SEARCH

SEARCH BY CITATION

Keywords:

  • EXERCISE;
  • PQCT;
  • BONE STRENGTH;
  • TIBIA;
  • RADIUS

Abstract

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

The long-term benefits of habitual physical activity during adolescence on adult bone structure and strength are poorly understood. We investigated whether physically active adolescents had greater bone size, density, content, and estimated bone strength in young adulthood when compared to their peers who were inactive during adolescence. Peripheral quantitative computed tomography (pQCT) was used to measure the tibia and radius of 122 (73 females) participants (age mean ± SD, 29.3 ± 2.3 years) of the Saskatchewan Pediatric Bone Mineral Accrual Study (PBMAS). Total bone area (ToA), cortical density (CoD), cortical area (CoA), cortical content (CoC), and estimated bone strength in torsion (SSIp) and muscle area (MuA) were measured at the diaphyses (66% tibia and 65% radius). Total density (ToD), trabecular density (TrD), trabecular content (TrC), and estimated bone strength in compression (BSIc) were measured at the distal ends (4%). Participants were grouped by their adolescent physical activity (PA) levels (inactive, average, and active) based on mean PA Z-scores obtained from serial questionnaire assessments completed during adolescence. We compared adult bone outcomes across adolescent PA groups in each sex using analysis of covariance followed by post hoc pairwise comparisons with Bonferroni adjustments. When adjusted for adult height, MuA, and PA, adult males who were more physically active than their peers in adolescence had 13% greater adjusted torsional bone strength (SSIp, p < 0.05) and 10% greater adjusted ToA (p < 0.05) at the tibia diaphysis. Females who were more active in adolescence had 10% larger adjusted CoA (p < 0.05), 12% greater adjusted CoC (p < 0.05) at the tibia diaphysis, and 3% greater adjusted TrC (p < 0.05) at the distal tibia when compared to their inactive peers. Benefits to tibia bone size, content, and strength in those who were more active during adolescence seemed to persist into young adulthood, with greater ToA and SSIp in males, and greater CoA, CoC, and TrC in females. © 2014 American Society for Bone and Mineral Research.


Introduction

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

Exercise interventions have been shown to improve bone mass and strength development in a site-specific and sex-specific manner in childhood and adolescence.[1, 2] Prospective observational evidence has further indicated that habitual physical activity (PA) is associated with enhanced bone mass accrual during periods of rapid growth in adolescence.[3] Importantly, these bone mass benefits seem to be carried over into young adulthood,[4] suggesting a long-term benefit from habitual PA in adolescence on bone mineral mass. Less is known, however, whether habitual PA during skeletal growth is related to long-term adaptation in bone structure and strength. This evidence of maintenance is essential to justify childhood and adolescent PA recommendations for enhancing bone mass and strength development, which could potentially prevent osteoporosis and related fractures later in life.[5, 6]

In comparison to the mature skeleton, the growing skeleton has heightened responsiveness to loading stimuli.[1, 2, 7] In addition to enhanced bone mineral accrual, adaptations in bone structure and size contribute to greater bone strength benefits when loading stimulus has been obtained during skeletal growth.[7] Habitual PA and sports participation during skeletal growth may cause permanent adaptations to bone structure, size, and tissue properties whereby even an interruption in training stimulus would not abolish all benefits.[8] This could explain why exercise-induced bone mass (or areal bone mineral density) has been sustained in children and adolescents up to 8 years after exercise interventions,[9-11] whereas the opposite has been reported in premenopausal and postmenopausal women.[12, 13] This evidence is supported by retrospective assessment of earlier sports participation in adult males[14] and athlete versus nonathlete comparisons, suggesting sustained skeletal benefits from high levels of childhood PA on adult bone mass, structure, and microarchitecture.[8, 15-21] However, as in any observational study, maintenance of bone mass and strength into adulthood may be influenced by an individual's genetic predisposition, such as greater muscle size and strength and/or current level of PA.

