In vivo validation of whole body composition estimates from dual-energy X-ray absorptiometry

Prior, Barry M., Kirk J. Cureton, Christopher M. Modlesky,Ellen M. Evans, Mark A. Sloniger, Michael Saunders, and Richard D. Lewis. In vivo validation of whole body composition estimates fromdual-energy X-ray absorptiometry. J. Appl.Physiol. 83(2): 623-630, 1997.We validated wholebody composition estimates from dual-energy X-ray absorptiometry (DEXA)against estimates from a four-component model to determine whetheraccuracy is affected by gender, race, athletic status, ormusculoskeletal development in young adults. Measurements of bodydensity by hydrostatic weighing, body water by deuterium dilution, andbone mineral by whole body DEXA were obtained in 172 young men(n = 91) and women(n = 81). Estimates of body fat(%Fat) from DEXA (%Fat_DEXA)were highly correlated with estimates of body fat from thefour-component model [body density, total body water, and totalbody mineral (%Fat_d,w,m);r = 0.94, standard error of theestimante (SEE) = 2.8% body mass (BM)] with no significantdifference between methods [mean of the difference ± SD ofthe difference = 0.4 ± 2.9 (SD) % BM,P = 0.10] in women and men. Onthe basis of the comparison with%Fat_d,w,m, estimates of%Fat_DEXA were slightly moreaccurate than those from body density(r = 0.91, SEE = 3.4%; mean of the difference ± SD of the difference = 1.2 ± 3.4% BM).Differences between %Fat_DEXA and%Fat_d,w,m were weakly related tobody thickness, as reflected by BMI (r = 0.34), and to the percentage of water in the fat-free mass(r = 0.51), but were notaffected by race, athletic status, or musculoskeletal development. Weconclude that body composition estimates from DEXA are accuratecompared with those from a four-component model in young adults whovary in gender, race, athletic status, body size, musculoskeletaldevelopment, and body fatness.

     

Effect of race and resistance training status on the density of fat-free mass and percent fat estimates.

The impact of race and resistance training status on the assumed density of the fat-free mass (D(FFM)) and estimates of body fatness via hydrodensitometry (%Fat(D)) vs. a four-component model (density, water, mineral; %Fat(D,W,M)) were determined in 45 men: white controls (W; n = 15), black controls (B; n = 15), and resistance-trained blacks (B-RT; n = 15). Body density by hydrostatic weighing, body water by deuterium dilution, and bone mineral by dual-energy X-ray absorptiometry were used to estimate %Fat(D,W,M). D(FFM) was not different between B and W (or 1.1 g/ml); however, D(FFM) in B-RT was significantly lower (1.091 +/- 0.012 g/ml; P < 0.05). Therefore, %Fat(D) using the Siri equation was not different from %Fat(D,W,M) in W (17.5 +/- 5.0 vs. 18.3 +/- 5.4%) or B (14.9 +/- 5.6 vs. 15.7 +/- 5.7%) but significantly overestimated %Fat(D,W,M) in B-RT (14.0 +/- 5.9 vs. 10.4 +/- 6.0%; P < 0.05). The use of a race-specific equation (assuming D(FFM) = 1.113 g/ml) did not improve the agreement between %Fat(D) and %Fat(D,W,M), resulting in a significantly greater mean (+/-SD) discrepancy for B (1.7 +/- 1.8% fat) and B-RT (6.2 +/- 4.3% fat). Thus race per se does not affect D(FFM) or estimates of %Fat(D); however, B-RT have a D(FFM) lower than 1.1 g/ml, leading to an overestimation of %Fat(D).     

Assessment of skeletal muscle mass in men with spinal cord injury using dual-energy X-ray absorptiometry and magnetic resonance imaging.

