You are here

Article Text

Article menu

Download PDFPDF

Original article
Evidence of a cumulative effect for risk factors predicting low bone mass among male adolescent athletes
  1. Michelle T Barrack1,
  2. Michael Fredericson2,
  3. Adam S Tenforde3,
  4. Aurelia Nattiv4
  1. 1California State University, Long Beach, Long Beach, California, USA
  2. 2Stanford University, Palo Alto, California, USA
  3. 3Spaulding Rehabilitation Hospital, Cambridge, Massachusetts, USA
  4. 4University of California Los Angeles, Los Angeles, California, USA
  1. Correspondence to Michelle T Barrack, Department of Family and Consumer Sciences, California State University, Long Beach, 1350 Bellflower Blvd, Long Beach, CA, 90840 USA; michelle.barrack{at}csulb.edu

Abstract

Background Limited research has evaluated risk factors for low bone mineral density (BMD) in male adolescent athletes.

Aims/objectives To evaluate predictors of low BMD (defined as BMD Z-score <−1.0) in a sample of male adolescent distance runner and non-runner athletes.

Methods Male adolescent athletes completed a survey characterising sports participation, nutrition, stress fracture history, dual energy X-ray absorptiometry (DXA)-measured BMD and body composition. Independent t-tests and analysis of covariance (ANCOVA) evaluated group differences; logistic regression evaluated low BMD risk factors.

Results Runners (n=51) exhibited a lower body weight (p=0.02), body mass index (BMI) (kg/m2) (p=0.02), per cent expected weight (p=0.02) and spine BMD Z-score (p=0.002) compared with non-runners (n=18). Single risk factors of low BMD included <85% expected weight (OR=5.6, 95% CI 1.4 to 22.5) and average weekly mileage >30 in the past year (OR=6.4, 95% CI 1.5 to 27.1). The strongest two-variable and three-variable risk factors included weekly mileage >30+ stress fracture history (OR=17.3, 95% CI 1.6 to 185.6) and weekly mileage >30+<85% expected weight + stress fracture history (OR=17.3, 95% CI 1.6 to 185.6), respectively. Risk factors were cumulative when predicting low BMD (including <85% expected weight, weekly mileage >30, stress fracture history and <1 serving of calcium-rich food/day): 0–1 risk factors (11.1%), 2 risk factors (42.9%), or 3–4 risk factors (80.0%), p<0.001).

Conclusions Male adolescent runners exhibited lower body weight, BMI and spine BMD Z-score values. The risk of low BMD displayed a graded relationship with increasing risk factors, highlighting the importance of using methods to optimise bone mass in this population.

  • Bone density
  • Body weight
  • Female athlete triad
  • Nutrition

https://doi.org/10.1136/bjsports-2016-096698

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Prior investigations have identified female endurance runners as a population with an elevated risk of developing low bone mineral density (BMD).1–4 Studies evaluating female collegiate athletes representing all sport types report a 10–11% prevalence of meeting criteria for low BMD, defined as a BMD Z-score5 or BMD T-score <−1.0.1 ,2 In contrast, Cobb et al3 and Barrack et al4 reported a 36% and 39% prevalence of low bone mass among a sample of collegiate and high school runners, respectively. Factors previously associated with low bone mass among female endurance runners include elevated dietary restraint, amenorrhea, a longer duration of participation (5 or more vs <5 seasons) in endurance running,4 higher intake of dietary fibre, phytic acid, vegetable protein and history of a stress fracture.6 ,7 Lower body mass index (BMI thresholds of 18.5 kg/m2 and 17.5 kg/m2) has been identified as a risk factor for low BMD.7 ,8 A majority of the risk factors are related to low energy availability (LEA), either indicating or contributing to an energy deficiency.

