Comparisons of two-, three-, and four-compartment models of body composition analysis in men and women.

This study compared the traditional two-compartment (fat mass or FM; fat free mass or FFM) hydrodensitometric method of body composition measurement, which is based on body density, with three (FM, total body water or TBW, fat free dry mass)- and four (FM, TBW, bone mineral mass or BMM, residual)-compartment models in highly trained men (n = 12), sedentary men (n = 12), highly trained women (n = 12), and sedentary women (n = 12). The means and variances for the relative body fat (%BF) differences between the two- and three-compartment models [2.2 +/- 1.6 (SD) % BF; n = 48] were significantly greater (P </= 0.02) than those between the three- and four-compartment models (0.2 +/- 0.3% BF; n = 48) for all four groups. The three-compartment model is more valid than the two-compartment hydrodensitometric model because it controls for biological variability in TBW, but additional control for interindividual variability in BMM via the four-compartment model achieves little extra accuracy. The combined group (n = 48) exhibited greater (P < 0.001) FFM densities (1.1075 +/- 0.0049 g/cm3) than the hydrodensitometric assumption of 1.1000 g/cm3, which is based on analyses of three male cadavers aged 25, 35, and 46 yr. This was primarily because their FFM hydration (72.4 +/- 1.1%; n = 48) was lower (P </= 0.001) than the hydrodensitometric assumption of 73.72%.