- Email: [email protected]
Maturity and activity-related differences in bone mineral density: Tanner I vs. II and gymnasts vs. non-gymnasts
Bone 39 (2006) 895 – 900 www.elsevier.com/locate/bone
Maturity and activity-related differences in bone mineral density: Tanner I vs. II and gymnasts vs. non-gymnasts☆ Jodi N. Dowthwaite a,⁎, James G. DiStefano a , Robert J. Ploutz-Snyder b , Jill A. Kanaley c , Tamara A. Scerpella a a
Department of Orthopedic Surgery, State University of New York Upstate Medical University, 550 Harrison St., Suite 128, Syracuse, NY 13202, USA b Center for Outcomes Research and Evaluation, Institute for Human Performance, State University of New York Upstate Medical University, 505 Irving Avenue, Syracuse, NY 13202, USA c Department of Exercise Science, Syracuse University, 820 Comstock Avenue, Syracuse, NY 13244, USA Received 5 December 2005; revised 30 March 2006; accepted 12 April 2006 Available online 6 June 2006
Abstract This study tests the hypotheses that (1) Tanner I and Tanner II girls comprise distinct maturational cohorts, exhibiting BMD differences that are not explained by age and body size alone; and (2) within these distinct maturational cohorts, BMD is higher in gymnasts than non-gymnasts, independent of age and body size. Premenarcheal artistic gymnasts (n = 28) and non-gymnasts (n = 28) were evaluated. Fan-beam DXA measured areal BMD (aBMD) at the forearm, femoral neck, and lumbar spine; fat free mass (FFM) was derived from whole-body scans. Height, weight, physical activity and calcium intake were assessed. Group means were compared using ANOVA; ANCOVA was used to adjust for age, height and FFM. For all 3 sites and both maturity levels, gymnasts had higher aBMD than non-gymnasts, independent of age and body size (7.2–20.8%, p b 0.04). After adjustment for age, height and FFM, Tanner II gymnasts demonstrated lower aBMD than Tanner I gymnasts at the femoral neck (7.6%, p b 0.05); no other maturity group comparisons yielded statistically significant differences independent of age and body size. In conclusion, for both Tanner groups, the osteogenic role of impact activity is evident at all three sites. Trends in Tanner group differences in aBMD were specific to gymnast and non-gymnast activity groups and therefore were not generalizable to all subjects. Overall, aBMD correlations and ANCOVA results differ by activity group, maturity level and site. These results highlight the need to consider both maturity and activity status in studies assessing bone accrual. © 2006 Elsevier Inc. All rights reserved. Keywords: Bone density; Gymnasts; Maturation; Physical activity
Introduction Prior studies have demonstrated greater BMD in artistic gymnasts compared to non-gymnasts prior to puberty, during puberty and in early adulthood, attributing the increased BMD to impact loading [1–6]. Because gymnasts tend to exhibit different anthropometric characteristics and patterns of growth and maturation than non-gymnasts [7–10], attempts to isolate the effects of gymnastics activity on bone accrual must account for ☆
Funding sources: NIH/NIAMS (#AR47613) and Orthopedic Research and Education Foundation (#97-015). ⁎ Corresponding author. Fax: +1 315 464 5221. E-mail address: [email protected] (J.N. Dowthwaite). 8756-3282/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2006.04.007
the covariates age, body size and physical maturity. Although some studies have considered these covariates carefully [1–5], others address age and body size differences, but group together subjects that exhibit different maturational stages [11–14]. In particular, we note the use of heterogeneous samples of Tanner I and Tanner II girls in studies that evaluate bone accrual in young gymnasts. The transition from Tanner stage I to II marks the threshold between childhood and adolescence. Tanner stage II heralds the onset of puberty, marked by hormonal increases and associated increases in growth velocity. Specifically, in girls, calciotropic hormones are higher in Tanner II than Tanner I, with associated increases in linear growth velocity [15,16]. In addition, higher estrogen levels in pubertal females may increase rates of endocortical apposition,
896
J.N. Dowthwaite et al. / Bone 39 (2006) 895–900
generating qualitative changes in bone structure [17]. If these stages truly represent distinct cohorts on the basis of growth and maturation, merging of Tanner I and Tanner II girls into a single cohort may confound comparisons between gymnasts and non-gymnasts. Therefore, this study was designed to test the hypotheses that (1) Tanner I and Tanner II girls comprise distinct maturational cohorts, exhibiting BMD differences that are not explained by age and body size alone; and (2) within these distinct maturational cohorts, BMD is higher in gymnasts than non-gymnasts, independent of age and body size. Materials and methods Experimental subjects The 56 subjects in this premenarcheal cohort are a subset of a larger sample (n = 122) of female artistic gymnasts and non-gymnasts who have participated in an ongoing longitudinal study, with subjects aged 7–12 years at study initiation. The institutional review board approved the study, and written informed consent was obtained from participants and their parents prior to enrollment. Gymnasts were recruited from local gymnastics clubs, and non-gymnasts were recruited from local grade schools. Premenarcheal subjects from the larger sample were included in the present analyses if they reported self-assessed Tanner stage I (Tanner breast I and Tanner pubic I) or Tanner stage II (Tanner breast II and Tanner pubic II), at the time of an annual DXA scan. Because data from the present analysis were taken from a longitudinal study for which multiple annual DXA scans were performed, most subjects had numerous data points. In cases where subjects reported the same Tanner status (I or II) for more than one annual DXA scan, the earliest of these measurements was used in these cross-sectional analyses. Subjects were excluded if they were non-Caucasian in origin, as there were insufficient numbers to evaluate racial differences. Gymnasts were excluded if they trained for less than 6 hours per week (h/week) in the 2-year period prior to measurement, based upon significant results from a previous study [6]. On this basis, 28 gymnasts and 28 non-gymnasts were evaluated.
Study design Subjects were measured and completed questionnaires every 3 to 6 months for the assessment of anthropometry, body composition, calcium intake and pubertal stage. Weekly hours of organized physical activity (including gymnastics) were recorded by interview at 3- to 6-month intervals; for most gymnasts, gymnastics training was also recorded in daily logs. Annual DXA scans were performed.
Measures Fan-beam DXA scans were performed on the same Hologic QDR 4500W scanner (Hologic Inc., Bedford, MA), using a standardized protocol to obtain areal BMD (aBMD, g/cm2), bone mineral content (BMC, g) and projected area (area, cm2) for the distal third of the non-dominant forearm (FABMD), as well as femoral neck (FNBMD) and lumbar spine vertebrae 2 and 3 (LS23BMD) in the antero-posterior plane. Hologic software (version 9.03D) was used to calculate non-bone, fat free mass (FFM) and percentage of body mass as fat (%BF) from whole body scans. The coefficient of variation for this DXA machine is 1%. The technician was blind to group membership for each subject. Standing height (Ht) was measured without shoes and with hands on hips, using a right angle and wall-mounted meter sticks. Weight (Wt) was determined with an electronic scale (Detecto Scale, Webb City, MO). Pubertal stage was self-assessed using Tanner diagrams of five pubertal stages of breast and pubic hair development. This method has been validated previously for use in this age group [18]. For non-gymnasts, mean hours per week of organized, weight-
bearing activity (excluding swimming) were reported by interview at 3- to 6month intervals; at the same intervals, gymnastics participation was recorded by interview, with most girls providing daily logs. For both groups, annual mean hours per week were calculated for the year prior to DXA measurement. Habitual calcium intake was assessed by semi-quantitative food frequency questionnaires, administered at 3- to 6-month intervals.