    In contrast to female athletes, few studies have evaluated BMD and risk factors for impaired bone health among male athletes. The limited research available provides evidence that male runners may also represent a population at risk for low BMD. Hind et al reported that 36% of male adult runners (16 of 44) exhibited BMD T-score <−1.0,9 while Tenforde et al found 21% of male adolescent runners (9 of 42) with BMD Z-score ≤ −1.0.7 Fredericson et al reported that among a sample of male adult athletes, right hip and lumbar spine BMD in endurance runners (n=15) was no different than sedentary controls (n=15) and significantly lower compared with soccer players (n=15).10

    While these investigations provide preliminary evidence of low BMD among male endurance runners, the studies are limited by their sample size. Only one study evaluated adolescent male runners,7 a population of increasing importance. Adolescence is a critical period for developing bone mass, especially during peak skeletal growth for both sexes.11 Additionally, while several investigations have evaluated variables associated with low bone mass in endurance runners, none reported on potential additive effects of multiple risk factors for the development of low BMD. To the best of our knowledge, there are no studies among adolescents comparing bone mass values between endurance runners and athletes participating in non-endurance running sports. Therefore, the current study aims to evaluate the proportion of athletes with low BMD (defined as a BMD Z-score <−1.0 in an athlete participating in a weight-bearing sport)5 and related single and multivariable risk factors previously associated with low bone mass among a sample of male adolescent runner and non-runner athletes.

    Methods

    Study design and participant characteristics

    This cross-sectional study evaluated bone mass, body weight, body composition, stress fracture history, training volume and dietary patterns among 69 male adolescent athletes (n=51 endurance runners, n=18 non-runner athletes). Eligible athletes were aged 13–19 years and without medical conditions known to affect bone health. Runners were recruited who participated on a high school cross-country or track and field team. Non-endurance runner athletes were currently participating in a ball or power sport.12 The 18 non-runner athletes participated in soccer (n=11), basketball (n=3) and volleyball (n=3) as their primary sport. Data were acquired at two study sites. All participants under the age of 18 years provided written assent and parental consent prior to undergoing study measures. Participants aged 18 years and older provided written consent. The study was approved by institutional review board (IRB) at each site.

    Data collection

    Subject participants were recruited from two sites including Southern California Los Angeles-Santa Monica and Northern California San Francisco Bay Area. In Southern California, study recruitment occurred during a team parent meeting. Athlete participants attended a 1-hour appointment at the Santa Monica-UCLA Medical Center where they had their height and weight measured and completed a dual energy X-ray absorptiometry (DXA) scan. Each athlete completed a survey inquiring about dietary intake, sports participation and training volume (runners), stress fracture history and perceptions regarding weight and body composition.

    Athletes recruited from Northern California were part of a larger investigation reported previously.7 ,13 Male runners who participated on their high school cross-country and track and field team were eligible to enrol in a study evaluating risk factors for stress fracture. Each subject completed an online questionnaire about baseline risk factors for stress fracture and low bone density including injury history, running history, training variables and diet. Subjects who completed the initial survey were invited to complete a DXA scan and 3-day food frequency questionnaire.14 At time of DXA, height and weight were measured.

    Anthropometric data: Height and weight without shoes were measured, to the nearest 0.5 inch and 0.5 pound, respectively. Data from the CDC growth charts for boys aged 2–20 years were used to calculate BMI Z-score.15 Underweight status was evaluated using two BMI thresholds (<17.5 kg/m2 and <18.5 kg/m2), BMI Z-score <−1.0 and <85% of expected weight.16–18

    Bone density and body composition: Areal BMD (lumbar spine and total body less head (TBLH)) bone mineral content (BMC) and body composition (per cent body fat, fat mass, fat-free mass (FFM) and lean body mass) were analysed using DXA by a certified technician. Participants were scanned with a Hologic QDR 4500A (n=27; 100% non-runner athletes, 18% runner athletes) (Hologic, Bedford, Massachusetts, USA) and GE Lunar iDXA (n=42; 82% runner athletes) (GE Healthcare, Chicago, Illinois, USA). Each site maintained a quality-assurance programme and the same trained and licensed technician at each site performed the DXA scans as previously described.7 ,8

    Merging databases

    Researchers followed a rigorous process to merge databases from the two populations of male adolescent athletes by evaluating and comparing survey tools item-by-item. This allowed for the identification of identical questions or questions addressing similar constructs.8 ,19 Only variables consistent among databases were used in the current study and analysis.

