IMPROVE study protocol, investigating post-stroke local muscle vibrations to promote cerebral plasticity and functional recovery: a single-blind randomised controlled trial

Abstract

Introduction

Spasticity is a frequent disabling consequence following a stroke. Local muscle vibrations (LMVs) have been proposed as a treatment to address this problem. However, little is known about their clinical and neurophysiological impacts when used repeatedly during the subacute phase post-stroke. This project aims to evaluate the effects of a 6-week LMV protocol on the paretic limb on spasticity development in a post-stroke subacute population.

Methods and analysis

This is an interventional, controlled, randomised, single-blind (patient) trial. 100 participants over 18 years old will be recruited, within 6 weeks following a first stroke with hemiparesis or hemiplegia. All participants will receive a conventional rehabilitation programme, plus 18 sessions of LMV (ie, continuously for 30 min) on relaxed wrist and elbow flexors: either (1) at 80 Hz for the interventional group or (2) at 40 Hz plus a foam band between the skin and the device for the control group.

Participants will be evaluated at baseline, at 3 weeks and 6 weeks, and at 6 months after the end of the intervention. Spasticity will be measured by the modified Ashworth scale and with an isokinetic dynamometer. Sensorimotor function will be assessed with the Fugl-Meyer assessment of the upper extremity. Corticospinal and spinal excitabilities will be measured each time.

Ethics and dissemination

This study was recorded in a clinical trial and obtained approval from the institutional review board (Comité de protection des personnes Ile de France IV, 2021-A03219-32). All participants will be required to provide informed consent. The results of this trial will be published in peer-reviewed journals to disseminate information to clinicians and impact their practice for an improved patient’s care.

Trial registration number

Clinical Trial: NCT05315726

Dataset

EUDRAct

Strengths and limitations of this study.

• This study is a bicentric controlled randomised, single-blind (participant) trial in an early subacute stroke population.

• The materials and methods used in this protocol allow quantitative evaluations of spasticity (ie, isokinetic dynamometer), spinal (ie, H-reflex), corticospinal (ie, transcranial magnetic stimulation) and cortical (ie, electroencephalogram) excitabilities and activities.

• The design of this protocol allows a short, mid-term and long-term follow-up of the participants, at different times to assess the recovery.

• The key methodological limitation of the study is that the investigators are not blinded. Standardised automatised quantitative procedures and treatments would improve the accuracy of the measures and would mitigate this issue.

Introduction

Stroke is one of the most common neurological disorders worldwide, affecting about 12 million people every year.¹ Following a stroke, a majority suffer from various symptoms of the upper motor neuron syndrome, such as motor weakness and spasticity.² Despite a partial recovery, patients often live with functional disabilities, impacting their daily activities (ie, washing, eating and walking) and quality of life, also affecting their relatives.

About 40% of patients with paresis will develop spasticity during the first year following a stroke.³ However, this prevalence is variable within the stroke population studied in the literature, questioning the quality of the measure and the definition of spasticity. Among many existing descriptions, reflecting its multifactorial aspects and its complex physiopathology, a consortium described spasticity as a sensorimotor control disorder ‘resulting from an upper motor neuron lesion, presenting as intermittent or sustained involuntary activation of muscles’.⁴ Spasticity is likely a result of an imbalance of excitatory and inhibitory signals at the spinal level from cortical tracts (mainly reticulospinal and corticospinal), which play an important role in the regulation of the spinal reflex activity.^{5 6} Alongside, the sprouting of new axonal fibres⁷ and local structural and morphological modifications of the disuse muscles² have been observed in patients who have spastic stroke.

The development of spasticity is progressive. It appears from the stroke onset after a period of flaccidity and hypotony,⁵ and its severity tends to increase during the first year, leading to more discomfort for patients.^{8 9} Spasticity is more frequent in the muscles of the contralateral upper limb from the lesion than in the contralateral lower limb.¹⁰ Patients presenting important paresis, hypoesthesia and low functional scores are more likely to develop spasticity after 1 month of stroke onset.^{8 11} Pharmacological therapy during this period has limited effectiveness and remanence to overcome the development of spasticity, with sometimes significant side effects.¹²

It has been proposed that early and intense rehabilitation treatments to enhance cerebral plasticity could decrease sensorimotor troubles, as the first months following stroke are the best therapeutic window to promote neuroplasticity.¹³ Several non-pharmacological treatments are also available for rehabilitation, including electrotherapy, shock waves, ultrasound, thermotherapy, neuromodulation, vibrations, passive mobilisations or stretching.¹⁴ However, limited scientific evidence exists to confirm their efficiency during this therapeutic period.

