Winning the Battle Against Childhood Physical Inactivity: The Key to Bone Strength?


  • Heather McKay,

    1. Department of Family Practice and Orthopaedics, University of British Columbia, Centre for Hip Health and Musculoskeletal Research, and Vancouver Coastal Health Research Institute, Vancouver, Canada
    Search for more papers by this author
  • Everett Smith

    1. Department of Population Health Sciences, University of Wisconsin Medical School, Madison, Wisconsin, USA
    Search for more papers by this author

  • The authors state that they have no conflicts of interest


In this issue, Gunter et al.(1) highlight the convergence of two factors that greatly promote bone health—the critical period of bone accrual during childhood and the importance of bone loading through specific physical activity. Bailey and McCulloch(2) and others(3,4) first alluded to the possibility that adult osteoporosis had its antecedents in childhood. In 2008, the verdict is in—well-designed childhood physical activity programs are likely critical for preventing osteoporosis in mature adults. Advances in novel imaging technology and analysis methods suggest that the early DXA-based bone and physical activity studies may have underestimated the importance of developing a greater peak bone mass in childhood through targeted physical activity,(5) which will benefit skeletal integrity for life.


The Oxford Evidence Based Medicine group assigns level 1 evidence to findings supported by systematic reviews of studies with populations that are homogenous. Also, individual randomized controlled trials (RCTs) with moderate to large numbers can provide level 1 evidence. Thus, in 2008, there is level 1 evidence that children who undertook weight-bearing physical activity interventions gained significantly more bone mass than did children in control groups who did not.(6)

We highlight three RCTs that provide strong evidence. In the longest randomized intervention to date, the University of British Columbia healthy bones trial administered a 10- to 12-min moderate impact circuit training program 3 times/wk to ∼400 elementary school children. The nearly 5% greater bone accrual (p < 0.05) observed at 20 mo in both girls(7) and boys(8) in the exercising schools was approximately double that reported at 7 mo.(9,10) Notably, hip structural analysis showed that early-pubertal girls increased bone strength by reduced endocortical expansion,(11) whereas boys in the same intervention study had increased periosteal apposition.(8)

To examine the question of exercise effects in relation to calcium, Iuliono-Burns et al.(12) administered a progressive moderate impact exercise program for 20 min, 3 times/wk, with or without calcium (434 mg/d). Children in the exercise-only group had ∼3% greater BMC at the tibia/fibula after 8.5 mo compared with the nonexercise group (p < 0.05). Exercise benefits at the femur were enhanced in calcium-supplemented children (7.1% difference, p < 0.05).

In the study that was the genesis for the follow-up study of Gunter et al. reported here, Fuchs et al.(13) intervened with an intense, repetitive box jumping program across 7 mo in boys and girls. This regimen elicited ground reaction forces 9 times body weight,(14) compared with ground reaction forces 3–5 times body weight during running(15) and jumping(16) and 10 times body weight in elite gymnastics training.(17) The magnitude of the difference in BMC between box-jumpers and controls after 7 mo was ∼4.5% at the hip and 3% at the spine.

The common thread among these three relatively straightforward interventions was that they could be delivered in a relatively short time (10–30 min) by untrained persons—physical education training was not required. Importantly, the positive adaptations to exercise were without adverse effects.


The early studies of exercise in children all had the same primary outcome—change in BMC or BMD as measured by DXA. Although DXA technology was instrumental in setting the course for pediatric bone science, it has a number of limitations because of its planar nature and low spatial resolution.(18) This instrument cannot distinguish between trabecular and cortical bone; it cannot capture changes in bone geometry or describe the surface-specific adaptations to exercise.(19) Thus, a paradigm shift away from assessing only bone mass to the important concept of “bone strength” was needed to capture the more subtle adaptations of children's bone to physical activity.


Ultimate bone strength depends on bone's material properties, quantity, dimensions (size and material distribution), quality, and microarchitecture. Although we would ideally measure each of these, it is impossible to do so in clinical studies. Therefore, to capture the adaptations of bone structure during growth and in response to exercise, investigators have adopted safe and precise imaging techniques such as MRI(20,21) and pQCT.(22,23) pQCT measures volumetric BMD (vBMD, mg/mm3) at both the compartment and total bone level. However, current noninvasive techniques cannot assess bone at the material level (degree of bone “mineralization”) per se, and vBMD includes both the tissue porosity and the average material density and does not distinguish between them.(24) Ultimately, we seek to understand bone's propensity to fail (failure load), which can only be determined ex vivo through mechanical testing. Thus, clinical studies must rely on predicted estimates of bone strength obtained from advanced imaging techniques.

