The Female Athlete Triad: Review of Current Literature

Abstract

Purpose of the review

Adolescence and young adulthood are a critical period in the life of women for optimizing long-term bone health. Young athletes lead a demanding lifestyle with increased dietary requirements to meet the robust demands of energy expenditure to maintain a state of energy balance. During a time of fast paced changes and unpredictable societal demands on young athletes, it is important to review the severe consequences of energy deficiency and options for adequate management.

Recent findings

This review focuses on hormonal adaptations that occur in energy deficient female athletes that lead to menstrual irregularities and impaired bone health, increasing the risk for stress and other fractures. We also describe management strategies to mitigate the consequences of limited energy availability on bone and other outcomes.

Summary

These strategies should help guide the management of young female athletes to prevent irreversible changes to their bone health. Identifying current knowledge should help increase awareness among medical providers, which can then be communicated to the sports community, parents, and athletes.

Introduction

The Female Athlete Triad describes a condition in physically active young women that includes low energy availability, menstrual dysfunction and low bone mineral density (BMD) (1–3). Low energy availability is often (though not always) associated with disordered eating or a frank eating disorder (1). An imbalance between caloric intake and metabolic demand results in a relative energy deficient state (2). Although exercise is associated with positive health related outcomes in most individuals, this can be deleterious if associated with low energy availability (3). Low energy availability leads to hormonal adaptations, which, in turn result in menstrual irregularities and compromised bone health.

Hypothalamic-Pituitary-Growth Hormone axis

Growth hormone (GH) secretion is higher in amenorrheic athletes compared to oligomenorrheic and eumenorrheic athletes as well as sedentary controls (4). Body fat, which signifies energy stores, plays an important role in regulating the hypothalamic-pituitary-GH axis. Low body fat implies a state of energy deficit, and GH levels are inversely related to total body fat (4) and also glucose concentrations. The latter is consistent with the gluconeogenic effects of GH, with an increase in GH in conditions of low energy availability being adaptive to maintain euglycemia.

However, higher GH concentrations do not translate to higher insulin like growth factor 1 (IGF-1) levels. In fact, IGF-1 levels are lower in amenorrheic athletes vs. non-athletes (5), consistent with a state of acquired GH resistance, also reported in conditions of very low energy availability such as anorexia nervosa (AN). In one study, IGF-1 levels were measured in swimmers as they progressed through various training protocols: recreation (4-week training free interval), endurance (3 km distance twice daily in addition to strength exercises or running), sprint (3–6 sprint training units per week and additional speed strength exercises) and finally after competing (6). IGF-1 levels were lower in swimmers during the sprinting portion of their training (associated with lower caloric intake) compared to other phases of training. Further, Zietz et al. reported that IGF-1 concentrations, along with leptin (an anorexigenic adipokine) and salivary cortisol, can represent the intensity of training (6). There is also evidence of lower IGF-1 and IGFBP-3 levels associated with overtraining (7).

Of note, levels of ghrelin, an orexigenic peptide and GH secretagogue, are higher in very energy deficient states, such as AN vs. controls, and are positively associated with GH levels (8). Similarly, higher ghrelin has been reported in amenorrheic vs. eumenorrheic athletes and non-athletes (9).

Hypothalamic-Pituitary-Cortisol axis

Amenorrheic athletes have higher cortisol concentrations compared to eumenorrheic athletes and non-athletes (4,10). Similar to GH, lower fat mass is associated with higher cortisol concentrations in amenorrheic exercisers (10) consistent with findings in low-weight AN (11). These higher cortisol concentrations are believed to be adaptive to maintain euglycemia in an energy deficient state (as cortisol is gluconeogenic). Hypercortisolemia is also inhibitory to LH pulsatile secretion (12), and higher cortisol concentrations in amenorrheic athletes are associated with lower overnight LH concentrations (10). Further, data suggest that lower leptin and higher ghrelin in amenorrheic athletes drive higher cortisol concentrations, which then inhibit LH secretion (10). These higher cortisol levels deleteriously impact bone health, as previously reported in AN (11). A study examining surrogate markers of bone turnover found that higher cortisol concentrations in non-athletes were associated with higher concentrations of C-telopeptide (CTX), a bone resorption marker, and lower concentrations of P1NP, a bone formation marker (10).

