Body weight and body composition are important performance factors in many sports (23). Low body weight and weight regulations have been associated with athletes competing in weight-sensitive sports (5,27), and the use of extreme weigh-loss methods has been related to health risks for different athletic groups (2,19,20,26,32).
Estimates of body composition characteristics are used in both athletic and nonathletic populations to identify health status (15). Quantifying body composition is an important assessment parameter for monitoring performance, training, and dieting regimens. In recent years, the use of body mass index (BMI (kg·m−2)) and, in particular, assessment of body fat to determine optimal body weight for athletes has increased. BMI is often used in studies investigating the association between body composition and biological markers in elite athletes (14,17) and as a cutoff value for allowing athletes to compete in different sports (16,29,33). Furthermore, calculation of the BMI is recommended by Bonci et al. (4) as the method for assessing body composition in “National Athletic Trainers’ Association Position Statement: Preventing, Detecting, and Managing Disordered Eating in Athletes.” The use of body weight by itself and/or BMI has, however, been criticized, especially in athletic populations (18,21,22). Ode et al. (21) concluded that BMI should be used cautiously when classifying fatness in college athletes and nonathletes. The BMI value does not discriminate between the different components of the body, and therefore, an individual with high fat-free mass (FFM) relative to height might have a high BMI value but not be obese. On the contrary, individuals may have BMI values in the normal range and be classified as nonrisk subjects in terms of health risk and mortality, while they simultaneously have unhealthy low levels of body fat (18,22). Furthermore, BMI does not consider shapes or leg lengths that differ from the norms (16). Despite these known limitations of BMI, it is still used to assess body composition in athletes, and it is even the basis of a rule in some sports (16,29,33).
In ski jumping, athletes with a BMI value below 20.5 and 21 kg·m−2 (including jumping boots and jumping suit) for male and female jumpers respectively, are issued skis that are shorter in proportion to their BMI. Shorter skis reduce aerodynamic lift to compensate for the advantage a lighter jumper would otherwise gain (16).
The use of measurement methods with poor validity and/or reliability has made it difficult to provide accurate body composition data in athletes. Furthermore, reliable and valid data related to body composition in elite athletes representing many different sports are lacking. Knowledge of the typical body composition of athletes in a specific sport group is helpful in determining suitable body composition or target weights, evaluating the effects of the training and/or eating behavior, deciding an optimal competitive weight class, and making sure the athlete is within a healthy body fat range. Considering the fact that dual-energy x-ray absorptiometry (DXA) measures bone mineral composition and density, DXA has been suggested as preferable to either hydrostatic or air displacement plethysmography 2-component models as a reference method for assessing body composition in athletes (9). To our knowledge, no studies have measured both BMI and percent body fat using DXA in randomly selected groups of female elite athletes representing all sports and nationally randomly selected nonathletic controls.
The World Health Organization (WHO) has, on the basis of statistics of weight and height of the general population, defined the different cutoff limits for BMI (34). This classification is based primarily on the association between BMI and mortality (34). Similar guidelines for body fat levels are lacking. When it comes to athletes, it has been recommended, on the basis of Behnke’s (3) theoretical concept of minimal body mass, that a body fat level below 12% may be unhealthy for females (8,28). In the other end of the body fat spectrum, Gallagher et al. (6) have stated that 33% is an acceptable cutpoint for overfatness for females, corresponding to a BMI of 25 kg·m−2 in young African-American and white adults age 20–39 yr. However, data related to the prevalence of female elite athletes classified with unhealthy low or unhealthy high body fat levels are lacking, and furthermore, more information about what characterizes the athletes with unhealthy low and high percent body fat is needed to guide athletes, sport support health care personnel, and coaches with target weights, healthy body fat levels, and dietary intervention.
Therefore, the aims of the present study were (a) to relate BMI to body fat percentage as measured by DXA in female elite athletes from different sports and nonathletic controls and (b) to investigate what characterizes the athletes with unhealthy low and high body fat values.
This study was conducted in three phases: 1) screening with a detailed questionnaire, 2) measurement of body composition, and 3) clinical interview.