Statistical analysis The cohort was divided into subgroups based upon maturational status (Tanner I and Tanner II) and activity status (gymnasts and non-gymnasts). Comparisons were made between gymnasts and non-gymnasts within each maturational stage (activity group comparisons) and between Tanner I and II girls within each activity group (maturity group comparisons). Analyses of variance (ANOVA) were used to test for significant differences between groups in age, body size and physical activity, setting two-tailed alpha to 0.05. The inter-relatedness of age, body size, physical activity and BMD variables was assessed using correlation analysis within each subgroup: Tanner I, Tanner II, gymnasts and non-gymnasts. Calcium intake and hours per week of physical activity were also assessed for correlation with aBMD; for significant results, partial correlations were used to adjust for age, height and FFM. Age, height and FFM were also used as covariates in ANCOVA, with simultaneous entry. We report means, standard deviations, significance levels (p values) and effect sizes (Cohen's f ). Cohen's conventions for effect size are f ≥ 0.1, 0.3 and 0.5 signifying small, medium and large effects, respectively [19].
Power calculation This study is a retrospective, cross-sectional analysis of data collected during a larger, ongoing study; therefore, the number of available subjects dictated sample size (n = 56). Projected power to detect group differences between gymnasts and non-gymnasts was based on a prior analysis of a similar cohort, in which the effect size for activity group was f = 0.83 [6]. Expecting similar differences between groups of gymnasts and non-gymnasts, power for the current study exceeds 80%. We are aware of no studies that specifically evaluate differences in mean BMD between Tanner I and II girls. Therefore, we relied on Cohen's conventions for small, medium and large effects to estimate power for the comparisons between maturity groups. Based upon the smallest subgroup size for each comparison, power to detect a significant Cohen's large effect ( f = 0.40) was 40–47%. Accordingly, we report effect sizes for all aBMD comparisons to
Table 1a Subject characteristics, independent variables Tanner I
Tanner II
Non-gymnasts Gymnasts (n = 10) (n = 12)
Non-gymnasts Gymnasts (n = 18) (n = 16)
Age (years) 10.4 ± 0.9 Ht (cm) 137.7 ± 8.4 Wt (kg) 31.7 ± 3.8 FFM (kg) 22.9 ± 2.9 % Body mass 22.2 ± 5.5 b as fat Calcium 1019.3 ± intake (mg) 308.9 Gymnastics N/A (h/week) Non-gym activity 4.6 ± 5.5 (h/week)
10.0 ± 1.0 134.8 ± 5.2 29.7 ± 3.2 23.1 ± 2.3 18.1 ± 3.0
11.0 ± 0.8 144.7 ± 6.1 a 38.8 ± 5.9 a 27.6 ± 3.4 a 24.8 ± 6.5
11.4 ± 0.9 a 141.9 ± 5.5 a 35.4 ± 4.2 a 26.9 ± 2.9 a 19.3 ± 4.7
990.5 ± 1068.1 ± 352.7 455.7 10.3 ± 2.4 N/A
970.19 ± 532.8 14.7 ± 5.3 a
N/A
N/A
3.0 ± 2.1
Unadjusted means and standard deviations are shown for each subgroup. Ht = height; Wt = weight; FFM = DXA fat-free mass; Non-gym activity = organized weight-bearing activity. a Indicates higher group mean for Tanner I vs. II, ANOVA p b 0.05. b Indicates higher group mean for gymnasts vs. non-gymnasts, ANOVA p b 0.05.
J.N. Dowthwaite et al. / Bone 39 (2006) 895–900
897
Table 1b Subject characteristics, unadjusted bone variables Tanner I
Tanner II
Non-gymnasts Gymnasts (n = 10) (n = 12)
Non-gymnasts Gymnasts (n = 18) (n = 16)
Forearm measures Areal BMD 0.372 ± 0.025 (g/cm2) BMC (g) 4.77 ± 0.60 Area (cm2) 12.79 ± 0.97 Lumbar spine measures Areal BMD 0.595 ± 0.056 (g/cm2) BMC (g) 10.27 ± 1.70 Area (cm2) 17.23 ± 2.11 Femoral neck measures Areal BMD 0.642 ± 0.075 (g/cm2) BMC (g) 2.44 ± 0.26 Area (cm2) 3.80 ± 0.28
0.414 ± 0.032 a 0.394 ± 0.029
0.434 ± 0.037 a
5.82 ± 0.81a 14.00 ± 1.33 a
7.17 ± 1.14 a, b 16.44 ± 1.50 a
5.92 ± 0.79 b 15.02 ± 1.45 b
0.693 ± 0.070 a 0.684 ± 0.087
0.746 ± 0.097
12.97 ± 2.02 a 18.67 ± 1.70
14.82 ± 3.23 19.67 ± 2.21
14.08 ± 3.06 b 20.41 ± 2.46 b
0.708 ± 0.042 a 0.682 ± 0.058 b 0.748 ± 0.095 a 2.65 ± 0.32 3.74 ± 0.37
2.84 ± 0.40 b 4.16 ± 0.34 b
2.98 ± 0.46 b 3.98 ± 0.31
Unadjusted means and standard deviations are shown for each subgroup. BMC = bone mineral content; Area = projected area. a Indicates higher group mean for gymnasts vs. non-gymnasts, ANOVA p b 0.05. b Indicates higher group mean for Tanner I vs. II, ANOVA p b 0.05. supplement significance levels. We also note overlap in 95% confidence intervals for statistically significant results.
Results Subject characteristics are presented in Tables 1a and 1b. As determined by correlation analysis, within most subgroups, age Table 2 Areal BMD correlations Forearm
Fig. 1. Adjusted group means and standard deviations are depicted for aBMD comparisons between gymnasts and non-gymnasts within Tanner I and II subgroups. All means are adjusted for age, height and FFM by ANCOVA. *Indicates significantly higher aBMD for gymnasts vs. non-gymnasts, p b 0.05.
Femoral neck Lumbar spine
Gymnasts (n = 28) Age 0.60*** 0.58*** Ht 0.66**** 0.50** FFM 0.60*** 0.63*** Gym h/week 0.40* 0.46* Ca intake 0.02 0.43* Non-gymnasts Age 0.27 0.14 (n = 28) Ht 0.25 0.26 FFM 0.26 0.42* PAL hrs/wk 0.06 0 Ca intake − 0.18 0 Tanner I (n = 22) Age 0.28 − 0.14 Ht 0.21 − 0.02 FFM 0.17 0.08 Ca intake 0.17 0.13 Tanner II (n = 34) Age 0.38 0.56** Ht 0.12 0.25 FFM 0.24 0.50** Ca intake − 0.21 0.21
0.66**** 0.73**** 0.67**** 0.56** −0.02 0.26 0.43* 0.60*** −0.04 0.15 0.12 0.18 0.21 −0.11 0.46** 0.38* 0.55** 0.06
Correlation coefficients (R) for areal BMD vs. age, body size, physical activity and calcium intake are listed for each subgroup. Statistical significance is indicated as follows: *p b 0.05, **p b 0.01, ***p b 0.001, ****p b 0.0001. Ht = standing height; FFM = DXA fat-free mass; Gym h/week = hours per week of gymnastic participation; PAL h/week = hours per week of organized physical activity; Ca intake = calcium intake.
and body size exhibited significant, positive correlations with aBMD, associating higher aBMD with higher age and larger body size (Table 2). In addition, for the gymnasts, there were significant, positive correlations between h/week of gymnastics and aBMD. However, after adjustment for age, height and FFM by partial correlation, only lumbar spine remained positively correlated with gymnastic hours per week (R = 0.34, p b 0.05). In contrast, hours per week of organized physical activity were not significantly correlated with aBMD for the non-gymnasts. Calcium intakes were only significantly correlated with aBMD for gymnasts at the femoral neck (R = 0.43, p b 0.05), and no significant positive correlation remained after adjustment for age, height and FFM. There were no statistically significant differences in calcium intakes between maturity or activity groups (Table 1a). Activity group comparisons Gymnasts and non-gymnasts did not differ significantly in age, height, weight or FFM within either maturity group, although non-gymnasts exhibited higher mean %BF in the Tanner I cohort (Table 1a). In the Tanner I group, aBMD was not correlated with
898
J.N. Dowthwaite et al. / Bone 39 (2006) 895–900
effect). At the lumbar spine, mean aBMD adjusted for age, height and FFM also tended to be greater for Tanner II than for Tanner I non-gymnasts (p = 0.26, f = 0.24, small effect); adjusted mean FNBMD did not differ between maturity groups (p = 0.73, f = 0.07, negligible effect size). In contrast, for gymnasts, adjusted means were greater for Tanner I than for Tanner II girls at all sites, but only femoral neck aBMD differences achieved statistical significance (p = 0.048–0.17, f = 0.30 to 0.44, small to medium effects; all 95% confidence intervals overlapped) (Fig. 2). Discussion
Fig. 2. Adjusted group means and standard deviations are depicted for aBMD comparisons between Tanner I and II girls within gymnast and non-gymnast subgroups. All means are adjusted for age, height and FFM by ANCOVA. *Indicates significantly higher aBMD for Tanner I vs. Tanner II, p b 0.05.
