Balancing Performance and Health in Elite Hungarian Athletes: The Relationship Among Disordered Eating Risk, Body Composition, and Nutrition Knowledge

The study employed a cross-sectional quantitative design, utilizing a combination of survey-based, self-reported assessments alongside body composition measures. This design allowed for the simultaneous collection of data on disordered eating risk, nutrition knowledge, and body composition within a single time frame, providing a snapshot of these variables among elite Hungarian athletes. By integrating self-reported behavioral data with objective body measurements, the study aims to provide a comprehensive picture of disordered eating risk contextualized in nutrition knowledge and physical health markers.

Data collection for this study was conducted from 10 February to 31 July 2024, targeting elite athletes over 18 years of age. Recruitment took place at the National Institute for Sports Medicine in Budapest, Hungary, using convenience sampling. Posters featuring a QR code were displayed on-site to encourage participation. By scanning the QR code, athletes could access the study information, including the study’s purpose, completion guidelines, and data security measures. Once the athletes had reviewed and accepted the study conditions (by checking a consent box), they were directed to complete the following questionnaires: the Hungarian version of the disordered eating screen for athletes (DESA-6H), the Eating Attitudes Test (EAT-26), and the Abridged Nutrition for Sport Knowledge Questionnaire (A-NSKQ) [44,45,46]. In addition, participants provided sociodemographic details (e.g., sex, age) and sport-related information (e.g., sport discipline, years in sport, training hours per week). Answering a total of 94 questions took an average of 27 (SD = 10) minutes. It is similar to a previous study by Trakman et al. (2018), where 89 questions took 25 min to answer [45,46]. After completing the survey, the participants registered at the outpatient dietetic care for physical measurements, including height measured with a Seca 200 device and body composition measured with an OMRON BF511 body composition measurement device, where their body weight, BMI, and body fat percentage (PBF) were recorded. Each athlete received an individual code to ensure anonymity. Data responses collected via Qualtrics were exported to Microsoft Excel 2007 for analysis.

This study was conducted following ethical standards and received approval from the first author’s institutional ethics board (ELTE PPK KEB 2024/10). Informed consent was obtained from all participants, who were fully briefed on the study’s purpose, procedures, and their right to withdraw at any time without consequences. Participation was voluntary, and athletes received no financial compensation; however, their body composition measurements were shared with them as part of their health information. After the data collection was completed, data were anonymized using individual codes to protect the participants’ identity, with all digital data securely stored on password-protected systems accessible only to authorized research personnel.

To mitigate against potential psychological distress, participants were provided with information on resources for mental health and nutrition counseling if needed. Participants were informed that, if the assessment identified an elevated risk of DE, they would be referred to appropriate support resources for follow-up care upon request.

Descriptive statistics, including mean (M), standard deviation (SD), and frequency (%), were used when presenting details on gender, age, BMI, PBF, at-risk and non-risk groups, scores of DESA-6H, EAT-26, and A-NSKQ. The normal distribution assumption was tested with the Shapiro–Wilk test. The possible relationship between DESA-6H and EAT-26 total scores, in addition to EAT-26 subscales with BMI, PBF, general and sport nutrition knowledge, and total nutrition knowledge score, was examined using Kendall’s tau correlation tests. Non-parametric Mann–Whitney U tests were used to investigate pairwise differences between the following variables: DESA-6H and EAT-26 scores by gender and nutritional status, at-risk and non-risk groups based on DESA-6H and EAT-26 scores in nutrition knowledge. Effect sizes were determined using Cohen-d effect size indicators [53,54] and interpreted for Mann–Whitney U tests, as follows: d < 0.3 (small effect), 0.3 ≤ d < 0.5 (medium effect), d ≥ 0.5 (large effect). Cohen’s κ was run to determine if there was agreement between DESA-6H and EAT-26 categorizations, and whether certain athletes were at risk or not was interpreted as follows: 0.01–0.20 none to slight, 0.21–0.40 fair, 0.41–0.60 moderate, 0.61–0.80 substantial, and 0.81–1.00 almost perfect agreement [55,56]. To explore the diagnostic accuracy of DESA-6H with respect to how well the test separated DE risk groups among elite Hungarian athletes, ROC (receiver operating characteristic) analysis and area under the curve (AUC) were used with Youden’s index (J) to determine an optimal cut-off value. For this analysis, the at-risk status was determined by the EAT-26, a tool which has been widely used in sport-related settings and beyond to screen for ED risks [57,58]. In the statistical analyses, p < 0.05 was considered statistically significant. All calculations were conducted using IBM SPSS 29.0.

