Competitive female athletes are subjected to significant training loads in both intensity and duration. The goal of training is to produce physical adaptations to achieve a higher level of athletic performance. Higher training loads (exercise energy expenditure [EEE]) demand greater amounts of energy intake (EI) for energy balance (EB) to be maintained. Adequate energy reserves are necessary to reduce the risk for nutritional repartitioning, as seen in both animal and human models (10,36). The maintenance of EB has been long theorized as an essential component to maximize sport performance. To date, little is known about the bioenergetic status, hormonal environment, and sport performance of elite female athletes during a competitive season, in which bioenergetic status refers to variables affecting the energy environment such as resting energy expenditure (REE), total triiodothyronine (TT3), insulin-like growth factor (IGF-1), and energy availability (EA).
Female athletes have been reported to exhibit symptoms of clinical significance secondary to the unsuccessful maintenance of EB and subsequent nutritional repartitioning that result in energy conservation (28,36). The clinically significant medical symptoms include impaired menstrual, bone, and cardiovascular health (5,20,28). Suppressed estradiol (E2) and progesterone (P4) concentrations have been reported in female athletes (5,20) and recreationally active women (9,10) in association with the above-mentioned medical sequelae. Cross-sectional findings clearly support the interplay between energy conservation (nutritional repartitioning) and suppressed E2 and P4 levels, whereas prospective studies provide causal evidence for the induction and progression of this condition to amenorrhea in both humans and animals (38,39). These human and animal models of amenorrhea have been associated with suppressed metabolic hormones including low TT3 (5,10,20).
To date, few investigators have prospectively explored the influence of suppressed ovarian hormones and bioenergetic status on sport performance. Therefore, the primary purpose of this study was to assess the impact of suppressed ovarian hormones, secondary to an energy deficit, on sport performance assessed by a 400-m time trial during a 12-wk competitive swimming season in junior national caliber athletes. We hypothesize that competitive female swimmers with suppressed ovarian hormones, secondary to an energy deficit, will experience minimal improvement in 400-m time trial swim velocity compared with competitive female swimmers without suppressed ovarian hormones. The secondary purpose of this study was to explore the association of bioenergetic variables (TT3, IGF-1, REE, EA, and EB) and 400-m swim velocity (sport performance) during a 12-wk swim season in junior national caliber athletes. We hypothesize that bioenergetic variables indicative of an energy deficiency (i.e., suppressed TT3 and IGF-1 concentrations) will be associated with poor sport performance.
Ten junior national caliber female swimmers (age 15–17 yr) participated in a 12-wk competitive swim training season, during which time this study was executed. The study was initiated at the beginning of the training season (week 0; baseline) and ended after the final major competition (week 12). The athletes were evaluated every 2 wk for the 12-wk competitive swim season (0, 2, 4, 6, 8, 10, and 12 wk). Each subject maintained a daily diary throughout the study. Daily questions included items regarding sleep, diet, injury, illness, stressors, and menstrual cycle. The athlete would respond to several questions such as, “Did you begin your period today?” and “Do you have your period today?” An additional question was placed on the first daily diary page asking, “When were your last two menstrual periods?” Menstrual cycle data were determined throughout the duration of the 12-wk season to document intermenstrual interval. Evaluations included measurements of ovarian hormones (E2 and P4), TT3, IGF-1, REE, EEE, EI, EA, body composition, and 400-m swim velocity (swim performance). The athletes were retrospectively classified into ovarian hormone groups based on serum P4 and E2 levels assessed at weeks 0 and 2, as well as self-reported menstrual status from daily logs. Athletes were classified as “ovarian suppressed” (OVS) or “cyclic” menstrual function (CYC). The primary variable of interest in this study was sport performance, assessed by the measurement of 400-m time trial velocity measured at weeks 0, 2, 4, 6, 8, 10, and 12. The study was approved by the Human Subjects Review Board at the University of Colorado, Colorado Springs, and each athlete and parent/guardian signed an informed consent document.
Total body mass was measured to the nearest 0.1 kg on a physician’s scale at weeks 0, 2, 4, 6, 8, 10, and 12. Participants were weighed in swimsuits. Height was measured to the nearest 1.0 cm. Body mass index (BMI) was calculated as the body mass divided by height squared (kg·m−2). Body composition was calculated from skinfold assessment (Lange calipers) at four sites (biceps, triceps, subscapular, and suprailiac) using the generalized equation for women by Durnin and Womersley (13). Skinfold measurements were taken by a single highly trained researcher. Although skinfold measurements are not the ideal methodology, serial dual-energy x-ray absorptiometry measures were not considered practical or safe for the study population. Therefore, standardized controls were used by the experienced investigator to collect the skinfold parameters.
