Basic Nutritional Investigation

Evidence of the validity and accuracy of dual x-ray absorptiometry (DXA) to measure in vivo body composition is limited. We compared DXA estimates made in prone and side positions with measurements of chemical composition of 20 pigs (10 barrows and 10 gilts) weighing 52–113 kg. DXA yielded similar estimates of body composition in prone and side positions. DXA estimates of body composition were significantly correlated with reference compositional values (r 5 0.927–0.998). No significant differences were found for determinations of body weight, fat mass (FM), fat free mass (FFM), bone-free, and fat-free mass (BFFFM) between DXA and chemical determinations. DXA significantly underpredicted percent fat (% fat); it underestimated FM (20%, P . 0.05). and overestimated FFM and BFFFM (6 and 9%, respectively, P . 0.05). Differences between individual determinations of FM and % fat by chemical analyses and DXA were significantly correlated with mean values. No significant correlations were found between the differences for weight, FM, % fat, FFM, and BFFFM and measurements of carcass breadth (19–28 cm) and width (15–25 cm). Total errors in determination of DXA body composition variables were similar with body thicknesses less than and greater than 24 cm. These findings indicate that DXA is a valid and accurate method for determination of soft tissue composition. Initial problems with DXA determinations of % fat apparently have been reconciled partially with revisions in soft tissue analytic software. Nutrition 1999;15:697–703. ©Elsevier Science Inc. 1999