BC methods

The most commonly used BC methods were skinfolds (with and without %BF formulae). Owing to the low-cost, convenient data collection of skinfolds as a field method,2 it is not surprising that this method was used more than any other method. Measuring skinfolds is practical and requires limited equipment, but reliable scores can be expected only if technicians are well trained and protocols are standardised.2 ,4 The paper by Lohman4 illustrates that for one BC method, large interinvestigator variation and investigator measurement error occur due to calliper and equation employed. Skinfold measurements produce more variability if %BF is calculated using one of the numerous equations available, which often can lead to confusion, even though the %BF figure may be easier for athletes and coaches to grasp. Equations are population-specific2 and because much of the developmental work was carried out several decades ago using densitometry as the criterion, it is likely there is no equation available for the current sample of interest. An alternative to these approaches includes that which is currently followed by ISAK. Indeed, ISAK has carefully controlled landmarking for locating measurement sites and provides training in technique in an effort to constrain measurement error as far as possible. While the ISAK methodology does not preclude the use of the raw data in estimating density and %BF, it emphasises the value of raw data as individual skinfolds, skinfold sum or skinfold ratios. As a result, this approach eliminates the additional error associated with converting skinfolds into body density and %BF. Further, ISAK training is mandatory at each of its levels, with more stringent error targets for both intratester and intertester error at higher levels. While intratester error is frequently reported as justification for the precision of measurements from a laboratory, in practice, ISAK is the only approach which systematically reports intertester error—where a candidate technician's data are compared with those of an anthropometry instructor or criterion anthropometrist.5 At any rate, skinfolds remain limited in validity when used to predict %BF both by the reference method used and its assumptions, and also by how representative the sites chosen are of total body adiposity. The method and skill of technicians in acquiring skinfolds is pivotal, and while many laboratories report low intratester errors, intertester errors are more important, and standardisation of technique is fundamental. In the current study, ISAK methodology was reported to be used as frequently as other skinfold methods overall, but ISAK appeared to be used significantly more often in Oceania and Europe than in the Americas (including Canada, the USA and South America) and ISAK was also more often used in weight-sensitive sports than non-weight-sensitive sports. ISAK was first established in 1986, and to date has provided training to more than 5000 anthropometrists in over 50 countries, and remains the only internationally recognised certification scheme for anthropometry. While ISAK training and certification are offered world-wide, the availability of courses may be limited by certified ISAK specialists, and thus not all countries offer courses at the same frequency. The observation that the ISAK method is more frequently used in weight-sensitive sports is encouraging, as standardisation of BC methods is especially critical where the risks associated with extremes of BC are high.

Besides skinfolds, DXA, BIA and air displacement plethysmography (ADP) were other frequently reported BC methods. Of those measuring BC in our sample, 38% reported using DXA, with significantly more practitioners using DXA in Oceania than on other continents. The use of DXA for BC assessment was also reported more at the international level and in weight-sensitive sports. This is probably due to the fact that DXA technology is expensive and is often not accessible to athletes at the regional and national level. For athletes competing at the international level, national sport institutes may have DXA available on campus. In the case of DXA, both fat and bone mass can be assessed, and fat-free soft tissue mass predicted by subtraction from total mass. DXA also enables BC to be ascertained in each body segment or separate region of interest, making it one of the most appealing approaches to measuring BC in athletes.2 Against a four-component model, DXA accurately measures BF within 1% of error in ethnically diverse male and female athletes participating in various sports, with errors likely to be greatest at extremes of body fatness and lean tissue hydration.6 Recently, Nana et al7 examined various protocols reported in the literature in DXA studies over the past 15 years. These authors found a lack of standardisation for DXA positioning of participants, giving rise to error. In their article, they present a detailed positioning procedure others may adopt for future studies to allow for improved positioning of arms and legs on the DXA table. Further, DXA measures are thought to be less sensitive to changes in total body water than other measures,2 ,8 although new research suggests a significant influence of food intake, exercise and hydration status on DXA reliability and concludes that to achieve a minimum technical error and to estimate BC changes over time, all DXA measurements should be made in the fasted, euhydrated and rested state following a standardised positioning protocol by the same operator.9 In addition, there can be a loss of accuracy at the extremes of leanness (less than 10% BF), and in the heavier, more muscular individuals10 although software settings (eg, thin mode) may minimise discrepancies in %BF estimates from DXA in very lean individuals.

