It is currently understood that osteoporosis may occur from accelerated bone loss in adulthood or from inadequate bone mineral accumulation during adolescence, a period during which individuals gain approximately half of their adult bone mineral content (BMC) (1,9,18,30). Longitudinal and cross-sectional studies report a sharp decline in the rate of bone deposition after age 16-18 yr (27,28). Later in life it may not be possible, without pharmaceutical intervention, to duplicate the hormonal milieu that promotes adolescents' bone mineral gains.
An elevated prevalence of low bone mass for age, as well as cross-sectional evidence of suppressed bone mineral accumulation, was recently recognized among a sample of adolescent endurance runners (2,3). Bone mass levels were lower among adolescent runners who exhibited a history of oligomenorrhea or amenorrhea, elevated dietary restraint, lower body mass index (BMI), or lean tissue mass levels and those with the longest history of participation in an endurance running sport (2). Each of these factors may be associated with an energy deficit and is therefore consistent with the results of several controlled laboratory studies that demonstrate a direct negative effect of low energy availability on factors that promote bone formation (12,15,21,23,24).
It remains to be determined whether low bone mass for age among adolescents is irreversible or if bone can undergo "catch-up" mineralization at the end of or after the second decade of life. Several studies among girls with anorexia nervosa (AN) have evaluated this possibility, yet they have yielded inconclusive results (4,14,25). However, some evidence points to the possibility that girls who recover weight and menstrual function, or who have late pubertal onset, may increase bone mass to near normal levels (4,17,20).
To our knowledge, no previous studies have longitudinally evaluated bone mass and variables influencing bone mass change among adolescent runners. Therefore, the current 3-yr follow-up study aimed to: 1) determine whether female adolescent runners with low bone mass for age at baseline exhibit a delayed period of rapid bone mineralization, allowing them to "catch up" with their peers and 2) evaluate factors positively or negatively associated with bone mass change.
METHODS
Participants.
In the summer of 2007, approximately 3 yr after an initial data collection period in the fall of 2004, a sample of 93 female high school endurance runners (2) was recruited to complete follow-up measures. Forty girls (aged 18.7 ± 0.2 yr) agreed to participate. Data collection from one girl was incomplete; therefore, the follow-up sample consisted of 39 runners. Girls who agreed to participate (n = 40) did not differ significantly in baseline characteristics from those who did not participate in the follow-up measures (n = 53). Written parental consent from girls <18 yr and subject assent from all participants were obtained before data collection. The study was approved by the San Diego State University's Institutional Review Board.
Data collection.
At the follow-up assessment, all baseline measures were repeated (2). Girls completed a questionnaire derived from an athletic preparticipation medical history form (29) to evaluate menstrual function, training volume, and sports participation. Participants underwent a dual-energy x-ray absorptiometry (DXA) scan to assess BMC, bone area, and bone mineral density (BMD) of the total hip, total body, lumbar spine, and body composition. Before the DXA scan, height and weight without shoes (to the closest 0.5 inch and 0.5 lb, respectively) were measured.
Menstrual status.
Girls reported information about their age at menarche, the regularity of their cycles, effect of training on their cycle, length between cycles, number of cycles they had during each of the 3 yr between baseline and follow-up, and their use of oral contraceptives (OC). Menstrual irregularity was defined as: primary amenorrhea (no onset of menarche by age 15 yr), secondary amenorrhea (absence of three consecutive menstrual cycles in the past year), or oligomenorrhea (interval between menses >35 d in the past year) (16). A composite variable representing mean menstrual function during the 3 yr was calculated by dividing the total number of cycles girls had between baseline and follow-up by 3. Normal and irregular 3-yr menstrual function was defined as an average of ≥10 or <10 cycles per year, respectively.
BMD.
BMC (g), bone area (cm2), and areal BMD (g·cm−2) at the spine (L1-L4), proximal femur and total body, and body composition were assessed by DXA using a Lunar DPX-NT densitometer (Lunar/GE Corp., Madison, WI). Quality assurance tests were performed each morning of testing. The coefficient of variation in BMD for our laboratory is 0.6% for the total hip, 1.2% for the spine (L1-L4), and 1.0% for the total body. Runners were classified with low bone mass for age if their lumbar spine and/or total body values were 1 or 2 SD or more below the age-matched, gender-specific reference data from the GE/Lunar pediatric database (26). Laboratory-specific values used to indicate a significant BMD increase at the total body, total hip, and lumbar spine were 0.034, 0.017, and 0.04 g·cm−2, respectively.
