Association of Amount of Physical Activity With Cortical Bone Size and Trabecular Volumetric BMD in Young Adult Men: The GOOD Study

Authors

  • Mattias Lorentzon MD, PhD,

    Corresponding author
    1. Center for Bone Research at the Sahlgrenska Academy (CBS), Department of Internal Medicine, Gothenburg University, Gothenburg, Sweden
    • Division of Endocrinology, Department of Internal Medicine, Gröna Stråket 8, Sahlgrenska University Hospital, 413 45 Gothenburg, Sweden
    Search for more papers by this author
  • Dan Mellström,

    1. Department of Geriatric Medicine, University of Gothenburg, Gothenburg, Sweden
    Search for more papers by this author
  • Claes Ohlsson

    1. Center for Bone Research at the Sahlgrenska Academy (CBS), Department of Internal Medicine, Gothenburg University, Gothenburg, Sweden
    Search for more papers by this author

  • The authors have no conflict of interest.

Abstract

In this population-based study, amount of PA was associated with cortical bone size (increased thickness and periosteal circumference) and trabecular vBMD, but not with cortical vBMD or length of the long bones in young men. The lowest effective amount of PA was ≥4 h/week.

Introduction: Physical activity (PA) is believed to have positive effects on the skeleton and possibly help in preventing the occurrence of osteoporosis. Neither the lowest effective amount of PA needed to induce an osteogenic response nor its effect on the BMD and size of the different bone compartments (i.e., trabecular and cortical bone) has yet been clarified.

Materials and Methods: In this population-based study, we investigated the amount of all types of PA in relation to areal BMD (aBMD), trabecular and cortical volumetric BMD (vBMD), and cortical bone size in 1068 men (age, 18.9 ± 0.02 years), included in the Gothenburg Osteoporosis and Obesity Determinants (GOOD) study. aBMD was measured by DXA, whereas cortical and trabecular vBMD and bone size were measured by pQCT.

Results and Conclusions: The amount of PA was associated with aBMD of the total body, radius, femoral neck, and lumbar spine, as well as with cortical bone size (increased thickness and periosteal circumference) and trabecular vBMD, but not with cortical vBMD or length of the long bones. The lowest effective amount of PA was ≥4 h/week. aBMD, cortical bone size, and trabecular vBMD were higher in subjects who started their training before age 13 than in subjects who started their training later in life. Our data indicate that ≥4 h/week of PA is required to increase bone mass in young men and that exercise before and during the pubertal growth is of importance. These findings suggest that PA is imperative for the augmentation of cortical bone size and trabecular vBMD but does not affect the cortical vBMD in young men.

INTRODUCTION

OSTEOPOROSIS-RELATED FRACTURES constitute a major public health concern in women as well as in men.(1) The fracture risk is dependent on several factors, including muscle strength, BMD, and bone geometry.(2,3) Each SD decrease in BMD has been associated with about a 2-fold increase in the age-adjusted hip fracture risk in postmenopausal women(4) and with a 3-fold risk increase in elderly men.(5) The maximal attained bone mass in life, peak bone mass (PBM), is dependent primarily on genetic factors but also on environmental factors such as physical activity (PA) and calcium intake.(6) PBM has been estimated to account for about one-half of the BMD variation up to 65 years of age and may influence the risk of future osteoporosis.(7,8) Several cross-sectional studies have shown higher areal BMD (aBMD) in physically active subjects than in controls, and some randomized intervention studies have revealed a positive effect of PA on BMD.(9–14)

Most previous studies of adequate sample size have been limited by the use of 2-D measurements (aBMD) of the bone, using DXA, which does not provide information about volumetric BMD (vBMD) and size of the different bone compartments (i.e., cortical and trabecular bone). Therefore, because of the limitations of the DXA methodology, it remains unclear whether the presumed effect of PA on bone mass is caused by an increase in vBMD or bone enlargement of the different bone compartments. To measure vBMD and bone size of the cortical and trabecular bone, a 3-D technique, such as pQCT, must be used. The few available reports in which this technique has been used suggest that PA is primarily associated with the cortical bone size and not vBMD, whereas the relationship between PA and trabecular vBMD remains uncertain.(15–17) Although PA is believed to increase bone mass,(18) and possibly help prevent bone loss and subsequent osteoporosis, the minimal amount of PA to induce higher bone mass in men remains insufficiently studied.(19) Information about the lowest effective amount of PA needed to induce an osteogenic response could serve as the foundation of lifestyle recommendations issued to the general population to optimize bone health.

In this population-based study, an association of PA amount with cortical bone size and trabecular vBMD is shown in an extensively phenotyped cohort of 1068 18- to 19-year-old Swedish men.