    Risk factors for low BMD

    Potential risk factors for low BMD in male adolescent athletes were determined a priori to include variables associated with impaired BMD in other populations: markers of LEA (ie, BMI<17.5 kg/m2, BMI<18.5 kg/m2, body weight <85% expected, BMI Z-score <−1.0, higher weekly running mileage), nutrition (<1 serving of a calcium-rich food/day, calcium and vitamin D supplement use), history of stress fracture and sports participation in the lean sport of running.4 ,7 ,8 ,15–18

    Statistical analyses

    Mean and SEM were reported for continuous variables and categorical variables (ie, low BMD) were expressed as per cent. Independent samples t-tests evaluated means and SEM between runner and non-runner athletes, while χ2 analyses assessed differences in categorical variables between the two groups. Between-group BMD and BMC value comparisons were adjusted for age and FFM, BMD Z-score comparisons were adjusted for FFM using analysis of covariance (ANCOVA). Logistic regression analysis was used to calculate ORs and 95% CIs and to identify risk factors for low BMD. Data were analysed using the SPSS software, V.23.0 (SPSS, Chicago, Illinois, USA).

    Results

    Descriptive characteristics for the runner and non-runner athletes are outlined in table 1. Runners exhibited a significantly lower body weight, BMI, BMI Z-score and per cent of expected weight (per cent expected weight calculated using the Hamwi equation) compared with non-runners (table 1). Body fat per cent did not significantly differ between groups (table 1).

    View this table:
    Table 1

    Descriptive characteristics among the endurance runner and non-runner athletes

    Measures of BMD, body composition and injury history are outlined in table 2 and table 3. A significantly higher per cent of runners compared with non-runner athletes met criteria for low body weight (<85% of expected weight) (table 2). Only runners (16.0% of the runner sample) reported a history of stress fracture (table 2). The proportion of runners exhibiting low BMD (defined as BMD Z-score <−1.0) was four times higher compared with non-runner athletes (table 2). Runners also exhibited significantly lower adjusted (FFM) and unadjusted lumbar spine BMD Z-score when compared with non-runner athletes (table 3).

    View this table:
    Table 2

    Proportion of runner and non-runner athletes meeting criteria for categorical body weight, calcium intake and bone health variables

    View this table:
    Table 3

    Total body and lumbar spine BMD, BMC and BMD Z-score values between the runner and non-runner athletes

    Table 4 describes predictors of TBLH and lumbar spine BMD, BMC and BMD Z-score in the sample. The linear regression models included age, FFM, running mileage, history of stress fracture and/or history of participating in a ball sport as covariates. FFM emerged as a significant predictor of total body BMD, BMC and BMD Z-score, lumbar spine BMC and BMD Z-score (standardised β coefficients ranging from 0.35 to 0.80) in the runner and non-runner athletes (table 4). Among the runner athletes, higher cumulative weekly running mileage over the past year was negatively associated with lumbar spine BMD and BMC values (table 4). The linear regression models explained 11–54% of the variability in bone mass for the runners and 51–75% of the variability in bone mass for the non-runner athletes as indicated by the adjusted R2 values (table 4).

    View this table:
    Table 4

    Multiple linear regression analyses assessing predictors of total body and lumbar spine bone mass values in the runner (n=51) and non-runner (n=18) athletes

    Results of the logistic regression analyses evaluating single and multivariable risk factors of low bone mass among the sample are outlined in table 5. Single-variable risk factors significantly associated with low BMD included low body weight (<85% expected weight), low BMI (BMI Z-score <−1.0) and cumulative average weekly running mileage >30 over the past year (table 5), with OR values ranging from 5.6 to 7.0.

    View this table:
    Table 5

    Single and multiple factor predictors of low BMD

    Combined two-variable risk factors significantly associated with low BMD with the highest OR values, include weekly mileage >30+ history of stress fracture (OR=17.3), <85% expected weight +<1 serving of calcium-rich food/day (OR=16.5), mean weekly mileage >30+<85% expected weight (OR=11.0) and mean weekly mileage >30+BMI Z-score <−1.0 (OR=8.0) (table 5). The three-variable risk factors most strongly associated with low BMD were a combination of mean weekly mileage >30+<85% expected weight + stress fracture history (OR (95% CI) 17.3 (1.6 to 185.6)) (table 5).