Statistical analyses.
Independent t-test, ANOVA, and ANCOVA assessed between-group differences among the sample. The Bonferroni correction was used to perform multiple comparison analyses. Repeated-measures ANOVA evaluated within-group baseline to follow-up BMC change. Pearson χ2 tests determined differences in prevalence estimates between groups. Multivariate regression analyses evaluated predictors of BMC change at the total body, total hip, and lumbar spine measurement sites. All values were expressed as mean ± SEM. All analyses were conducted using the Statistical Package for Social Sciences (version 16.0; SPSS, Inc., Chicago, IL).
RESULTS
Whole-sample descriptive characteristics and prevalence estimates.
Runners in the follow-up sample were 18.7 ± 0.2 yr old. Their age at menarche and gynecological age (GA) was 12.8 ± 0.3 and 5.9 ± 0.3 yr, respectively. The sample, as a whole, significantly increased in height (0.9 ± 0.2, P < 0.001), body weight (4.7 ± 0.7 kg, P < 0.001), BMI (1.5 ± 0.2 kg·m−2, P < 0.001), and percent body fat (5.6 ± 0.6 kg, P < 0.001) between the baseline and follow-up analyses. On average, subjects ran 2538.6 ± 255.2 miles between assessments. During the months that girls trained, they ran 28.7 ± 1.7 miles·wk−1. Thirty-nine percent of the girls participated in at least one competitive endurance running season during each of the 3 yr between baseline and follow-up. Five (26% of the 19 girls in college) competed on an intercollegiate cross-country team for at least one season.
None of the runners in our sample reported use of OC at baseline. However, 11 (28.2%) reported OC use during at least 1 yr between baseline and follow-up. The mean length of OC use (among those taking OC) was 1.6 ± 0.2 yr. Baseline or follow-up age, anthropometric variables, training volume, bone mass, or bone mass change was not significantly different between those reporting (n = 11) or not reporting (n = 28) OC use.
The prevalence of menstrual irregularity at the baseline and follow-up assessments was 28.2% and 25.6%, respectively (Fig. 1). At the follow-up, six girls reported using OC to regulate their menses (denoted "OC" in Fig. 1). Five additional girls reported OC use at follow-up but not for regulating menses; these girls had normal menses at baseline and were categorized in the normal-menses group at both time points (Fig. 1). A higher percentage of girls with menstrual irregularity compared with normal menses at baseline had menstrual irregularity or were taking OC to regulate their menstrual function at the follow-up (81.8% vs 25.0%, χ2 = 10.5, P < 0.005). More runners with secondary amenorrhea compared with oligomenorrhea or normal menses at baseline reported secondary amenorrhea for ≥1 yr between assessments (85.7% compared with 50.0% and 14.3%, respectively, Pearson χ2 = 14.2, P < 0.005).
;)
FIGURE 1: Runners' menstrual status at baseline and 3-yr follow-up. ‡Number of runners in each group reporting secondary amenorrhea during one or more of the 3 yr between baseline and follow-up. *Normal BMD at follow-up after low BMD at baseline. **Low BMD at follow-up after normal BMD at baseline. Normal, normal menses; OC, use of OC to regulate menses; PA, primary amenorrhea; SA, secondary amenorrhea.
At baseline, 15 girls were classified with "low" bone mass (BMD z-score ≤ −1.0) and 24 girls were classified with "normal" bone mass (BMD z-score > −1.0). Eighty-seven percent of runners with low bone mass at baseline met criteria for low bone mass at follow-up (Fig. 2). Two girls, both with a 1.10 lumbar spine z-score increase, moved from the low- to the normal-BMD classification (Fig. 2). One runner, with normal BMD at baseline, exhibited a lumbar spine z-score decline of 1.50, causing her to become newly categorized with low BMD (Fig. 2).
FIGURE 2: Classification of runners (n = 39) into normal- or low-BMD groups at baseline and at the 3 yr follow-up.
Bone mass groups.