Currently, little is known about the long-term skeletal benefits of habitual PA during childhood and adolescence. Prospective follow-up data from the Saskatchewan Pediatric Bone Mineral Accrual study (PBMAS) showed a 9% to 17% greater bone mineral accrual during a 2-year period around peak height velocity in those children who were more physically active than their peers.[3] These skeletal benefits of adolescent PA were maintained into young adulthood, with active adolescent males having 8% to 10% more adjusted bone mass at the total body, hip, and femoral neck compared to their inactive peers.[4] Active adolescent females had 9% and 10% more adjusted bone mass at the total hip and femoral neck, respectively.[4] These results suggest that benefits in bone mass accrued in physically active adolescents persist into young adulthood. Although long-term benefits in bone mineral accrual have been identified, it is unclear if habitual PA during childhood and adolescence in this cohort could lead to benefits in bone structure, density, and estimated strength in young adulthood.

The objective of the present study was to assess if physically active adolescents had greater tibia and radius bone size, (volumetric) bone mineral density, mineral content, and estimated bone strength in young adulthood. We hypothesized that more physically active adolescent males and females would have greater bone strength and bone densitometric parameters in young adulthood when compared to their peers who were less physically active in adolescence.

Subjects and Methods

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

Participants and assessment

The Saskatchewan Pediatric Bone Mineral Accrual Study (PBMAS) used a mixed longitudinal study design that has been described in detail.[3, 4] In brief, in 1991, participants aged between 8 and 15 years were recruited from two elementary schools within the city of Saskatoon; of the 375 eligible students, parents of 228 children (113 boys and 115 girls) provided written consent for participation and 220 were scanned using dual-energy X-ray absorptiometry (DXA). From 1992 to 1993 an additional 31 children were recruited and scanned. After 7 years of data collection, 109 males and 121 females had been measured on two or more occasions (median 6 scans). Between 2002 and 2006, 169 participants were measured on at least 1 further occasion (age range, 18.2–27.5 years). Between 2009 and 2011, 122 participants (49 males and 73 females) aged between 24 and 34 years had bone parameters assessed with peripheral quantitative computed tomography (pQCT). For the present analysis we included PBMAS participants with radius and/or tibia pQCT and physical activity measurements in young adulthood and measures of PA over the entire adolescent period (n = 122).

Anthropometry

Anthropometric measurements were taken during each laboratory visit by trained personnel following standard protocols. Stretch height was measured without shoes to the nearest 0.1 cm using a wall-mounted stadiometer (Holtain Ltd., Crosswell, UK) and weight to the nearest 0.01 kg was measured on a calibrated mechanical weighing scale (Toledo Scale Company, Ontario, Canada). The length of the lower leg and forearm were measured to the nearest 0.1 cm using an anthropometric tape measure. Lower leg length was measured as the distance from the tibial joint line to the distal edge of the medial malleolus. Forearm length was measured as the distance from the distal aspect of the radial styloid process to the proximal head of the radius.

pQCT

The nondominant (self-reported) forearm and ipsilateral lower leg were scanned using pQCT (Stratec XCT 2000; Stratec Medical, Pforzheim, Germany). Scout scans were performed over the wrist and ankle joint, with the scanner reference line positioned at the most proximal aspect of the endplate of the distal tibia and medial tip of the distal radius. Cross-sectional pQCT images were acquired at the following sites, proximal to the reference line: 4% and 66% of the tibia length; and 4% and 65% of the radius length. The pQCT scanning parameters included: 2.4-mm slice thickness, 0.4-mm pixel size, and 20-mm/s scanning speed.

