Higher Prevalence of Eating Disorders among Adolescent Elite Athletes than Controls : Medicine & Science in Sports & Exercise

Journal Logo


Higher Prevalence of Eating Disorders among Adolescent Elite Athletes than Controls


Author Information
Medicine & Science in Sports & Exercise: June 2013 - Volume 45 - Issue 6 - p 1188-1197
doi: 10.1249/MSS.0b013e318281a939
  • Free


Most elite athletic participation and competition take place during adolescence or early adulthood (6). This is also when eating disorders (ED) often develop, when females, especially, experience a rapid change in the body composition and shape. For athletes, these possible changes might affect not only their attitudes toward weight and shape but also their athletic performance (6).

For high school athletes, most studies do not report a greater risk for the development of an ED than among age-matched controls (15,17,26,29). However, most studies have used only questionnaires to assess symptoms associated with ED rather than clinical interviews. Questionnaires have limitations, and it has been argued that clinical interviews are needed to obtain accurate prevalence data (20,32). The few studies available that have used a two-tiered approach (questionnaire screening followed by clinical interviews) have reported that adult elite athletes tend to underreport ED symptoms to a higher degree than controls, and that when clinical interviews are used, a higher proportion of elite female and male athletes actually have ED than that estimated from questionnaires (Diagnostic and Statistical Manual of Mental Disorders (DSM-IV)) (35,38,39).

Among the elite female and male adolescent athletes, no previous study has examined the prevalence of ED using clinical interviews. However, it has been reported that young athletes are at substantial risk for ED (27,29), and in one of our previous studies, most of the adult elite athletes who met the criteria for an ED reported having started dieting and developing an ED during puberty or adolescence (32). Thus, there is a need for a study that includes a representative population of young elite athletes representing a wide range of sports examining the prevalence of athletes classified as “at risk” for ED and then identifying how many meet the diagnostic criteria. Therefore, the aim of this study was to examine the prevalence of clinical ED in adolescent female and male elite athletes and age-matched controls by using a two-tiered approach, including a questionnaire screening followed by clinical interviews.


We conducted this study in two phases: self-reported questionnaire (part I) and clinical interview (part II).


We invited adolescent elite athletes attending Elite Sport High Schools (n = 16) in Norway and age-matched controls from two randomly selected regular high schools in Buskerud county in Norway to participate. The inclusion criteria for part I of the study were enrollment as a first-year student and a birth date in 1992. Norwegian Elite Sport High Schools are private and public high schools with programs designed for talented athletes. We invited 1211 first-year students, 711 athletes, and 500 controls, to participate in the initial screening during the 2008–2009 school year. Of these, 34 athletes and 79 controls were excluded (because of age: 29 athletes and 60 controls; because of failure to obtain parental consent: 5 athletes and 19 controls). Of the remaining 677 athletes and 421 controls, 66 athletes and 66 controls did not participate because they were not able to attend school during the test day. The reasons reported for the athletes not attending school that day were training camp, competition, or illness. When it comes to the controls, no specific reasons for not attending school were reported. This resulted in a final sample of 611 athletes (response rate, 90%) and 355 controls (84%), representing 50 different sports/disciplines, and for part of the analysis and in accordance with recent research, they were classified into weight-sensitive and less weight-sensitive sports (Table 1) (1).

Classification of the 50 different sports into weight-sensitive and less weight-sensitive sports.

The Regional Committee for Medical and Health Science Research Ethics in Southern Norway and the Norwegian Social Science Data Services approved the study. The respondents and their parents provided written consent to participate. We also obtained permission to collect data from each school principal.

Assessment Procedures

Part I: Screening.

We asked the respondents to complete a questionnaire that included a battery of assessment questions regarding training history, physical activity and nutritional patterns, menstrual history, oral contraceptive use, dieting and weight fluctuation history, use of pathogenic weight control methods, injuries, self-report of previous and/or current ED and the standardized questionnaire Eating Disorders Inventory-2 (EDI-2), and the short version of Hopkins Symptom Checklist (SCL-5). The questionnaire has been described in detail previously (17), except for the SCL-5. The SCL-5 is a five-item scale designed to measure symptoms of anxiety and depression. These five items are scored on a four-point scale ranging from 1 = “not at all” to 4 = “extremely.” The item score is calculated by dividing the total score of the number of items answered (ranging between 1 and 4), and a cutoff score of 2 indicates psychological distress (30). The SCL-5 has been proven reliable, and it is significantly correlated to the original instrument (37). In addition, a comparison study of several instruments found the reliability of this short version acceptable, with the recommendation to use the short version if included in a comprehensive questionnaire (30).

