Potential of Whole-Body Vibration in Parkinson’s Disease: A Systematic Review and Meta-Analysis of Human and Animal Studies

Simple Summary Exercise has shown to have a positive impact on both motor and non-motor functions in Parkinson’s Disease patients. However, particularly in later stages of the disease, reduced cognitive function and motor capacity may lead to an inability to stay physically active. Therefore, alternative strategies for patients with Parkinson’s Disease are necessary to minimize burden for patients, their families and public health care. Whole-Body Vibration could be such an alternative. Whole-Body Vibration is an exercise or treatment method in which subjects are exposed to a mechanical vibration while sitting, standing or exercising on a vibrating platform. Whole-Body Vibration is currently used for physiotherapy, sports and rehabilitation purposes. Whole-Body Vibration treatment is interesting because it affects both the body and brain. The potential of Whole-Body Vibration for, specifically, Parkinson’s Disease patients should be clarified for further application. For this purpose, we conducted an extensive systematic review of the articles investigating the effects of Whole-Body Vibrations (1) on animals and humans with Parkinson’s Disease and (2) on neuropathological Parkinson’s Disease mechanisms. The results show some potential of Whole-Body Vibration for Parkinson’s Disease patients. The recommendations provided by this review can be used by researchers and rehabilitative practitioners implementing Whole-Body Vibration as a treatment for Parkinson’s Disease patients. Abstract (1) Background: When the severity of Parkinson’s Disease (PD) increases, patients often have difficulties in performing exercises. Whole-Body Vibration (WBV) may be a suitable alternative. This systematic review aims to clarify if WBV shows potential as rehabilitative therapy for PD patients. (2) Methods: We searched several databases for controlled trials investigating the effects of WBV (1) on PD populations and (2) PD neuropathological mechanisms. We included both human and animal studies and performed meta-analyses. (3) Results: The studies on PD populations (14 studies) show an overall significant, but small, effect in favor of WBV (Hedges’ g = 0.28), for which the effects on stability (Hedges’ g = 0.39) and balance (Hedges’ g = 0.30) are the most prominent. The studies on the neuropathological mechanisms (18 studies) show WBV effects on neuroinflammation (Hedges’ g = –1.12) and several specific WBV effects on neurotransmitter systems, growth factors, neurogenesis, synaptic plasticity and oxidative stress. (4) Conclusions: The effects of WBV on human PD patients remains inconclusive. Nevertheless, WBV protocols with sufficient duration (≥3 weeks), session frequency (≥3 sessions/week) and vibration frequency (≥20 Hz) show potential as a treatment method, especially for motor function. The potential of WBV for PD patients is confirmed by the effects on the neuropathological mechanisms in mostly non-PD populations. We recommend high-quality future studies on both PD patients and PD mouse models to optimize WBV protocols and to examine the neuropathological mechanisms in PD populations.


Simple Summary:
Exercise has shown to have a positive impact on both motor and non-motor functions in Parkinson's Disease patients. However, particularly in later stages of the disease, reduced cognitive function and motor capacity may lead to an inability to stay physically active. Therefore, alternative strategies for patients with Parkinson's Disease are necessary to minimize burden for patients, their families and public health care. Whole-Body Vibration could be such an alternative. Whole-Body Vibration is an exercise or treatment method in which subjects are exposed to a mechanical vibration while sitting, standing or exercising on a vibrating platform. Whole-Body Vibration is currently used for physiotherapy, sports and rehabilitation purposes. Whole-Body Vibration treatment is interesting because it affects both the body and brain. The potential of Whole-Body Vibration for, specifically, Parkinson's Disease patients should be clarified for further application. For this purpose, we conducted an extensive systematic review of the articles investigating the effects of Whole-Body Vibrations (1) on animals and humans with Parkinson's Disease and (2) on neuropathological Parkinson's Disease mechanisms. The results show some potential of Whole-Body Vibration for Parkinson's Disease patients. The recommendations provided by this review can be used by researchers and rehabilitative practitioners implementing Whole-Body Vibration as a treatment for Parkinson's Disease patients.
Abstract: (1) Background: When the severity of Parkinson's Disease (PD) increases, patients often have difficulties in performing exercises. Whole-Body Vibration (WBV) may be a suitable alternative. This systematic review aims to clarify if WBV shows potential as rehabilitative therapy for PD patients.
(2) Methods: We searched several databases for controlled trials investigating the effects of WBV (1) on PD populations and (2) PD neuropathological mechanisms. We included both human and animal studies and performed meta-analyses. (3) Results: The studies on PD populations (14 studies) show an overall significant, but small, effect in favor of WBV (Hedges' g = 0.28), for which the effects on stability (Hedges' g = 0.39) and balance (Hedges' g = 0.30) are the most prominent. The studies on the neuropathological mechanisms (18 studies) show WBV effects on neuroinflammation (Hedges' g = -1.12) and several specific WBV effects on neurotransmitter systems, growth factors, neurogenesis, synaptic plasticity and oxidative stress. (4) Conclusions: The effects of WBV on human PD patients remains inconclusive. Nevertheless, WBV protocols with sufficient duration (≥3 weeks), session frequency (≥3 sessions/week) and vibration frequency (≥20 Hz) show potential as a treatment method, especially for motor function. The potential of WBV for PD patients is confirmed by the effects on the neuropathological mechanisms in mostly non-PD populations. We recommend high-quality future studies on both PD patients and PD mouse models to optimize WBV protocols and to examine the neuropathological mechanisms in PD populations.

