Body mass index and athletic performance in elite female gymnasts.

Sherman, Robert Trattner ; Thompson, Ron A. ; Rose, Jennifer S. 等

Recently, considerable emphasis has been given in the literature to eating disorders in athletes (Brownell & Rodin, 1992; Thompson & Sherman, 1993). Much of that emphasis has focused on female gymnasts, presumably because the appearance and performance demands of gymnastics lead these athletes to diet in an effort to attain or maintain suboptimal weights (Rosen & Hough, 1988; Thompson & Sherman, 1993; Yates, 1991). Dieting and attempted weight loss play a critical role in the development of eating disorders (Garfinkel & Garner, 1982; Hsu, 1990). Thus, attempts to lose weight increase the athlete's risk of developing such a disorder.

Concern among female gymnasts regarding their weight has been well documented (Harris & Greco, 1990; Loosli, Benson, Gillien, & Bourdet, 1986; O'Connor, Lewis, & Kirchner, 1995; Rosen & Hough, 1988). This concern is reflected in a trend over the past thirty years showing that U.S. Olympic female gymnasts have become significantly smaller in terms of body size and weight (Nattiv & Mandelbaum, 1993). In order to comply with the size and weight standards that are emphasized in gymnastics, many gymnasts are attempting to lose weight. Harris and Greco (1990) reported that 61% of their sample were trying to lose weight, while Rosen and Hough (1988) found that all of the gymnasts in their study were dieting. Relatedly, other studies with gymnasts have found the frequent use of pathogenic weight control methods (Petri & Stoever, 1993; Rosen, McKeag, Hough, & Curley, 1986).

Many of the gymnasts in the aforementioned studies reported that they were dieting or attempting to lose weight in order to improve their athletic performance. However, the literature with respect to the relationship between decreased body weight and performance is less than conclusive. At least part of the equivocal findings in this area can probably be attributed to the fact that some studies have been concerned with "body weight," while others have looked at the athlete's body composition or level of "body fat." Although the two terms are related, they are not synonymous or interchangeable.

The issues with respect to body weight vs. body fat and athletic performance have probably been addressed most thoroughly by Wilmore (1992a; 1992b). He maintains that a negative relationship exits between body weight or body composition and athletic performance where jumping for height or distance or running for speed or endurance are involved. Wilmore suggests that the higher the body weight, generally the poorer the athletic performance. He clarifies the issue somewhat by saying that, although weight can limit performance, body fat tends to be the crucial factor and that the leaner athlete typically performs better. Studies by Cureton and Sparling (1980) and by Pate, Barnes, and Miller (1985), using distance runners, support his contention. A study by Clark, Nelson, and Evans (1988), however, challenges the belief that low body weight enhances performance for distance runners. The investigators found no relationship between the fastest racing times and either body weight or body mass index in the 93 elite distance runners they surveyed.

Although several studies have investigated anthropometric characteristics of outstanding or elite gymnasts (Caldarone, Leglise, Giampietro, & Berlutti, 1986; Claessens et al., 1991; Pool, Binkhorst, & Vos, 1969), there is a paucity of research that has investigated the relationship between body weight or body composition and gymnastics performance. A study by Pool, Binkhorst, and Vos (1969) looked at physiological performance variables in elite gymnasts, such as running speed, jumping height, and hand strength, which are related to gymnastics skills rather than to gymnastics performance per se. They found that body weight did not correlate with the performance of these behaviors. Only a study by Falls and Humphrey (1978) looked at the relationship between body fat composition and actual gymnastics performance. Specifically, Falls and Humphrey found that college gymnasts who placed first, second, or third in a national competition had significantly lower body fat than those who did not place. Interestingly, the gymnasts who placed in the competition did not differ in terms of height and weight from those who did not place.

The purpose of the present study was to determine the nature of the relationship between thinness and performance in elite gymnasts. That is, do gymnasts need to be thin in order to perform well? More specifically, is there a relationship between thinness and gymnastics performance? If so, what is the nature of that relationship? Also, can gymnasts be too thin with respect to performance?

Certainly, an ethical question regarding the appropriateness of asking or allowing an athlete to engage in the potentially harmful process of losing weight can be raised. That question has been asked and addressed previously (Thompson & Sherman, 1993) and will not be recounted here.

Method

Subjects

Subjects were the top 36 finishers in the all-around competition at the 1991 World Gymnastics Championships held in Indianapolis, Indiana, on September 6-15. These athletes represented 17 countries. The competition was composed of a field of 233 competitors representing 40 countries.

