Predictive capacity of different bioelectrical impedance analysis devices, with and without protocol, in evaluation of adolescents

Objective: this study was performed to determine the predictive capacity of four different bioelectrical impedance analysis (BIA) devices in the assessment of adolescents, with and without a protocol. Methods: a cross-sectional study was performed with 215 adolescents aged 10 to 14 years, of both genders, evaluated through anthropometry and body composition by dual energy X-ray absorptiometry (DXA) and by four different BIA devices, with and without a protocol. The following tests were used: Kolmogorov-Smirnov's, chi-squared, Student's t or Mann-Whitney's, Kruskal-Wallis's, Wilcoxon's, and kappa index. The ROC curves were constructed and the sensitivity, specificity, and positive and negative predictive values were calculated. Results: of the 215 adolescents, 44.2% had excessive body fat. The tetrapolar BIA device equipped with eight tactile electrodes showed more sensitivity and results that were closer to those obtained by DXA (area under the ROC curve [AUC] = 0.964 with protocol and AUC = 0.973 without protocol, p < 0.001), as well as greater agreement (k = 0.67 with protocol and k = 0.63 without protocol, p < 0.001). The evaluation without protocol was similar to that by DXA in most investigated situations (p > 0.05). Conclusion: BIA is capable of predicting alterations in adolescents' body composition. When it is impossible to perform the assessment with a protocol, its results may be useful in population studies. (C) 2013 Sociedade Brasileira de Pediatria. Published by Elsevier Editora Ltda. All rights reserved.