SEARCH

SEARCH BY CITATION

Keywords:

  • CHILDREN;
  • GYMNASTICS;
  • HIP STRUCTURAL ANALYSIS;
  • LONGITUDINAL;
  • BONE STRUCTURAL STRENGTH

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

Gymnastics, a high-impact weight-bearing physical activity, has been shown to be highly osteogenic. Previously in this cohort, bone mass development (bone mineral content accrual [BMC]) was shown to be positively associated with low-level (recreational) gymnastics exposure (1 to 2 hours per week); however, BMC is only one single component of bone strength. Bone strength is influenced not only by bone mineralization but also bone geometry, bone architecture, and the imposing loads on the bone. The aim of this study was to investigate whether low-level gymnastics training influenced the estimated structural geometry development at the proximal femur. A total of 165 children (92 gymnasts and 73 non-gymnasts) between the ages of 4 and 6 years were recruited into this study and assessed annually for 4 years. During the 4 years, 64 gymnasts withdrew from the sport and were reclassified as ex-gymnasts. A dual-energy X-ray absorptiometry (DXA) image of each child's hip was obtained. Values of cross-sectional area (CSA), section modulus (Z), and cortical thickness (CT) at the narrow neck (NN), intertrochanter (IT), and shaft (S) were estimated using the hip structural analysis (HSA) program. Multilevel random-effects models were constructed and used to develop bone structural strength development trajectories (estimate ± SEE). Once the confounders of body size and lifestyle were controlled, it was found that gymnasts had 6% greater NN CSA than non-gymnasts controls (0.09 ± 0.03 cm2, p < 0.05), 7% greater NN Z (0.04 ± 0.01 cm3, p < 0.05), 5% greater IT CSA (0.11 ± 0.04 cm3, p < 0.05), 6% greater IT Z (0.07 ± 0.03 cm3, p < 0.05), and 3% greater S CSA (0.06 ± 0.03 cm3, p < 0.05). These results suggest that early exposure to low-level gymnastics participation confers benefits related to geometric and bone architecture properties during childhood and, if maintained, may improve bone health in adolescence and adulthood. © 2013 American Society for Bone and Mineral Research.


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

The amount of bone gained during childhood and adolescence impacts greatly on lifetime skeletal health, and it is well accepted that physical activity during growth increases bone acquisition.[1-4] Gymnastics training results in a unique mechanical loading to the skeleton and, therefore, provides an excellent model for assessing the effects of weight-bearing physical activity on bone development.[5] Previous studies in young adolescent competitive female gymnasts have shown that they have 8% to 23% greater dual-energy X-ray absorptiometry (DXA)-derived areal bone mineral density (aBMD, cm/g2) at the total body, lumbar spine, and hip, respectively.[6-9] A recent review of prepubertal gymnastics participation found that gymnasts had denser bones in the lower body compared with non-gymnasts;[10] however, the authors suggested that the DXA-derived aBMD of gymnasts may have underestimated the effect of gymnastics participation on bone strength because the gymnasts within the review displayed a trend toward increased bone mineral content (BMC) and had a smaller bone area than non-gymnasts in the lower body.[10] Although bone strength generally trends in the same direction as aBMD and BMC, this is not always the case; bone strength has been shown to significantly improve with exercise training independent of changes in aBMD.[11] In addition, these parameters are not themselves properties that govern strength. Bone strength is determined by structural dimensions (e.g., bone geometry) and material properties.[12] Thus, bone strength may be better assessed by determining structural strength measures such as cross-sectional area (CSA) and section modulus (Z), which provide information about the structural dimensions of bone.

During the growing years, bone can respond to increased mechanical load by the apposition of bone to the endosteal as well as periosteal surface or by diminished resorption at the endosteal surface.[11, 13] Such geometric adaptations would positively affect bone structural strength. Hip structural analysis (HSA) is a program that allows for the estimation of geometric properties at the hip. Elite prepubertal female gymnasts have been found to have significantly greater strength indices at the hip as assessed by HSA;[9] however, competitive gymnastics is a high-level competitive sport and participation is limited to a select number of skilled individuals. Recreational gymnastics, on the other hand, is attainable by most children and does not require a high level of training. Recreational gymnastics involves the development of spatial and body awareness, muscular strength, and neuromuscular coordination. A paucity of information, however, remains about the effect of recreational-level gymnastics participation and bone structural strength during childhood.

Our group has previously reported that the gymnasts in the present study had greater size-adjusted total body and femoral neck (hip) BMC compared with non-gymnasts (3% and 7%, respectively),[14] although, as previously stated, aBMD and BMC may underestimate the influence of gymnastics participation on bone structural strength. There is also a paucity of research examining the effect of gymnastics participation at both the competitive and recreational level on bone strength parameters in males. Therefore, the aim of the present study was to compare bone structural strength, as assessed through geometric indices, at the hip in young children with a current or past participation history in recreational gymnastics against non-gymnastic controls. We hypothesized that young male and female gymnasts would have greater geometric indices of bone structural strength at the hip compared with children with no past history of gymnastics participation.