Introduction
Parkinson's Disease (PD) is a common neurodegenerative disease affecting millions of people worldwide, especially the aging population [1]. The hallmark pathogenesis of PD is defined by the progressive degeneration of dopaminergic nigrostriatal systems, manifesting in a wide range of clinical phenotypes, including motor (bradykinesia, tremor, muscle rigidity and postural instability) and non-motor (depression, anxiety, fatigue, cognitive dysfunction and dementia) functions [2,3]. The demographical transition from predominantly young to older populations, causes an increase in the prevalence and incidence of PD [4]. Therefore, the incorporation of rehabilitative strategies for PD is necessary to minimize the burden for the patient, their families and public health care [1].
The treatment guidelines for PD management include physical exercise as an adjuvant to pharmacological treatments and surgery [5]. Exercise has shown to have a positive impact on both motor and non-motor functions by ameliorating clinical symptoms and modulating brain maintenance and plasticity in PD patients [6,7]. However, particularly in the later stages, reduced cognitive function and motor capacity may lead to the inability to stay physically active [8]. When PD patients are not (cap)able of performing exercise, an alternative approach is needed to fully benefit from its positive effects. Whole-Body Vibration (WBV) may be a suitable alternative.
WBV has been considered as an alternative for active exercise. A major advantage of WBV is that it can be used by people who cannot or have severe difficulties to participate in active exercise programs. Participants may be too frail or may not have the physical condition to perform active exercise, next to motivational or mental problems. In general, WBV does not take much effort; WBV procedures are relatively brief, and can be performed indoors at any moment of the day with limited or without supervision. The benefits of WBV on motor performance [9][10][11][12] are, to a large extent, comparable to those reported after active exercise; hence, it can be viewed as a form of passive exercise.
During WBV, a vibration is transmitted through the body by maintaining an active (e.g., performing squats) or passive posture (e.g., sitting or lying) on a mechanically vibrating platform ( [13] and references therein). The vibrations can be categorized into harmonic or random oscillations with synchronous vertical motions or motions around the sagittal axis [14] (Figure 1). The vibrations are defined further by their frequency (generally between 5-60 Hz), amplitude or peak-to-peak displacement (generally between 0.5-4 mm peak-topeak) and temporal aspects, such as number (generally between 1-10 bouts) and durations of bouts (generally between 30 s and 10 min, also depending on the number of bouts). Patients with PD may be potential beneficiaries of WBV in view of the multidimensional effects of WBV (see, for review, [15]). Several studies have shown that WBV improves muscle strength [16], postural control [17], bone health [18] and balance [19] in older adults. It has also been suggested that WBV can improve bone fractures [20] and wound healing [21] in animals. However, not much is known about the impact of WBV on foot health. In PD patients, proper foot health is in general relevant for good motor performance [22,23]. In people with type II diabetes mellitus, a protocol of eight weeks of WBV improved foot health [24], but so did the controls that followed the same procedure without the actual vibrations. Hence, the improvement did not depend on the actual vibrations, and it should be stressed that WBV did not interfere with foot health improvements. Furthermore, both animal and human studies [25][26][27] provide empirical support for a potential value of WBV in improving cognition. Moreover, WBV research has been performed in healthcompromised individuals with both motor and non-motor deficits, such as stroke [28], multiple sclerosis [29,30] and dementia [31], indicating that WBV intervention is feasible in health-compromised patients.
In line with this, reviews have been published on the effects of WBV in PD patients [32][33][34][35][36][37]. However, these studies have a limited scope of view because they are solely based on human research with outcome measures restricted to the sensorimotor and functional levels. Given the wide range of clinical phenotypes ranging from mild motor disturbances to severe motor, cognitive, affective and sleep disturbances [2], an overarching perspective of the potential of WBV exercise in PD patients is needed.
The purpose of this study is to provide a systematic overview with meta-analysis in order to clarify if WBV shows potential as rehabilitative therapy for PD patients. We examine two research questions: (1) What are the effects of WBV on humans and animal models with PD? (2) Are there potential ameliorative effects of WBV on the neuropathological mechanisms of PD? Outcome measures on cellular, brain and functional levels are included to provide a complete perspective.

Methods
We followed the PRISMA guidelines for reporting the systematic reviews [38]. This review was registered in Prospero under ID number CRD42022351823.

Information Sources
We conducted a systematic search using the following digital online databases: PubMed, EMBASE, Scopus and Web of Science. These databases were accessed online in March 2021.