The entire population of female elite athletes in Norway age 13–39 yr (n = 938) and controls in the same age group (n = 900) were invited to participate in the study. All participants in the study gave written consent to participate. We defined an elite athlete as a member of a junior or senior national team or a member of a recruiting squad for those teams. The athletes represented 66 different sports/events. An overview of these sports has been published elsewhere (31). Thirty-one athletes did not receive the questionnaire in the mail because of the fact that they were representing teams in other countries or were traveling, and 149 were excluded (76 had ended their career, 35 did not compete at the national level, 15 were injured, eight were pregnant and did not plan to continue their athletic career after delivery, five were older than 39 yr, nine did not complete the questionnaire satisfactorily, and one athlete competed in two different sport groups). A total of 758 athletes were available for participation; 89 were nonresponders (subjects who, for unknown reasons, never returned the questionnaire), resulting in 669 responders (88.3%).
The control group (n = 900) was selected by the Norwegian Bureau of Statistics from the total population of Norwegian females age 13–39 yr. Twenty-three controls did not receive the questionnaire in the mail because of problems finding their addresses, and 12 were excluded (9 did not understand the Norwegian language, and three were severely ill and were unable to fill in the questionnaire). A total of 865 controls were available for participation; 258 were nonresponders (subjects who, for unknown reasons, never returned the questionnaire), resulting in 607 responders (70.2%). The study was approved by the regional committee for medical research ethics.
Part 1. Screening.
A questionnaire including questions regarding training and/or physical activity patterns; menstrual, dietary, and weight history; oral contraceptive use; and disordered eating behavior was sent to all athletes and controls.
Selection for parts 2 and 3.
On the basis of the initial selection, a random sample of athletes (n = 300) and controls (n = 300) was selected and invited to participate in parts 2 and 3 of the study. This sample was stratified on the basis of age group (13–19, 20–29, and 30–39 yr) and “risk profile” for the female triad (two groups: “at risk” and “not at risk”) (32). No pregnant women were included in parts 2 and 3 of the study. Altogether, 186 athletes (62%) and 145 controls (48%) participated in all parts of the study.
Part 2. Assessment of body composition.
Body composition was measured with DXA (Prodigy version 5.6; GE Lunar Corp, Madison, WI). The coefficient of variance for body composition measurements was 1.0% for fat percentage, 0.4% for FFM, and 0.3% for total body mass. Body height and weight were measured twice, and results are given as a mean.
Part 3. Clinical interview.
The same sample that participated in part 2 also participated in part 3 of the study. A Ph.D. specially trained in eating disorders conducted all the interviews. Participants meeting the DSM-IV criteria (1) for anorexia nervosa, bulimia nervosa, or eating disorder not otherwise specified were categorized as subjects with clinical eating disorders.
BMI and body fat classifications.
In accordance with the WHO (34), a BMI value less than 18.5 kg·m−2 was considered as underweight, between 18.5 and 24.9 kg·m−2 was considered as normal weight, between 25.0 and 29.9 kg·m−2 was considered as overweight, and at 30 kg·m−2 and above was considered as obese.
Because of the lack of similar categories for body fat, we used criteria related to body composition for optimal health. It has been recommended that a body fat level below 12% may be unhealthy for females (8,28) and that 33% is an acceptable cutpoint for overfatness for females, corresponding to a BMI of 25 kg·m−2 in young African-American and white adults age 20–39 yr (6). Therefore, in our study, females with body fat levels below 12% were classified as underfat, whereas females with body fat levels at or above 33% were classified as overfat.
Training and physical activity.
Physical activity was defined as the weekly total hours of training and competition for the athletes (presented as a mean of the training and competition periods during the previous year). The physical activity for the controls was defined as the weekly total hours of physical education, recreational sports, and activities (31).
Other health parameters.
As in our previous studies (31,32), athletes reporting present primary amenorrhea, secondary amenorrhea, oligomenorrhea and short luteal phase, or a lifetime prevalence of primary or secondary amenorrhea were all classified with menstrual dysfunction. The question asking for a history of stress fracture was as follows: “Have you ever had one or more stress fractures?” Pathogenic weight control methods (PWCM) included reported use of one or more of the following methods: diet pills, hunger-repressive pills, laxatives, diuretics, or vomiting. In accordance with earlier publications (31,32), a low bone mineral density (BMD) was indicated by a z-score below −1.0 with a positive diagnosis in at least one of five measurement areas (total body, lumbar spine (L2–L4), femur neck, femur trochanter, or total femur).