To be classified as “at risk” for ED, subjects had to meet at least one of the following criteria: a) drive for thinness (DT) score ≥15 for girls and ≥10 for boys; b) body dissatisfaction (BD) score ≥14 for girls and ≥10 for boys; c) body mass index (BMI) corresponding to the underweight value (8); d) trying to lose weight now; e) tried to lose weight before three times or more; e) current and/or previous use of pathogenic weight control methods: use of diet pills, laxatives, diuretics, or vomiting to reduce weight; or f) self-reported menstrual dysfunction: primary amenorrhea or secondary amenorrhea (previous 6 months). The athletes and controls completed the questionnaire at school during school hours in the presence of one or two members of the research group.

Physical activity was defined as the total hours of physical activity per week, including physical education lessons, recreational sports, and active daily living like walking. We classified those who reported ≥7 h of physical activity per week as physically active (equivalent to the Norwegian physical activity recommendations of 1 h·d−1 of moderate activity) (19). Training volume was reported as the number of training hours per week in the following categories: ≤5, 6–10, 11–15, 16–20, or ≥21.

Part II: Clinical interview.

Based on the data from part I, 102 female and 51 male athletes (n = 153, 25.0%) and 113 female and 67 male controls (n = 180, 50.7%) were classified as being “at risk” with symptoms associated with ED (Fig. 1). We invited all “at-risk” athletes (n = 153) and a randomly selected sample (50%) of “at-risk” controls (n = 91, 57 girls and 34 boys) to participate in a clinical interview. In addition, to match the “at-risk” subjects, a random sample of athletes (102 girls and 51 boys) and controls (22 girls and 66 boys) not classified as “at risk” after the screening examination was also invited to attend a clinical interview.

Number and percentage of subjects participating in parts I and II of the study.

The clinical interview was based on the Eating Disorder Examination 16:0 (EDE) (14), including sport-specific questions regarding the suggested predisposing, precipitating, and perpetuating factors related to ED risk, as well as the Eating Disorder Examination Questionnaire 6.0 (EDE-Q) (12). The EDE is an investigator-based interview that assesses ED psychopathology, i.e., undue importance of weight and shape in determining self-worth, as well as key ED behaviors. It is generally considered the best established instrument for assessing ED and is used for diagnostic purposes (7,13,40). The EDE-Q is a 36-item self-report measure derived from EDE focusing on the past 28 d and uses the same seven-point rating scale (0–6). Both EDE and EDE-Q are composed of four scales: restraint, weight concern, shape concern, and eating concern, with higher score representing greater severity of psychopathology. Both EDE and EDE-Q have been translated into Norwegian-language versions and validated (23,25).

We included the clinical interview to determine whether athletes and controls classified as being “at risk” for ED during the screening met the DSM-IV criteria for anorexia nervosa (AN), bulimia nervosa (BN), or ED not otherwise specified (EDNOS) (2) and to determine the presence and frequency of core ED symptoms. Four people did the interviews, one psychiatrist and three people trained in using the EDE. All had clinical experience with ED patients and special competence in physical activity and sports. A random selection of the 37 subjects (16 who fulfilled and 21 who did not meet the ED criteria) was interviewed again by a second interviewer. In all cases, there was complete agreement between the two interviewers concerning the diagnostic classification.

Statistical Analysis and Data Presentation

The statistical analyses were carried out using PASW Statistics 18 for Windows (IBM Corporation, Route, Somers, NY). The results are expressed as absolute numbers (N) and percentages (%) for categorical data and mean values with their SD for continuous data. We classified the athletes into weight-sensitive and less weight-sensitive sports as described previously (17). To compare the mean differences, an independent sample t-test was used, whereas we used chi-square tests to compare the categorical frequencies. We calculated the sensitivity as the proportion of the athletes and controls correctly identified as having ED and specificity as the proportion correctly identified as healthy. The negative predictive value (true negative) was calculated among the proportion of the athletes and controls classified as not “at risk” correctly identified as healthy. We estimated the total prevalence of ED from the prevalence of “at-risk” and healthy subjects participating in the clinical interview, factoring the distribution of “at-risk” and not “at-risk” subjects in the total sample. This was done separately for the athletes and controls with the use of the following formula: Ppop = P1W1 + P2W2, where Ppop = estimated prevalence in the total population, P1 = prevalence of the disorder in “at-risk” subjects in part I, W1 = weighted distribution of at-risk subjects in part I, P2 = prevalence of the disorder in subjects classified as not “at-risk” subjects in the clinical interview, W2 = weighted distribution of not at-risk subjects in part I.

The significance level was set to 0.05.


Part I

Subject characteristics.

