Are associations between physical activity and bone mineral density in adults sex- and age-dependent? An analysis of the UK Biobank study

Gallin Montgomery; Mohamed Yusuf; Rachel Cooper; Alex Ireland

doi:10.1093/jbmr/zjae017

. 2024 Jan 29;39(4):399–407. doi: 10.1093/jbmr/zjae017

Are associations between physical activity and bone mineral density in adults sex- and age-dependent? An analysis of the UK Biobank study

Gallin Montgomery ¹, Mohamed Yusuf ², Rachel Cooper ^3,⁴, Alex Ireland ^5,^✉

PMCID: PMC11207656 PMID: 38477821

Abstract

Although physical activity (PA) is recognized as a key bone mass determinant across life, athlete studies suggest that it may be less effective in women and older individuals. This has not been explored within the general population. We aimed to address this knowledge gap using data from the UK Biobank Study, a large population-based study of middle-aged and older adults. Free-living PA data collected at 100 Hz for 7 d using wrist-worn accelerometers were classified as sedentary behavior (0–29 milligravities [mg]), light (30–124 mg), or moderate-to-vigorous PA (125 + mg). LS and FN-BMD were assessed using DXA. The associations between PA and BMD were assessed using linear regression models, with formal assessments of sex and age interactions undertaken and adjustments made for accelerometer wear time, height, body mass index, education, ethnicity, disability, and (in women only) menopausal status. In total, 15 133 UK Biobank participants (52% women) had complete PA, bone, and covariate data. In this sample, greater overall and moderate-to-vigorous PA was associated with higher LS BMD. In women, these associations were typically weaker in older individuals, for example, regression coefficients in women aged 70 yr or older were ~50% lower than at 45–54 yr (age-by-PA interactions P < .01 in all models). Similar associations were observed in basic but not full models for FN BMD. Greater sedentary time was associated with lower LS BMD in men only, and greater light PA and sedentary time were associated with higher and lower FN BMD, respectively, in both sexes. These results suggest that associations between PA and bone health at clinically-relevant sites are weaker in older than younger women. That positive associations are evident between overall and moderate-vigorous PA and FN BMD even in women ≥70 yr suggests that PA for bone health should still be promoted in older women.

Keywords: exercise, mechanoadaptation, mechanostat, DXA

Introduction

Osteoporosis is the age-related deterioration of bone mass and architecture, which leads to bone fragility and increased fracture risk.¹ Osteoporosis and associated fractures are common health outcomes that are burdensome at the individual and societal levels. For example, over 500 000 fragility fractures occur in the UK each year with associated healthcare costs of approximately €5bn, a figure that is expected to grow to around €7bn by 2030.² Hip and spine fractures result in large increases in mortality rates particularly in the early period after fracture³ and have higher associated healthcare costs than other types of osteoporotic fractures.⁴ The incidence of osteoporosis and associated fractures is highly sex and age-dependent,⁵ with around 75% of osteoporotic fractures occurring in women aged over 65 yr.⁶ Bone mass, assessed clinically as BMD, is a key predictor of fracture risk across multiple sites in older adults of both sexes.⁷ Given this, maintenance of bone mass in older adults (particularly women) is an important public health goal.

A key determinant of bone mass across life is mechanical loading by reaction and muscle forces during physical activity (PA). Long-term participation in exercise is associated with up to 40% greater bone mass in athletes,⁸^,⁹ whereas long-term unloading following spinal cord injury is associated with site-specific bone mass losses of up to 50%.¹⁰ In the broader population, a number of studies have shown that greater levels of participation in habitual PA are associated with higher bone mass in people of different ages, including older adults.^11–14 Moderate and vigorous activities such as running and team sports appear to be particularly beneficial for bone mass.^11–14 However, little is known about whether these associations differ by sex or age in older adults. This is important, given the substantial age and sex dependency of osteoporosis and fractures.

Animal studies suggest that the mechanosensitivity of bone decreases with age,¹⁵ while the age-related loss of muscle function also reduces the stimulus applied to the bone during PA. In addition, the proposed modulating role of estrogen on bone’s response to mechanical loading¹⁶ suggests that the adaptive response of bone may be impaired in post-menopausal women. Evidence from athlete studies suggests a reduced advantage in bone mass in older individuals despite the maintenance of PA levels.¹⁷^,¹⁸ In contrast, the findings on sex differences are mixed,¹⁷^,¹⁸ but could be explained by self-selection biases inherent in athlete-control studies. In a study of associations between PA and tibia bone strength indicators, no sex-specific associations were observed.¹¹ To the best of our knowledge, none of the studies that have objectively assessed PA and bone in adults, for example,¹¹^,¹²^,¹⁹^,²⁰ have investigated whether associations between habitual PA and bone health differ by age or, whether sex-specific associations exist for clinically-relevant sites such as the spine and hip.

To address this important knowledge gap, we utilized data from the UK Biobank study, a large population-based study of middle-aged and older adults to examine associations between accelerometer-assessed measurements of habitual PA and clinically-relevant DXA assessments of bone. The large sample size facilitated the testing of 2- and 3-way interactions between PA, sex and age, in relation to BMD. We hypothesized that associations between PA and BMD would be stronger in younger adults compared to older adults, and in men compared with women.

Materials and methods

Data were obtained from UK Biobank, a prospective cohort study of over 500 000 participants (5.5% of those invited) aged 37–69 yr when recruited in 2006–2010.²¹ Ethical approval was given by the National Information Governance Board for Health and Social Care and the National Health Service North West Centre for Research Ethics Committee (Ref: 11/NW/0382), and participants provided written informed consent at each data collection wave.

