| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrine Care |
Department of Medicine and Obstetrics and Gynecology, Columbia College of Physicians and Surgeons, New York, New York 10032; and Nutrition Research Center, St. Luke’s-Roosevelt Hospital Center, New York, New York 10025
Address all correspondence and requests for reprints to: Michelle P. Warren, M.D., PH 16-127, Department of Obstetrics and Gynecology, Columbia University, 622 West 168th Street, New York, New York 10032. E-mail: . mpw1{at}columbia.edu
Abstract
Osteopenia, which is correlated with amenorrhea and poor nutritional habits, has been well documented in elite ballet dancers. Estrogen replacement therapy and recovery from amenorrhea have not been associated with normalization of bone density. Thus, the osteopenia may be related to changes brought about by chronic dieting or other factors, such as a hypometabolic state induced by poor nutrition. The purpose of this study was to investigate the relationship of chronic dieting and resting metabolic rate (RMR) to amenorrhea and bone density.
RMR, bone density, eating disorder assessments, leptin levels, and complete menstrual and medical histories were determined in 21 elite ballet dancers and in 27 nondancers (age, 20–30 yr). No significant correlations were found between high EAT26 scores, a measure of disordered eating, and RMR, bone densities, body weight, body fat, or fat-free mass. However, when RMR was adjusted for fat-free mass (FFM), a significant positive correlation was found between RMR/FFM and bone density in both the arms (P < 0.001) and spine (P < 0.05) in ballet dancers, but not in the normal controls. The dancers also demonstrated significantly higher EAT scores (22.9 ± 10.3 vs. 4.1 ± 2.4; P < 0.001) and lower RMR/FFM ratios (30.0 ± 2.2 vs. 32.05 ± 2.8; P < 0.01). The only variable to predict lower RMR/FFM in the entire sample was ever having had amenorrhea; this group had significantly higher EAT scores (18.0 ± 13.5 vs. 10.3 ± 10.2; P < 0.05), lower leptin levels (4.03 ± 0.625 vs. 7.10 ± 4.052; P < 0.05), and lower bone mineral density in the spine (0.984 ± 0.11 vs. 1.10 ± 0.13; P < 0.05) and arm (0.773 ± 0.99 vs. 0.818 ± 0.01; P < 0.05).
We hypothesize that the correlation between low RMR and lower leptin levels and bone density may be more strongly related to nutritional habits in ballet dancers, causing significant depression of RMR, particularly for those with a history of amenorrhea.
POSTMENOPAUSAL BONE LOSS due to estrogen deficiency has been well documented. Estrogen replacement therapy has been shown to slow the rate of bone loss in postmenopausal women. Estrogen acts by decelerating bone resorption, which leads to a slower rate of bone turnover and thus prevents bone loss (1). Studies in premenopausal young women show a strong relationship between amenorrhea and low bone density (2, 3, 4, 5, 6, 7, 8, 9), with lack of normalization with return of menses (10, 11) and no response or a minimal response to estrogen-progestin therapy (12, 13). Recent work, in addition, has challenged the concept that estrogen deficiency is the primary cause of osteopenia, particularly in amenorrheic athletes, as the mechanisms of bone turnover appear to differ from those of hypoestrogenism (14, 15, 16, 17). Such studies suggest that factors other than lack of estrogen may be operative.
Recent evidence suggests that women with hypothalamic amenorrhea, particularly those with an exercise-induced disorder, may be in a calorically depressed state (18, 19, 20, 21, 22). Low energy intake is known to be associated with a decrease in resting metabolic rate (RMR) (23, 24, 25, 26). Work on obesity suggests that the RMR may remain depressed in obese individuals who maintain a body weight that is lower than their natural (obese) weight (27) and restrict their caloric intake to do so. Even women who maintain a normal weight may suffer an energy deficit; their normal weight is probably maintained by a decrease in metabolic rate (28). Animal models indicate that the repeated gaining and losing of weight or the combination of fasting and refeeding causes a heightened food efficiency that persists beyond the periods of deprivation, and studies of male and female athletes have confirmed this (29, 30). A decrease in metabolic rate may be one of the mechanisms of the heightened food efficiency. As the RMR has been reported to be depressed in athletic amenorrheics (28), and hypoleptinemia has been associated with low bone mineral density (BMD) in hypothalamic amenorrhea (17), the purpose of this study was to determine whether changes in the RMR are related to restrictive eating patterns and low bone density in athletic amenorrheic subjects, a group known to maintain low body weight. This also may explain the depression in bone density seen even in women of normal weight who have experienced amenorrhea (17).