age or body size. After adjustment for age, height and FFM by ANCOVA, mean aBMD at all three sites was greater for gymnasts than for non-gymnasts within both maturity groups (for Tanner II LSBMD and FNBMD, p b 0.04, f N 0.40, medium effect with overlapping 95% confidence intervals; for all other comparisons, p b 0.03, f = 0.58 to 1.08, large effects; 95% confidence intervals overlapped for Tanner I FNBMD) (Fig. 1). Maturity group comparisons There were differences between maturity groups for height, weight and FFM within both gymnast and non-gymnast subgroups (p b 0.05). Age was higher in Tanner II gymnasts than Tanner I gymnasts (p b 0.001), with a similar trend for nongymnasts (p = 0.057, f = 0.39, medium effect) (Table 1a). Overall, early pubertal girls were older, taller and heavier than their prepubertal counterparts. For non-gymnasts, aBMD differences between Tanner I and II girls varied by site (Fig. 2). After adjustment for age, height and FFM, mean forearm aBMD demonstrated a trend for greater values in Tanner II than in Tanner I non-gymnasts, but this did not achieve statistical significance (p = 0.16, f = 0.40, medium
At both maturity levels, gymnasts exhibited higher aBMD at all sites than non-gymnasts (7.2–20.8%), independent of age and body size. Thus, in this stringent analysis, accounting for influential covariates, including maturational status, there is strong evidence for the osteogenic role of impact activity in the growing skeleton. Results for maturity group comparisons were less conclusive. Marking the initiation of puberty, Tanner stage II is associated with a changing hormonal milieu, accelerated linear growth, increased markers of bone turnover, and higher rates of endocortical apposition [15–17]. On this basis, one would expect to find differences in aBMD between Tanner stage I and II subjects. However, after adjustment for age and body size, the present study did not demonstrate significantly greater mean aBMD for Tanner II vs. Tanner I cohorts (gymnasts or nongymnasts) at any site. Nonetheless, in non-gymnasts, a trend for greater aBMD in Tanner II than Tanner I girls was identified at the lumbar spine and forearm (6.4% and 5.4%, respectively). It is important to note that, despite statistical adjustment for age and body size, this trend for greater aBMD in Tanner II vs. Tanner I non-gymnasts may be attributed to larger bone size, rather than (or in addition to) higher vBMD. In gymnasts, after adjustment for age and body size, aBMD was lower in Tanner II than Tanner I girls (4.6–7.6%), despite greater gymnastic participation in the Tanner II subgroup (nearly 50% higher hours per week). Asynchrony between linear growth and bone mass accumulation may partially explain this finding [16]. Other investigators have provided evidence of a transient phase of reduced BMD, coincident with high linear growth velocity [20] and Tanner stages II and III [16]. In the present study, Tanner I gymnasts demonstrate disproportionately high aBMD for their age and body size; this aBMD elevation is attributed to impact activity. We speculate that the comparatively lower adjusted aBMD seen in Tanner II gymnasts is a result of the asynchrony noted above; with limited mineral resources, bone length has increased without a commensurate increase in bone density. Thus, for the Tanner II gymnasts, it appears that acceleration of growth in bone length is prioritized over preservation of disproportionately high aBMD. It is possible that higher calcium intakes among Tanner II gymnasts might have allowed greater mineral deposition and higher body size-adjusted aBMD in this pubertal subgroup. Nonetheless, despite the results of these maturity group comparisons, it important to note that both Tanner I and Tanner II gymnasts have greater adjusted aBMD at all sites relative to their non-gymnast counterparts.
J.N. Dowthwaite et al. / Bone 39 (2006) 895–900
In non-gymnasts, asynchrony between linear growth and bone density accumulation was less apparent, with trends toward greater relative aBMD in the Tanner II cohort for both FABMD and LSBMD. We speculate that because aBMD in Tanner I nongymnasts was not disproportionately elevated, acceleration of linear growth could occur in the Tanner II cohort without compromising relative aBMD. In Tanner II non-gymnasts, bone mineral resources may have been sufficient to fuel both linear growth and proportional aBMD increases. Variation in bone geometry may explain some of the differences observed between maturity and activity groups. Bone width, cortical thickness and bone shape change during normal growth and may be modulated by physical activity, possibly in a dose-dependent manner [21–23]. Because three-dimensional morphology is assessed as two-dimensional aBMD, fan beam DXA cannot distinguish between differences in density and geometry. Thus, a region of high bone width but moderate volumetric BMD (vBMD) may appear equivalent to a region of high vBMD but moderate bone width. In the present study, it is likely that Tanner II girls have wider bones than Tanner I girls and possible that gymnasts have wider bones for their body size than non-gymnasts; higher aBMD in these groups may not necessarily indicate greater vBMD. To address this issue, bone mineral apparent density (BMAD) can be calculated using DXA-derived bone width and aBMD [24]. Although this procedure is acceptable for pencil-beam DXA, when fan-beam DXA is used, magnification error affects area and BMC measurement more adversely than it affects aBMD, precluding calculation of BMAD [25,26]. Thus, in this study, aBMD provides a more accurate outcome measure than BMC, area, or BMAD. As used here, aBMD provides a measure of overall bone quantity; material and geometric properties are not isolated from each other. Unadjusted projected area comparisons suggest that at both Tanner stages, high forearm aBMD in gymnasts may be partially due to large bone size. As fan-beam error is likely to be negligible at the forearm, these observed differences require further study, specifically investigating the association between impact/weight-bearing activity and bone geometry. The osteogenic effect of body weight is a well-described determinant of BMD and explains some of the variation in density within and between individuals [27,28]. Fat-free mass correlates even more closely with BMD [29], accounting for up to 80% of variance in BMD in adolescent girls [30,31]. In the present study, FFM was highly correlated with FNBMD and LSBMD, accounting for up to 45% of variation in aBMD in all subgroups except Tanner I. The lack of correlation between FFM and aBMD in Tanner I was unexpected. This result may be due to small sample size, or to minimal intra-subject variation, as age and body size differences were purposefully limited by the study design. Alternatively, hormonal status may be instrumental in the relationship between FFM and aBMD, only yielding a significant correlation in early puberty. Forearm aBMD was not correlated with FFM in non-gymnasts. This is not surprising, as non-gymnasts do not bear weight at this site. For gymnasts, in whom the forearm functions as a weight-bearing and impact-loaded limb, aBMD was strongly correlated with FFM. The association between FFM and aBMD
899
was more pronounced in gymnasts than in non-gymnasts at all three sites. It is likely that high levels of impact loading in the gymnast subgroup enhanced these correlations; greater FFM generates greater force per impact, compounding skeletal stresses. On this basis, ANCOVA adjustment of aBMD for FFM may have diminished the effect of impact activity in Tanner II gymnasts compared to their Tanner I peers, as FFM was higher in the Tanner II subgroup. This association probably contributed to the relative reduction in Tanner II adjusted aBMD. The cross-sectional design of this study is limiting, especially when evaluating differences between Tanner I and II girls. These maturity group comparisons contrast distinct cohorts and do not represent changes through time or growth of specific individuals. Longitudinal studies must be used to elucidate growth processes. In addition, to characterize skeletal responses to the interactions of growth and activity, longitudinal studies that accurately measure both geometric and volumetric properties are necessary. The maturity group comparisons were further limited by small cell sizes and resultant low power. Nonetheless, the use of effect sizes allowed us to identify aBMD trends, distinguishing between the two maturity groups. Furthermore, we observed a directional difference in these maturity group trends for gymnasts vs. non-gymnasts. Larger numbers are necessary to establish the significance of these findings. In conclusion, for this sample of premenarcheal girls, impact activity was associated with higher mean aBMD at all sites, independent of age and body size at both maturational stages. Trends in Tanner group differences in aBMD were specific to gymnast and non-gymnast activity groups and therefore were not generalizable to all subjects. These results highlight the need to consider both maturity and activity status in studies assessing bone accrual. Further studies are necessary to elucidate the extent to which bone accrual is affected by the interaction of maturation and physical activity. Acknowledgments The authors gratefully acknowledge the effort and commitment of the subjects and their parents, without whom this study would not have been possible. We also acknowledge Susan Hemingway, study coordinator, as well as current and former colleagues: Kay Bruening, Jacqueline Cole, Moira Davenport, Nicole Gero, Michael Mincolla, Christina Morganti and Marjolein van der Meulen. This project was funded by grants from the Orthopaedic Research and Education Foundation and from the National Institute of Arthritis and Musculoskeletal and Skin Diseases. References [1] Bass S, Pearce G, Bradney M, Hendrich E, Delmas PD, Harding A, et al. Exercise before puberty may confer residual benefits in bone density in adulthood: studies in active prepubertal and retired female gymnasts. J Bone Miner Res 1998;13:500–7. [2] Cassell C, Benedict M, Specker B. Bone mineral density in elite 7- to 9-yr-old female gymnasts and swimmers. Med Sci Sports Exerc 1996;28:1243–6. [3] Courteix D, Lespessailles E, Jaffre C, Obert P, Benhamou CL. Bone mineral acquisition and somatic development in highly trained girl gymnasts. Acta Paediatr 1999;88:803–8.