Materials and Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

Participants

Participants were drawn from the University of Saskatchewan's Young Recreational Gymnast Study. Details of the Young Recreational Gymnast Study have been published elsewhere.[14] In brief, 178 participants were recruited as part of a mixed longitudinal study. Data were collected annually for 4 years between the fall of 2006 and the spring of 2010. Participants' measures were collected approximately 12 months apart; however, it should be noted that not all participants were measured at every occasion. The gymnast cohort consisted of individuals who were participating in recreational and/or precompetitive gymnastics programs at three competitive gymnastics clubs in Saskatoon, Saskatchewan, and at the University of Saskatchewan's recreational gymnastics program. The gymnasts participated in gymnastic-specific exercises for 45 minutes or more per week for at least one term (4 months) at the study initiation and continued participation throughout the duration of the study. Gymnasts who participated in gymnastics at study initiation (baseline) but did not participate in subsequent follow-up measures were classified as ex-gymnasts. Non-gymnasts were recruited from other local recreational sport programs and had no exposure to a gymnastics stimulus. These individuals were classified as non-gymnastic controls. Participants were excluded from the current study if they had any conditions that prevented them from performing exercise safely or had any conditions known to affect bone development. To be included in the present analysis, a valid proximal femur scan on two or more assessment occasions was required; 13 participants did not fulfill this requirement. This resulted in the inclusion of 165 participants (92 gymnasts [45 males, 47 females] and 73 non-gymnastic controls [36 males, 37 females]) covering the ages of 4 to 10 years. Of the 165 participants, 96% were white, 2% were Asian, and 2% were other, as self-identified by a demographics questionnaire. By the end of 4 years, 64 participants had dropped out of gymnastics (36 males, 28 females) and were redefined as ex-gymnasts. Informed consent was obtained from all parents or guardians and verbal assent was obtained from all children. All procedures and protocols were approved by the University of Saskatchewan's Biomedical Research Ethics Board (Bio 06-111).

Anthropometry

Anthropometric measures included height and weight and were assessed annually following the anthropometric standards outlined by Ross and Marfell-Jones.[15] Height was recorded without shoes to the nearest 0.1 cm using a wall-mounted stadiometer (Holtain Limited, Crymych, UK). Weight was measured on a calibrated digital scale to the nearest 0.5 kg (Model 1631, Tanita Corp., Tokyo, Japan). All measurements were performed twice, and if the difference was greater than 0.4, a third measure was recorded. The mean or median then was reported depending on whether two or three measures were recorded, respectively.[16] All measurements were performed by the same Canadian Society for Exercise Physiology certified exercise physiologist (MCE).

Bone measures

At each measurement occasion, participants underwent a DXA scan of the total body, lumbar spine, and proximal femur following the procedures outlined in Hologic operator's manual (Hologic Discovery 4500, Bedford, MA, USA). Data for DXA-derived BMC for this cohort have been previously published.[14] For the current study, only proximal femur DXA scans were used and all bone measures were derived using the hip structural analysis (HSA) program (Hologic Apex Software Version 3.0). The HSA program has been previously reported elsewhere in greater detail.[17] In brief, the HSA program estimates the structural geometry of the proximal femur from DXA-derived images of the hip. Employing the principles originally reported by Martin and Burr,[18] a line of pixel values traversing the bone axis in a bone mass image is a projection of the corresponding cross section and its dimensions. To determine these cross sections, the HSA algorithms divide the pixel mass values in g/cm2 by the effective mineral density of fully mineralized adult cortical bone. Cross sections are evaluated by averaging geometry over five parallel lines spaced 1 mm apart at three regions of the proximal femur: 1) narrow neck (NN), 2) the intertrochanteric (IT), and 3) the femoral shaft (S).[19] The HSA program locates these regions on the DXA proximal femur bone mineral image and then derives the estimates of structural geometry. From each region, the HSA program produces 10 output variables, of which two were assessed for this study: 1) cross-sectional area (CSA, cm2), the estimated amount of bone surface area in the cross section after excluding all the trabecular and soft tissue space and 2) section modulus (Z, cm3), an indicator of bending strength calculated as the cross-sectional moment of inertia/the maximum distance from the center of mass to outer cortex.[19-21] The short-term precision for CSA and Z derived using a Hologic QDR 4500 hip scan in osteoporotic women ranges from 2.3% to 2.8% and 2.8% to 3.4%, respectively.[22] All HSA analyses were completed by a single technologist (SAJ) and derived from proximal femur scans using a Hologic Discovery 4500.

Physical activity assessments

Physical activity was assessed annually using the Netherlands Physical Activity Questionnaire (NPAQ). Details of the NPAQ as used in this cohort have been described in detail elsewhere.[14] In brief, the NPAQ ranks individuals based on parental reports of their child's current activity preferences and everyday activity choices.[23] The questionnaire response ranges from 7 (low-level physical activity) to 35 (high-level physical activity). The NPAQ has been documented to be a reliable and valid assessment of childhood physical activity levels.[23]

Dietary assessment

Calcium (mg) and vitamin D (IU) intake were assessed annually by a 24-hour food recall questionnaire. Dietary data were analyzed using the Food Processor and Nutritional Software (ESHA Research Software, Salem, OR, USA; Version 8.5). The use of 24-hour food recall has been documented in this cohort previously and been suggested as an appropriate method to assess nutrient intake in children.[24]

Statistical analysis

All variables were assessed for normality, and violations were adjusted using logarithmic transformations. Anthropometrics, dietary intake, and physical activity levels were assessed cross sectionally at each age between gymnasts, ex-gymnasts, and non-gymnastic controls using analyses of variance (ANOVA). All cross-sectional analyses were assessed using SPSS 20 for Windows (Statistical Package for Social Sciences [SPSS], Chicago, IL, USA). For the longitudinal analyses, multilevel (hierarchical) random effects models were constructed using a multilevel modeling approach (MlwiN version 2.25, Multilevel Models Project; Institute of Education, University of London, UK[25]). A detailed description of the multilevel modeling procedures are presented elsewhere.[26] In brief, bone parameters (CSA and Z) were measured repeatedly in individuals (level 1 of the hierarchy) and between individuals (level 2 of the hierarchy). Analysis models that contain variables measured at different levels of a hierarchy are known as multilevel regression models. Specifically, the following additive random effects multilevel regression models were adopted to describe the developmental changes in bone parameters with chronological age.