Subjects and Methods
Subjects
Twenty-one elite, professional ballet dancers were recruited from highly competitive ballet companies and studios in New York City. The dancers were either members of a professional ballet company or, if not employed, danced more than 20 h/wk. Subjects were categorized into three groups: those who were normally menstruating, those with amenorrhea in the past, and those currently amenorrheic. Those with amenorrhea in the present or past were defined as "ever having had" a history of amenorrhea. Amenorrhea was defined as cessation of menstruation for longer than 3 months, and all had a history of amenorrhea for at least 6 months. The dancers’ ages ranged from 20–30 yr, with a mean age of 23.2 ± 2.8 yr. All dancers were screened before entering the study, and those with any known illness that could affect body composition or bone density were excluded. Three dancers were taking oral contraceptives, and one dancer was a light smoker. Of those taking oral contraceptives, one had taken them for 5 months, one for 3 yr, and the third for an unspecified period of time. An effort was made to exclude both dancers and controls with a history of eating disorders. However, we enrolled one dancer who described a history of anorexia 6 yr before the study.
Twenty-seven normal controls were recruited from a local university and hospital staff. All were healthy and without significant medical histories, and two were taking birth control pills. One had taken birth control pills for less than 1 yr, and the other for an unknown length of time. Controls were matched with dancers according to height and weight. Their ages ranged from 20–30 yr, with a mean age of 24.5 ± 2.6 yr. Four controls had a history of amenorrhea and were considered a separate subset. The protocol was approved by the institutional review board, and procedures were followed according to approved ethical guidelines. All subjects gave written informed consent.
Bone density, body fat, and fat-free mass (FFM)
Each subject underwent one total body bone density scan by dual energy x-ray absorptiometry. The reports from the DPX scanner (Lunar Corp., Madison, WI) were analyzed using version 3.6 software and were used to determine regional (arm, leg, and spine) and total body bone densities, total body bone mineral content, and total body fat percentage (31). When measured by dual energy x-ray absorptiometry, percent body fat is independent of BMD, as this value is measured directly by recognized standard means in fat depots at sites where bone is not present (31). FFM is calculated by the following: fat mass = weight x fat % and FFM = weight - fat mass. Bone density measurements can be artificially altered by increased tissue thickness, but are minimally affected in lean subjects (31).
RMR
Each subject’s RMR was measured in a respiratory chamber-indirect calorimeter. The chamber was equipped with a high precision gas (oxygen and carbon dioxide) exchange measurement system. A linear state space model converts gas exchange measurements to energy expenditure estimates. RMR was measured early in the morning, with subjects in a fasted state for 12–15 h before the experiment (32). Subjects remained in the chamber for 1 h, and RMR calculations were based on the average of 3 10-min readings after 30 min had elapsed. As differences in RMR have been documented depending on the phase of the menstrual cycle (33), in both groups resting metabolic measurements were taken during the luteal phase of subjects who were menstruating normally (16 dancers and 26 controls). Ovulation was predicted based on menstrual cyclicity. As the metabolic rate depends on metabolically active tissues, all of them contained in the FFM, we controlled for differences in FFM between dancers and controls by using the RMR/FFM ratio.
Eating disorder assessment
Each subject answered a questionnaire, the EAT26 (34, 35), designed to assess abnormal eating behaviors, which was used to measure the existence and severity of anorexic thoughts and behaviors. The dancers also completed the Mizes Anorectic Cognition Scales (MACS) (36) to verify the accuracy of the EAT26 scores. The EAT26, the questionnaire used for our analysis, separates its questions into three groups: dieting behavior, bulimic behavior, and oral control. It measures dieting behavior through questions concerning avoidance of fattening foods and obsessions with losing weight, bulimic attitudes through questions concerning binge/purge behavior and preoccupation with food, and oral control through questions concerning the ability to control eating and the perceived pressure from others to gain weight. Each question receives a score from 1–6, with 1 meaning "this is never true of me" and 6 meaning "this is always true of me." The score is totaled based on a weighted scale, an answer of 4–6 being weighted and an answer of 1–3 not being weighted. A score above 30 on the EAT26 represents pathological behavior. With increasing scores, the behavior becomes more pathological. This scale addresses present behavior and does not detect eating disorders in the past. The MACS questionnaires assess anorexic and bulimic thoughts and behaviors (36).