Mechanical oscillatory vibrations, applied on a tendon or the muscle belly, provoke strong proprioceptive stimuli, particularly through the activation of multiple cutaneous receptors and muscle spindles.¹⁵ Set at a frequency of 80 Hz, vibrations are harmonic with the discharge frequency of the afferent fibres Ia.¹⁵ On a relaxed and hidden limb, they create a movement illusion, as if the vibrated muscle is stretched,¹⁶ and could enhance the cerebral excitability of the sensorimotor cortex during and after a vibration session.^{17 18} Recent studies showed that local muscle vibrations (LMVs) decrease, momentarily, the H-reflex of the vibrated muscle in healthy volunteers.^{19 20} As this reflex is an indicator of the excitatory state at the spinal level,²¹ such reduction highlights a neuromuscular plasticity mechanism, with a probable interest in treating spasticity.^{22 23}

LMVs of the upper and lower limbs have been tested in patients who have chronic stroke, showing a momentary decreased spasticity and increased cortical excitability.^{24 25} Yet, it is known that cortical plasticity is more important in the first 3–6 months following a stroke.²⁶ Only one preliminary study investigated their impact on the acute phase (72 hours post-stroke, three sessions of 10 min of vibrations for 3 days), showing a significant functional improvement of the patients in the vibration treatment group.²⁷ To date, the neurophysiological quantification of the effects of LMV in acute or subacute stages post-stroke is scarce, as described in two recent literature reviews, and the variability within the protocols used does not allow the definition of the optimal LMV modalities.^{24 28} This protocol proposes to investigate the effect of a 6-week LMV protocol on the development of spasticity in a subacute post-stroke population.

Study objectives

Primary objective

This project aims to evaluate the effects of a 6-week LMV protocol on the paretic limb on the spasticity development between interventional and placebo (SHAM)groups at 6 weeks.

Secondary objectives

Four secondary objectives are defined:

1. Describe the short-term and long-term (ie, 6 months) effects of a 6-week LMV protocol on the onset and behaviour of spasticity in a subacute stroke population between interventional and SHAM groups.

2. Investigate the cortical and spinal neuroplasticity associated with LMV.

3. Explore the effects of LMV on the patient’s sensorimotor performance.

4. Evaluate the correlations between spasticity and several spinal and cortical markers to highlight potential predictive markers of spasticity.

Trial design

This protocol is an interventional, bicentric, controlled, randomised, single-blinded (patient) trial. Participants will be enrolled for 7.5 months, with five visits along and the intervention, 3 times per week for 6 weeks, as shown in figure 1. The inclusion time will last 36 months, leading to a total time for this trial of 43.5 months.

Figure 1.

Study design and timeline.

All the visits are conducted by an investigator of the study, either a physical therapist or medical doctor, and a trained clinical research technician.

Participants

We aim to include 100 participants, 50 per group, from the rehabilitation centre of Dijon University Hospital, France, and of Marguerite Boucicaut in Chalon sur Saône, France, during their inpatient or outpatient hospitalisation stay. Inclusion and non-inclusion criteria are detailed in table 1.

Table 1.

Inclusion and non-inclusion criteria of the participants

Inclusion criteria	Non-inclusion criteria
Adult patient >18 years old	Significant pain on mobilisation of the wrist or elbow (VAS >5/10)
Able to follow a rehabilitation programme on medical opinion (no medical issues or acute intercurrent medical events)	Presence of other neurological, muscular or osteoarticular conditions altering upper-limb function
First stroke ever <6 weeks, confirmed by imaging	Significant cognitive impairments: inability to understand simple instructions or give consent of any kind (not included if LAST scores <5/7 in comprehension and if YES/NO answers are unreliable)
Hemiparesis or hemiplegia of the upper limb (FMA-UE score <48), particularly in the wrist and elbow flexors	Apparent wound, which may postpone inclusion, or very fragile skin
Requiring inpatient or outpatient hospitalisation in a rehabilitation centre	Not covered by national health insurance
Having given oral consent	Being pregnant or breastfeeding
	Being under guardianship or curatorship
	Person subject to a measure of legal protection

FMA-UE, Fugl-Meyer assessment of upper extremity; LAST, LAnguage Screening Test; VAS, visual analogue scale.

