In line with the updated definition, Hoch et al. [19] assessed the prevalence of all three Triad components in a sample of 80 female high school varsity athletes across multiple sports. Around one third (36 %) of the athletes were classified as having low daily energy availability (≤45 kcal/kg of lean body mass), with 6 % having energy availability of less than 30 kcal/kg of lean body mass. In addition to the gold standard of energy availability, eating pathology was also measured, with only 4 % of athletes classified as at risk of disordered eating based on having EAT-26 scores of greater than or equal to 15. Over half of athletes (54 %) reported menstrual dysfunction, with 30 % reporting secondary amenorrhea and 15 % reporting oligomenorrhea, both operationalized using the 2007 ACSM definition [5]. Hormonal contraception was assessed and reported, but results were not stratified by use. Serum hormones were also assessed to eliminate other endocrinologic or gynecologic causes of menstrual dysfunction. BMD was assessed using the 2007 ACSM [5] definitions and using DXA technology: 3 % of athletes had Z-scores of less than −2, and 13 % had Z-scores between −1 and −1.9. Overall, the authors found that 1 % of the sample had all three Triad conditions, between 4 and 18 % had any two Triad conditions, and between 16 and 54 % had any one Triad condition.

Pollock et al. [20] assessed the prevalence of the conditions of the Triad in a sample of 44 elite female endurance runners (mean age 22.9 years, SD = 6.0 years). BMD was measured at several locations on the body, with Z-scores varying by location. Low BMD, as measured by Z-scores of between −1 and −2, was characteristic of 34.2 % of the sample at the lumbar spine, 13.8 % at the femoral neck, 29.6 % at the radius, and 4.9 % for the total body. Z-scores below −2 were characteristic of 7.3 % of the sample at the lumbar spine, 33.3 % at the radius, and 0 % at the femoral neck and for the total body. Energy availability was not assessed. Rather, disordered eating was assessed using the Three-Factor Eating Questionnaire (TFEQ), a self-report measure of disordered eating cognitions, including cognitive restraint. Athletes scoring in the upper quartile for this sample on any of the three TFEQ subscales were classified as engaging in disordered eating. Secondary amenorrhea or oligomenorrhea, assessed using a self-report questionnaire, were present in 52.3 % of the sample. While information on hormonal contraceptive use was reported, results were not stratified by its use. Considering the sample as a whole, 15.9 % were classified as having all three components of the Triad, with menstrual dysfunction, disordered eating and low BMD.

Coehla et al. [17] also used the 2007 Triad definition to assess prevalence of the Triad in a sample of 24 adolescent female tennis players. Although the participants were from only one sport and the sample size was small—thus producing imprecise estimates with wide confidence intervals—this study is notable because it is the first and only study to date to estimate the prevalence of the Triad using the spectrum concept. The authors divided the Triad into Stage I and Stage II to reflect graded severity. Stage I was considered to be “moderately severe” and was operationally defined as having daily energy intake of less than or equal to 45 kcal/kg of lean body mass, presence of primary or secondary amenorrhea or oligomenorrhea, and a BMD Z-score of less than or equal to −1.0. Stage II was considered to be “severe” and was operationally defined as meeting a clinical threshold for at least one of three validated self-report measures of disordered eating (Eating Attitudes Test-26 > 20, BSQ > 80, Bulimic Investigatory Test Edinburgh > 10), presence of amenorrhea, and a BMD Z-score less than or equal to −2.0. Of note is that building from the ACSM’s definition, a gold standard measure of a “severe” classification should include having daily energy intake ≤30 kcal/kg of lean body mass and not solely the presence of eating pathology [5]. Nonetheless, using these definitions, 4.2 % of the athletes met criteria for all three components of Stage I of the Triad, with 5.0 % having low energy availability, 33.3 % having menstrual irregularity, and 25.0 % falling one or more standard deviations below the age- and sex-adjusted mean. No athletes in this sample met criteria for all three components of Stage II of the Triad, with 50 % meeting criteria for disordered eating, 8.3 % classified as amenorrheic, and 0 % having BMD Z-score ≤ −2.0. There may be disagreement about whether a two-stage approach and the choice of measures and thresholds at each stage in this study were most appropriate to represent the spectrum of risk; however, this study represents an important starting point for the design of future studies to assess the prevalence and severity of the Female Athlete Triad.

Since the 2007 update of the ACSM definition of the Triad to encompass a spectrum of energy availability, with or without disordered eating, few studies have assessed the prevalence of energy availability among female athletes using validated methods. In addition to the work of Coelho et al. [17] and Hoch et al. [19] as described above, Reed et al. [22] and Da Costa et al. [23] measured energy availability among female athletes. In a sample of 77 adolescent swimmers (age 11–19 years), Da Costa and colleagues found 16 (20.8 %) to have daily energy availability below 45 kcal/kg of lean body mass, and 6.5 % had daily energy availability below 30 kcal/kg of lean body mass [23]. In a rare longitudinal study, Reed et al. [22] measured how energy availability varied across the competitive season on a team of 19 female collegiate soccer players. The percentage of players with daily energy availability below 30 kcal/kg of lean body mass was 26.3 % pre-season, 33.3 % mid-season, and 11.8 % post-season. The authors found this difference to be driven by lower dietary energy intake at lunch and dinner during mid-season as compared to other points in the season.

Because most studies available on the prevalence of the Triad (and components of the Triad) were published prior to the 2007 ACSM definition update, many more studies report the prevalence of disordered eating than energy availability. In the studies reviewed by Gibbs et al. [16], disordered eating and eating disorders were assessed using structured clinical interviews or validated survey measures that measure eating pathology or risk factors for eating pathology. In summary, 35 studies of the prevalence of clinical or subclinical disordered eating or eating disorders were either included in the review of Gibbs et al. [16], or were published later and met their inclusion criteria. The prevalence of subclinical disordered eating among all exercising women included in the studies reviewed ranged from 2.9 to 60 % [4, 16]. The prevalence of clinical disordered eating, such having an EAT-26 score of ≥20, ranged from 7.1 to 89.2 % [16, 17, 23]. Estimates of the prevalence of clinical eating disorders ranged from 0.9 to 40.8 % for bulimia nervosa and from 0 to 48.0 % for anorexia nervosa [16, 24]. While these rates of eating disorders and disordered eating suggest that eating pathology, and likely inadequate energy intake, may be endemic in certain populations of female athletes, they do not provide concrete information about energy availability according to the 2007 ACSM standards.