Osteoporosis affects skeletal integrity, whereby bone mineral density (BMD) is decreased in addition to microarchitectural deterioration (Kanis et al., 2019). This results in an increased risk of fractures impacting mobility (Kanis et al., 2019) and quality of life (Rhodes et al., 2000), with repercussions at an individual and societal level worldwide (Rashki Kemmak et al., 2020; Salari et al., 2021). In Europe, fracture rates and treatment costs are projected to surge by 25% (Willers et al., 2022). Recent European statistics have also demonstrated a stark contrast in the numbers of osteoporotic cases between males and females, with females affected nearly four times as much (Willers et al., 2022). This increased risk in women is largely attributed to the effect of menopause (The ESHRE Capri Workshop Group, 2010), and despite advances in prevention, screening and management of this condition, osteoporosis continues to remain a significant challenge worldwide (Salari et al., 2021; Willers et al., 2022). For this reason, females are the primary focus of this study.

Several pharmacological treatments and exercise interventions are relied upon to manage osteoporosis (Beck, 2022; Palacios, 2022). However, the difficulty with medication is finding the appropriate balance between evidence-based medicine and proven treatment strategies (Ragucci and Shrader, 2011; Gregson et al., 2022). This is further complicated when integrating suitable exercise interventions for the management of this condition. Previous studies have highlighted the effectiveness of exercise in the management of osteoporosis (Gupta and March, 2016; Koshy et al., 2022). However, significant consideration is required since exercise can substantially increase fall risk and needs to be adapted to the individual (Gupta and March, 2016; Benedetti et al., 2018). Furthermore, prevention with physical activity interventions has not always yielded positive effects (Rhodes et al., 2000; Borer, 2005; Ma et al., 2013). As a result, modalities such as whole-body vibration (WBV) were proposed to harness the potential of physical activity by combining simplicity, ease of administration and encouraging adherence. Since WBV can be easily adapted into an everyday routine, it has garnered a lot of attention over the years as a potential treatment modality for osteoporosis prevention, particularly in postmenopausal women (Marin-Puyalto et al., 2018). WBV involves the transmission of mechanical stimuli delivered via different vibration platforms (vertical, rotational, or lateral planes) that transfer forces to skeletal segments like other forms of exercise (Marin-Puyalto et al., 2018). The effects of WBV have been extensively investigated (Verschuren et al., 2004; Gusi et al., 2006; Von Stengel et al., 2011a; Von Stengel et al., 2011b; Slatkovska et al., 2011; Verschueren et al., 2011; Lai et al., 2013; Stolzenberg et al., 2013; Slatkovska et al., 2014; Liphardt et al., 2015; de Oliveira et al., 2019) resulting in disparities between studies.

Following seminal studies by Rubin et al. (2004) to evidence the impact of WBV on bone, abundant studies ranging from short to long-duration exposure to WBV have been tested (Gómez-Cabello et al., 2014; Kiel et al., 2015). In the 11-week study published by Gomez-Cabello et al. (2014), bone parameters in a relatively small population were examined using high-amplitude and high-frequency vibration stimuli. Particular focus was on dual-energy X-ray absorptiometry (DXA) derived parameters at the femoral neck and lumbar spine along with added peripheral quantitative computer tomography (pQCT) parameters examining the cortical and trabecular BMD of the radius and tibia. Despite this combination, no conclusive effects on bone were identified. Similarly, studies 6-month in length showed parallels in signal characteristics observed in the 11-week study; however, muscle strength and hip density were also incorporated into the protocol (Verschueren et al., 2004; Verschueren et al., 2011).

Interestingly, two studies of the same 6-month duration explored effects on bone turnover markers (cross-linked C-telopeptide of type I collagen and osteocalcin) alongside DXA outcomes using similar high-frequency and high-amplitude stimuli. Both reached similar conclusions with varying bone and serum results despite utilising similar vibration platforms (Verschuren et al., 2004; Sen et al., 2020). Some studies of much longer duration (12 months or greater) have shifted the focus more towards incorporating advancements in 3D imaging modalities, moving away from the reliance on traditional 2D DXA parameters to draw more in-depth conclusions associated with longitudinal changes in bone (Slatkovska et al., 2011). Studies have experimented with a range of frequencies (Slatkovska et al., 2011) and exposure intervals (Liphardt et al., 2015), using High Resolution Peripheral Computer Tomography (HR-pQCT) (Slatkovska et al., 2011; Liphardt et al., 2015) or novel Magnetic Resonance Imagining (MRI) techniques (Rajapakse et al., 2021) to examine the effects of WBV on bone geometry and microstructural parameters. However, the extensive focus on 3D imaging modalities has exacerbated differences of opinion regarding the efficacy of WBV on bone.