Hypothalamic-Pituitary-Thyroid axis

Amenorrheic as well as anovulatory eumenorrheic hyperexercisers (vs. non-exercisers) have lower total triiodothyronine (T3) concentrations (13). Furthermore, based on data from extreme energy deficient states such as AN, these changes are consistent with the sick euthyroid syndrome secondary to decreased energy availability. Low total T3 levels are positively associated with low body mass index (BMI) and leptin levels and inversely with ghrelin and cortisol concentrations (8). This is consistent with adaptations to conserve energy in states of low energy availability given the impact of T3 on resting energy expenditure (8). Thyroid stimulating hormone (TSH) concentrations are in the range of low normal to normal while free thyroxine levels are variable (8). There are no recommendations at this time to continue monitoring thyroid labs in AN or in amenorrheic athletes unless there are symptoms concerning for weight loss secondary to hyperthyroidism, and no indications to treat with levothyroxine unless there is definitive evidence for hypothyroidism (14).

Adipokines and appetite regulating hormones

Hormones regulate energy balance via two sets of hypothalamic neurons: the neuropeptide Y/agouti-related protein neurons, which when stimulated have orexigenic or food seeking effects, and the pro-opiamelanocortin/cocaine and amphetamine related transcript neurons, which when stimulated have an anorexigenic effect. By means of a complex interaction between these neuronal pathways and hormones involved in energy balance, the body seeks to achieve energy homeostasis.

Ghrelin is an orexigenic hormone secreted by the fundus of the stomach and duodenum, and its concentrations are inversely related to fat mass. Conversely, leptin is an anorexigenic adipokine that directly correlates with fat mass, and acts as a sensor of energy availability. Studies have reported lower leptin levels across all elite athletes independent of menstrual status compared with recreational athletes, associated with lower body fat (15). Further, leptin levels are lower in amenorrheic vs. eumenorrheic athletes and non-athletes (9,15) again believed to be an adaptive mechanism to normalize feeding patterns (9,16,17). Consistent with this, studies in amenorrheic exercisers and AN show a decrease in ghrelin and an increase in leptin once energy balance is achieved or following weight regain (8,16,18). As previously indicated, higher ghrelin and lower leptin concentrations may drive lower LH secretion in amenorrheic vs. eumenorrheic athletes and non-athletes (9), likely mediated by higher cortisol concentrations in amenorrheic athletes (10).

Peptide YY (PYY), secreted from the distal gut by the endocrine L cells, increases shortly after a meal signaling satiety. PYY levels are also determined by nutritional status, and levels are higher in amenorrheic vs. eumenorrheic athletes (19,20) and in AN vs. controls (21). In these studies, PYY concentrations were inversely associated with fat mass and BMI, and in AN, higher PYY concentrations were associated with lower food intake (21,22). Higher PYY levels are certainly not an adaptive response to low energy availability, and this may, in fact, contribute to restrictive eating patterns. Furthermore, PYY inhibits osteoblastic activity, and higher PYY levels are associated with lower BMD in athletes (19) and AN (21).

Oxytocin, a bone anabolic hormone, modulates metabolic rate as well as appetite. Studies have shown lower nighttime oxytocin in states of higher energy expenditure, as in athletes vs. non-athletes (23). Moreover, in amenorrheic athletes, lower oxytocin concentrations correlate with abnormal cortical and trabecular microarchitecture at non-weight bearing sites, such as the radius, where benefits of mechanical loading are less evident. Thus lower oxytocin may contribute to abnormal bone microarchitecture in hypoestrogenic states (23). Lower oxytocin concentrations are also associated with lower resting energy expenditure in athletes but not non-athletes (24), and with lower energy availability in amenorrheic and not eumenorrheic athletes, suggesting a role for oxytocin in energy homeostasis in low energy states (24).

