Whole body vibration training is promoted as a potentially safe, low-impact alternative to current therapeutic modalities to combat bone loss in older adults. Whole body vibration is without the potential side effects of pharmacological intervention or risks associated with high impact or strenuous exercise. The mechanism of vibration stimulus on bone is not wholly understood; however, it is hypothesized that the anabolic effect of vibration on bone may be a result of stress exerted on bone resulting in increased fluid flow [
Animal studies investigating the effects of vibration stimulus on bone reported positive results, with improvements in bone strength, formation rates, bone mineral content, and bone mineral density (BMD) following low-magnitude, high-frequency vibration exposure [
In this study, we specifically examined the dose-response relationship (frequency of exposure) of vibration stimulus on markers of bone turnover in postmenopausal women in a community setting. We hypothesized the application of low-frequency, low-magnitude vibration stimulus for 8 weeks would increase the bone formation marker, ALP, and decrease the bone resorption marker, NTx/Cr, in a dose-dependent relationship compared with a sham vibration exposure.
Postmenopausal women were recruited from northern suburbs of Sydney, Australia by means of advertisements in local newspapers, posters in local GPs surgeries, letterbox drops, and advertisements to existing members of a large suburban community club via newsletters and posters within the club.
Participants were considered eligible if they were postmenopausal for at least 12 months; willing to continue taking any bone altering medications or supplements they were previously taking for the duration of the study, including calcium, vitamin D, bisphosphonates, and hormone replacement therapy (HRT); able to stand unassisted for sustained periods of time (i.e., 20 minutes); willing to attend testing and training sessions as determined by the researchers; able to travel independently to and from the testing/training venue. Exclusion criteria included cognitive impairment, contraindications to vibration platform training (including pacemaker and fracture within the past six months), and the diagnosis of diseases other than osteoporosis affecting bone. The study protocol was approved by the Human Research Committee, University of Sydney, Australia (June 2007) and was registered with the Australian Clinical Trials Registry (ACTRN12607000491460). All participants gave informed written consent prior to enrolment.
The study was a randomized placebo-controlled clinical trial, with a complete case statistical analysis design and secondary intention-to-treat analytic strategy. The vibration protocol consisted of an 8-week whole body vibration exposure program. Serum and urine markers of bone turnover were measured at baseline and at the conclusion of the study. All testing and training was conducted at Freshwater Health and Fitness, Harbord Diggers-Mounties Group, Harbord, NSW, Australia. Participants were recruited and trained over one season (late autumn to winter 2007).
Participants were randomized into the Sham group, 1×/week group, or 3×/week group following completion of baseline assessments. Randomization was performed offsite by a researcher who was not involved in testing and training of participants using computer-generated randomly permuted blocks (
Participants who were asked to remain in a standing position on a whole body vibration device. The purpose built device produced a synchronous vibration resulting in vertical acceleration, engineered by Australian Catholic University. Low-frequency, low-magnitude vibration was applied (12 Hz, 0.5 mm peak-to-peak displacement, 0.3 g), consistent with current research that suggests that low magnitude is anabolic to bone [
Participants in the 1×/week and 3×/week groups were required to attend sessions either once or three times a week over eight weeks for 20 minutes of intermittent vibration, respectively, during which time, vibration was applied for one minute, followed by one minute of rest. The osteogenic effect of vibration is reported to be greater when rest breaks are incorporated into the vibration stimulus, as intermittent exposure provides osteoblasts and mechanoreceptors of bone at rest from vibration stimulus, preventing insensitivity to vibration that can occur during prolonged vibration exposure [
Outcome measures were conducted at baseline (before randomization) and at followup (completion of 8 sessions for Sham and 1×/week groups and 24 sessions for 3×/week group).
All outcomes were blindly assessed and analyzed at baseline, as assessment occurred prior to randomization. Outcomes assessed included anthropometric measures, demographic characteristics (including socioeconomic status, smoking habits, and caffeine consumption), habitual physical activity, 25-OH vitamin D status, and body composition measurement by bioelectric impedance analysis (fat-free mass (FFM), skeletal muscle mass (SMM), fat mass (FM)) using standard equations for older adults [
Markers of bone formation (bone-specific alkaline phosphatase (ALP)) and resorption (N-telopeptide X/Creatinine (NTx/Cr)) were assessed via blood and urine tests performed without batching by an independent laboratory that was blinded to group allocation. Blood samples were collected between 72 and 120 hours following the last vibration exposure session to standardize previously reported acute bout effects on markers of bone metabolism [
Compliance for each participant was calculated as a percentage of the prescribed sessions attended. Throughout the eight week study period, a weekly questionnaire was completed in person or via phone calls, in order to monitoring possible adverse effects from the vibration exposure and any changes in health status.