900
J.N. Dowthwaite et al. / Bone 39 (2006) 895–900
[4] Courteix D, Lespessailles E, Peres SL, Obert P, Benhamou CL. Effect of physical training on bone mineral density in prepubertal girls: a comparative study between impact-loading and non-impact-loading sports. Osteoporos Int 1998;8:152–8. [5] Lehtonen-Veromaa M, Mottonen T, Nuotio I, Heinonen OJ, Viikari J. Influence of physical activity on ultrasound and dual-energy X-ray absorptiometry bone measurements in peripubertal girls: a cross-sectional study. Calcif Tissue Int 2000;66:248–54. [6] Scerpella TA, Davenport M, Morganti CM, Kanaley JA, Johnson LM. Dose related association of impact activity and bone mineral density in prepubertal girls. Calcif Tissue Int 2003;72:24–31. [7] Bass S, Bradney M, Pearce G, Hendrich E, Inge K, Stuckey S, et al. Short stature and delayed puberty in gymnasts: influence of selection bias on leg length and the duration of training on trunk length. J Pediatr 2000;136:149–55. [8] Claessens AL, Malina RM, Lefevre J, Beunen G, Stijnen V, Maes H, et al. Growth and menarcheal status of elite female gymnasts. Med Sci Sports Exerc 1992;24:755–63. [9] Georgopoulos NA, Markou KB, Theodoropoulou A, Bernadot D, Leglise M, Vagenakis AG. Growth retardation in artistic compared with rhythmic elite female gymnasts. J Clin Endocrinol Metab 2002;87:3169–73. [10] Markou KB, Mylonas P, Theodoropoulou A, Kontogiannis A, Leglise M, Vagenakis AG, et al. The influence of intensive physical exercise on bone acquisition in adolescent elite female and male artistic gymnasts. J Clin Endocrinol Metab 2004;89:4383–7. [11] Dyson K, Blimkie CJR, Davison KS, Webber CE, Adachi JD. Gymnastic training and bone density in pre-adolescent females. Med Sci Sports Exerc 1997;29:443–50. [12] Falk B, Bronshtein Z, Zigel L, Constantini NW, Eliakim A. Quantitative ultrasound of the tibia and radius in prepubertal and early-pubertal female athletes. Arch Pediatr Adolesc Med 2003;157:139–43. [13] Laing EM, Massoni JA, Nickols-Richardson SM, Modlesky CM, O'Connor PJ, Lewis RD. A prospective study of bone mass and body composition in female adolescent gymnasts. J Pediatr 2002;141:211–6. [14] Nickols-Richardson SM, Modlesky CM, O'Connor PJ, Lewis RD. Premenarcheal gymnasts possess higher bone mineral density than controls. Med Sci Sports Exerc 2000;32:63–9. [15] Maalouf J, Mahfoud Z, Arabi A, Nabulsi M, El-Hajj Fuleihan G. Calciotropic hormones, bone and mineral metabolism across puberty. Bone 2005;36:S59–60. [16] Fournier P-E, Rizzoli R, Slosman D-O, Theintz G, Bonjour J-P. Asynchrony between the rates of standing height gain and bone mass accumulation during puberty. Osteoporos Int 1997;7:525–32.
[17] Schoenau E, Neu CM, Rauch F, Manz F. The development of bone strength at the proximal radius during childhood and adolescence. J Clin Endocrinol Metab 2001;86:613–8. [18] Duke PM, Litt JF, Gross RT. Adolescents' self-assessment of sexual maturation. Pediatrics 1980;66:918–20. [19] Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988. p. 532. [20] McKay HA, Bailey DA, Mirwald RL, Davison KS, Faulkner RA. Peak bone mineral accrual and age at menarche in adolescent girls: a 6-year longitudinal study. J Pediatr 1998;133:682–7. [21] Binkley T, Specker B. Increased periosteal circumference remains present 12 months after an exercise intervention in preschool children. Bone 2004;35:1383–8. [22] Wang QJ, Suominen H, Nicholson PHF. Influence of physical activity and maturation status on bone mass and geometry in early pubertal girls. Scand J Med Sci Sports 2005;15:100–6. [23] Duncan CS, Blimkie CJR, Kemp A, Higgs W, Cowell CT, Woodhead H, et al. Mid-femur geometry and biomechanical properties in 15- to 18-yr-old female athletes. Med Sci Sports Exer 2002;34:673–81. [24] Carter DR, Bouxein ML, Marcus R. New approaches for interpreting projected bone densitometry data. J Bone Miner Res 1992;7:137–45. [25] Pocock NA, Noakes KA, Majerkovic Y, Griffiths MR. Magnification error of femoral geometry using fan beam densitometers. Calcif Tissue Int 1997;60: 8–10. [26] Cole JH, Scerpella TA, van der Meulen MCH. Fan-beam densitometry of the growing skeleton: are we measuring what we think we are? J Clin Densitom 2005;8:57–64. [27] Rubin LA, Hawker GA, Peltekova VD, Fielding LJ, Ridout R, Cole DE. Determinants of peak bone mass: clinical and genetic analyses in a young female Canadian cohort. J Bone Miner Res 1999;14:633–43. [28] Welten DC, Kemper HC, Post GB, Van Mechelen W, Twisk J, Lips P, et al. Weight-bearing activity during youth is a more important factor for peak bone mass than calcium intake. J Bone Miner Res 1994;9:1089–96. [29] Cadogan J, Blumsohn A, Barker ME, Eastell R. A longitudinal study of bone gain in pubertal girls: anthropometric and biochemical correlates. J Bone Miner Res 1998;13:1602–12. [30] Witzke K, Snow CM. Effects of plyometric jump training on bone mass in adolescent girls. Med Sci Sports Exerc 2000;32:1051–7. [31] Faulkner RA, Bailey DA, Drinkwater DT, Wilkinson AA, Houston CS, McKay HA. Regional and total body bone mineral content, bone mineral density, and total body tissue composition in children 8–16 years of age. Calcif Tissue Int 2003;53:7–12.
