Are elite female soccer athletes at risk for disordered eating attitudes, menstrual dysfunction and stress fractures?

Introduction

The combination of disordered eating, amenorrhea, and osteoporosis in female athletes was first recognized and defined as The Female Athlete Triad (Triad) in 1992 by the American College of Sports Medicine.[1, 2] The updated position statement in 2007 defined the Triad as a spectrum of energy availability, menstrual function and bone health.[3] Menstrual dysfunction and restrictive eating in adolescence is associated with low bone mineral density, which has been shown to persist at 4.7 and 6 years despite resumption of normal menses.[4, 5] The prevalence of all 3 components of the Triad ranges from 1-14%. Up to 78% of female athletes[6] have at least one of the three components of the Triad.[6–12] Burrows and colleagues[13] suggest the Triad components do not identify all females at risk and suggest that exercise related menstrual dysfunction, eating disorders, and osteopenia may be more appropriate markers of risk.

Studies have surveyed female athletes involved in various sports and levels of play regarding the components of the Triad[6–12, 14] but little is known about components of the Triad as they pertain specifically to soccer. In a study of 17 Norwegian national soccer team athletes, 24% had an eating disorder, 9.3% had menstrual dysfunction and 13% had a history of stress fracture. [15] All athletes had normal bone mineral density. The percentage with menstrual dysfunction is higher than that reported in the general population in data from the National Health Interview Survey, which demonstrated a prevalence of 5.3% for menstrual dysfunction in US women between 1984 and 1992.[16] Although the number of reports in the literature of lower extremity stress fractures in female soccer athletes is small,[17–22] a study by Yildirim and colleagues[23] found 21 female soccer athletes to have 32 focal tracer uptake changes in the femur and tibia on three-phase bone scan consistent with bone stress injuries. The authors concluded stress fractures are likely under diagnosed and treated.

Of the 18 million Americans who play soccer, 78% are under 18 and over 40% are female.[24] Since the inception of Title IX, female participation in the National College Athletic Association (NCAA) sports program has increased by greater than 5-fold from 1972 to 2008.[25, 26] Women’s collegiate soccer increased from 1,855 athletes on 80 teams during the 1981-82 season to 22,682 athletes on 956 teams during the 2007-08 season making women’s soccer the NCAA sport with the greatest number of athletes. [25] Concurrently, the number of girls playing high school soccer in the U.S. has also grown from 700 in 1972-73 to 344,534 in 2008-09. [24] Given the increasing prevalence of female soccer participation in the United States, particularly among youths and adolescents, a large number of female athletes may be at risk for developing components of the Triad or Triad itself. The purpose of this study is to determine the prevalence of stress fractures within the previous 2 years, menstrual dysfunction one year following the onset of menses, and disordered attitudes towards body perception and eating attitudes in an elite group of female soccer athletes.

Methods:

After obtaining Institutional Review Board (IRB) approval, female soccer athletes were recruited from a youth soccer club, an NCAA Division I university team, and a women’s professional team. Athletes were recruited at the beginning of their respective competitive seasons 2009-2010. One team from each year of age for the grade school, middle school and high school girls were recruited. These athletes were considered elite for their age group because they were competing at the highest possible level of play including the youth soccer club athletes that competed regionally and nationally. Athletes completed questionnaires regarding their age, height, weight, age at menarche, menstrual function, history of an eating disorder and a detailed history of musculoskeletal injuries, including stress fractures. All responses were self-reported. Athletes were asked their age at onset menses. For any year following the first year of menses, athletes were asked if they had missed 3 menstrual cycles within a 12 month period. Stress-fractures were recorded if they had been confirmed by a physician. All athletes reported soccer to be their only or primary sport. The grade school and middle school athletes did participate in other sports and play.

The Eating Attitudes Test (EAT-26) was administered to the athletes to assess their body image perception, attitudes towards eating, and eating behaviors. The EAT-26, validated by Garner et al. [27] assigns 0 to 3 points to each of the 26 questions for a maximum of 78 possible points. A score of 20 or above indicates an at-risk individual and a score of 10-19 indicates an intermediate risk individual.