We used the pQCT manufacturer's software (version 6.0; Stratec Medical) to acquire structural and densitometric parameters of bone and muscle. At the shaft sites of tibia (66% of limb length) and radius (65% of limb length), total area (ToA, mm2) was derived using contour mode 1 with a threshold of 280 mg/cm, which delineated the periosteal (outer) surface of bone at the diaphysis. Cortical bone area (CoA, mm2), cortical density (CoD, mg/cm3), and cortical content (CoC, mg/mm) were derived using separation mode 4 with a threshold of 480 mg/cm3. Density-weighted polar section modulus (SSIp, mm3) was also derived to estimate diaphyseal bone resistance to torsional loading.[22] For the distal tibia and radius (4% of limb length), total bone area (ToA, mm2) and density (ToD, mg/cm3) were defined using a contour mode 1 with an outer threshold of 169 mg/cm3. Trabecular area (TrA, mm2), trabecular density (TrD, mg/cm3), and trabecular content (TrC, mg/mm) were defined using peel mode 2 and threshold of 480 mg/cm3. To estimate bone strength in compression, a bone strength index (BSIc, mg2/mm4) was calculated by multiplying ToA by ToD squared (BSIc = ToA × ToD2).[23] All scans were performed by a single operator and analyzed by another two technicians (regions of interest were manually defined by one of these technicians and confirmed by another). The short-term precision error (coefficient of variation [CVrms]) for the reported bone outcomes in our laboratory ranged from 1.4% to 6.1% at the radius and 0.7% to 2.1% at the tibia.[24] We were unable to obtain five tibia shaft scans and three distal tibia scans from 5 female participants because their leg did not fit comfortably in the gantry. Scans with movement artifacts were omitted from the analysis.

We assessed muscle area (MuA, mm2) from the shaft sites of both lower and upper extremities. Muscle tissue was defined as voxels with a density greater than 40 mg/cm3 (differentiating muscle from subcutaneous fat) and less than 280 mg/cm3 (differentiating muscle from bone). In our laboratory, precision error (CVrms) for muscle area was 1.4% in the forearm and 3.7% in the lower leg in duplicate measures of 49 adults.[25]

PA assessment

The assessment of PA has previously been reported.[4] In brief, childhood and adolescence PA was assessed via the Physical Activity Questionnaires for Children (PAQ-C) and Adolescents (PAQ-A).[26] These nine-item questionnaires provided an estimate of moderate and vigorous PA in grades 3 and up using a five-point scale.[26] Higher scores on the questionnaires indicate higher PA levels. PA was defined as “sports, games, or dance that make you breath hard, make your legs feel tired, or make you sweat.”[26] In completing the questionnaire, participants were asked to rate their level of PA over the past 7 days. For the present analysis, childhood and adolescence physical activity scores were calculated for each individual and sex-specific and age-specific means were determined. Individual PA scores were then used to calculate Z-scores. The Z-scores were based on the mean and SD for the entire sample of similar chronological age. Individual Z-scores were averaged over the entire adolescent period to determine a PA score during childhood and adolescence. Based on their PA Z-score, participants were categorized into quartiles (inactive = quartile 1; average = quartiles 2 and 3; and active = quartile 4).[3, 4]

In young adulthood, PA was assessed via the Physical Activity Questionnaire for Adults (PAQ-AD), a seven-item version of the PAQ-C and PAQ-A designed for use after secondary education was completed.[27] For the present analysis, adulthood PA scores determined in 2009–2011 were used as a continuous variable. As reported,[27-29] the PAQ-C and PAQ-A has good test-retest reliability and the PAQ-AD has a favorable validity and moderate relationship to the ratings of leisure time activity scale, self-reported physical activity recall, and motion sensors.

Dietary analysis

Dietary calcium and vitamin D intake was assessed during 2009–2011 using two 24-hour recalls conducted at the time of the pQCT scanning. Food intake, from the 24-h recalls, was analyzed using a nutritional assessment software package (Food Processor version 10; ESHA Research, Salem, OR, USA) that contained foods from the 2007 Canadian Nutrient File. All 24-hour recalls were processed by the same research assistant and checked for consistency by a second research assistant according to procedures described elsewhere.[30, 31] In addition to 24-hour recalls, we collected data on usual vitamin D and calcium supplement use. In this study, we used data on daily average intake of vitamin D (IU) and calcium (mg) from food and supplement for descriptive purposes owing to the small number of participants with valid data (n = 77).