    Figure 1 presents results from the regression model outlining the additive relationship between risk factor variables and low BMD, with risk factor variables including body weight <85% of expected, stress fracture history, mean weekly mileage >30 and consuming <1 serving of a calcium-rich food per day. OR (95% CI) values associated with low BMD among the male high school athletes exhibiting two or 3–4 risk factor variables (compared with 0-1 risk factors) were 6.0 (95% CI 1.1 to 33.5) and 32.0 (95% CI 3.1 to 335.5), respectively.

    Figure 1
    Figure 1

    Male high school athletes grouped according to risk factors including body weight <85% of expected, stress fracture history, running ≥30 miles per week over the past year, and consuming <1 serving of a calcium rich food per day. Logistic regression analysis yielded odds ratio values demonstrating the additive relationship between number of risk factors and likelihood of low BMD (*P<0.005, §P<0.05).

    Discussion

    Our study aimed to identify predictors of low BMD in a sample of male adolescent runner and non-runner athletes. While a standard definition for low bone mass has not been established for male athletes participating in a weight-bearing sport, we classified ‘low bone mass’ as a BMD Z-score value of <−1.0 combined with associated risk factors, similar to what the American College of Sports Medicine has defined for the at-risk female athletes.5 ,18 Distance runners in the current study exhibited a lower BMD Z-score value compared with non-runner athletes and a higher proportion of runners exhibited low bone mass. Runners were also the only athlete group exhibiting BMD Z-score values ≤−2.0 (n=2), the BMD Z-score used by the International Society for Clinical Densitometry to define low bone mass in the general population of children and adolescents aged 5–19 years.20 ,21 Furthermore, a higher proportion of runners compared with non-runners exhibited low body weight (defined as <85% expected weight), a factor previously associated with low BMD in females.8 Prior stress fractures were reported among one-sixth of endurance runners and none in the non-running athletes. Our results suggest that male adolescent runners may represent an athlete group with increased risk for low bone mass and related characteristics.

    Differences between runner and non-runner athletes

    The higher proportion of male adolescent runners with low BMD Z-score values compared with non-running athletes is consistent with studies in adult male athletes,9 ,10 female adolescent and collegiate athletes.22 ,23 Male adolescent runners also exhibited an overall lower body weight, per cent expected weight and BMI compared with the non-runner athletes, a finding consistent with a prior study conducted among female adolescent athletes.22 Further research is recommended to evaluate risk factors for impaired bone health in male and female adolescent runners and to assess differences compared with non-athletes.

    Predictors of low BMD

    To the best of our knowledge, this serves as the first investigation outlining key predictors of low bone mass in a sample of adolescent athletes. According to logistic regression analyses, one-variable risk factors significantly contributing to the prediction of low BMD (defined as a BMD Z-score <−1) included low body weight (<85% expected weight and BMI Z-score <−1.0) and a higher mean weekly running mileage over the past year (>30 miles/week vs <30 miles/week). The strongest two-variable and three-variable risk factors (OR values 16.5 and 17.3) also included variables related to low calcium intake (<1 serving of calcium/day) and history of stress fracture.

    The relationship between risk factors identified in our investigation and low BMD are consistent with earlier reports. Hind et al9 and Kemmler et al24 reported moderate negative associations between running training volume and lumbar spine BMD in male adult runners. Roberts et al and Wheeler et al25 ,26 also reported a link between increased endurance training and reductions in gonadal hormones in male adult runners, which may negatively affect low bone mass. Barrack et al found an inverse association between longer participation in endurance running (≥5 vs <5 seasons) and lumbar spine BMD in female adolescent runners.4