The athletes were slightly younger than the controls at the time of data collection (16.5 ± 0.3 vs 16.9 ± 0.3, P < 0.001). A higher number of female controls than female athletes were classified as overweight according to age- and sex-adjusted BMI criteria (P = 0.003) (8). Among the controls, 55% reported being physically active. Among the athletes, 1.5% reported training for ≤5 h·wk−1, 16.5% for 6–10 h·wk−1, 46.2% for 11–15 h·wk−1, 25.7% for 16–20 h·wk−1, and 10.2% for ≥21 h·wk−1. Moreover, 17% of the athletes were selected for the national team (recruit, junior, or senior level). We observed no differences in the anthropometry between the controls participating in part I compared with those in part II.

Prevalence of symptoms associated with ED.

A higher proportion of controls than athletes were classified as “at risk” for ED (50.7% vs 25.0%, P < 0.001) (Fig. 1). Furthermore, a higher proportion of females compared with males were classified as “at risk” for ED in both groups (athletes: 46.2% vs 13.1%, P < 0.001; controls: 72.4% vs 33.7%, P < 0.001). We observed no difference in the proportion classified as being “at risk” for ED between weight-sensitive (26.4%) and less weight-sensitive sports (24.6%, P = 0.65).

Part II

Prevalence of athletes and nonathletes meeting the DSM-IV criteria for ED.

Because we gave priority to those classified as being “at risk” based on the questionnaire screening, there was a difference in the time from finishing the data collection from the screening part to attending the interview between those “at risk” and not “at risk” (12.1 ± 4.0 (SD) vs 16.3 ± 3.1 months, P < 0.0001). There was no difference between the athletes and nonathletes (14.2. ± 5.0 months vs 14.0 ± 2.2, P = 0.435).

The number of the athletes and controls who participated in the clinical interviews and met the DSM-IV criteria is shown in Figure 1. As seen, 24 of 96 female athletes and 5 of 48 male athletes classified as “at risk” based on the questionnaire fulfilled the criteria for ED. For the female controls, four of 57 “at risk” who participated in the clinical interview met the criteria (Fig. 1). Among the 34 athletes who met the criteria, 73.5% were diagnosed with EDNOS (20 females, 5 males), 23.5% with BN (7 females, 1 male), and 2.9% with AN (1 female), whereas all the female controls (n = 4) diagnosed with an ED were diagnosed with EDNOS.

As part of the analysis, the prevalence of the athletes and controls meeting the criteria for ED was standardized according to the total distribution of “at-risk” and not at-risk subjects for ED in part I of the study (Fig. 1). In this way, we estimated that 7.0% of the athletes and 2.3% of the controls met the criteria for an ED, with a difference of 4.7% (95% confidence interval (CI), 3.4–6.0; P = 0.001). The estimated prevalence was higher among the female athletes (14.0%) compared with the female controls (5.1%), with a difference of 8.9% (95% CI, 6.0–11.8; P = 0.005). Among the male athletes, 3.2% were estimated to meet the diagnostic criteria, with a difference from the female athletes of 10.8% (95% CI, 8.3–13.3; P < 0.001). We detected no difference in the prevalence of ED between the females representing weight-sensitive and less weight-sensitive sport groups (19.7% vs 11.9%, P = 0.136). Because of the few cases of ED among the male athletes, no subgroup comparisons were possible.

Five of 136 (3.7%) athletes and no controls classified as healthy from the screening examination were diagnosed with an ED (false negatives). The sensitivity of the screening examination was 0.85 versus 1.00 for the athletes and controls, respectively, whereas the specificity was 0.53 vs 0.50.

Characteristics of athletes meeting the criteria for an ED.

Table 2 shows that there were significant differences between the female and male athletes with ED compared with the athletes classified as healthy on several variables reported to be associated with ED. The athletes diagnosed with an ED had higher EDE-Q global score and subscale scores than healthy athletes (Table 3). The shape concern was the subscale with the highest mean score among both the athletes and controls with an ED. As expected, all of the 34 athletes with ED reported current dieting, 25 reported current and/or previous use of pathogenic weight control methods, and among these, 22 reported vomiting regularly at present. However, only eight reported binging and vomiting to such extent that they met the BN criteria (Table 4). Further characteristics of the severity and morbidity among the different athletes and controls with ED are given in Tables 4 and 5.

Characteristics of female and male adolescent elite athletes diagnosed with clinical ED or as healthy after the clinical interview.
EDE-Q subscale scores for healthy and eating disordered elite athletes and controls given as mean ± SD.
BMI, menstrual function, and dieting behavior in subjects meeting the DSM-IV criteria for AN, BN, or EDNOS.
Psychological and behavioral characteristics of subjects meeting the DSM-IV criteria for AN, BN, or EDNOS.