Insulin increases peripheral glucose uptake and is anorexigenic, while adiponectin increases insulin sensitivity and increases with weight loss. Higher adiponectin levels contribute to lower BMD by impacting both bone formation and resorption; while insulin is osteoanabolic. Girls with AN have lower insulin levels than controls associated with lower levels of the bone turnover markers (25). Elite amenorrheic athletes have lower serum insulin concentrations compared to normally cycling athletes and athletes taking oral contraceptives, with a positive correlation between insulin and leptin secretion suggesting that insulin may play a role in regulating leptin by sensing changes in the body’s energy balance (15).

Hypothalamic-Pituitary-Gonadal axis

Preservation of gonadal function is dependent on energy availability, which in athletes can be relatively low because of insufficient calorie intake. Menstrual dysfunction in athletes with low energy availability may range from anovulatory cycles and luteal phase defects to frank oligo-amenorrhea (26). While amenorrheic athletes have lower LH pulsatile secretion (pulse mass and/or frequency), than eumenorrheic athletes and non-athletes (9), one study reported higher daytime testosterone concentrations in oligomenorrheic athletes than other groups, including amenorrheic and normally cycling athletes as well as controls, with LH secretion similar to that in eumenorrheic peers (4). Thus, in this study, there was an evident pattern of energy conservation in amenorrheic athletes that was absent in oligomenorrheic individuals. Overall studies have implicated a role for lower leptin, insulin and IGF-1, and higher ghrelin, cortisol, PYY and adiponectin concentrations in suppressing the hypothalamic-pituitary gonadal (HPG) axis in low energy availability states (4,12,16,27,28).

Energy availability is defined as the difference between energy intake and exercise energy expenditure divided by fat-free mass (or lean body mass (LBM), which accounts for the body’s most metabolically demanding tissues) (2). Some studies have attempted to establish a cutoff for energy availability that can identify women at increased risk for menstrual irregularities. Loucks et al. showed that LH pulsatility is not affected when energy availability is maintained at 45 kcals/kg LBM/day or higher, but is altered below a threshold of 30 kcals/kg LBM/day (29). These findings are supported by another study showing that a decrease in energy availability from 38 to 28 kcal/kg FFM/day over a period of 3 months led to a decrease in LH pulse frequency and HPG axis suppression (30). However, conflicting studies argue that even though there is a relationship between low energy availability and menstrual dysfunction, there is no clear threshold for alteration in LH secretion (31–33). In fact, each woman may have her own threshold for energy availability at which LH pulsatile secretion is disrupted. Overall, suppression of the HPG axis during periods of low energy availability is consistent with an adaptive effect to prevent pregnancy, which would otherwise result in diversion of available energy to the growing fetus, rather than conserving this for essential body functions.

Prevalence of menstrual dysfunction among athletes depends on the nature of sport, intensity of training and nutritional status of the athlete. Traditionally, endurance runners are known to be at higher risk for functional hypothalamic amenorrhea (from hypogonadotropic hypogonadism) than other athletes, associated with overall lower body weight (34). A more recent study showed that runners were more aware of the Triad than dancers and figure skaters, and although two-thirds of participants were at risk for developing the Triad, the risk for dancers was two times as high (35). Thus, there is a critical need to educate dancers and other athletes regarding the Triad. Young female collegiate athletes engaged in long-distance sports compared to swimmers are at a higher risk of bone stress injuries and bursitis (risk categorization based on presence or absence of amenorrhea, low BMI and BMD) (36). Among high school female runners, up to 25.8% report irregular menses (37), while in adult runners this number is as high as 62%. Another study found that 20% of athletes met criteria for disordered eating and 20.1% reported irregular menses. Oligo-amenorrheic athletes more frequently reported disordered eating and athletes with disordered eating were twice as likely to report irregular periods. Moreover, irregular menses were more common among athletes classified as having a lean build (38). Further, baseline disordered eating and shape concern has been associated with fewer menstrual cycles per year (39).