The mean DESA-6H score across the sample was 0.7 (SD = 1.6), with women scoring slightly higher 0.8 (SD = 1.9) than men 0.5 (SD = 1.1). Based on DESA-6H scores, nine athletes (one male and eight females) were assigned to the risk group for disordered eating. For the EAT-26, the mean score in the total sample was 11.2 (SD = 7.0), with men averaging at 8.5 (SD = 4.2) and women at 12.9 (SD = 8.0). Overall, 23 athletes (five males and 18 females) scored in the risk zone for eating disorders. Figure 1 shows the characteristics of the athletes classified in the risk group based on DESA-6H and EAT-26 scores.

The most affected athletes were female team athletes (n = 5) and those in aesthetic, endurance, and weight-dependent sports (n = 3 in each group).

For EAT-26, we examined the individual factors as well. In case of dieting behavior, scores ranged between 0 and17, with 2 (n = 16, 22.5%), 0 (n = 11, 15.4%), and 8 (n = 6, 8.4%) being the most frequent in the sample. In the oral control subscale, athletes scored between 0 and 10, but over half scored 0 (n = 17, 24%), 2 (n = 16, 22.5%), or 3 (n = 13, 18.3%). Finally, examining bulimia and the food preoccupation scale, scores ranged from 2 to 14, with the most frequent score being 3 (n = 51, 71.8%), 4 (n = 4, 5.6%), or 9 (n = 3, 4.2%).

Body composition classifications were based on the reference ranges provided by the World Health Organization (WHO) and the American Council on Exercise (ACE) [59,60]. Most athletes (74%) fell within the normal BMI range. One male swimmer and two females (one orienteer and one dancer) were classified as underweight, while one male and one female racewalker approached the lower limit of the normal range (18.7 and 18.6 kg/m², respectively). Four male athletes (two shooters, one discus thrower, and one javelin thrower) and three female athletes (a shooter, a water polo player, and a canoeist) were classified as overweight. Regarding PBF, four males (one each from swimming, canoeing, dancing, and orienteering) had low BF, while others in shooting, archery, discus throwing, and volleyball showed high levels. Among female athletes, one athlete each from orienteering, pentathlon, rhythmic gymnastics, and long jump had a low PBF, while one water polo player exhibited high BF.

On the A-NSKQ questionnaire, the whole sample scored an average of 56.1% in general nutrition knowledge and 42.1% in sports nutrition knowledge, resulting in an overall score of 49.1%. The greatest gender difference (with higher scores among women) was observed in relation to the general nutrition knowledge, while scores on sport-specific questions were nearly equivalent (Figure 2).

The relationships between the measures used to assess athletes’ DE risks, body composition, and nutrition knowledge are shown in Table 1.

Regarding disordered eating risk, the DESA-6H significantly correlated with EAT-26 total score across the total sample. For body composition, a significant correlation was observed with DESA-6H and BMI, but no relationship was found between DESA-6H and PBF, nor between EAT-26 total scores and BMI or PBF. With respect to nutrition knowledge, DESA-6H showed significant correlations with sport nutrition knowledge and A-NSKQ total scores. Similar relationships were shown between EAT-26 scores and general-, sport nutrition knowledge, and A-NSKQ total scores, respectively. Statistically significant, strong, positive correlations were only found within the nutrition knowledge measures (0.51–0.85). Across domains, correlations ranged between 0.22 and 0.29, suggesting positive but weak relationships.

Statistically significant differences with a large effect size were found in relation to the EAT-26 total scores of men and women, with females scoring higher than males. However, no significant differences showed between genders in the case of DESA6-H scores (Table 2).

When investigating the differences between nutritional status categories, we set three groups based on the BMI results (underweight, normal, overweight, and obese). Only one participant was categorized as obese (BMI = 33.4 kg/m²) and was added to the overweight group.

Due to the small sample size, a comparison among the three groups was not possible. Differences between the two groups (normal vs. overweight and obese), however, showed a statistically significant difference with a large effect size between the nutritional status and DESA-6H, with overweight and obese athletes scoring higher. There were no significant differences in the EAT-26 total scores. Our results are detailed in Table 3. For most variables, the overweight and obese group had less favorable scores.

The number of participants identified as ‘at risk’ and ‘non-risk’ based on their DESA-6H scores were unbalanced (at-risk, n = 9, mean score: 4.11, SD = 0.92; non-risk, n = 62, mean score: 0.24, SD = 1.02)

No significant differences in nutrition knowledge and its factors were found between DESA-6H-based at-risk and non-risk groups (Table 4).

