University Cl~ildret~'~ Ilospital. Ilci(lc~lhc~r,y, Gcrrtlro~y /F.S., ;\I.G., K.S../ ut~d hr~rcI.ar Rcscw-ch Coltrc~.Kurl.tr~rlrc, Gcrr)lunjv [..I .%.I ABSTRACT. Despite the increasing use of tetrapolar TBK, total body potassium whole-body bioelectric impedance (BI) analysis in the as-CV, coefficient of variation sessment of body composition, its usefulness in estimating RhISE, root mean square error fat-free mass (FFR.1) has not been evaluated in comparison with conventional skinfold anthropometry in children. We therefore compared I) the intraobserver and interobserver reproducibility of BI and skinfold measurements and the Within the past few years, BI analysis has gained wide use in derived FFhl estimates, and 2) the predictability of FFhI the assessment of body composition. The method appears para s calculated from measurements of total body potassium ticularly suitable for use in children, because the measurements (TBK) using ' "K spectrometry by equations based on either are fast, noninvasive, painless, and require little subject cooper-BI or skinfold measurements in 112 healthy children, ation. The BI technique is based on the difference in specific adolescents, and young adults aged 3.9 to 19.3 y. best-resistance to an electrical current between aqueous and nonfitting equation to predict TBK-derived FFhl from BI and aqueous compartments within the body. Validation studies have other potential independent predictors was developed and shown the accuracy and precision of the technique to be at least cross validated in two randomly selected subgroups of the comparable with conventional anthropometric methods such as study population by stepwise multiple regression analysis. skinfold thickness measurements in both adults and children Although the technical error associated with BI measure-(1-4). ments was much smaller than that of skinfold measure-The few studies performed so far in children suggest that the mcnts, the reproducibility of BI-derived FFhl estimates changing proportions of body fluid compartments during child-(intraobserver coefficient of variation [CV], 0.39%; inter-hood may require the use of specific pediatric prediction equaobserver CV, 1.2370) was only slightly better than that of tions (4-7). However, the validation methods used so far are FFhI estimates obtained by use of weight and two skinfold difficult to apply in smaller children or rely on a number of measurements (0.62% and 1.39%, respectively). 'l'he cross unverifiable assumptions. Densitometric underwater weighing validation procedure yielded the following best-fitting pre-and respirometric assessment of residual lung volume require diction equation: FFhI = 0.65. (height2/impedance) + full subject cooperation. In addition, the specific density of the 0.68.age + 0.15 (Rz = 0.975, root mean square error = FFM seems to change during childhood; variable-density FFM 1.98 kg, CV = 5.8%, 95% limits of agreement = -1 1.1% models have been applied but are hypothetical (4, 8). Isotope to +12.470). Conventional anthropometry, using publi...