MATERIALS AND METHODS

Subjects

The Gothenburg Osteoporosis and Obesity Determinants (GOOD) study was initiated with the aim to determine both environmental and genetic factors involved in the regulation of bone and fat mass. Study subjects were randomly identified using national population registers, contacted by telephone, and asked to participate in this study. A total of 1068 men, 18.9 ± 0.02 years of age, from the greater Gothenburg area were included. To be included in the GOOD study, subjects had to be >18 and <20 years of age and willing to participate in the study. There were no other exclusion criteria; 48.6% of the contacted study subject candidates agreed to participate and were included in this study. A standardized questionnaire was used to collect information about amount of present PA (hours per week), age of PA start (years), duration of present PA (years), nutritional intake (dairy products), and smoking. Calcium intake was estimated from dairy product intake and semiquantitated into quintiles.

The GOOD study was approved by the ethics committee at Gothenburg University. Written and oral informed consent was obtained from all study participants. To determine whether the GOOD cohort was representative of the general young male population in Gothenburg, we compared the height and weight of the GOOD subjects with a group of 624 age-matched, randomly selected conscripts (conscripts represent 86% of all men at this age in the Gothenburg area) living in the same area as the GOOD subjects. There was no difference in height, weight, or BMI (using an independent-samples t-test) between these two cohorts, showing that the GOOD cohort is representative of the general young male population of Gothenburg.

PA

A total of 678 subjects participated in PA (physically active), whereas 390 subjects did not participate in any PA (sedentary). Physically active subjects were divided into quartiles according to the amount (hours per week) trained. Quartiles were as follows: first, >0 and <4 h/week (n = 164); second, ≥4 and <6 h/week (n = 157); third, ≥6 and <8.5 h/week (n = 184); fourth, ≥8.5 h/week (n = 173). In Swedish boys, peak height velocity occurs at the average age of 13.5 years and commences within 2 years after the start of puberty.(20) Among the training subjects, 288 began their present PA before age 13, and 390 began their present PA at age 13 or later. PA type (strain) was categorized according to peak strain score, as previously described.(21) Activities with jumping actions (e.g., basketball) were assigned a strain score of 3; activities involving repetitive sprinting and turning (e.g., soccer) were given a strain score of 2; all other types of weight-bearing PAs were assigned a strain score of 1, and non-weight-bearing PA (e.g., swimming) were assigned a strain score of 0.

Anthropometrical measurements

Height and weight were measured using standardized equipment. The CV values were <1% for these measurements.

DXA

aBMD (g/cm2) of the whole body, femoral neck (of the left leg), lumbar spine, and the left and right radius were assessed using the Lunar Prodigy DXA (GE Lunar, Madison, WI, USA). The CVs for the aBMD measurements ranged from 0.5% to 3%, depending on application.

pQCT

A 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, respectively. A 2-mm-thick single tomographic slice was scanned with a voxel size of 0.50 mm. The cortical vBMD (not including the bone marrow; mg/cm3), cortical BMC (mg/mm), cortical cross-sectional area (CSA, mm2), endosteal and periosteal circumference (EC and PC), 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. Trabecular vBMD was measured using a scan through the metaphysis (at 4% of the bone length in the proximal direction of the distal end of the bone) of these bones. 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 CVs were <1% for all pQCT measurements.

Statistical analysis

Differences in anthropometric characteristics and bone parameters were studied using an independent samples t-test or ANOVA with Bonferroni's correction for multiple comparisons. The independent predictors of the various bone parameters were tested using multiple linear regression analysis. The percentage of the variation, of each bone parameter, explained (R2) by amount of PA alone and by amount of PA together with all covariates was calculated, using the linear regression model. χ2 tests were used to determine whether the distribution of smokers differed between the physically active and sedentary groups. The side-to-side evaluation between the radius of the dominant and nondominant arm was calculated using a paired t-test. To determine whether the differences in aBMD, between fourth quartile and sedentary subjects, were greater for the femoral neck than for the radius, delta values (Δ) for each bone site were calculated. Delta values for the aBMD of each bone site were calculated as percent differences between average aBMD values of the sedentary subjects and the aBMD values of the fourth quartile subjects. Differences in Δ values between aBMD of the radius and femoral neck were calculated using a paired samples t-test. This procedure was also used to determine if the differences between fourth quartile and sedentary subjects were greater for the tibia CSA than for the radius CSA. A p value <0.05 was considered significant.