Hyperandrogenism/Polycystic Ovarian Syndrome (PCOS) needs to be considered in addition to functional hypogondadotropic hypogonadism as a potential cause of menstrual dysfunction in young female athletes (40). There is some evidence of higher testosterone levels in oligomenorrheic women (4). Further, an underlying predisposition for PCOS may represent an added risk factor for menstrual dysfunction in young athletes (28,41). The prevalence of PCOS in elite Iranian female athletes 13–37 years old with oligo-amenorrhea was 15.5% (42). In another study, 17% of athletes with hypogonadotropic hypogonadism presumed to be secondary to energy deficiency also had hyperandrogenism. Oligo-amenorrheic women with hyperandrogenism had higher weight, BMI, body fat, insulin, leptin, free androgen index and LH/follicle-stimulating hormone ratio than oligo-amenorrheic peers without hyperandrogenism (41).

Consequences of hypogonadotropic hypogonadism

Bone density, microarchitecture and strength estimates:

The major consequence of hypoestrogenism in conditions of low energy availability is low BMD and increased fracture risk. The higher areal BMD Z-scores at the femoral neck, total hip and whole body observed in eumenorrheic athletes vs. non-athletes (consistent with the known adaptive effect of mechanical loading on bone in athletes), is not seen in oligo-amenorrheic athletes, indicating a loss of these effects when there is associated hypoestrogenism (43). Oligo-amenorrheic athletes also have lower spine BMD Z-scores than eumenorrheic athletes (44). Prolonged estrogen deficiency impacts not only areal BMD, but also bone microarchitecture and strength estimates, assessed using high resolution peripheral quantitative computer tomography (HRpQCT) and microfinite element analysis (μFEA). At the non-weight bearing radius, oligo-amenorrheic athletes have greater cortical porosity, and lower stiffness and failure load than eumenorrheic athletes and non-athletes. At the weight-bearing tibia, they have higher cortical porosity than non-athletes and lower stiffness than eumenorrheic athletes and non-athletes (44). Finally, fracture risk (for all fractures and particularly stress fractures) is higher in oligo-amenorrheic athletes vs. eumenorrheic athletes and non-athletes, with differences being mostly driven by stress fractures (44). Among oligo-amenorrheic athletes, women who sustained two or more stress fractures had lower spine and whole-body BMD Z-scores, and lower radial volumetric BMD (vBMD) and radial and tibial stiffness and failure load than those who sustained less than two stress fractures.

Further, in a study comparing adolescents and young adults with AN, normal-weight oligo-amenorrheic athletes and eumenorrheic controls, the oligo-amenorrheic athletes had the highest stress fracture prevalence despite similar areal BMD at the hip and whole body less head (WBLH) to controls, suggesting that oligo-amenorrheic athletes require higher BMD than controls to prevent fractures (45).

Another study comparing triathletes with regular periods or on oral contraceptives with healthy non-athletic controls reported that athletes with low energy availability did not have the pronounced effects of exercise on bone seen in athletes with adequate energy availability such as on cortical perimeter, vBMD and trabecular microstructure, but continued to have better parameters than controls. Stress fractures were more frequent in athletes with low energy availability than those with adequate energy availability (46).

Importantly, there are concerns that oligo-amenorrheic athletes may not reach optimal peak bone mass because of a lack of “catch up” effects during the narrow pubertal window of peak bone accrual (47). In addition to hypoestrogenism, other factors that contribute to impaired bone health include lower concentrations of IGF-1, insulin, leptin and oxytocin, and higher concentrations of cortisol, PYY and adiponectin.