Search Strategy
The search strategy was designed to examine (1) the effects of WBV on humans and animal models with PD, and (2) the potential ameliorating effects of WBV on neuropathological mechanisms of PD. Key search terms related to WBV training were combined with terms specified to PD itself (research question 1) or terms related to the neuropathology of PD (research question 2).
The search terms for the neuropathology of PD were determined by empirical evidence of dopaminergic and non-dopaminergic neurotransmitter losses [2,[39][40][41], programmed cell death [42], oxidative stress [2,43] and Lewy Bodies [44] in humans and PD animal models. Furthermore, 'cellular', 'brain' and "blood" search terms were added to amplify the search for potential mechanisms.
We used the following key search terms for PubMed and adapted this strategy for the other databases (Table A1)

Study Selection
We removed duplicates from the records identified by the databases with the Mendeley reference management tool. Two authors (Y.L.A.A. and G.A.) screened the titles and abstracts based on the eligibility criteria. Subsequently, one author (Y.L.A.A.) screened full texts for inclusion in the qualitative synthesis based on the same eligibility criteria. In case of doubts, the other author (G.A.) screened the full texts as well. Disagreements on the overall article selection were resolved through mutual consensus by all co-authors. Finally, we manually searched the reference list of each selected article and review articles to identify other relevant articles.

Eligibility Criteria
We followed the PICO framework [45] to formulate the eligibility criteria. The intervention (I) and control (C) eligibilities were the same for both research questions. The criteria of population (P) and outcome (O) were specified separately for each research question. Starting with the intervention (I), articles were included if (1) WBV was applied as a mechanical vibration transmitted through a platform or chair, (2) the design consisted of randomized or non-randomized controlled trials with single and/or multiple WBV treatment sessions and (3) acute, short-and/or long-term effects were measured. Related to the control (C), the comparison condition or group consisted of no treatment, pseudo WBV, sham WBV (exposure of WBV to a negligible low frequency, or noise only) or another form of exercise. Furthermore, we included articles with a publication date till March 2021. Articles were excluded if (1) any type of vibration therapy other than WBV (e.g., vibration pads, ultrasound, electrical stimulation) was examined and, (2) articles were not written in English or (3) articles were review articles.
Separate criteria were designed for each research question regarding the population (P) and outcome (O) measures. To examine the effects of WBV on humans and animal models with PD (research question 1), each selected article had to satisfy the following criteria: (1) the study population consisted of patients diagnosed with PD or PD animal models, (2) no limitations were set on disease severity, age or medication and (3) no restrictions were set on outcome measures.
To examine the potential ameliorating effects of WBV on neuropathological mechanisms (research question 2), articles were included if (1) humans or animal models were examined, and (2) outcome measures were related to neuropathological mechanisms of PD as described in the key search terms.

Data Extraction
One author (Y.L.A.A.) extracted data from each selected article. For each research question, the characteristics and results of the study were presented separately. Characteristics of the studies included the target population, number of participants, intervention and control groups, study design, intervention duration and WBV specifications. In accordance with the guidelines of van Heuvelen et al. (2021), the WBV specifications included type of vibration (synchronous vertical or side-alternating), frequency (Hz), amplitude (mm), temporal aspects (number of sessions per week, number of bouts per session, bout duration) and posture(s).
The results were categorized into the specific outcome measures and into animal and human results. The results included the examined areas with specific outcome measures, pre-test and/or post-test results, within-group effects, between-group effects and main findings. If the authors only presented graphed data, we extracted the data points with a validated web-based tool (WebPlotDigitizer) [46]. To create a general consensus regarding the terminology, we referred to "acute effects" when outcome measures were measured following one WBV session, whereas "chronic effects" referred to outcomes measured over a longer period of time following WBV intervention.

Quality Assessment
One author (Y.L.A.A.) assessed the methodological quality of the studies included. The standardized and validated PEDro scale was used to evaluate the external and internal validities of the controlled, clinical, human studies [47]. Trials were rated with a score ranging from 0-11 with higher scores representing a superior methodological quality. A Scores ranging from 0 to 4 indicated "poor", 5-6 "fair", 7-9 "good" and 10-11 "excellent" qualities.
The ARRIVE essential 10 guidelines were used to determine the methodological quality of animal studies [48]. The aforementioned guidelines ensured the rigorous design and reporting of animal research with a score ranging from 0-18. Scores ranging from 0-7 indicated "poor", 8-12 "fair", 13-15 "good" and 16-18 "excellent" qualities.

Data Synthesis for Meta-Analyses
We used Review Manager Software (Revman) 5.4.1 to conduct a meta-analysis [49] for each research question separately. We used the inverse variance method with a randomeffect model. Hedges' adjusted g effect sizes with 95% confidence intervals (CIs) were calculated for each continuous variable (benchmarks 0.20, 0.50 and 0.80 for, respectively, small, medium and large effects). In case of a two-group pre-post design study, we included the post-pre or pre-post (for inversed scales) difference scores with the pooled standard deviations (square root of the mean of pre-and post-variances) into the analyses. We performed subgroup analyses for each outcome domain. If a study included more than one outcome measure, we incorporated each outcome measure separately into the analyses. I 2 was interpreted as the main heterogeneity measure with the following benchmarks: 0-40, not important; 30-60, moderate; 50-90, substantial and 75-100, considerable heterogeneity [50].