Based on results from previous research investigating the anabolic effects of exercise on bone markers [
Statistical analyses were performed using StatView statistical software package (Version 5.0 SAS Institute, Cary, NC). Data distributions were inspected for normality. Normally distributed data were described using mean ± SD and non-normally distributed data using median and ranges. Complete case analyses compared the differences in primary and secondary outcomes between the 3×/week, 1×/week, and Sham control group using all available data independent of compliance level. A secondary sensitivity analysis was performed using an intention-to-treat analysis with conservative imputation of results via last value carried forward method. Analysis of covariance (ANCOVA) models were constructed to compare 3×/week, 1×/week, and sham control groups, using the percentage change scores (posttest − pretest/pretest × 100) as the dependent variable, adjusted for baseline values of NTx/Cr and ALP. Variables that were potentially related to the outcome of interest such as age, month of assessment, percent change scores for fat-free mass (FFM), skeletal muscle mass (SMM), body weight, vitamin D, habitual physical activity, use of HRT, and height at baseline were all identified a priori as potential confounders of the treatment effect that could be used as covariates in analysis of covariance (ANCOVA) models of change scores. Variables with clinically importance between group differences were included in ANCOVA models. Post-hoc least significant difference (LSD)
Recruitment and enrolment of 46 participants occurred from May to July, 2007. Participant flow through the study is presented in Figure
Recruitment flowchart.
Baseline participant characteristics are presented in Table
Demographics and health status.
Variable | Whole cohort ( | Sham ( | 1×/wk ( | 3×/wk ( |
---|---|---|---|---|
Age (yrs) | 59.8 ± 6.2 | 59.8 ± 5.2 | 60.9 ± 6.5 | 58.9 ± 7.1 |
Height (cm) | 160.2 ± 5.8 | 156.9 ± 4.9 | 162.5 ± 4.6 | 161.6 ± 6.2 |
Body weight (kg) | 62.80 | 62.15 | 64.67 | 62.88 |
Body mass index (kg/m2)§ | 24.38 | 25.22 | 24.64 | 23.62 |
History of smoking (%) | 32.6 | 28.6 | 31.3 | 37.5 |
Number of medications/day | 3.3 ± 2.6 | 3.2 ± 3.2 | 3.8 ± 3.3 | 2.5 ± 3.5 |
Vitamin D prescription (%) | 17.4 | 12.5 | 14.3 | 25.0 |
Calcium prescription (%) | 43.5 | 50.0 | 28.6 | 50.0 |
Bisphosphonate prescription (%) | 13.0 | 6.3 | 14.3 | 18.8 |
HRT prescription (%) | 8.7 | 0.0 | 14.3 | 12.5 |
Years since menopause | 7.25 | 9.75 | 6.25 | 6.50 |
Participation in regular structured exercise (%) | 71.7 | 68.8 | 71.4 | 75.0 |
History of osteoporosis, osteopenia, or osteoporotic fracture | 30.4 | 18.8 | 28.6 | 43.8 |
All data presented as mean ± SD for normally distributed data or median (range) for nonnormally distributed data unless otherwise specified.
§Body mass index: an indicator of body fat calculated by weight (kg)/Height (m)2. Normal values range from 18.5 to 24.9 kg/m2. Values ≥25 kg/m2 are considered overweight and ≥30 kg/m2 obese.
Vitamin D status varied across the cohort, with the majority of participants within the normal range (80%); however, 20% were either classified as mildly or moderately deficient. Following the 8-week study period, Vitamin D levels decreased in all groups, with only 68% of participants classified as normal and an increase in the number of women who were mildly or moderately deficient (33%). This pattern is consistent with expected seasonal change, in which vitamin D levels tend to decrease over winter months [
The NTx/Cr levels were not different between women stratified by years since menopause (
Following eight weeks of vibration exposure, NTx/Cr levels changed differentially among the three groups. Specifically, NTx/Cr decreased in the 1×/week and 3×/week groups and increased in the sham group, indicating a significant effect of group allocation on bone resorption when compared by repeated measures ANOVA (
Primary outcomes.
Sham ( | 1×/week ( | 3×/week ( | Repeated measures ( | Effect size (95% CI) | |||||||
Pre | Post | Pre | Post | Pre | Post | Time | Group × time | S versus 1× | S versus 3× | 1× versus 3× | |
ALP (nmol/L) | 13.35 | 13.20 | 14.85 | 13.60 | 14.01 | 13.30 | .08 | .27 | 0.13 | −0.19 | −0.25 |
NTx/ | 42 | 41 | 38 | 42 | 42 | 36.5 | .21 | .03* | −0.26 | −0.96 | −0.28 |
All data presented as median (range).
*Indicates a significant difference between the three groups (
Group effect: Absolute changes in NTx/Cr (nm/mmCr) values presented as Mean ± SD. Analysis by ANCOVA model with changes score as dependent variable and group assignment, baseline value of Ntx/Cr, age, height, use of HRT, and years since menopause as independent variables. Group effect for model was
The bone formation marker, ALP, tended to decrease in all groups over the eight-week period (
Due to low numbers not handled by stratification, the number of participants taking bisphosphonates and on HRT therapy were uneven between groups. In order to correct for any bias due to the potential confounding effect of bisphosphonate and HRT use on bone metabolism, secondary analysis was conducted excluding those participants (
In the three study groups, the reported median vibration training compliance was 100 (0–100%), with 87.5% of randomized participants completing the study. One participant in the sham group withdrew from the study due to family illness, one participant in the 1×/week group withdrew due to illness and family problems, and two participants in the 3×/week group withdrew prior to their first vibration exposure session due to inability to commit to study requirements.
