Objective: Body mass index (BMI) is the surrogate measure of adiposity most commonly employed in children and adults. Waist circumference (WC) and the waist circumference-to-height ratio (WCHt) have been proposed as markers of adiposity-related morbidity in children. However, no study to date has compared WCHt, WC, BMI and skinfolds thickness for their ability to detect body adiposity.
Aim: To compare WCHt, WC, BMI and skinfolds for their accuracy in predicting percent body fat (PBF), percent trunk fat (PTF) and fat mass index (FMI) in a large sample of children and adolescents.
Design, setting and participants: We studied 2339 children and adolescents aged 8-18 years from the US National Health and Nutrition Examination Survey 2003/2004. Body fat was measured using dual-energy X-ray absorptiometry. Multivariable regression splines were used to model the association between PBF, PTF, FMI and the predictors of interest.
Results: WCHt alone explained 64% of PBF variance as compared with 31% for WC, 32% for BMI and 72% for the sum of triceps and subscapular skinfolds (SF2) (P<0.001 for all). When age and gender were added to the predictors, the explained variance increased to 80% for the WCHt model, 72% for the WC model, 68% for the BMI model and 84% for the SF2 model. There was no practical advantage to add the ethnic group as further predictor. Similar relationships were observed with PTF and FMI.
Conclusions: WCHt is better than WC and BMI at predicting adiposity in children and adolescents. It can be a useful surrogate of body adiposity when skinfold measurements are not available.