Decrease in fat mass (FM) is a one of the aims of pediatric obesity treatment; however, measurement techniques suitable for routine clinical assessment are lacking. The objective of this study was to validate whole-body bioelectrical impedance analysis (BIA; TANITA BC-418MA) against the three-component (3C) model of body composition in obese children and adolescents, and to test the accuracy of our new equations in an independent sample studied longitudinally. A total of 77 white obese subjects (30 males) aged 5-22 years, BMI-standard deviation score (SDS) 1.6-3.9, had measurements of weight, height (HT), body volume, total body water (TBW), and impedance (Z). FM and fat-free mass (FFM) were calculated using the 3C model or predicted from TANITA. FFM was predicted from HT(2)/Z. This equation was then evaluated in 17 other obese children (5 males) aged 9-13 years. Compared to the 3C model, TANITA manufacturer's equations overestimated FFM by 2.7 kg (P < 0.001). We derived a new equation: FFM = -2.211 + 1.115 (HT(2)/Z), with r(2) of 0.96, standard error of the estimate 2.3 kg. Use of this equation in the independent sample showed no significant bias in FM or FFM (mean bias 0.5 +/- 2.4 kg; P = 0.4), and no significant bias in change in FM or FFM (mean bias 0.2 +/- 1.8 kg; P = 0.7), accounting for 58% (P < 0.001) and 55% (P = 0.001) of the change in FM and FFM, respectively. Our derived BIA equation, shown to be reliable for longitudinal assessment in white obese children, will aid routine clinical monitoring of body composition in this population.