Study Selection
The flowchart of the article selection process is presented in Figure 2. The initial database searches produced a total of 1904 citations. Two additional articles were identified through other sources. Among 1904 articles, 106 articles were identified as potentially relevant and were subjected to full-text screening. A total of 32 articles was sufficient for meeting the eligibility criteria for research questions 1 (14 studies) and 2 (18 studies). The data for 29 studies were synthesized in the meta-analyses. The interrater reliability revealed a 91% agreement between the two authors (Y.L.A.A. and G.A.).

Effects of WBV on PD Populations
The effects of WBV on PD populations were investigated in 14 studies. Thirteen studies included human participants diagnosed with PD [51][52][53][54][55][56][57][58][59][60][61][62][63] and one study in a PD mouse model following brain decapitation [64]. Table 1 presents the characteristics of the studies. Twelve studies had an RCT design and two studies used a non-RCT design, including a non-randomized clinical trial and a cross-over study. In total, 476 human participants diagnosed with PD (mean age: 66.1 years; M/F: 299/149, no gender was reported for 28 participants) and 26 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) mice, a commonly used animal model for PD, were included. The average severity of the disease in human participants ranged from mild to moderate bilateral dysfunctions without serious balance disability (range mean Hoehn and Yahr stage: 2.0-3.3) [65]. The control intervention, intervention duration and WBV parameters varied across all studies. The control interventions comprised sham vibration [53,54], placebo [51,59,61,63], rest [56,57], listening to music [55], moderate walking [52] and conventional (physical) therapies [58,60,62]. The intervention duration varied from one day to four weeks. Nine studies used vertical and four studies side-alternating (one study not reported) vibrations. The number of WBV sessions ranged from 1 to 28. WBV frequency ranged from 6 Hz to 30 Hz and amplitude: 2-14 mm. The number of bouts per session varied between 1 to 15 per session, and the bout duration ranged from 20 s to 15 min.

Effects of WBV on PD Neuropathological Mechanisms
Potentially ameliorating effects of WBV on PD neuropathological mechanisms were investigated in eighteen studies: seven studies on non-PD human populations [66][67][68][69][70][71][72] and eleven on non-PD animal populations [25,[73][74][75][76][77][78][79][80][81][82]. Unlike Zhao et al. (2014; see  Tables 1 and 2), we did not obtain studies on the PD-ameliorating mechanisms of WBV in PD populations. Table 3 presents the characteristics of the included studies. Two cross-over studies and 16 RCTs were performed. A total of 147 human participants (range of average age: 23,4-73 years, M/F: 34/102; no gender was reported for 9 participants) and 321 healthy or diseased modeled mice or rats were included. Intervention duration varied between 1 day and 12 weeks, 12 studies applied vertical vibrations and 3 studies side-alternating vibrations. The vibration frequency varied from 10 to 45 Hz. The amplitude ranged between 2.5 and 4 mm in human studies and between 0.014 and 5 mm in animal studies. The number of sessions varied from 1 single session to 7 sessions/week during 12 weeks, the number of bouts per session ranged from 1 to 10 bouts and the bout duration from to 20 s to 240 min.
Regarding the growth factors, a significant upregulation of BDNF following chronic WBV was observed in humans [66] and animals (BDNF as well as pTrK-B [79]; BDNF [78]) Additionally, higher levels of VEGF were observed following chronic WBV in humans [70]. Chronic WBV increased IGF-1 levels further in CRS rats [78], but decreased in mice models with atherosclerosis [82].
The research on brain-related changes during WBV showed an increased cortical activation in humans [71]. In animals, chronic WBV reduced neurodegeneration [78], increased neurogenesis [75] and increased synaptic plasticity in the hippocampus [74]. No changes in brain glucose uptake following chronic WBV were observed [25]. For oxidative stress markers, acute WBV improved the oxidant and antioxidant parameters in humans [67], and chronic WBV attenuated oxidative stress in mice [76]. Figure 4 presents the results of the meta-analysis. Overall, and in each outcome domain, except for brain-related changes, heterogeneity appeared considerable (81 ≤ I 2 ≤ 91%). For the brain-related changes, heterogeneity was substantial (I 2 = 71%). The accumulated effect size was only significant for inflammatory markers with a considerable reduction in levels following WBV (Hedges' g = −1.12).            "The results of this study reveal that the positive effects of WBV on attention may be (at least in part) mediated by an increased activity of the NBM cholinergic system. WBV could therefore be a suitable intervention strategy in conditions where a reduced cholinergic forebrain activity plays a role".     "WBV induces a significant reduction in inflammatory markers and infarct volume with significant increases in brain-derived neurotrophic factor and improvement in functional activity after tMCAO in middle-aged female rats that were treated with WBV as compared to the non-WBV group". "The 18F-FDG PET data does not reveal any significant difference in brain uptake ratio due to WBV. There was a small but not significant increase in the pseudo WBV group post-treatment".   "WBV attenuates oxidative stress to ameliorate liver steatosis and thus improves insulin resistance in db/db mice. Therefore, WBV administration is a promising treatment for individuals who suffered from central obesity and IR". 1 Hedges' g effect sizes with a negative and positive values indicate a (relative) decrease and increase, respectively, in the mean score of the parameter concerning WBV (vs. control). 2 If the standard error was reported, the standard deviations were calculated with: standard deviation = standard error* √ n. 3 In the WBV + squatting group, no significant differences in pre-post-tests were observed for all parameters. 4 Significant differences were observed between the control and CRS groups for all parameters. Abbreviations: BDNF = brainderived neurotrophic factor; FM = fibromyalgia; sTNFR1 = soluble tumur necrosis factor receptor 1; sTNFR2 = soluble tumor necrosis factor receptor 2; TLR2 = Toll-like receptor 2; TLR4 = Toll-like receptor 4 TNFa = tumor necrose factor 1; pTrK-B = tropomyosin-related kinase B receptor; IL-6/GAPDH = interleukin 6/glyceraldehyde 3-phosphate dehydrogenase; IGF-1 = insulin-like growth factor-1; OxyHb = oxygen hemoglobin; F-FDG = fludeoxyglucose; Neu/BrdU = neuron/bromodeoxyuridine; F-Jade C = fluoro-jade stain C; TBARS = thiobarbituric acid reactive substances; FRAP = ferric reducing ability of plasma; SOD = superoxide dismutase antioxidant enzyme activity; CAT = catalase; GSH = glutathione; CRS = chronic restraint stress model; ChAT = choline acetyltransferase. * p < 0.05 significance versus pre-test data. ** p < 0.01 significance versus pre-test data.