  • display math

where: y is the bone parameter (CSA or Z) on measurement occasion i in the j-th individual; α is a constant; βjxij is the slope of the bone parameter over time (in this model age is centered around 7 years, the average age of the sample) for the j-th individual; and k1 to kn are the coefficients of various explanatory variables (i.e., age, height, weight, physical activity, calcium intake, vitamin D intake, sex, and gymnastic grouping) at assessment occasion i in the j-th individual. These are the fixed parameters in the model. Both μj and ϵij are random quantities, whose means are equal to zero; they form the random parameters in the model. They are assumed to be uncorrelated and follow a normal distribution and thus their variances can be estimated; μj is the level 2 (between subjects variance) and ϵij the level 1 residual (within individual variance) for the i-th assessment of the bone parameter in the j-th individual. Models were built in a stepwise procedure, i.e., predictor variables (κ-fixed effects) were added one at a time, and the log likelihood ratio statistics was used to judge the effects of including further variables. Predictor variables (κ) were accepted as significant if the estimated mean coefficient was greater than twice the standard error of the estimate (SEE, i.e., p < 0.05). If the retention criteria were not met, the predictor variable was discarded. To allow for the non-linearity of growth, age-centered power functions were introduced into the linear models (e.g., age centered 2). These variables are introduced to shape the developmental curve. The predictor variable coefficients were used to predict CSA and Z development with chronological age, height, weight, physical activity levels, and calcium and vitamin D intake controlled in the prediction equations using population averages at chronological age category for each gymnastic grouping (i.e., gymnasts, ex-gymnasts, non-gymnasts). A total of six independent multilevel (hierarchical) random effects models were constructed for each bone parameter (CSA and Z) at each assessment site of the proximal femur (NN, IT, and S).

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

Participant characteristics

The descriptive characteristics for the gymnasts, ex-gymnasts, and non-gymnastic controls from 4 to 10 years of age are presented in Table 1. Cross-sectional analyses revealed that there were no significant differences at any age category in height, weight, calcium, vitamin D intake, or physical activity score between individuals classified as gymnasts, ex-gymnasts, and non-gymnastic controls, with the exception of ex-gymnasts being significantly taller than gymnasts at 6 years of age (Table 1). Significant differences were observed in the hours of gymnastic training. Naturally, the recreational gymnasts had significantly greater hours of gymnastic training at all age categories compared with both ex-gymnasts and non-gymnast controls (Table 1).

Table 1. Descriptive Characteristics of Chronological Age-Related Anthropometry, Dietary Intake, and Lifestyle Data for Gymnasts, Ex-Gymnasts, and Non-Gymnast Controls (Presented as Means ± Standard Deviation)
 Age category (years)
45678910
  • PA score = physical activity score; Hours = hours of gymnastics training per week.

  • a

    Significant difference from the gymnasts (p < 0.05).

  • b

    Significant difference from the ex-gymnasts (p < 0.05).

Controls
n (males/females)11/1518/2130/3131/3320/2810/181/7
Height (cm)105.07 ± 4.24110.05 ± 4.68116.47 ± 5.33121.49 ± 5.77126.27 ± 5.57133.67 ± 6.89134.89 ± 4.72
Weight (kg)17.52 ± 1.4319.73 ± 2.3921.66 ± 2.8423.67 ± 3.1226.13 ± 3.4530.71 ± 4.7933.97 ± 4.05
Calcium (mg)1406.63 ± 330.481717.99 ± 951.371572.29 ± 406.831613.84 ± 502.371633.71 ± 539.581720.73 ± 455.461462.59 ± 384.38
Vitamin D (UI)194.35 ± 149.60231.66 ± 149.77192.25 ± 129.13187.33 ± 136.02185.26 ± 179.15211.79 ± 129.44202.01 ± 150.53
PA score24.19 ± 4.0324.77 ± 3.5024.75 ± 3.8925.47 ± 3.7324.84 ± 4.1724.77 ± 4.9625.57 ± 2.51
Hours (hours/week)0.00a,b0.00a,b0.00a,b0.00a,b0.00a,b0.00a,b0.00a,b
Gymnasts
n (males/females)4/76/84/154/134/81/60/3
Height (cm)101.05 ± 4.57107.12 ± 4.97113.48 ± 4.09b119.88 ± 5.23124.94 ± 5.12128.93 ± 5.74133.95 ± 5.29
Weight (kg)17.05 ± 3.2218.72 ± 2.7520.91 ± 3.2023.83 ± 4.6126.80 ± 5.4928.93 ± 9.0328.53 ± 3.06
Calcium (mg)1751.82 ± 522.351770.52 ± 498.391523.07 ± 396.471697.50 ± 466.761481.67 ± 381.452064.56 ± 613.561427.73 ± 366.18
Vitamin D (UI)184.32 ± 115.72197.54 ± 109.68203.08 ± 148.15158.65 ± 122.20129.48 ± 95.71115.82 ± 91.40318.29 ± 137.98
PA score24.91 ± 3.1725.69 ± 2.9025.37 ± 2.8725.76 ± 2.7026.00 ± 5.0824.43 ± 3.7826.67 ± 4.16
Hours (hours/week)0.93 ± 0.12b1.50 ± 1.45b3.14 ± 2.35b4.31 ± 3.37b7.65 ± 3.00b7.57 ± 6.59b7.83 ± 5.79b
Ex-gymnasts
n (males/females)1/213/918/1414/2420/2019/109/5
Height (cm)103.84 ± 5.30111.12 ± 6.17117.87 ± 6.20a122.66 ± 6.52129.23 ± 6.77133.41 ± 7.56134.15 ± 7.27
Weight (kg)17.97 ± 2.7020.55 ± 3.5723.26 ± 4.5124.81 ± 4.0628.08 ± 5.2529.92 ± 5.6429.90 ± 4.82
Calcium (mg)1533.98 ± 532.961521.78 ± 432.921585.01 ± 507.451577.84 ± 419.951565.25 ± 463.281925.41 ± 529.121315.37 ± 289.63
Vitamin D (UI)240.64 ± 169.53223.43 ± 138.66227.88 ± 202.60206.33 ± 155.07198.13 ± 153.90197.63 ± 75.34158.70 ± 78.83
PA score24.40 ± 2.8624.88 ± 2.6525.00 ± 3.1825.78 ± 3.4025.97 ± 3.5526.25 ± 2.9326.4 ± 3.2
Hours (hours/week)0.46 ± 0.46a0.57 ± 0.92a0.48 ± 0.98a0.42 ± 1.32a0.33 ± 1.64a0.06 ± 0.25a0.21 ± 0.58a