Statistical analysis

All data was first checked for normality using the Kolmogorov-Smirnov test and outliers (Z-score greater than 3.29) were removed. Sex-specific models were developed because of the reported sex-differences in bone structural and densitometric parameters[32, 33] and the observed sex-differences in the current sample (data not shown). Differences in anthropometric characteristics, nutritional data, and PA levels were compared across the three physical activity groups in both sexes using analysis of variance (ANOVA). To assess differences in young adult bone parameters (ToA, ToD, CoD, CoA, CoC, TrD, TrA, and TrC) and strength estimates (SSIp and BSIc) between adolescence PA groupings (inactive, average, active), site-specific (tibia diaphysis, distal tibia, radius diaphysis, and distal radius) analyses of covariance (ANCOVAs) were performed for each sex separately. The covariates were adult height, MuA, and adult PA. Selection of covariates was based on the literature[25, 34] and observed associations between the covariates and bone outcomes in the current sample (data not shown). If significant between-group differences across adolescent physical activity groups were observed in ANCOVA, post hoc pairwise comparisons with Bonferroni adjustments were used to establish individual group differences. Adjusted means, standard errors (SE), and percentage differences between the adolescent PA groups from the pairwise comparisons are reported. An alpha level of p < 0.05 was considered statistically significant. All statistical analyses were performed with the Statistical Package for Social Science (IBM SPSS, Version 20 for Windows; SPSS Inc., Chicago, IL, USA).

Results

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

Young adult anthropometric characteristics, PA scores, and dietary calcium and vitamin D by adolescent activity group (within each sex) are presented in Table 1. There were no significant differences in young adulthood in age, height, weight, lower leg MuA, calcium or Vitamin D intakes, or adult PA in sex-separated comparisons between adolescence activity groups (Table 1). Forearm MuA was higher in females who were active in adolescence when compared to their inactive peers (p = 0.009) while there was no differences in the forearm MuA across the male adolescent activity groups (Table 1).

Table 1. Young Adults' Anthropometric, PA, Daily Calcium, and Vitamin D Intakes by Sex and Adolescence PA Group
MalesInactive (n = 12)Average (n = 25)Active (n = 12)
Chronological age (years)29 ± 329 ± 230 ± 1
Height (cm)179.7 ± 8.5179.6 ± 6.4179.6 ± 6.6
Weight (kg)86.4 ± 11.189.9 ± 15.284.8 ± 13.8
MuA lower leg (mm2)8678 ± 13118617 ± 10698444 ± 1617
MuA forearm (mm2)4572 ± 7564860 ± 7464876 ± 1064
Adolescence PA (Z-score)−1.3 ± 0.5−0.1 ± 0.41.4 ± 0.5a
Adult PA (score)2.2 ± 0.62.4 ± 0.42.5 ± 0.5
 Inactive (n = 8)Average (n = 18)Active (n = 10)
Calcium (mg/d)904 ± 210903 ± 511903 ± 493
Vitamin D (IU/d)188 ± 107162 ± 143111 ± 89
FemalesInactive (n = 18)Average (n = 37)Active (n = 18)
Chronological age (years)29 ± 230 ± 229 ± 3
Height (cm)168.4 ± 7.8166.1 ± 6.2165.0 ± 8.3
Weight (kg)69.1 ± 18.078.0 ± 24.372.2 ± 14.5
MuA lower leg (mm2)6001 ± 9226812 ± 13726950 ± 1201
MuA forearm (mm2)2549 ± 2962733 ± 4442938 ± 515a
Adolescence PA (Z-score)−1.3 ± 0.7−0.2 ± 0.31.2 ± 0.5a
Adult PA (score)2.0 ± 0.72.2 ± 0.52.4 ± 0.7
 Inactive (n = 9)Average (n = 22)Active (n = 10)
  1. Values are mean ± SD: ANOVA, post hoc pairwise comparisons with a Bonferroni adjustment.