Accelerometry

A total of 236 519 (47.1%) participants who completed the baseline assessment in 2006–2010 were invited to wear an accelerometer for 7 d between February 2013 and December 2015. Participant’s email addresses were randomly selected with the exception of those in the North West of England. As participants from this region had been recruited for new Biobank projects being trialed earlier in the study’s development, they were excluded from this wave due to concerns about participant burden. The consenting 103 053 participants (43.6% of those invited) agreed to wear the Axivity AX3 triaxial accelerometers (AX3, Axivity) on their wrist for 7 d (range: ±8 g). Acceleration data were sampled at 100 Hz and calculated as mean vector magnitude values over 5 s epochs. After excluding data from participants with <72 h wear time and where data failed to meet calibration requirements, final data from 96 600 participants (93.7% of those that consented) were available.²²

Accelerometer data were calculated as the Euclidean norm for all 3 axes (x,y,z) and processed accounting for stationary time, interrupted data, non-wear time, and signal noise as previously described.²² PA was calculated as the mean of all measured and imputed values and expressed in milli-gravity units (mg), having removed 1 g from the signal and calculating all negative values as zero. Non-wear data segments were imputed by the core Biobank analysis team, as previously described.²² Accelerometer data were also used to calculate estimates of time spent in sedentary behavior (0 ≤ mg < 30, minus self-reported sleep duration), light PA (30 ≤ mg < 125), and moderate-vigorous PA (mg ≥ 125), which were expressed in minutes duration in each range.²³

Musculoskeletal imaging

In 2014, 46 848 participants attended an imaging visit where DXA scans of the proximal femur (n = 39 431) and spine (from L4 to T4, n = 38 842) (iDXA GE-Lunar) were obtained using a standardized protocol.²⁴ Daily quality control and local calibration using a phantom were performed in accordance with the manufacturer’s guidelines (GE-Lunar). Cross site calibration was performed intermittently using a European spine phantom for consistency in recordings.²⁵

Covariates

A number of potential confounding variables was selected a priori. Height was measured at baseline using a stadiometer (Seca) and body mass with a Tanita BC418ma (Tanita Europe) scale, from which BMI (kg/m²) was calculated. Highest educational qualification and whether participants had a long-standing illness or disability were recorded via questionnaire. In women, menopausal status was self-reported as a categorical variable, coded as premenopausal, postmenopausal, or hysterectomy. Because of the small number of participants from minority ethnic groups in the UK Biobank, self-reported ethnicity was categorized as a binary variable: white (encompassing white British, Irish, or other white ethnicity) or other. Wear time was recorded from accelerometer data capture.

Statistical analysis

Multivariate regression models were used to assess the relationships between each of the 4 main accelerometry measures (ie, time spent sedentary, average PA and time spent in light and moderate-vigorous PA) and BMD at the LS (L1–L4) and FN regions. To assess deviations from linearity, quadratic terms for each PA variable were entered into sex-stratified basic models. This was done in view of non-linear associations between loading volume and bone gain identified in animal models.²⁶ Where a quadratic term was statistically significant, we visually inspected regression coefficients from models with the PA variable grouped into tenths. We decided a priori to analyze data from men and women separately, given evidence of sex-specific associations between PA and bone in younger populations and in older athletes.¹⁷^,²⁷ In basic models with adjustments for height, BMI, age, and accelerometer wear time, we formally tested the relationships between PA and BMD with age interactions. Full models were additionally adjusted for the highest level of education, ethnicity and long-term illness, or disability. Models for data in women were also adjusted for menopausal status.

Where there was a statistically significant interaction between PA and age, subsequent models were stratified by age group. In these cases, sex by PA interactions could not be assessed as the degree of difference between sexes would be dependent upon the age at which it was examined. Where there was no PA by age interaction for men or women, models were adjusted for age and examined for sex by PA interactions. Likelihood ratio tests were used to compare the fit of sex-specific models with those where sexes were combined, and a sex-by-age-by-PA interaction term was included. Where substantial attenuation of regression coefficients was observed following adjustment for covariates in full models, additional analyses were performed with each individual covariate introduced separately to assess their individual impact on associations.

All analyses were run on the sample with complete accelerometry, imaging, and covariate data. Characteristics of the UK Biobank participants included in analyses were compared with those who had incomplete data using unpaired t-tests. All analyses were performed using RStudio (R Foundation for Statistical Computing 2021, Version 1.4.1717).

Results

A total of 15 133 participants had both valid accelerometry and bone imaging data in addition to complete covariate data (Figure 1), their basic characteristics are described in Table 1. Although the majority of participants completed their accelerometry measurements prior to the DXA imaging, 9.4% completed them after. The proportions of men and women in the latter group were similar (9.5% and 9.3% respectively, P = .8), but participants completing the accelerometry measure after imaging were slightly younger (62 ± 7 yr) than those who completed it before (64 ± 8 yr, P < .001). When compared with the characteristics of the UK Biobank participants who were not included in our analytical sample, those who were included were slightly less likely to be women (53.4% complete cases vs 54.4% incomplete cases), more likely to be of white ethnicity (97.5% vs 94.5%), and more likely to have a college or university degree (47.4% vs 32.3%). Within the analytical sample, women were younger and shorter with lower body mass, BMI, LS, and FN BMD compared with men. Ethnicity distribution was similar in both sexes (P = .5), and women had a lower level of long-standing illness or disability. Women had higher levels of overall, light, and moderate-vigorous PA than men, with less sedentary time recorded. For both men and women, overall and moderate-vigorous PA was lower in older individuals (Table 2, all P < .001). Although sedentary time and light PA were lower in older than younger women, the same trend was not observed in men (both P > .3).

Flow diagram showing participant n at each stage of data preparation.

Table 1.

Characteristics of the UK Biobank participants included in analyses (N = 15 133) stratified by sex, presented as mean (SD) or N (%).

graphic file with name zjae017fx1.jpg

Open in a new tab

Table 2.

PA characteristics, stratified by age and sex and presented as median (IQR).