RESULTS

Anthropometric characteristics

The 678 physically active men (training, 6.9 ± 0.20 [SE] h/week) did not differ from the 390 sedentary men in height (181.4 ± 0.3 versus 181.5 ± 0.3 cm), body weight (74.3 ± 0.4 versus 73.1 ± 0.7 kg), or BMI (22.5 ± 0.1 versus 22.2 ± 0.2 kg/m2). The physically active subjects had a greater daily calcium intake (1162 ± 29 versus 980.0 ± 34 mg; p < 0.0001) and smoked less (4.9% versus 14.8%; p < 0.001) than the sedentary subjects. Physically active subjects were divided into quartiles according to amount of PA. Third and fourth quartile subjects were slightly younger and had a greater calcium intake than the sedentary subjects (Table 1). Subjects in higher quartiles of amount of PA were associated with lower age of PA onset than subjects in lower quartiles of amount of PA.

Table Table 1.. Anthropometric Characteristics of the Entire GOOD Cohort (A) and According to Quartiles of Amount of Present PA (B)
original image

PA amount association with aBMD

The differences in aBMD between sedentary and physically active subjects, divided into quartiles according to amount of PA (hours per week), were studied using ANOVA (Fig. 1). Subjects in the first quartile (training >0 and <4 h/week) did not have significantly higher aBMD at any site measured, whereas subjects in the fourth quartile (training ≥ 8.5 h/week) had higher aBMD of the total body (6.4%), lumbar spine (10.5%), femoral neck (14.0%), and radius (3.0%) than their sedentary counterparts. The percent difference between fourth quartile and sedentary subjects was greater for the femoral neck (Δ = 14.0 ± 1.2%) than for the radius (Δ = 3.0 ± 0.7%; p < 0.0001). aBMD of the radius of the dominant arm was higher than the aBMD of the nondominant arm (0.59 ± 0.0 versus 0.58 ± 0.0 g/cm2; p < 0.0001).

Figure FIG. 1..

Association between PA amount and aBMD of the (A) total body, (B) lumbar spine, (C) femoral neck, and (D) radius. Subjects were divided into sedentary and quartiles of PA, according to hours per week trained. A > sedentary, B > first quartile, and C > second quartile. Capital and lowercase letters represent p < 0.01 and p < 0.05, respectively.

The independent predictors of aBMD were calculated using a multiple linear regression model including age, weight, height, smoking, calcium intake, age of PA onset, and present amount of PA (Table 2). PA amount was found to be an independent predictor of aBMD of the total body, lumbar spine, femoral neck, and radius.

Table Table 2.. Amount of Present Physical Activity as an Independent Predictor of aBMD, Cortical Bone Size and Trabecular Volumetric BMD
original image

Association of PA amount with cortical bone size and trabecular vBMD

To evaluate the association between PA and vBMD and size of the different bone compartments (i.e., trabecular and cortical bone), pQCT measurements were used. The differences in cortical bone parameters and trabecular vBMD between sedentary and physically active subjects, divided into quartiles according to amount of PA (hours per week), were studied (Figs. 2A-2G). In general, subjects in quartiles with high amounts of PA (third and fourth) had greater cortical bone size (greater cortical thickness, caused by increased PC but not EC) and higher trabecular vBMD, but not cortical vBMD, than the sedentary men. Subjects in the first quartile (training <4 h/week) did not have significantly higher cortical bone size or trabecular vBMD of the tibia than sedentary individuals. The largest differences in cortical BMC (10.3%) and CSA (10.5%) were seen between the fourth quartile and the sedentary cohort (Figs. 2A and 2C). Cortical vBMD of the tibia was not different between sedentary subjects and men in the various quartiles (Fig. 2B). Trabecular vBMD of the tibia was higher in second, third, and fourth quartile subjects, than in nontraining subjects (Fig. 2G). Similar results were obtained for the radius (data not shown).

Figure FIG. 2..

Association between PA amount and (A) cortical BMC, (C) cortical cross-sectional area, (D) cortical thickness, (E) periosteal circumference, and (G) trabecular vBMD, but not with (B) cortical vBMD or (F) endosteal circumference of the tibia. Subjects were divided into sedentary and quartiles of PA, according to hours per week trained. A > sedentary, B > first quartile, and C > second quartile. Capital and lowercase letters represent p < 0.01 and p < 0.05, respectively.

The percent difference between fourth quartile and sedentary subjects was greater for the cortical CSA of the tibia (Δ = 10.5 ± 1.0%) than of the radius (Δ = 6.3 ± 1.0%; p < 0.0001).

The independent predictors of cortical and trabecular bone parameters, as measured using pQCT, were studied using linear regression, including age, weight, height, smoking, calcium intake, age of PA onset, and present amount of PA (Table 2). PA amount was a strong independent predictor of cortical BMC, CSA, thickness, PC, and trabecular vBMD of both the radius and tibia (Table 2).