Differences in nutrition knowledge between the EAT-26 risk groups (at-risk, n = 47, mean score: 11.18 (SD = 7.02); non-risk, n = 24, mean score: 19.08 (SD = 6.48)) are presented in Table 5.

A statistically significant, medium-effect-size difference was found in relation to sport nutrition knowledge and a large-effect-size difference was found in relation to the nutrition knowledge total score; however, there was no difference in general nutrition knowledge between the EAT-26-based non-risk and at-risk groups.

To bring all the results together, Figure 3 depicts the observed profiles for at-risk and non-risk athletes.

As shown in Figure 3, 12.7% of the athletes was identified as ‘at-risk’ based on DESA6-H and 32.4% based on EAT-26, with men scoring less severely. There was a statistically significant fair agreement between the two variables, κ = 0.294, p = 0.03. For non-risk participants, a 63.3% overlap was found between DESA-6H and EAT-26, whilst only 24.03% of the at-risk athletes were identified by both screening tools. Thus, the remaining 12.67% was not categorized in the same group (at-risk or non-risk). Interestingly, during body composition measurements, there was not a large proportion of abnormal results, as only 4.2% of the athletes were underweight and 10% of them were overweight. Similarly, 11.2% had a low and 7.0% high body fat percentage. When assessing nutrition knowledge, females scored higher in both fields (general and sport nutrition knowledge); however, the average scores did not meet the acceptable level neither for males nor for females.

To establish evidence that DESA-6H is a good screening tool to distinguish between at-risk and non-risk groups, a ROC analysis was executed. At-risk and non-risk groups were based on EAT-26, where athletes with a score of 12 or higher were identified as being ‘at risk’. When examining the diagnostic value—accuracy of the test—for DESA-6H, the true positive rate (sensitivity) was measured against the false positive rate (1-specificity) with the ROC curve. In this present case, under the curve values (AUC = 0.82, 95%CI: 0.71–0.93, p < 0.001) showed that DESA-6H was able to identify at-risk athletes for disordered eating (Figure 4).

The Youden’s Index (J) is often used, as it provides the best balance between sensitivity and specificity. In our case, the DESA-6Hmax (J = 0.54) cut-off was set to 0.50, with a sensitivity of 83.3% and a specificity of 70.2%. According to international recommendations for DESA-6H, the cut-off point is ≥3 (J = 0.25), with a sensitivity of 29.1% and a specificity of 95.7%. In the present sample, we took this as a basis for the previously presented analyses. Table 6 presents the statistical values according to the two cut-off definitions in terms of separating the DE groups. Based on the two cut-off points, it can be seen that the international standard had a high specificity and a low sensitivity in our case. The best balance that we indicate has a higher sensitivity and a lower specificity compared to the international standard. However, it is worth noting that there was no meaningful difference in accuracy between the results of the analyses based on the two cut-off points.

The study revealed that female athletes, particularly those in team, aesthetic, endurance, and weight-dependent sports, were more likely to score in the risk zone for DE in both the DESA-6 and EAT-26 questionnaires. This is consistent with previous research that identifies female athletes as more vulnerable to DE behaviors due to increased pressures around body image and weight [63,64,65,66]. For example, prior studies have shown that female bodybuilders, gymnasts, and endurance athletes face a heightened risk of DE, with some even reaching clinical thresholds for eating disorders [67,68]. Cultural comparisons, such as the study by Okano et al. (2005) on Japanese and Chinese athletes, suggest that vulnerability to DE may also be shaped by sociocultural factors, with prevalence rates varying between countries and even among athletes in the same sport [69]. These findings emphasize the need for DE prevention strategies that address both sport-specific and cultural pressures. When comparing nutrition knowledge of at-risk and non-risk athletes, at-risk individuals were more informed around nutrition with large effect sizes for sports nutrition knowledge and the A-NSKQ total score. This result aligns with a systematic review which showed that athletes’ nutrition knowledge does not necessarily influence conscious food choices [40].

DESA-6 is a relatively new assessment tool for screening disordered eating risk; therefore, there is a limited number of studies implementing this screening tool for athletes. Although results, so far, have shown that DESA-6 is a reliable questionnaire [68], the recommended cut-off value for at-risk status could be challenged on practical grounds. Whilst the recommended score of 3 or higher leads to a high level of specificity, our results suggest that it may lead to missing athletes who could also be ‘at risk’ and in need of support and intervention.