Recent studies more accurately estimate changes in bone strength and “deconstruct DXA” by identifying the components that underpin DXA-based changes. Racquet sport studies that used MRI(20,21) or pQCT(25,26) showed that bend-ing strength was primarily gained through apposition of relatively small amounts of bone at specific sites on the periosteal surface rather than through increased bone mass. There have been very few physical activity intervention studies in children that described changes in bone geometry, density, or estimated strength using pQCT.(27–29) We highlight three of these. Heinonen et al.(27) observed significant increases in BMC by DXA at the femoral neck (4%) after 9 mo of intense training in premenarcheal girls; this did not translate into a positive change in cortical area, density, or density-weighted section modulus at the tibial midshaft (by pQCT). Also, a 16-mo cluster randomized controlled trial of short bouts of primarily classroom-based jumping activity in ∼500 children showed significantly greater, but modest, gains in bone strength (polar strength strain index, mm3) at the tibia in boys (but not girls) whose teachers were at least 80% compliant with the intervention.(28) Furthermore, a 12-mo RCT evaluated the effect of calcium supplementation and physical activity on cortical bone properties at the tibial shaft (20% site) in 3- to 5-yr-old children. At completion of the trial, children in the exercise arm had significantly greater periosteal and endosteal circumferences (p < 0.05) compared with controls.(29)

These studies of bone structure provided the first glimpses of the intricate tapestry that results when bone surfaces of growing bone adapt to physical activity. Increased bone strength was a function of bone apposition on the periosteal surface and/or apposition or decreased resorption on the endosteal surface.(27–29) Once again, the specific adaptation depended on both the sex and maturity of the child.


These clinically relevant findings of site-specific changes with loading are mirrored in the laboratory studies that allow further mechanistic study of the increased mechanical demands of loading. The research of Robling et al.(30) provided two important insights into better understanding mechanical loading of bone. They used two loading regiments: group 1, a single bout of 360 cycles/d at 2 Hz, with a haversine waveform and a peak force of 17 N (∼3300 μϵ of compression for 3 d/wk; group 2, four bouts of 90 cycles separated by 3 h, 3 d/wk. These investigators found that after 16 wk of cyclic loading, the loaded right ulna significantly increased BMC and ultimate strength. However, in group 1, BMC increased 6.9% more in the loaded ulna compared with the contralateral control left ulna, and ultimate force increased 64% more. By comparison, the ulna subjected to four loading bouts, 3 h apart, increased BMC by 7.8% and ultimate force by 87%, compared with the control limb. This study was originally designed to show that rest insertions between loading bouts increased bone formation and strength. However, it very clearly showed that a small percent increase in BMC induced by bone loading significantly increased bone strength. Importantly, for each 1% increase in BMC in the single loaded ulna, ultimate strength increased 9.3%. In comparison, in the four loading cycle regimen, for each 1% increase in BMC, strength increased 11.5%. This contrasts with the PTH study of Sato et al.,(31) which also used the rat ulna model. They observed a 36% increase in BMC after PTH therapy and an ultimate strength gain of 50%. However, for each increase in BMC, they reported only a 1.4% increase in ultimate strength.

The greater skeletal strength conferred by exercise-related changes in bone geometry may explain why there is an overlap between fracture and nonfracture subjects who have similarly low BMC. The rather moderate immediate increase in BMC induced by physical activity in humans may relate to the fact that physical activity increases strength first by internal architectural changes and then by adding BMC along lines of strain.

In addition to changes in the distribution of bone about a defined axis, regional variations in vBMD and cortical thickness (measured with pQCT) within localized areas of the bone cross-section were observed in a nonhuman primate model(32) and in postmenopausal women.(33) These variations were consistent with expected patterns of remodeling and structural adaptations induced by bending. Applied to pediatric exercise studies, this suggests that a regional analysis of bone geometry that is consistent with the predominant loading pattern might provide a more relevant assessment of bone structural adaptation to exercise. There is preliminary evidence of a regional effect in a pediatric school-based intervention study. Macdonald et al.(34) reported increased bending strength (Imax) at the tibial midshaft in young boys in the intervention group and a tendency for quadrant-specific changes in cortical area and cortical thickness that were consistent with expected patterns of bone formation induced by anterior-posterior bending loads.