This is the first study to present prevalence data on ED among adolescent male and female athletes representing almost all European sports and a sample of adolescent controls. The novelty of this study is the two-tiered design including both the questionnaire screening and clinical interview and the high prevalence of clinical ED among adolescent elite male and female athletes.

The Prevalence of ED Is High among Adolescent Elite Athletes

Our finding revealed a high prevalence of ED among adolescent elite athletes. The prevalence was higher than among the controls, although more controls than athletes self-reported symptoms associated with ED. In addition, many athletes underreported the use of unhealthy behaviors such as pathogenic weight control methods at the screening compared with the reported use in the clinical interview, which is consistent with previous studies on adult elite athletes (18,31,36).

Among the athletes, it has been reported that some look at their disordered eating (DE) behavior as a natural part of their sport and are not necessarily generally dissatisfied with their body shape or weight (6,21,36), do not realize that they have a problem (4), and thus do not report DE problems. In addition, it has been argued that the DE behavior may reflect a rational response to pressure to achieve a body shape that will ensure optimal performance and reflect dedication to their sport rather than psychopathology (6). Therefore, it seems reasonable that the specific psychological concerns of adolescent elite athletes are somewhat different to those of nonathletes presenting a similar clinical picture, and that prevalence studies based on self-reported questionnaires will not be able to distinguish these differences. Consequently, by using the clinical interview EDE and sport-specific questions, we were able to get more accurate information about the intensity and duration of the most severe symptoms being essential to diagnose ED among both the athletes and the controls. Case number 2 (Tables 4 and 5) is a good illustration of this. This female athlete had BN, but at the screening, she reported no use of pathogenic weight control methods, had scores below the cutoff for “at risk” at the EDI subscales BD and DT, and was only classified as “at risk” because of the present dieting. However, at the clinical interview, it was clear that her dieting behavior primarily was related to sport-specific weight concern. Moreover, her thoughts about weight, shape, and food were at such an intensity that they affected her self-worth and concentration and had resulted in key ED behaviors such as binging and self-induced vomiting. As a result, she met the DSM-IV criteria for BN. Subsequently, our use of a two-tiered approach confirms the difficulties that exist when it comes to identifying athletes at risk for ED, or those who already have developed an ED, and explains why it is important to use clinical interviews to assess the specific psychopathology of ED.

Because there are no previous studies that have used a two-tiered approach when investigating ED in adolescent elite athletes, it is difficult to compare the prevalence reported in our study with other studies. In general, comparing prevalence data is difficult because most studies use a variety of methodological approaches. Different populations are investigated (adolescent vs adult, collegiate vs elite vs recreational), and the number of athletes studied (n = size), the number and type of sports included, and which part of the training or competition season the data were collected vary between studies (36). Taking this into account, our findings are in accordance with the studies on older elite athletes using a two-tiered approach (7,35,38). However, the prevalence was, as expected, lower than found among older Norwegian elite athletes in 2004 (35). In contrast to our present study, Sundgot-Borgen and Torstveit (35) included the subclinical ED anorexia athletica (AA) in their prevalence study. Still, it may be argued that these athletes would have been diagnosed with EDNOS if AA had not been included. Furthermore, if AA was eliminated as a diagnosis and these athletes were classified as healthy, the percentage of ED would still be higher among both female (16.3%) and male (7.4%) athletes when using the same estimate calculation as we used in this present study. Similarly, Torstveit et al. (38) found a higher percentage of ED among adult female elite athletes as compared with the prevalence among the females in our study. Thus, although adolescence is considered to be the period of greatest vulnerability and the time when ED typically occurs, the prevalence of ED seems to be higher among adult elite athletes.

Several explanations may be relevant for the higher prevalence of ED among the adult elite athletes than the young elite athletes in this study, for example, longer duration of sport-specific demands such as competing in sports with weight class system and where the procedure during competitive events may allow use of extreme weight loss methods, physique-revealing uniforms, or competing in sports in which appearance and performance are being judged. In addition, frequent weight fluctuation has been suggested to be an important trigger factor for the development of ED in athletes (32). In our study, the athletes are in the beginning of their athletic career, whereas in Sundgot-Borgen and Torstveit’s study (35) the athletes were older and had been competing at the highest national and international level for years. Furthermore, and in contrast to previous studies on adult elite athletes (7,35,38), we found no difference in the prevalence of ED among athletes in weight-sensitive sports compared with less weight-sensitive sports. Although BD is considered an important symptom of ED (2), 50% of the athletes in our study diagnosed with an ED had BD scores above clinical values, but all reported sport-specific BD. Byrne and McLean (7) suggest that the demands of a sport to meet a particular body requirement alone, even without a high level of BD, may be enough to lead to the development of an ED. In addition, Torstveit et al. (38) found that even though the prevalence of ED was highest among athletes competing in weight-sensitive sports, fewer were dissatisfied with their bodies compared with athletes in less weight-sensitive sports and controls. This indicate that other factors than general BD may be essential for the development of ED among athletes, and it may be speculated that the higher prevalence among adult elite athletes is due to a longer duration of exposure to sports environment factors that can contribute to heightened concerns regarding weight and promote pathogenic weight control behaviors.

