SEARCH

SEARCH BY CITATION

Keywords:

  • PERIPHERAL QUANTITATIVE COMPUTED TOMOGRAPHY;
  • FINITE ELEMENT ANALYSIS;
  • CORTICAL BONE;
  • EXERCISE;
  • OLD MEN

ABSTRACT

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

Previous studies have reported an association between exercise during youth and increased areal bone mineral density at old age. The primary aim of this study was to investigate if exercise during growth was independently associated with greater cortical bone size and whole bone strength in weight-bearing bone in old men. The tibia and radius were measured using both peripheral quantitative computed tomography (pQCT) (XCT-2000; Stratec) at the diaphysis and high-resolution pQCT (HR-pQCT) (XtremeCT; Scanco) at the metaphysis to obtain cortical bone geometry and finite element–derived bone strength in distal tibia and radius, in 597 men, 79.9 ± 3.4 (mean ± SD) years old. A self-administered questionnaire was used to collect information about previous and current physical activity. In order to determine whether level of exercise during growth and young adulthood or level of current physical activity were independently associated with bone parameters in both tibia and radius, analysis of covariance (ANCOVA) analyses were used. Adjusting for covariates and current physical activity, we found that men in the group with the highest level of exercise early in life (regular exercise at a competitive level) had higher tibial cortical cross-sectional area (CSA; 6.3%, p < 0.001) and periosteal circumference (PC; 1.6%, p = 0.011) at the diaphysis, and higher estimated bone strength (failure load: 7.5%, p < 0.001; and stiffness: 7.8%, p < 0.001) at the metaphysis than men in the subgroup with the lowest level of exercise during growth and young adulthood. Subjects in the group with the highest level of current physical activity had smaller tibial endosteal circumference (EC; 3.6%, p = 0.012) at the diaphysis than subjects with a lower current physical activity, when adjusting for covariates and level of exercise during growth and young adulthood. These findings indicate that exercise during growth can increase the cortical bone size via periosteal expansion, whereas exercise at old age may decrease endosteal bone loss in weight-bearing bone in old men. © 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

Osteoporosis is a disease characterized by reduced bone mass and microarchitectural deterioration of bone tissue leading to an increased risk of fracture.[1] The variance in bone mass is mostly genetically determined,[2-4] but exercise with loading of the bone also has a major impact on bone mass[5-7] as well as on bone strength.[5, 8, 9] The adaptive response to mechanical loading is highly site-specific; only those bones that are actually loaded will adapt. This has been shown in several studies in racquet-sport players, where the arm holding the racquet had significantly greater bone mass and size than the contralateral nonplaying arm.[5, 8-10] Dynamic load in excess of loads encountered in daily life has the most favorable effect on the skeleton.[11, 12] A study on older twin pairs suggested that the relative importance of physical activity compared to genetic factors for structural bone strength was greater for the weight-bearing tibia than for the non–weight-bearing radius.[13] In comparison to the mature skeleton, the young and growing skeleton seems to be more adaptive and has a higher responsiveness to loading stimuli by exercise.[5, 9, 14] Physical activity has been suggested as an intervention strategy to promote an optimal peak bone mass during growth and to reduce the rate of bone loss during adulthood.[15, 16] In addition, high peak bone mass is associated with reduced risk of osteoporotic fractures later in life.[15] Physical exercise during skeletal growth may cause beneficial adaptations to bone size and structure, but it is still debated whether these benefits will be maintained with reduction in activity level.[14] Some studies demonstrate that the benefits of physical activity are lost after its cessation.[17, 18] In contrast, several other studies have shown that the benefits of previous training will remain when the level of activity is decreased, even after complete cessation of training.[19-28] In the large majority of studies, bone properties have been measured using dual-energy X-ray absorptiometry (DXA).[20, 24-28] Bone density measured by DXA is areal, can be confounded by differences in bone size, and cannot determine whether changes in areal bone mineral density (aBMD) are due to volumetric BMD (vBMD) or bone geometrical parameters.[29] Therefore, it is possible that studies which have observed bone mass by DXA reflect changes in bone size rather than changes in trabecular or cortical vBMD. The clinical importance of exercise-induced skeletal benefits could also be questioned if the benefits are not maintained into late adulthood, when fragility fractures occur. The mechanical strength of the bone and resistance against fracture have been reported to be dependent on bone size, volumetric density,[30, 31] and trabecular bone architecture.[32]

We have reported an association between exercises during youth and increased aBMD in old men.[20] However, using the DXA technique to measure aBMD, we could not determine whether the changes found in aBMD were caused by preserved BMD or bone geometrical parameters. The primary objective of this study was to investigate if exercise during growth and young adulthood was associated with cortical bone geometry, bone microstructure, and whole-bone strength in weight-bearing and non–weight-bearing bone in old men. The secondary objective was to assess if current physical activity was associated with these bone traits. We hypothesized that men who exercised at a high level during growth and young adulthood would have greater cortical bone size and whole-bone strength at old age in weight-bearing bone than men who were less physically active, and that this association would be independent of current physical activity.