Maturity and activity-related differences in bone mineral density: Tanner I vs. II and gymnasts vs. non-gymnasts☆ Jodi N. Dowthwaite a,⁎, James G. DiStefano a , Robert J. Ploutz-Snyder b , Jill A. Kanaley c , Tamara A. Scerpella a a
Department of Orthopedic Surgery, State University of New York Upstate Medical University, 550 Harrison St., Suite 128, Syracuse, NY 13202, USA b Center for Outcomes Research and Evaluation, Institute for Human Performance, State University of New York Upstate Medical University, 505 Irving Avenue, Syracuse, NY 13202, USA c Department of Exercise Science, Syracuse University, 820 Comstock Avenue, Syracuse, NY 13244, USA Received 5 December 2005; revised 30 March 2006; accepted 12 April 2006 Available online 6 June 2006
Abstract This study tests the hypotheses that (1) Tanner I and Tanner II girls comprise distinct maturational cohorts, exhibiting BMD differences that are not explained by age and body size alone; and (2) within these distinct maturational cohorts, BMD is higher in gymnasts than non-gymnasts, independent of age and body size. Premenarcheal artistic gymnasts (n = 28) and non-gymnasts (n = 28) were evaluated. Fan-beam DXA measured areal BMD (aBMD) at the forearm, femoral neck, and lumbar spine; fat free mass (FFM) was derived from whole-body scans. Height, weight, physical activity and calcium intake were assessed. Group means were compared using ANOVA; ANCOVA was used to adjust for age, height and FFM. For all 3 sites and both maturity levels, gymnasts had higher aBMD than non-gymnasts, independent of age and body size (7.2–20.8%, p b 0.04). After adjustment for age, height and FFM, Tanner II gymnasts demonstrated lower aBMD than Tanner I gymnasts at the femoral neck (7.6%, p b 0.05); no other maturity group comparisons yielded statistically significant differences independent of age and body size. In conclusion, for both Tanner groups, the osteogenic role of impact activity is evident at all three sites. Trends in Tanner group differences in aBMD were specific to gymnast and non-gymnast activity groups and therefore were not generalizable to all subjects. Overall, aBMD correlations and ANCOVA results differ by activity group, maturity level and site. These results highlight the need to consider both maturity and activity status in studies assessing bone accrual. © 2006 Elsevier Inc. All rights reserved. Keywords: Bone density; Gymnasts; Maturation; Physical activity
Introduction Prior studies have demonstrated greater BMD in artistic gymnasts compared to non-gymnasts prior to puberty, during puberty and in early adulthood, attributing the increased BMD to impact loading [1–6]. Because gymnasts tend to exhibit different anthropometric characteristics and patterns of growth and maturation than non-gymnasts [7–10], attempts to isolate the effects of gymnastics activity on bone accrual must account for ☆
Funding sources: NIH/NIAMS (#AR47613) and Orthopedic Research and Education Foundation (#97-015). ⁎ Corresponding author. Fax: +1 315 464 5221. E-mail address: [email protected] (J.N. Dowthwaite). 8756-3282/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.bone.2006.04.007
the covariates age, body size and physical maturity. Although some studies have considered these covariates carefully [1–5], others address age and body size differences, but group together subjects that exhibit different maturational stages [11–14]. In particular, we note the use of heterogeneous samples of Tanner I and Tanner II girls in studies that evaluate bone accrual in young gymnasts. The transition from Tanner stage I to II marks the threshold between childhood and adolescence. Tanner stage II heralds the onset of puberty, marked by hormonal increases and associated increases in growth velocity. Specifically, in girls, calciotropic hormones are higher in Tanner II than Tanner I, with associated increases in linear growth velocity [15,16]. In addition, higher estrogen levels in pubertal females may increase rates of endocortical apposition,
896
J.N. Dowthwaite et al. / Bone 39 (2006) 895–900
generating qualitative changes in bone structure [17]. If these stages truly represent distinct cohorts on the basis of growth and maturation, merging of Tanner I and Tanner II girls into a single cohort may confound comparisons between gymnasts and non-gymnasts. Therefore, this study was designed to test the hypotheses that (1) Tanner I and Tanner II girls comprise distinct maturational cohorts, exhibiting BMD differences that are not explained by age and body size alone; and (2) within these distinct maturational cohorts, BMD is higher in gymnasts than non-gymnasts, independent of age and body size. Materials and methods Experimental subjects The 56 subjects in this premenarcheal cohort are a subset of a larger sample (n = 122) of female artistic gymnasts and non-gymnasts who have participated in an ongoing longitudinal study, with subjects aged 7–12 years at study initiation. The institutional review board approved the study, and written informed consent was obtained from participants and their parents prior to enrollment. Gymnasts were recruited from local gymnastics clubs, and non-gymnasts were recruited from local grade schools. Premenarcheal subjects from the larger sample were included in the present analyses if they reported self-assessed Tanner stage I (Tanner breast I and Tanner pubic I) or Tanner stage II (Tanner breast II and Tanner pubic II), at the time of an annual DXA scan. Because data from the present analysis were taken from a longitudinal study for which multiple annual DXA scans were performed, most subjects had numerous data points. In cases where subjects reported the same Tanner status (I or II) for more than one annual DXA scan, the earliest of these measurements was used in these cross-sectional analyses. Subjects were excluded if they were non-Caucasian in origin, as there were insufficient numbers to evaluate racial differences. Gymnasts were excluded if they trained for less than 6 hours per week (h/week) in the 2-year period prior to measurement, based upon significant results from a previous study [6]. On this basis, 28 gymnasts and 28 non-gymnasts were evaluated.
Study design Subjects were measured and completed questionnaires every 3 to 6 months for the assessment of anthropometry, body composition, calcium intake and pubertal stage. Weekly hours of organized physical activity (including gymnastics) were recorded by interview at 3- to 6-month intervals; for most gymnasts, gymnastics training was also recorded in daily logs. Annual DXA scans were performed.
Measures Fan-beam DXA scans were performed on the same Hologic QDR 4500W scanner (Hologic Inc., Bedford, MA), using a standardized protocol to obtain areal BMD (aBMD, g/cm2), bone mineral content (BMC, g) and projected area (area, cm2) for the distal third of the non-dominant forearm (FABMD), as well as femoral neck (FNBMD) and lumbar spine vertebrae 2 and 3 (LS23BMD) in the antero-posterior plane. Hologic software (version 9.03D) was used to calculate non-bone, fat free mass (FFM) and percentage of body mass as fat (%BF) from whole body scans. The coefficient of variation for this DXA machine is 1%. The technician was blind to group membership for each subject. Standing height (Ht) was measured without shoes and with hands on hips, using a right angle and wall-mounted meter sticks. Weight (Wt) was determined with an electronic scale (Detecto Scale, Webb City, MO). Pubertal stage was self-assessed using Tanner diagrams of five pubertal stages of breast and pubic hair development. This method has been validated previously for use in this age group [18]. For non-gymnasts, mean hours per week of organized, weight-
bearing activity (excluding swimming) were reported by interview at 3- to 6month intervals; at the same intervals, gymnastics participation was recorded by interview, with most girls providing daily logs. For both groups, annual mean hours per week were calculated for the year prior to DXA measurement. Habitual calcium intake was assessed by semi-quantitative food frequency questionnaires, administered at 3- to 6-month intervals.
Statistical analysis The cohort was divided into subgroups based upon maturational status (Tanner I and Tanner II) and activity status (gymnasts and non-gymnasts). Comparisons were made between gymnasts and non-gymnasts within each maturational stage (activity group comparisons) and between Tanner I and II girls within each activity group (maturity group comparisons). Analyses of variance (ANOVA) were used to test for significant differences between groups in age, body size and physical activity, setting two-tailed alpha to 0.05. The inter-relatedness of age, body size, physical activity and BMD variables was assessed using correlation analysis within each subgroup: Tanner I, Tanner II, gymnasts and non-gymnasts. Calcium intake and hours per week of physical activity were also assessed for correlation with aBMD; for significant results, partial correlations were used to adjust for age, height and FFM. Age, height and FFM were also used as covariates in ANCOVA, with simultaneous entry. We report means, standard deviations, significance levels (p values) and effect sizes (Cohen's f ). Cohen's conventions for effect size are f ≥ 0.1, 0.3 and 0.5 signifying small, medium and large effects, respectively [19].
