A comparison between composition and density of the fat-free mass of youngadult Singaporean Chinese and Dutch Caucasians

Background: The body composition was measured in 30 female Singapore Chines e and 34 female Dutch Caucasians aged (mean +/- SD) 22.7 +/- 3.6 years and in 39 male Singapore Chinese and 39 male Dutch Caucasians aged 23.3 +/- 3.2 years. Methods: The body fat percentage (BF%) was measured using densitome try and deuterium oxide dilution as well as a chemical four-compartment mod el, including fat mass, water, mineral, and protein. The chemical compositi on of the fat-free mass (FFM) and its density were calculated. Results: Mea n body mass index (kg/m(2)) and BF% based on the four-compartment model in the four subgroups were 20.3 +/- 1.5 and 29.7 +/- 5.3 for the Singapore Chi nese females, 22.1 +/- 1.2 and 28.9 +/- 5.2 for the Dutch females, 21.5 +/- 2.5 and 19.4 +/- 6.1 for the Singapore males, and 22.2 +/- 1.8 and 15.8 +/ - 4.9 for the Dutch males. The contributions to the FFM of water, protein, and mineral were, respectively, 72.8 +/- 1.5, 19.2 +/- 1.8, and 7.9 +/- 0.8 % for the Singaporean females and 74.2 +/- 1.4, 18.3 +/- 1.5, and 7.6 +/- 0 .5% for the Dutch females. In males these figures were 73.2 +/- 1.7, 19.6 /- 1.7, and 7.3 +/- 0.5% for the Singaporeans and 72.9 +/- 1.4, 20.5 +/- 1. 4, and 6.6 +/- 0.5% for the Dutch. In females, these figures were for all c omponents significantly different between the ethnic groups; in males the d ifferences were significant for protein and mineral only. The calculated de nsities (kg/l) of the FFM were 1.1074 +/- 0.0057, 1.1012 +/- 0.0051, 1.1027 +/- 0.0065, and 1.1004 +/- 0.0055 for the Chinese females, Caucasian femal es, Chinese males, and Caucasian males, respectively. Conclusions: Although it cannot be excluded that the differences in density and composition of t he FFM are due to methodological differences between the study sites, the f indings may have consequences for the validity of single (reference) method s due to violation of normally used assumptions. The findings explain the o bserved underestimation of BF% from density using Siri's formula in the Chi nese subjects, especially in females. The use of deuterium oxide as a singl e method to determine the BF% resulted in a smaller bias than densitometry (overall 1.6 +/- 2.0% for densitometry, 0.3 +/- 1.7% for deuterium dilution ), suggesting that the deuterium oxide method may be more suitable as a sin gle method in comparative body composition studies. Copyright (C) 2000 S. K arger AG, Basel.