  2. PA = physical activity; n = number of individuals in the adolescent PA group with valid data; MuA = muscle area.

  3. a

    Significant difference from the inactive and average adolescent PA groups (p < 0.05).

Calcium (mg/d)690 ± 298848 ± 334847 ± 398
Vitamin D (IU/d)130 ± 104158 ± 102231 ± 302

At the tibia diaphysis in males, individuals who were classified as active in adolescence had 10% larger adjusted adult ToA (p = 0.021) and 13% greater adjusted adult torsional bone strength (SSIp, p = 0.023) when compared to males who were classified as inactive in adolescence (Table 2). Males who were classified as average active in adolescence did not differ significantly from other groups (p > 0.05) (Table 2). No significant between-group differences were found in other adjusted bone outcomes in males at the tibia diaphysis (Table 2).

Table 2. Young Adult Male Adjusted Bone Parameters by Adolescence PA Group
Diaphysis tibiaaInactive (n = 12)Average (n = 24)Active (n = 12)
ToA (mm2)688 ± 17723 ± 12755 ± 17b
CoA (mm2)423 ± 11425 ± 8439 ± 11
CoD (mg/cm3)1097 ± 71083 ± 51089 ± 7
CoC (mg/mm)464 ± 12460 ± 9478 ± 12
SSIp (mm3)3249 ± 1093410 ± 763687 ± 110b
Distal tibiacInactive (n = 12)Average (n = 25)Active (n = 12)
ToA (mm2)1222 ± 481354 ± 331366 ± 48
ToD (mg/cm3)356 ± 11345 ± 8344 ± 11
TrA (mm2)1045 ± 501187 ± 341196 ± 50
TrC (mg/mm)305 ± 16348 ± 11352 ± 16
TrD (mg/cm3)293 ± 9295 ± 6294 ± 9
BSIc (mg2/mm4)155 ± 11162 ± 8163 ± 11
Diaphysis radiusdInactive (n = 12)Average (n = 25)Active (n = 11)
ToA (mm2)156 ± 5166 ± 3170 ± 5
CoA (mm2)113 ± 3114 ± 2116 ± 3
CoC (mg/mm)127 ± 4126 ± 3129 ± 4
CoD (mg/cm3)1117 ± 81111 ± 51110 ± 8
SSIp (mm3)381 ± 18418 ± 12443 ± 19
Distal radiuseInactive (n = 12)Average (n = 25)Active (n = 12)
  1. Values are mean ± SE.

  2. PA = physical activity; n = number of individuals in the adolescent activity group with valid data; ToA = total area; CoA = cortical area; CoD = cortical density; CoC = cortical content; SSIp = stress strain index; ToD = total density; TrA = trabecular area; TrC = trabecular content; TrD = trabecular density; BSIc = bone strength index; MuA = muscle area; ANCOVA = analysis of covariance.

  3. a

    Adjusted for height 180 cm, lower leg MuA 8233 mm2, adult PA score 2.3.

  4. b

    ANCOVA, post hoc pairwise comparison with a Bonferroni adjustment indicates a significant difference (p < 0.05) from the inactive adolescence PA group.

  5. c

    Adjusted for height 180 cm, lower leg MuA 8590 mm2, Adult PA score 2.3.

  6. d

    Adjusted for height 179 cm, forearm MuA 4757 mm2, Adult PA score 2.3.

  7. e

    Adjusted for height 180 cm, forearm MuA 4793 mm2, Adult PA score 2.3.