Statistical Analysis and Presentation of the Data
After the randomized selection of athletes was conducted for parts 2 and 3 of the study, 46 different sports/events were represented. In accordance with previous studies, athletes were divided into two groups: athletes competing in leanness and nonleanness sports. The former sports included endurance, aesthetic, weight class, and gravitational sports, and the latter sports included technical, ball game, and power sports (30,32).
All analyses were performed using PASW (Predictive Analytics SoftWare) Statistics 18 for Windows (IBM Corporation, Somers, NY). Results are expressed as mean value and SD. Comparisons between groups were carried out using a two-sample Student’s t-test for continuous data and a chi-square test for categorical data. The Fisher exact test was carried out when the cells had expected counts of less than five. All tests were two tailed. Comparisons of the seven different sport groups were carried out using one-way ANOVA. To prevent against type I error, the Bonferroni method of adjustment was used when describing differences between the seven sports groups. Correlation analyses were carried out using Pearson R, and the SEE was calculated to investigate the accuracy of the predictions. Differences were considered statistically significant for P values equal to or less than 5%.
In part 1 of the study, a total of 669 athletes and 607 controls participated. Self-reported mean BMI was 21.6 ± 2.4 and 23.4 ± 4.3 kg·m−2 for athletes and controls, respectively. Among the athletes in part 1 of the study, 7.8% were underweight, 85.1% were normal weight, 6.3% were overweight, and 0.8% were obese according to the WHO classification system. Among the controls, the percentages were 5.0%, 70.3%, 17.6%, and 7.1%, respectively. In this article, only anthropometric data from the clinical part 2 of the study are included in the analysis and presented. Detailed information about data based on part 1 of the study is published elsewhere (31).
Subject characteristics and anthropometric data.
The athletes differed from the controls in all the anthropometric variables, and the athletes representing leanness sports had a lower BMI and body fat percentage as compared with athletes competing in nonleanness sports (Table 1).
Body fat and BMI in athletes representing different sport groups.
Athletes competing in endurance and aesthetic sports had a lower BMI and body fat percentage compared with athletes in ball game sports and controls. Athletes competing in gravitational sports had a lower body fat percentage as compared with athletes competing in technical sports. The controls had a higher body fat percentage compared with athletes in all the different sport groups, except athletes in technical sports (Fig. 1).
BMI and body fat categories.
On the basis of BMI measurements, more athletes than controls were classified as underweight and normal weight (P < 0.01), whereas more controls than athletes were classified as overweight and obese (P < 0.001) (Table 2). A total of 63.4% of the controls had a body fat percentage at or above 33%, and this was more than in the athletic group (11.3%) (P < 0.001) (Table 2).
Of those athletes with normal BMI values (n = 150), 2.0% (n = 3) were classified with low body fat levels (<12%), and 6.7% (n = 10) were classified with obese body fat levels (≥33%). The median value was 24.3% body fat (Fig. 2). Of those athletes classified as overweight (n = 17), 58.8% had a body fat percentage at or above 33%.
For the controls with normal BMI values (n = 96), none were classified with low body fat levels (<12%), and 50% (n = 48) were classified with obese body fat levels (≥33%). The median value was 33.1% body fat (Fig. 3). Of those controls classified as overweight (n = 34), 97.1% had a body fat percentage at or above 33%.
Correlation between BMI and body fat.
The correlation between BMI and body fat percentage was 0.740 (P < 0.01) for all subjects, 0.671 (P < 0.01) (SEE = 5.3%) for the athletes, and 0.813 (P < 0.01) (SEE = 4.1%) for the controls. For the leanness sport athletes, the correlation was 0.638 (P < 0.01) (SEE = 5.6%), and for the nonleanness athletes, the correlation was 0.559 (P < 0.01) (SEE = 4.8%).
Characteristics of the underfat and overfat athletes.