    While the current investigation did not report energy availability specifically, the link between higher levels of endurance training and lower BMD may be due to the potential contribution of higher levels of training with LEA, particularly if the energy expended from exercise was not adequately replenished. In women, LEA reduces gonadal hormones and other biomarkers associated with bone mass.27 Although our research methodology did not calculate energy availability or assess hormonal markers in the male adolescent athlete, parallels may exist in the mechanisms leading to low BMD in the male athlete as in the female athlete triad and need to be further studied.28 ,29 Factors contributing to LEA may also contribute to hypogonadotrophic hypogonadism and low testosterone (rather than menstrual deficits) in male athletes, a finding supported by prior research among endurance athletes.25 ,26 ,28 Low body weight may also occur as a result of chronic LEA and serves as a key factor in the diagnosis of anorexia nervosa. Prior studies document low body weight as a predictor of low bone mass in samples of active girls and women,8 and female adolescent athletes.30

    Tenforde et al identified a link between drinking less milk and lower TBLH BMD Z-score values in female adolescent runners.13 Barrack et al identified consuming <1300 mg calcium/day as predictive of elevated bone turnover/resorption in female adolescent endurance runners.31 These findings support the importance of adequate calcium intake through diet, a key mineral involved in the mineralisation of bone, paired with optimal energy availability and levels of running training, in maintaining bone mass in young male runners.

    Cumulative effect of risk factors predicting low BMD

    Furthermore, our results suggest a cumulative effect between the four risk factors (body weight <85% expected weight, mean weekly running mileage >30, stress fracture history and consuming <1 serving of calcium/day) and risk of low bone mass, with the prevalence increasing from 11.1% to 42.9% to 80.0% between athletes exhibiting 0–1 compared with 2 and 3–4 of these key risk factors linked to low BMD in this sample of male adolescent endurance runners. This finding is similar to results from an investigation from Gibbs et al which found the prevalence of low BMD increasing for participants meeting criteria for zero (10.2% prevalence) compared with one, two, three or four (61.5% prevalence) female athlete triad-related risk factors (including low BMI, late age at menarche, elevated dietary restraint, lean sport/activity participation, oligo/amenorrhea) among a sample of active girls and women.8 Low body weight (ie, low BMI<18.5 kg/m2 and body weight <85% expected weight) and participation in a leanness sport or related activity (ie, endurance running mileage) were linked to low BMD in Gibbs et al and our current investigation. Comparison of other risk factors identified including dietary restraint, stress fracture history, calcium intake were not evaluated in both investigations and therefore, could not be compared. The additive effect of the risk factors most associated with low BMD in male adolescent athletes has not been previously reported and underscores the importance of developing multifactor screening tools to address adolescent athletes' full risk profile in male athletes as well as female athletes.18 ,28 ,29 Results from this study and future investigations may guide development of such screening tools and methods for optimising bone mineral accrual in young male athletes.

    In the current study, multiple linear regression analyses found running training volume to significantly and inversely contribute to the prediction of BMD and BMC in the male adolescent athlete, particularly at the lumbar spine. Bone mass at the lumbar spine may be more vulnerable to effects of LEA or other related factors than other body sites, as the vertebrae consists primarily of the more metabolically active trabecular bone, which is more responsive to changes in hormone status than cortical bone.32 This characteristic of the trabecular bone may also explain why runners exhibited lower lumbar spine BMD compared with the non-runner athletes, whereas both athlete groups exhibited TBLH values that did not significantly differ.

    The linear regression analyses also indicated that FFM positively contributed to the prediction of BMD, BMC and BMD Z-score values at each bone site in the runner and non-runner athletes. FFM has been previously linked to BMD, BMC and BMD Z-score values in adolescent athletes.22 Higher levels of FFM and, therefore, muscle mass may exert a larger degree of external mechanical forces to bone, thus further stimulating bone mineral accumulation.

    Limitations

    Potential limitations of our current investigation include use of self-report questionnaires that can be prone to recall bias. The non-leanness sport comparison group was smaller than the runner group, reducing statistical power to detect differences between groups. The non-runner athlete group consisted primarily of soccer athletes, rather than athletes participating in a variety of non-leanness sports. Our data were drawn from a multisite study with measurements completed by different investigators at two study locations using different DXA machines. However, we used Z-scores to standardise BMD values and performed careful measures to merge variables from the two studies to ensure consistency and optimise validity of results using similar methods as previously reported.8 ,19 Furthermore, bone age, which may be used to adjust bone mass values among adolescents, was not measured in the current study. Larger investigations among male adolescent athletes are recommended to further evaluate potential predictors of low bone mass.

