass in children.