Men	Overall acceleration average (mg)	Sedentary time (min)	Light PA (min)	Moderate-to-vigorous PA (min)	Wear time (d)
45–54, N = 969	30 (25, 36)	659 (600, 717)	279 (239, 320)	82 (62, 109)	6.91 (6.61, 7.00)
55–59, N = 1001	29 (24, 36)	657 (589, 723)	278 (233, 321)	81 (56, 105)	6.93 (6.66, 7.00)
60–64, N = 1303	28 (23, 33)	651 (590, 718)	279 (237, 318)	73 (52, 99)	6.95 (6.76, 7.00)
65–69, N = 1705	26 (22, 32)	651 (588, 712)	278 (236, 320)	66 (48, 92)	6.95 (6.80, 7.00)
70+, N = 2253	25 (21, 30)	657 (591, 719)	276 (238, 315)	59 (42, 82)	6.95 (6.83, 7.00)
P-value	< .001	.389	.903	< .001	< .001
Women
45–54, N = 1127	30 (25, 36)	630 (569, 695)	300 (261, 338)	82 (60, 108)	6.90 (6.57, 6.96)
55–59, N = 1366	30 (25, 35)	625 (561, 690)	302 (264, 340)	81 (59, 107)	6.91 (6.66, 7.00)
60–64, N = 1707	28 (24, 34)	630 (567, 696)	300 (264, 338)	73 (53, 101)	6.91 (6.67, 7.00)
65–69, N = 1897	27 (23, 32)	631 (568, 696)	304 (265, 346)	68 (48, 94)	6.91 (6.74, 7.00)
70+, N = 1805	26 (22, 31)	636 (571, 703)	308 (266, 350)	62 (43, 86)	6.91 (6.74, 7.00)
P-value	< .001	.044	.002	< .001	< .001

Open in a new tab

The relationships between light PA and BMD at the LS and FN were non-linear. For these analyses, light PA was modeled as a quadratic relationship following assessment of log likelihood ratios from different model structures. In women, greater levels of overall and moderate-vigorous PA were associated with greater LS and FN BMD (Figures 2 and 3). The associations with LS were age-dependent, being typically weaker in older women (age by PA interactions P ≤ .001). For example, in the final models for overall activity, when compared with the standardized regression coefficient for the youngest age group of 45–55 yr (0.057, 95% CI, 0.001 to0.114, P = .044), values were over 50% lower in women over 70 (0.026, −0.02 to 0.072, P = .256) for whom there was no clear evidence of an association. Similar age-dependent associations were observed for moderate-vigorous activity, in which coefficients in women over 70 (0.025, −0.02 to 0.071, P = .28) were around 50% lower than those in younger women (0.047, −0.008 to 0.103, P = .094).

Associations between PA measurements and LS BMD, in basic and fully-adjusted models. Where age-by-PA interactions were identified, data are presented stratified by age group. Basic model: height, BMI, age, accelerometer wear time. Full model: basic model + highest level of education, ethnicity, disability, and (for women only) menopausal status. Error bars indicate 95% CI.

Associations between PA measurements and FN BMD, in basic and fully-adjusted models. Where age-by-PA interactions were identified, data are presented stratified by age group. Basic model: height, BMI, age, accelerometer wear time. Full model: basic model + highest level of education, ethnicity, disability, and (for women only) menopausal status. Error bars indicate 95% CI.

For the FN in women, the relationships between higher overall and moderate-vigorous PA and higher BMD in basic models also became weaker with age. Regression coefficients were around 50% lower in the oldest than in the youngest age group, although this age-dependency was attenuated in full models (Supplementary Tables 3 and 4, age-by-PA interactions P = .113 and P = .057, respectively). Sensitivity analyses showed that this attenuation was almost entirely attributable to adjustment for menopause status. In contrast to results in the LS, there was still evidence of higher FN BMD in individuals with greater levels of both overall (0.062, 0.017 to 0.108, P = .008) and moderate-vigorous PA (0.052, 0.006 to 0.098, P = .025) in the oldest group. Light and sedentary activity was not associated with LS BMD in either model (all P > .6). However, at the FN, greater time spent sedentary was associated with lower BMD (full model -0.032, -0.052 to -0.012, P = .002), and greater time in light PA was associated (full model 0.028, 0.008 to 0.048, P = .005) with higher BMD with little difference between models.

Similar to findings in women, in men, higher levels of overall and moderate-vigorous PA were associated with greater LS and FN BMD. There was little (<10%) attenuation of regression coefficients following adjustments for covariates, therefore throughout the rest of the text we report only the full model results for males (both sets of results are shown in Figures 2 and 3). Unlike in women, there was no evidence that relationships between any PA measure and BMD at either site differed by age (all age by PA interactions P > .1).

For the LS, higher levels of overall (standardized regression coefficient: 0.057, 95% CI, 0.033 to 0.081, P < .001) and moderate-vigorous PA (0.048, 0.024 to 0.072, P < .001) were associated with lower BMD. Greater levels of sedentary behavior (−0.037, −0.059 to −0.014, P = .001) were associated with lower BMD, with weak evidence of a sex interaction (P = .07) suggestive of a stronger association in men than women. No association was observed for light activity (both models P > .15). Similar but stronger associations were observed at the FN, with higher levels of overall (0.109, 0.086 to 0.132, P < .001) and moderate-vigorous PA (0.089, 0.066 to 0.112, P < .001) associated with greater BMD and higher levels of sedentary behavior associated (−0.048, −0.070 to −0.026, P < .001) with lower BMD. However, at this site, higher levels of light activity were also associated with higher BMD (0.029, 0.008 to 0.051, P = .008). There was no evidence of sex differences in association between sedentary behavior or light activity and FN BMD (all interactions P > .2).

For both LS and FN interactions described above, likelihood ratio tests were used to compare the fit of sex-specific models with those where sexes were combined, and a sex-by-age-by-PA interaction term was included. In all cases, the fit was better with models including an interaction term, supporting the finding of sex-specific associations indicated in the sex-stratified analyses.

Discussion

The aim of this study was to examine whether associations between different intensities of PA and bone density at 2 clinically relevant sites (LS and FN) differed by sex or age in a large population-based cohort. Higher levels of overall and moderate-vigorous PA were found to be associated with greater BMD at both sites in men and women in all models. In line with previous studies, the effect size of these associations was relatively modest. For women only, these associations were age-dependent and were approximately 45%–55% weaker in the oldest than youngest age group. For the LS, these associations were robust to adjustment for potential confounders. Greater time spent sedentary was associated with lower LS BMD in men only. However, no sex-specific associations with sedentary time were observed in the FN or between light PA and bone outcomes.