Of the seven athletes with body fat levels below 12%, four competed in endurance sports, two competed in aesthetic sports, and one competed in a gravitational sport. Of the 21 athletes with body fat levels at or above 33%, 4 competed in technical sports, three competed in endurance sports, 2 competed in aesthetic sports (freestyle dancers), and 12 competed in ball game sports. Between 19% and 71% of the overfat athletes self-reported unhealthy weight loss methods and disorders associated with the female athlete triad. Almost half of these athletes (47.6%) were diagnosed with a clinical eating disorder (Table 3). Among the underfat athletes, 42.9% self-reported stress fractures or use of PWCM, such as diet pills, hunger-repressive pills, laxatives, diuretics, or vomiting, whereas 57.1% were diagnosed with a clinical eating disorder or low BMD. Also, 85.7% self-reported menstrual dysfunction in this group of underfat athletes (Table 3).
BMI and body fat percentage correlated in both athletes and nonathletes. However, the SEE is large, and the discrepancy in the number of athletes and controls characterized as under-, normal-, and overweight by the BMI classification and the number classified with unhealthy low or high body fat levels by DXA reveals that BMI is not a valid measure for assessing or monitoring body composition in female elite athletes. On the basis of these findings, we would argue against using BMI as a measure for assessing or monitoring body composition and/or health or as a selection criterion for participation in competitive events. Furthermore, it should be used carefully in female nonathletes. Finally, females representing both under- and overfat athlete groups reported weight fluctuation, use of PWCM, menstrual disorders, and stress fractures and had low BMD and clinical eating disorders.
Correlation between BMI and body fat.
In accordance with findings from American college athletes and nonathletes (21), we found significant correlations between body fat percentage and BMI in both female athletes and nonathletes. The correlation coefficient was somewhat lower for the athletes (0.671, SEE = 5.3%) as compared with the controls (0.813, SEE = 4.1%) in our study, indicating that the association between BMI and body fat percentage may be better in female nonathletic populations than among competitive female athletes. Furthermore, we found a slightly higher correlation (0.638, SEE = 4.8%) between BMI and body fat percentage for the leanness athletes as compared with the nonleanness sport athletes (0.559, SEE = 5.6%). In this context, BMI should be considered as an indicator of heaviness because BMI describes the quantity of mass and not the quality (7). Therefore, when lean mass is high, as in the leanness group, the only variable that could influence BMI is fat content, suggesting that BMI indirectly can correlate, to a higher degree, with body fat content in this group of elite leanness athletes as compared with values in the nonleanness group. Finally, the SEE for athletes was 5.3%, indicating a high prediction error when BMI is used to estimate body fat for an individual athlete.
BMI and body fat in athletes and nonathletes.
The athletes in our study had a lower weight, lower BMI, lower body fat percentage, and a higher FFM as compared with the nonathletic controls. This is in contrast to the study by Ode et al. (21), where their female college athletes had a higher BMI than the nonathletes. However, the 77 female college athletes in their study participated in basketball, crew, and softball, and they are therefore not directly comparable to our elite athletes representing 46 different sports/events.
We found that our athletes as a group had a 9.6% lower BMI and as much as a 31.8% lower body fat percentage than the controls. The athletes’ FFM constituted 72.4% of the athletes’ total mass, whereas among the controls, only 62.0% of the total body mass was FFM (P < 0.001). These findings support the assumption that the athletic body is constituted of a higher FFM-to–fat mass ratio compared with nonathletes.
A further analysis of the athletic group showed that, as expected, athletes competing in leanness sports had a lower BMI and body fat percentage and a higher FFM as compared with athletes competing in nonleanness sports. As much as 75.5% of the total body mass of the leanness athletes was FFM, which was more than among the nonleanness sport athletes (69.5%). Leanness sport athletes represent sports where leanness and/or a low body weight is considered important for performance, such as aesthetic sports, endurance sports, weight class sports, and gravitational sports. Another investigation of female competitors in different sports, where female body builders, pentathletes (competing in pentathlons), and distance runners have the largest FFM-to–fat mass ratio values, whereas the smallest FFM-to–fat mass ratio values emerge for nonleanness athletes, such as shot-putters and discus throwers, as well as ball game athletes, such as basketball and volleyball players, supports our findings (13).