Competitive swim training season
The coach designed swim training program included the following elements: (a) an early season, which consisted primarily of aerobic training (weeks 0–2); (b) a midseason characterized by rapidly increasing training distance with some increased intensity (“threshold pace”) work (weeks 4–8); (c) a late season characterized by reductions in training distance but significant increases in intensity (weeks 9–10); and (d) the taper period characterized primarily by high intensity (speed), low distance training (week 12).
Training sessions were analyzed by members of the research group. For each training session, the following training parameters were recorded for each swimmer in the study: (a) swim distance, (b) stroke or drill type, (c) swim time for each repetition, (d) recovery intervals, and (e) heart rate for each repetition. The training distance and intensity for both groups varied during the course of the season in a pattern typical of competitive swim training (34,35).
Participants were retrospectively grouped according to ovarian status at baseline into either (i) cyclic menstrual function (CYC) group or (ii) ovarian-suppressed (OVS) menstrual function group. Athletes were classified based on serum P4 levels (weeks 0 and 2) and self-reported menstrual status. Cyclic menstrual function (CYC) was defined as serum P4 concentrations during the luteal phase of at least 5 ng·mL−1 (15.9 nmol·L−1) or greater at week 0 or week 2. Ovarian-suppressed (OVS) was defined as serum P4 concentrations less than 5 ng·mL−1 (15.9 nmol·L−1) at both week 0 and week 2, and the absence of cyclical increases in E2. An increase in E2 was deemed cyclical if the concentration increased above a threshold of 25 pg·mL−1 during a 4-wk period.
The characterization of “cyclic” in this study does not imply ovulation but implies only that these swimmers did not display ovarian-suppressed cycles with an acyclic pattern of release. An acyclic pattern of release was defined as failure of P4 to rise during a 4-wk period above 2 ng·mL−1 (18). The P4 criteria are based on clinical studies that suggest a single sample serum concentration of at least 5 ng·mL−1 (15.9 nmol·L−1) is likely to be indicative of adequate luteinization (18). Low luteal phase serum P4 levels of <2 ng·mL−1 in regularly cycling women and exercising women are predictive of ovulation disorders and luteal suppression (10,18,27,33). Although single serum sampling is not an ideal methodology, it is, however, consistent with clinical studies that have used these methods to categorize menstrual status (18,27,33). Gynecological age (current age − age at menarche) was calculated from recall. Intermenstrual interval was recorded from menstrual logs to determine menstrual cycle length throughout the duration of the study.
Maximal time trial performance (400-m swim)
The athletes were required to perform one maximal 400-m freestyle swim effort (time trial) during the afternoon practice session on two consecutive days at weeks 0, 2, 4, 6, 8, 10, and 12. The average velocity for the two swims was calculated (m·s−1) and used for statistical analysis. The swimmer was allowed to warm up for 10 min, followed by a 5-min stretching routine. The athlete would then swim a 400-m freestyle time trial from a dive start to simulate race conditions. A decoy or “rabbit” competitor was used to better simulate the competitive swim environment for each trial. A standard swim timing system was used (gun start and wall timing pads) to record the swimmer’s performance time. The time trial was completed at the start of the afternoon training session, the 2 d after the REE and fasting blood draw, but within the diet record period. To evaluate improvement during the 12-wk season, the change in performance (Δ400-m time trial velocity) was calculated as the difference between the week 0 baseline performance and the week 12 posttaper performance.
Dietary EI assessments
Diet records were kept for 3 d during the first portion of each 2-wk measurement period (weeks 0, 2, 4, 6, 8, 10, and 12). Twenty-one days of diet records were obtained for each athlete throughout the study. Each swimmer was trained to complete food records and how to use the measuring devices (scales and measuring cups) by an experienced registered dietitian. To ensure accuracy of recording, the dietary records were also reviewed by the lead investigator with each swimmer that included a 24-h dietary recall for the last day of the diet record period. Diet records were evaluated using the Nutritionist IV database (San Bruno, CA) by a single investigator. Recipes and food labels for food items not included in the database were provided for further evaluation. Total daily EI (kcal·d−1) was determined for each swimmer at each time point measured. In addition, macronutrient content as a percentage of the total daily EI was determined for carbohydrates, proteins, and fats. The Goldberg cutoff was applied to EI data for each subject (14).
Energy expenditure assessments
REE was measured by indirect calorimetry using a Sensormedics Vmax® metabolic cart (Yorba Linda, CA) under standardized conditions (72°F–74°F and 26%–29% humidity). After a 45-min rest period, a ventilated hood was placed on the subject and REE was measured for 30 min. REE was calculated from steady state values as follows: (a) <10% fluctuation in oxygen (V˙O2; mL·min−1) and carbon dioxide (V˙CO2; mL·min−1) values and (b) <5% fluctuation in respiratory quotient values. A minimum of 20 min of data was used to determine REE according to the Weir equation (37). The Harris–Benedict equation was used to calculate predicted REE (pREE) (16). A ratio of REE:pREE was calculated to assess the presence or absence of an energy deficit (8). We have previously published data using operationally defined energy deficiency as a ratio of REE:pREE less than 0.90 (7,11). The REE:pREE cutoff ( <0.90) was used in the current study as our operational definition of energy deficiency to best discriminate the exercising women who may present with an energy deficiency from those exercising women that are potentially energy replete.