ED Diagnosis and the Severity

As expected, the most frequent ED was EDNOS, which is the most common ED encountered in outpatient (24) and inpatient settings (10). EDNOS is sometimes called the residual category in the DSM-IV classification because it is a “catch-all” category for eating problems that are considered to be of clinical severity but do not fulfill the criteria for AN or BN (22). EDNOS is therefore sometimes viewed as a “less severe” form of ED, but a recent study observed a crude mortality of 5.2%, similar to those found in AN (9). However, EDNOS is a residual category, and some of the athletes fulfilling the EDNOS criteria closely resemble cases of AN and BN, whereas others differ in terms of psychopathology and medical condition. As an example, one of the female athletes diagnosed with EDNOS reported fear of weight gain, loss of concentration due to bodyweight and shape and food concern, high BD, binge eating at least twice a week, and daily compulsive exercise (in addition to her planned sport-specific training) (case number 19, Tables 4 and 5). Because she reported vomiting only once a week for the last 3 months, she did not fulfill the criteria for BN. Another athlete (case number 12, Tables 4 and 5) did not report use of pathogenic weight control methods and had regular menstruation. However, she reported binge eating four times a week and fulfilled the psychological criteria for EDNOS (the binge ED). The many hours of training (16–20 h·wk−1) may explain her normal weight for age and height, despite the binge eating. Finally, case number 15 (Tables 4 and 5) fulfilled all the criteria for AN except that she had a normal weight for her age and height, and therefore, the correct diagnosis was EDNOS (partial or subclinical AN). In sum, this means that those who fulfill the EDNOS criteria differ in the degree of psychological and medical severity and chronicity. In addition to this, the characteristics of their competitive sport and the duration and intensity of the symptoms will influence the consequences of their ED behavior on acute and long-term health as well as performance (34).

Frequent binging and vomiting are examples of behavior that can lead to both health and performance consequences on a short- and long-term basis (4). For instance, bingeing generally results in gastric distention where the severity increases with an increase in the amount of food consumed during the binge (4), whereas purging by self-induced vomiting, laxative abuse, or diuretic abuse may lead to dehydration, acid–base and electrolyte imbalances, cardiac arrhythmia as well as different chronic physical problems (4). Furthermore, the combination of binging and compulsive exercise for young athletes will increase the risk of overtraining, chronic fatigue, illness, overuse injuries, and menstrual dysfunction (4).

ED athletes in this study reported use of pathogenic weight control methods ranging from not at all to daily. The same range was found for binge eating. The use of these methods and/or the extreme weight and body shape (related) stress experienced underline the acute need for treatment. Consequently, some athletes should be withdrawn from competition as well as training, whereas others can continue training during treatment (11). These challenge the type of treatment and prevention strategies that should be chosen, as well as the balance between taking risk behavior seriously, but do not morbidly challenge the behavioral or psychological traits that are common and accepted in some sports needed for success. It is therefore essential to determine whether the athletes’ abnormal eating and dieting behaviors are transient, safely managed behaviors associated with the specific demands of the sport or if the symptoms are more stable and represent a clinical ED (11,33).

Higher Prevalence of ED among Athletes Than Controls

In our study, we found a higher prevalence of ED among the athletes compared with the controls and general prevalence estimates (28). In addition to the unique pressure athletes face compared with nonathletic controls, the Elite Sport High School setting may be an additional factor that increases the risk for DE and the development of an ED. Some of the factors that may lay a high pressure on these young athletes to perform on the sport field as well as in the classroom are greater personal responsibility and maturation, perceived loss of social support, competing with the best athletes in the country, evaluation by coaches on an almost daily basis, increased training volume, lack of performance, and not being the best anymore. Some might also feel indebted to their family to achieve good results. For some, such sudden changes and high pressure might be experienced as traumatic events and serve as trigger conditions for the onset of ED. In addition, Barker and Galambos (3) found that moving away from home to attend college increased the likelihood of binge eating among young women. Other factors such as loss of coach or an illness or injury might also be conceptualized as traumatic events and trigger conditions (32).