To the best of our knowledge, age-related differences in the associations between PA and bone have not previously been formally tested in adults, and sex differences in association have only been examined at less clinically-relevant sites.¹¹ In the previous study, no sex interactions were observed which may be related to the much lower participant numbers and hence statistical power than in our analyses. This may also explain our finding of associations between greater light PA and higher FN BMD which were not evident in previous smaller studies in older adults.¹¹^,¹² The results agree in part with previous studies in master athletes, whereby both advantages in lower limb bone in runners and racquet arm bone advantages in tennis players were less pronounced in older players.¹⁷^,¹⁸ Our findings of a weaker association between PA and bone in older women than men also concur with similar findings in the tennis players.¹⁷ However, opposing associations were observed in the tibiae of master athletes compared with controls.¹⁸ This discrepancy may be explained by self-selection biases evident in the case–control design of the latter study that group differences may be explained by differences in genetics, nutrition, or other factors between athletes and controls. For the tennis players, the non-racquet arm could be considered as a more robust internal control. Sex-stratified analyses of age interactions were not completed in these previous studies, so we do not know if the sex-specific age-by-PA interactions observed in the current study were evident in these other populations. A previous study examined heel ultrasound measures of bone in a larger sample of women from the UK Biobank, but did not examine clinical DXA measurements, sex effects, or the interaction between sex, age, and PA.²⁰

There are a number of possible mechanisms which could explain the sex-specific age-by-PA interactions observed in this study. Hormonal status, particularly estrogens, appear to have an important role in mechanosensitivity,¹⁶ and postmenopausal changes could contribute to impaired efficacy of exercise on bone health. This is supported by data showing that the ratio of bone mass to lean mass (a proxy marker of mechanical loading) is lower in postmenopausal than in premenopausal women.²⁸ This has also been demonstrated in an interventional model, whereby postmenopausal women had showed a weaker bone response to the same high-impact exercise intervention than premenopausal women.²⁹ Menopausal status did not influence PA and bone associations independent of age. However, we only had basic information on menopausal and hysterectomy status rather than detailed information on age at which either occurred or detailed information on hormone or other treatments. This may have limited our ability to fully capture the effects of menopause on bone. An age-related decrease in mechanosensitivity has also been observed in robust animal studies in male animals,¹⁵ so it is unclear why the associations in men were not also affected by age.

The intensity-specific nature of the observations, for example, being evident only at both sites in moderate-vigorous and overall activity, may be a result of reduced statistical power to detect these associations due to the much weaker associations between light activity or sedentary behavior and bone. Alternatively, it could be related to the greater inter-individual variation evident in moderate-vigorous than other PA behaviors in this cohort. Generally, associations were stronger for FN than LS BMD. This may be due to the former site’s proximity to ground contact and hence reduced attenuation of impact reaction forces. In addition, larger, stronger muscles are acting around the hip than the spine. Given the key role of muscular forces in bone loading, the age-related loss of muscle force and mass has been suggested to contribute to reduced effectiveness of exercise in bone.¹⁷ Unfortunately, the count-based measures of PA used in this study do not allow for a more detailed assessment of changes in high-intensity activity loading likely to be osteogenic. Hence, the contribution of muscle weakness in older age to the observed site-specific associations cannot be explored.

A substantial minority of osteoporotic fractures also occur in men,³⁰ and the condition is undertreated in this population, with relatively few exercise intervention studies for better bone health available.³¹The current results suggest that PA-based interventions may better retain their effectiveness in older men. Alternatively, that the moderate-vigorous activities engaged in by women may not be those activities such as high-impact exercise known to be particularly beneficial to bone. Despite the weakened associations in older women, higher levels of moderate-vigorous and light activity were associated with greater FN BMD even in the oldest age group, and therefore PA for bone health should still be encouraged in this group. These results could also emphasize the importance of exercise earlier in life, where benefits at clinically-relevant regions may be greater.¹⁷ In men, benefits to bone geometry accrued during growth persist several decades after cessation of exercise.³² Although this has not been explored over the same time period in women, good maintenance of exercise benefits to bone has been observed after 5 yr of reduced exercise volume in premenopausal women.³³ Although a number of interventional exercise trials for bone health have been conducted, relatively few have involved males,³¹ and overall effect sizes in women have been moderate.³⁴ This prevents well-powered investigation of moderation of exercise effects by sex and age, which to the best of our knowledge have not been performed across both sexes.

This study used a large population-based cohort with detailed objective measures of PA and clinically-relevant measures of bone health. This gave us substantial statistical power to perform the interaction analyses which were the focus of this study. The UK Biobank cohort is not broadly representative of the UK population,³⁵ particularly with regards to socioeconomic position and ethnicity which may limit the generalizability of these results. Further to this, the subset of individuals who had complete data and were included in analyses differed in basic characteristics from the broader UK Biobank cohort. Moreover, as a cross-sectional observational study, we cannot rule out the possibility of reverse causality. An exposure should ideally be assessed prior to an outcome. In UK Biobank, the larger accelerometry study (February 2013-December 2015) took place around the imaging study (conducted in 2014) such that for a small proportion of participants temporality was reversed. The accelerometer measures available in UK Biobank assess PA at the wrist. Although this is strongly correlated with values at the hip, measures from this site would have increased relevance for bone strength especially at the FN. Some activities such as resistance training known to be beneficial for bone would be detected as light behavior. Although only a small minority of older adults participate in resistance training,³⁶ and fewer still in high-magnitude exercise likely to benefit bone, we must acknowledge this as a limitation of the study. Conversely, the thresholds for moderate-vigorous activity used in Biobank were not specifically chosen for their relevance for bone health. Therefore, some activities classified as moderate-vigorous may not be osteogenic, which may have weakened observed associations. Also, more detailed assessment of activity magnitude through ,for example, identification of individual impacts¹²^,³⁷ would have allowed further interrogation of the age and sex-specific associations between PA and bone. We cannot rule out residual confounding by exposures not included in analyses, including more detailed assessment of clinical conditions and use of medications. In Supplementary Tables 1–8), we reported observed associations for each covariate, which given the large resultant number of assessments could justify adjustment for multiple comparisons. However, we only had one exposure of interest (PA, plus its interactions with age and sex). The other exposures were only included as potential confounding covariates, and were not examined or reported in the manuscript. Therefore, as multiple testing adjustment for confounders is not conventional practice, we did not include this.