Power calculation This study is a retrospective, cross-sectional analysis of data collected during a larger, ongoing study; therefore, the number of available subjects dictated sample size (n = 56). Projected power to detect group differences between gymnasts and non-gymnasts was based on a prior analysis of a similar cohort, in which the effect size for activity group was f = 0.83 [6]. Expecting similar differences between groups of gymnasts and non-gymnasts, power for the current study exceeds 80%. We are aware of no studies that specifically evaluate differences in mean BMD between Tanner I and II girls. Therefore, we relied on Cohen's conventions for small, medium and large effects to estimate power for the comparisons between maturity groups. Based upon the smallest subgroup size for each comparison, power to detect a significant Cohen's large effect ( f = 0.40) was 40–47%. Accordingly, we report effect sizes for all aBMD comparisons to
Table 1a Subject characteristics, independent variables Tanner I
Tanner II
Non-gymnasts Gymnasts (n = 10) (n = 12)
Non-gymnasts Gymnasts (n = 18) (n = 16)
Age (years) 10.4 ± 0.9 Ht (cm) 137.7 ± 8.4 Wt (kg) 31.7 ± 3.8 FFM (kg) 22.9 ± 2.9 % Body mass 22.2 ± 5.5 b as fat Calcium 1019.3 ± intake (mg) 308.9 Gymnastics N/A (h/week) Non-gym activity 4.6 ± 5.5 (h/week)
10.0 ± 1.0 134.8 ± 5.2 29.7 ± 3.2 23.1 ± 2.3 18.1 ± 3.0
11.0 ± 0.8 144.7 ± 6.1 a 38.8 ± 5.9 a 27.6 ± 3.4 a 24.8 ± 6.5
11.4 ± 0.9 a 141.9 ± 5.5 a 35.4 ± 4.2 a 26.9 ± 2.9 a 19.3 ± 4.7
990.5 ± 1068.1 ± 352.7 455.7 10.3 ± 2.4 N/A
970.19 ± 532.8 14.7 ± 5.3 a
N/A
N/A
3.0 ± 2.1
Unadjusted means and standard deviations are shown for each subgroup. Ht = height; Wt = weight; FFM = DXA fat-free mass; Non-gym activity = organized weight-bearing activity. a Indicates higher group mean for Tanner I vs. II, ANOVA p b 0.05. b Indicates higher group mean for gymnasts vs. non-gymnasts, ANOVA p b 0.05.
J.N. Dowthwaite et al. / Bone 39 (2006) 895–900
897
Table 1b Subject characteristics, unadjusted bone variables Tanner I
Tanner II
Non-gymnasts Gymnasts (n = 10) (n = 12)
Non-gymnasts Gymnasts (n = 18) (n = 16)
Forearm measures Areal BMD 0.372 ± 0.025 (g/cm2) BMC (g) 4.77 ± 0.60 Area (cm2) 12.79 ± 0.97 Lumbar spine measures Areal BMD 0.595 ± 0.056 (g/cm2) BMC (g) 10.27 ± 1.70 Area (cm2) 17.23 ± 2.11 Femoral neck measures Areal BMD 0.642 ± 0.075 (g/cm2) BMC (g) 2.44 ± 0.26 Area (cm2) 3.80 ± 0.28
0.414 ± 0.032 a 0.394 ± 0.029
0.434 ± 0.037 a
5.82 ± 0.81a 14.00 ± 1.33 a
7.17 ± 1.14 a, b 16.44 ± 1.50 a
5.92 ± 0.79 b 15.02 ± 1.45 b
0.693 ± 0.070 a 0.684 ± 0.087
0.746 ± 0.097
12.97 ± 2.02 a 18.67 ± 1.70
14.82 ± 3.23 19.67 ± 2.21
14.08 ± 3.06 b 20.41 ± 2.46 b
0.708 ± 0.042 a 0.682 ± 0.058 b 0.748 ± 0.095 a 2.65 ± 0.32 3.74 ± 0.37
2.84 ± 0.40 b 4.16 ± 0.34 b
2.98 ± 0.46 b 3.98 ± 0.31
Unadjusted means and standard deviations are shown for each subgroup. BMC = bone mineral content; Area = projected area. a Indicates higher group mean for gymnasts vs. non-gymnasts, ANOVA p b 0.05. b Indicates higher group mean for Tanner I vs. II, ANOVA p b 0.05. supplement significance levels. We also note overlap in 95% confidence intervals for statistically significant results.
Results Subject characteristics are presented in Tables 1a and 1b. As determined by correlation analysis, within most subgroups, age Table 2 Areal BMD correlations Forearm
Fig. 1. Adjusted group means and standard deviations are depicted for aBMD comparisons between gymnasts and non-gymnasts within Tanner I and II subgroups. All means are adjusted for age, height and FFM by ANCOVA. *Indicates significantly higher aBMD for gymnasts vs. non-gymnasts, p b 0.05.
Femoral neck Lumbar spine
Gymnasts (n = 28) Age 0.60*** 0.58*** Ht 0.66**** 0.50** FFM 0.60*** 0.63*** Gym h/week 0.40* 0.46* Ca intake 0.02 0.43* Non-gymnasts Age 0.27 0.14 (n = 28) Ht 0.25 0.26 FFM 0.26 0.42* PAL hrs/wk 0.06 0 Ca intake − 0.18 0 Tanner I (n = 22) Age 0.28 − 0.14 Ht 0.21 − 0.02 FFM 0.17 0.08 Ca intake 0.17 0.13 Tanner II (n = 34) Age 0.38 0.56** Ht 0.12 0.25 FFM 0.24 0.50** Ca intake − 0.21 0.21
0.66**** 0.73**** 0.67**** 0.56** −0.02 0.26 0.43* 0.60*** −0.04 0.15 0.12 0.18 0.21 −0.11 0.46** 0.38* 0.55** 0.06
Correlation coefficients (R) for areal BMD vs. age, body size, physical activity and calcium intake are listed for each subgroup. Statistical significance is indicated as follows: *p b 0.05, **p b 0.01, ***p b 0.001, ****p b 0.0001. Ht = standing height; FFM = DXA fat-free mass; Gym h/week = hours per week of gymnastic participation; PAL h/week = hours per week of organized physical activity; Ca intake = calcium intake.
and body size exhibited significant, positive correlations with aBMD, associating higher aBMD with higher age and larger body size (Table 2). In addition, for the gymnasts, there were significant, positive correlations between h/week of gymnastics and aBMD. However, after adjustment for age, height and FFM by partial correlation, only lumbar spine remained positively correlated with gymnastic hours per week (R = 0.34, p b 0.05). In contrast, hours per week of organized physical activity were not significantly correlated with aBMD for the non-gymnasts. Calcium intakes were only significantly correlated with aBMD for gymnasts at the femoral neck (R = 0.43, p b 0.05), and no significant positive correlation remained after adjustment for age, height and FFM. There were no statistically significant differences in calcium intakes between maturity or activity groups (Table 1a). Activity group comparisons Gymnasts and non-gymnasts did not differ significantly in age, height, weight or FFM within either maturity group, although non-gymnasts exhibited higher mean %BF in the Tanner I cohort (Table 1a). In the Tanner I group, aBMD was not correlated with
898
J.N. Dowthwaite et al. / Bone 39 (2006) 895–900
effect). At the lumbar spine, mean aBMD adjusted for age, height and FFM also tended to be greater for Tanner II than for Tanner I non-gymnasts (p = 0.26, f = 0.24, small effect); adjusted mean FNBMD did not differ between maturity groups (p = 0.73, f = 0.07, negligible effect size). In contrast, for gymnasts, adjusted means were greater for Tanner I than for Tanner II girls at all sites, but only femoral neck aBMD differences achieved statistical significance (p = 0.048–0.17, f = 0.30 to 0.44, small to medium effects; all 95% confidence intervals overlapped) (Fig. 2). Discussion
Fig. 2. Adjusted group means and standard deviations are depicted for aBMD comparisons between Tanner I and II girls within gymnast and non-gymnast subgroups. All means are adjusted for age, height and FFM by ANCOVA. *Indicates significantly higher aBMD for Tanner I vs. Tanner II, p b 0.05.