    Summary and conclusions

    In conclusion, our study provides additional evidence that a subset of male adolescent endurance runners may represent a group with an increased risk of low bone mass compared with adolescent athletes participating in non-leanness sports (ie, soccer, volleyball, basketball). We identified four risk factors, including body weight <85% expected weight, average weekly mileage >30, stress fracture history and consuming <1 serving of calcium/day, that exhibited a cumulative relationship with risk for low BMD. The proportion of athletes meeting criteria for low BMD increased from 11.1% to 80.0% for athletes meeting 0–1 to 3–4 risk factors, respectively. The strongest three-variable risk factor combination was average weekly mileage >30+<85% expected weight + stress fracture history. These findings emphasise the importance of implementing comprehensive screening and preventive measures that address key risk factors associated with low bone mass in male adolescent athletes.

    What are the findings?

    • Male adolescent distance runners exhibited lower body weight, body mass index (BMI), BMI Z-score and lumbar spine bone mineral density (BMD) Z-score values compared with non-runner athletes participating in non-leanness sports.

    • Single-variable risk factors contributing to the prediction of low bone mass, defined as a BMD Z-score <−1.0 in the adolescent athlete sample, included low body weight (<85% expected weight), BMI Z-score <−1.0 and running >30 miles/week over the past year.

    • The strongest three-variable risk factor contributing to the prediction of low bone mass included running >30 miles/week over the past year +<85% expected weight + stress fracture history.

    • A cumulative effect was observed between the four risk factors (body weight <85% expected weight, running >30 miles/week over the past year, stress fracture history and consuming <1 serving of calcium/day) and risk of low bone mass.

    Acknowledgments

    We thank our study participants and support from the Harvard-Westlake Institute for Scholastic Sports Science and Medicine. The Sports Medicine Center and Boswell Human Performance Laboratory at Stanford provided access to the DXA scanner and Phil Cutti performed the DXA scans, without which this study would not have been possible. We also thank the UCLA Metabolic Bone Disease and Osteoporosis Center, UCLA Department of Orthopaedic Surgery for the DXA scans.