EEE was calculated from training records using swimming specific energy expenditure tables (15). Energy expenditure from 7 d of training records was averaged to determine EEE for each athlete. Each subject maintained a daily activity log from which all nontraining activities were determined. Nontraining EEE (non-EEE) was calculated using standard methods and values as per Ainsworth et al. (1). The thermic effect of food (TEF) was calculated as EI multiplied by 0.1 (kcal·d−1) (31).
Caloric equivalents in fat and fat-free mass change were determined using the coefficients described by Del Corral et al. (12). We used the energy coefficients of 9.3 kcal·g−1 (fat mass) and 1.1 kcal·g−1 (fat-free mass) to convert losses in fat mass or fat-free mass, respectively. Gains in fat mass or fat-free mass were calculated using the energy coefficients 9.3 or 1.8 kcal·g−1, respectively (12). Total daily energy expenditure (TDEE) was adjusted using these daily values per the recommendations of de Jonge et al. (6). TDEE (kcal·d−1) was calculated as TDEE = REE + TEF + EEE + non-EEE. EB was calculated as EI minus TDEE. Negative EB values > 200 kcal·d−1 were considered to be an energy deficit due to the potential variability in measuring both EI and TDEE (21). EA was operationally defined as EA = EI – EEE then calculated relative to fat free mass (FFM) (EA per kilogram FFM) (22).
Athletes reported to the laboratory after an overnight fast. After the REE measurement, blood (15 mL) was drawn using standard clinical procedures by a single researcher. Whole blood was collected in serum separator tubes (10 mL) and lithium heparin (5 mL) plasma tubes. Tubes were centrifuged (for 15 min at 2500 RPM) and serum or plasma was removed and stored at −80°C for later analysis.
Blood was collected using standard clinical procedures. The reproductive hormones, P4 (sensitivity 0.3 ng·mL−1) and E2 (sensitivity 10 pg·mL−1) were analyzed using microparticle enzyme immunoassays (MEIA) (Abbott Laboratories, Abbott Park, IL). All interassay CVs were less than 7% and intra-assay CVs ranged from 3.6% to 5.1%. IGF-1 levels were measured using IRMA (immunoradiometric assays) assays (Diagnostic System Laboratories, Webster, TX). IGF-1 had an intra-assay CV of 6% at 0.80 ng·mL−1. Coated tube RIAs (radioimmunoassay) were used for T3 (Diagnostic System Laboratories, Webster, TX). Total T3 has an interassay CV of 6.5% and an intra-assay CV of 3.2% at 0.3 ng·dL−1.
Data were analyzed for nonnormality and homogeneity of variance. Demographic, hormonal, metabolic, bioenergetic and performance data were analyzed using a two-way (group × time) repeated-measures (time) ANOVA. When a significant interaction was obtained, post hoc analyses (Scheffe) were performed where significant F-ratios were found. Correlations between hormones and the seasonal change in swim performance (400-m velocity) data were completed using week 12 data for hormones by Pearson product moment correlations. All data are reported as mean ± SEM. Significance was accepted at P < 0.05. All data were analyzed using SPSS for Windows (version 12.0, Chicago, IL) statistical software.
The CYC (n = 5) and the OVS group (n = 5) were similar in chronological and gynecological age as illustrated in Table 1. All athletes were postmenarcheal. BMI was significantly different between groups at week 12 (CYC = 20 ± 0.4 kg·m−2; OVS = 24 ± 1.0 kg·m−2; P = 0.05). Fat mass was significantly lower in the CYC athletes at week 12 (CYC = 10.6 + 3.0 kg; OVS = 14.4 + 2.3 kg; P = 0.05). Body mass, body fat, lean body mass, height and weight were similar between the CYC and OVS groups across the entire study.
Training history was similar between groups (P > 0.05). CYC and OVS athletes participated in organized swim training for approximately 5 yr. Both groups trained similarly; each group performed 13.5 h·wk−1 of pool training and 1.6 h·wk−1 of dryland (resistance) training (P > 0.05). Current season training (m·wk−1; P = 0.98) was similar between the groups.
Injury and illness records (taken from daily logs) were similar (P > 0.05) between the two groups (data not shown). The athletes studied, attended 97% of the practice sessions throughout the season and were highly motivated to perform at their season’s major competition.
Ovarian hormone and menstrual characteristics.