We also had the opportunity to look further into BMI and body fat percentage data in the seven different sport groups. This classification system is also used by the Norwegian Olympic Training Centre and has been used in several previous studies (27,31,32). When comparing the BMI values, we found that the controls had a higher BMI only compared with athletes competing in endurance and aesthetic sports. In contrast, when looking at the body fat percentage, the controls had higher values compared with six of the seven sport groups. Athletes competing in technical sports did not differ from the nonathletes in terms of body fat or BMI. This is not very surprising because low weight and/or leanness is not considered important for optimal performance in these technical sports (31). Furthermore, when comparing the seven different sport groups, we found that the athletes competing in aesthetic and endurance sports had a lower body fat percentage as compared with athletes competing in technical and ball game sports. Furthermore, the athletes competing in gravitational sports had a lower body fat percentage as compared with athletes competing in technical sports. According to Loucks (12), a sport-specific (or position-specific in team sports) optimal body composition may have a great effect on athletic performance. Optimal body composition is important for performance; however, our results call attention to the differences in body composition in and also within different sport groups.
BMI and body fat categories.
Because of the relatively high prevalence of overweight young females in Norway (10), it was hypothesized that the majority of both athletes and controls would be categorized as normal weight and overweight. These assumptions were, to a large degree, met, but more athletes than controls were classified as underweight and normal weight, whereas more controls than athletes were classified as overweight and obese. It was, however, surprising that as many as 63.4% of the controls had a body fat percentage at or above 33%, indicating overfatness and increased health risk. This is especially alarming when looking at the BMI data for this group showing that only 7.6% of the controls were categorized as obese. A total of 66.2% of the controls had normal BMI values, but half of these had body fat levels at or above 33%. The median body fat value in the normal-weight group was surprisingly as high as 33.1%. These findings underline the importance of measuring body fat more directly to evaluate body composition and health risk.
Also, the finding that more than 1 of every 10 elite athletes (11.3%) was classified as overfat was unexpected. A further concern is associated with the fact that 3.8% of our athletes had body fat percentages below 12%, also indicating a health risk (8,28). These findings show that, in total, as many as 15% of our female elite athletes have unhealthy levels of body fat. The characteristics of these particular athletes are discussed later in this article.
The use of BMI measurement in athletes has been criticized because of the fact that the influence of large muscle mass on BMI may misclassify these athletes as overweight and obese (21,22). Of our athletes with normal BMI values (n = 150), 2.0% (n = 3) were classified with body fat levels below 12%, and 6.7% were overfat (body fat ≥ 33%). The median body fat value in the normal-weight athletes was 24.3%. Of those athletes classified as overweight, almost 60% actually had a body fat percentage at or above 33% and were classified as overfat. The corresponding value for the nonathletes was 97%, indicating that more athletes than controls are falsely categorized as overweight, probably because of a higher FFM-to–fat mass ratio as compared with nonathletes.
To illustrate the limitations in using BMI as the measure for body composition, four cases from this study are presented below. Case 1 had a BMI value of 20.2 kg·m−2 and a body fat percentage of 10.4%. Case 2 had a BMI value of 19.0 kg·m−2 and a body fat percentage of 5.9%. Case 3 had a BMI value of 21.6 kg·m−2 and a body fat percentage of 35.9%, and finally, case 4 had a BMI value of 23.7 kg·m−2 and a body fat percentage of 39.2%. By using only BMI as a measurement method for body composition in female athletes, we take a risk of falsely classify them as “normal” or “healthy,” when the opposite may very well be the truth. As mentioned in the introduction, in ski jumping, a BMI value of 20.5 and 21 kg·m−2 is the cutoff for male and female athletes to be allowed to compete (then equipment is included). Also, these ski jumpers are allowed to cut their skis if they are too thin and do not reach the cutoff for the BMI. The International Ski Federation implemented the BMI system in 2004 to care for the athletes’ health (16). In the aforementioned case (case 2), she would almost be accepted for participation in ski jumping with a fat percentage of 5.9%, although such a low fat mass actually indicates symptoms associated with an eating disorder.
Characteristics of the underfat and overfat athletes.