In the current study, the use of BIA was reported as widespread with 29% of those measuring BC in athletes choosing this method. There was a significant difference in the reported frequency of BIA use among continents, with Europe employing BIA more frequently than any other continent. BIA technology varies greatly: some units use single frequency and some use only upper extremities, lower extremities, while others use the whole body. Some devices incorporate multifrequency using the whole body. In addition, there are many BIA equations and only a limited number have been cross-validated in specific athletic populations.11 Currently, there is no evidence that BIA methodology is valid in athletes, especially considering the lack of generalised equation using a total body water-based, multiple component method.11 Furthermore, limited research supports BIA's validity to track changes in BC over time.12 Most importantly, the sensitivity of BIA to hydration status is much greater than any other BC method2 and variation in hydration status, due to sweat loss and dehydration from exercise in athletes has a significant effect on its validity. As with other BC methods, BIA may produce acceptable results if standardised protocols and raw data scores are used in the appropriate equation by body type, sex, age and sport.11

Of those measuring BC, 17% also reported the use of ADP. For this two-component method, body density depends on the given athlete having both normal BMD as well as a normal hydration status at the time of measurement. Variation in both areas can influence the fat free component and this is well documented in athletes,13 ,14 thus limiting the use of ADP and other densitometry techniques such as underwater weighing for the athletic population.2

Assessing BC over time is a common practice of those working with athletes. The results of this study showed that most practitioners select the frequency of BC measures based on individual circumstances (figure 4), but many also reported making 2–4 measurements/year, and some reported more frequent assessments. However, changes in BF content of 1–3% are not reliably detected by any method, so measuring changes by any method more than three or four times a year in the same individual is not recommended. Measurement of BM can be performed more often, where BM loss or gain is expected and needs to be monitored. As previously mentioned, most of the respondents to the survey reported that problems can arise with the focus on BM and BC in sport. Frequent BM and BC assessments were highlighted as a problem, as this can undoubtedly encourage the athlete to focus more on BM and BC than on performance. While some sports require frequent BM and BC assessment even during competitions, measuring BC on or close to the day of competition is discouraged as this could affect an athlete's mental make-up and negatively influence performance.

Standardisation of BC methods

Measuring hydration status is one element of standardisation among others (eg, pretesting protocol, calibrated equipment, certified tester; see tables 3 and 4) reported by survey participants. It was surprising to find that only 42 respondents reported the use of hydration measures prior to BC assessment. Considering the likelihood of non-responders to this question also not assessing hydration status, in all probability, less than one-fifth of our sample reported measuring hydration status prior to BC assessment. While data show that hydration status at the time of BC measurement has a smaller effect on skinfolds and DXA8 ,15 ,16 for BIA and ADP, hydration status can decrease the accuracy of BC estimates17 and is a major limiting factor when applying these methods to athletes.

Also critical to standardisation is the tester. It is apparent from the survey that sports scientists and dietitians were the professionals most commonly assessing BC, with the sports scientists more often measuring athletes competing at the international level and especially in Oceania. This may be due to the interdisciplinary approach to sport science and medicine at the elite level. Sport dietitians and sports scientists were also more likely to measure BC in weight-sensitive sports. Currently, BC certification exists only for ISAK, and thus there is concern regarding the various approaches to standardisation, even within one method such as DXA. Thus, publications such as the one by Nana et al9 appear essential and are needed also for other BC methods. The finding that others, including coaches, undertake measurement of BC was seen by some respondents as a major concern because of the potential for problems related to BM and BC testing in sports (box 2). While coaches, including strength and conditioning staff, may be equally trained in BC assessment as others in the sport science team, there is still a concern if BC is measured by the coaching staff. Ideally, BC assessment should be part of sport science assessment and best be linked to a sport dietitian who can directly follow-up with the athlete to assist with nutritional strategies if data indicate BC manipulations (see box 1). If coaches are directly involved with BC/BM assessment there must be careful consideration given to the sensitivity of such an environment and possible risks should be minimised. Further, BC is more often regarded as a performance-modifying tool than as a factor also affecting general health. There is no doubt that BC is a performance-determining factor in many sports; however, the dialogue less likely includes possible health risks associated with the focus on BC manipulation in sports. In addition, BC does not assume equal prominence with other physiological tests of athletes, especially those that are sport-specific and conducted frequently across the periodised training plan. Health professionals and coaching staff should treat each athlete's BC assessment, and subsequent goal setting and training/nutrition-related strategies as unique and a mindful risk assessment each time BC is measured is necessary by those testing the athlete's BC to reduce risk. Finally, integrating BC data with other performance-related parameters as well as subjective input from the athlete will likely offer the most diverse feedback to the interdisciplinary team working towards the common goal of athlete success.

From referral for BC assessment to data dissemination

Another aspect of standardisation relates to the process of BC requests and data dissemination. Various scenarios were reported from initiation of BC requests to what data get reported and to whom. Almost half of the respondents did not answer the question related to the dissemination of results. It may be that this depends on each situation or that those testing are also the practitioners reporting the information back to the athlete. However, it may also mean that there is no clearly identified pathway for providing a report of results to the athlete.