Procedure

All-around scores were selected as the measure of performance. In gymnastics, all around scores are derived by adding the scores for the four individual events - vault, balance beam, uneven bars, and floor exercise. All-around scores were selected for use in order to eliminate any possible bias of a particular body size in any single event. Additionally, the gymnast's rank in the competition was determined by her all-around score.

Height, weight, age, and nationality of each of the 36 competitors were provided by each gymnast on an athlete biographical information identification form, turned into the 1991 World Gymnastics Championships' organizers prior to competition. A measure of thinness - Quetelet's Body Mass Index (BMI) - was computed (weight in kilograms/height in meters squared) for each athlete from these self-reported heights and weights. BMI was chosen as the measure of thinness for several reasons. Most important, BMI is not only a measure of weight; it also correlates highly (0.8) with direct measures of body fat (Bray, 1986). Additionally, BMI correlates minimally with height (Billewicz, Kemsley, & Thomson, 1962; Neggers, Stitt, & Roseman, 1989) and allows for comparisons within a population (Clark et al., 1988).

Results

Descriptive Statistics

Complete data were available for 34 of the top 36 gymnasts. BMI's for these 34 gymnasts ranged from 14.63 to 20.04 with a mean of 17.20 and a standard deviation (sd) of 1.52, The mean age of the gymnasts was 17 (sd = 2.04), and ages ranged between 13 and 22 years. A bivariate scatterplot indicated that there was a curvilinear relationship between BMI and rank.

Predicting Performance from BMI

Hierarchical regression analysis was used to predict gymnasts' rank performance. The dependent variable, rank, was ordinally-scaled. Thus, in order to transform it into a variable with a normal distribution, rank was normalized following the procedure outlined by Cohen and Cohen (1983). Because age was related to both BMI (r(34) = .44, p [less than] .01) and rank ([r.sub.s] (34) = .23, p = .09) in this sample, it was entered as a predictor in the first step of the hierarchical regression analysis predicting rank. This was done to evaluate the relationship between BMI and rank, controlling for age. The linear predictor, BMI, was entered in the second step of the analysis, followed by the quadratic predictor (BMI squared) in the third step. The quadratic predictor was entered to evaluate the curvilinear relationship between BMI and rank. This was assessed by examining the change in the proportion of variance accounted for in rank when the quadratic term was entered into the equation.

Results of the hierarchical regression analysis are shown in Table 1. Neither age nor the linear relationship between BMI and rank was significant. However, the quadratic effect, representing the curvilinear relationship between BMI and rank, was significant (p [less than] .05), and the proportion of variance accounted for in rank was significantly increased when the quadratic term was entered into the equation. Age and the linear BMI predictor explained 14% of the variance in rank. The variance in rank explained by all three predictors (age, linear BMI, and quadratic BMI) was 29%.

The significant, positive regression coefficient for the quadratic effect suggests that BMI is indeed a significant predictor of rank, but that the relationship is curvilinear. This indicates that a lower BMI appears to be related to better performance up to a point in this sample, but that performance becomes more negative as BMI becomes very low. Within the BMI range of this sample, those with higher and lower BMI did not do as well as those in the middle of the range. For example, the top three finishers ranked 27th, 8th, and 19th, respectively in terms of BMI. Similarly, the two competitors with the lowest BMI's finished 29th and 30th of the 34 competitors for whom complete BMI and Athletic data were available. They finished 31st and 32nd of the original 36 competitors.

[TABULAR DATA FOR TABLE 1 OMITTED]

Discussion

The purpose of the present study was to determine the nature of the relationship between body mass index and performance among elite female gymnasts. The findings of the present study suggested that the nature of the relationship between thinness and performance in the present study was curvilinear. Although there was a trend toward thinner athletes performing better, the athletes who performed best were neither the thinnest nor the heaviest. The effects of thinness apparently reached a point of diminishing returns. That is, a lower BMI appeared to be related to better performance, but performance was more negative as BMI became very low.

With regard to the possibility that performance can be negatively affected as BMI becomes too low, Wimore (1992a) maintains that, even though weight loss typically results in enhanced athletic performance, there is a point beyond which continued weight loss will lead to deterioration of performance. This decrement in performance can probably be explained by the fact that the lower the individual's body fat, the greater the likelihood that weight loss will be at the expense of lean tissue and body fluid. Based on results from studies with wrestlers who have experienced such losses (e.g., Webster, Rutt, & Weltman, 1990), we would expect a decrease in strength and endurance. At some point, the athlete may simply be too unhealthy and weak to perform well.