All athletes from each team participated in the study. Data was analyzed based on competitive level of experience: grade school/middle school, high school, collegiate, and professional levels. Menstrual dysfunction was defined as a disruption in the normal menstrual cycle or onset of menarche at 15 years of age or older.[28] [29] Athletes one year past the onset of menarche were asked if they had a history of 3 or more missed periods in the past year.[29] The percentages and mean for the entire group and each subgroup were determined for the EAT-26, menstrual function, body mass index (BMI), and self-reported history of stress fractures. Comparisons between subgroups were completed for the EAT-26 using a chi-square analysis. Unless otherwise noted, data are reported as mean ± standard deviation. All analyses were performed by a research statistician who used SAS software (SAS Institute Inc., Cary, NC, USA).

Results:

All athletes from each team invited to participate completed the study. Two hundred twenty female soccer athletes ranging from pre-adolescent (10 years old) to adult (30 years old) with a mean age of 16.4±4 years were enrolled and completed the questionnaires. On average, athletes participated in 9.5±6 (range 5.2+ 2 to 17.9 + 5) hours of organized soccer training per week. Seventy-five athletes (34.0%) age 10-14 years were in the grade school/middle school group; 81 athletes (36.8%) age 15-17 years were in the high school group; 28 athletes (12.7%) age 18-21 years were in the collegiate group and 36 athletes (16.3%) age 19-30 years were in the professional group. The average BMI for all groups was 20.8+2 kg/m²⁾ and there was no significant difference by age (Table 1). Only 3 athletes’ (1.6%) BMI were considered underweight. (BMI of less than the 5^th percentile for their age up to age 18 or <17.5 kg/m²) [30, 31] for their age (Tables 2 and 3).

Table 1.

Average age in years and body mass index (BMI) for each group of athletes with cutoffs for low BMI of less than <5^th percentile for age for 10-19 yo and adult BMI low range for 20 yo and older)

	Number athletes in group	Average age (years)	Average BMI in kg/m2	Low BMI for age in < kg/m2
Grade school/Middle school (10-14 years)	75	12.3	18.6±2	<5th percentile for age 10 yo < 14 11 yo< 14.4 12 yo <14.8 13 yo < 15.3 14 yo < 15.8
High school (15-17 years)	81	15.8	20.9±2	<5th percentile for agea 15 yo < 16.4 16 yo < 16.8 17 yo < 17.2
College (18-21 years)	28	19.2	22.2±2	<5th percentile for age 18 yo < 17.5 19 yo < 17.5 20 Low BMI 20 yo BMI <17.5 kg/m2
Professional (19-30)	36	24.1	22.8±2	Low BMI <17.5 kg/m2

Table 2.

Stress fractures, body mass index (BMI), menstrual dysfunction and eating attitudes

Age / Level of Soccer	No.	BMI (mean) kg/m2	No. athletes with menstrual dysfunction	EAT −26 score (mean)	No. athletes EAT-26 score		Lower extremity stress fracture (no. athletes)
					≥10	≥20
Grade & Middle School	75	18.6±2	45* (64%)	4.1	4 (5.3%)	0	1 (0.1%)
High school	81	20.9±2	16 (19.0%)	4.4	5 (6.1%)	1 (1.9%)	11 (13.6%)
College	28	22.2±2	5 (17.9%)	5.9	5 (17.8%)	0	2 (7.1%)
Professional	36	22.8±2	7 (19.4%)	3.3	3 (8.3%)	0	5 (13.8%)
All groups	220	20.8±2	28/145 (19.3%)	4.4	17 (7.7%)	1 (0.05%)	19 (8.6%)

^*Not menstruating, do not count in the totals for this column

Table 3.