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

Subjects

The study subjects were initially enrolled in the population-based Osteoporotic Fractures in Men (MrOS) study from Gothenburg, Sweden.[33] All study subjects in the original study were contacted and invited to participate in this 5-year follow-up study. Out of the original 1010 subjects, 597 men (79.9 ± 3.4 years of age) were included in the present study (Table 1). To be included in the original MrOS study, subjects had to be men between 69 to 81 years of age. All subjects were randomly sampled from the Swedish national population register for Gothenburg and invited to participate on a voluntary basis. To be eligible for the study, the subjects had to be able to walk indoors unaided and to understand questions and instructions in Swedish. To determine whether the cohort of the present study was representative of the initial population, we compared the age, height, and weight (all variables measured at the time of inclusion in the original MrOS study) of the included subjects (n = 597) with the subjects that did not participate (n = 413) in the present study. The included subjects were younger than the excluded subjects at the first MrOs examination (74.4 ± 2.9 versus 76.6 ± 3.2 years; p < 0.001), but at that occasion there were no significant differences between the included and excluded subjects in, height (175.9 ± 6.4 versus 175.5 ± 6.4 cm; p = 0.378), or weight (81.3 ± 11.7 versus 80.5 ± 13.0 kg; p = 0.317) (using an independent-samples t test (mean ± SD). The regional ethical review board at the University of Gothenburg approved the study. Written and oral informed consent was obtained from all study participants.

Table 1. Characteristics and Bone Macrostructure and Microstructure of the Total Cohort
Number of subjects597
  1. Values are given as mean ± SD.

  2. Ct.CSA = cortical cross-sectional area; Ct.Th = cortical thickness; PC = periosteal circumference; EC = endosteal circumference; D.Ct = volumetric cortical bone density; BV/TV = bone volume/total volume; Tb.N = trabecular number; Tb.Th = trabecular thickness; Tb.Sp = trabecular separation; Ct.Pm = cortical perimeter.

Age (years)79.9 ± 3.4
Height (cm)175.0 ± 6.5
Weight (kg)79.6 ± 11.7
Calcium intake (mg/d)958 ± 448
Smoking (%)5.0
Tibia pQCT at the diaphysis (n = 575) 
Ct.CSA (mm2)267.7 ± 33.6
Ct.Th (mm)4.23 ± 0.58
PC (mm)76.9 ± 4.4
EC (mm)50.3 ± 6.1
Tibia HR-pQCT at the metaphysis (n = 457) 
Failure load (kN)12.2 ± 2.2
Stiffness (kN/mm)240 ± 46
Percent load trabecular distal (%)64.0 ± 7.4
Percent load trabecular proximal (%)41.5 ± 8.0
Ct.CSA (mm2)120 ± 35
Ct.Th (mm)0.99 ± 0.31
Ct.Pm (mm)122 ± 9
D.Ct (mg/cm3)779 ± 75
BV/TV (%)14.9 ± 2.8
Tb.N (mm−1)1.97 ± 0.30
Tb.Th (µm)76.0 ± 11.6
Tb.Sp (µm)444 ± 83
Radius pQCT at the diaphysis (n = 590) 
Ct.CSA (mm2)93.0 ± 14.5
Ct.Th (mm)2.36 ± 0.39
PC (mm)47.0 ± 3.6
EC (mm)32.2 ± 4.6
Radius HR-pQCT at the metaphysis (n = 372) 
Failure load (kN)4.7 ± 1.1
Stiffness (kN/mm)91 ± 22
Percent load trabecular distal (%)64.0 ± 6.8
Percent load trabecular proximal (%)27.1 ± 7.7
Ct.CSA (mm2)54 ± 18
Ct.Th (mm)0.61 ± 0.21
Ct.Pm (mm)90 ± 8
D.Ct (mg/cm3)768 ± 77
BV/TV (%)13.8 ± 3.2
Tb.N (mm−1)2.06 ± 0.29
Tb.Th (µm)66.6 ± 10.6
Tb.Sp (µm)430 ± 92

Assessment of physical activity

A self-administered questionnaire, based on a validated physical activity questionnaire,[34] was used to collect information about patterns of physical exercise from the perspective of a lifetime. The original questionnaire designed for interview was adapted to suit the self-administered form. Briefly, this questionnaire divided the lifespan into five prior age periods: 10 to 20, 21 to 30, 31 to 50, 51 to 70, and ≥71 years of age. For each age period the participants were asked to identify the types of exercise and on what level they performed this exercise during each period. Bivariate correlations between the levels of exercise among the participants during the different age periods were tested using Pearson's coefficient of correlation (r). Correlations ranged from r = 0.15 to 0.71 (p < 0.001, respectively) with the highest correlation between the first to periods (10 to 20 years and 20 to 30 years of age) and the lowest correlation between the first and last period (10 to 20 years and ≥71 years of age). The identified levels of exercise during the first two periods (10 to 30 years of age) divided the subjects into four groups: level 1 = low level (hardly any exercise or exercise sporadically a couple of times per month); level 2 = moderate level (regular exercise but only part of the year); level 3 = high level (regular exercise on a fixed time every week); and level 4 = very high level (regular exercise on competitive level). A detailed description of the most common types of physical exercise and number of subjects participating in each type of activity is shown in Table 2. Current physical activity was assessed using the Physical Activity Scale for the Elderly (PASE), a validated self-reporting questionnaire designed to measure physical activity in individuals aged 65 years or older.[35] This scale comprises 12 items regarding physical activity during a 7-day time frame prior to the assessment, including walking outside; light, moderate, and strenuous sport and recreational activities; muscle strengthening; light and heavy housework, home repairs, lawn work, or yard care; caring for another person; and work for pay or as a volunteer. The PASE is a valid and reliable instrument that reflects the types of activities in which that older adults commonly participate.[35] The total PASE score was computed by multiplying the amount of time spent in each activity (hours per week) or participation in an activity (yes/no) by empirically derived weights and then summing the product for all 12 items. The obtained total PASE score was stratified in quartiles that divided the subjects into four groups for statistical analyses according to the levels of current physical activity.