For majority of the analyses, runners were stratified by baseline bone mass. Runners with low compared with normal bone mass at baseline exhibited a significantly lower body weight, BMI, and percent body fat at the baseline and follow-up assessments; however, baseline to follow-up anthropometric changes did not significantly differ between groups (Table 1). The total number of menses reported during the 3 yr between assessments did not differ significantly between girls with normal baseline bone mass and low baseline bone mass. This was true for the sample as a whole (29.2 ± 2.1 vs 26.3 ± 2.7, P = 0.40, n = 39) and for girls not reporting use of OC (28.8 ± 2.8 vs 26.5 ± 3.4, P = 0.60, n = 28). Runners with low compared with normal baseline bone mass reported participating in significantly fewer months of a non-lean-build impact sport at both time points (Table 1). Further, they reported higher off-season mileage at baseline (248.1 ± 39.6 vs 147.8 ± 31.3, P < 0.05) and ran significantly more total miles (3165.6 ± 395.8 vs 2146.8 ± 312.9, P < 0.05) and more months (26.1 ± 2.6 vs 19.5 ± 2.0, P < 0.05) between assessments.
TABLE 1: Descriptive characteristics among runners with normal (n = 24) or low (n = 15) baseline bone mass.
As a whole, runners' mean changes in total body, total hip, and lumbar spine BMC were 182.4 ± 20.0, 0.6 ± 0.2, and 3.6 ± 0.4 g, respectively. Both the normal- and low-bone mass groups exhibited significant 3-yr increases in total body and lumbar spine BMC (Fig. 3), and the amount that runners' total body, total hip, and lumbar spine BMC changed between the baseline and follow-up did not significantly differ between groups (Table 1). However, runners with low bone mass at baseline had a significantly lower adjusted total body BMC, total hip BMC, lumbar spine BMC, total body BMD z-score, and lumbar spine BMD z-score at both the initial and the 3-yr follow-up assessments (Table 1 and Fig. 4).
FIGURE 3: Within-group total body (A), total hip (B), and lumbar spine (C) BMC change between the baseline and follow-up assessments among runners with low or normal baseline bone mass. a P < 0.001, b P < 0.05, indicated significant baseline to follow-up differences, repeated-measures ANOVA.
FIGURE 4: Comparisons of adjusted baseline (adjusted for baseline height, weight, and GA) and follow-up (adjusted for follow-up height, weight, and GA) total body and lumbar spine BMD z-score, a P < 0.05 between low- and normal-bone mass groups, ANCOVA.
Variables associated with bone mass change.
In Table 2, runners were stratified by baseline bone mass and according to whether they exhibited a significant 3-yr BMD increase at two or more bone sites ("significant increase" group). Among both the normal- and low-baseline bone mass groups, girls who significantly increased BMD were younger, had a lower GA, and reported a lower off-season running mileage at baseline than those who did not significantly increase their bone mass (Table 2). Runners who exhibited a significant increase in bone mass additionally had a higher 3-yr increase in height, body weight, lean tissue mass, and percent body fat and trended toward running fewer total miles between the baseline and follow-up assessments (Table 2).
TABLE 2: Traits among runners with low or normal bone mass at baseline, who did or did not significantly increase bone mass during the 3 yr between assessments.
In the sample as a whole, runners with GA ≤3 yr compared with >3 yr at baseline exhibited higher adjusted BMC increases at each bone site (Table 3). When grouping the sample into tertiles according to the change in their lean tissue mass, runners in the highest compared with two lower tertiles had significantly higher adjusted increases in total body, total hip, and lumbar spine BMC (Table 3). Further, runners with a GA ≤3 yr at baseline and regular, compared with irregular, menstrual function between baseline and follow-up had larger 3-yr increases in adjusted total hip and lumbar spine BMC (Table 3).
TABLE 3: Variables associated with 3-yr change in BMC among the runners (n = 39).
Multiple linear regression analyses indicated that chronological age at baseline, GA, and the total miles girls ran between baseline and the follow-up were negative predictors of girls' 3-yr BMC change (Table 4). In contrast, the 3-yr difference in lean tissue mass and percent body fat, baseline BMI, and the number of menses girls had between baseline and follow-up positively predicted BMC differences between assessments (Table 4). These variables accounted for 29%-54% of the variability in the 3-yr BMC change (Table 4).
TABLE 4: Multiple linear regression analysis to identify variables that significantly contribute to the prediction of change in BMC at each bone site.