Minimum weight

Both the sum of skinfolds (ISAK) and percentage of fat from selected skinfold equations that have been cross-validated are widely used to determine BC and associated minimum weight. Since the minimum %BF that is appropriate for both optimum performance and health varies between and within athletes, either approach (a minimum level of sum of skinfolds or a minimum %BF) may be appropriate, provided certain safeguards of equipment and professional training are followed. Limitations in current practice, including a lack of training in skinfold measurements and using generalised or inappropriate %BF formulae for a given athlete, appear prevalent in the field. Much of the current variation could be minimised with appropriate (ISAK) training, use of discrete skinfolds or use of combinations including all eight ISAK sites18 or for %BF to be derived from skinfolds using a formula validated on athletes such as the Jackson-Pollock sum of four or seven skinfold equations with a certified Lange calliper.19

This survey set out to identify BC methods in current use and their standardisation, and to gather information on data dissemination pathways in the sports medicine/science community around the world. It is apparent from the results that great variability exists in all of these constructs worldwide. Almost all respondents felt that problems exist with respect to the focus on BM and BC in sport. Several respondents were, at the time of the survey, working for national institutes of sport, Olympic training centres, universities and professional sport clubs. Box 2 and table 6 show the most frequent problems and solutions identified by the participants. Several also mentioned that there were barriers to implementation. Some of these barriers included economic, personnel (if coach and medical personnel are not committed) and regulatory issues.

While this original research assists with formulating conclusions about the current use of BC methods around the world, the Ad Hoc Research Working group proposes recommendations based on this and previously published papers2 related to BC, health and performance in sport, and these are summarised below.

Recommendations

The Ad Hoc Working Group on Body Composition, Health and Performance was established to explore opportunities to reduce unhealthy practices related to weight management in sport. This study aimed to describe current BC practices used around the world and to complement solutions recommended by responding experts. Based on the study's results, standardisation with regard to (1) method, (2) tester, (3) frequency, (4) hydration testing along with other health/performance parameters and (5) pathways of requests for BC assessment to data dissemination is needed. If a minimum weight is set, using targets for %BF and/or sum of eight skinfolds (ISAK), it is essential to align methods and related standardisation protocols regionally, nationally and internationally. How sporting organisations deal with BM and BC assessment and pathways from requests for BC testing to data dissemination, including communication to athletes, should be described in internal policy papers, as these issues appear critical in raising the risk of unhealthy practices related to BM and BC in sport. Further, organisations, on the local and global level, should aim to develop new resources and share existing educational materials to improve knowledge surrounding BM and BC and their impact on performance and health for everyone involved (ie, athlete, coach, parents and sport science and medical personnel). In addition, training of all sports medicine/science and coaching staff should include increased awareness and understanding of risk factors leading to unhealthy weight management behaviours and the ability to identify athletes at risk, as well as to initiate appropriate referral and management processes. Thus, scientists and practitioners in the field of BC have an educational mandate for shifting the focus from the rather troubling weight management approaches to a more robust BC management concept in sport. Facing current challenges and identifying new solutions, and advancing research to identify the validity and reliability of emerging BC assessment field techniques, such as ultrasound, are all part of this BC management concept in the near future.2 ,20 ,21

At present, if all practitioners were certified to use either the skinfold method or DXA method (subject to standardisation of hydration status and other potential errors) to estimate values of %BF or skinfold thickness in athletes, there would undoubtedly be much less variation in the results obtained. However, it should be clear that there is currently no good method for the assessment of BC in the field with better accuracy than 3% under the most standardised conditions. Advancement of field BC methodology today pertains mostly to its standardisation (eg, ISAK) and the development of athlete-specific and sport-specific normative data, as carried out by many national sport institutes. Furthermore, laboratory techniques, such as DXA, can give precise assessments of %BF if used repeatedly with the same system in the same laboratory2 and using a standardised protocol.9 Moreover, the assessment of other parameters pertaining to performance and the general health of the athlete should be included in any assessment to minimise the possibility of misinterpretation of BC results. Progress related to both field and laboratory BC assessments is being achieved world-wide, but practice varies widely among countries and institutions. Thus, provided the limitations of commonly used BC methods are known and appreciated, BC assessment remains highly useful in sport, especially when standardised techniques in lab (ie, DXA) and field (ie, anthropometry) are employed and other data are integrated when interpreting results and developing intervention strategies.