Table 2. Number of Participants for Each Type of Physical Exercise During Growth and Young Adulthood
Type of exerciseSubjects (n)
  1. All activity with ≥10 participating subjects is shown. Subjects may have participated in several types of exercise.

Badminton18
Bandy20
Bicycling79
Cross-country skiing89
Gardening55
Gymnastics66
Handball74
Middle and long distance running72
Orienteering24
Soccer149
Swimming47
Table tennis12
Tennis22
Track and field34
Walking86

Assessment of cortical bone geometry

A peripheral quantitative computed tomography (pQCT) device (XCT-2000; Stratec Medizintechnik, Pforzheim, Germany) was used to scan the distal leg (tibia) and the distal arm (radius) of the nondominant leg and arm in 575 and 590 subjects, respectively. A 2-mm-thick single tomographic slice was scanned with a pixel size of 0.50 mm. The cortical cross-sectional area (CSA, mm2), endosteal and periosteal circumference (EC and PC, mm), and cortical thickness (mm) were measured using a scan through the diaphysis (at 25% of the bone length in the proximal direction of the distal end of the bone) of the radius and tibia. Tibia length was measured from the medial malleolus to the medial condyle of the tibia, and length of the forearm was defined as the distance from the olecranon to the ulna styloid process. The coefficients of variation (CVs) for the bone measurements used were obtained by three repeated measurements according to the standardized protocol, including repositioning between the scans, on one subject (male 79 years of age). The CVs ranged from 0.1% to 1.1% of the tibia (cortical CSA, 0.8%; EC, 0.8%; PC, <0.1%; and cortical thickness, 1.1%) and from 0.3% to 4.7% of the radius (cortical CSA, 0.5%; EC, 4.7%; PC, 1.9%; and cortical thickness, 3.4%).

Bone microarchitectural measurement

A high-resolution pQCT device (HR-pQCT) (XtremeCT; Scanco Medical AG, Zürich, Switzerland) was used to scan the metaphysis of the distal tibia and the metaphysis of the distal radius of the left leg and arm in 478 and 479 subjects, respectively. Anatomically formed carbon fiber shells, especially designed for each type of limb (Scanco Medical AG) were used to immobilize the subject's arm or leg during the scan. The measurements of the volume of interest (VOI) in the metaphysis of the distal tibia and radius were carried out according to a standardized protocol, as described.[36] Briefly, a reference line was manually placed at the center of the endplate of the distal tibia and distal radius. The first CT slice started 22.5 mm and 9.5 mm proximal to the reference line for the tibia and radius, respectively. A total of 110 parallel CT slices, with a nominal isotropic resolution of 82 µm, were obtained at each skeletal site, delivering a three-dimensional representation of approximately 9 mm section of both the tibia and radius in the proximal direction. At each skeletal site, the entire VOI was automatically separated into a cortical and a trabecular region. From this separation and by previously described methods to process the data,[37] we obtained cortical cross-sectional area (Ct.CSA, mm2), cortical thickness (Ct.Th, mm), cortical periosteal perimeter (Ct.Pm, mm), volumetric cortical bone density (D.Ct, mg/cm3), trabecular bone volume fraction (BV/TV, %), trabecular number (Tb.N, mm−1), trabecular thickness (Tb.Th, µm), and trabecular separation (Tb.Sp, µm).

The quality of the measurements on the tibia and radius were assessed by a five-item graded scale, recommended by the manufacturer (Scanco Medical AG), where 1 had the highest quality, 2 to 3 acceptable quality (included in the analyses), and grade 4 to 5 had unacceptable quality (excluded from the analyses) due to artifacts caused by inadequate limb fixation. A total of 22 measurements of the leg and 107 measurements of the arm were considered to have unacceptable quality (grade 4 and 5), leaving 456 subjects for further analyses of the tibia and 372 for further analyses of the radius. The CVs for the bone measurements used were obtained by three repeated measurements according to the standardized protocol, including repositioning between the scans, on two subjects (one male 43 years of age and one female 38 years of age). The CVs ranged from 0.1% to 1.6% of the tibia and from 0.3% to 3.9% of the radius, as described in detail.[38] The same device, software, and operator were used throughout the whole study.