Why more controls than athletes reported symptoms associated with ED without having an ED may be explained by the higher percentage of overweight (BMI), higher percentage with BD, and percentage reporting dieting (17). The fact that more female controls than athletes are classified as overweight in this study, with the focus on the importance of a lean body and the general experienced pressure of a thin ideal body, might have influenced the controls’ perception of their bodies and partly explain the high number with BD and DT scores above the cutoff for the “at-risk” criteria. Previous findings also indicate that females especially are dissatisfied with their bodies and tend to report DE or have high scores on the EDI without being actively dieting or using extreme methods to reduce or control weight (32). Also, the questionnaire being used may have had an effect. The EDI subscale BD is for instance not sport specific and assesses dissatisfaction with areas of the body such as stomach and hips that well-trained athletes most likely are generally satisfied with. Furthermore, the reasons for dieting or being dissatisfied with one’s own body may be different for males compared with females (16). In male athletes, high BD does not necessarily mean a desire to lose weight and or getting thinner. For instance, only 4 of our 17 male athletes with a BD score over the cutoff (≥10) present dieting, and only two of them were diagnosed with an ED (Table 4). The many false-positive male athletes and one false-negative athlete diagnosed with EDNOS also indicate that the questionnaires being used do not have the sensitivity to differentiate between male athletes who are dissatisfied with their body because they want to be more muscular, athletes who are dissatisfied because of the extra fat, or those with ED.

Methodological Considerations

As expected, the use of a wide range of at-risk criteria for ED led to a high number of false positives, but more strict criteria could have resulted in too many false negatives (38). Because this study is part of an intervention aiming to prevent the development of ED, a high number of false negative would have been less advisable.

A cross-sectional study design has its limitations, but our use of a two-tiered approach by including the clinical interview EDE is considered to be the “gold standard” for diagnosing ED (7,13). Furthermore, all measurements have some weaknesses limiting their validity and reliability, but until more sport-specific measurements are developed, a two-tiered design is a good solution.

Because of ethical consideration, we prioritized to interview those classified as being “at risk” based on the questionnaire screening first. This resulted in a time difference from the screening to attending the interview between those “at risk” and not “at risk” and the athletes and controls. Because certain features of diagnostic importance in the interview are assessed over a 3-month period, and also a random selection of those diagnosed and not diagnosed with an ED was interviewed again by the second interviewer with complete agreement; we believe that this time difference most likely has not had any effect on the results. Because so many of the participants classified as healthy at the screening attended the clinical interview and only 5 of 218 were false negatives (all of them were athletes), we believe answering our questionnaire did not lead to focus more on weight, body shape, and dieting among the participants. Furthermore, when interviewing these five false-negative athletes, they all reported a history of ED, indicating that they were already experiencing ED when they fulfilled the questionnaire at screening.

The causes reported for athletes not attending school at the test day indicate that there are no reasons to believe that those not attending differed from the sample of athletes who participated in the study. Because we have no specific reasons for controls not attending school at the test day, we could speculate that some did not attend because of the test related to DE. However, it could also be due to ordinary reasons for not attending school. Based on this, we believe that the nonresponders did not skew the sample in any direction. Furthermore, the high response rate, the large number of sports included, and the fact that a representative sample of both those classified as “at risk” and healthy attended the clinical interview show that these data are generalizable to other adolescent elite athletes and controls. However, the lack of instruments designed for athletes may have had an effect on the prevalence of athletes classified as “at risk.” For instance, the BD scale might not be suited for those who are dissatisfied because they want to gain weight (increase muscle mass), which is of special interest for the male athletes representing ice hockey, team handball, alpine skiing, and some athletic events as well as discriminate between general BD and more sport-specific BD. Furthermore, because dieting is usually seen as the primary precursor to the development of DE and ED, we included the present dieting as one of our at-risk criteria. However, dieting in this special setting should maybe more correctly be considered as a “collection” factor for other symptoms not detected. Finally, it is worth noting that when comparing female athletes with ED in weight-sensitive sports and less weight-sensitive sports, we calculated the lowest detectable difference to 17.1% when the observed was 7.8%. The difference in these populations could only be detected if the number per group had been fourfold if the power had been at least 80%.


Our results support two previous important statements: 1) a clinical interview is the best available method to identify ED among both elite athletes and controls, and 2) there is a need for further development and validation of screening instruments designed for athletes. A sport-specific instrument should be designed for athletes in different kinds of sports and competition levels. Until this is established, clinical interviews should be used for accurate identification of ED.

