Introduction: The technique for measuring body composition by bioelectrical impedance analysis (BIA) loses precision in case of thinness or severe obesity, because of the limitations of the mathematical model which assumes that human body has a cylindrical shape. Segmental BIA minimizes this source of error since it modelizes the body as a collection of five cylinders. New devices, whose use becomes easier, become now available. We evaluated the interest and accuracy of a new segmental BIA device, the Biacorpus RX4000, compared with DEXA. Methods and results: In 36 subjects aged 15 to 69 years covering a range of BMI from 13 to 41 kg/m in whom body composition was measured by DEXA compared with the conventional BIA. The composition of each segment is accurately predicted (r = 0.726 to 0.987), as well as truncal fat mass (r = 0.9575; average error 2.25 [95% CI: 1.41-3.09]). For the whole body fat mass prediction is excellent (r2 = 0.9755; average error 0.532 [CI: -0.495-1.56]) and remains high in lean subjects (r = 0,925; slope: 0.953, 0.46 orderly; average error 0015 [CI: -1.1-1.13]). In 20 persons investigated by standardized breakfast with mathematical modeling, insulin sensitivity determined by the oral minimal model (OMM-SI) is more correlated with the truncal fat mass (r = -0686) than to total body fat mass or fat mass of other body segments. Conclusion: This segmental approach increases the accuracy of BIA compared to previous BIA techniques, even in very lean subjects.
© Copyright 2013 Science & Sports. Elsevier. All rights reserved.