We observed that although greater overall and moderate-vigorous PA was associated with higher LS and FN BMD, in women only, these associations were age dependent being ~45%–55% weaker in the oldest compared to the youngest individuals. This may relate to reduced mechanosensitivity secondary to postmenopausal hormonal changes, or alternatively to the type of activity undertaken. Alternatively, to altered muscular loading during PA, which we could not explore further in this study. However, in women over 70, both light and moderate-to-vigorous PA were still associated with higher BMD at the FN, which is a key osteoporotic fracture site. This suggests that PA for bone health should still be promoted in older women, and that it may be consistently effective for men across later adult life.

Supplementary Material

Biobank_PA_paper_supplementary_material_zjae017

Biobank_PA_paper_supplementary_material_zjae017.docx^{(29KB, docx)}

Acknowledgments

This research has been conducted using data from UK Biobank, a major biomedical database, under application number 71242, and we are therefore grateful to the participants of UK Biobank.

Contributor Information

Gallin Montgomery, Department of Sport and Exercise Sciences, Musculoskeletal Science and Sports Medicine Research Centre, Manchester Metropolitan University, Manchester M1 5GD, United Kingdom.

Mohamed Yusuf, Faculty of Epidemiology and Population Health, Department Infectious Disease Epidemiology and International Health, London School of Hygiene and Tropical Medicine, London WC1E 7HT, United Kingdom.

Rachel Cooper, AGE Research Group, Faculty of Medical Sciences, Translational and Clinical Research Institute, Newcastle University, Newcastle upon Tyne NE4 5PL, United Kingdom; NIHR Newcastle Biomedical Research Centre, Newcastle University, Newcastle upon Tyne Hospitals NHS Foundation Trust and Cumbria, Northumberland, Tyne and Wear NHS Foundation Trust, Newcastle upon Tyne NE4 5PL, United Kingdom.

Alex Ireland, Musculoskeletal Science and Sports Medicine Research Centre, Manchester Metropolitan University, Manchester M1 5GD, United Kingdom.

Author contributions

Gallin Montgomery (Data Curation, Formal analysis, Writing—Original Draft, Writing—Review & Editing), Mohamed Yusuf (Data Curation, Formal analysis, Writing—Review & Editing), Rachel Cooper (Conceptualization, Writing—Review & Editing, Supervision), and Alex Ireland (Conceptualization, Formal analysis, Writing—Original Draft, Writing— Review & Editing, Supervision)

Funding

UK Biobank data access fees were supported by Manchester Metropolitan University.

R.C. acknowledges support from the National Institute for Health and Care Research (NIHR) Newcastle Biomedical Research Centre based at Newcastle upon Tyne Hospitals NHS Foundation Trust, Cumbria, Northumberland, Tyne and Wear NHS Foundation Trust and Newcastle University.

R.C. also receives support as part of a generous donation made by the McArdle family to Newcastle University for research that will benefit the lives of older people in the UK.

The views expressed in this publication are those of the authors and not necessarily those of the National Institute for Health and Care Research, the Department of Health and Social Care, or the McArdle family.

Conflicts of interest

G.M., M.Y., R.C., and A.I. declare that they have no conflicts of interest relevant to this work.

Data availability

UK Biobank data are available through an access procedure described at https://biobank.ndph.ox.ac.uk/showcase/exinfo.cgi?src=AccessingData.