Finite element analysis

Biomechanical properties of the bone were derived by finite element analysis (FEA) as described.[39] The FE models were created by Finite Element software from Scanco (version V5.11/FE-V01.15), incorporated into the manufacturer's analysis software. To summarize, cortical and trabecular bone, measured by an HR-pQCT device (XtremeCT), were first separated by a script provided in the software. Converting each voxel in the model to an equally-sized brick element created the FE models.[40] Both the cortical and trabecular elements were regarded as isotropic and linear-elastic, and according to the method established by Pistoia and colleagues,[41] a Young's modulus of 10 GPa and a Poisson ratio of 0.3 was used for all elements. In the FE simulation, uniaxial compression was applied in the longitudinal direction of the bone, at the radius, corresponding to a fall from standing on an outstretched hand, representing the type of trauma involved in Colles fracture.[42] Following the failure criterion established by Pistoia and colleagues,[41] failure load (N) was defined as the load at which at least 2% of the bone elements surpassed 7000 microstrain. The same failure criterion has been used in a study on adolescent boys.[43] The FEA simulations were performed in the same manner at both the radius and tibia. FEA-derived stiffness (kN/mm) and percentage of load carried by the trabecular bone at the distal and proximal surface of the VOI (percent load trabecular distal and percent load trabecular proximal, respectively) were also reported. All FEAs were done using the FE solver integrated in the manufacturer's analysis software (Scanco Medical AG), as described.[39] The CVs for the bone measurements used were obtained by three repeated measurements according to the standardized protocol, including repositioning between the scans, on two subjects (one male 43 years of age and one female 38 years of age). The CVs were 0.2% and 3.1% for failure load, 0.3% and 3.9% for stiffness, 3.0% and 0.9% for percent load trabecular distal, and 0.7% and 0.8% for percent load trabecular proximal at the tibia and radius, respectively.[39]

Assessment of CVs

Height and weight were measured using standardized equipment. The CV values were <1% for these measurements. A self-administered questionnaire was used to collect information about calcium and smoking habits (yes/no). Calcium intake (mg/d) was estimated from dairy product intake.

Statistical analysis

All data was analyzed using SPSS software, version 21.0 for Windows (IBM Corp. Armonk, NY). Differences in characteristics and bone parameters between subjects divided according to the level of physical exercise during growth and young adulthood (10 to 30 years of age), or the level of current physical activity (stratified total PASE score), were calculated using ANOVA or ANCOVA followed by least significant difference post hoc test for continuous variables and χ2 test was used for categorical variables (Tables 3 and 4). Weight was not normally distributed and was logarithmically transformed before entered into the ANCOVA model. Bivariate correlations were tested using Pearson's coefficient of correlation.

Table 3. Characteristics and Bone Macrostructure and Microstructure of the Cohort According to Levels of Exercise During Growth and Young Adulthood (10 to 30 Years of Age)
 Levels of exercise during growth and young adulthood (n = 597)
Level 1Level 2Level 3Level 4p (ANOVA)
  • Unadjusted characteristics and adjusted bone parameter values are given as mean ± SD. Bone parameters are adjusted for weight and levels of current physical activity (stratified PASE score). Differences between groups tested by ANOVA or ANCOVA followed by least significant difference post hoc test for continuous variables or by χ2 for categorical variables.

  • Level 1 = no or sporadic recreational exercise; Level 2 = seasonally intermittent recreational exercise; Level 3 = weekly regular recreational exercise; Level 4 = exercise at competitive level; PASE = Physical Activity Scale for the Elderly; Ct.Th = cortical thickness; Ct.CSA = cortical cross-sectional area; Ct.Pm = cortical perimeter; D.Ct = volumetric cortical bone density; BV/TV = bone volume/total volume; Tb.N = trabecular number; Tb.Th = trabecular thickness; Tb.Sp = trabecular separation; PC = periosteal circumference; EC = endosteal circumference.

  • a

    p < 0.05 versus Level 1.

  • b

    p < 0.05 versus Level 2.

  •  cp < 0.05 versus Level 3.

  • d

    p < 0.01 versus Level 1.

  • e

    p < 0.001 versus Level 1.

