*3.1. DXA*

The following practical considerations when using DXA to assess body composition are important to generate valid and reproducible data: (1) machine and software types, (2) legal and ethical considerations and (3) technical standardisation.

## 3.1.1. Machine and Software Types

As highlighted earlier, DXA was originally developed to measure BMD within the general population and therefore quantifies FM and LM as a secondary measure. DXA scanning assesses the composition of photographic pixels, directly distinguishing bone (usually 40–45% of pixels) from other soft tissues (FM and LM). The remaining pixels are used to calculate the remaining soft tissue using a FM:LM ratio [75]. Where no bone is present, the ratio of the attenuation of the two beams is linearly proportional to fat within the soft tissue, with this relationship used to estimate FM and LM, respectively [76]. There are also variations between manufacturers, such as energy levels emitted, pixel size, beam path, software algorithms and scanning frequencies [67].

A further consideration often not considered are both inter-machine and inter-manufacturer variability, causing issues with athletes who are required to travel and may be scanned at different locations. Additionally, even when the same model of DXA machine is used, results may vary significantly between hardware and software version. For example, Table 2 shows data collected on elite female soccer players assessed on units produced by the same DXA manufacturer, ye<sup>t</sup> with different models. These measurements were made within weeks of each other and despite players reporting with the same total BM, there was an 18% increase in FM and a 4% decrease in LM. Furthermore, the algorithm used within differing software packages can be modified with numerous iterations available for scan analysis, all of which may produce conflicting values. Within the published literature it is not common for authors to report which algorithm they have used for analysis, rather, only the make and model of machinery used. Furthermore, the software on DXA scanning systems often only allow allocation of athletes to a single racial classification, thereby not allowing further options for individuals who may be of multi-ethnic backgrounds. Such issues do not detract from the use of DXA; however, they must be taken into consideration when comparing athletes who may have been scanned at differing sites and even at the same site if the software on the machine has been updated.


Dual-energy X-ray absorptiometry (DXA); bone mineral content (BMC); skinfolds (SF).

**Table 2.** Example of the differences observed in practice using real-world data derived from English Premier League and Women's Super League soccer players. This