PA amount alone could explain 12.7% of the total variation in aBMD of the femoral neck, 10.8% of the aBMD of the lumbar spine, and 10.0% of the tibia CSA. When the other covariates (age, height, weight, calcium intake, smoking, age of PA onset) were included together with amount of PA, the whole regression model could explain 26.3% of the aBMD of the femoral neck, 22.3% of the aBMD of the lumbar spine, and 39.6% of the tibia CSA.

Association of PA type (strain) with aBMD, cortical bone size, and trabecular vBMD

A detailed description of the different types of PA, the average amount of hours per week of each PA, the peak strain score, and the number of subjects participating in each type of PA is shown in Table 3.

Table Table 3.. Physical Activity (PA) Type, Present Amount of PA (h/week), Strain Score, and Number of Subjects Participating in the Type of PA
original image

The peak strain score was 1.3 ± 0.1 for the first quartile, 1.5 ± 0.1 for the second quartile, 1.6 ± 0.1 for the third quartile, and 1.8 ± 0.1 for the fourth quartile of amount of PA. Higher quartiles were associated with higher peak strain scores than lower quartiles of amount of PA (p < 0.001).

To determine if PA type influenced the association of PA amount with bone parameters, peak strain score was (besides age, height, weight, calcium intake, smoking, age of PA onset, and amount of present PA) included in the multiple linear regression model. Peak strain score was an independent positive predictor of bone parameters in the weight-bearing part of the skeleton (femoral neck aBMD [β = 0.39, p < 0.001] and cortical CSA of the tibia [β = 0.38, p < 0.001]) but not in the non-weight-bearing part of the skeleton (aBMD [β = 0.03, p = 0.58] or the cortical CSA [β = 0.03, p = 0.63] of the radius). In this model, PA amount was still a strong independent predictor of the aBMD at all sites measured and of cortical bone size, both in the weight-bearing and in the non-weight-bearing part of the skeleton (aBMD femoral neck: β = 0.17, p < 0.001; aBMD radius: β = 0.11, p = 0.02; radius CSA: β = 0.18, p < 0.001; tibia CSA: β = 0.12, p = 0.003).

PA before and during pubertal growth is associated with aBMD and cortical bone size

Subjects who started their present PA before age 13 were compared with subjects who began their present PA at age 13 or later to determine the possible contribution of training during puberty and growth to present BMD and bone size. Second to fourth quartile subjects, who began their present PA before age 13, had higher aBMD of the femoral neck (Fig. 3A), cortical CSA of the tibia (Fig. 3B), and trabecular vBMD of the tibia (Fig. 3C) than subjects who began their present PA at age 13 or later.

Figure FIG. 3..

Physical activity before age 13 is associated with increased bone mass. (A) aBMD of the femoral neck, (B) cortical cross-sectional area, and (C) trabecular vBMD of the tibia are shown separately for subjects who began their present PA before age 13 (dotted line) and the subjects who began their present PA at age 13 or later (unbroken line). Subjects were divided into sedentary and quartiles of amount of PA, according to hours per week trained. A > sedentary, B > first quartile, and C > second quartile. Capital and lowercase letters represent p < 0.01 and p < 0.05, respectively. *p < 0.05 and **p < 0.01 represent significant differences between subjects who started training before age 13 and subjects who started training at age 13 or later, for their respective quartile of present amount of PA.

PA and bone size

There were no differences in height or in length of the radius or the tibia between the training and sedentary men (Table 1).

DISCUSSION

PA is believed to have positive effects on the skeleton and possibly help in preventing the occurrence of osteoporosis.(18,22) Neither the lowest effective amount of PA needed to induce an osteogenic response nor its effect on vBMD and size of the different bone compartments has yet been clarified.

In this study, amount of PA was associated with aBMD of the total body, lumbar spine, femoral neck, and radius. The greatest differences in aBMD were seen at the weight-loaded bone sites (6.4–14.0%) between subjects in the fourth quartile and the sedentary group. When adjusting for body size parameters and environmental factors known to affect aBMD, it was shown that PA amount was an independent strong predictor of aBMD at all sites measured. Interestingly, no significant differences in aBMD at any site measured between subjects in the first quartile (training <4 h/week) and the sedentary group were seen, suggesting that most recreational training of <4 h/week is not sufficient to induce increased bone mass. In many previous reports where PA could not stimulate an osteogenic response, the amount of weekly PA exerted was well under the lowest effective amount (4 h/week or more) of PA seen in this study, offering an explanation to previous contradictory findings and failure to prove PA as stimulatory for the skeleton.(23,24) The lowest effective amount of PA associated with increased BMD has been insufficiently studied. Karlsson et al.(19) showed that male soccer players had higher aBMD in weight-loaded bones than what sedentary men had, but they could not see any differences in aBMD between subjects training 12, 8, or 6 h/week. However, the small sample size (67 soccer players and 24 controls) and bone measurements limited to DXA, in this study, could have influenced these results.