Subjects (n)363746694 
Age (years)80.0 ± 3.480.0 ± 3.480.3 ± 3.779.4 ± 3.30.332
Height (cm)174.9 ± 6.6174.2 ± 5.8176.0 ± 6.2175.2 ± 6.50.423
Weight (kg)78.9 ± 11.378.4 ± 11.882.9 ± 12.0a,b81.2 ± 12.60.026
Calcium intake (mg/d)962 ± 463943 ± 4371018 ± 441914 ± 4110.537
Smoking (%)5.51.43.07.6
Current physical activity (PASE score)354 ± 225406 ± 221396 ± 255381 ± 1980.184
     p (ANCOVA)
Tibia pQCT at the diaphysis (n = 572)     
Ct.Th (mm)4.16 ± 0.544.26 ± 0.554.30 ± 0.614.39 ± 0.59e<0.001
Tibia HR-pQCT at the metaphysis (n = 451)     
Failure load (kN)11.9 ± 2.012.3 ± 1.812.2 ± 1.812.8 ± 2.0e0.005
Stiffness (kN/mm)235 ± 42244 ± 39241 ± 38254 ± 41e0.007
Percent load trabecular distal (%)64.1 ± 7.464.6 ± 7.363.3 ± 7.364.1 ± 7.00.827
Percent load trabecular proximal (%)41.5 ± 7.942.2 ± 7.840.4 ± 8.041.6 ± 7.50.691
Ct.CSA (mm2)117 ± 34119 ± 34122 ± 38126 ± 320.172
Ct.Th (mm)0.97 ± 0.311.00 ± 0.301.01 ± 0.331.03 ± 0.280.364
Ct.Pm (mm)122 ± 8121 ± 8122 ± 9123 ± 80.574
D.Ct (mg/cm3)775 ± 75777 ± 75779 ± 82787 ± 640.671
BV/TV (%)14.7 ± 2.915.4 ± 2.614.8 ± 2.115.6 ± 2.7d0.028
Tb.N (mm−1)1.97 ± 0.291.99 ± 0.281.98 ± 0.262.01 ± 0.300.753
Tb.Th (µm)74.6 ± 11.378.1 ± 11.1a75.5 ± 10.778.0 ± 10.9a0.030
Tb.Sp (µm)446 ± 80439 ± 76442 ± 72430 ± 770.484
Radius pQCT at the diaphysis (n = 587)     
Ct.CSA (mm2)92.7 ± 12.491.7 ± 13.094.1 ± 14.693.8 ± 14.90.648
Ct.Th (mm)2.37 ± 0.372.31 ± 0.362.37 ± 0.402.37 ± 0.390.713
PC (mm)46.9 ± 3.547.1 ± 3.347.4 ± 3.547.1 ± 3.20.738
EC (mm)32.1 ± 4.732.6 ± 4.332.5 ± 4.632.2 ± 4.10.778
Radius HR-pQCT at the metaphysis (n = 369)     
Failure load (kN)4.6 ± 1.04.7 ± 0.84.7 ± 1.24.8 ± 1.00.778
Stiffness (kN/mm)90 ± 2192 ± 1791 ± 2393 ± 210.779
Percent load trabecular distal (%)64.1 ± 6.763.9 ± 6.364.0 ± 7.964.1 ± 7.00.999
Percent load trabecular proximal (%)27.4 ± 7.426.6 ± 7.426.2 ± 9.027.9 ± 7.70.657
Ct.CSA (mm2)53 ± 1756 ± 1556 ± 2055 ± 170.615
Ct.Th (mm)0.59 ± 0.210.63 ± 0.180.63 ± 0.240.62 ± 0.210.530
Ct.Pm (mm)90 ± 789 ± 789 ± 991 ± 70.372
D.Ct (mg/cm3)764 ± 76778 ± 64771 ± 93766 ± 690.725
BV/TV (%)13.6 ± 3.114.3 ± 2.913.8 ± 3.114.3 ± 3.00.319
Tb.N (mm−1)2.05 ± 0.282.04 ± 0.272.07 ± 0.302.15 ± 0.280.130
Tb.Th (µm)65.8 ± 10.570.1 ± 12.166.5 ± 8.966.5 ± 10.30.092
Tb.Sp (µm)434 ± 95433 ± 83430 ± 91408 ± 700.285
Table 4. Characteristics and Bone Macrostructure and Microstructure of the Cohort According to Level of Current Physical Activity
 Level of current physical activity (stratified PASE score) (n = 596)
Group 1Group 2Group 3Group 4p (ANOVA)
  1. Unadjusted characteristics and adjusted bone parameter values are given as mean ± SD. Bone parameters are adjusted for age, smoking, and levels of exercise during growth and young adulthood. Differences between groups tested by ANOVA or ANCOVA followed by least significant difference post hoc test for continuous variables or by χ2 for categorical variables. Subjects in the four groups (Group 1, Group 2, Group 3, Group 4) were determined by quartiles of the total PASE score.

  2. PASE = Physical Activity Scale for the Elderly; Ct.Th = cortical thickness; Ct.CSA = cortical cross-sectional area; Ct.Pm = cortical perimeter; D.Ct = volumetric cortical bone density; BV/TV = bone volume/total volume; Tb.N = trabecular number; Tb.Th = trabecular thickness; Tb.Sp = trabecular separation; PC = periosteal circumference; EC = endosteal circumference.

  3. a

    p < 0.05 versus Group 1.

  4. b

    p < 0.05 versus Group 2.

  5. c

    p < 0.01 versus Group 1.

  6. d

    p < 0.001 versus Group 1.

  7. e

    p < 0.001 versus Group 2.