Generally, the CYC group demonstrated an ovarian steroid pattern reflective of follicular and luteal cyclicity throughout the season, as depicted in Figure 1 (left side panels; CYC1-5). The mean luteal concentration of P4 in the CYC group typically exceeded 10 nmol·L−1, a concentration associated with adequate luteinization and ovulation (10,18,27). All subjects in the CYC group reported that they were currently experiencing eumenorrheic menstrual cycles of 28–31 d at baseline and 26–33 d at week 12. In contrast, the OVS group (Fig. 1; right side panels OVS1-5) exhibited a suppressed pattern in both ovarian steroids compared with CYC swimmers. Additional evidence for suppressed ovarian function in the OVS group is indicated by the between-group differences for E2 at weeks 2 (P = 0.006), 6 (P = 0.003), and 10 (P = 0.002). An evaluation of self-reported menstrual patterns (prestudy) revealed that in the OVS group, all but one swimmer was oligomenorrheic (inconsistent length cycles greater than 36 d), and no swimmers were amenorrheic. Menstrual history records revealed that cycle length of the two cycles before week 0 (from self-report) was 29 ± 2 d for CYC and 40 ± 3 d for OVS. Prestudy, none of the OVS subjects reported amenorrhea, but they did report prolonged and irregular cycles of 33–46 d. Average cycle length across the study was significantly greater in the OVS compared with CYC swimmers (OVS: 86 ± 2 d, CYC: 29 ± 1 d). During the study, the OVS swimmers failed to menstruate after the first 2 wk of training (as indicated in Fig. 1), and the duration of the absence of menses during the 12-wk study approached the clinical definition of no menses for 90 d or more. The occurrence of menses poststudy was not recorded, but it is highly likely that the OVS women may have met the criteria for amenorrhea if the study exceeded the 12-wk training period.
The CYC group had significantly greater dietary EI than the OVS group throughout the study (P ≤ 0.001) (Table 2). EI was not significantly different across the season for either group (P = 0.98). Although both groups increased their EI from baseline to week 6, the increases were not significant. Macronutrient content (protein, carbohydrates, and fats) of the diet was similar between the groups (P = 0.98). Carbohydrate content (as a percentage of total EI kcal) was 57% ± 1% for the CYC and 59% ± 1% for the OVS athletes. Protein content was 15% ± 1% for each group, whereas fat content was 26% ± 1% for the OVS athletes and 28% ± 2% for the CYC swimmers.
A significant main effect of group (P = 0.001) was seen for absolute REE (kcal·d−1). Furthermore, the group–time interaction was significant (P = 0.045) (Table 2). REE:pREE in the CYC group ranged from 95% to 104%, supporting a significant main effect of group (P < 0.001; Table 2), whereas OVS values were less than 85% throughout the study, indicative of energy conservation.
Patterns in EEE, EA, and EB for CYC (panel A) and OVS (panel B) are illustrated in Figure 2. EEE was similar between CYC and OVS (P = 0.99), increasing from week 0 through weeks 6 and 8 (peak values) in both groups. A significant time effect (P < 0.001) was noted, as expected. Measures of EA were significantly greater in the CYC group compared with the OVS group throughout the study. A negative EB was seen in the OVS athletes at all weeks (Fig. 2B). The CYC swimmers were in a positive EB at weeks 2 and 4. All other time points measured in the CYC swimmers reflected a negative EB. The OVS athletes had energy deficits two times (EA per kilogram FFM) and five times (EB) greater than the CYC women throughout the study (main effect group; P < 0.001 for both EA per kilogram FFM and EB).
At baseline, TT3 concentrations were similar between groups (Fig. 3A). TT3 concentrations were significantly lower in the OVS group compared with the CYC athletes at weeks 6 through 12 (P < 0.001). TT3 concentration declined significantly from concentrations observed at baseline and week 2 compared with levels observed at weeks 4 through 12 in the OVS swimmers. The suppression in TT3 concentration progressively worsened (12% decline) from week 4 to week 12. IGF-1 was 9% higher in the CYC group compared with OVS at week 0, and this difference increased to 23% (P = 0.0001) by week 12 (Fig. 3B). Between-group differences in IGF-1 were significant beginning at week 4 (initiation of heavy training) and continuing throughout the season. The OVS group showed a marked decline in IGF-1 during the final 6 wk of training. IGF-1 declined during the season in the CYC athletes with week 12 significantly lower than weeks 0 (P = 0.001) and 4 (P = 0.002).