All seven athletes with body fat levels below 12% competed in leanness sports, whereas of the 21 athletes with body fat levels at or above 33%, 16 competed in nonleanness sports and 5 competed in leanness sports. Almost half of the underfat athletes and more than one of five of the overfat athletes were diagnosed with clinical eating disorders and/or low BMD or self-reported one or more of the following disorders or behaviors: menstrual dysfunction, stress fractures, weight loss, or use of PWCM. These data show that females with disorders such as clinical eating disorders, low BMD, and menstrual dysfunction are represented both in the underfat group and the overfat group. A myth in the sport environment is that athletes with eating disorders are underweight and skinny. Among our overfat athletes, as many as 47.6% of them were diagnosed with a present or past clinical eating disorder, either a bulimia nervosa or an eating disorder not otherwise specified (Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria) (1). The increased level of fat mass in some of these eating-disordered athletes might be due to a history of restrictive eating, a later development of binge eating, and a more bulimic type of eating disorder as suggested in some of our earlier publications (27,30).
Strengths and limitations.
All female athletes representing national teams or recruiting squads for such a team were invited to participate in our study; the response rate was high, and thus, the athletic population should be representative for the target population—female elite athletes. The results from this study should therefore be considered generalizable to other athletes competing at the Olympic and/or World Cup level or who are participants in international or national competitions in a junior or senior national team or as recruits for such a team. Furthermore, in the present study, all European sports were included, and most of these are commonly known also in other parts of the world. Thus, the results concerning the athletic population should be generalizable to national-level female athletes in other cultures or countries.
The random selection of the controls and the fact that few differences were detected between those controls participating in part 1 and those participating in parts 2 and 3 of the study strengthen the assumption that a representative sample participated in all parts of the study. There are no reasons to suspect that the randomly selected and nationally representative sample of controls participating in this study is not comparable to nonathletic females in general in other Western cultures and countries. Thus, it is likely that the results are generalizable to nonathletic females age 13–39 yr in the Western culture. However, it should be noted that the relatively high level of physical activity in the controls may not apply to all other populations and also that genetic differences may be present.
The participation rates in the clinical parts of the study were lower than expected, especially in the control group. One of the reasons might be that the clinical parts of the study were conducted in the capital of Norway (Oslo), and because the control sample was representative for the whole population, some of the subjects had to travel long distances to participate in the study. This was, by many, reported as their main reason for not participating. The athletes also represented the whole country, and many had to travel long distances to participate. In addition, many of the invited athletes were out traveling, competing in other countries, etc., and it was difficult for them to meet for clinical evaluation in Oslo.
A possible consequence of the low participation rate is systematic selection bias. However, because all the subjects invited to participate in the clinical parts of this study already had participated in the questionnaire study, it was possible to conduct a “dropout analysis” comparing the participants and the dropouts regarding many relevant factors. The dropout analysis showed that the only factors that distinguished the control participants from the dropouts were geographical location and height. When looking at the BMI values, which are of specific interest in this study, we found that the mean BMI values for the total sample of elite athletes (21.6 kg·m−2) and controls (23.4 kg·m−2) and the clinically investigated group of athletes (21.7 kg·m−2) and controls (24.0 kg·m−2) were similar, and among the athletes even identical.
Each body composition measurement technique has different advantages and disadvantages in terms of sites that can be measured, clinical utility, radiation dose, availability, cost, and ease of use. DXA is not a direct measurement method, but it is seen as an indirect method where a surrogate parameter is measured to predict tissue composition. When compared with the multicomponent model, results vary from study to study confounded by changes in hardware for DXA, especially fan beam versus pencil beam, and also by company. Various investigators have shown systematic differences of 1% to 4% (11,14,24,25). However, despite these limitations, if body composition assessment is used for health-related questions or performance criteria, the DXA might be used until we have techniques without problems related to methodology or in the assumptions they make.
Finally, although we have clinical data from as many as 186 athletes and 145 controls, it would be interesting to investigate these research questions in larger samples and in other countries.
BMI and body fat percentage were significantly correlated in female elite athletes and nonathletes. The discrepancy in number of athletes and controls characterized as under-, normal-, and overweight by the BMI classification and the number classified with unhealthy low or high body fat levels by DXA reveal that BMI is not a valid measure for assessing or monitoring body composition in female elite athletes, and it should be used carefully in female nonathletes. We found that body composition varies among athletes representing different sports. Both athletes with low (underfat, <12% body fat) and high body fat levels (overfat, ≥33% body fat) reported weight fluctuation, use of PWCM, menstrual disorders, and stress fractures. Because low BMD and clinical eating disorders are common among both under- and overfat (weight) athletes, it is important to look for these serious health problems not only in the skinny athletes but also in athletes representing the whole spectrum of body fat and/or body weight.