Leptin levels
Leptin was measured by an immunoradiometric assay (Linco Research, Inc., St. Charles, MO). The interassay coefficient of variation was 3.6–4.0%, the intraassay coefficient of variation was 3.04–4.03%, and the lower limit of sensitivity was 0.5 ng/ml. Midafternoon cortisol was measured in serum by RIA (Diagnostic Products, Los Angeles, CA). In this assay, the intra- and interassay coefficients of variation were 2.5–8.0% and 4.5–6.36%, respectively. The detection limit was less than 0.2 g/dl (17).
Medical and menstrual histories
Each subject also filled out a medical questionnaire. The questionnaire was designed to assess general medical and menstrual history with questions concerning age of menarche, periods of amenorrhea, and body weight at that time.
Statistics
We used factorial ANOVAs with repeated measures to evaluate differences within and between groups while controlling for possible differences between individual subjects, as well as group t tests to localize between-group differences. Multiple comparisons were made using a modified least squares difference method available in SPSS software (Statistical Package for Social Sciences, SPSS, Inc., Chicago, IL). Correlations between variables and bone mass changes were tested using the Pearson linear correlation. Results were analyzed considering the amenorrheic groups as well as examining dancers as a separate group.
Results
Anthropometric measures for our subjects and data for RMR, bone densities, and EAT scores are depicted in Table 1
. There was no difference in any of the anthropometric measures between dancers and controls, including weight, height, body mass index, and FFM. Mean RMR and RMR/FFM did not differ in the two groups as a whole. EAT scores were higher in dancers (P < 0.001), but BMDs did not differ. We found significant correlations in dancers between RMR and arm BMD (ABD; P < 0.001), spine BMD (SBD; P < 0.01), leg BMD (P < 0.05), and total bone mineral content (P < 0.001). Controls also showed correlations of densities measured: ABD (P < 0.01), SBD (P < 0.05), leg BMD (P < 0.05), and total bone mineral content (P < 0.05; Table 2). RMR also correlated with weight (P < 0.05) and percent fat mass in the entire group. However, when controlling for weight and percent fat mass, the three BMD values remained significantly correlated (P < 0.05) with RMR only in the dancers.
|
|
and Figs. 1 and 2).
|
|
More ballet dancers had amenorrhea at the time of the study (23.8%, or 5 of 21) than did controls (4%, or 1 of 27; P < 0.05). Means for RMR and RMR/FFM did not differ when present amenorrheic subjects were compared with normal subjects. When including subjects with a past history of amenorrhea, more dancers (11 of 21, 52%) than controls (4 of 27, 14.8%) reported ever having had amenorrhea (P < 0.05; Table 3). The only variable that predicted lower RMR/FFM in the entire sample was ever having had amenorrhea. This group also had significantly higher EAT scores (18.0 ± 13.5 vs. 10.3 ± 10.2; P < 0.05) with lower RMR/FFM (29.0 ± 2.5 vs. 31.6 ± 2.4; P < 0.001) and lower BMD in the spine (0.984 ± 0.11 vs. 1.10 ± 0.13; P < 0.05) and arm (0.773 ± 0.99 vs. 0.818 ± 0.01; P < 0.05) as well as lower leptin levels (4.03 ± 0.625 vs. 7.10 ± 4.052; P < 0.05). When the group of amenorrheic dancers were divided based on past, present, or never having manifested amenorrhea, those with past amenorrhea had significantly lower RMR/FFM than those with no history of amenorrhea (28.4 ± 2.4 vs. 31.3 ± 1.1), although EAT scores did not differ in these groups (19.9 ± 9.8, 27.8 ± 13.7, and 22.3 ± 8.4, respectively). Thus, apart from EAT scores, a past history of amenorrhea may selectively affect RMR/FFM.