Discussion
This systematic review aimed to determine if WBV training exhibited potential as treatment for PD patients. In total, 32 human and animal studies were included. Fourteen studies investigated the effects of WBV on PD populations and 18 studies investigated the potential ameliorating effects of WBV on neuropathological mechanisms of PD in non-PD populations. Prior systematic reviews only focused on human PD studies with outcomes restricted to the domains of motor impairments, balance, gait and/or mobility [32][33][34][35][36][37]. New insights into the etiology of PD and mode of actions of WBV required the reconsideration of the potential of WBV as a valuable intervention for PD. With the inclusion of animal studies, non-motor outcomes in PD and PD-relevant neuropathological mechanisms in non-PD populations, we extended the search considerably. Moreover, our broader view will facilitate the identification of the relevant, but unexplored, outcome domains.

WBV in Human PD Populations
At present, WBV studies on human PD populations (13 studies) do not provide a conclusion as to whether WBV is effective. Although the vast majority of the studies revealed pre-versus post-improvements following WBV, only four studies provided statistical evidence for the superior beneficial effects of WBV vs. control [52,56,58,59], while four studies did not report on WBV vs. control [51,54,61,62]. The meta-analysis revealed a significant, but small, effect in favor of WBV (Hedges' g = 0.28, p < 0.0001). However, given the considerable variation of control interventions, WBV protocols and settings, and outcome measures, WBV's lack of a convincing superiority over control should be regarded with caution.
The control interventions include active (e.g., walking or physical therapy) and nonactive (sham or placebo) interventions. Active-control interventions may lack contrast with WBV training, limiting the chance of finding differential effects of WBV vs. control. However, a comparison to the conventional treatments may support the applicability of WBV. The fact that WBV and physical therapy have comparable beneficial effects on balance, gait and motor PD symptoms [60] can be considered as promising, since WBV is easier to apply and requires less effort [62]. In line with this, WBV is well tolerated [83] and feasible [84] to apply to patients with disorders.
The majority of the studies included motor outcome measures, including balance, stability, mobility, gait (with a variety of tests) or specific motor PD symptoms. The subgroup analysis of the meta-analysis showed significant overall effects for balance (Hedges' g = 0.30, p < 0.05) and stability (Hedges' g = 0.39, p < 0.01), but not for mobility (Hedges' g = 0.32, p > 0.05, gait (Hedges' g = 0.19, p > 0.05) or specific motor PD symptoms (Hedges' g = 0.18, p > 0.05). These values obtained for Hedges' g are comparable to those observed in another recent review of the meta-analysis [32]. On the level of individual studies, the significant beneficial effects of WBV vs. control were observed for balance/stability [52,58,59], mobility [59] and motor PD symptoms [56]. In addition, Kaut et al. (2011) observed improvements in motor PD symptoms (including rigidity, postural stability and bradykinesia) following WBV and not after sham vibration with a considerable effect (Hedges' g = 0.76), but without reporting statistical testing. Only one study included non-motor outcomes related to depression, anxiety and fatigue, and observed similar pre-post improvements for WBV and listening to music (Hedges' g =, respectively, -0.21, -0.10 and -0.07) [55]. Based on the current studies, we cannot conclude that heterogeneity in outcome measures plays a major role, but non-motor outcomes, including cognitive function, are hardly represented.
WBV protocols and settings varied considerably across the studies. In four studies, the short-term effects were examined following a single session of WBV with mixed results. The other studies focused on the chronic effects. Overall, longer intervention durations (3-5 weeks) with sufficient session frequencies (3-4 sessions per week) might be more promising, although one study with a duration of 5 weeks and 2-3 sessions per week did not observe any effects of WBV compared to the placebo [63]. The latter result might be related to the relatively low vibration frequency (6Hz), although a study with comparable duration, session frequency and vibration frequency [59] revealed significant positive effects of WBV vs. placebo. Only one study applied several vibration frequencies and concluded that higher frequencies seem to present a greater improvement [61]. Indeed, Guadarrama-Molina and co-authors (2020) applied a frequency of 20 Hz during 5 weeks and observed positive effects of WBV vs. conventional therapy, and Ebersbach et al. (2008) applied a frequency of 25 Hz during 3 weeks and observed similar effects of WBV and physical therapy. The stronger effects of higher vibration frequencies are in agreement with results findings of a review of COPD patients [85] and may be explained by enhanced neuromuscular activity [86].