Subjects (n)149149149149 
Age (years)80.5 ± 3.580.2 ± 3.579.9 ± 3.579.2 ± 3.0a,b0.009
Height (cm)174.5 ± 6.6175.4 ± 7.2175.7 ± 6.0174.3 ± 5.90.206
Weight (kg)80.9 ± 12.879.8 ± 12.380.0 ± 10.877.9 ± 10.70.175
Calcium intake (mg/d)910 ± 437995 ± 497945 ± 398972 ± 4590.400
Smoking (%)4.68.94.42.2b
Current physical activity (PASE score)140 ± 51268 ± 34390 ± 43679 ± 190
     p (ANCOVA)
Tibia pQCT at the diaphysis (n = 571)     
Ct.Th (mm)4.13 ± 0.574.21 ± 0.604.22 ± 0.524.34 ± 0.57b,c0.015
Tibia HR-pQCT at the metaphysis (n = 451)     
Failure load (kN)11.8 ± 1.912.0 ± 2.212.3 ± 2.212.4 ± 2.10.129
Stiffness (kN/mm)233 ± 40235 ± 46244 ± 45246 ± 430.082
Percent load trabecular distal (%)64.4 ± 7.764.4 ± 7.764.4 ± 7.163.0 ± 6.60.338
Percent load trabecular proximal (%)42.0 ± 8.441.8 ± 8.741.9 ± 7.740.2 ± 6.80.245
Ct.CSA (mm2)111 ± 36116 ± 35121 ± 30a127 ± 34b,d0.003
Ct.Th (mm)0.92 ± 0.330.96 ± 0.310.99 ± 0.261.06 ± 0.30b,d0.005
Ct.Pm (mm)123 ± 10122 ± 10123 ± 9121 ± 90.541
D.Ct (mg/cm3)761 ± 87768 ± 69783 ± 69a794 ± 64b,d0.003
BV/TV (%)14.8 ± 2.614.9 ± 3.114.9 ± 2.615.0 ± 2.80.959
Tb.N (mm−1)2.03 ± 0.302.01 ± 0.321.96 ± 0.311.93 ± 0.290.050
Tb.Th (µm)73.6 ± 11.174.0 ± 10.576.5 ± 10.6a78.2 ± 11.9d,e0.004
Tb.Sp (µm)430 ± 71437 ± 85447 ± 84453 ± 870.137
Radius pQCT at the diaphysis (n = 586)     
Ct.CSA (mm2)92.4 ± 13.692.6 ± 15.093.8 ± 14.593.3 ± 13.60.846
Ct.Th (mm)2.33 ± 0.372.36 ± 0.402.37 ± 0.362.39 ± 0.370.543
PC (mm)47.3 ± 3.747.0 ± 3.547.2 ± 3.946.8 ± 3.40.595
EC (mm)32.7 ± 4.632.1 ± 4.532.3 ± 4.731.7 ± 4.30.374
Radius HR-pQCT at the metaphysis (n = 368)     
Failure load (kN)4.6 ± 1.14.6 ± 1.24.7 ± 1.04.7 ± 1.00.913
Stiffness (kN/mm)89 ± 2389 ± 2590 ± 2191 ± 200.902
Percent load trabecular distal (%)64.5 ± 7.164.0 ± 7.063.9 ± 6.864.1 ± 6.70.943
Percent load trabecular proximal (%)28.7 ± 7.027.8 ± 8.326.7 ± 7.426.2 ± 7.70.151
Ct.CSA (mm2)52 ± 1752 ± 2055 ± 1655 ± 170.467
Ct.Th (mm)0.58 ± 0.190.59 ± 0.240.61 ± 0.170.62 ± 0.220.515
Ct.Pm (mm)90 ± 890 ± 890 ± 890 ± 80.976
D.Ct (mg/cm3)758 ± 66758 ± 88775 ± 65769 ± 760.314
BV/TV (%)13.9 ± 3.113.4 ± 3.613.7 ± 2.814.0 ± 3.00.565
Tb.N (mm−1)2.08 ± 0.282.02 ± 0.332.09 ± 0.252.07 ± 0.290.383
Tb.Th (µm)66.6 ± 10.365.7 ± 11.065.1 ± 9.567.5 ± 9.10.348
Tb.Sp (µm)423 ± 77447 ± 130420 ± 67426 ± 800.200

Results

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

Characteristics of the total cohort and the cohort divided according to the level of exercise during growth and young adulthood (henceforth also reported as the level of exercise early in life) or the level of current physical activity (stratified PASE score) are shown in Tables 1, 3, and 4, respectively.

Correlation between cortical bone variables at the diaphysis and the metaphysis

Bivariate correlation were used to analyze the correlation between the corresponding cortical bone variables measured with the pQCT at the diaphysis and HR-pQCT at the metaphysis of the tibia and the radius. All the corresponding cortical bone variables at the diaphysis and metaphysis of the tibia and the radius were correlated; Ct.CSA r = 0.63 (p < 0.001) and r = 0.69 (p < 0.001), Ct.Th r = 0.71 (p < 0.001) and r = 0.63 (p < 0.001), and PC/perimeter r = 0.75 (p < 0.001) and r = 0.60 (p < 0.001), respectively.

Associations between exercise early in life and bone geometry, structure, and strength

There were no significant differences in age, height, daily calcium intake, smoking, or level of current physical activity between the four subgroups divided by level of exercise during growth and young adulthood (Table 3). However, subjects in the subgroup with the second-highest level of exercise early in life (level 3) had higher body weight than subjects with the lower level of exercise (levels 1 and 2) (Table 3).