The purpose of this study was to determine whether the proportion of skeletal muscle in the fat-free soft tissue mass (FFST) is the same in men with spinal cord injury (SCI) and able-bodied controls. Skeletal muscle mass and FFST of the midthigh were determined by using magnetic resonance imaging and dual-energy X-ray absorptiometry, respectively, in men with long-term (>2 yr) complete SCI (n = 8) and able-bodied controls of similar age, height, and weight (n = 8). Muscle mass (1.36 +/- 0.77 vs. 2.44 +/- 0.47 kg) and FFST (1.70 +/- 0.94 vs. 2.73 +/- 0.80 kg) were lower in the SCI group than in the controls (P < 0.05), but the lower ratio of muscle to FFST in the SCI group (0.80 +/- 0.09 vs. 0.91 +/- 0.10, P < 0.05) suggested that they had a lower proportion of muscle in the FFST than in controls. This notion was supported by analysis of covariance, in that the mean muscle adjusted to the mean FFST of the groups combined was lower in the SCI group. Despite the lower proportion of muscle in the FFST of the SCI group, the relation between muscle and FFST was strong in the SCI group (r = 0.99) and controls (r = 0.96). The findings suggest a disproportionate loss of muscle in the paralyzed thighs after SCI relative to other nonfat constituents, which may be accurately estimated in men with long-term SCI by dual-energy X-ray absorptiometry if the lower proportion of muscle in the FFST (approximately 15%) is taken into account.     

Muscularity and the density of the fat-free mass in athletes.

The purpose of this study was to use estimates of body composition from a four-component model to determine whether the density of the fat-free mass (D(FFM)) is affected by muscularity or musculoskeletal development in a heterogenous group of athletes and nonathletes. Measures of body density by hydrostatic weighing, body water by deuterium dilution, bone mineral by whole body dual-energy X-ray absorptiometry (DXA), total body skeletal muscle estimated from DXA, and musculoskeletal development as measured by the mesomorphy rating from the Heath-Carter anthropometric somatotype were obtained in 111 collegiate athletes (67 men and 44 women) and 61 nonathletes (24 men and 37 women). In the entire group, D(FFM) varied from 1.075 to 1.127 g/cm3 and was strongly related to the water and protein fractions of the fat-free mass (FFM; r = -0.96 and 0.89) and moderately related to the mineral fraction of the FFM (r = 0.65). Skeletal muscle (%FFM) varied from 40 to 68%, and mesomorphy varied from 1.6 to 9.6, but neither was significantly related to D(FFM) (r = 0.11 and -0.14) or to the difference between percent fat estimated from the four-component model and from densitometry (r = 0.09 and -0.16). We conclude that, in a heterogeneous group of young adult athletes and nonathletes, D(FFM) and the accuracy of estimates of body composition from body density using the Siri equation are not related to muscularity or musculoskeletal development. Athletes in selected sports may have systematic deviations in D(FFM) from the value of 1.1 g/cm3 assumed in the Siri equation, resulting in group mean errors in estimation of percent fat from densitometry of 2-5% body mass, but the cause of these deviations is complex and not simply a reflection of differences in muscularity or musculoskeletal development.     

Relation of bone mineral density and content to mineral content and density of the fat-free mass.

Differences in the mineral fraction of the fat-free mass (M(FFM)) and in the density of the FFM (D(FFM)) are often inferred from measures of bone mineral content (BMC) or bone mineral density (BMD). We studied the relation of BMC and BMD to the M(FFM) and D(FFM) in a heterogeneous sample of 216 young men (n = 115) and women (n = 101), which included whites (n = 155) and blacks (n = 61) and collegiate athletes ( n = 132) and nonathletes (n = 84). Whole body BMC and BMD were determined by dual-energy X-ray absorptiometry (DXA; Hologic QDR-1000W, enhanced whole body analysis software, version 5.71). FFM was estimated using a four-component model from measures of body density by hydrostatic weighing, body water by deuterium dilution, and bone mineral by DXA. There was no significant relation of BMD to M(FFM) (r = 0.01) or D(FFM) (r = -0.06) or of BMC to M(FFM) (r = -0.11) and a significant, weak negative relation of BMC to D(FFM) (r = -0.14, P = 0.04) in all subjects. Significant low to moderate relationships of BMD or BMC to M(FFM) or D(FFM) were found within some gender-race-athletic status subgroups or when the effects of gender, race, and athletic status were held constant using multiple regression, but BMD and BMC explained only 10-17% of the variance in M(FFM) and 0-2% of the variance in D(FFM) in addition to that explained by the demographic variables. We conclude that there is not a significant positive relation of BMD and BMC to M(FFM) or D(FFM) in young adults and that BMC and BMD should not be used to infer differences in M(FFM) or D(FFM).