It is clear from numerous animal studies that there is much more to ultimate bone strength than aBMD and BMC. In a 2-yr rat study, Warden et al.(35) showed a significant difference in aBMD and BMC between exercised and nonexercised ulnas at the end of a 7-wk program. However, this BMC difference did not persist over the 92 wk of the study. Conversely, this study very nicely showed that the intervention induced lifelong benefits in bone structure, strength, and fatigue fracture resistance. Structural changes included a 25.4% increase in minimum second moment of area (Imin) that histology determined was caused by bone apposition on the periosteal surface. The importance of bone structure was also illustrated by the minimum second moments of area, which accounted for 76% of the variance in ultimate force compared with 29% for BMC. However, whereas the exercised ulna had a greater ultimate force to fracture, they were more brittle and exhibited a lower postyield displacement. This was expected because the exercised ulna had a greater ash content and thus greater mineralized tissue than the nonexercised ulna. Differences in mineralization were further verified using synchrotron infrared microspectroscopy that showed the exercised ulna had a greater phosphate to protein and carbonate to protein ratio. Thus, exercise during growth not only increased BMC but also induced long-term structural changes that contributed significantly to bone strength, independent of changes in BMC. Finally, the exercised ulna exhibited a 10-fold greater resistance to fatigue fracture, although there was no difference in BMC. This is consistent with the 23% greater ultimate force required to fracture the exercised ulna.

If we were able to directly transfer the results of these three animal studies to the human model, one might conclude that exercise-induced bone formation during childhood would reduce fractures in the older adult. Two hypotheses might be generated from this conclusion: (1) exercise in childhood will increase bone quantity in the short term and bone quality (distribution of mineral, phosphorus to protein ratio, level of mineralization, crystal size) and bone structure in the long term, and these changes will last a life time; and (2) periosteal and endosteal diameters along with trabecular architecture modeled during the growing years will result in a structural framework formed under the strain of loading, and these changes will last a lifetime. Thus, animal models provide excellent evidence as to how bone responds to mechanical loading and they also provide a sound foundation from which to design human studies.

However, in humans, we do not yet know whether bone mass and strength benefits achieved in childhood persist beyond the growing years into later life when fractures are more likely to be sustained. The studies that would definitely answer these questions would extend well beyond most reader's lifetimes.

Nevertheless, it is inconceivable that a sedentary lifestyle, after a relatively short term of even intense physical activity, could enhance bone health over the long term. Indeed, it is critical to determine the specific exercise regimen or the “minimum effective dose” of exercise required to either maintain or enhance bone health across the lifespan. Human nature suggests that jumping 5 times/d would be more palatable than an hour of intense circuit training 4 times/wk. Although the definitive studies to address this question have yet to, and may never, be completed, we are beginning to accrue powerful near term evidence from human studies(25,26,36,37) and, as discussed previously, long-term evidence from animal studies(35) to support the notion that short-term exercise programs at a key developmental time point may indeed provide sustained bone health benefits. There is also some retrospective evidence that suggests higher levels of bone attained in childhood athletes is maintained into young adulthood.(38,39)

Notably, epidemiological evidence supports an association between physical activity during the growing years and reduced fracture risk in later life.(40–42) A number of cross-sectional studies showed that individuals who continue to be physically active over their lifespan also maintain a higher BMC and reduce their risk of fracture.(43,44) Smith et al.(45) and others(46) showed that physical activity increased BMD, even in the seventh and eighth decade of life. Perhaps, even more importantly,(18) it is now well known that physical activity reduces both falls risk factors(47,48) and falls.(49,50) It seems most likely that by effectively targeting both poor bone health and increased risk of falls together, we will in the future reduce the risk of fracture.

All of that said—perhaps the more relevant question is how does physical inactivity influence our genetic blueprint for bone health? What is the cost of physical inactivity to the health of children's bones? Whatever the cost—it is one that our society can not afford to pay. Therefore, the significant question becomes how do we engage individuals in physical activity early in their lives to increase the likelihood of continued physical activity during the challenging adolescent period, especially for girls,(51) into adulthood and during later life.