In order to determine if the level of exercise during growth and young adulthood was independently associated with bone parameters in both tibia and radius, ANCOVA was used. Analyses by ANCOVA, adjusted for weight and levels of current physical activity (stratified PASE score), revealed that men in the group with the highest level of exercise early in life (level 4, exercise at competitive level) had greater tibial Ct.CSA (6.3%), PC (1.6%), and Ct.Th (5.4%) at the diaphysis than men in the subgroup with the lowest level of exercise during growth and young adulthood (level 1) (Fig. 1A, B; Table 3). Using HR-pQCT, we observed that men in the group with the highest level of exercise early in life (level 4) had higher estimated bone strength (failure load: 7.5%, stiffness: 7.8%) and trabecular bone volume fraction (BV/TV) (6.5%) at the metaphysis than men in the subgroup with the lowest level of exercise during growth and young adulthood (level 1); this was also true when adjusted for covariates and current physical activity (Table 3). Furthermore, exercise early in life was associated with trabecular thickness (Tb.Th) at the metaphysis (Table 3).

image

Figure 1. Level of physical exercise during growth and young adulthood associated with cortical bone size, cortical cross-sectional area (A) and periosteal circumference (B), in weight-bearing bone (tibia) in elderly men. Subjects are divided into groups according to level of exercise early in life (10 to 30 years of age): level 1 (no or sporadic recreational exercise, n = 347), level 2 (seasonally intermittent recreational exercise, n = 72), level 3 (weekly regular recreational exercise, n = 63), and level 4 (exercise at competitive level, n = 90). Values adjusted for weight and levels of current physical activity (stratified PASE score) are given as mean ± 95% CI. Differences between groups tested by ANCOVA followed by least significant difference post hoc test. PASE = Physical Activity Scale for the Elderly.

Download figure to PowerPoint

Level of exercise during growth and young adulthood was not associated with any measured bone trait at the radius (Table 3).

Associations between current physical activity and bone geometry, structure, and strength

There were no significant differences in height, weight, or daily calcium intake between the four subgroups divided by level of current physical activity (Table 4). Subjects in the subgroup with the highest level of current physical activity (group 4) were younger than subjects in the subgroups with the lowest level of activity (groups 1 and 2) (Table 4). Furthermore, subjects in group 4 were less frequently smokers than subjects in group 2 (Table 4).

In order to determine if level of current physical activity was independently associated with bone parameters in both tibia and radius, ANCOVA was used. Analyses by ANCOVA, adjusted for age, smoking, and levels of exercise during growth and young adulthood, revealed that level of current physical activity (stratified PASE score) was independently associated with tibial cortical thickness and endosteal circumference at the diaphysis (Table 4, Fig. 2C). Subjects in the group with the highest level of current physical activity (group 4, stratified PASE score) had smaller tibial EC (3.6%) at the diaphysis than subjects in the subgroup with the lowest stratified PASE score (group 1) (Fig. 2C). Using HR-pQCT we observed that the level of current physical activity was associated with cortical bone geometry and structure (Ct.CSA, Ct.Th, and D.Ct), and Tb.Th at the metaphysis, and also when adjusting for covariates and exercise early in life (Table 4).

image

Figure 2. Level of current physical activity was associated with endosteal circumference (C), in weight-bearing bone (tibia) in elderly men. Subjects in the different groups were divided by quartiles based on total PASE score; group 1 with the lowest total PASE score (PASE score 140 ± 51, n = 138), group 2 (PASE score 268 ± 34, n = 142), group 3 (PASE score 390 ± 43, n = 147), and group 4 with the highest total PASE score (PASE score 679 ± 190, n = 144). Values adjusted for age, smoking, and levels of exercise during growth and young adulthood are given as mean ± 95% CI. Differences between groups tested by ANCOVA followed by least significant difference post hoc test. PASE = Physical Activity Scale for the Elderly.

Download figure to PowerPoint

The level of current physical activity was not associated with any measured bone trait at the radius (Table 4).

Discussion

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

The role of regular weight-bearing exercise in the prevention of osteoporosis remains controversial because the long-term effects of exercise on bone structure, size, and strength have not been well assessed. Our results indicate that physical activity can influence cortical bone structure in different ways at different stages throughout life in men. We showed that exercise during growth and young adulthood was independently associated with cortical bone size, in terms of greater periosteal circumference and cortical thickness at the tibial diaphysis, in 80-year-old Swedish men. In addition, we found that current physical activity was independently associated with cortical bone structure, in terms of lower endosteal circumference and greater cortical thickness at the tibial diaphysis, in the same population.

The results of the present study are consistent with previous findings in populations of both young adult men and women, where association between previous physical activity during childhood and adolescence and cortical bone size were found.[21, 44, 45] Consistent with our previous results from a cohort of young men, the group with the lowest level of previous exercise had smaller cortical bone size, indicating that only high amounts of exercise can enhance cortical bone structure.[21] In a thorough study (161 men between 50 and 85 years of age),[46] associations were found between parameters of cortical bone size of the mid-femur in men aged over 50 years and lifetime as well as mid-adulthood (19 to 50 years) physical activity. However, this study did not find any association between adolescent sport and leisure activity participation and bone structure and strength in old men. Furthermore, in this previous study there was no evidence that long-term participation in weight-bearing physical activities could reduce bone loss at the endocortical surface.[46] However, experimental studies have argued for lifelong benefits in cortical bone properties from skeletal loading when young.[47, 48]

