Validation of body composition estimates in male and female distance runners using estimates from a four-component model

This study used estimates of body composition from a four-component model ( %Fat(d,w,m)) to determine whether the assumed density (D-ffm) and the compo sition of the fat-free mass (FFM), and estimates of body composition from m ethods based on two- and three-component models are valid in distance runne rs. Measures of body density (D-b) by underwater weighing, total body water by deuterium dilution, and bone mineral by dual-energy X-ray absorptiometr y (DXA) were obtained in 10 female and 12 male runners and an equal number of controls matched for age, height, weight, gender, and ethnicity. D-ffm o f the runners did not differ from 1.1 g.cm(-3) or from the controls even th ough the composition of the FFM differed from that assumed. Therefore, perc entage of body fat (%Fat) from densitometry did not differ from %Fat(d,w,m) in the runners, although individual variation was substantial (-5.3 to 3.0 % body mass, SDdiff = 2.1% body mass). Three-component model estimates of % Fat from Db and body water agreed quite closely with %Fat(d,w,m) ((x) over bar less than or equal to 1.6% body mass, SDdiff = 0.5% body mass), whereas those from Db and body mineral ((x) over bar(diff) = 1.6-2.9% body mass, S Ddiff - 2.3% body mass) and from DXA ((x) over bar(diff) = 2.2-4.0% body ma ss, SDdiff = 2.1% body mass) were less accurate and underestimated %Fat(d,w ,m). we conclude that D-ffm does not differ from 1.1 g.cm(-3) and that dens itometry is a valid method to estimate group mean body composition, but tha t measurement of D-b and total body water are necessary for accurate indivi dual estimates of body composition in male and female distance runners. (C) 2000 Wiley-Liss, Inc.