DISCUSSION
To our knowledge, this is the first study to evaluate physiological indicators of body mass, menstrual function, training volume, and bone mass at two time points among a sample of female adolescent endurance runners. We found that approximately 90% of girls classified with low bone mass for chronologic age at baseline also had low bone mass at the 3-yr follow-up. Higher levels of training in an endurance running sport and having an older chronological and GA were negatively associated with the change in runners' BMC. Change in lean tissue mass, BMI, and percent body fat as well as having more normal menses positively related to girls' 3-yr total body, total hip, and/or lumbar spine BMC gains.
As a whole, runners with both normal and low bone mass for age at baseline exhibited significant increases in total body, total hip, and lumbar spine BMC between assessments. These BMC increases were anticipated among this sample, whose mean age at baseline was 15.9 ± 0.2 yr, since, under normal conditions, approximately 10%-15% of adult bone mass is accrued between age 16 yr and early adulthood (18,30). Adjusted total body, total hip, and lumbar spine BMC increases were not significantly different between runners with low compared with those with normal bone mass at baseline. Further, the amount that body weight and BMI increased and the number of menses girls had between baseline and follow-up did not significantly differ between the low- and normal-baseline bone mass groups. Body mass and menstrual function were significantly and positively associated with BMC change in this and previous samples (2,6,7,13). These similar characteristics may explain, at least in part, why girls with low and normal bone mass at baseline did not significantly differ in the amount that their total hip and lumbar spine BMC changed during the 3 yr between assessments.
Although BMC among runners with low bone mass at baseline significantly increased during the 3 yr between assessments, as a whole, their BMC increases did not occur at a rate that allowed them to "catch up" to their peers with normal bone mass. This is true, although runners with low bone mass at baseline exhibited traits known to positively influence bone during the 3 yr between baseline and follow-up. Specifically, body weight among girls in this group significantly increased ∼5 kg, their mean BMI rose from ∼19 to 21 kg·m−2, and their lean tissue mass increased ∼0.4 kg. However, during the 3-yr period between assessments, girls with low bone mass at baseline also exhibited traits known to negatively affect bone. More than 45% had menstrual irregularity or were taking OC to regulate their menses at the follow-up and 40% reported secondary amenorrhea for at least 1 yr between assessments. Thus, factors associated with their menstrual disturbances may have limited their bone mass gains.
Findings from previous follow-up studies among recovered AN patients and young adult endurance runners are consistent with our observations. Recovered AN patients who increased their body weight have been shown to prevent further decreases in their bone mass (25). However, despite a significant postrecovery increase in BMD, recovered AN patients exhibit lower bone mass levels as compared with controls without a history of AN (8,14). In addition, a follow-up study among a sample of young adult endurance runners found that those who previously had chronic amenorrhea compared with those who were eumenorrheic exhibited significantly lower bone mass levels at baseline and after a 6- to 10-yr follow-up (22). These findings suggest that low bone mass that occurs early in life may be irreversible and emphasize the importance of accruing a normal amount of bone mineral during the developmental years.
In contrast to the majority of girls with low bone mass at baseline, two runners exhibited a 1.10-SD lumbar spine BMD increase, changing their BMD classification from low to normal. Both girls had normal menstrual function and increased their BMI between the baseline and follow-up assessments. This observation is consistent with two case reports that documented a near 2-SD increase in lumbar spine BMD z-score among two young adult competitive endurance runners who, during adolescence, exhibited disordered eating, high levels of training, and amenorrhea (17,20). Their bone mass gains occurred after a 6- to 8-yr period of weight gain, normal menses, and reduced training (17,20). These findings suggest that a prolonged increase in energy status may promote significant BMC accrual in some young adults with amenorrhea as adolescents. However, since gaining significant bone mass during the third decade of life is rare, it seems preferable to focus on optimizing adolescent bone mineral gains.
Among girls with both low and normal bone mass at baseline, bone mineral increases were associated with age, developmental stage, and behavioral and physical traits associated with energy status. In girls, the rate of bone mineral accumulation peaks between ages 12 and 14 yr and then declines during the remaining adolescent years (19,23,27,28). Consequently, runners of a younger chronological and GA may have exhibited higher gains as they may have had more bone mass to accrue between assessments compared with the older, more mature runners. Further, runners' changes in bone mass were associated with higher increases in body fat, lean tissue mass, more normal menses, and a lower training volume. These characteristics either indicate or promote an increased energy status and thus support the benefit of consuming adequate calories for optimal bone mineral gains.