WBV Related to the Neuropathological Mechanisms of PD
None of the studies on human PD patients investigated the neuropathological PD mechanisms, and solely one animal study on PD mice investigated the neuropathological mechanisms. However, the neuropathological mechanisms of PD in non-PD animal (11 studies) or human (7 studies) populations may provide further insight into the potential of WBV for this patient group. Animal studies are valuable [87] because biological processes can be examined in detail, whereas examining mechanisms in humans is expensive, stressful for the patient or even impossible/unethical. In addition, in animals, processes can be studied in less time because of the shorter lifespan of frequently used species (e.g., mice and rats), and within strains, genetic variations and variations in lifestyle regimes are considerably reduced.
Only one study included a PD population, in which MPTP mice were exposed to WBV during four weeks on a daily basis [64]. The vibration training almost completely prevented the MPTP-induced loss of dopamine neurons in the substantia nigra and upregulated BDNF. This is an important result, since it suggests that WBV could act upon the primary cause of PD. It may also explain the improvement to specific motor symptoms (UPDRS-III score) following a WBV intervention [53][54][55]60,63].
All other WBV studies related to neuropathological PD mechanisms were conducted on non-PD populations. Overall, the meta-analysis presented no effect of WBV (Hedges' g = 0.04, p > 0.05), but this observation should be interpreted in light of considerable heterogeneity (I 2 = 86%). Therefore, it is more meaningful to consider the subgroup results or even the individual study results.
For the neurotransmitter domain, the meta-analysis resulted in an overall moderate to large effect with higher neurotransmitter levels following WBV, but the effect was nonsignificant (Hedges' g = 0.62, p = 0.09). However, considerable heterogeneity also played a major role (I 2 = 81%) here, which caused us to consider specific neurotransmitters and individual studies. Two non-PD studies investigated the effects on dopaminergic systems. In these studies, Wistar rats were exposed to single bouts of WBV. Following 90 min of WBV, the dopamine levels increased in the nucleus accumbens (part of ventral striatum) and frontal cortex (20 Hz; [77]), but dopamine levels in the cortex and striatum appeared unaffected after 240 min (20 Hz; [81]. Although the first study may support a beneficial effect on the dopaminergic system, the value for exercise studies in which the WBV bouts are much shorter is limited. These long bout-duration studies mainly focus on safety issues regarding "bad vibrations" associated with detrimental effects [15,88]. Other neurotransmitter systems are affected by WBV as well, with enhanced levels of epinephrine and norepinephrine in human blood plasma levels following a single session of WBV (26 Hz and 10 × 60 s; [72]). Following lengthy exposure to WBV (240 min), norepinephrine levels decreased in the specific parts of the brain of Wistar rats [73,81], while serotonin levels also increased [81]. Heesterbeek et al. (2017) suggested that a five-week WBV intervention (30 Hz) activated the cholinergic system in C57BI/6J mice. Altogether, WBV has the potential to affect neurotransmitter systems. This may be promising for PD patients, since not only is the dopaminergic system impaired in PD, but also the adrenergic, serotonergic and cholinergic systems [2]. However, this should be confirmed in the studies including PD patients.
Chronic neuroinflammation is one of the hallmarks of PD pathophysiology and therefore an important therapeutic target [89]. Overall, WBV resulted in reduced levels of inflammatory markers (Hedges' g = -1.12, p < 0.01), but with considerable heterogeneity (I 2 = 91%). Three human and two animal studies examined the effects of WBV on pro-inflammatory markers (e.g., cytokines, such as TNFa and interleukin 1β, Toll-like receptors TLR2 and TLR4, tumor necrosis factor TNFa or receptor sTNFR1), anti-inflammatory markers (e.g., adiponectin, sTNFR2, interleukin 10) or markers with both pro-and antiinflammatory properties (interleukin 6). The studies on the chronic effects of a WBV intervention consistently revealed significant decreases in several pro- [68,79] and antiinflammatory markers [79,82]. Acute effects, however, showed enhanced levels of some pro-inflammatory markers [69,70] with a decrease [69] or increase [70] in anti-inflammatory markers. These results suggest that WBV may evoke an acute inflammatory response. An acute inflammatory response is an essential and protective response in, for example, injured tissues. This response can restore the tissues to their pre-injury state [90]. The chronic inflammatory response may indicate that the inflammation balance is altered, with an overall lower level of chronic inflammation.
Indeed, a recent paper confirmed that WBV attenuated traumatic brain injury-related damage through the regulation of neuroinflammation [91]. Whether this is also true for PD should be examined in either research on animals or humans.
Changes in the growth factors (BDNF, IGF-1) are considered to be an underlying mechanism of the cognitive effect of exercise [92]. The meta-analysis showed an overall moderate to large, but non-significant, increase in growth factor level after WBV (Hedges' g = 0.67, p > 0.05) with considerable heterogeneity (I 2 = 86%). Chronic increases in plasma BDNF [66] and brain BDNF [79] concentrations were observed after a WBV intervention [66,79], but not for a single session [69]. For IGF-1, the results are mixed with chronically increased [78] or decreased [82] levels. The difference may be explained by heterogeneity in frequency (respectively, 30 Hz and 15 Hz) or animals (respectively, mice and rats) used. The potential of WBV to affect the brain is supported further by some animal studies, showing increased hippocampal synaptic plasticity [74], increased neurogenesis [75] and decreased neural degeneration [78]. The meta-analyses revealed a small to moderate, but non-significant, effect (Hedges' g = 0.35, p > 0.05) with substantial heterogeneity (I 2 = 71%). Nevertheless, since PD goes together with cognitive decline [2,43] these findings may be promising, but should be confirmed.
Oxidative stress, an imbalance between increased levels of reactive oxygen species (free radicals) and low activity of antioxidant defense, is considered to modulate the development of PD [93]. Only two studies examined WBV in relation to oxidative stress markers resulting in an overall less than small effect (Hedges' g = 0.11, p > 0.05) with considerable heterogeneity (I 2 = 83%). Notwithstanding, the findings from both a single session of WBV in humans (women with fibromyalgia; [67]) and a 12-week WBV intervention in mice [76] support a beneficial effect of WBV on oxidative stress. However, the translational value to PD is unknown, so these findings should be interpreted with caution.

