| The purpose of this study was to evaluate the diagnostic accuracy of various childhood obesity definitions using bioelectrical impedance analysis (BIA) as a criterion measure of body fat. To achieve such an objective, the study involved three phases. First, the validity of BIA in measuring body composition in Chinese children and adolescents was evaluated. Then, examination on measuring accuracy of several popular BIA scales was performed. Finally, mass testing of body fat levels using an accurate BIA scale, as well as other anthropometric measures, was conducted to evaluate the diagnostic accuracy of childhood obesity definitions. A total of 255 healthy Chinese children and adolescents aged 9 years to 19 years participated in the first and second phases of the study. Body composition was measured from BIA using a criterion device (Biodynamics 310) and four popular commercial BIA scales and compared with measurements from dual-energy X-ray absorptiometry (DEXA). BIA provided excellent predictions of body composition among Chinese children and adolescents. However, all of the previously developed BIA equations yielded biased estimation. The four commercial BIA scales had similar diagnostic accuracy in measuring body fat. Subsequently, another 2,134 boys and girls were recruited to take part in body fat measurements using the best BIA scale identified in the second phase with adjusted equations, as well as other anthropometric measurements. The diagnostic accuracy of all existing definitions of childhood obesity was poor in both genders. Moreover, Receiver Operating Characteristics (ROC) analysis found that childhood obesity definitions using weight-and-height based indices [Body mass index (BMI) and weight-for-height) had superior sensitivity and specificity in identifying obese children compared with waist circumference-based indices. Therefore, with adjusted cut-off criteria for weight-and-height indices, the diagnostic accuracy of childhood obesity would be improved. |