Figure 4 illustrates the 400-m time trial results for both groups. Both groups began the season (weeks 0 and 2) with similar swim velocity (P = 0.98), followed by a rapid decline in swim velocity at mid season (CYC week 4 = 10%, OVS week 6 = 12%). CYC improved in response to training throughout the season with a faster time trial performance (P < 0.001) compared with OVS at week 12. In contrast, OVS athletes’ time trial performance at week 12 declined (P = 0.001) by 10% compared with week 4. The absolute change in 400-m time trial performance was improved by 0.10 m·s−1 or 8.2% in the CYC athletes compared with a −0.13 m·s−1 or 9.8% decline in performance in the OVS group (P ≤ 0.001). Figure 4 illustrates individual athlete 400-m swim velocity data at weeks 0 and 12. At baseline, both the OVS and the CYC athletes demonstrated similar swim velocity. At week 12, each OVS athlete showed a decline in 400-m velocity, whereas the CYC swimmers improved 400-m performance compared with baseline. The group performance overlaps at baseline with each OVS athlete showing a decline in 400-m velocity at week 12, whereas the CYC swimmers improved performance.
Relationships between metabolic factors, ovarian status, and swim performance
Correlation coefficients demonstrated associations between bioenergetic, hormone, and performance parameters. Significant relationships were found between week 12 and Δ400-m time trial velocity (%) in P4 (r = 0.899, P < 0.01), E2 (r = 0.843, P = 0.002), IGF-1 (r = −0.855, P < 0.001), TT3 (r = 0.643, P = 0.04, EI (r = 0.823, P = 0.003), and EA (r = 0.760, P = 0.01).
Our study is the first to document significant decrements in elite athlete sport performance during a 12-wk competitive season that are associated with chronic ovarian suppression secondary to an energy deficit, which was confirmed by several measures of energy status. The ovarian suppression is clearly linked to a cascade of metabolic and bioenergetic perturbations likely responsible for reduced sport performance in the OVS women. TT3 was 19% lower in the OVS athletes, REE:pREE was 27% lower and EA was 90% lower than CYC. We are the first to demonstrate that athletes who are in an energy deficit with ovarian suppression experience effects that translate to poor sport performance that worsens if the energy deficit remains. In fact, sport performance declined by 9.8% in the ovarian-suppressed group, whereas performance improved by 8.2% in the cycling group. The strength of this study is that it broadens the scope and magnitude of concern beyond medical outcomes to include the deleterious effect of an energy deficit and ovarian suppression on sport performance. Communication of this message to coaches and athletes is essential because many individuals promote energy restrictive practices with the goal of improved performance.
The OVS swimmers exhibited suppressed E2 and P4 throughout the season. The suppression of ovarian hormonal fluctuations has been reported in cross-sectional studies of female athletes (5), and active females (9,10), prospective studies of sedentary ovulatory women undergoing several months of exercise training (4,40), and in longitudinal work with primates with diet restriction and/or exercise (38,39). Although these young women did not exhibit the most severe menstrual disturbance of amenorrhea, it seems that the subtle ovarian suppression and long cycles characteristic of oligomenorrhea were coupled with reductions in sport performance. The CYC swimmers demonstrated cyclic fluctuations in their ovarian hormones typical of a pattern of release observed in ovulatory menstrual cycles. It is interesting to note that the OVS swimmers exhibited a higher body fat percentage than the CYC swimmers throughout the study, which was different than that observed in ovarian-suppressed women in land-based sports (7–11,20,28). The higher FM in the OVS swimmers would not have been an inherently negative performance issue because of the positive relationship of fat mass and buoyancy (34). It is important to remember that both groups of swimmers were considered elite for their age, yet the OVS athlete’s performance improvement was reduced in conjunction with the energy deficient and hormonally suppressed environment. In female swimmers of equivalent training and skill level, the observed differences in energy and hormonal status likely represent key factors associated with performance deficits as seen in the OVS athletes.
CYC and OVS athletes expended between 900 and 1225 kcal·d−1 in EEE when training for ∼2 h·d−1 during the 12-wk season. The pattern of training volume was typical for competitive swimmers (Fig. 4D), with values corresponding to a periodization protocol observed in junior elite athletes (34,35). The determination of EEE in free-living elite caliber athletes is critical to our understanding of bioenergetics and performance. A direct evaluation of training parameters, coupled with sport specific conversion factors provides a more complete characterization of the bioenergetic demands in the current study. TDEE in both groups was similar during the course of the season but was lower than that reported previously (∼5600 kcal·d−1) in female swimmers using doubly labeled water methods (32,34). It is important to note that both doubly labeled water studies assessed older (mean age = 19.1 yr), collegiate swimmers with more rigorous training schemes.