We believe that coaches, leaders, athletes, governing bodies, and other important actors in the sport environment have to continue the work related to preventing athletes from practicing unhealthy weight control methods and thereby achieving the expected “competition weight.” To date, there seems to be no universally applicable criterion or “gold standard” methodology for body composition assessment, and it is the authors’ opinion that there is a need for development of improved approaches to body composition assessment. However, BMI should not be used for assessing or monitoring body composition in elite athletes. BMI does not necessarily correctly determine when an elite athlete is too fat or too thin. DXA has limitations but is considered practical and usable for the assessment of body composition in athletes. Thus, there is a need for further research within this area to determine the most valid, reliable, and practical laboratory and field methods for body composition assessment. Finally, we also believe that knowledge of body composition in athletes representing different sports is important in helping medical personnel in their regular surveillance of the athlete’s physical and mental health.
This study was supported by research grants from the Norwegian Olympic Committee and Confederation of Sport and the Norwegian Foundation for Health and Rehabilitation.
The authors thank Professor Ingar Holme for his assistance with the statistical analysis and Professor Timothy G. Lohman for his helpful comments on this article.
There are no potential conflicts of interest.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington (DC): American Psychiatric Association; 1994. p. 539–50.
2. Beals KA, Hill AK. The prevalence of disordered eating, menstrual dysfunction, and low bone mineral density among US collegiate athletes. Int J Sport Nutr Exerc Metab. 2006; 16 (1): 1–23.
3. Behnke AR. New concepts of height–weight relationships. In: Wilson NL, editor. Obesity. Philadelphia (PA): F.A. Davis; 1969. p. 25–53.
4. Bonci CM, Bonci LJ, Granger LR, et al.. National Athletic Trainers’ Association Position Statement: preventing, detecting, and managing disordered eating in athletes. J Athl Train. 2008; 43 (1): 80–108.
5. Byrne S, McLean N. Elite athletes: effects of the pressure to be thin. J Sci Med Sport. 2002; 5 (2): 80–94.
6. Gallagher D, Heymsfield SB, Heo M, Jebb SA, Murgatroyd PR, Sakamoto Y. Healthy percentage body fat ranges: an approach for developing guidelines based on body mass index
. Am J Clin Nutr. 2000; 72 (3): 694–701.
7. Garrido-Chamorro RP, Sirvent-Belando JE, Gonzalez-Lorenzo M, Martin-Carratala ML, Roche E. Correlation between body mass index
and body composition
in elite athletes. J Sports Med Phys Fitness. 2009; 49 (3): 278–84.
8. Heymsfield S, Lohman T, Wang Z, Going S. Human Body Composition
. 2nd ed. Champaign (IL): Human Kinetics; 2005. p. 523.
9. Heyward VH, Wagner DR. Applied Body Composition
Assessment. 2nd ed. Champaign (IL): Human Kinetics; 2004. p. 271.
10. Kjaer IGH, Lohne-Seiler H, Kolle E, et al.. Muskuloskeletal and motor fitness in Norwegian adults and elderly; status and the development of reference values. KAN 1. Med Sci Sports Exerc. 2011; 43 (5 suppl): S653.
11. Lohman TG, Chen Z. Dual-energy x-ray absorptiometry. In: Heymsfield S, Lohman T, Wang Z, Going S, editors. Human Body Composition
. 2nd ed. Champaign (IL): Human Kinetics; 2005. p. 63–78.
12. Loucks AB. Energy balance and body composition
in sports and exercise. J Sports Sci. 2004; 22: 1–14.
13. McArdle WD, Katch FI, Katch VL. Physique, performance, and physical activity. In: McArdle WD, Katch FI, Katch VL, editors. Exercise Physiology. Energy, Nutrition, and Human Performance. 6th ed. Baltimore (MD): Lippincott Williams & Wilkins; 2007. p. 811–33.
14. Milic R, Colombini A, Lombardi G, Lanteri P, Banfi G. Estimation of glomerular filtration rate by MDRD equation in athletes: role of body surface area. Eur J Appl Physiol. 2011; 111: 9–16.