Longitudinal multilevel modeling results

Tables 2 to 4 summarize the results from multilevel models for the development of CSA and Z for males and females at the NN, IT, and S sites of the proximal femur, respectively. In the multilevel model tables, age centered (centered around 7 years of age) was added as both a fixed and a level 2 random variable. The random effects coefficients describe the two levels of variance (within individuals [level 1 of the hierarchy] and between individuals [level 2 of the hierarchy]). For all models (Tables 2 to 4), the significant variances at level 1 of the models indicate that CSA and/or Z was increasing significantly at each measurement occasion within individuals (E > 2*SEE; p < 0.05). The between-individuals variance matrix (level 2) for each model indicates that individuals had significantly different CSA and/or Z growth curves, both in terms of their intercepts (constant/constant, p < 0.05) and the slopes of their lines (age centered/age centered, p < 0.05). To shape the individual curves, and thus make the models nonlinear, power functions of age centered (age centered 2) were added as fixed effects. Although these power functions were not significant in all models, they were included in all models because they improved the model fit as indicated by log likelihood ratio statistics (data not presented).

Table 2. Multilevel Regression Models for Cross-Sectional Area (CSA) and Section Modulus (Z) at the Narrow Neck (NN) Site
VariableNN CSANNZ
  • Fixed effect values are estimated mean coefficients ± SEE (standard error estimate) of cross-sectional area (CSA; cm2) and sectional modulus (Z; cm3) at the narrow neck (NN). Random effects values are estimated mean variance ± SEE.

  • Age centered is age in years centered around 7 years of age. Sex (male = 0, females = 1). Gymnastic activity: gymnasts compared with nongymnasts (GYM versus CON) and ex-gymnasts compared with non-gymnasts (EX-GYM versus CON).

  • a

    Numerical values multiplied by 10−3.

  • Numerical values are all significant, p < 0.05 (mean > 2*SEE). NS = not significant and variable removed from the final model.

Fixed effects
Constant−0.76 ± 0.24−0.80 ± 0.13
Age centeredNSNS
Age centered2NSNS
Height (cm)13.70 ± 2.40a8.42 ± 1.35a
Weight (kg)24.30 ± 3.30a11.37 ± 1.93a
Vitamin D (UI)NSNS
Calcium (mg)NSNS
Physical activity6.80 ± 2.50a3.80 ± 1.37a
Sex−0.14 ± 0.01−66.53 ± 9.80a
Hours trained per weekNSNS
GYM versus CON95.60 ± 26.00a37.19 ± 14.30a
EX-GYM versus CONNSNS
Random effects
Level 1  
Constant13.90 ± 3.40a4.92 ± 1.18a
Level 2ConstantAge centeredConstantAge centered
Constant15.90 ± 3.80aNS4.35 ± 1.34a0.70 ± 0.30a
Age centeredNSNS0.70 ± 0.30aNS
Table 3. Multilevel Regression Models for Cross-Sectional Area (CSA) and Section Modulus (Z) at the Intertrochanteric (IT) Site
VariableIT CSAIT Z
  • Fixed effect values are estimated mean coefficients ± SEE (standard error estimate) of cross-sectional area (CSA; cm2) and sectional modulus (Z; cm3) at the narrow neck (NN). Random effects values are estimated mean variance ± SEE.

  • Age centered is age in years centered around 7 years of age. Sex (male = 0, females = 1). Gymnastic activity: gymnasts compared with non-gymnasts (GYM versus CON) and ex-gymnasts compared with non-gymnasts (EX-GYM versus CON).

  • a

    Indicated numerical value multiplied by 10−3.

    Numerical values are all significant, p < 0.05 (mean > 2*SEE). NS = not significant and variable removed from the final model.

