Pattern of fat distribution rather than obesity is of importance for cardiovascular morbidity and mortality. The accurate measurement of total and regional fat mass requires sophisticated and often expensive methods that have limited applicability in the clinical setting. The aim of this study is to evaluate body fat distributions by ultrasound (US) as a gold standard method for measuring visceral, preperitoneal and subcutaneous fat layers and comparing with anthropometric results, and then to find the most reliable anthropometric measurement in childhood obesity. Study group of 51 obese children (21 F, 30 M) (mean age+/-s.d.: 11.5+/-2.6 years) and control group of 33 non-obese children (17 F, 16 M) (mean age+/-s.d.: 12.2+/-2.7 years) were recruited for this study. Anthropometric measurements as body mass index (BMI), waist circumference (WC), waist/hip ratio (WHR), triceps and subscapular skinfold thicknesses were taken from all the participants. Abdominal preperitoneal (P), subcutaneous (S) fat at their maximum (max) and minimum (min) thickness sites, visceral (V), triceps (TrUS) and subscapular (SsUS) fat thicknesses were also measured ultrasonographically. In the obese group, BMI was significantly correlated with US measurements of fat thicknesses, except Pmin and SsUS, whereas in the control group, BMI was significantly correlated with all US fat measurements. The relation of US measurements with skinfold thickness and WC was more significant in the control than in the obese group. No relation between WHR and US fat thickness measurements was found in both groups. Multiple regression analysis, using V as the dependent variable and anthropometric parameters, gender and the group as the independent variables, revealed BMI was the best single predictor of V (R(2): 0.53). This study suggests that the validity of the anthropometric skinfold thickness in the obese children is low. Despite the limitations reported in the literature, in our study, BMI provides the best estimate of body fat. WHR in children and adolescents is not a good index to show intra-abdominal fat deposition.