|
We have observed a strong relationship between bone density and RMR and postulate that the osteopenia in this group is related to a chronic adaptation to low energy intake. The association of BMD with metabolic status has not been previously reported. We also report that a depressed metabolic rate is associated with lower BMD and leptin levels in women ever having had amenorrhea. We believe that a depressed metabolic state, low bone density, and low leptin levels may all be a manifestation of the same underlying cause, poor nutrition and long-term restricted energy intake, resulting from the high propensity of ballet dancers to diet. The subdivision of the dancers and controls into those with a history of amenorrhea and those without, was done to further investigate the effects of chronic nutritional deprivation on RMR and BMD, as the dietary restriction is especially profound in dancers with amenorrhea. We report significantly lower metabolic rate and BMD in this subgroup; however, the number of subjects was too small to measure a relationship between RMR/FFM and BMD in the amenorrheics. A similar effect could be expected in amenorrheic nondancers; however, our subject group was small. Restriction of nutritional intake has also been associated with stress fractures in ballet dancers and may be affecting bone quality as well. This may explain the tendency of dancers to fracture, even when menstrual function is reported to be normal (38), and has important clinical implications. Although the number of subjects in this study is small, our results are significant and merit further study, as published data suggest that women with a history of amenorrhea continue to be at risk for osteopenia even after the initial insult. This history is now a risk factor for osteoporosis (39).
We hypothesized that the EAT26 and/or the MACS questionnaires could detect long-term dieters with pathological eating behaviors and correlate this with RMR, as chronic food restriction is known to depress the metabolic rate (27). Our EAT26 scores correlated highly with the scores from the MACS, but both questionnaires may accurately describe the present without regard to a long-term history of dieting behavior. If long-term dieting affects the RMR, this may explain why the metabolic rate does not correlate with present EAT or MACS scores. The study would have been strengthened by the inclusion of a measure of long-term dieting and a food record to assess nutritional status. The differences in metabolic rate and leptin levels found between those with past amenorrhea and those never having had this problem suggest that long-term dieting may have a more profound effect on metabolic rate than current dietary behavior, as past amenorrhea is strongly associated with dieting behavior. The low metabolic rates found in amenorrheic dancers were particularly surprising, as FFM or lean body mass is the major determinant of metabolic rate (37), and lean body mass tends to be higher in athletic groups as it was in our dancers (although not significantly so).
A history of amenorrhea, particularly past history, was the only predictor of lower RMR/FFM, suggesting that this problem affects RMR long-term, even with the recovery of cyclicity. This is consistent with reports that prior menstrual history is the best predictor of current BMD and that a history of amenorrhea is associated with lower BMD (40, 41). The long-term effects of dieting and associated secondary menstrual disturbance would also explain the failure to normalize bone mass with return of menstrual cyclicity (10) if adaptation to malnutrition was the dominant factor. Our studies suggest that a decrease in RMR is highly predictive of menstrual dysfunction as well as the associated severe side-effect, osteopenia. Although the etiology of this problem is unclear, our data suggest that restrictive eating behavior may be the dominant causal factor. The continued depression of the metabolic rate beyond the amenorrheic insult suggests resulting persistent effects on metabolism.
The complication of osteopenia is much more prevalent in ballet dancers and other amenorrheic athletes (2) than in the normal population (13, 42). As amenorrhea is also frequent in dancers, and estrogen deficiency is associated with both eating disorders and ballet dancing and can also be associated with osteopenia, the hypothesis that estrogen deficiency was a primary cause of bone deficiency has been explored, with largely negative results. Estrogen replacement does not normalize the low bone densities of dancers when replacement is given experimentally (13).
In addition, the metabolic profile of bone turnover seen with the osteopenia characteristic of amenorrhea is not consistent with the high turnover typical of hypoestrogenism. Studies of biochemical markers of bone turnover in exercise-induced amenorrhea suggest a pattern different from estrogen deficiency (14, 15, 16). Estrogen deficiency is associated with increased bone turnover with excessive resorption that responds to estrogen replacement (43). Recent studies suggest low bone turnover in exercise-induced athletic amenorrhea and relate these depressed bone markers to signs of energy deficiency (14).