Fixed effects
ConstantNS−1.30 ± 0.3
Age centered63.18 ± 18.00a36.90 ± 13.88a
Age centered2NS10.18 ± 3.96a
Height (cm)NS12.87 ± 3.03a
Weight (kg)15.96 ± 3.28a33.36 ± 4.42a
Vitamin D (UI)NSNS
Calcium (mg)NSNS
Physical activityNSNS
Sex−52.41 ± 28.33aNS
Hours trained per weekNSNS
GYM versus CON113.95 ± 41.68a72.97 ± 31.92a
EX-GYM versus CONNSNS
Random effects
Level 1  
Constant21.15 ± 5.15a12.47 ± 3.11a
Level 2ConstantAge centeredConstantAge centered
Constant66.41 ± 8.37a8.41 ± 2.43a36.7 ± 4.92a5.78 ± 1.51a
Age centered8.41 ± 2.43aNS5.78 ± 1.51aNS
Table 4. Multilevel Regression Models for Cross-Sectional Area (CSA) and Section Modulus (Z) at the Femoral Shaft (S) Site
VariableS CSAS Z
  • Fixed effect values are estimated mean coefficients ± SEE (standard error estimate) of cross-sectional area (CSA; cm2) and sectional modulus (Z; cm3) at the narrow neck (NN). Random effects values are estimated mean variance ± SEE.

  • Age centered is age in years centered around 7 years of age. Sex (male = 0, females = 1). Gymnastic activity: gymnasts compared with non-gymnasts (GYM versus CON) and ex-gymnasts compared with non-gymnasts (EX-GYM versus CON).

  • a

    Indicated numerical value multiplied by 10−3.

  • Numerical values are all significant, p < 0.05 (mean > 2*SEE). NS = not significant and variable removed from the final model.

Fixed effects
ConstantNS−0.76 ± 0.12
Age centered44.70 ± 10.43a11.07 ± 5.46a
Age centered2NS3.46 ± 1.61a
Height (cm)6.32 ± 2.34a6.97 ± 1.20a
Weight (kg)39.09 ± 3.27a18.68 ± 1.72a
Vitamin D (UI)NSNS
Calcium (mg)NSNS
Physical activityNS2.31 ± 1.19a
Sex−35.16 ± 17.80a−27.38 ± 8.79a
Hours trained per weekNSNS
GYM versus CON54.45 ± 26.40aNS
EX-GYM versus CONNSNS
Random effects
Level 1  
Constant6.49 ± 1.61a13.9 ± 3.4a
Level 2ConstantAge centeredConstantAge centered
Constant21.91 ± 3.51aNS15.9 ± 3.8a0.59 ± 0.02a
Age centeredNSNS0.59 ± 0.02aNS

Narrow neck

Table 2 summarizes the results from the multilevel models for the development of CSA and Z at the NN for males and females. It was observed that height, weight, physical activity, and sex significantly contributed to the prediction of NN CSA and Z. This significant contribution of sex to the prediction models indicates that female participants, regardless of grouping, had significantly less NN CSA (−0.14 ± 0.01 cm2) and NN Z (−0.07 ± 0.01 cm3) than their male counterparts. Once age, height, weight, vitamin D, calcium intake, physical activity, sex, and the hours of gymnastics training were accounted for, it was observed that individuals involved in the recreational gymnastic training had significantly greater NN CSA (+0.09 ± 0.03 cm2) and NN Z (+0.04 ± 0.01 cm3) than the non-gymnast controls (p < 0.05, Fig. 1). In contrast, former recreational gymnasts (ex-gymnasts) were noted to have no significant difference in the development of NN CSA and NN Z compared with the non-gymnast controls (p > 0.05, Fig. 1).

image

Figure 1. Predicted cross-sectional area (CSA) and section modulus (Z) at the narrow neck site of the proximal femur for male and female gymnasts (GYM), ex-gymnasts (EX GYM), and non-gymnasts controls (CON). Values are predicted from age-appropriate bone geometry values derived from the multilevel longitudinal models (Table 2); mean values at each age for height (cm); weight (kg); vitamin D (UI); calcium (mg); physical activity score; and hours trained (hours/week).

Download figure to PowerPoint

Intertrochanter

The result of the multilevel models for CSA and Z in males and females is summarized in Table 3. Age centered, weight, and sex significantly contributed to the prediction of IT CSA, whereas age centered, age centered2, height, and weight contributed to prediction of IT Z. From the models, it was observed that there were significant sex differences in the development of IT CSA, with females having developed significantly less IT CSA (−0.05 ± 0.02 cm2, p < 0.05) than their male peers. No significant sex difference was observed for the development of IT Z (p > 0.05, Table 3).

Once adjusted for age, height, weight, vitamin D, calcium intake, physical activity, sex, and the hours of gymnastics training, recreational gymnasts were observed to have developed significantly greater IT CSA (+0.11 ± 0.04 cm2) and IT Z (+0.07 ± 0.03 cm3) than their non-gymnast counterparts (p < 0.05, Fig. 2). Similar to the observation at the NN, no significant differences were observed in the development of IT CSA and IT Z between ex-gymnasts and the non-gymnast controls (p > 0.05, Fig. 2).

image

Figure 2. Predicted cross-sectional area (CSA) and section modulus (Z) at the intertrochanteric site of the proximal femur for male and female gymnasts (GYM), ex-gymnasts (EX GYM), and non-gymnast controls (CON). Values are predicted from age-appropriate bone geometry values derived from the multilevel longitudinal models (Table 3); mean values at each age for height (cm); weight (kg); vitamin D (UI); calcium (mg); physical activity score; and hours trained (hours/week).