Comparisons with the earlier study by Falls and Humphrey (1978) are difficult for several reasons. Their collegiate sample was not only heavier (x = 55.1 kg) than our elite international one (x = 38.77 kg), it also most certainly had a higher mean percent body fat. Although body fat percents were not available for our elite athletes, it has been suggested that a range of body fat of 8-16k is representative of elite female gymnasts (Wilmore, 1992b). Comparatively, the gymnasts in the Falls and Humphrey study who placed in the competition had a mean body fat percent of 16.82, whereas the nonplacers had a mean body fat percent of 18.41. If weight or leanness is as important as it appears to be in gymnastics, gymnasts at different weight or body composition levels may experience different levels of pressure regarding weight loss. Also, our elite sample was younger (x = 16.99 years) than Falls and Humphrey's college group (x = 19.4), and age may also affect the need or pressure to lose weight.

An additional factor affecting generalizability of our results is that this sample was composed of gymnasts competing at the elite international level. An elite sample was used primarily because, at least intuitively, it appears to be the most at risk for a need to be inordinately thin. At the elite level, one-tenth of a point in a competition may mean the difference between making and not making a national team. Or, it may mean the difference between winning an international competition or finishing without a medal. In such cases, an athlete and her coach may believe that the loss of a few pounds or a percent loss in body fat could provide the necessary margin, thus making it worth any health risk that might be incurred through dieting or another weight loss endeavor. Given the issue of competition pressure, as well as those of thinness and age already discussed, replication of this study with gymnasts at lower levels of competition is certainly warranted.

A possible limitation of the present study involves the use of self-reported heights and weights to derive BMI measures. It is difficult to determine how reliable the self-reported data were. At the same time, these measures were extracted from official forms completed by competitors for the championships' organizers. The competitors were unaware that this information was being used for any purpose. There was no obvious advantage to the competitors to provide inaccurate data on these forms. Also, in at least one study investigating perceived weight status in college students (Sciacca, Melby, Hyner, Brown, & Femea, 1991), correlation coefficients and slopes of regression lines computed between reported height and actual height, as well as between reported weight and actual weight, indicated no differences between reported and actual heights and weights. Additionally, a recent study by Sherman, Rose, and Thompson (1995) found that there was no difference in self-report and actual height and weight between those college students who were informed that actual height and weight measurements would be taken and those who were not informed. Furthermore, the correlation between perceived height and weight and actual height and weight in this sample (with conversion to BMI) was very high (r = .93).

The findings of the present study have implications for elite gymnasts' perceived need to be thinner for performance reasons. Many athletes, coaches, and athletic personnel believe that "thinner is better." Our results suggest that thinner may not always be better. That is, low BMI does not ensure enhanced performance. In fact, it may be related to decreased performance levels. Perhaps this information could be used by athletic personnel in making a decision whether to have an athlete begin a weight- or fat-reducing process. In essence, it could decrease the credence currently given to the often used rationale that weight or body fat loss will result in enhanced athletic performance. Most importantly, it might preclude at least some athletes from pursuing potentially unhealthy suboptimal weights and engaging in pathogenic weight loss methods often used to attain them. This seems particularly salient for the elite gymnast in that she tends to be inordinately thin. In the present study, 6 (17.65%) of the 34 gymnasts had BMI's below 16, and a BMI below 16 has been reported to be indicative of undernutrition (Beumont, Al-Alami, & Touyz, 1988).

References

Beumont, P., Al-Alami, M., & Touyz, S. (1988). Relevance of a standard measure of undernutrition to the diagnosis of anorexia nervosa: Use of Quetelet's body mass index (BMI). International Journal of Eating Disorders, 7, 399-405.

Billewicz, W.Z., Kemsley, W.F.F., & Thomson, A.M. (1962). Indices of adiposity. British Journal of Preventive and Social Medicine, 16, 183-188.

Bray, G.A. (1986). Effects of obesity on health and happiness. In K.D. Brownell & J.P. Foreyt (Eds.), Handbook of eating disorders (pp. 3-44). New York: Basic Books.

Brownell, K.D., & Rodin, J. (1992). Prevalence of eating disorders in athletes. In K.D. Brownell, J. Rodin, & J.H. Wilmore (Eds.), Eating, body weight, and performance in athletes: Disorders of modern society (pp. 128-145). Philadelphia: Lea & Febiger.