Leptin, a hormone secreted by the fat cells, is disproportionally lowered by fasting (44, 45, 46, 47) and is an independent regulator of metabolic rate (44). Leptin receptors have been found in bone, may be important to osteoblastic function (48, 49), and have been found to be depressed in hypothalamic amenorrhea (17). Thus, dieting may indirectly affect bone metabolism through a mechanism involving a leptin pathway.
Low leptin levels have been associated with amenorrhea (17, 50) as well as with disordered eating (17). The recent discovery of leptin receptors on bone (48) suggests that leptin may be involved in skeletal regulation and possibly be centrally controlled, most likely in the hypothalamus (51, 52, 53). It may even be locally controlled (54), perhaps by local adipocytes (52, 53, 55). This provides a possible mechanism for an interaction between the nutritional and metabolic bone axes and suggests that leptin may function as a physiological regulator of bone mass. Therefore, a mechanism other than estrogen deficiency may also account for the low BMD seen in women whose amenorrhea is associated with caloric deficiency, nutritional insults, and exercise (17, 38, 56, 57). The suppression of leptin levels may explain the persistence of the osteopenia beyond the period of menstrual cyclicity when the hypoestrogenism has reversed.
We explore here an alternate hypothesis of osteopenia due to chronic dieting to relate the osteopenia in ballet dancers to a hypometabolic state induced by severe dieting. Given that extensive exercise, including the resistance exercises favored by osteoporosis treatment regimens, are an unavoidable aspect of a dancer’s training, this group should be less, not more, susceptible to osteopenia. Our group of ballet dancers suggests that ballet dancers may be exhibiting osteopenia due to a more basic mechanism relating to chronic malnutrition induced by severe nutritional restriction, as has recently been suggested by others (14, 15, 16). The diet of dancers is also chronically low in calcium (58, 59, 60), but a study with calcium supplements did not normalize bone density (13). Vitamin D levels were measured in one study and found to be normal (42). Thus, we suggest that other metabolic factors are operative, possibly a metabolic signal to the bone triggered by malnutrition.
Young women who consistently diet or exist in a calorically deprived state may slow their metabolic rates (23, 27, 28), an alteration that could lead to factors affecting bone metabolism. Metabolic indexes of bone turnover have not been consistently examined in these young women, but a report of ballet dancers with osteopenia showed a decrease in bone turnover (61). In addition, hypothalamic amenorrhea of the type seen in dancers is associated with low leptin levels (17, 62).
Ballet dancers consistently weigh 10–12% below ideal weight and diet to maintain this low weight. The measured decrease in relative RMR is probably a reflection of this behavior, which is, in turn, related to low bone density. Further research is necessary to identify the exact mechanism causing the osteopenia associated with amenorrhea, as this may open new avenues for treatment and prevention.
Acknowledgments
Footnotes
This work was supported in part by NIH/NICHHD Grants R01-HD-22171-01 to -07.
Abbreviations: ABD, Arm bone mineral density; BMD, bone mineral density; FFM, fat-free mass; MACS, Mizes Anorectic Cognition Scales; RMR, resting metabolic rate; SBD, spine bone mineral density.
Received August 27, 2001.
Accepted March 7, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  J. Dominguez, L. Goodman, S. Sen Gupta, L. Mayer, S. F. Etu, B T. Walsh, J. Wang, R. Pierson, and M. P Warren Treatment of anorexia nervosa is associated with increases in bone mineral density, and recovery is a biphasic process involving both nutrition and return of menses Am. J. Clinical Nutrition, July 1, 2007; 86(1): 92 - 99. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. W. Hamrick Leptin and Bone: A Consensus Emerging? IBMS BoneKEy, March 1, 2007; 4(3): 99 - 107. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Ruggiero and L. Ferrucci The endeavor of high maintenance homeostasis: resting metabolic rate and the legacy of longevity. J. Gerontol. A Biol. Sci. Med. Sci., May 1, 2006; 61(5): 466 - 473. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. De Souza and N. I. Williams Physiological aspects and clinical sequelae of energy deficiency and hypoestrogenism in exercising women Hum. Reprod. Update, September 1, 2004; 10(5): 433 - 448. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