ToA (mm2)442 ± 20473 ± 14480 ± 20
ToD (mg/cm3)383 ± 16366 ± 11373 ± 16
TrA (mm2)349 ± 22383 ± 15388 ± 22
TrC (mg/mm)96 ± 6105 ± 4107 ± 6
TrD (mg/cm3)275 ± 8274 ± 6277 ± 8
BSIc (mg2/mm4)64 ± 463 ± 367 ± 4

At the tibia diaphysis in females, individuals who were classified as active in adolescence had 10% larger adjusted adult CoA (p = 0.008) and 12% higher adjusted adult CoC (p = 0.003) when compared to females classified as inactive in adolescence (Table 3). Similar to males, females who were classified as average active in adolescence did not differ significantly from other groups (Table 3). Unlike males, there were no significant differences in adjusted SSIp between female adolescent activity groups at the tibia diaphysis (Table 3).

Table 3. Young Adult Female Adjusted Bone Parameters by Adolescence PA Group
Diaphysis tibiaaInactive (n = 17)Average (n = 29)Active (n = 17)
ToA (mm2)532 ± 13549 ± 10548 ± 13
CoA (mm2)304 ± 6319 ± 5333 ± 6b
CoD (mg/cm3)1100 ± 51103 ± 41108 ± 5
CoC (mg/mm)334 ± 7352 ± 5369 ± 7b
SSIp (mm3)2147 ± 662265 ± 492325 ± 65
Distal tibiacInactive (n = 17)Average (n = 32)Active (n = 18)
ToA (mm2)1004 ± 241063 ± 171061 ± 23
ToD (mg/cm3)289 ± 9288 ± 6300 ± 8
TrA (mm2)907 ± 29964 ± 20958 ± 27
TrC (mg/mm)221 ± 6236 ± 5243 ± 6b
TrD (mg/cm3)245 ± 6246 ± 4254 ± 6
BSIc (mg2/mm4)84 ± 589 ± 396 ± 4
Diaphysis radiusa,dInactive (n = 18)Average (n = 35)Active (n = 18)
ToA (mm2)118 ± 3119 ± 2117 ± 3
CoA (mm2)83 ± 284 ± 283 ± 2
CoC (mg/mm)93 ± 396 ± 294 ± 3
CoD (mg/cm3)1122 ± 71133 ± 51133 ± 7
SSIp (mm3)251 ± 10257 ± 7252 ± 10
Distal radiuseInactive (n = 18)Average (n = 32)Active (n = 18)
  1. Values are mean ± SE.

  2. PA = physical activity; n = number of individuals in the adolescent activity group with valid data; ToA = total area; CoA = cortical area; CoD = cortical density; CoC = cortical content; SSIp = stress strain index; ToD = total density; TrA = trabecular area; TrC = trabecular content; TrD = trabecular density; BSIc = bone strength index; MuA = muscle area; ANCOVA = analysis of covariance.

  3. a

    Adjusted for height 166 cm, lower leg MuA 6697 mm2, adult PA score 2.2.

  4. b

    ANCOVA, post hoc pairwise comparison with a Bonferroni adjustment indicates a significant difference (p < 0.05) from the inactive adolescence PA group.

  5. c

    Adjusted for height 168 cm, lower leg MuA 6721 mm2, adult PA score 2.2.

  6. d

    Adjusted for height 166 cm, forearm MuA 2738 mm2, adult PA score 2.2.

  7. e

    Adjusted for height 166 cm, forearm MuA 2737 mm2, adult PA score 2.2.

ToA (mm2)348 ± 8354 ± 6351 ± 8
ToD (mg/cm3)301 ± 9285 ± 6309 ± 9
TrA (mm2)295 ± 9305 ± 7292 ± 9
TrC (mg/mm)66 ± 366 ± 265 ± 3
TrD (mg/cm3)226 ± 6218 ± 4221 ± 6
BSIc (mg2/mm4)32 ± 229 ± 134 ± 2

At the distal tibia, the only significant between-group difference was noted in adjusted TrC in females, with females classified as active in adolescence having 3% higher adjusted TrC when compared to females classified as inactive in adolescence (p = 0.047) (Table 3).