The use of minimum levels of %BF is a critical goal to ensure health and safety in certain sports, but its implementation is complex. Earlier work22 proposing % BF thresholds for athletes to be healthy, has been widely accepted, but several important factors have since become apparent. The tools for BC measurement are neither as precise nor accurate2 as had been assumed, and individual differences inevitably raise concerns about regulations that potentially exclude a healthy athlete from competition. Minimum BF levels of 5–7% in men and 12–14% in women used most often are based on the science of BC methodology and are generally accepted by most practitioners as a condition for participation in competitive sport. However, these numbers are purely arbitrary, because they are not based on an established reference method which would enable investigators to distinguish between these values. At best, the accuracy of the four-component model, as discussed by Ackland et al2 is roughly 1.5% when standardised, while field methods, such as skinfolds or BIA, are worse at 3.5% and when not standardised at 5% and higher. Thus, from a practical point of view, the accuracy of field methods may only be around 4%, which makes it highly problematic when setting minimum values for %BF in any athletic individual or sporting group. An athlete's individual health and performance profile should dictate his/her optimum weight where a minimum value for %BF is appropriate for the athletic event. Thus, 7% BF may be the minimum value estimate for one athlete, while for another 6% or 5% may allow enhanced performance while preserving the health of the athlete. Whereas, in anatomical terms, the physique of an athlete might track the periodisation of the training programme, there is a time-lag to this process, and care must be taken when aligning BC measurements to performance. As the optimal fatness for performance is generally less than that for optimal health, there is room for considerable variability in BF levels throughout the training year. Ultimately, many athletes will be walking the tightrope between optimum training and overtraining, and consider short-term performance to a greater extent than longer term health.

Further, few studies show enhanced performance when athletes reduce BM and fat, due to the often concomitant loss of lean mass.23 However, BM and fat loss can be successfully achieved in the absence of compromised lean mass and performance when using a supervised diet and resistance exercise intervention.24 Thus, BC management in athletes can be achieved effectively and safely, but it usually requires a long-term commitment. In the short term, weight fluctuations are likely to reflect water and glycogen changes in athletes, underscoring the importance of standardising nutrition and hydration status when assessing BC.

Leanness is an important performance-determining factor in sports; however, excessive leanness should not be viewed as a performance-enhancing factor by athletes, coaches or staff. There are many examples of athletes, where extreme leanness may have increased performance over the short term, but in many cases resulted in illness when training and competition continued at high levels. In fact, athletes such as those highlighted in the ‘Famous Sufferers: Eating Disorders in Athletes’25 are not exceptions, and there are athletes in many sports (eg, ski jumping, running, gymnastics, figure skating, Nordic skiing, triathlon, cycling) in which specific examples could be named. Having appropriate tools and a professional framework, including standardisation for assessing BC, policies for screening, detection and management of underweight athletes, and educational schemes that de-emphasise excessive leanness and low BM already in young athletes, therefore, represent key elements in the wider domain for both sports performance and athlete well-being.

BC measures are ideally scheduled systematically throughout the year during the planning phases. To avoid a high frequency of measures, a general guideline would include at least 2–6 months between measurements, depending on the method and specific athletic situation, with DXA measurement administered no more than two times per year and preferably not close to or during competition unless needed as part of the qualification process (ie, weight category sports). Further, specific BC requests should not be initiated by the coach or the athlete but rather originate from within an interdisciplinary team and/or preparticipation physicals. Moreover, the interdisciplinary team should also delineate what information (eg, fat and lean variables) is reported back to the athlete and in what format (report, verbal, email) and what information may go to the coach. Finally, the pathway from BC assessment requests to data dissemination should clearly outline who reports the data back to the athlete. In the case of ISAK, we recommend this to be an ISAK-certified sports scientist or sport dietitian, as they may evaluate changes over time based on their known intratester reliability and established sport-specific database. It is important to appreciate individual performance differences in relation to an athlete's optimum BC. Excessive leanness might compromise health and performance in one individual, while the same composition in another individual might enhance performance without adversely affecting health.

Limitations

Survey research has various limitations, which are acknowledged. First, data were collected from a convenience sample and we recruited those individuals inherently interested in BC assessment in athletes. Second, survey research may also be biased due to the subjective nature of responses. This survey provided an option for respondents not to answer a question and move on to the next area, which affected the number of responses for some questions. Owing to the nature of the topic and the anonymity of the survey, deliberate false answers were not expected. Many respondents added extra information as expert opinion and many respondents expressed encouragement regarding the urgency of this topic. Thus, there is little reason to doubt the validity and reliability of the responses. The results of this study apply the use of BC methods of practitioners working with athletes and are not generalisable to other populations. Finally, this survey was sent out in English only, which may have limited the ability of personnel in some countries to respond.

Summary

This study assessed the current use of BC methodology in practitioners from sport organisations around the world. The results show great variability in BC assessment methods and a lack of standardisation. Most respondents also reported that the current focus on BM and BC in sport is problematic and has the potential to increase the risk of unhealthy weight management practices. This study was conducted by the Ad Hoc Working Group on Body Composition, Health and Performance under the auspices of the IOC Medical Commission. We recommend better standardisation of BC methods, more opportunities for training, greater access to resources to increase awareness and understanding of risk factors leading to unhealthy weight control practices and more research related to BC and performance.