Download figure to PowerPoint

Femoral shaft

Table 4 summarizes the multilevel models for CSA and Z in males and females at the femoral shaft site of the proximal femur. Age centered, height, weight, and sex significantly contributed to the prediction of S CSA, whereas age centered, age centered 2, height, weight, physical activity, and sex significantly contributed to the prediction of S Z. Paralleling the observations at the NN and IT, significant sex differences were also noted in the development of S CSA and S Z, with females continuing to have developed significantly less S CSA (−0.04 ± 0.02 cm2) and S Z (−0.03 ± 0.01 cm3) than their male equivalents (p < 0.05, Table 4).

When comparing the development of S CSA and S Z between gymnasts and non-gymnast controls, it was observed that recreational gymnasts developed significantly greater S CSA (+0.06 ± 0.03 cm2, Fig. 3), whereas no differences were observed in S Z between gymnasts and non-gymnast controls (p > 0.05, Fig. 3). Additionally, no significant differences were observed in either the development of SCSA or S Z between ex-gymnasts and non-gymnast controls (p > 0.05, Fig. 3).

image

Figure 3. Predicted cross-sectional area (CSA) and section modulus (Z) at the femoral shaft site of the proximal femur for male and female gymnasts (GYM), ex-gymnasts (EX GYM), and non-gymnasts controls (CON). Values are predicted from age-appropriate bone geometry values derived from the multilevel longitudinal models (Table 4); mean values at each age for height (cm); weight (kg); vitamin D (UI); calcium (mg); physical activity score; and hours trained (hours/week).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

The aim of this study was to investigate whether recreational-level gymnastics participation was associated with the development of estimated structural geometry at the proximal femur in young males and females, and whether current or past participation history influenced this structural strength development. It was observed that those individuals who participated in recreational-level gymnastics had significantly greater bone structural strength at all sites of the proximal femur compared with individuals previously involved in training and nongymnastic controls. This is the first study, to our knowledge, to investigate the effects of recreational-level gymnastics participation on bone structural strength at the proximal femur in both males and females during early childhood.

Gymnastics participation has been previously shown to be associated with a 3% to 28% benefit in aBMD and BMC at the total body, lumbar spine, and hip when compared with non-gymnast.[6-9, 14] This magnitude of benefits is similar to the effects documented in the current study. Our results show that those individuals currently involved in recreational gymnasts had approximately 3% to 7% greater CSA and Z at the proximal femur compared with both ex-gymnasts and non-gymnast controls, when adjusted for age, height, weight, vitamin D, calcium intake, physical activity, sex, and the hours of gymnastics training. These observed effects are similar to those previously reported for BMC in the same cohort.[14] This finding would suggest that low levels of continued recreational gymnastics participation provide benefits not only to BMC but also to geometric bone structural strength at the proximal femur. Thus, despite the low levels of gymnastics exposure, the dynamic loading nature of the jumping and tumbling activities associated with the recreational gymnastic training may serve as adequate stimulus for initiating bone structural strength adaptations, supporting previous assertions by Laing and colleagues.[6] Laing and colleagues found that 4- to 8-year-old females participating in recreational gymnastics had significantly greater lumbar spine aBMD and forearm bone area than non-gymnast controls, suggesting that beginner-level gymnastics skills were adequate for stimulating gains to bone mass and area.

These advantages were not observed in the ex-gymnasts involved in the current study. The longitudinal models revealed that those individuals defined as ex-gymnasts did not possess any advantages to CSA and Z at any site of the proximal femur compared with non-gymnast controls, after adjusting for age, height, weight, vitamin D, calcium intake, physical activity, sex, and the hours of gymnastics training. These results would suggest that previous involvement in recreational gymnastics does not confer any advantages to childhood bone geometry. These conclusions, however, may be influenced by the length of time ex-gymnasts were involved in gymnastics. Greater gymnastic exposure, in either number of hours trained or years of training, have been documented to increase total and regional aBMD, suggesting a dose-response relationship between loading and bone mass.[27, 28] Given that the ex-gymnasts were involved in the recreational gymnastics training for <1 to 2 years (data not shown), the shortened exposure time to the recreational gymnastics may be insufficient to precipitate detectable bone structural changes.

The non-gymnasts controls were also a highly active group, evidenced by the non-significant difference in physical activity scores between groups; their similar activity levels to the ex-gymnasts may have resulted in contemporaneous structural strength benefits to a frequently loaded region such as the proximal femur. Supporting this conjecture is the finding that gymnastic exposure appears to provide smaller effects at the femoral shaft site compared with the more proximal NN and IT regions. This may be influenced by the way the skeletal load is perturbed by gymnastic loads at these regions, with greater bending effects being elicited at the two proximal regions and more axial effects at the shaft. These site-specific differences, however, may be better evident at sites uniquely loaded by the recreational gymnastic training, such as the distal radius. Recently, Erlandson and colleagues[29] using peripheral quantitative computed tomography (pQCT), in the same cohort, reported that ex-gymnasts had 5% to 11% greater adjusted cortical bone content, cortical area, and polar strength strain index at the forearm compared with non-gymnasts controls, but these variables were not significantly different at the weight-bearing tibia. Similarly, Ward and colleagues[30] compared bone strength parameters in the upper and lower limbs of young children (8 to 9 years of age) using pQCT to investigate differences between the peripheral axial skeleton of precompetitive gymnasts and age-matched controls. They observed that at the mid radius, gymnasts had significantly greater bone total area, periosteal circumference, cortical area, and bending strength than age-matched controls; however, no significant differences were observed between gymnasts and controls for any bone measures at the 65% tibia. Therefore, the recreational gymnastics in the current study may provide site-specific skeletal benefits to the ex-gymnasts that were not currently assessed. Additionally, retired gymnasts studies have observed that although competitive gymnastics may continue to confer skeletal advantages many years after gymnastic senescence, these benefits experience a similar rate of natural decline over time.[31] Thus, any advantages that may have been accrued by the ex-gymnasts during their short exposure to recreational gymnastics may have ceased to be evident by the end of the current study's investigation period. Further research is necessary to support these speculations and to elaborate on the longevity of any site-specific skeletal advantages and potential dose-response relationship resulting from recreational gymnastics exposure, particularly in these young growing populations.