At the radius, there were no between-group differences in adjusted bone parameters (ToA, ToD, CoD, CoA, CoC, TrD, TrA, and TrC) or strength indices (SSIp and BSIc) in either sex (p > 0.05) (Tables 2 and 3).

Discussion

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

The results of this study add to increasing evidence that positive benefits to the skeleton through childhood and adolescent PA persist into young adulthood. When controlling for adult height, MuA, and PA, males who were more physically active during adolescence had 13% greater adjusted torsional bone strength (SSIp) and 10% larger adjusted total area at the tibia diaphysis compared with their peers who were inactive during adolescence. Although there were no significant differences in torsional bone strength at the tibia diaphysis between activity groups in females, adult females who were more physically active during adolescence had 10% larger adjusted cortical area and 12% more adjusted cortical content when compared to their inactive adolescence peers. The only difference in trabecular bone was observed in the distal tibia; women who were active in adolescence had 3% greater trabecular content when compared to their peers who were inactive during adolescence. This suggests that bone structural adaptation though childhood PA can be maintained, especially at the weight-bearing cortical bone sites into young adulthood in both males and females.

This study suggests that habitual PA in adolescence can result in long-term benefits in bone structure, mineral content, and strength at loading sites independent of current PA. According to our knowledge, this is the first evidence of a positive relationship between adolescent physical activity and adult bone structure and strength from a healthy nonathletic longitudinal cohort. These findings are consistent with athlete loading models that have shown sustained skeletal benefits from high levels of childhood PA on adult bone mass, structure, and architecture.[8, 15-20] Furthermore, the observed differences were similar to earlier reported benefits in bone mass.[4] The current study adds to this evidence, highlighting that increased loading at the weight-bearing tibia diaphysis during adolescence may lead to 10% to 13% larger adjusted bone size and greater strength into adulthood, independent of current PA.

The present study identified surface specific differences in bone structural adaptation according to sex. At the tibial diaphysis, males who were more physically active during adolescence had greater adjusted ToA and torsional bone strength (SSIp) than males who were inactive in adolescence, whereas females who were more physically active during adolescence had higher adjusted CoA and CoC in young adulthood when compared to their peers who were inactive in adolescence. Although there were no differences in torsional bone strength, differences in CoA and CoC may indicate higher axial and bending strength in females who were more physically active during adolescence.[23] We speculate that the male and female differences in bone formation may have occurred from adolescence during the pubertal estrogen surge.[35] In girls, the estrogen surge during puberty has been shown to suppress the osteogenic effect of exercise at the periosteal surface and lead to slower expansion of the medullary area or endocortical contraction,[36, 37] whereas in boys the surge of testosterone during puberty may promote periosteal expansion directly, or indirectly via increases in muscle size and activity, resulting in a stronger bone.[36] Our study findings of greater total bone size and strength in males and greater cortical area in females, when compared to their adult peers who were inactive during adolescence, support this concept. In addition, differences in the skeletal loading—such as type and intensity of physical activities—between sexes may explain some of the observed differences in bone adaptation. However, the amount of PA in adolescence did not differ between the sexes in this cohort when sexes were aligned by maturity instead of chronological age.[38]

Interestingly, there were no differences in mean adult PA levels according to adolescence PA groups. PA levels have been shown to decrease but track with increasing age in both sexes, as observed in this cohort.[38, 39] Based on this evidence, the level of PA needed to maintain bone size and strength benefits, obtained by a physically active lifestyle in adolescence, is likely less than the activity needed for structural adaptation. Previous observations from retired athletes[7, 8, 34] and animal experiments[40-42] support this concept. Experimental studies have indicated lifelong benefits in cortical bone from loading stimulus obtained during skeletal growth.[41, 42] Altogether, this evidence suggests that developing larger cortex during growth may provide lifelong antifracture benefits by priming the skeleton to offset the cortical bone thinning and trabecularization of the endocortical bone associated with aging.[43] It is currently unknown if the same mechanism could provide long-term benefits in cortical bone structure at the distal radius (and other long bone ends) prone to fractures. Imaging tools with high resolution, such as high-resolution (HR)-pQCT, will enable addressing this question in the future studies.