Body composition results suggest that athletes competing in sports where low body weight is emphasized (lean sports), such as aesthetic and endurance sports, had lower PBFs. This aligns with the classification of lean and non-lean sports by Mancine et al. (2020) [8] and underscores the physical demands and aesthetic expectations that drive athletes in these disciplines toward lower body weight and fat. Interestingly, no significant relationship was observed between PBF and DE risk scores (DESA-6 and EAT-26) in our sample. This contrasts with Magee et al. (2023), who found a negative correlation between low BF and ED risk, particularly in female athletes [39]. Our findings suggest that, within this sample, DE risk may be more influenced by psychological and behavioral factors than by body composition alone. Future research could further explore this complex relationship, potentially examining the roles of psychological factors and the individual’s experience of pressure related to body composition.

Our findings indicate that female athletes scored higher on both general and sports nutrition knowledge questions (as assessed by A-NSKQ), with no significant differences between the study groups. These findings align with previous studies suggesting that female athletes often exhibit greater nutrition awareness than their male counterparts [37,38,39]. This discrepancy highlights the need for future studies to explore how knowledge is applied in athletes’ dietary practices and whether specific educational interventions can effectively reduce DE risk in high-risk sports. It may also be interesting and beneficial to explore possible differences between sport disciplines, an approach which was not possible in this study due to the heterogeneity of the sample. Concerning this present study, it is also notable that, except for nutrition knowledge, males had preferable (less risky) results in the assessments.

Limited information exists on the disordered eating risk, body composition, and nutrition knowledge of Hungarian elite athletes. This study addresses this gap, providing valuable insights and paving the way for future research in this area. The results suggest that DE risk in athletes is multifactorial, influenced by a combination of sport-specific demands (appearance or weight), body composition expectations for better athletic performance, and, potentially, an inadequate level of nutrition knowledge. Matching this complexity, our results also suggest that identifying athletes at risk of developing DE requires simultaneously using multiple complimentary assessment tools.

The novelty of our research is that we found associations with relatively newly validated questionnaires (DESA-6 and A-NSKQ, between 2018 and 2021) that have not been found in the scientific literature yet. Previous studies have reported that symptoms of DE and ED are unlikely to improve without appropriate treatment [23,24,25]; therefore, early detection is imperative. In this study, although a significant small-to-medium correlation was found between DESA-6H and EAT-26 total scores, DESA-6H identified fewer athletes as being at risk of DE compared to the EAT-26. This association is weaker than the original validation study [43], where a larger sample size and fewer sport disciplines yielded more statistically significant results. These findings highlight the importance of employing multiple validated assessment tools that account for the diversity in athlete populations [12]. However, according to a systematic review of ED screening tools for athletes, DESA-6 was found to be a suitable questionnaire based on validity, reliability, sensitivity, specificity, and receiver–operator–curve (ROC) statistics [70]. Our ROC statistics showed that DESA-6H is a questionnaire that shows high specificity and low sensitivity based on the internationally recommended cut-off point. When choosing different cut-off values, the researcher can decide what to focus on: being sensitive, so that it has a high chance of detecting potential cases (at the cost of misidentifying some athletes who are not actually at risk), or being very specific to avoid “misidentification”, but at the cost of missing real cases. The higher sensitivity cut-off point identified by us in our own sample is a potentially good basis for further testing in larger Hungarian and international samples. In line with other international research, DESA-6 was found to be a suitable tool for assessing DE risk with a considerable diagnostic accuracy [43].

It must be taken into account that the development of DE and ED is influenced by complex, sport-specific pressures around body image and performance [71]. In another study measuring an extended Hungarian sample of elite (n = 91) and recreational athletes (n = 90), correlations of DESA-6H with EAT-26 and EAT-26 subscales showed quite similar results, as there was a significant medium correlation between DESA-6H and the EAT-26 total score, DESA-6H and EAT-26 dieting, and DESA-6H and the bulimia subscale [44]. When comparing these results across gender groups, women scored higher (and had less favorable results) compared to men, with a significant effect size in relation to the EAT-26 total score and the dieting scale. The results by nutritional status showed that, across most variables, normal-weight individuals had less worrisome results compared to their overweight and obese counterparts, with a significant effect size in the case of DESA-6H. The most prominent result was obtained from the comparison of DESA-6H scores between the healthy and the overweight and obese groups, with a very large effect size, which is also the most meaningful result of this present study.