15. Moon JR, Eckerson JM, Tobkin SE, et al.. Estimating body fat in NCAA Division I female athletes: a five-compartment model validation of laboratory methods. Eur J Appl Physiol. 2009; 105 (1): 119–30.
16. Müller W. Determinants of ski-jump performance and implications for health, safety and fairness. Sports Med. 2009; 39 (2): 85–106.
17. Nelson AE, Howe CJ, Nguyen TV, et al.. Influence of demographic factors and sport type on growth hormone–responsive markers in elite athletes. J Clin Endocrinol Metab. 2006; 91 (11): 4424–32.
18. Nevill AM, Stewart AD, Olds T, Holder R. Relationship between adiposity and body size reveals limitations of BMI. Am J Phys Anthropol. 2006; 129 (1): 151–6.
19. Nichols JF, Rauh MJ, Lawson MJ, Ji M, Barkai HS. Prevalence of the female athlete triad syndrome among high school athletes. Arch Pediatr Adolesc Med. 2006; 160 (2): 137–42.
20. O’Connor H, Caterson I. Weight loss and the athlete. In: Burke L, Deakin V, editors. Clinical Sports Nutrition. 3rd ed. New York (NY): McGraw-Hill; 2006. p. 135–74.
21. Ode JJ, Pivarnik JM, Reeves MJ, Knous JL. Body mass index
as a predictor of percent fat in college athletes and nonathletes. Med Sci Sports Exerc. 2007; 39 (3): 403–9.
22. Prentice AM, Jebb SA. Beyond body mass index
. Obes Rev. 2001; 2 (3): 141–7.
23. Rodriguez NR, DiMarco NM, Langley SDietitians of Canada; American College of Sports MedicineAmerican Dietetic Association. Position of the American Dietetic Association, Dietitians of Canada, and the American College of Sports Medicine: nutrition and athletic performance. J Am Diet Assoc. 2009; 109 (3): 509–27.
24. Santos DA, Silva AM, Matias CN, Fields DA, Heymsfield SB, Sardinha LB. Accuracy of DXA in estimating body composition
changes in elite athletes using a four compartment model as the reference method. Nutr Metab (Lond). 2010; 7: 22.
25. Silva AM, Minderico CS, Teixeira PJ, Pietrobelli A, Sardinha LB. Body fat measurement in adolescent athletes: multicompartment molecular model comparison. Eur J Clin Nutr. 2006; 60 (8): 955–64.
26. Sundgot-Borgen J. Prevalence of eating disorders in elite female athletes. Int J Sport Nutr. 1993; 3: 29–40.
27. Sundgot-Borgen J, Torstveit MK. Prevalence of eating disorders in elite athletes is higher than in the general population. Clin J Sport Med. 2004; 14: 25–32.
28. Sundgot-Borgen J, Torstveit MK. Aspects of disordered eating continuum in elite high-intensity sports. Scand J Med Sci Sports. 2010; 20 (2 suppl): 112–21.
29. Thompson RA, Sherman RT. Eating disorders and disordered eating in sport: a review of the literature. In: Thompson RA, Sherman RT, editors. Eating Disorders in Sport. New York (NY): Routledge, Taylor & Francis; 2010. p. 29–56.
30. Torstveit MK, Rosenvinge J, Sundgot-Borgen J. Prevalence of eating disorders and the predictive power of risk models in female elite athletes: a controlled study. Scand J Med Sci Sports. 2008; 18 (1): 108–18.
31. Torstveit MK, Sundgot-Borgen J. The female athlete triad: are elite athletes at increased risk? Med Sci Sports Exerc. 2005; 37 (2): 184–93.
32. Torstveit MK, Sundgot-Borgen J. The female athlete triad exists in both elite athletes and controls. Med Sci Sports Exerc. 2005; 37 (9): 1449–59.
33. Watts PB, Joubert LM, Lish AK, Mast JD, Wilkins B. Anthropometry of young competitive sport rock climbers. Br J Sports Med. 2003; 37 (5): 420–4.
34. World Health Organization. Obesity: Preventing and Managing the Global Epidemic: Report of a WHO Consultation. Geneva (Switzerland): World Health Organization; WHO Technical Report Series No. 894; 2000. p. 265.