In this study, subjects training ≥8.5 h/week had 14.0% higher aBMD (>1 SD) of the femoral neck than sedentary individuals, suggesting that high amounts of weekly PA could be associated with two to three times reduced risk of hip fracture if the higher aBMD is maintained until old age.(4,5) Our regression analyses (including body constitution parameters, environmental factors such as calcium intake and smoking, age, amount of PA, and age of PA onset) could explain 26.3% of the total variation in aBMD of the femoral neck. This finding is in agreement with earlier reports showing that environmental factors could be responsible for 20–40%, whereas genetic factors are believed to be able to explain up to 60–80%, of the variation in aBMD.(8,25,26) In our cohort, amount of PA alone was found to be responsible for almost one-half (12.7%) of the explainable variation in aBMD of the femoral neck.

A few previous reports using pQCT suggested that PA is associated with cortical bone size but not vBMD.(15–17) In this study, we confirmed these reports in our large cohort of young men and also provided novel findings, in showing that PA is associated with cortical bone size but not with cortical vBMD. The larger CSA accompanied by a high amount of PA was associated with cortical thickening and periosteal augmentation, whereas the EC was unchanged. These findings indicate that PA results in radial bone growth caused by periosteal bone formation. In our cohort, CSA of the tibia was 10.5% higher in the fourth quartile (training ≥8.5 h/week), but no difference could be seen in subjects in the first quartile (training <4 h/week) than in the sedentary cohort, further supporting the notion that a high amount of PA is required to affect the size of the cortical bone.

Only a few reports studying the PA association with trabecular vBMD have been published, and the available data are limited and conflicting.(15–17) In this study, a relationship between PA amount and trabecular vBMD of both the radius and tibia was found. It was also found that PA is a strong independent predictor of cortical bone size and trabecular vBMD when taking environmental factors, such as calcium intake and smoking, and body size parameters into account.

It has previously been suggested that PA before and during puberty in girls has profound stimulatory effects on adult BMD.(27,28) In this study, we showed that aBMD, cortical bone size, and trabecular vBMD are higher in subjects who started their training before age 13 than in subjects who started their training later in life, suggesting that exercise before and during the pubertal growth is of importance for the male skeleton.

We believe that this study constitutes the most comprehensive and detailed investigation of the PA association with bone mass in young men thus far reported. The limitation of this study is the cross-sectional design, which does not allow direct cause-effect conclusions. Thus, the associations found between PA and cortical size in these young men could possibly arise if men with larger bones are more prone to exercise. Strongly arguing against this hypothesis are the present findings that the bone lengths of the radius and tibia, as well as height, did not differ between the physically active and the sedentary men. Furthermore, the PA-associated difference in cortical bone size was greater in the weight-loaded tibia than in the non-weight-loaded radius. Finally, aBMD of the radius in the arm being exposed to more PA (dominant) was significantly greater than the aBMD of the arm being exposed to less PA (nondominant).

In conclusion, we show here in the largest cohort of young men ever studied, using both DXA and pQCT, that PA amount is strongly positively associated with aBMD at all sites measured, as well as with cortical bone size and trabecular vBMD, but not with cortical vBMD. These findings suggest that PA increases cortical bone size through periosteal apposition, without changing the material property of the cortical bone. Our findings indicate that most regular recreational exercise, involving <4 h of training per week, is not conducive to the young male skeleton, but higher amounts of weekly PA can have substantial effects in stimulating bone mass accretion and possibly reduce future fracture risk. However, our results are based on young men and can not therefore be generalized and interpreted in regard to possible risk prevention in the population primarily subjected to osteoporosis (i.e., elderly women and men). We suggest that our findings can be useful when outlining recommendations to the public regarding the lowest effective amount of weekly PA needed to induce an osteogenic response in young men.

Acknowledgements

The authors thank the SWEGENE Center for Bio-Imaging (CBI) and Göterborg University for technical support regarding image analysis. This study was supported by the Swedish Research Council, the Swedish Foundation for Strategic Research, European Commission Grant QLK4-CT-2002–02528, the Lundberg Foundation, the Torsten and Ragnar Söderberǵs Foundation, Petrus and Augusta Hedlunds Foundation, and the Novo Nordisk Foundation.

Ancillary