Although there is a growing body of literature in male gymnasts, most studies investigating the effects of gymnastics training on bone strength measures have focused on female populations. Of the limited male research, it has been observed that male gymnasts have greater calcaneus BMD, total body and femoral neck BMC, and proximal femur CSA and cross-sectional moments of inertia (CSMI).[5, 14, 32] These observations parallel the findings of the current study, which suggests that continued participation in recreational gymnastics bears significant advantages to bone structural strength at the proximal femur in both males and females. In the current study, it was also observed that sex was a significant predictor in the multilevel models for all bone structural strength measures except IT Z, identifying males as having greater CSA at the NN, IT, and S, and greater Z at the NN and S sites. This would suggest that alongside the positive effects of recreational gymnastics participation at this young age to both sexes, sex-specific advantages to bone structural strength may also be apparent regardless of gymnastic participation. This supports the sex-specific effects reported in previous studies that have documented greater cortical thickness, total body BMC, and femoral neck BMC in young male precompetitive gymnasts.[14, 30] These sex-specific differences may be owing to increases in endocortical and periosteal apposition seen in prepubertal boys.[33] Given that type of exercises and their sex-specific adherence were not assessed, follow-up research is required to support the supposition of sex-specific benefits of gymnastics training.

Despite the novel findings of the current study, there are a number of study shortcomings. The observational mixed longitudinal design does not allow for any cause-effect assessments of current or past gymnastics participation on bone parameters. Controlled prospective studies before the initiation of gymnastics training are necessary to definitively address whether cumulative recreational gymnastic exposure is responsible for the observed greater bone structural strength. Physical activity was assessed for all participants in this study; however, the highly active nature of non-gymnasts, who participated in other recreational sports such as soccer, T-ball, basketball, and karate, may reduce or even mask the potential effects of the recreational gymnast participation. Further, the HSA procedure is also associated with a variety of limitations. DXA images are often noisy and blurred, resulting in the difficulty of locating precise edge margins.[21] In addition, the assessment of CSA and Z assumes the mineralization of adult bone. This assumption, however, results in an underestimation of CSA and Z in pediatric populations.[19, 21] Thus, any effects observed may undervalue true differences. Finally, HSA is sensitive to small changes in the positioning of femur.[20, 21] All DXA scans were performed by qualified technologists familiar with proper positioning of the proximal femur to ensure hip scans were performed with care to limit these potential errors; nevertheless, it is difficult to position the hip consistently in repeated measures over time.

In summary, when compared with other physically active children, males and females participating in recreational gymnastics had significantly greater CSA and Z development at the proximal femur. Previous history of any recreational gymnastics participation may provide skeletal benefits at the load-bearing proximal femur; potential advantages may be site and exposure time specific; however, further longitudinal investigations of this cohort are necessary to substantiate these notions.

Acknowledgments

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References

We acknowledge all the study participants and their families for their constant enthusiasm and commitment to this project. This study was supported in part by funding from the Canadian Institutes of Health Research (CIHR: MOP 57671) and The Saskatchewan Health Research Foundation (SHRF).

