Comparison of body fat-free masses calculated from hand-to-foot and foot-to-foot resistances with DXA measurements.

This article compares the determination of body fat-free-mass (FFM) by impedance, using either hand-to-foot resistance (R13) or foot-to-foot one (R34) from comparison with dual X-ray absorptiometry (DXA) measurements in a normal population. The first goal was to see if the foot-to-foot resistance used in body fat analysers provides less accurate information for body FFM than the hand-to-foot one used by medical impedance-meters. Another goal was to compare the prediction accuracy of six different regression equations of FFM for each sex and for each resistance relatively to DXA. The impedancemeter used in this study was a Tefal prototype with 4 plantar electrodes and 4 additional electrodes for the hands and providing hand-to-foot and foot-to-foot resistances. Coefficients of these correlations were determined by comparison with FFM measured by DXA in a 1st cohort of 170 healthy adults. For an independent validation, these equations were tested in a 2nd cohort of 86 adults who underwent the same impedance and DXA protocols, using Student's paired t-tests. The accuracy of FFM prediction increased generally with the number of physiologic parameters included in the regression, but none of our equations gave FFM predictions significantly different from DXA. FFM calculated from the foot-to-foot resistance were closer to DXA values than those calculated from hand-to-foot resistance, as their average P-value of comparison with DXA was higher at 0.695 against 0.387 for R13.

A comparison between two methods for measuring limb resistances with wrist and ankle electrodes.

Segmental measurements of limb resistances permit to evaluate the regional distribution of fat-free-mass (FFM). As shown by Organ et al. (J Appl Physiol 1994;77:98-112) "virtual" limb resistances can be measured with only peripheral electrodes at ankle and wrist, which is faster and does not require undressing. This paper presents a different method for obtaining "virtual" limb resistances from peripheral electrodes and compares it with that of Organ et al. and with segmental measurements using same protocol. It is found that virtual arm resistances of both methods were overestimated as compared to segmental values, by 6.8% in men and 5.8% in women for our method and by 7.4% and 8% for Organ et al. one. Virtual leg resistances were found, for both methods, to be slightly lower than segmental leg resistances measured with proximal waist electrodes, which were a little overestimated. But after correcting this overestimation using Organ et al. data, we evaluated the overestimation of virtual leg resistances to be 2.3% in men and 2.8% in women by our method versus 4.4% and 5.9%, respectively, for Organ et al. one. An important finding was the strong correlation between virtual resistances of both methods and segmental ones, as their ratio had a small standard deviation between 0.019 and 0.024 for arms and legs. Thus, the overestimation of limb resistance is automatically corrected when FFM is determined by an equation calibrated with DXA segmental measurements.