General Considerations
In summary, based on human studies on PD patients, we cannot definitively confirm the effectiveness of WBV for PD patients considering the significant, but low, overall effect size. However, there may be potential for WBV protocols with sufficient duration (at least three weeks), session frequency (at least three sessions per week) and vibration frequency (at least 20 Hz), especially for outcomes related to motor function. Studies on neuropathological mechanisms of PD further confirm the potential of WBV to reduce neuroinflammation and promote specific neurotransmitter systems, increase the expression of specific growth factors, increase neurogenesis and synaptic plasticity and reduce specific markers of oxidative stress.
However, all results should be interpreted with caution. Optimal settings are not definitively defined yet, since high-quality studies with a direct comparison of different settings are still scarce. Furthermore, optimal settings may vary across desired outcomes and individual patients. Whether the effects depend on posture remains uncertain, since the majority of the human studies used a static semi-squat posture. Furthermore, the peakto-peak displacement (or amplitude) does not vary strongly across the studies. However, in most studies, the peak-to-peak displacement of the vibrations remains unclear, since they are not defined and/or not verified with actual measurements. Since the manufacturer's settings and measurement of the vibration often lead to different outcomes, it is important to report adequate information here [13]. With current studies, we cannot yet elaborate on the moderating effect of peak-to-peak displacement. Whether the effects depend on the type of vibrations (vertical vs. side-alternating) is still unknown since there is a paucity of studies with side-alternating WBV in PD populations. Although several outcome measures validated for PD were used (i.e., UPDRS-III, 8 m or 10 m walk test, Tinetti scale and Berg balance scale), other outcome measures were only validated in other populations (e.g., timed up and go test, BDI depression scale, ISQ anxiety scale or FFS fatigue scale) or, as far as we know, not yet validated (e.g., freezing test and step-walk-turn test. Subsequently, the studies regarding the neuropathological mechanisms of PD in PD populations is very limited, with only one study on MPTP mice [64]. The translational value of study results obtained from non-PD to PD populations and from animal to human populations in the context of WBV is still unsure. In addition, several studies with long WBV exposure, investigating "bad vibrations", are less relevant for WBV in the context of exercise or treatment. Finally, several PD relevant outcomes have not yet been investigated, such as cognitive function in PD patients, Lewy bodies or alpha-synuclein aggregation. Compared to the previous systematic reviews [32][33][34][35][36][37], our broader and novel scope resulted in a more detailed identification of WBV settings and protocols delivering further perspectives for PD patients. Furthermore, it strengthened the evidence of the potential of WBV considering the effects observed on neuropathological mechanisms, such as neuroinflammation and specific neurotransmitter systems. Finally, it resulted in the identification of (relatively) unexplored research domains (e.g., neuropathological mechanisms in PD and non-motor outcomes, such as cognition).

