A Mathematical Method to Estimate Body Composition in Tall Individuals Using DXA

Abstract

The purpose of this investigation was to determine whether whole body composition could be estimated in tall individuals using summed measures from two partial DXA scans. Using a pencil-beam DXA, a convenience sample of young individuals (N = 19) were scanned three times. Two partial scans of the upper and lower body were combined to predict bone area, bone mineral content (BMC), bone mineral density (BMD), fat mass (FM), fat-free soft tissue mass (FFST), and percent fat (%Fat), from a complete whole body scan. Two different methods, dividing the body at the hip (HIP) or at the top of the shoulders (NECK), were used to determine optimal method of summing. There were no significant differences in BMC, BMD, FM, FFST, and %Fat comparing the complete scan and summed scans regardless of summing method. There was excellent agreement between complete and summed values of BMC, BMD, FM, FFST, and %Fat values as indicated by regression analysis (r value range: 0.992-1.00) and nonsignificant intercepts. Bland-Altman plot analysis indicated that a small systemic bias occurred in estimates of FM and %Fat using the HIP method and FFST using the NECK method; however, all biases were negligible. The findings suggest that accurate estimates of whole body composition from a complete scan can be obtained by summing two partial scans. Although both summing methods provide good estimates of body composition, dividing the body at the neck provides more accurate estimates of bone and soft tissue composition than dividing the body at the hip.