    References

    1. ↵
      1. Beals KA,
      2. Hill AK
      . The prevalence of disordered eating, menstrual dysfunction, and low bone mineral density among US collegiate athletes. Int J Sport Nutr Exerc Metab 2006;16:1–23. doi:10.1123/ijsnem.16.1.1
      OpenUrlPubMedWeb of Science
    2. ↵
      1. Torstveit MK,
      2. Sundgot-Borgen J
      . Low bone mineral density is two to three times more prevalent in non-athletic premenopausal women than in elite athletes: a comprehensive controlled study. Br J Sports Med 2005;39:282–7. doi:10.1136/bjsm.2004.012781
      OpenUrlAbstract/FREE Full Text
    3. ↵
      1. Cobb KL,
      2. Bachrach LK,
      3. Greendale G, et al
      . Disordered eating, menstrual irregularity, and bone mineral density in female runners. Med Sci Sports Exerc 2003;35:711–19. doi:10.1249/01.MSS.0000064935.68277.E7
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Barrack MT,
      2. Rauh MJ,
      3. Nichols JF
      . Prevalence of and traits associated with low BMD among female adolescent runners. Med Sci Sports Exerc 2008;40:2015–21. doi:10.1249/MSS.0b013e3181822ea0
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Nattiv A,
      2. Loucks AB,
      3. Manore MM, et al
      . The American college of sports medicine position stand on the female athlete triad. Med Sci Sports Exerc 2007;39:1867–82. doi:10.1249/mss.0b013e318149f111
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Barron E,
      2. Cano Sokoloff N,
      3. Maffazioli GD, et al
      . Diets high in fiber and vegetable protein are associated with low lumbar bone mineral density in young athletes with oligoamenorrhea. J Acad Nutr Diet 2016;116:481–9. doi:10.1016/j.jand.2015.10.022
      OpenUrl
    7. ↵
      1. Tenforde AS,
      2. Sainani KL,
      3. Sayres LC, et al
      . Identifying sex-specific risk factors for low bone mineral density in adolescent runners. Am J Sports Med 2015;43:1494–504. doi:10.1177/0363546515572142
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Gibbs JC,
      2. Nattiv A,
      3. Barrack MT, et al
      . Low bone density risk is higher in exercising women with multiple triad risk factors. Med Sci Sports Exerc 2014;46:167–76. doi:10.1249/MSS.0b013e3182a03b8b
      OpenUrlCrossRefPubMed
    9. ↵
      1. Hind K,
      2. Truscott JG,
      3. Evans JA
      . Low lumbar spine bone mineral density in both male and female endurance runners. Bone 2006;39:880–5. doi:10.1016/j.bone.2006.03.012
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Fredericson M,
      2. Chew K,
      3. Ngo J, et al
      . Regional bone mineral density in male athletes: a comparison of soccer players, runners and controls. Br J Sports Med 2007;41:664–8. doi:10.1136/bjsm.2006.030783
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Bailey DA,
      2. Martin AD,
      3. McKay HA, et al
      . Calcium accretion in girls and boys during puberty: a longitudinal analysis. J Bone Miner Res 2000;15:2245–50. doi:10.1359/jbmr.2000.15.11.2245
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Torstveit MK,
      2. Sundgot-Borgen J
      . The female athlete triad: are elite athletes at increased risk? Med Sci Sports Exerc 2005;37:184–93. doi:10.1249/01.MSS.0000152677.60545.3A
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Tenforde AS,
      2. Sayres LC,
      3. McCurdy ML, et al
      . Identifying sex-specific risk factors for stress fractures in adolescent runners. Med Sci Sports Exerc 2013;45:1843–51. doi:10.1249/MSS.0b013e3182963d75
      OpenUrlCrossRefWeb of Science
    14. ↵
      1. Smucker R,
      2. Block G,
      3. Coyle L, et al
      . A dietary and risk factor questionnaire and analysis system for personal computers. Am J Epidemiol 1989;129:445–9.
      OpenUrlAbstract/FREE Full Text
    15. ↵
      1. Kuczmarski RJ,
      2. Ogden CL,
      3. Grummer-Strawn LM, et al
      . CDC growth charts: United States. Adv Data 2000;314:1–27.
      OpenUrlPubMed
    16. ↵
      1. Hamwi GJ
      . Therapy: changing dietary concepts. New York (NY): American Diabetes Association 1964.
    17. ↵
      1. Barrack MT,
      2. Van Loan MD,
      3. Rauh MJ, et al
      . Body mass, training, menses, and bone in adolescent runners: a 3-yr follow- up. Med Sci Sports Exerc 2011;43:959–66. doi:10.1249/MSS.0b013e318201d7bb
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. De Souza MJ,
      2. Nattiv A,
      3. Joy E, et al
      . 2014 female athlete triad coalition consensus statement on treatment and return to play of the female athlete triad: 1st international conference held in San Francisco, California, May 2012 and 2nd international conference held in Indianapolis, Indiana, May 2013. Br J Sports Med 2014;48:289. doi:10.1136/bjsports-2013-093218