age or body size. After adjustment for age, height and FFM by ANCOVA, mean aBMD at all three sites was greater for gymnasts than for non-gymnasts within both maturity groups (for Tanner II LSBMD and FNBMD, p b 0.04, f N 0.40, medium effect with overlapping 95% confidence intervals; for all other comparisons, p b 0.03, f = 0.58 to 1.08, large effects; 95% confidence intervals overlapped for Tanner I FNBMD) (Fig. 1). Maturity group comparisons There were differences between maturity groups for height, weight and FFM within both gymnast and non-gymnast subgroups (p b 0.05). Age was higher in Tanner II gymnasts than Tanner I gymnasts (p b 0.001), with a similar trend for nongymnasts (p = 0.057, f = 0.39, medium effect) (Table 1a). Overall, early pubertal girls were older, taller and heavier than their prepubertal counterparts. For non-gymnasts, aBMD differences between Tanner I and II girls varied by site (Fig. 2). After adjustment for age, height and FFM, mean forearm aBMD demonstrated a trend for greater values in Tanner II than in Tanner I non-gymnasts, but this did not achieve statistical significance (p = 0.16, f = 0.40, medium
At both maturity levels, gymnasts exhibited higher aBMD at all sites than non-gymnasts (7.2–20.8%), independent of age and body size. Thus, in this stringent analysis, accounting for influential covariates, including maturational status, there is strong evidence for the osteogenic role of impact activity in the growing skeleton. Results for maturity group comparisons were less conclusive. Marking the initiation of puberty, Tanner stage II is associated with a changing hormonal milieu, accelerated linear growth, increased markers of bone turnover, and higher rates of endocortical apposition [15–17]. On this basis, one would expect to find differences in aBMD between Tanner stage I and II subjects. However, after adjustment for age and body size, the present study did not demonstrate significantly greater mean aBMD for Tanner II vs. Tanner I cohorts (gymnasts or nongymnasts) at any site. Nonetheless, in non-gymnasts, a trend for greater aBMD in Tanner II than Tanner I girls was identified at the lumbar spine and forearm (6.4% and 5.4%, respectively). It is important to note that, despite statistical adjustment for age and body size, this trend for greater aBMD in Tanner II vs. Tanner I non-gymnasts may be attributed to larger bone size, rather than (or in addition to) higher vBMD. In gymnasts, after adjustment for age and body size, aBMD was lower in Tanner II than Tanner I girls (4.6–7.6%), despite greater gymnastic participation in the Tanner II subgroup (nearly 50% higher hours per week). Asynchrony between linear growth and bone mass accumulation may partially explain this finding [16]. Other investigators have provided evidence of a transient phase of reduced BMD, coincident with high linear growth velocity [20] and Tanner stages II and III [16]. In the present study, Tanner I gymnasts demonstrate disproportionately high aBMD for their age and body size; this aBMD elevation is attributed to impact activity. We speculate that the comparatively lower adjusted aBMD seen in Tanner II gymnasts is a result of the asynchrony noted above; with limited mineral resources, bone length has increased without a commensurate increase in bone density. Thus, for the Tanner II gymnasts, it appears that acceleration of growth in bone length is prioritized over preservation of disproportionately high aBMD. It is possible that higher calcium intakes among Tanner II gymnasts might have allowed greater mineral deposition and higher body size-adjusted aBMD in this pubertal subgroup. Nonetheless, despite the results of these maturity group comparisons, it important to note that both Tanner I and Tanner II gymnasts have greater adjusted aBMD at all sites relative to their non-gymnast counterparts.
J.N. Dowthwaite et al. / Bone 39 (2006) 895–900
In non-gymnasts, asynchrony between linear growth and bone density accumulation was less apparent, with trends toward greater relative aBMD in the Tanner II cohort for both FABMD and LSBMD. We speculate that because aBMD in Tanner I nongymnasts was not disproportionately elevated, acceleration of linear growth could occur in the Tanner II cohort without compromising relative aBMD. In Tanner II non-gymnasts, bone mineral resources may have been sufficient to fuel both linear growth and proportional aBMD increases. Variation in bone geometry may explain some of the differences observed between maturity and activity groups. Bone width, cortical thickness and bone shape change during normal growth and may be modulated by physical activity, possibly in a dose-dependent manner [21–23]. Because three-dimensional morphology is assessed as two-dimensional aBMD, fan beam DXA cannot distinguish between differences in density and geometry. Thus, a region of high bone width but moderate volumetric BMD (vBMD) may appear equivalent to a region of high vBMD but moderate bone width. In the present study, it is likely that Tanner II girls have wider bones than Tanner I girls and possible that gymnasts have wider bones for their body size than non-gymnasts; higher aBMD in these groups may not necessarily indicate greater vBMD. To address this issue, bone mineral apparent density (BMAD) can be calculated using DXA-derived bone width and aBMD [24]. Although this procedure is acceptable for pencil-beam DXA, when fan-beam DXA is used, magnification error affects area and BMC measurement more adversely than it affects aBMD, precluding calculation of BMAD [25,26]. Thus, in this study, aBMD provides a more accurate outcome measure than BMC, area, or BMAD. As used here, aBMD provides a measure of overall bone quantity; material and geometric properties are not isolated from each other. Unadjusted projected area comparisons suggest that at both Tanner stages, high forearm aBMD in gymnasts may be partially due to large bone size. As fan-beam error is likely to be negligible at the forearm, these observed differences require further study, specifically investigating the association between impact/weight-bearing activity and bone geometry. The osteogenic effect of body weight is a well-described determinant of BMD and explains some of the variation in density within and between individuals [27,28]. Fat-free mass correlates even more closely with BMD [29], accounting for up to 80% of variance in BMD in adolescent girls [30,31]. In the present study, FFM was highly correlated with FNBMD and LSBMD, accounting for up to 45% of variation in aBMD in all subgroups except Tanner I. The lack of correlation between FFM and aBMD in Tanner I was unexpected. This result may be due to small sample size, or to minimal intra-subject variation, as age and body size differences were purposefully limited by the study design. Alternatively, hormonal status may be instrumental in the relationship between FFM and aBMD, only yielding a significant correlation in early puberty. Forearm aBMD was not correlated with FFM in non-gymnasts. This is not surprising, as non-gymnasts do not bear weight at this site. For gymnasts, in whom the forearm functions as a weight-bearing and impact-loaded limb, aBMD was strongly correlated with FFM. The association between FFM and aBMD
899
was more pronounced in gymnasts than in non-gymnasts at all three sites. It is likely that high levels of impact loading in the gymnast subgroup enhanced these correlations; greater FFM generates greater force per impact, compounding skeletal stresses. On this basis, ANCOVA adjustment of aBMD for FFM may have diminished the effect of impact activity in Tanner II gymnasts compared to their Tanner I peers, as FFM was higher in the Tanner II subgroup. This association probably contributed to the relative reduction in Tanner II adjusted aBMD. The cross-sectional design of this study is limiting, especially when evaluating differences between Tanner I and II girls. These maturity group comparisons contrast distinct cohorts and do not represent changes through time or growth of specific individuals. Longitudinal studies must be used to elucidate growth processes. In addition, to characterize skeletal responses to the interactions of growth and activity, longitudinal studies that accurately measure both geometric and volumetric properties are necessary. The maturity group comparisons were further limited by small cell sizes and resultant low power. Nonetheless, the use of effect sizes allowed us to identify aBMD trends, distinguishing between the two maturity groups. Furthermore, we observed a directional difference in these maturity group trends for gymnasts vs. non-gymnasts. Larger numbers are necessary to establish the significance of these findings. In conclusion, for this sample of premenarcheal girls, impact activity was associated with higher mean aBMD at all sites, independent of age and body size at both maturational stages. Trends in Tanner group differences in aBMD were specific to gymnast and non-gymnast activity groups and therefore were not generalizable to all subjects. These results highlight the need to consider both maturity and activity status in studies assessing bone accrual. Further studies are necessary to elucidate the extent to which bone accrual is affected by the interaction of maturation and physical activity. Acknowledgments The authors gratefully acknowledge the effort and commitment of the subjects and their parents, without whom this study would not have been possible. We also acknowledge Susan Hemingway, study coordinator, as well as current and former colleagues: Kay Bruening, Jacqueline Cole, Moira Davenport, Nicole Gero, Michael Mincolla, Christina Morganti and Marjolein van der Meulen. This project was funded by grants from the Orthopaedic Research and Education Foundation and from the National Institute of Arthritis and Musculoskeletal and Skin Diseases. References [1] Bass S, Pearce G, Bradney M, Hendrich E, Delmas PD, Harding A, et al. Exercise before puberty may confer residual benefits in bone density in adulthood: studies in active prepubertal and retired female gymnasts. J Bone Miner Res 1998;13:500–7. [2] Cassell C, Benedict M, Specker B. Bone mineral density in elite 7- to 9-yr-old female gymnasts and swimmers. Med Sci Sports Exerc 1996;28:1243–6. [3] Courteix D, Lespessailles E, Jaffre C, Obert P, Benhamou CL. Bone mineral acquisition and somatic development in highly trained girl gymnasts. Acta Paediatr 1999;88:803–8.