Authors' roles: Study design: ME and ABJ. Data collection: ME, RGR, and SJ. Data analysis: SJ, RGR, ME, and ABJ. Data interpretation: RGR, ME, SJ, and ABJ. Drafting and revising manuscript: RGR, ME, SJ, and ABJ. All authors approved the final version of this manuscript.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  • 1
    Bailey DA, McKay HA, Mirwald RL, Crocker PR, Faulkner RA. A six-year longitudinal study of the relationship of physical activity to bone mineral accrual in growing children: the university of Saskatchewan bone mineral accrual study. J Bone Miner Res. 1999; 14:16729.
  • 2
    Bailey DA. The Saskatchewan Pediatric Bone Mineral Accrual Study: bone mineral acquisition during the growing years. Int J Sports Med. 1997; 18(Suppl 3):S1914.
  • 3
    Baxter-Jones AD, Kontulainen SA, Faulkner RA, Bailey DA. A longitudinal study of the relationship of physical activity to bone mineral accrual from adolescence to young adulthood. Bone. 2008; 43:11017.
  • 4
    Baxter-Jones AD, Faulkner RA, Forwood MR, Mirwald RL, Bailey DA. Bone mineral accrual from 8 to 30 years of age: an estimation of peak bone mass. J Bone Miner Res. 2011; 26:172939.
  • 5
    Daly RM, Rich PA, Klein R, Bass S. Effects of high-impact exercise on ultrasonic and biochemical indices of skeletal status: a prospective study in young male gymnasts. J Bone Miner Res. 1999; 14:122230.
  • 6
    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; 141:2116.
  • 7
    Nickols-Richardson SM, Modlesky CM, O'Connor PJ, Lewis RD. Premenarcheal gymnasts possess higher bone mineral density than controls. Med Sci Sports Exerc. 2000; 32:639.
  • 8
    Zanker CL, Gannon L, Cooke CB, Gee KL, Oldroyd B, Truscott JG. Differences in bone density, body composition, physical activity, and diet between child gymnasts and untrained children 7-8 years of age. J Bone Miner Res. 2003; 18:104350.
  • 9
    Faulkner RA, Forwood MR, Beck TJ, Mafukidze JC, Russell K, Wallace W. Strength indices of the proximal femur and shaft in prepubertal female gymnasts. Med Sci Sports Exerc. 2003; 35:5138.
  • 10
    Burt LA, Greene DA, Ducher G, Naughton GA. Skeletal adaptations associated with pre-pubertal gymnastics participation as determined by DXA and pQCT: a systematic review and meta-analysis. J Sci Med Sport. 2013; 16:2319.
  • 11
    Robling AG, Castillo AB, Turner CH. Biomechanical and molecular regulation of bone remodeling. Annu Rev Biomed Eng. 2006; 8:45598.
  • 12
    Bonnick SL. Noninvasive assessments of bone strength. Curr Opin Endocrinol Diabetes Obes. 2007; 14:4517.
  • 13
    Petit MA, McKay HA, MacKelvie KJ, Heinonen A, Khan KM, Beck TJ. A randomized school-based jumping intervention confers site and maturity-specific benefits on bone structural properties in girls: a hip structural analysis study. J Bone Miner Res. 2002; 17:36372.
  • 14
    Erlandson MC, Kontulainen SA, Chilibeck PD, Arnold CM, Baxter-Jones AD. Bone mineral accrual in 4- to 10-year-old precompetitive, recreational gymnasts: a 4-year longitudinal study. J Bone Miner Res. 2011; 26:131320.
  • 15
    Ross W, Marfell-Jones M. Kinanthropometry. In: Green H, editor. Physiological testing of the high performance athlete. Champaign, IL: Human Kinetics; 1991.
  • 16
    International Standards for Anthropometric Assessment. International Society for the Advancement of Kinanthropometry. Sydney, Australia: UNSW Press; 2001.
  • 17
    Beck TJ, Ruff CB, Warden KE, Scott WW Jr, Rao GU. Predicting femoral neck strength from bone mineral data. A structural approach. Invest Radiol. 1990; 25:618.
  • 18
    Martin RB, Burr DB. Non-invasive measurement of long bone cross-sectional moment of inertia by photon absorptiometry. J Biomech. 1984; 17:195201.
  • 19
    Beck TJ. Hip Structural Analysis (HSA) Program (BMD and structural geometry methodology): as used to create NHANES III dataset. 2002.
  • 20
    Beck T. Measuring the structural strength of bones with dual-energy X-ray absorptiometry: principles, technical limitations, and future possibilities. Osteoporos Int. 2003; 14:818.
  • 21
    Beck TJ. Extending DXA beyond bone mineral density: understanding hip structure analysis. Curr Osteoporos Rep. 2007; 5:4955.
  • 22
    Khoo BC, Beck TJ, Qiao QH, Parakh P, Semanick L, Prince RL, Singer KP, Price RI. In vivo short-term precision of hip structure analysis variables in comparison with bone mineral density using paired dual-energy X-ray absorptiometry scans from multi-center clinical trials. Bone. 2005; 37:11221.
  • 23
    Janz KF, Broffitt B, Levy SM. Validation evidence for the Netherlands physical activity questionnaire for young children: the Iowa bone development study. Res Q Exerc Sport. 2005; 76:3639.
  • 24
    Whiting SJ, Shrestha R. Dietary assessment of elementary school aged children and adolescents. J Can Diet Assoc. 1993; 54:1936.
  • 25
    Goldstein H, Browne W, Rasbash J. Multilevel modeling of medical data. Stat Med. 2002; 21:3291315.
  • 26
    Baxter-Jones A, Mirwald RL. Multilevel modeling. In: Hauspie RC, Cameron N, Molinari L, editors. Methods in human growth research. Cambridge, MA: Cambridge University Press; 2004; p.30630.
  • 27
    Scerpella TA, Davenport M, Morganti CM, Kanaley JA, Johnson LM. Dose related association of impact activity and bone mineral density in pre-pubertal girls. Calcif Tissue Int. 2003; 72:2431.
  • 28
    Laing EM, Wilson AR, Modlesky CM, O'Connor PJ, Hall DB, Lewis RD. Initial years of recreational artistic gymnastics training improves lumbar spine bone mineral accrual in 4- to 8-year-old females. J Bone Miner Res. 2005; 20:50919.
  • 29
    Erlandson MC, Kontulainen SA, Baxter-Jones AD. Precompetitive and recreational gymnasts have greater bone density, mass, and estimated strength at the distal radius in young childhood. Osteoporos Int. 2011; 22:7584.
  • 30
    Ward KA, Roberts SA, Adams JE, Mughal MZ. Bone geometry and density in the skeleton of pre-pubertal gymnasts and school children. Bone. 2005; 36:10128.
  • 31
    Kudlac J, Nichols DL, Sanborn CF, DiMarco NM. Impact of detraining on bone loss in former collegiate female gymnasts. Calcif Tissue Int. 2004; 75:4827.
  • 32
    Hind K, Gannon L, Whatley E, Cooke C, Truscott J. Bone cross-sectional geometry in male runners, gymnasts, swimmers and non-athletic controls: a hip-structural analysis study. Eur J Appl Physiol. 2011; 112:53541.
  • 33
    Bradney M, Pearce G, Naughton G, Sullivan C, Bass S, Beck T, Carlson J, Seeman E. Moderate exercise during growth in prepubertal boys: changes in bone mass, size, volumetric density, and bone strength: a controlled prospective study. J Bone Miner Res. 1998; 13:181421.