Every year, about 45 patients who had stroke come to the rehabilitation centre of Dijon and 45 to the rehabilitation centre of Chalon sur Saone for inpatient rehabilitation care. We estimate that 40% of them will meet the inclusion criteria and will accept to participate in the trial, being 36 participants per year that could be recruited in this trial.

During a screening visit, patients who fulfil inclusion criteria are informed about the protocol and the two groups (ie, SHAM and intervention). If all the criteria are met, and with the consent of the patient, the trial starts with the preparation of visit 1.

To calculate the number of participants necessary for this study, we based on several studies. The meta-analysis of Zeng et al ³ showed a spasticity prevalence of 40% in a hemiparetic subacute stroke population following a first stroke. In a chronic stroke population, Marconi et al ²⁵ found a reduction of spasticity parallel to an increase in the arm function (+11 points, Motricity Index) after 3 days of vibrations. Moreover, in a study of an acute stroke population, Toscano et al ²⁷ found an increased arm function (+18.5 points, Motricity Index) after 3 days of vibrations. As we know that the neuroplasticity window is more important during the first months post-stroke, we expect that the effects on spasticity would be proportional to those found on arm function in the previous studies (18.5/11=1.68); therefore, –1.68 points on the modified Ashworth scale (MAS), in comparison with –1 point.²⁵

Thus, we hypothesised that 15% of our patients in the intervention group would develop spasticity, in comparison with 40% in the control group. Moreover, this effect could be strengthened as our protocol lasts 6 weeks, much longer than the protocols used in the previously cited studies. Based on this hypothesis, with a 0.05 alpha risk and beta risk of 20%, 46 volunteers are necessary per group. With a risk of having a 10% dropout, 100 participants (50 per group) will be included in a 1:1 randomised schema.

Exclusion and randomisation criteria

Participants will be randomly assigned to one group, either the placebo group or intervention group, at the end of visit 1, using a computer-generated stratified randomisation on age and sex. The blinding will be removed at the end of the last visit of the study (visit 4).

Botulinum toxin injections in the upper-limb muscles are not allowed during the protocol except for a contrary medical opinion. In that case, the participant stops the protocol and will be assigned to a failure of the protocol.

Intervention

Both groups will receive LMV from the same device: VIBRAMOOV Physio (TECHNOCONCEPT, Manosque, France, CE 93/42, ISO 13485: 2016), amplitude 0.2–0.5 mm.

The interventional group will receive the conventional rehabilitation programme therapy of the rehabilitation centre (about 4 hours of physiotherapy and occupational therapy, five times per week), plus LMV at 80 Hz on the elbow and wrist flexors of the paretic arm continuously for 30 min, 3 times per week. The control group will receive the conventional rehabilitation programme therapy of the rehabilitation centre, plus SHAM LMV at 40 Hz (ie, without neurophysiological impact), plus a foam band between the skin and the device, on the elbow and wrist flexors continuously for 30 min, 3 times per week. All participants must attend all 18 vibration sessions during the 6 weeks.

Outcomes

All outcome measures are described in table 2.

Table 2.

Outcome measures throughout the study

	Inclusion visit	Visit 1	Visit 2	Visit 3	Visit 4	Intervention
			Visit 1+3 weeks	Visit 1+6 weeks	Visit 3+6 months	For 6 weeks
LAST	X
Medical evaluation	X				X
Spasticity: MAS entire upper limb	X			X	X
Spasticity: MAS elbow and wrist flexors	X	X	X	X	X
Spasticity: dynamometer		X	X	X	X
FMA-UE	X	X		X	X
Transcranial magnetic stimulation		X	X	X	X
H-reflex/M-wave		X	X	X	X
VAS: pain and tiredness		X	X	X	X
Local muscle vibrations (3×30 min/week)						X
Sensations of vibration and illusion						X
EEG (1/week)						X
Quantity of rehabilitation time (minutes)

EEG, electroencephalogram; FMA-UE, Fugl-Meyer assessment of upper extremity; LAST, LAnguage Screening Test; MAS, modified Ashworth scale; VAS, visual analogue scale.

