OBJECTIVE: To investigate the association of fatness in children with dyslipoproteinemia and high blood pressure, with the purpose of proposing standards for childhood obesity which are directly related to intermediate biological parameters that predict future disease. DESIGN: A cross-sectional study of a large, nationally representative sample of Australian schoolchildren. SUBJECTS: 1834 children aged 9 or 15 years, with skinfolds blood lipid measurements on 1144 and with skinfolds and blood pressure measurements on 1757. MEASUREMENTS: Skinfolds thicknesses measured at four locations (triceps, biceps, subscapular and suprailiac) using holtain calipers, percent body fat calculated from the sum of four skinfolds, Quetelet's index calculated from weight and height, waist and hip circumferences, plasma total cholesterol and triglycerides determined using a Technicon Autoanalyser II, high density lipoprotein cholesterol (HDLC) analysed following precipitation with heparin manganese, and systolic blood pressure (SBP) measured using a standard mercury sphygmomanometer. RESULTS: For 9 years old girls, 15 year old girls and 9 year old boys, dichotomising their HDLC and SBP measurements by percent body fat defined the two groups most homogeneous in terms of a measure of within-group variation. The cut-points in percent body fat were in the ranges 29-35% (girls) and 17-20% (boys). CONCLUSION: It is feasible to use the current biomedical status of individual children to define criteria for obesity. A cut-off point of 30% body mass as fat for girls and 20% for boys appears to be an appropriate standard.