The first line in managing impaired bone health in oligo-amenorrheic athletes is addressing the state of relative energy deficiency/low energy availability. This involves an assessment of caloric intake and expenditure and improving caloric intake to meet the needs of exercise energy expenditure (1). This is best addressed by a sports dietician. Particularly, it is important to address caloric needs before and after periods of intense energy expenditure, and guidelines recommend intake of healthy fat (as in nuts, avocado and fatty fish) and adding extra olive oil to food, in addition to increased protein intake. Normalization of energy status should allow for recovery of the HPG axis as well as normalization of other hormonal axes, and an improvement in bone measures with a reduction in fracture risk. It is also important to optimize calcium and vitamin D intake, and to maintain 25-hydroxy vitamin D levels above 32 ng/ml (1).

After 6–12 months of intensive lifestyle intervention, if there is no evidence of menstrual recovery, the next step in management is estrogen replacement therapy (48). However, estrogen administration as oral combined hormonal contraception (an oral contraceptive pill) is not effective in improving bone outcomes in oligo-amenorrheic athletes. In contrast, physiologic estrogen administration as the transdermal 17β-estradiol patch (given with a cyclic progestin) is effective in this regard. This was demonstrated in a study of 121 athletes with oligo-amenorrhea 14–25 years old randomized to receive 100 mcg of transdermal 17β-estradiol patch daily (given with cyclic oral micronized progesterone at a dose of 200 mg for 12 days of every month), a combined oral contraceptive pill (30 mcg ethinyl estradiol and 0.15 mg desogestrel), or no estrogen/progestin for one year. The 17β-estradiol patch arm (but not the oral pill arm) demonstrated improved lumbar spine, hip and femoral neck BMD Z-scores, measured using dual-energy X-ray absorptiometry (49). Similarly, the transdermal patch group (vs. the oral pill group) demonstrated greater increases in total and trabecular vBMD, trabecular number, and cortical area and thickness at the distal tibia (50). They also had greater percent increases in tibial total vBMD and radial cortical vBMD vs. the oral pill group, and greater percent increase in cortical area and thickness at both sites vs. the oral pill and no-estrogen groups. Lack of effects of the oral pill were likely because of IGF-1 suppression (an osteoanabolic hormone) and an increase in sex hormone binding globulin (with a decrease in bioavailable gonadal hormones) in the pill arm because of hepatic first-pass metabolism (not seen in the patch and no-estrogen arms), and significant reductions in sclerostin, preadipocyte factor-1 and brain-derived neurotrophic factor (that otherwise inhibit osteoblastic activity) in the transdermal 17β-estradiol arm (51). Consistent with this, N-terminal propeptide of type 1 procollagen, P1NP, a marker of bone formation, decreased significantly in young women treated with the pill, but not the transdermal 17β-estradiol patch. Finally, the increase in estradiol observed over 12 months was associated with increases in BMD at multiple sites (51). A study by Nose-Ogura et al. similarly reported an increase in BMD in athletes with low body weight receiving transdermal 17β-estradiol with cyclic progestin for 12 months; however, increases in BMD were more marked in athletes who attained spontaneous menstrual recovery (52).

Positive effects of transdermal 17β-estradiol on BMD have also been reported in adolescents with AN in an 18-month randomized controlled trial (53). Interestingly, adding rhIGF-1 in replacement doses to transdermal 17β-estradiol did not result in further benefits to BMD above that observed with transdermal 17β-estradiol alone (54), despite the described deleterious impact of low IGF-1 levels on bone.