Adverse effects were reported by five participants over the course of the study. Three participants reported dizziness while standing on the vibration platform (sham,
In this study, we investigated the frequency dose-response relationship of whole body vibration stimulus on markers of bone turnover in postmenopausal women in a community setting. We showed, for the first time, that low-magnitude, low-frequency vibration exposure (0.3 g) 3 times per week reduced a marker of bone resorption (NTx/Cr) in postmenopausal women. The net decrease in NTx/Cr in the 3×/week group compared with the sham group was 34.6%, which was not only statistically significant, but also potentially
Only one previous study by Iwamoto et al. has investigated the effects of vibration on urinary NTx, reporting a nonsignificant decrease following high frequency, low-magnitude vibration exposure over 12 months [
Our study suggests a dose-response relationship between vibration exposure and reduced bone resorption. Specifically, vibration exposure 3×/week significantly reduced NTx/Cr when compared with the sham group, but this difference was not seen in the 1×/week group. A large effect size suggests that sample size was adequate to highlight group differences between the 3×/week and sham group (ES = −0.96). However, a small effect size was obtained when comparing the 1×/week group to the sham group, with low statistical power a potential factor in this lack of significance. Larger studies are needed to confirm this lack of efficacy of once-weekly training. However, it remains unknown whether vibration stimulus 3×/week is optimal or whether similar or greater changes in NTx/Cr could be achieved with higher exposure to vibration (number of days per week or minutes per session).
The bone formation marker ALP decreased similarly in all groups over the study period, which may be explained by the late autumn to winter period of recruitment into the study [
The lack of improvement in bone formation markers observed in our study is consistent with previous research in older adults. The effect size for changes in ALP between groups in our study were small, suggesting that a larger sample size was required to find a group effect, if one existed, on bone formation levels. The significant reduction in the bone resorption marker we observed is a novel finding after vibration exposure in humans. Notably, the biological plausibility of this finding is supported by a recent study of murine osteocytes in which low-magnitude (0.3 g), high-frequency vibration similar to our protocol increased soluble inhibitors of osteoclast formation at both the transcript and protein level. Resorption activity of the exposed osteocytes was also lowered in response to this vibration stimulus [
Large consumption of caffeine is associated with increased risk of osteoporosis and decreased bone formation [
Compliance for traditional treatments for osteoporosis, including pharmacological aids, have been shown to be low in previous studies [
Limitations to this study include sample size and intervention. The intervention period of eight weeks may not have been sufficient to observe larger effects on markers of bone metabolism. Previous studies have shown changes in bone outcomes typically occur over longer intervention periods, especially if architectural changes in bone are to be observed [
Vibration exposure protocols have varied widely in all studies to date investigating the effect of vibration on bone outcomes. While our study compared two vibration intervention groups differing in the number of sessions per week, no human studies have directly investigated if any differences exist in bone outcomes between continuous and intermittent vibration or if a difference exists in the effect of vibration on bone in participants standing still on the platform compared with participants performing dynamic and static exercise while on the platform. Another underexplored aspect of the research is the direction of vibration that elicits the greatest gains in bone outcomes, that is, reciprocal or oscillating. At present, no standard approach exists to counteract the attenuation effects of wearing footwear while on the vibration platform and differences between receiving vibration exposure while wearing shoes or while barefoot, as well as with locked, soft, or flexed knee positions should be investigated. Future research needs to establish whether vibration exposure effects on bone are retained after the vibration stimulus is removed, and the time course of detraining. It is unknown whether gender, nutritional status, hormonal status, bone density, or use of medications affecting bone remodeling influences the efficacy of vibration exposure, and studies comparing responses across such cohorts are required. Finally, the molecular mechanisms which may underlie vibration effects on bone turnover are largely unknown.
Low-frequency, low-magnitude vibration (12 Hz, 0.3 g) three times a week leads to a potentially clinically meaningful 34.6% reduction in NTx/Cr, a marker of bone resorption, whereas one day per week exposure appears insufficient. Further studies are required to extend and confirm these findings, determine the optimal dose of vibration exposure, and determine whether this decreased resorption is sustained and ultimately leads to increased bone density, tensile strength, and reduced risk of fragility fracture. The effect of long-term vibration on bone resorption may be clinically relevant if the changes we observed after 24 sessions were to be sustained or magnified with greater vibration dosage.
The authors declare that there is no conflict of interests.
The authors are grateful to Dr. Evan Atlantis for assistance with baseline assessments and all the participants who volunteered their time. This study was performed in the facilities at the Faculty of Heath Sciences, University of Sydney and at the Freshwater Health and Fitness, NSW, Australia. This study was performed in partial fulfillment of requirements for BAppSc (Hon) degree for Sarah Turner. The results of this paper were presented in part at The Annual Meeting for the Gerontological Society of America, 2008, and at the IOF World Congress on Osteoporosis, 2008. This study was supported by the Discipline of Exercise and Sport Science, University of Sydney, $700. Vibration platforms were engineered and provided by the Australian Catholic University.