3.1.3. Technical Standardisation Requirements

Despite DXA scanner hardware/software limitations and the legal/ethical considerations, whole-body DXA measurements can be a reliable measure of body composition in athletic populations when standardised protocols are used [79], with technical error of measurements of ~0.1% for total mass, ~0.4% for total LM, ~1.9% for total FM and ~0.7% for total BMC [69]. Whilst measurements should always be taken from the whole body within the limits of the bed, this can be problematic with athletes such as rugby players, who may be both wider and taller than the bed. Taller athletes cannot always fit within the ~195 cm limit of the scanning area [80], which can significantly alter the outcome results. An example of this is indicated in Table 3, where a 201 cm tall athlete was placed on the DXA scanner on the same day in a number of positions including head off/feet on, head on/feet off and head and feet on, but with knees bent to 90 degrees. These varying positions altered FM by over 3 kg and total body mass by 7 kg, with BF% ranging from 13.8% to 16.6% (Table 3). Indeed, DXA may fail to accurately assess athletes who are often considered within the 'extremes' of physiology, such as tall and extremely muscular athletes. Specifically, historical data have shown issues with those who have a chest depth >25 cm [81] and particularly lean athletes, where negative fat values were found for the torso [82]. Given that many sports teams often set arbitrary non-discriminatory BF% targets for athletes (often 15% with sanctions in place when this is not achieved), positioning can lead to errors and thus major consequences for the athlete. To date, there is no approved standardisation procedure for tall athletes (in our experience most teams opt for head off feet on) and we would sugges<sup>t</sup> that, as a minimum, this needs to be considered and reproduced for each scan, whilst future studies should establish best practice for such situations.

Table 4 summarises a number of emerging technical considerations that must be accounted for and/or controlled to generate reliable DXA body composition data. It is important that athletes have fasted overnight when being scanned, which can be challenging when assessing large squads of individuals [79]. Given that a typical DXA scan can take 10–15 min per individual, when testing an entire sporting squad in one day it is possible that some athletes may need to be fasted until the late afternoon, which is often not feasible. Given that it is unusual for a sports team to own their own DXA unit, assessment can involve visiting a local University, which poses challenges when ensuring pre-scanning standardisation procedures. Indeed, studies have suggested that eating a meal can lead to as much as a 2.6% increase in FM [83]. Another key consideration is the effect of muscle glycogen on the reliability of DXA body composition measures. It is now common practice for athletes to periodise carbohydrate intake, commencing some training sessions with low and often competing with high muscle glycogen, utilising the

'fuel for the work required' concept [84–87]. Glycogen depletion significantly affects DXA results, with a 2.5% increase in LM when glycogen super compensated [83]. Conversely, this could have major implications for athletes who may feel they have lost LM, when in reality they may simply have depleted glycogen stores due to training. Similarly, creatine loading can have major effects on the reliability of the data obtained [88]. Other factors to consider include time of day, hydration status and previous exercise activity [83,89]. Careful planning is required for the use of DXA to give the best possibility of reproducible data, which from an applied perspective is often difficult to control on a regular basis. Perhaps the strongest example of the effects of diet and exercise on DXA data veracity is in a series of case studies from our research group, summarised in Table 4. These reports involved professional combat sport athletes 'making weight' for competition utilising acute weight loss (AWL) strategies, inclusive of the manipulation of carbohydrate intake and hydration status. Post AWL, LM and FM was reduced by 17.5% and 10.4% respectively, across a 4 day period of intense energy and water restriction. These values rebounded by 25.4% and 40.6%, respectively following a two week period of rehydration, refuelling and recovery, inclusive of total cessation of training activity [9]. In addition, when assessing another weight making athlete one day prior and following weigh in, LM and FM were both reduced by 4.0% and 5.8%, followed by a 10.0% and 4.6% rebound [10]. These values are physiologically impossible and highlight the effects that acute changes in nutrition and exercise status can have on DXA body composition data, further illustrating the importance of pre scan standardisation to enhance reproducibility of the outcome data. It is crucial coaches are aware of these limitations and create conditions where such controls can be implemented. Indeed, if it is not possible to implement such controls, it would seem unethical to subject an athlete to a radiation dose, albeit a very low concentration, if the data produced have limited validity and could result in inaccurate practical outcomes. It could therefore be argued that if it is not necessary to assess BMC and/or limb-specific measurements of FM and FFM, then skinfold thickness measures may provide a suitable and more reliable alternate in free-living conditions. Where using DXA as a technique to measure body composition, it is imperative to ensure standardised, best practice protocols are followed [69] and all information considered within the Committee on Medical Aspects of Radiation in the Environment's (COMARE) report considered [90].

**Table 3.** An example of failure to fit different athletic body types within the confinements of the DXA bed and how this affects results: (a) head on, feet off; (b) head off, feet on; (c) head on, feet on with 90◦ bend of the knee.


All measurements collected one after the other by one of the research team. dual-energy X-ray absorptiometry (DXA); bone mineral content (BMC).


**Table 4.** A summary of differences in DXA scan results over the course of a habitual day [83]; following creatine supplementation [88]; exercise activity [89]; glycogen
