Background/objectives: Body composition techniques are required for monitoring response to treatment in individual obese children, and assessing the efficacy of weight loss programmes. Densitometry is readily undertaken, using air displacement plethysmography (ADP), but requires appropriate information on the density of lean tissue (D(LT)). The aims of this study were to develop predictive equations for D(LT) in obese children and adolescents, and to test the accuracy of ADP when using such predicted D(LT) values in an independent longitudinal sample using the four-component model as the reference method.
Subjects/methods: Equations for the prediction of D(LT) from age, gender and body mass index standard deviation score were developed in 105 children (39 boys). Accuracy of ADP, when incorporating predicted D(LT) values, was tested for baseline body composition and its change over time in a separate sample of 51 children (20 boys).
Results: The predictive equation explained 33% of the variance in D(LT). Fat mass obtained from ADP using such predicted values had a mean (s.d.) bias of 0.32 (1.39) kg, nonsignificant, whereas change in fat mass had an error of -0.25 (1.38) kg, nonsignificant. Hydration was strongly correlated with D(LT).
Conclusions: Use of ADP with predicted D(LT) values was associated with nonsignificant bias when estimating fat mass and its change over time. This study aids the application of ADP in childhood obesity research and clinical practise. The limits of agreement (±2.8 kg) relative to four-component values are moderately better than those for X-ray absorptiometry (±3.2 kg). Further improvement to accuracy would require assessment of lean tissue hydration by bioelectrical impedance.