Body weight in relation to stress fractures, eating attitudes and menstrual dysfunction

Body weight (based on age)	No. athletes	EAT-26 score (mean)	No. athletes EAT score		Menstrual dysfunction	Stress fracture
			≥10	≥20
Underweight	3	3.0	0	0	0	0
Appropriate	166	4.3	17	1	27	24
Overweight	3	2.7	0	0	1	2

NOTE: Only includes athletes with established menarche (175/220).

Forty-eight (64%) of athletes in the youngest group had not yet reached menarche at the time of this study. The average onset of menarche reported by the older subgroups was 13+1 (range 10-17) years. Sixteen (19%) high school athletes, 5 (17.9%) college athletes, and 7 (19.4%) professional players reported menstrual dysfunction.

The mean EAT-26 score for all athletes was 4.4+4 and were considered low risk for an eating disorder. A total of 18 athletes (8.1%) were considered at risk for an eating disorder. One athlete scored >20 points and 17 (7.7%) athletes scored between 10 and 19 points. The percentage of at risk athletes was highest in the collegiate group (5/28, 17.8%) compared to the other groups (grade school 5.3%, high school 6.1%, professionals 8.3%) (p=0.03). Athletes with an EAT-26 score ≥ 10 points had a significantly higher prevalence of menstrual dysfunction in the past year compared to athletes with an EAT-26 score of less than 10 (p=.02).

Overall, 19 athletes (8.6%) reported a history of 19 stress fractures of the lower extremity including the hip, femur, tibia, fibula, and foot. Two of the athletes with an EAT score ≥ 10 had a history of lower extremity stress fracture. None of the athletes classified as underweight reported a history of stress fracture (Table 3). There was no difference in the reported incidence of stress fractures between athletes with menstrual dysfunction as compared to those with no menstrual dysfunction.

Discussion:

This is the first report to describe the prevalence of low BMI, disordered eating attitudes, and stress fracture history in a large cohort of female soccer athletes across different ages. Only 3 athletes were found to have a BMI considered underweight for their age and 18 athletes were considered to be at risk for an eating disorder based on the EAT-26 score. Therefore, the majority of athletes had a BMI and Eat-26 score that fell within a normal range for their age and gender. Nevertheless, this cohort of elite female soccer athletes reported a history of menstrual dysfunction (19.3%) and lower extremity stress fractures (8.6%).

This is the first study to describe a delay in the onset of menarche in a cohort of soccer athletes. The mean age of onset of menarche for the soccer athletes was slightly delayed at 13+1 years of age as compared to the United States national average of 12.43 years[32] and more recent data based on ethnicity similar to our cohort of non-Hispanic Caucasian girls 12.77 +/− 0.09.[33] This older age at onset is similar to that reported in runners, gymnasts, tennis athletes, ballet dancers and swimmers.[34–40] Further, once menses was established, all groups reported similar menstrual dysfunction percentages ranging from 18% to 20%. This is compared to 5.3% menstrual dysfunction in the general US population.[41] A prior study by Warren and colleagues[42] attributed the wide incidence range (6-79%) of menstrual dysfunction in female athletes to be specific to sport and level of competition. Our data demonstrate that regardless of age, almost one in five female soccer athletes participating at a high level of competition have menstrual dysfunction.

A study of 300 collegiate level long distance female runners reported a 19.4% incidence of eating disorders, a 23.0% prevalence of menstrual dysfunction, and a 29.1% prevalence of deficient calcium intake.[43] However, the incidence of stress fractures was not reported. Compared to endurance athletes, the soccer athletes in this study had a lower prevalence of eating disorders and menstrual dysfunction, but still presented with a history of stress fractures. This may represent unintentional energy availability deficits. However, at the collegiate level, one in five soccer athletes were at risk for eating disorders based on their EAT-26 score. The collegiate population may be particularly at risk for this aspect of the triad. A separate retrospective analysis found 31.3% incidence of stress fractures among collegiate female track and field athletes.[44] This same study reported an incidence of stress fracture of 2.6% among collegiate female soccer athletes, lower than the incidence of stress fractures in our cohort (8.6%). These lower incidences in stress fractures may be related to the specific sport. Fredericsson and colleagues found athletes participating in ball sports developed fewer stress fractures than athletes participating in repetitive low impact sports such as running. They proposed that ball sports require multidirectional loading with high ground reaction forces that may result in more fracture-resistant bones.[45, 46]

