Total body water in health and disease: Have anthropometric equations any meaning?

Abstract

The accurate measurement of total body water (TBW) requires isotopic dilution techniques that are not easily applicable to the clinical setting. Therefore, indirect methods of estimating TBW are commonly employed, such as bioelectrical impedance analysis (BIA) and anthropometry. In the human body, >90% of the measured impedance is composed of resistance (R). The aim of the present study was to compare TBW estimated by means of two anthropometric equations (by Watson and Hume) with TBW obtained by BIA (equations proposed by Sun et al.) in a group of white disease-free individuals (n = 3625, 1860 men and 1765 women) and white haemodialysis (HD) patients (n = 688, 443 men and 245 women). They underwent one single-frequency BIA measurement, on the nondominant side of the body, injecting an 800-muA and 50-kHz alternating sinusoidal current with a standard tetrapolar technique. The BIA variable measured was R. Among them, a selection of disease-free individuals (n = 481) and HD patients (n = 270), pair-matched by age, body weight and height, after stratification by gender, was made. When comparing the four pair-matched groups, it was found that (1) TBW was not different (disease-free men versus HD men; disease-free women versus HD women) when using anthropometric equations, which utilize quite identical parameters (age, body weight and height); (2) R was statistically significantly different in the four groups (511 +/- 58 SD Omega in disease-free men versus 558 +/- 80 in HD men, P < 0.0001; 593 +/- 70 Omega in disease-free women versus 615 +/- 100 in HD women, P < 0.02) and (3) therefore, TBW was statistically significantly different only when applying BIA equations (P < 0.0001 and 0.05, respectively). The present study demonstrates that anthropometric equations for the estimation of TBW can be used only within a specific population in order to assess individual differences; they cannot be used in order to compare two different populations.