Other agents currently being considered as treatments for low BMD include human recombinant parathyroid hormone (PTH) and bisphosphonates. Teriparatide (or human recombinant PTH 1–34) has a bone anabolic effect by inhibiting osteoblast apoptosis and enhancing osteoblast function. It is commonly used in postmenopausal osteoporosis but has not been extensively studied in young females with functional hypothalamic amenorrhea, particularly athletes. Furthermore, teriparatide is contraindicated in pregnancy, which limits its use in young female patients (1,55). Of note, the requirement for a black box warning for use of teriparatide in patients with increased baseline risk for osteosarcoma, including children with open epiphyses, was recently lifted by the FDA. Studies in older adult women with AN who received teriparatide over 6 months did demonstrate an increase in posteroanterior spine and lateral spine BMD. Additionally, P1NP levels increased after 3 months and remained elevated until the end of treatment in the teriparatide but not the placebo group (56). A case study of two premenopausal women with stress fractures showed improved pain control and faster healing after four weeks of treatment with teriparatide (57). More recently, a study reported an increase in spine, femoral neck and hip BMD in young women with AN, ages 18–35 years, with confirmed closed epiphysis and very low BMD or low BMD with history of at least one fragility fracture, treated with teriparatide over 2 years. However, this study reported using HR-pQCT that radial cortical vBMD and thickness decreased in the teriparatide vs. placebo group with no changes in radius trabecular bone or tibial parameters (58).

Bisphosphonates inhibit osteoclastic bone resorption and studies in adult women with AN have demonstrated an improvement in spine BMD with risedronate (vs. placebo) for a year (59). A similar improvement in spine vBMD was not observed in adolescents with AN randomized to alendronate vs. placebo for a year, although small increases were noted at the femoral neck (60). There are concerns regarding the use of bisphosphonates in young women given their very long half-life, particularly around possible teratogenicity; although data to date are reassuring (61).

Cognitive function and factors driving eating behavior:

Data suggest a deleterious effect of hypoestrogenism on verbal memory and executive function, specifically cognitive flexibility, in oligo-amenorrheic athletes compared to eumenorrheic athletes and/or non-athletes (62). Further, estrogen replacement, particularly as the transdermal 17β-estradiol patch with cyclic oral progesterone given for a 6-month period was noted to improve both verbal memory and cognitive flexibility (62). A subsequent study reported a significant improvement in drive for thinness and body dissatisfaction scores, and a reduction in uncontrolled eating following 12 months of transdermal 17β-estradiol replacement therapy (with cyclic progesterone) (63). This is relevant given the high prevalence of disordered eating or eating disorders among amenorrheic athletes. Most recently, studies have shown that sport body image ideals and the perceived power dynamic between the athlete and their coach may influence risk for disordered eating (64).

Conclusion:

Studies examining long-term effects of energy deficiency in female athletes have demonstrated deleterious effects on bone health, cognitive outcomes and eating behaviors. These effects are particularly problematic during the critical adolescent and young adult years of peak bone accrual and brain maturation. Evidence points to these effects being secondary to hormonal changes adaptive to the low energy availability state; thus normalizing the latter is essential to improving bone, cognitive and eating outcomes. Future research should be aimed at identifying individuals at risk of earlier onset of Triad symptoms. Furthermore, research into novel treatment options to improve bone outcomes in younger patients lags behind research in older women and is important to implement.

Key points:

• The Female Athlete Triad includes the triad of low energy availability, menstrual dysfunction and low bone mineral density.

• Hormonal changes in low energy availability states are mostly adaptive and include a state of growth hormone resistance with low IGF-1 concentrations despite increased GH secretion, hypercortisolemia, lower T3 concentration, higher ghrelin and lower leptin levels, and lower LH pulsatile secretion than in controls with normal energy availability.

• A major consequence of these hormonal changes is low bone mineral density, impaired bone geometry and microarchitecture and reduced bone strength, associated with a higher risk of fractures, particularly stress fractures.

• Treatment includes increasing caloric intake to meet the needs of exercise energy expenditure, reducing exercise activity, or both. However, if menses do not resume after 6–12 months of lifestyle changes, estrogen replacement therapy, via the transdermal route, needs to be considered (with a cyclic progestin). At this time, there is insufficient evidence for the use of therapies such as teriparatide (a recombinant parathyroid hormone) and bisphosphonates, in young female athletes.