Cortical bone size is an important determinant of bone strength and resistance against fracture.[30, 31] Because the resistance of bone to bending and torsion forces is related exponentially to its diameter, even a small difference in the outer circumference could make a substantial contribution to its strength and resistance to fracture.[30] In the present study, we found that exercise during growth and young adulthood (age 10 to 30 years) was related to augmented cortical bone size (CSA) via actions on the outer cortical envelope (periosteal circumference) at old age, indicating attained benefits of physical activity in enlarging the cortical shell, even though physical activity has decreased later in life. In postmenopausal women, each SD decrease in cortical CSA has been associated with a 3.6-fold increased prevalent fractures at the radius.[49] If these results could be translated to elderly men in this study, where the differences in tibia CSA between subjects in the group with the highest level of exercise and men in the group with the lowest level of activity early in life are equal to 0.6 SD, previous physical activity could result in more than a halved risk of future fracture at old age. Previous physical activity was associated with cortical bone size owing to increased periosteal circumference/perimeter at the diaphysis but not at the metaphysis of the tibia. Current physical activity was associated with cortical CSA and thickness at the metaphysis and endosteal circumference at the diaphysis but not periosteal perimeter/circumference at either the metaphysis or the diaphysis of the tibia. The constitution of the skeleton is clearly different at the different bone sites we have measured using pQCT and HR-pQCT. The diaphysis measured by pQCT (at 25% of the bone length) almost exclusively contains cortical bone, whereas the metaphysis measured by HR-pQCT mainly contains trabecular bone and a thin cortex. By using HR-pQCT, we could reveal that level of exercise during growth and young adulthood but not current physical activity was independently associated with estimated bone strength (failure load and stiffness) and trabecular bone volume fraction in weight-bearing bone. However, we found no association between level of exercise during growth and young adulthood and FEA-derived percentage of load carried by the trabecular bone (Table 3), indicating that both cortical and trabecular bone structure contribute to the association found with bone strength at this bone site. Using bivariate correlations between the corresponding measurements at the metaphysis and diaphysis, we found that the various measurements of the cortex at the different bone sites were only moderately correlated (r = 0.63 to r = 0.75). Although we were not able to calculate FEA-derived bone strength at the diaphysis, cortical bone most likely contributes substantially more to the total bone strength at the diaphysis (25% of the bone length) than at the metaphysis.

There are limitations associated with the present study. The results from the present study are derived from investigations on 80-year-old men, and therefore may not be applicable to women and other age groups. Present and former physical activity habits were assessed using a retrospective self-reporting questionnaire, which could have limited the ability of the subjects to recall their history of physical activity and cause bias and misclassification. However, by using a self-administered questionnaire, based on a validated questionnaire concerning physical activity habits in a lifetime perspective,[34] we believe that we have increased our possibilities to collect accurate information about physical activity habits. We also assessed calcium intake using a questionnaire, but our questions were focused on calcium intake from dairy products, which have been shown to be the source of about 65% of the daily calcium intake.[50] Thus, the calculated calcium intake in our population is most likely underestimated. In addition, we have not been able to control for supplements of either calcium or other minerals or vitamins, which could have confounded the presented analyses. The cross-sectional design does not allow for direct cause-effect relationships to be established. In cross-sectional studies, such as the present study, there is always a risk of selection bias; ie, men who are bigger and stronger could be more likely to be more successful in and participate to a higher extent in exercise. However, we could not find any difference in height between subjects in the different groups with different levels of exercise early in life. Subjects in the group with the second highest level of exercise early in life had higher body weight. Thus, we controlled for body weight in ANCOVA analyses and found that associations between exercise early in life and bone parameters remained. Patterns of physical activity earlier in life may also influence patterns of physical activity later in life. However, we did not find any significant difference in level of current physical activity between the different groups of subjects with different levels of exercise during growth and young adulthood.

In this large cohort of elderly men, exercise during growth and young adulthood was independently associated with larger periosteal circumference, and higher bone strength, whereas current exercise was independently associated with lower endosteal expansion at the tibia. These findings indicate that exercise during growth and young adulthood can increase the cortical bone size via periosteal expansion, whereas exercise in old age can decrease endosteal bone loss in weight-bearing bone in old men.

Acknowledgments

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

This work was supported by the Swedish Research Council, the Swedish Foundation for Strategic Research, European Commission, the Lundberg Foundation, the Torsten and Ragnar Söderberg's Foundation, Petrus and Augusta Hedlund's Foundation, the ALF/LUA grant from the Sahlgrenska University Hospital, the Novo Nordisk Foundation, and Gustaf V:s och Drottning Victorias Frimurarstiftelse.

Authors' roles: Study design: ML, MN, DM, MK, and CO. Study conduct: ML, DM, and MN. Data collection: ML, MN, DM, and DS. Data analysis: MN and ML. Data interpretation: MN and ML. Drafting manuscript: MN and ML. Revising manuscript content: ML, MN, DM, CO, MK, and DS. Approving final version of manuscript: ML, MN, DM, CO, MK, and DS. MN and ML take 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