References

1.Consensus development conference on osteoporosis. Hong Kong, April 1-2, 1993. Am J Med. 1993;95:1S–78S. [PubMed] [Google Scholar]
2.Borgström F, Karlsson L, Ortsäter G, et al. Fragility fractures in Europe: burden, management and opportunities. Arch Osteoporos. 2020;15(1):59. 10.1007/s11657-020-0706-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Johnell O, Kanis JA, Odén A, et al. Mortality after osteoporotic fractures. Osteoporos Int. 2004;15(1):38–42. 10.1007/s00198-003-1490-4. [DOI] [PubMed] [Google Scholar]
4.Weycker D, Li X, Barron R, Bornheimer R, Chandler D. Hospitalizations for osteoporosis-related fractures: economic costs and clinical outcomes. Bone Rep. 2016;5:186–191. 10.1016/j.bonr.2016.07.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Curtis EM, van der Velde R, Moon RJ, et al. Epidemiology of fractures in the United Kingdom 1988-2012: variation with age, sex, geography, ethnicity and socioeconomic status. Bone. 2016;87:19–26. 10.1016/j.bone.2016.03.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Burge R, Dawson-Hughes B, Solomon DH, Wong JB, King A, Tosteson A. Incidence and economic burden of osteoporosis-related fractures in the United States, 2005-2025. J Bone Miner Res. 2007;22(3):465–475. 10.1359/jbmr.061113. [DOI] [PubMed] [Google Scholar]
7.Johnell O, Kanis JA, Oden A, et al. Predictive value of BMD for hip and other fractures. J Bone Miner Res. 2005;20(7):1185–1194. 10.1359/JBMR.050304. [DOI] [PubMed] [Google Scholar]
8.Ireland A, Maden-Wilkinson T, McPhee J, et al. Upper limb muscle-bone asymmetries and bone adaptation in elite youth tennis players. Med Sci Sports Exerc. 2013;45(9):1749–1758. 10.1249/MSS.0b013e31828f882f. [DOI] [PubMed] [Google Scholar]
9.Wilks DC, Winwood K, Gilliver SF, et al. Bone mass and geometry of the tibia and the radius of master sprinters, middle and long distance runners, race-walkers and sedentary control participants: a pQCT study. Bone. 2009;45(1):91–97. 10.1016/j.bone.2009.03.660. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Rittweger J, Goosey-Tolfrey VL, Cointry G, Ferretti JL. Structural analysis of the human tibia in men with spinal cord injury by tomographic (pQCT) serial scans. Bone. 2010;47(3):511–518. 10.1016/j.bone.2010.05.025. [DOI] [PubMed] [Google Scholar]
11.Johansson J, Nordström A, Nordström P. Objectively measured physical activity is associated with parameters of bone in 70-year-old men and women. Bone. 2015;81:72–79. 10.1016/j.bone.2015.07.001. [DOI] [PubMed] [Google Scholar]
12.Hannam K, Deere KC, Hartley A, et al. Habitual levels of higher, but not medium or low, impact physical activity are positively related to lower limb bone strength in older women: findings from a population-based study using accelerometers to classify impact magnitude. Osteoporos Int. 2017;28(10):2813–2822. 10.1007/s00198-016-3863-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Elhakeem A, Heron J, Tobias JH, Lawlor DA. Physical activity throughout adolescence and peak hip strength in young adults. JAMA Netw Open. 2020;3(8):e2013463. 10.1001/jamanetworkopen.2020.13463. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Deere K, Sayers A, Rittweger J, Tobias JH. Habitual levels of high, but not moderate or low, impact activity are positively related to hip BMD and geometry: results from a population-based study of adolescents. J Bone Miner Res. 2012;27(9):1887–1895. 10.1002/jbmr.1631. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Rubin CT, Bain SD, McLeod KJ. Suppression of the osteogenic response in the aging skeleton. Calcif Tissue Int. 1992;50(4):306–313. 10.1007/BF00301627. [DOI] [PubMed] [Google Scholar]
16.Klein-Nulend J, van Oers RF, Bakker AD, Bacabac RG. Bone cell mechanosensitivity, estrogen deficiency, and osteoporosis. J Biomech. 2015;48(5):855–865. 10.1016/j.jbiomech.2014.12.007. [DOI] [PubMed] [Google Scholar]
17.Ireland A, Maden-Wilkinson T, Ganse B, Degens H, Rittweger J. Effects of age and starting age upon side asymmetry in the arms of veteran tennis players: a cross-sectional study. Osteoporos Int. 2014;25(4):1389–1400. 10.1007/s00198-014-2617-5. [DOI] [PubMed] [Google Scholar]
18.Wilks DC, Winwood K, Gilliver SF, et al. Age-dependency in bone mass and geometry: a pQCT study on male and female master sprinters, middle and long distance runners, race-walkers and sedentary people. J Musculoskelet Neuronal Interact. 2009;9(4):236–246. [PubMed] [Google Scholar]
19.Chastin SF, Mandrichenko O, Helbostadt JL, Skelton DA. Associations between objectively-measured sedentary behaviour and physical activity with bone mineral density in adults and older adults, the NHANES study. Bone. 2014;64:254–262. 10.1016/j.bone.2014.04.009. [DOI] [PubMed] [Google Scholar]
20.Stiles VH, Metcalf BS, Knapp KM, Rowlands AV. A small amount of precisely measured high-intensity habitual physical activity predicts bone health in pre- and post-menopausal women in UK Biobank. Int J Epidemiol. 2017;46(6):1847–1856. 10.1093/ije/dyx080. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Sudlow C, Gallacher J, Allen N, et al. UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 2015;12(3):e1001779. 10.1371/journal.pmed.1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Doherty A, Jackson D, Hammerla N, et al. Large scale population assessment of physical activity using wrist worn accelerometers: the UK Biobank study. PLoS One. 2017;12(2):e0169649. 10.1371/journal.pone.0169649. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Kandola AA, Del Pozo CB, Osborn DPJ, Stubbs B, Choi KW, Hayes JF. Impact of replacing sedentary behaviour with other movement behaviours on depression and anxiety symptoms: a prospective cohort study in the UK Biobank. BMC Med. 2021;19(1):133. 10.1186/s12916-021-02007-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Littlejohns TJ, Holliday J, Gibson LM, et al. The UK Biobank imaging enhancement of 100,000 participants: rationale, data collection, management and future directions. Nat Commun. 2020;11(1):2624. 10.1038/s41467-020-15948-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Kim HS, Jeong ES, Yang MH, Yang SO. Bone mineral density assessment for research purpose using dual energy X-ray absorptiometry. Osteoporos Sarcopenia. 2018;4(3):79–85. 10.1016/j.afos.2018.09.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Umemura Y, Ishiko T, Yamauchi T, Kurono M, Mashiko S. Five jumps per day increase bone mass and breaking force in rats. J Bone Miner Res. 1997;12(9):1480–1485. 10.1359/jbmr.1997.12.9.1480. [DOI] [PubMed] [Google Scholar]
27.Deere K, Sayers A, Rittweger J, Tobias JH. A cross-sectional study of the relationship between cortical bone and high-impact activity in young adult males and females. J Clin Endocrinol Metab. 2012;97(10):3734–3743. 10.1210/jc.2012-1752. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Capozza RF, Cointry GR, Cure-Ramírez P, Ferretti JL, Cure-Cure C. A DXA study of muscle-bone relationships in the whole body and limbs of 2512 normal men and pre- and post-menopausal women. Bone. 2004;35(1):283–295. 10.1016/j.bone.2004.03.010. [DOI] [PubMed] [Google Scholar]
29.Bassey EJ, Rothwell MC, Littlewood JJ, Pye DW. Pre- and postmenopausal women have different bone mineral density responses to the same high-impact exercise. J Bone Miner Res. 1998;13(12):1805–1813. 10.1359/jbmr.1998.13.12.1805. [DOI] [PubMed] [Google Scholar]
30.Johnell O, Kanis JA. An estimate of the worldwide prevalence and disability associated with osteoporotic fractures. Osteoporos Int. 2006;17(12):1726–1733. 10.1007/s00198-006-0172-4. [DOI] [PubMed] [Google Scholar]
31.Bolam KA, van Uffelen JG, Taaffe DR. The effect of physical exercise on bone density in middle-aged and older men: a systematic review. Osteoporos Int. 2013;24(11):2749–2762. 10.1007/s00198-013-2346-1. [DOI] [PubMed] [Google Scholar]
32.Warden SJ, Mantila Roosa SM, Kersh ME, et al. Physical activity when young provides lifelong benefits to cortical bone size and strength in men. Proc Natl Acad Sci U S A. 2014;111(14):5337–5342. 10.1073/pnas.1321605111. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Kontulainen S, Kannus P, Haapasalo H, et al. Good maintenance of exercise-induced bone gain with decreased training of female tennis and squash players: a prospective 5-year follow-up study of young and old starters and controls. J Bone Miner Res. 2001;16(2):195–201. 10.1359/jbmr.2001.16.2.195. [DOI] [PubMed] [Google Scholar]
34.Mohebbi R, Shojaa M, Kohl M, et al. Exercise training and bone mineral density in postmenopausal women: an updated systematic review and meta-analysis of intervention studies with emphasis on potential moderators. Osteoporos Int. 2023;34(7):1145–1178. 10.1007/s00198-023-06682-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Fry A, Littlejohns TJ, Sudlow C, et al. Comparison of sociodemographic and health-related characteristics of UK Biobank participants with those of the general population. Am J Epidemiol. 2017;186(9):1026–1034. 10.1093/aje/kwx246. [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Burton E, Farrier K, Lewin G, et al. Motivators and barriers for older people participating in resistance training: a systematic review. J Aging Phys Act. 2017;25(2):311–324. 10.1123/japa.2015-0289. [DOI] [PubMed] [Google Scholar]
37.Hannam K, Deere KC, Hartley A, et al. A novel accelerometer-based method to describe day-to-day exposure to potentially osteogenic vertical impacts in older adults: findings from a multi-cohort study. Osteoporos Int. 2016;28(3):1001–1011. 10.1007/s00198-016-3810-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Biobank_PA_paper_supplementary_material_zjae017