These studies alone highlight several gaps in the literature. Since the osteogenic response to physical activity reduces with age (Rubin et al., 1992), alterations in the length and frequency of sessions could be necessary (Gómez-Cabello et al., 2014). Weekly sessions have varied between one and three, with some varying the training intensity by modifying the signal characteristics and rest periods (Beck and Norling, 2010; Slatkovska et al., 2011; Sen et al., 2020). Despite modifications to signal attributes seen in studies to date, the role of amplitude in WBV bone response is still in question. Physical activity is known to produce a range of amplitudes depending on the type of exercise performed (Deere et al., 2016). Since WBV is a substitute for physical activity, it is essential to replicate this aspect as closely as possible. Frequency variations (high/low), varying amplitudes (<1 g/> 1 g), as well as exposure intervals to the vibration stimuli, have significantly differed between studies and are critical elements of any vibration signal (Prisby et al., 2008). The lack of studies investigating combined amplitude creates a new research opportunity. Physical activity levels prior to intervention were also alluded to as a possible explanation for the lack of observable effect (Rajapakse et al., 2021), which has not been a consideration in studies thus far. Accounting for this might reveal a relationship between WBV and osteogenic potential. Finally, several studies have already explored the response of formation and resorption of bone markers but have not investigated sclerostin, which is known to increase with age and fluctuate based on activity levels (Robling et al., 2008; Amrein et al., 2012). The inclusion of this could capture the subtle responses to combined amplitude stimuli.

In light of the differences outlined above, this study aimed to investigate the role of high-frequency and combined amplitude stimuli over 12 months on physically inactive postmenopausal women. The primary research objective was to investigate changes specific to the femoral neck, while the secondary objective was to measure bone geometry and microarchitectural changes when exposed to such stimuli. In addition, a 6-month follow-up after vibration training was added to observe any sustained osteogenic benefits. It was hypothesised that bone parameters would be improved, with positive osteogenic effects sustained following WBV therapy.

This non-randomised clinical trial matched participants according to their FRAX scores for major osteoporotic fracture (MOF) 10 years-absolute risk ranging from (3%–35%) to ensure group comparability. Statistical analysis was performed using 197 participants (Figure 1) having four timepoints (baseline, 6, 12 and 18 months). For all 197 participants, HR-pQCT image quality and common region were (≤ 3) and (≥ 70%) respectively (Table 1; Figure 1). Participants’ baseline characteristics can be found in (Tables 2, 3). Vibration protocol adherence was good with 110.02 ± 14.25 (mean ± SD) vibration sessions from a maximum of 130 completed by the vibration group. All within groups differences are presented in (Tables 4 –7) and significant between group differences are presented in (Table 8).

This non-randomised clinical trial matched participants according to their FRAX scores for major osteoporotic fracture (MOF) 10 years-absolute risk ranging from (3%–35%) to ensure group comparability investigated the effects of WBV on 197 postmenopausal women over 12 months, with an additional 6-month follow-up to observe post-vibratory outcomes. Treatment responses were measured using DXA, HR-pQCT, and serum analysis. Fat, lean body, and total body mass assessed by DXA very marginally decreased in the vibration group during the vibration phase. No significant differences at the femoral neck or for all other DXA and serum parameters were observed except for an 11% increase in CTX levels in the vibration group, a non-clinically relevant variation which remained present at month-18. Furthermore, no demonstrable benefits from vibration were detected for all bone geometry and microarchitectural parameters measured by HR-pQCT. Finite element analysis demonstrated a continuous decrease in failure load, and disproportionate stress loads were observed for the cortical and trabecular compartments in both groups. Finally, no changes in muscle parameters evaluated by HR-pQCT at the distal tibia were detected.