Previous work with anorexic women indicates that a decreased REE as a percentage of pREE (60%–80%) represents a suppressed metabolic status (30). De Souza et al. (8) reported a significantly reduced ratio of REE:pREE in exercising women with severe menstrual dysfunction and metabolic evidence of hypometabolism. A significant main effect (P < 0.001) was observed for REE:pREE in the OVS group consistently less than 85% (77%–84%), indicative of energy conservation. Our findings support De Souza et al. (8) in active women with suppressed reproductive function and high drive for thinness. Our data support the theory that competitive female athletes often train while in a chronic energy deficit, likely contributing to impaired sport performance given our reported outcomes. Exacerbation of an energy deficient state, observed in the OVS athletes, seems to play a pivotal role in the decrements in swim performance. The OVS swimmers had negative EB values coupled with low REE, indicating a chronic energy deficit in this group. It is important to note that the OVS women maintained constant body mass and body composition in the face of a significant energy deficit. OVS women were metabolically suppressed at the start of the study; therefore, their body weight may have already declined to a level supported by their “new homeostasis” of habitually low EB. Athletes can maintain a chronic energy deficit for varied lengths of time with continued sport success; however, the chronic nature of the deficit resulted in the maladaptation to training seen in the current study.
The difference in EI between groups (∼700 kcal·d−1 throughout the study) was similar to that reported by other groups (9,40). However, OVS data illustrate a reduction in EI compared with data from previous work. Kabasakalis et al. (19) reported dietary intakes (EI) averaging approximately 2360 kcal·d−1 in elite Greek swimmers. A small but insignificant increase in EI between pretraining of approximately 1910 kcal·d−1 to during 2389 kcal·d−1 at week 10 of training was reported (19). It is important to note that these women were considered to have normal menstrual cycles based on cycle length. The current findings for CYC athletes are similar to previous data from American female swimmers reported by Trappe et al. (32) (midtraining or pretaper EI ∼2700 kcal·d−1, end of season EI ∼2400 kcal·d−1), VanHeest (34,35) (heavy training EI = 2200–3700 kcal·d−1, taper EI = 2000–3500 kcal·d−1), or Ousley-Pahnke et al. (29) (EI ∼2300 kcal·d−1), which did not report menstrual status in their swimmers. In an attempt to determine those athletes who may have underreported in our study, the data were analyzed using the Goldberg cutoff (3,14), and all subjects met the criteria, suggesting that the repeated dietary reports were indicative of the athletes’ overall dietary patterns.
The chronic energy deficit in our study was supported by decreased TT3 and IGF-1 concentrations. These typical markers of a nutritional deficit (i.e., TT3 and IGF-1) were not significantly reduced until an additional stress or energy demand was placed on the OVS group during the midseason, heavy training period (weeks 4–8). The hypometabolic state supported by the decreased TT3, IGF-1, EA, and REE:pREE was coupled with suppressed ovarian hormones resulting in decreased performance. Previous studies have evaluated the link between training stress and thyroid hormone status (2,23). These investigations have reported suppressed thyroid hormones in response to training and/or negative EB. TT3 data in the OVS group support this previous work (2,9,23).
Loucks et al. (17,24,25,26) have demonstrated that decreased LH pulsatility is associated with decreased EA in sedentary women. Specifically, Loucks and Thuma (24) have suggested that EA levels lower than a threshold of 30 kcal·kg−1 LBM per day results in negative alterations in LH pulse characteristics. In the present study, the ovarian-suppressed athletes had large energy deficits coupled with an altered reproductive hormone profile. The OVS athletes presented with an energy deficiency concomitant with suppressed reproductive hormones, which if sustained for a prolonged period may have detrimental outcomes for both performance and menstrual status. The long-term implications of energy deficiency and reproductive dysfunction on performance are unknown. In addition, thresholds indicative of performance decrements associated with EA and menstrual status do not currently exist. The chronically reduced EA (both kilocalories per day and kilocalories per kilogram FFM) in the OVS group indicated a pathologic physiological environment, resulting in perturbations in the ovarian axis coupled with the thyroid axis. As suggested by Loucks et al. (22), the suppressed physiological state experienced in the face of chronically low EA resulted in maladaptation to sports training. The altered metabolic and hormonal conditions resulted in an inappropriate training response leading to impaired performance outcomes. The strong relationships between end of season (week 12) TT3, P4, E2, and IGF-1 and the Δ400-m time trial performance support a theory of maladaptation to training in the OVS swimmers.
Both groups of swimmers periodized their EI to some extent during the season (i.e., 200 kcal·d−1 increase weeks 0–4 for CYC group); however, the OVS athletes exhibited a smaller increase in EI (∼100 kcal·d−1) compared with the CYC group. Coaches and athletes should be educated on the appropriate strategies to sync training periodization and energy periodization techniques. Further evaluation of competitive performance in chronically training female athletes, coupled with markers of bioenergetics, nutritional status, metabolic hormones, and reproductive hormones, is critical in providing appropriate recommendations to competitive female athletes. Short-term exercise interventions typically seen in the literature, although important, may not adequately reflect the alterations in homeostatic physiological conditions in chronically undernourished female athletes. The present study supports the widely held claim that a chronic energy deficit coupled with ovarian suppression can result in compromised performance in competitive female athletes.