Biobank_PA_paper_supplementary_material_zjae017.docx^{(29KB, docx)}

Data Availability Statement

UK Biobank data are available through an access procedure described at https://biobank.ndph.ox.ac.uk/showcase/exinfo.cgi?src=AccessingData.

[ref1] 1.Consensus development conference on osteoporosis. Hong Kong, April 1-2, 1993. Am J Med. 1993;95:1S–78S. [PubMed] [Google Scholar]

[ref2] 2.Borgström F, Karlsson L, Ortsäter G, et al. Fragility fractures in Europe: burden, management and opportunities. Arch Osteoporos. 2020;15(1):59. 10.1007/s11657-020-0706-y. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref3] 3.Johnell O, Kanis JA, Odén A, et al. Mortality after osteoporotic fractures. Osteoporos Int. 2004;15(1):38–42. 10.1007/s00198-003-1490-4. [DOI] [PubMed] [Google Scholar]

[ref4] 4.Weycker D, Li X, Barron R, Bornheimer R, Chandler D. Hospitalizations for osteoporosis-related fractures: economic costs and clinical outcomes. Bone Rep. 2016;5:186–191. 10.1016/j.bonr.2016.07.005. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref5] 5.Curtis EM, van der Velde R, Moon RJ, et al. Epidemiology of fractures in the United Kingdom 1988-2012: variation with age, sex, geography, ethnicity and socioeconomic status. Bone. 2016;87:19–26. 10.1016/j.bone.2016.03.006. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref6] 6.Burge R, Dawson-Hughes B, Solomon DH, Wong JB, King A, Tosteson A. Incidence and economic burden of osteoporosis-related fractures in the United States, 2005-2025. J Bone Miner Res. 2007;22(3):465–475. 10.1359/jbmr.061113. [DOI] [PubMed] [Google Scholar]

[ref7] 7.Johnell O, Kanis JA, Oden A, et al. Predictive value of BMD for hip and other fractures. J Bone Miner Res. 2005;20(7):1185–1194. 10.1359/JBMR.050304. [DOI] [PubMed] [Google Scholar]

[ref8] 8.Ireland A, Maden-Wilkinson T, McPhee J, et al. Upper limb muscle-bone asymmetries and bone adaptation in elite youth tennis players. Med Sci Sports Exerc. 2013;45(9):1749–1758. 10.1249/MSS.0b013e31828f882f. [DOI] [PubMed] [Google Scholar]

[ref9] 9.Wilks DC, Winwood K, Gilliver SF, et al. Bone mass and geometry of the tibia and the radius of master sprinters, middle and long distance runners, race-walkers and sedentary control participants: a pQCT study. Bone. 2009;45(1):91–97. 10.1016/j.bone.2009.03.660. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref10] 10.Rittweger J, Goosey-Tolfrey VL, Cointry G, Ferretti JL. Structural analysis of the human tibia in men with spinal cord injury by tomographic (pQCT) serial scans. Bone. 2010;47(3):511–518. 10.1016/j.bone.2010.05.025. [DOI] [PubMed] [Google Scholar]

[ref11] 11.Johansson J, Nordström A, Nordström P. Objectively measured physical activity is associated with parameters of bone in 70-year-old men and women. Bone. 2015;81:72–79. 10.1016/j.bone.2015.07.001. [DOI] [PubMed] [Google Scholar]

[ref12] 12.Hannam K, Deere KC, Hartley A, et al. Habitual levels of higher, but not medium or low, impact physical activity are positively related to lower limb bone strength in older women: findings from a population-based study using accelerometers to classify impact magnitude. Osteoporos Int. 2017;28(10):2813–2822. 10.1007/s00198-016-3863-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref13] 13.Elhakeem A, Heron J, Tobias JH, Lawlor DA. Physical activity throughout adolescence and peak hip strength in young adults. JAMA Netw Open. 2020;3(8):e2013463. 10.1001/jamanetworkopen.2020.13463. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref14] 14.Deere K, Sayers A, Rittweger J, Tobias JH. Habitual levels of high, but not moderate or low, impact activity are positively related to hip BMD and geometry: results from a population-based study of adolescents. J Bone Miner Res. 2012;27(9):1887–1895. 10.1002/jbmr.1631. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref15] 15.Rubin CT, Bain SD, McLeod KJ. Suppression of the osteogenic response in the aging skeleton. Calcif Tissue Int. 1992;50(4):306–313. 10.1007/BF00301627. [DOI] [PubMed] [Google Scholar]

[ref16] 16.Klein-Nulend J, van Oers RF, Bakker AD, Bacabac RG. Bone cell mechanosensitivity, estrogen deficiency, and osteoporosis. J Biomech. 2015;48(5):855–865. 10.1016/j.jbiomech.2014.12.007. [DOI] [PubMed] [Google Scholar]

[ref17] 17.Ireland A, Maden-Wilkinson T, Ganse B, Degens H, Rittweger J. Effects of age and starting age upon side asymmetry in the arms of veteran tennis players: a cross-sectional study. Osteoporos Int. 2014;25(4):1389–1400. 10.1007/s00198-014-2617-5. [DOI] [PubMed] [Google Scholar]