The primary objective was to assess the effect of WBV on the femoral neck BMD, given that this parameter is a strong predictor of hip and spine fractures in this type of population (Crandall et al., 2021). Some studies have found neutral effects of WBV at the hip and femoral neck (Lau et al., 2011; Ma et al., 2016; Oliveira et al., 2016; Jepsen et al., 2017; Luo et al., 2017) while one demonstrated positive effects (Sitjà-Rabert et al., 2012). Although the objective of previous studies has been on frequency variation and exposure length, none have incorporated combined amplitude in the experimental protocol. Therefore, unique to this study was the incorporation of high and low-amplitude stimuli with an incremental increase throughout the vibration protocol. By further tailoring the signal characteristics that have generally been sufficient to stimulate bone formation (Santos et al., 2017), this protocol was expected to better replicate the role of mechanical stimulation, in particular the ground reaction forces produced during regular jumping and physical activity and to demonstrate its tolerance in this population. However, despite this combination of amplitude and frequency, no discernible benefits were observed at the femoral neck, spine, or HR-pQCT parameters. The continued reduction in failure load with unequal load distributions between cortical and trabecular compartments is consistent with elevated fracture risk (Boutroy et al., 2008) in this population, not significantly negated by vibration therapy.

The results from this study are difficult to compare, given the unique signal characteristics. However, considering studies 12 months in duration, Slatkovska et al. (2011) share several similarities (population size, age range and 20-min training sessions with predominantly European participants) with the current study. Although Slatkovska et al. (2011) mainly focused on frequency variation (90 Hz and 30 Hz), the frequency range in the present study (30 Hz–50 Hz) fell between the two training groups. This is a particular strength as it helps to further expand upon the range of frequencies previously tested while incorporating the application of combined amplitude. However, both Slatkovska et al. (2011) and the present study failed to see any benefits on bone after 12 months of WBV exposure. Another one-year longitudinal study (Liphardt et al., 2015), which incorporated a follow-up period to observe sustained osteogenic benefits of WBV, also failed to detect changes in the bone and observed no improvement in load distribution between groups, which indicates no improvement in fracture risk. Though these studies employed different platforms and signal attributes, the consistency in HR-pQCT results of bone geometry and microarchitecture remains a crucial finding.

The current results can be further contrasted with other studies using the PowerPlate device. Sen et al. (2020) and Verschueren et al. (2004) utilised high-frequency stimuli for 6 months to measure BMD changes at one or more sites (hip, femoral neck, and lumbar region) along with serum bone turnover markers (CTX and osteocalcin). In both studies, the vibration groups demonstrated an improvement in BMD; however, bone turnover marker results were inconsistent. In the present study, no significant benefits were observed at the crucial DXA sites. However, from baseline to the end of the training period, CTX was elevated in both groups with marginally higher concentrations in the vibration group. Nevertheless, when comparing months 12 and 18 alone these differences were not statistically different.

Several explanations could account for the marginal increase in CTX concentrations. Firstly, the two previous studies (Verschueren et al., 2004; Sen et al., 2020) were of much shorter duration, hence CTX would not have changed as it did in this study. Secondly, when taken together, the increased BAP activity with the rising CTX concentrations is indicative of the stimulated bone remodeling process in postmenopausal women (Eastell et al., 2016). Finally, the marginal elevation in CTX concentrations in the vibration group could also be linked with the small reduction in body mass resulting from WBV exposure. The impact of weight loss attributed to dieting and/or exercise on bone turnover markers (CTX) has been previously reported (Eastell and Szulc, 2017).

Sen et al. (2020) and Verschueren et al. (2004) examined the response of WBV on osteocalcin, as a bone formation marker. No group differences were reported by Verschueren et al. (2004) while a decrease was reported by Sen et al. (2020). In the present study, the effects on bone formation were studied using P1NP which was not altered due to vibration training. Elevation in P1NP was observed by Corrie et al. (2015) in a study using a PowerPlate device. However, the age of the participants was considerably higher (79–82 years) than the current study, the training duration was only twelve weeks and focused on both men and women in a small population. It is also important to note that no BMD measurements were taken during this study.