Our findings differed from that of the Norwegian team study with regards to the incidence of aberrant eating attitudes and could be attributed to differences in methods of measurement of eating attitudes and culture. The Norwegian study used both a written questionnaire and a clinical interview which may be more sensitive than administering only a questionnaire. Of the 24% of female Norwegian soccer athletes determined to have an eating disorder based on the clinical interview, only 5.9% reported having an eating disorder on the written questionnaire. The same Norwegian study also reported that 21.2% of an aged-matched control group met criteria for an eating disorder, which is significantly higher than the 3-5% previously reported in both America and Norway for female athletes in general.[13, 20, 47, 48] Cultural differences between female athletes in the United States and Norway exist and may affect the findings on questionnaires regarding eating attitudes.

While this study and the Norwegian study differed in the number of athletes at risk for eating disorders, the prevalence of stress fractures and menstrual dysfunction were similar. The rates of menstrual dysfunction in both studies indicate soccer athletes may be at risk to develop stress fractures, other problems associated with hypoestrogenism including cardiovascular disease and infertility.[49, 50] Our data is consistent with that of the Norwegian study that suggest components of the female athlete triad are unrecognized and therefore likely undertreated in female soccer athletes. Athletes that are underweight may receive more attention regarding investigation of the Triad risk factors from healthcare providers based on their appearance. Our data complements the previous study that suggests despite having a BMI in the appropriate range for age and gender; female athletes may be at risk for menstrual dysfunction and stress fractures.

Differing from participants in aesthetic and endurance sports the soccer athletes in this study had a lower prevalence aberrant eating attitudes associated with eating disorders and menstrual dysfunction. Greydanus and colleagues reported rates of menstrual dysfunction to be as high as 44% of ballet dancers, 50% of female triathletes and 51% of endurance runners.[37] Despite the lower prevalence of aberrant eating attitudes and menstrual dysfunction in the soccer athletes as compared to reports involving aesthetic and endurance athletes, the soccer athletes reported a greater percentage of stress fractures than previously known.

The current study has several limitations. Height and weight were self-reported which may result in error because of inaccurate recall. Eating attitudes were reported only by written questionnaire and not by an eating journal or 3-day recall of consumed food. An individual interview may have revealed other athletes at risk. Athletes were asked to recall being diagnosed with a stress fracture by a physician. This retrospective recall may result in under reporting. Athletes were considered to have had a stress fracture if the diagnosis was made by a physician. We cannot determine if imaging confirmed all fractures. Age is a risk factor for developing stress fractures. We did not assess the exact age of the athlete at the time of the fracture but asked them to a report a history of a fracture within the past 2 years. Furthermore, bone mineral density was not measured in these athletes. Finally, this investigation was conducted at a single metropolitan area and may not reflect regional, age, and competition level-related differences across the United States.

In conclusion, elite female soccer athletes without aberrant attitudes towards eating and food have delayed onset of menarche, menstrual dysfunction, and a history of stress fracture. Increased awareness by athletes, parents, coaches, athletic trainers and physicians may help identify athletes with components of the female athlete triad in soccer and perhaps other team sports. Education about the detrimental effects of menstrual dysfunction and the importance of adequate energy balance and nutritional requirements should be encouraged to minimize the risk for poor bone health, manifesting as stress fracture in the short term and osteoporosis over the long term in these athletes. In female soccer athletes, education regarding appropriate energy intake may be particularly relevant as unintentionally inadequate nutrient intake is a potential driving force behind elevated risk for menstrual irregularity and stress fracture in these athletes with largely normal eating attitudes. Given the large numbers females participating in soccer, further investigation regarding energy balance is important.