900
J.N. Dowthwaite et al. / Bone 39 (2006) 895–900
[4] Courteix D, Lespessailles E, Peres SL, Obert P, Benhamou CL. Effect of physical training on bone mineral density in prepubertal girls: a comparative study between impact-loading and non-impact-loading sports. Osteoporos Int 1998;8:152–8. [5] Lehtonen-Veromaa M, Mottonen T, Nuotio I, Heinonen OJ, Viikari J. Influence of physical activity on ultrasound and dual-energy X-ray absorptiometry bone measurements in peripubertal girls: a cross-sectional study. Calcif Tissue Int 2000;66:248–54. [6] Scerpella TA, Davenport M, Morganti CM, Kanaley JA, Johnson LM. Dose related association of impact activity and bone mineral density in prepubertal girls. Calcif Tissue Int 2003;72:24–31. [7] Bass S, Bradney M, Pearce G, Hendrich E, Inge K, Stuckey S, et al. Short stature and delayed puberty in gymnasts: influence of selection bias on leg length and the duration of training on trunk length. J Pediatr 2000;136:149–55. [8] Claessens AL, Malina RM, Lefevre J, Beunen G, Stijnen V, Maes H, et al. Growth and menarcheal status of elite female gymnasts. Med Sci Sports Exerc 1992;24:755–63. [9] Georgopoulos NA, Markou KB, Theodoropoulou A, Bernadot D, Leglise M, Vagenakis AG. Growth retardation in artistic compared with rhythmic elite female gymnasts. J Clin Endocrinol Metab 2002;87:3169–73. [10] Markou KB, Mylonas P, Theodoropoulou A, Kontogiannis A, Leglise M, Vagenakis AG, et al. The influence of intensive physical exercise on bone acquisition in adolescent elite female and male artistic gymnasts. J Clin Endocrinol Metab 2004;89:4383–7. [11] Dyson K, Blimkie CJR, Davison KS, Webber CE, Adachi JD. Gymnastic training and bone density in pre-adolescent females. Med Sci Sports Exerc 1997;29:443–50. [12] Falk B, Bronshtein Z, Zigel L, Constantini NW, Eliakim A. Quantitative ultrasound of the tibia and radius in prepubertal and early-pubertal female athletes. Arch Pediatr Adolesc Med 2003;157:139–43. [13] Laing EM, Massoni JA, Nickols-Richardson SM, Modlesky CM, O'Connor PJ, Lewis RD. A prospective study of bone mass and body composition in female adolescent gymnasts. J Pediatr 2002;141:211–6. [14] Nickols-Richardson SM, Modlesky CM, O'Connor PJ, Lewis RD. Premenarcheal gymnasts possess higher bone mineral density than controls. Med Sci Sports Exerc 2000;32:63–9. [15] Maalouf J, Mahfoud Z, Arabi A, Nabulsi M, El-Hajj Fuleihan G. Calciotropic hormones, bone and mineral metabolism across puberty. Bone 2005;36:S59–60. [16] Fournier P-E, Rizzoli R, Slosman D-O, Theintz G, Bonjour J-P. Asynchrony between the rates of standing height gain and bone mass accumulation during puberty. Osteoporos Int 1997;7:525–32.
[17] Schoenau E, Neu CM, Rauch F, Manz F. The development of bone strength at the proximal radius during childhood and adolescence. J Clin Endocrinol Metab 2001;86:613–8. [18] Duke PM, Litt JF, Gross RT. Adolescents' self-assessment of sexual maturation. Pediatrics 1980;66:918–20. [19] Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988. p. 532. [20] McKay HA, Bailey DA, Mirwald RL, Davison KS, Faulkner RA. Peak bone mineral accrual and age at menarche in adolescent girls: a 6-year longitudinal study. J Pediatr 1998;133:682–7. [21] Binkley T, Specker B. Increased periosteal circumference remains present 12 months after an exercise intervention in preschool children. Bone 2004;35:1383–8. [22] Wang QJ, Suominen H, Nicholson PHF. Influence of physical activity and maturation status on bone mass and geometry in early pubertal girls. Scand J Med Sci Sports 2005;15:100–6. [23] Duncan CS, Blimkie CJR, Kemp A, Higgs W, Cowell CT, Woodhead H, et al. Mid-femur geometry and biomechanical properties in 15- to 18-yr-old female athletes. Med Sci Sports Exer 2002;34:673–81. [24] Carter DR, Bouxein ML, Marcus R. New approaches for interpreting projected bone densitometry data. J Bone Miner Res 1992;7:137–45. [25] Pocock NA, Noakes KA, Majerkovic Y, Griffiths MR. Magnification error of femoral geometry using fan beam densitometers. Calcif Tissue Int 1997;60: 8–10. [26] Cole JH, Scerpella TA, van der Meulen MCH. Fan-beam densitometry of the growing skeleton: are we measuring what we think we are? J Clin Densitom 2005;8:57–64. [27] Rubin LA, Hawker GA, Peltekova VD, Fielding LJ, Ridout R, Cole DE. Determinants of peak bone mass: clinical and genetic analyses in a young female Canadian cohort. J Bone Miner Res 1999;14:633–43. [28] Welten DC, Kemper HC, Post GB, Van Mechelen W, Twisk J, Lips P, et al. Weight-bearing activity during youth is a more important factor for peak bone mass than calcium intake. J Bone Miner Res 1994;9:1089–96. [29] Cadogan J, Blumsohn A, Barker ME, Eastell R. A longitudinal study of bone gain in pubertal girls: anthropometric and biochemical correlates. J Bone Miner Res 1998;13:1602–12. [30] Witzke K, Snow CM. Effects of plyometric jump training on bone mass in adolescent girls. Med Sci Sports Exerc 2000;32:1051–7. [31] Faulkner RA, Bailey DA, Drinkwater DT, Wilkinson AA, Houston CS, McKay HA. Regional and total body bone mineral content, bone mineral density, and total body tissue composition in children 8–16 years of age. Calcif Tissue Int 2003;53:7–12.