Whilst this study provides important insights into disordered eating (DE) risk, body composition, and nutrition knowledge among elite Hungarian athletes, several limitations should be acknowledged. Firstly, the sample size was relatively small, and, while the study included athletes from diverse sport disciplines, the distribution was uneven across categories and did not cover the full spectrum of sport disciplines. Also, the relatively small sample size did not allow for granularity at the sport discipline or sport group (e.g., power, endurance, skill-based, and team sports) level. This limited sample diversity may have restricted the generalizability of the findings to other athlete populations, especially those outside of Hungary or in sports not represented in our sample. Future research could benefit from larger, more diverse samples to better capture variations in DE risk and nutrition knowledge across diverse cultural contexts. When mentioning the sample size, a problem with the response rate occurred which had been previously mentioned over NSKQ and A-NSKQ development [45,46]. In our study, 16 responses were omitted due to age. Additionally, our reliance on self-report questionnaires, including the DESA-6H, EAT-26, and A-NSKQ, may have introduced response bias, as participants might underreport or misinterpret certain behaviors and knowledge levels. Incorporating clinical interviews or objective measures of dietary intake and eating behaviors could help verify and expand upon these self-reported data [12,72]. Results from DESA-6H and the general nutrition knowledge scale of A-NKSQ should be interpreted with caution due to their less than satisfactory internal consistency reliability.

Further, although body composition measurements were included, we did not assess all physiological indicators of low energy availability (LEA), such as hormonal profiles, which are valuable for understanding the full scope of DE’s impacts on athletes’ health. Petrie and Greenleaf emphasize that the prevalence of eating disorders in athletes is greatly influenced by the methods and the sample selection used [69]. Therefore, it is crucial to conduct additional studies similar to previous ones, which also examined hormonal profiles, serum iron levels, and other metabolic parameters, to better assess the risk of DE and ED [73,74,75]. Future studies may integrate additional physiological assessments (e.g., blood test, body composition measurements, etc.) to provide a more comprehensive overview of DE-related health consequences. Based on our findings and previous research, the EAT-26 and DESA-6 questionnaires are effective screening tools and can form a reliable part of the diagnostic process for assessing ED and DE. Recent studies have also shown that peculiar chest wall conformations may also play a role in the development of ED [76], a factor which we did not take into account. This evidence indicates that individuals with a concave-shaped chest wall conformation and/or pectus excavatum have an increased prevalence of mitral valve prolapse and anxiety disorders [77]; thus, among athletes, those with a concave-shaped chest wall conformation might have an increased prevalence of anxiety and concomitant DE/ED in comparison with those having a more spheroidal chest shape. Incorporating this variable in future studies could enhance the precision of identifying individuals at risk of DE/ED.

Our study also highlights a need to monitor the development of DE risk, as well as exploring the long-term effects of DE risk on athletes’ health and performance. The cross-sectional design does not allow for these. Future research should employ longitudinal designs to monitor athletes over time, examining whether initial DE risk scores predict a later development of eating disorders (ED), health complications, or performance outcomes. Because our findings suggest gaps in nutrition knowledge among athletes, future studies should also examine the impact of targeted nutrition education programs on DE risk and overall athlete well-being, exploring whether improved nutrition knowledge directly mitigates DE behaviors and enhances health and performance in elite sports contexts.

It is crucial to start learning about adequate nutrition as early as possible; thus, in case of junior athletes, involving parents can also increase effectiveness. This may contribute to avoiding/reducing the development of nutrition-related problems (mainly DE and ED). The support of athletes is important not only for sports performance, but also for long-term health preservation. The results, so far, show that knowledge does not reach the acceptable level, and the risk of DE and ED is also present; therefore, urgent intervention is needed which aims to increase athletes’ nutrition knowledge and prevent DE and ED. The sooner we provide appropriate care to those who are already at risk, the greater the chance of reversing the condition.

Athletes and their entourage (e.g., team doctor, nutritionist, psychologist, coach, etc.) should work together to find a balance between the desired bodyweight for optimal athletic performance and health. In doing so, sport nutritionists and dietitians should prioritize tailored nutritional education and support for athletes, especially for those in high-risk sports. Coaches and athletic trainers should also be aware of these risks to better monitor and guide their athletes, helping foster healthier attitudes and behaviors around food and body composition. Sports governing bodies and healthcare providers working with athletes should consider incorporating regular DE risk assessments and nutrition knowledge training as part of standard athletic care.

All data are included in the study.