This research has specific limitations related to study design and analysis methods. First, the observational nature of this study makes it inherently vulnerable to uncontrolled factors such as selection bias and reverse causality. Although the PA assessments were prospectively monitored, pQCT-derived bone measures relied on cross-sectional data observed in adulthood only; it is therefore not possible to determine if those adults who had been more physically active in adolescence already had increased bone strength at or prior to adolescence. It could be argued that children with a well-developed skeletal structure prior to adolescence are those who are naturally more active during adolescence and later life. Second, the results from this study may not be generalized to other populations because the small cohort of white adolescents followed into young adulthood was drawn from a small regional population living in Saskatchewan, Canada. Longitudinal studies in other populations and follow-ups after exercise interventions are needed to confirm the findings of the current observational study. Third, although the PAQ is a reliable and valid method for assessing PA in children, adolescents, and adults,[27-29] it provides little information on discriminating the nature of the activity. The PAQ focuses predominantly on activities that promote loading at weight-bearing sites and lacks information on upper body activities. Furthermore, the subjective measure of PA may lead to recall and reporting bias. The limitations related to the PAQ may explain the lack of findings at the radius in the current study. Because the types of PA and the loading site of the activity may play a role on the transfer of skeletal gains from adolescence to adulthood, it is important for future studies to develop more robust assessments for upper extremity PA. Moreover, these activities need to be classified separately from weight-bearing activities to determine site-specific gains in bone strength. In addition, randomized control trials are needed to identify which upper extremity activities during childhood will maintain bone strength at the radius into adulthood. Fourth, another limitation of this study is the self-reported dietary intakes, which can lead to underestimation due to underreporting. Using data from the PBMAS participants, we have previously shown that underreporting was greater in females, particularly older females, than in males.[44] However, despite these limitations, the current study provides novel findings that may have future implications for promoting PA in children and adolescents. Increasing PA during adolescence will not only promote early-life bone structural adaptation but will also help to maintain the associated benefits into young adulthood, thus potentially reducing future risk for fracture later in life.

In conclusion, this study indicates that adolescent PA will have positive implications on bone structure and strength in young adulthood. Bone benefits, especially in cortical bone size, mineral content, and strength, seemed to persist into young adulthood, even after accounting for young adult PA.

Acknowledgments

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

This study was supported by the Canadian Institute of Health Research (CIHR) (MOP 98002; MOP57671) and the Saskatchewan Health Research Foundation. We thank Ashlee McLardy, Clark Mundt and Lauren Sherar for coordinating the PBMAS study (2002–2011); Andrew Franks for the pQCT data acquisition and Megan Labas and Emma Burke for pQCT data analysis (2009–2011); Preston O'Brien and Carly Phinney for the nutritional analysis, and Marta Erlandson, Kana Nemoto, and Joel Krentz for their assistance in data collection (2009–2011). We also acknowledge that SK is a CIHR Regional Partnership Program New Investigator.

Authors' roles: Study design: ABJ, JJ, DC, and SK. Study conduct: ABJ and SK. Data collection: ABJ and SK. Data analysis: RD, ABJ, HV, and SK. Data interpretation: RD, ABJ, and SK. Drafting manuscript: RD, ABJ, and SK. Revising manuscript content: RD, ABJ, JJ, HV, DC, and SK. Approving final version of manuscript: RD, ABJ, JJ, HV, DC, and SK. SK takes responsibility for the integrity of the data analysis.

References

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