[ref18] 18.Wilks DC, Winwood K, Gilliver SF, et al. Age-dependency in bone mass and geometry: a pQCT study on male and female master sprinters, middle and long distance runners, race-walkers and sedentary people. J Musculoskelet Neuronal Interact. 2009;9(4):236–246. [PubMed] [Google Scholar]

[ref19] 19.Chastin SF, Mandrichenko O, Helbostadt JL, Skelton DA. Associations between objectively-measured sedentary behaviour and physical activity with bone mineral density in adults and older adults, the NHANES study. Bone. 2014;64:254–262. 10.1016/j.bone.2014.04.009. [DOI] [PubMed] [Google Scholar]

[ref20] 20.Stiles VH, Metcalf BS, Knapp KM, Rowlands AV. A small amount of precisely measured high-intensity habitual physical activity predicts bone health in pre- and post-menopausal women in UK Biobank. Int J Epidemiol. 2017;46(6):1847–1856. 10.1093/ije/dyx080. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref21] 21.Sudlow C, Gallacher J, Allen N, et al. UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 2015;12(3):e1001779. 10.1371/journal.pmed.1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref22] 22.Doherty A, Jackson D, Hammerla N, et al. Large scale population assessment of physical activity using wrist worn accelerometers: the UK Biobank study. PLoS One. 2017;12(2):e0169649. 10.1371/journal.pone.0169649. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref23] 23.Kandola AA, Del Pozo CB, Osborn DPJ, Stubbs B, Choi KW, Hayes JF. Impact of replacing sedentary behaviour with other movement behaviours on depression and anxiety symptoms: a prospective cohort study in the UK Biobank. BMC Med. 2021;19(1):133. 10.1186/s12916-021-02007-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref24] 24.Littlejohns TJ, Holliday J, Gibson LM, et al. The UK Biobank imaging enhancement of 100,000 participants: rationale, data collection, management and future directions. Nat Commun. 2020;11(1):2624. 10.1038/s41467-020-15948-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref25] 25.Kim HS, Jeong ES, Yang MH, Yang SO. Bone mineral density assessment for research purpose using dual energy X-ray absorptiometry. Osteoporos Sarcopenia. 2018;4(3):79–85. 10.1016/j.afos.2018.09.003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref26] 26.Umemura Y, Ishiko T, Yamauchi T, Kurono M, Mashiko S. Five jumps per day increase bone mass and breaking force in rats. J Bone Miner Res. 1997;12(9):1480–1485. 10.1359/jbmr.1997.12.9.1480. [DOI] [PubMed] [Google Scholar]

[ref27] 27.Deere K, Sayers A, Rittweger J, Tobias JH. A cross-sectional study of the relationship between cortical bone and high-impact activity in young adult males and females. J Clin Endocrinol Metab. 2012;97(10):3734–3743. 10.1210/jc.2012-1752. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref28] 28.Capozza RF, Cointry GR, Cure-Ramírez P, Ferretti JL, Cure-Cure C. A DXA study of muscle-bone relationships in the whole body and limbs of 2512 normal men and pre- and post-menopausal women. Bone. 2004;35(1):283–295. 10.1016/j.bone.2004.03.010. [DOI] [PubMed] [Google Scholar]

[ref29] 29.Bassey EJ, Rothwell MC, Littlewood JJ, Pye DW. Pre- and postmenopausal women have different bone mineral density responses to the same high-impact exercise. J Bone Miner Res. 1998;13(12):1805–1813. 10.1359/jbmr.1998.13.12.1805. [DOI] [PubMed] [Google Scholar]

[ref30] 30.Johnell O, Kanis JA. An estimate of the worldwide prevalence and disability associated with osteoporotic fractures. Osteoporos Int. 2006;17(12):1726–1733. 10.1007/s00198-006-0172-4. [DOI] [PubMed] [Google Scholar]

[ref31] 31.Bolam KA, van Uffelen JG, Taaffe DR. The effect of physical exercise on bone density in middle-aged and older men: a systematic review. Osteoporos Int. 2013;24(11):2749–2762. 10.1007/s00198-013-2346-1. [DOI] [PubMed] [Google Scholar]

[ref32] 32.Warden SJ, Mantila Roosa SM, Kersh ME, et al. Physical activity when young provides lifelong benefits to cortical bone size and strength in men. Proc Natl Acad Sci U S A. 2014;111(14):5337–5342. 10.1073/pnas.1321605111. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref33] 33.Kontulainen S, Kannus P, Haapasalo H, et al. Good maintenance of exercise-induced bone gain with decreased training of female tennis and squash players: a prospective 5-year follow-up study of young and old starters and controls. J Bone Miner Res. 2001;16(2):195–201. 10.1359/jbmr.2001.16.2.195. [DOI] [PubMed] [Google Scholar]

[ref34] 34.Mohebbi R, Shojaa M, Kohl M, et al. Exercise training and bone mineral density in postmenopausal women: an updated systematic review and meta-analysis of intervention studies with emphasis on potential moderators. Osteoporos Int. 2023;34(7):1145–1178. 10.1007/s00198-023-06682-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref35] 35.Fry A, Littlejohns TJ, Sudlow C, et al. Comparison of sociodemographic and health-related characteristics of UK Biobank participants with those of the general population. Am J Epidemiol. 2017;186(9):1026–1034. 10.1093/aje/kwx246. [DOI] [PMC free article] [PubMed] [Google Scholar]

[ref36] 36.Burton E, Farrier K, Lewin G, et al. Motivators and barriers for older people participating in resistance training: a systematic review. J Aging Phys Act. 2017;25(2):311–324. 10.1123/japa.2015-0289. [DOI] [PubMed] [Google Scholar]

[ref37] 37.Hannam K, Deere KC, Hartley A, et al. A novel accelerometer-based method to describe day-to-day exposure to potentially osteogenic vertical impacts in older adults: findings from a multi-cohort study. Osteoporos Int. 2016;28(3):1001–1011. 10.1007/s00198-016-3810-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Are associations between physical activity and bone mineral density in adults sex- and age-dependent? An analysis of the UK Biobank study

Gallin Montgomery

Mohamed Yusuf

Rachel Cooper

Alex Ireland

Abstract

Introduction