Education about health and performance-related nutrition and body composition should be administered at an earlier age than high school. Females should especially be educated on the health and performance consequences of menstrual irregularities and the importance of seeking timely medical intervention at the first sign of abnormalities (5). Coaches as well as other athletic staff members are in a prime position to monitor their athletes’ behaviors and reactions, and therefore, mandatory education programs should be implemented on an annual basis. Finally, sports-governing organizations and federations should give support to the coaches and provide education for coaches, athletic staff, and athletes (36).


This study is the first to confirm that the prevalence of ED is higher among adolescent elite athletes than controls and higher in female than male adolescent elite athletes. It also confirms the difficulties that exist when it comes to identifying male and female athletes “at risk” for ED and the need for a clinical interview to identify athletes who already have developed an ED.

The authors thank Professor Roald Bahr for his helpful comments on this article and Professor Ingar Holme for his assistance with the statistical analysis. We also thank all the subjects and their coaches for full cooperation and support during this study, and Audun Erikson, Guro Røen, Laila Sundgot Schneider, Runi Børresen and Kristin Brinchman Lundestad with regard to the collection of data. The Oslo Sports Trauma Research Center has been established at the Norwegian School of Sport Sciences through generous grants from the Royal Norwegian Ministry of Culture, the South-Eastern Norway Regional Health Authority, the International Olympic Committee, the Norwegian Olympic Committee and Confederation of Sport, and Norsk Tipping AS. The project has been established through a grant from The Norwegian Olympic Sports Center (Olympiatoppen).

There are no potential conflicts of interest.

The results of the study do not constitute endorsement by the American College of Sports Medicine.