Change in lean tissue mass, percent body fat, and BMI as well as the number of menses girls had between baseline and follow-up positively predicted BMC change. The idea that body weight is the largest single factor influencing variability in adult bone mass (18) is consistent with the positive associations we observed between indicators of body mass and bone mass gains. An increase in weight during adolescence is a normal developmental process that occurs concurrently with the onset of menses and other pubertal changes. Adequate weight and body fat increases among otherwise healthy adolescents promote "normal" synthesis and secretion of hormones (i.e., estradiol, insulin-like growth factor 1, and leptin) that regulate both menstrual function and bone metabolism (5,12). Higher lean tissue mass increases may have additionally benefited bone as muscle contractions exert a pull on the bone surface and this strain seems to facilitate processes that increase bone formation (10,11).
Although higher levels of training in an endurance running sport were inversely associated with bone mass in our sample, we do not interpret this to suggest that endurance running had a direct negative effect on bone mass. Instead, the effect of higher training on the possible development of a chronic energy deficit may partly explain the relationship between running mileage and lower BMC gains. In contrast, among runners with normal baseline bone mass, running training volume trended toward being a positive predictor of lumbar spine BMC change. Thus, the effect of running on bone mass may vary depending on behavioral or physiological factors.
At the initial assessment, girls with low bone mass were shorter, lighter, leaner, and had a later age at menarche. Therefore, it is possible that girls with baseline BMD z-scores <1.0 had BMD levels in the bottom 16% (according to a normal population distribution) due to genetic factors (i.e., their behaviors were optimal, but their genetic potential was below normal). Although this may apply to some girls, as a group, those with low compared with normal baseline bone mass had a significantly higher prevalence of oligomenorrhea or amenorrhea at baseline (2). This suggests that the bone mass differences between girls with normal or low bone mass at baseline were not due solely to normal genetic variation but also due, at least in part, to consequences of the known antiosteogenic hormone changes that occur with functional hypothalamic menstrual disturbances.
Our study was limited by its relatively small sample size. Although the design was longitudinal, the method required girls to retrospectively report menstruation and training during the 3 yr between assessments. This may have introduced an additional source of error, although runners in our sample did not report difficulty with the recall of this information. A dietary evaluation would have provided insight into the additional contribution of nutrition on the runners' BMC change over time.
CONCLUSIONS
Approximately 90% of adolescent runners with low bone mass for age at baseline failed to increase their bone mass to normal levels during a 3-yr period, thus emphasizing the importance of optimizing bone mineral accrual during the early to midadolescent years. Our findings suggest that low bone mass at baseline was either largely irreversible or that behaviors between assessments did not allow for potential bone mass gains. Genetic factors may have additionally influenced bone mass differences between groups. Among our sample, a higher increase in lean tissue mass and percent body fat as well as having more menses emerged as positive predictors of runners' increase in BMC. A higher level of endurance running training (particularly among runners with low baseline bone mass) and being of a younger chronological and GA negatively contributed to BMC gains. To further understand factors that affect bone mineral accrual among adolescent runners, future large-scale prospective studies that evaluate associations between exercise participation, dietary intake, pubertal development, and bone mineral change are recommended.
Funded by grants from the National Athletic Trainers Association and the San Diego State University Fred Kasch Endowment.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
REFERENCES
1. Bachrach LK. Acquisition of optimal bone mass in childhood and adolescence.
Trends Endocrinol Metab. 2001;12:22-8.
2. Barrack MT, Rauh MJ, Nichols JF. Prevalence of and traits associated with low BMD among female adolescent runners.
Med Sci Sports Exerc. 2008;40(12):2015-21.
3. Barrack MT, Rauh MJ, Nichols JF. Cross-sectional evidence of suppressed bone mineral accrual among female adolescent runners.
J Bone Miner Res. 2010;25:1850-7.
4. Bass SL, Saxon L, Corral AM, et al. Near normalisation of lumbar spine bone density in young women with osteopenia recovered from adolescent onset anorexia nervosa: a longitudinal study.
J Pediatr Endocrinol Metab. 2005;18:897-907.
5. Bates SH, Dundon TA, Seifert M, et al. LRb-STAT3 signaling is required for the neuroendocrine regulation of energy expenditure by leptin.