WBV response on bone turnover markers has not been extensively studied, however, the few that have investigated its effects (Russo et al., 2003; Verschueren et al., 2004; Corrie et al., 2015; Kiel et al., 2015) focused mainly on markers of formation and resorption. Even fewer have explored the role of sclerostin in the context of WBV. Sclerostin has emerged over the years as a marker essential to skeletal physiology and homeostasis and is found to be mostly expressed by osteocytes (Poole et al., 2005). It interacts with the Wnt signalling pathway (Baron and Kneissel, 2013) leading to the formation or resorption of bone. Moreover, sclerostin concentrations have been shown to increase or decrease based on the response to mechanical stimuli (Robling et al., 2008) and are also known to increase with age (Amrein et al., 2012). The goal of measuring sclerostin was to evaluate if the signal characteristics in this study could be detected. Since no differences between groups were observed, this further highlights that the effects of combined amplitude stimuli were undetected. However, it does pose the question of whether combined amplitude exposure of longer duration could produce significant group differences.

Previous research has also explored age-related responses to WBV (Slatkovska et al., 2011; Marín-Cascales et al., 2018); however, the relevance within the context of high-frequency and combined amplitude training has not been tested. Moreover, since fracture risk increases substantially beyond age 64 (Arlot et al., 1997; Emaus et al., 2006), this breakpoint provides an opportunity to ascertain dose and bone response associated with age, resulting from WBV exposure. These age-related differences were explored using retrospective statistical analysis by incorporating age groups (≤ 64 and > 64) into the fixed effects of the aforementioned statistical model to further examine whether any effects of WBV could be detected (Supplementary Material S8–S11). This was done to elucidate any preventative effects given that WBV is relied upon more as a preventive treatment modality. Despite such attempts, no significant differences between age groups were noted. Both bone and serum parameters demonstrated no significant improvement, and more importantly, parameters including P1NP, which is a marker sensitive to bone formation (Gillett et al., 2021), exhibited no significant improvement resulting from vibration therapy. As a result, this further substantiates that WBV has neither osteogenic benefits nor negates fracture risk in this type of physically inactive population regardless of the signal characteristics.

The minimal reductions in fat, lean body and total body mass observed towards the end of the vibration protocol could be linked with a few different possibilities. Firstly, the vibration group was aware that they were undergoing treatment, and the possible impact of this awareness cannot be discounted as an influencing factor. As a result, participants could have been influenced in their dietary habits. Nevertheless, aside from this possibility, the results indicate an interesting effect between groups in that the control group saw more significant weight increases over time than the vibration group. The present study is not the only one to report this effect since the impact of vibration on weight has previously been reported by many (Cristi-Montero, 2013). Several hypotheses exist and involve the role of the sympathetic nervous system (Alavinia et al., 2021), osteocalcin and sclerostin (Wang et al., 2021). A recent pilot study also suggested the role of SMP30 on fat mass, given its role in lipid regulation (Pérez-Gómez et al., 2020). Further studies are needed to fully understand the mechanism as it remains largely unknown and poorly understood.

To the best of our knowledge, this study was the first to comprehensively investigate the effects of WBV by including an additional 6-month follow-up period to examine bone geometry and microarchitecture, muscle (HR-pQCT distal tibia), serum and DXA with a particular focus on the femoral neck, using high-frequency and combined amplitude training. Despite this, there are several limitations:

1) This study investigated the effects of WBV in a group of Caucasian European postmenopausal women; however, future research will have to explore the effect of menopause on WBV using broader clinical criteria by accounting for the age, stage of menopause, level of exercise and diverse ethnicities.

In conclusion, this 18-month non-randomised clinical trial matched participants according to their (MOF) FRAX scores for major osteoporotic fracture and used WBV for 12 months with combined amplitude and high-frequency vibration to improve bone geometry and bone microarchitecture. This study failed to significantly detect an improvement in bone outcomes of physically inactive osteopenic postmenopausal women. Although the 12-month training protocol was well tolerated, DXA BMD measurements of the lumbar spine, femoral neck and hip were not significantly improved compared to control. Moreover HR-pQCT analysis of cortical and trabecular compartments, including finite element analysis did not demonstrate significant improvement resulting from WBV training. Serum markers (P1NP, BAP and sclerostin) also showed no response to mechanical stimuli. A marginal increase in CTX concentrations was observed; however, there was no indication that this could have resulted from vibration exposure. Nevertheless, future studies could expand upon these results by accounting for stratification in age, and stage of menopause to draw further conclusions.