Development and cross-validation of predictive equation for estimating total body lean in children.

AIM: To develop and cross-validate a predictive equation for estimating lean body mass (LBM) in children, based on bioelectrical impedance analysis (BIA) and anthropometric parameters, and to compare our equation with other predictive methods. METHODS: We evaluated body composition of 155 children (age 5-14 years) by BIA and dual energy X-ray absorptiometry (DXA). Children were divided in two groups: developing set and cross-validation set. Statistical analysis as multiple regression and Bland-Altman methods were performed. RESULTS: From developing set (105 children) a predictive LBM (kg) equation was created. From the cross-validated set (35 children), our equation was seen to agree with LBM (kg) by DXA. On the contrary, LBM predicted by Schaefer's equation and BIA did not agree with the criterion measure. CONCLUSIONS: This cross-validated equation can be useful in epidemiological studies and also in clinical practice, permitting a better definition and follow up of children's body composition.