1. Ackland TR, Lohman TG, Sundgot-Borgen J, et al. Current status of body composition assessment in sport: review and position statement on behalf of the ad hoc research working group on body composition health and performance, under the auspices of the I.o.C. Medical commission. Sports Med. 2012; 42 (3): 227–49.
2. APA. Eating Disorders. Diagnostic and Statistical Manual of Mental Disorders: DSM-IV. Washington (DC): American Psychiatric Association; 1994. pp. 539–50.
3. Barker ET, Galambos NL. Exploring criterion confounding in the measurement of binge eating symptoms and negative affect. J Psychopathol Behav Assess. 2009; 31 (1): 43–50.
4. Beals KA. Disordered Eating among Athletes: A Comprehensive Guide for Health Professionals. Champaign (IL): Human Kinetics; 2004. p. 255.
5. 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.
6. Byrne S, McLean N. Eating disorders in athletes: a review of the literature. J Sci Med Sport. 2001; 4 (2): 145–59.
7. Byrne S, McLean N. Elite athletes: effects of the pressure to be thin. J Sci Med Sport. 2002; 5 (2): 80–94.
8. Cole TJ, Flegal KM, Nicholls D, Jackson AA. Body mass index cut offs to define thinness in children and adolescents: international survey. Br Med J. 2007; 335 (7612): 194–7.
9. Crow SJ, Peterson CB, Swanson SA, et al. Increased mortality in bulimia nervosa and other eating disorders. Am J Psychiatry. 2009; 166 (12): 1342–6.
10. Dalle GR, Calugi S. Eating disorder not otherwise specified in an inpatient unit: the impact of altering the DSM-IV criteria for anorexia and bulimia nervosa. Eur Eat Disord Rev. 2007; 15 (5): 340–9.
11. Drinkwater B, Loucks A, Sherman R, Sundgot-Borgen J, Thompson R. International Olympic Committee Medical Commission Working Group Women in Sport: position stand on the female athlete triad. International Olympic Committee 2005; [cited 2010 September 4]. Available from: http://www.olympic.org/Documents/Reports/EN/en_report_917.pdf
12. Fairburn CG, Beglin S. Eating Disorder Examination Questionnaire (EDE-Q 6.0). In: Fairburn CG, editor. Cognitive Behavior Therapy and Eating Disorders. New York: The Guilford Press; 2008. pp. 309–14.
13. Fairburn CG, Bohn K. Eating disorder NOS (EDNOS): an example of the troublesome “not otherwise specified” (NOS) category in DSM-IV. Behav Res Ther. 2005; 43 (6): 691–701.
14. Fairburn CG, Cooper Z, O’Connor ME. Eating Disorder Examination (Edition 16.0D). In: Fairburn CG, editor. Cognitive Behavior Therapy and Eating Disorders. New York: The Guilford Press; 2008. pp. 265–308.
15. Fulkerson JA, Keel PK, Leon GR, Dorr T. Eating-disordered behaviors and personality characteristics of high school athletes and nonathletes. Int J Eat Disord. 1999; 26 (1): 73–9.
16. Glazer JL. Eating disorders among male athletes. Curr Sports Med Rep. 2008; 7 (6): 332–7.
17. Martinsen M, Bratland-Sanda S, Eriksson AK, Sundgot-Borgen J. Dieting to win or to be thin? A study of dieting and disordered eating among adolescent elite athletes and non-athlete controls. Br J Sports Med. 2010; 44 (1): 70–6.
18. Nattiv A, Loucks AB, Manore MM, Sanborn CF, Sundgot-Borgen J, Warren MP. American College of Sports Medicine. Position Stand: the female athlete triad. Med Sci Sports Exerc. 2007; 39 (10): 1867–82.
19. Norwegian Directorate for Health and Social Affairs. Physical Activity and Health. Recommendations. Oslo: Sosial- og helsedirektoratet; 2000. p. 82.
20. Oppliger RA, Landry GL, Foster SW, Lambrecht AC. Bulimic behaviors among interscholastic wrestlers—a statewide survey. Pediatrics. 1993; 91 (4): 826–31.
21. Owens RG, Slade PD. Running and anorexia nervosa—an empirical study. Int J Eat Disord. 1987; 6 (6): 771–5.
22. Palmer B, Norring C. EDNOS—the other eating disorders. In: Norring C, Palmer B, editors. EDNOS: Eating Disorders Not Otherwised Specified. London: Routledge; 2005. pp. 1–9.
23. Reas DL, Wisting L, Kapstad H, Lask B. Convergent validity of the eating disorder examination and the eating disorder examination-questionnaire among university women in Norway. Eur Eat Disord Rev. 2011; 19 (4): 357–61.
24. Ricca V, Mannucci E, Di Bernardo M, et al. Psychopathological and clinical features of outpatients with an eating disorder not otherwise specified. Eat Weight Disord. 2001; 6 (3): 157–65.
25. Ro O, Reas DL, Lask B. Norms for the eating disorder examination questionnaire among female university students in Norway. Nord J Psychiatry. 2010; 64 (6): 428–32.
26. Rosendahl J, Bormann B, Aschenbrenner K, Aschenbrenner F, Strauss B. Dieting and disordered eating in German high school athletes and non-athletes. Scand J Med Sci Sports. 2009; 19 (5): 731–9.
27. Sigman GS. Eating disorders in children and adolescents. Pediatr Clin North Am. 2003; 50 (5): 1139–77.
28. Smink FR, van HD, Hoek HW. Epidemiology of eating disorders: incidence, prevalence and mortality rates. Curr Psychiatry Rep. 2012; 14 (4): 406–14.
29. Smolak L, Murnen SK, Ruble AE. Female athletes and eating problems: a meta-analysis. Int J Eat Disord. 2000; 27 (4): 371–80.
30. Strand BH, Dalgard OS, Tambs K, Rognerud M. Measuring the mental health status of the Norwegian population: a comparison of the instruments SCL-25, SCL-10, SCL-5 and MHI-5 (SF-36). Nord J Psychiatry. 2003; 57 (2): 113–8.
31. Sundgot-Borgen J. Prevalence of eating disorders in elite female athletes. Int J Sport Nutr. 1993; 3 (1): 29–40.
32. Sundgot-Borgen J. Risk and trigger factors for the development of eating disorders in female elite athletes. Med Sci Sports Exerc. 1994; 26 (4): 414–9.
33. Sundgot-Borgen J. Atypical eating disorders in female athletes. In: Norring C, Palmer B, editors. EDNOS: Eating Disorders Not Otherwise Specified. London: Routledge; 2005. pp. 266–76.
34. Sundgot-Borgen J, Garthe I. Elite athletes in aesthetic and Olympic weight-class sports and the challenge of body weight and body composition. J Sports Sci. 2011; 29: S101–14.
35. 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 (1): 25–32.
36. Sundgot-Borgen J, Torstveit MK. Aspects of disordered eating continuum in elite high-intensity sports. Scand J Med Sci Sports. 2010; 20: 112–21.
37. Tambs K, Moum T. How well can a few questionnaire items indicate anxiety and depression. Acta Psychiatr Scand. 1993; 87 (5): 364–7.
38. Torstveit MK, Rosenvinge JH, 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.
39. 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.
40. Vrabel KR, Hoffart A, Ro O, Martinsen EW, Rosenvinge JH. Co-occurrence of avoidant personality disorder and child sexual abuse predicts poor outcome in long-standing eating disorder. J Abnorm Psychol. 2010; 119 (3): 623–9.


©2013The American College of Sports Medicine