Recommendations for Further Research
For further research, we recommend high-quality human studies on PD patients with an intervention duration of at least three weeks with at least three sessions per week, a vibration frequency of at least 20 Hz and with sufficient contrast between WBV and control group(s) (e.g., WBV, sham WBV and normal care). Furthermore, we recommend RCTs with different levels of frequency and/or peak-to-peak displacement and with sidealternating vibrations in order to make progress in identifying the optimal settings. We recommend the use of outcome measures validated for PD and adding outcomes related to non-motor functions, such as cognitive and affective functions, next to general motor function and PD-specific motor symptoms. We recommend animal studies using PDmouse models to investigate the PD-relevant mechanisms in more detail and to investigate different durations and settings relevant for exercise and treatment with WBV, in order to shape human research to identify the optimal settings. Finally, we recommend using the guidelines of van Heuvelen et al. (2021) [13] to promote correct, complete and consistent WBV reporting.

Recommendations for Sports and Rehabilitation Practices
Based on the current knowledge, we cannot recommend WBV as a primary treatment for PD patients in the exercise domain. If moderate-to-high intensity conventional exercise, such as resistance and aerobic training, is possible, this kind of exercise still has priority. However, for those who are not able to perform active exercises, WBV may be a valuable alternative. Furthermore, WBV can be performed in addition to conventional exercises to enhance the beneficial effects of exercise for PD patients. Sufficient intervention duration, session frequency and vibration frequency are important as well as guidance by a qualified trainer for correct performance and to promote efficacy and adherence [94].

Limitations
This systematic review had some limitations. First, we limited our review to animal and human studies, excluding cell culture studies. The inclusion of cell culture studies may provide additional insights into the potential and underlying molecular mechanisms of WBV for PD patients. More specifically, the use of human-induced pluripotent stem cells (hiPSCs) isolated from PD patients might be of great benefit. Second, with regard to non-PD populations, we restricted our search to direct PD neuropathological mechanisms. We considered indirect mechanisms (e.g., mechanisms on muscle level indirectly affecting the brain) as beyond the scope of this review. Finally, although we performed a meta-analysis on the effects on neuropathological PD mechanisms in non-PD populations, the generally considerable heterogeneity required us to interpret the overall results with caution.

Conclusions
Our review revealed the significant but minor beneficial overall effect of WBV on motor and non-motor outcomes in PD patients. Therefore, the potential of WBV for PD patients seemed limited, although specific protocols performed with sufficient frequency and duration may show potential and legitimizes further research. The overall reduced levels of neuroinflammation and specific WBV effects on other PD-relevant neuropathological mechanisms observed in non-PD populations strengthen the urgency for further research into the effects of WBV on PD patients. The recommendations for practice and further research are summarized in Table 5.

results
after duplication removal with PubMed (583)

results
after duplication removal with PubMed (658)

results
after duplication removal with PubMed (759) Note. Criteria PEDro scale: 1 = eligibility criteria were specified; 2 = participants were randomly allocated to groups; 3 = allocation was concealed; 4 = groups were similar at baseline; 5 = blinding of all participants; 6 = blinding of all therapists; 7 = blinding of all assessors; 8 = measures obtained from more than 85% of the participants; 9 = all participants received treatment/control condition or "intention to treat" analysis; 10 = comparison of between-group results reported; 11 = point measures and measures of variability are provided. a Quality assessment not applicable for design study with PEDro scale. ARRIVE guidelines: 1a = Are all experimental and control groups clearly identified?; 1b = is the experimental unit (e.g., an animal, litter or cage of animals) clearly identified?; 2a = Is the exact number of experimental units in each group at the start of the study provided (e.g., in the format 'n=')?; 2b = Is the method by which the sample size was chosen explained? 3a = Are the criteria used for including and excluding animals, experimental units or data points provided; 3b = Are any exclusions of animals, experimental units or data points reported, or is there a statement indicating that there were no exclusions?; 4 = Is the method by which experimental units were allocated to control and treatment groups described?; 5 = Is it clear whether researchers were aware of, or blinded to, the group allocation at any stage of the experiment or data analysis?; 6 = For all experimental outcomes presented, are details provided of exactly what parameter was measured?; 7a = Is the statistical approach used to analyze each outcome detailed?; 7b = Is there a description of any methods used to assess whether data met statistical assumptions?; 8a = Are all species of animal used specified?; 8b = Is the sex of the animals specified?; 8c = Is at least one of age, weight or developmental stage of the animals specified?; 9a = Are both the timing and frequency with which procedures occurred specified?; 9b = Are details of acclimatization periods to experimental locations provided?; 10a = Are descriptive statistics for each experimental group provided, with a measure of variability (e.g., mean and SD, or median and range)?; 10b = Is the effect size and confidence interval provided?