To truly capture the nuances of growth, longitudinal studies are imperative. A landmark study from the University of Saskatchewan assessed BMC accrual across 6 yr of growth in normally active children. The most active children accrued more bone over the measurement timeframe and had greater BMC at maturity compared with the least active children.(52) Forwood et al.(53) used hip structural analysis to study the influence of physical activity on femoral neck bone strength (represented by section modulus, Z) in this same cohort. In both boys and girls, physical activity was a significant predictor of Z and cross-sectional area but not subperiosteal width (SPW). The authors suggested that for a change in Z of ∼4%, only a 0.5% increase in SPW would be required. Detection of such a small increase in SPW is below the detection limit of current DXA technology.

In this issue of JBMR, Gunter et al.(1) undertook an 8-yr prospective follow-up study using DXA technology and showed that benefits observed at the proximal femur after an intensive 7-mo jumping trial were maintained in children (7–8 yr old at baseline) after 8 yr. This challenging study was innovative and important for a number of reasons: its longitudinal design, its use of advanced statistical modeling techniques to manage factors (most importantly maturity) that differed between children, its retention of participants for almost 8 yr, and most notably, for its focus on physical activity. To account for repeated measures within individuals and individual growth characteristics, sex-specific hierarchical random-effects models were created using a multilevel modeling approach. Importantly, age at peak height velocity was determined for every child and very elegantly used to control for the substantial maturational differences that can occur between children of the same chronological age.


Although researchers generally study the effect of physical activity on specific systems (e.g., bone health, cardiovascular health), we would vastly undervalue physical activity if we compared it with medication that benefits only one biological system. In children, appropriate physical activity improves cognitive development and academic achievement,(54–56) reduces cardiovascular risk factors,(57) and pro- motes strength and balance. Physical activity often provides a context for social interactions and fosters mutual understanding among children. Sadly, we live in an age when children are increasingly sedentary,(58) and childhood obesity garners global headlines. Why do so many children in developed countries remain inactive? What will remove the dissonance between what we know and what we do?

To be effective and sustained, physical activity must be woven into the fiber of society. This can be done. Escalating concerns about childhood inactivity and obesity has prompted public health efforts to identify and adopt population-based strategies that positively influence physical activity behaviors.(59) Because children spend most of their waking hours in school, and schools cross sociodemographic boundaries, active school models may be the best means to enhance childhood physical activity. A number of school-based studies that incorporated physical activity breaks in the classroom and adopted an environmental (whole school) approach to promoting physical activity showed a positive impact on children's physical activity levels.(60–62)

RCTs of school-based physical activity models very clearly support the notion that “a little goes a long way” when it comes to augmenting children's bone health through exercise.(18,28,63) Importantly, school-based models that encouraged children to “snack on physical activity” during the school day and demanded only a few minutes of teacher's curricular time showed positive health benefits.(64) Given the competing demands for teacher's time, this may be key to sustaining school-based physical activity practices. However, teachers can not achieve this alone. Sustained investment by organizations and governments to promote and extend physical activity research and to implement evidence-based physical activity practices in schools is urgently needed if we are to win the war against childhood physical inactivity(65–67) and promote and sustain child bone health.


We dedicate this commentary to our trainees who have contributed their curiosity, commitment, and passion to a myriad of research projects over the years and who have made the world a better place for being in it. We thank Dr Heather Macdonald, whose doctoral work drove pediatric bone health research forward beyond DXA to new frontiers of bone structure and strength, and to Yasmin Ahamed, MSc, who guided us through to completion of this commentary. We further thank Sylvana García Rodriguez for her commitment to research in bone biomechanics and contributions and Prof Heidi Ploeg for help with this commentary. HM is supported by the Canadian Institutes of Health Research and by a Michael Smith Foundation for Health Research Senior Scholar Award. We also acknowledge Frank Booth's pivotal review paper, “Waging War Against Physical Inactivity” (J Appl Physiol 2000), for its emphasis on physical inactivity as a medical condition. This paper taught us the importance of the paradigm shift from exercise [an optional extra] to physical inactivity [the major driver of health costs in the developed world]. ES thanks the AO Foundation 02-S52 and the University of Wisconsin Graduate School for support.