To date, there have been no studies assessing metabolic and reproductive hormones, bioenergetic factors and sport performance outcomes in competitive female athletes. In response to the chronic energy deficit, nutrient partitioning occurs resulting in suppression of the hypothalamic–pituitary–ovarian axis. Previous work has clearly linked low EA with suppressed ovarian axis function and metabolic markers in habitually sedentary women (17,23–26), recreationally active women (9,10), and exercising animals (36,38,39).
It is evident that the hormonal environment was different between the OVS and the CYC swimmers. The OVS athletes were in a hypometabolic state (reduced TT3 and IGF-1, suppressed E2 and P4, and decreased REE:pREE). These characteristics have been described by several groups in exercising women (9,17,20,24). The hormonal conditions reported previously in cross-sectional studies of female athletes with varying degrees of reproductive abnormalities (i.e., anovulation, luteal phase deficiency and amenorrhea) are generally similar to the results of this longitudinal evaluation of junior national caliber female swimmers. The continued hypometabolic state coupled with the altered hormonal environment resulted in impaired sport performance in the OVS group. This unique study provides important new insights, while supporting previous work (10,17,24). These data underscore the need for adequate fuel intake to maintain an optimal hormonal and metabolic environment to maximize sport performance outcomes. Future studies are necessary to investigate components of the behavioral, physiological, and metabolic phenotypes in this model.
The authors would like to thank the athletes and coaches for their participation in this study.
There is neither funding nor any conflicts of interest to report for this study.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc. 2000; 32 (9 Suppl): S498–516.
2. Baylor LS, Hackney AC. Resting thyroid and leptin hormone changes in women following intense, prolonged exercise training. Eur J Appl Physiol. 2003; 88 (4–5): 480–4.
3. Black AE. Critical evaluation of energy intake using the Goldberg cut-off for energy intake:basal metabolic rate. A practical guide to its calculation, use and limitations. Int J Obes Relat Metab Disord. 2000; 24 (9): 1119–30.
4. Bullen BA, Skrinar GS, Beitins IZ, von Mering G, Turnbull BA, McArthur JW. Induction of menstrual disorders by strenuous exercise in untrained women. N Engl J Med. 1985; 312 (21): 1349–53.
5. Constantini NW, Warren MP. Menstrual dysfunction in swimmers: a distinct entity. J Clin Endocrinol Metab. 1995; 80 (9): 2740–4.
6. de Jonge L, DeLany JP, Nguyen T, et al. Validation study of energy expenditure and intake during calorie restriction using doubly labeled water and changes in body composition. Am J Clin Nutr. 2007; 85 (1): 73–9.
7. De Souza MJ, Hontscharuk R, Olmsted M, Kerr G, Williams NI. Drive for thinness score is a proxy indicator of energy deficiency in exercising women. Appetite. 2007; 48 (3): 359–67.
8. De Souza MJ, Lee DK, VanHeest JL, Scheid JL, West SL, Williams NI. Severity of energy-related menstrual disturbances increases in proportion to indices of energy conservation in exercising women. Fertil Steril. 2007; 88 (4): 971–5.
9. De Souza MJ, Miller BE, Loucks AB, et al. High frequency of luteal phase deficiency and anovulation in recreational women runners: blunted elevation in follicle-stimulating hormone observed during luteal-follicular transition. J Clin Endocrinol Metab. 1998; 83 (12): 4220–32.
10. De Souza MJ, Van Heest J, Demers LM, Lasley BL. Luteal phase deficiency in recreational runners: evidence for a hypometabolic state. J Clin Endocrinol Metab. 2003; 88 (1): 337–46.
11. De Souza MJ, West SL, Jamal SA, Hawker GA, Gundberg CM, Williams NI. The presence of both an energy deficiency and estrogen deficiency exacerbate alterations of bone metabolism in exercising women. Bone. 2008; 43 (1): 140–8.
12. Del Corral P, Chandler-Laney PC, Casazza K, Gower BA, Hunter GR. Effect of dietary adherence with or without exercise on weight loss: a mechanistic approach to a global problem. J Clin Endocrinol Metab. 2009; 94 (5): 1602–7.
13. Durnin JV, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr. 1974; 32 (1): 77–97.
14. Goldberg GR, Black AE, Jebb SA, et al. Critical evaluation of energy intake data using fundamental principles of energy physiology: 1. Derivation of cut-off limits to identify under-recording. Eur J Clin Nutr. 1991; 45 (12): 569–81.