      OpenUrlAbstract/FREE Full Text
    19. ↵
      1. Barrack MT,
      2. Gibbs JC,
      3. De Souza MJ, et al
      . Higher incidence of bone stress injuries with increasing female athlete triad-related risk factors: a prospective multisite study of exercising girls and women. Am J Sports Med 2014;42:949–58. doi:10.1177/0363546513520295
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Lewiecki EM,
      2. Gordon CM,
      3. Baim S, et al
      . International society for clinical densitometry 2007 adult and pediatric official positions. Bone 2008;43:1115–21. doi:10.1016/j.bone.2008.08.106
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Gordon CM,
      2. Leonard MB,
      3. Zemel BS
      . 2013 Pediatric position development conference: executive summary and reflections. J Clin Densitom 2014;17:219–24. doi:10.1016/j.jocd.2014.01.007
      OpenUrlCrossRefPubMed
    22. ↵
      1. Barrack MT,
      2. Rauh MJ,
      3. Nichols JF
      . Cross-sectional evidence of suppressed bone mineral accrual among female adolescent runners. J Bone Miner Res 2010;25:1850–7. doi:10.1002/jbmr.63
      OpenUrlCrossRefPubMed
    23. ↵
      1. Mudd LM,
      2. Fornetti W,
      3. Pivarnik JM
      . Bone mineral density in collegiate female
      athletes: comparisons among sports      . J Athl Train 2007;42:403–8.
      OpenUrlPubMed
    24. ↵
      1. Kemmler W,
      2. Engelke K,
      3. Baumann H, et al
      . Bone status in elite male runners. Eur J Appl Physiol 2006;96:78–85. doi:10.1007/s00421-005-0060-1
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Roberts AC,
      2. McClure RD,
      3. Weiner RI, et al
      . Overtraining affects male reproductive status. Fertil Steril 1993;60:686–92. doi:10.1016/S0015-0282(16)56223-2
      OpenUrlPubMedWeb of Science
    26. ↵
      1. Wheeler GD,
      2. Singh M,
      3. Pierce WD, et al
      . Endurance training decreases serum testosterone levels in men without change in luteinizing hormone pulsatile release. J Clin Endocrinol Metab 1991;72:422–5. doi:10.1210/jcem-72-2-422
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Ihle R,
      2. Loucks AB
      . Dose-response relationships between energy availability and bone turnover in young exercising women. J Bone Miner Res 2004;19:1231–40. doi:10.1359/JBMR.040410
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Tenforde A,
      2. Barrack MT,
      3. Nattiv A, et al
      . Parallels with the female athlete triad in male athletes. Sports Med 2016;46:171–82. doi:10.1007/s40279-015-0411-y
      OpenUrl
    29. ↵
      1. Mountjoy M,
      2. Sundgot-Borgen J,
      3. Burke L
      , et al. The IOC consensus statement: beyond the female athlete triad—Relative Energy Deficiency in Sport (RED-S). Br J Sports Med 2014;48:491–7. doi:10.1136/bjsports-2014-093502
      OpenUrlAbstract/FREE Full Text
    30. ↵
      1. Thralls KJ,
      2. Nichols JF,
      3. Barrack MT, et al
      . Body mass related predictors of the female athlete triad among adolescent athletes. Int J Sport Nutr Exerc Metab 2016;26:17–25. doi:10.1123/ijsnem.2015-0072
      OpenUrl
    31. ↵
      1. Barrack MT,
      2. Van Loan MD,
      3. Rauh MJ, et al
      . Physiological and behavioral indicators of an energy deficiency among female adolescent runners with elevated bone turnover. Am J Clin Nutr 2010;92:652–9. doi:10.3945/ajcn.2009.28926
      OpenUrlAbstract/FREE Full Text
    32. ↵
      1. Borer KT
      . Physical activity in the prevention and amelioration of osteoporosis in women: interaction of mechanical, hormonal and dietary factors. Sports Med 2005;35:779–830. doi:10.2165/00007256-200535090-00004
      OpenUrlCrossRefPubMedWeb of Science
    View Abstract

    Footnotes

    • Contributors MTB led the manuscript writing, data analysis, participated in study design and project coordinating at the UCLA site. MF served as PI at the Stanford site, AN served as PI at the UCLA site, AT was project coordinator at the Stanford site. AN, MF and AT participated in writing, reviewing and editing the manuscript.

    • Funding The study received financial support from the UCLA Clinical and Translational Science Institute (grant #UL1TR000124) and the 2010 Richard S Materson Education Research Fund New Investigator Research Grant, Stanford Medical Scholars Research Program, Education Research Fund for Physical Medicine and Rehabilitation Medical Student Research Grant (awarded to AST).

    • Competing interests None declared.

    • Ethics approval UCLA IRB, Stanford University IRB.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    Linked Articles