Diabetes. 2004;53:3067-73.
6. Beals KA, 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.
7. Bennell KL, Malcolm SA, Khan KM, et al. Bone mass and bone turnover in power athletes, endurance athletes, and controls: a 12-month longitudinal study.
Bone. 1997;20:477-84.
8. Bolton JG, Patel S, Lacey JH, White S. A prospective study of changes in bone turnover and bone density associated with regaining weight in women with anorexia nervosa.
Osteoporos Int. 2005;16:1955-62.
9. 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.
10. Burr DB. Muscle strength, bone mass, and age-related bone loss.
J Bone Miner Res. 1997;12:1547-51.
11. Burr DB, Robling AG, Turner CH. Effects of biomechanical stress on bones in animals.
Bone. 2002;30:781-6.
12. Chan JL, Mantzoros CS. Role of leptin in energy-deprivation states: normal human physiology and clinical implications for hypothalamic amenorrhoea and anorexia nervosa.
Lancet. 2005;366:74-85.
13. Cobb KL, Bachrach LK, Greendale G, et al. Disordered eating, menstrual irregularity, and bone mineral density in female runners.
Med Sci Sports Exerc. 2003;35(5):711-9.
14. Compston JE, McConachie C, Stott C, et al. Changes in bone mineral density, body composition and biochemical markers of bone turnover during weight gain in adolescents with severe anorexia nervosa: a 1-year prospective study.
Osteoporos Int. 2006;17:77-84.
15. Davies JH, Evans BA, Gregory JW. Bone mass acquisition in healthy children.
Arch Dis Child. 2005;90:373-8.
16. Diaz A, Laufer MR, Breech LL. Menstruation in girls and adolescents: using the menstrual cycle as a vital sign.
Pediatrics. 2006;118:2245-50.
17. Fredericson M, Kent K. Normalization of bone density in a previously amenorrheic runner with osteoporosis.
Med Sci Sports Exerc. 2005;37(9):1481-6.
18. Heaney RP, Abrams S, Dawson-Hughes B, et al. Peak bone mass.
Osteoporos Int. 2000;11:985-1009.
19. Heaney RP, Weaver CM. Newer perspectives on calcium nutrition and bone quality.
J Am Coll Nutr. 2005;24:574S-81S.
20. Hind K. Recovery of bone mineral density and fertility in a former amenorrheic athlete.
J Sports Sci Med. 2008;7:415-8.
21. Ihle R, Loucks AB. Dose-response relationships between energy availability and bone turnover in young exercising women.
J Bone Miner Res. 2004;19:1231-40.
22. Keen AD, Drinkwater BL. Irreversible bone loss in former amenorrheic athletes.
Osteoporos Int. 1997;7:311-5.
23. MacKelvie KJ, Khan KM, McKay HA. Is there a critical period for bone response to weight-bearing exercise in children and adolescents? A systematic review.
Br J Sports Med. 2002;36:250-7; discussion 257.
24. Mann DR, Plant TM. Leptin and pubertal development.
Semin Reprod Med. 2002;20:93-102.
25. Misra M, Prabhakaran R, Miller KK, et al. Weight gain and restoration of menses as predictors of bone mineral density change in adolescent girls with anorexia nervosa-1.
J Clin Endocrinol Metab. 2008;93:1231-7.
26. Nattiv A, Loucks AB, Manore MM, et al. American College of Sports Medicine Position Stand: the female athlete triad.
Med Sci Sports Exerc. 2007;39(10):1867-82.
27. Sabatier JP, Guaydier-Souquieres G, Benmalek A, Marcelli C. Evolution of lumbar bone mineral content during adolescence and adulthood: a longitudinal study in 395 healthy females 10-24 years of age and 206 premenopausal women.
Osteoporos Int. 1999;9:476-82.
28. Theintz G, Buchs B, Rizzoli R, et al. Longitudinal monitoring of bone mass accumulation in healthy adolescents: evidence for a marked reduction after 16 years of age at the levels of lumbar spine and femoral neck in female subjects.
J Clin Endocrinol Metab. 1992;75:1060-5.
29. Van de Loo DA, Johnson MD. The young female athlete.
Clin Sports Med. 1995;14:687-707.
30. Weaver CM. Adolescence: the period of dramatic bone growth.
Endocrine. 2002;17:43-8.