15. Grandjean AC. Nutrition for swimmers. Clin Sports Med. 1986; 5 (1): 65–76.
16. Harris JA, Benedict FG. A biometric study of human basal metabolism. Proc Natl Acad Sci U S A. 1918; 4 (12): 370–3.
17. Hilton LK, Loucks AB. Low energy availability, not exercise stress, suppresses the diurnal rhythm of leptin in healthy young women. Am J Physiol Endocrinol Metab. 2000; 278 (1): E43–9.
18. Jordan J, Craig K, Clifton DK, Soules MR. Luteal phase defect: the sensitivity and specificity of diagnostic methods in common clinical use. Fertil Steril. 1994; 62 (1): 54–62.
19. Kabasakalis A, Kalitsis K, Tsalis G, Mougios V. Imbalanced nutrition of top-level swimmers. Int J Sports Med. 2007; 28 (9): 780–6.
20. Laughlin GA, Yen SS. Nutritional and endocrine-metabolic aberrations in amenorrheic athletes. J Clin Endocrinol Metab. 1996; 81 (12): 4301–9.
21. Loucks AB. Low energy availability in the marathon and other endurance sports. Sports Med. 2007; 37 (4–5): 348–52.
22. Loucks AB, Kiens B, Wright HH. Energy availability in athletes. J Sports Sci. 2011; 29 (1 Suppl): S7–15.
23. Loucks AB, Laughlin GA, Mortola JF, Girton L, Nelson JC, Yen SS. Hypothalamic-pituitary-thyroidal function in eumenorrheic and amenorrheic athletes. J Clin Endocrinol Metab. 1992; 75 (2): 514–8.
24. Loucks AB, Thuma JR. Luteinizing hormone pulsatility is disrupted at a threshold of energy availability in regularly menstruating women. J Clin Endocrinol Metab. 2003; 88 (1): 297–311.
25. Loucks AB, Verdun M. Slow restoration of LH pulsatility by refeeding in energetically disrupted women. Am J Physiol. 1998; 275 (4 pt 2): R1218–26.
26. Loucks AB, Verdun M, Heath EM. Low energy availability, not stress of exercise, alters LH pulsatility in exercising women. J Appl Physiol. 1998; 84 (1): 37–46.
27. McNeely MJ, Soules MR. The diagnosis of luteal phase deficiency: a critical review. Fertil Steril. 1988; 50 (1): 1–15.
28. 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.
29. Ousley-Pahnke L, Black DR, Gretebeck RJ. Dietary intake and energy expenditure of female collegiate swimmers during decreased training prior to competition. J Am Diet Assoc. 2001; 101 (3): 351–4.
30. Polito A, Fabbri A, Ferro-Luzzi A, et al. Basal metabolic rate in anorexia nervosa: relation to body composition and leptin concentrations. Am J Clin Nutr. 2000; 71 (6): 1495–502.
31. Reed GW, Hill JO. Measuring the thermic effect of food. Am J Clin Nutr. 1996; 63 (2): 164–9.
32. Trappe TA, Gastaldelli A, Jozsi AC, Troup JP, Wolfe RR. Energy expenditure of swimmers during high volume training. Med Sci Sports Exerc. 1997; 29 (7): 950–4.
33. van Zonneveld P, te Velde ER, Koppeschaar HP. Low luteal phase serum progesterone levels in regularly cycling women are predictive of subtle ovulation disorders. Gynecol Endocrinol. 1994; 8 (3): 169–74.
34. VanHeest JL. Energy expenditure of elite female swimmers measured using doubly labeled water. In. Swimming Science VII: Chapman & Hall: London, United Kingdom; 1996. p. 241–6.
35. VanHeest JL, Mahoney CE. Female athletes: factors impacting successful performance. Curr Sports Med Rep. 2007; 6 (3): 190–4.
36. Wade GN, Schneider JE, Li HY. Control of fertility by metabolic cues. Am J Physiol. 1996; 270 (1 Pt 1): E1–E19.
37. Weir JB. New methods for calculating metabolic rate with special reference to protein metabolism. J Physiol. 1949; 109 (1–2): 1–9.
38. Williams NI, Caston-Balderrama AL, Helmreich DL, Parfitt DB, Nosbisch C, Cameron JL. Longitudinal changes in reproductive hormones and menstrual cyclicity in cynomolgus monkeys during strenuous exercise training: abrupt transition to exercise-induced amenorrhea. Endocrinology. 2001; 142 (6): 2381–9.
39. Williams NI, Helmreich DL, Parfitt DB, Caston-Balderrama A, Cameron JL. Evidence for a causal role of low energy availability in the induction of menstrual cycle disturbances during strenuous exercise training. J Clin Endocrinol Metab. 2001; 86 (11): 5184–93.
40. Williams NI, Young JC, McArthur JW, Bullen B, Skrinar GS, Turnbull B. Strenuous exercise with caloric restriction: effect on luteinizing hormone secretion. Med Sci Sports Exerc. 1995; 27 (10): 1390–8.