Caldarone, G., Leglise, M., Giampietro, M., & Berlutti, G. (1986). Anthropometric measurements, body composition, biological maturation and growth predictions in young female gymnasts of high agonistic level. Journal of Sports Medicine, 26, 263-273.

Claessens, A. L., Veer, F. M., Stijnen, V., Lefevre, J., Maes, H., Steens, G., & Beunen, G. (1991). Anthropometric characteristics of outstanding male and female gymnasts. Journal of Sports Sciences, 9, 53-74.

Clark, N., Nelson, M., & Evans, W. (1988). Nutrition education for elite female runners. Physician and Sportsmedicine, 16, 124-136.

Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences. New Jersey: Lawrence Erlbaum.

Cureton, K.J., & Sparling, P.B. (1980). Distance running performance and metabolic responses to running in men and women with excess weight experimentally equated. Medicine and Science in Sports and Exercise, 12, 288-294.

Falls, H.B., & Humphrey, L.D. (1978). Body type and composition differences between placers and nonplacers in an AIAW gymnastics meet. Research Quarterly, 49, 3843.

Garfinkel, P.E., & Garner, D.M. (1982) Anorexia nervosa: A multidimensional perspective. New York: Brunner/Mazel.

Harris, M.B., & Greco, D. (1990). Weight control and weight concern in competitive female gymnasts. Journal of Sport and Psychology, 12, 427-433.

Hsu, L. K G. (1990). Eating disorders. New York: Guilford.

Loosli, A.R., Benson, J., Gillien, D.M., & Bourdet, K. (1986). Nutrition habits and knowledge in competitive adolescent female gymnasts. Physician and Sportsmedicine, 14, 118-130.

Nattiv, A., & Mandelbaum, B.R. (1993). Injuries and special concerns in female gymnasts. Physician and Sportsmedicine, 7, 69-82.

Neggers, Y.H., Stitt, K.R., & Roseman, J.M. (1989). Obesity: Problems with definition and prevalence. The Journal of Obesity and Weight Regulation, 8, 119-135.

O'Connor, P.J., Lewis, R.D., & Kirchner, E.M. (1995). Eating disorder symptoms in female college gymnasts. Medicine and Science in Sports and Exercise, 27, 550-555.

Pate, R.R., Barnes, C., & Miller, W. (1985). A physiological comparison of performance-matched female and male distance runners. Research Quarterly in Exercise and Sport, 56, 245-250.

Petri, T.A., & Stoever, S. (1993). The incidence of bulimia nervosa and pathogenic weight control behaviors in female collegiate gymnasts. Research Quarterly for Exercise and Sport, 64, 238-241.

Pool, J., Binkhorst, R.A., & Vos, J.A. (1969). Some anthropometric and physiological data in relation to performance of top female gymnasts. Internationale Zeitschrift fur Angewandte Physiologie einschliesslich Arbeitsphysiologie, 27, 329-338.

Rosen, L.W., & Hough, D.O. (1988). Pathogenic weight-control behaviors of female college gymnasts. Physician and Sportsmedicine, 16, 141-144.

Rosen, L.W., McKeag, D.B., Hough, D.O., & Curley, V. (1986). Pathogenic weight-control behavior in female athletes. Physician and Sportsmedicine, 14, 79-86.

Sciacca, J.P., Melby, C.L., Hyner, G.C., Brown, A.C., & Femea, P.L. (1991). Body mass index and perceived weight status in young adults. Journal of Community Health, 16, 159-168.

Sherman, R.T., Rose, J.S., & Thompson, R.A. (1995). Evaluation of a bogus pipeline procedure for eating disorder research. Unpublished manuscript.

Thompson, R.A., & Sherman, R.T. (1993). Helping athletes with eating disorders. Champaign, IL: Human Kinetics.

Webster, S., Rutt, R., & Weltman, A. (1990). Physiological effects of a weight loss regimen practiced by college wrestlers. Medicine and Science in Sports and Exercise, 22, 229-234.

Wilmore, J.H. (1992a). Body weight and body composition. In K.D. Brownell, J. Rodin, & J.H. Wilmore (Eds.), Eating, body weight, and performance in athletes: Disorders of modern society (pp. 77-93). Philadelphia: Lea & Febiger.

Wilmore, J.H. (1992b). Body weight standards and athletic performance. In K.D. Brownell, J. Rodin, & J.H. Wilmore (Eds.), Eating, body weight, and performance in athletes: Disorders of modern society (pp. 315-329). Philadelphia: Lea & Febiger.

Yates, A. (1991). Compulsive exercise and the eating disorders. New York: Brunner/Mazel.