All visits start with the evaluation of spasticity, first using the modified Ashworth scale and second by the isokinetic dynamometer. Then the Fugl-Meyer assessment of upper extremity (FMA-UE) is realised, during visits 1, 3 and 4. The neurophysiological evaluations come after, starting with the measure of M-wave. The order of the other tests (H-reflex or transcranial magnetic stimulation (TMS)) is randomised prior to visit 1 for each participant and kept identical for visits 2–4.

Primary outcome measure

Spasticity will be assessed by the modified Ashworth scale, the current clinical gold standard and frequently used tool.²⁹ This scale categorises spasticity within six scores, considering the angle of catch and the resistance intensity felt by the clinician when passively and quickly stretching the joint. The scores are described as follows: 0, no increased resistance; 1, slightly increased resistance (catch followed by relaxation or minimal resistance at the end of the range of motion); 1+, slightly increased resistance (catch followed by minimal resistance throughout less than half of the range of motion); 2, clear resistance throughout most of the range of motion; 3, strong resistance (passive movement is difficult); and 4, rigid flexion or extension.³⁰ During the inclusion visit and visits 3 and 4, spasticity of the following muscle groups at the shoulder (ie, flexors, extensors, abductors, adductors and internal and external rotators), elbow (ie, flexors, extensors, pronators and supinator), wrist (ie, flexors and extensors) and fingers (ie, flexors, extensors, adductors and abductors) is evaluated. During visits 1 and 2, spasticity only at the wrist and elbow flexors is assessed with this scale. For one participant, the test is performed by the same experimented therapist each time.

Secondary outcome measure

The sensorimotor function of the upper limb will be measured by the FMA-UE,³¹ a specific test evaluating the sensorimotor performance of the upper limbs in a stroke population. Hoonhorst and collegue³² suggested the following scores grading: 0–22 for no upper-limb motor capacity, 23–31 for poor capacity, 32–47 for limited capacity, 48–52 for notable capacity and 53–66 for full upper-limb capacity.

Pain and tiredness will be evaluated by two different visual analogue scales (0–100 with 0 for non-pain and 100 for the worst imaginable pain and 0 for non-tiredness and 100 for the worst imaginable tiredness) during all tests and visits carried out in this protocol.

Spasticity will be also assessed by an isokinetic dynamometer (Biodex, Shirly NY) that moves the articulation of the wrist or the elbow in a standardised position within physiological joint limits. The flexor muscle group is stretched at 180°/s.³³ The dynamometer records the angle of the catch and end as well as the torque of movement-related response.

Corticospinal tract (CST) excitability will be assessed by TMS. Linked to a stimulator (Magstim, Roseville, MN), the coil apposed on the top of the skull skin sends a simple stimulation of 1 ms and 1 Tesla, or a double stimulation, depolarising the underlying neurons. The surface electrodes apposed on the skin of the muscle (in this study: on the flexor carpi radialis (FCR)) will assess the electrical activity called motor evoked potentials. By using the single stimulation, the excitability state of the CST of the participants is evaluated. The integrity of the CST in the ipsilesional hemisphere has been proposed as a predictive biomarker of upper-limb recovery following stroke.³⁴ By using double stimulation, the short interval cortical inhibition (SICI) is evaluated. In the literature, an increased SICI in the ipsilesional hemisphere has been correlated to reduced spasticity in the spastic muscle^{22 35}

Spinal excitability will be measured using peripheral neural percutaneous stimulation, allowing depolarisation of the underlying axonal structures with a constant current stimulator (Digitimer, UK). In this study, we will stimulate the median nerve and record the muscle activity of the FCR with surface electrodes (a short motor response M and a later one from the reflex H). The ratio Hmax/Mmax, H80/Mmax, with H80 corresponding to 80% of Hmax during the ascending part of the recruitment curve, and the H-reflex post-activation depression will be evaluated.^{23 36} To ensure an equivalent stimulus condition (ie, same pool of motor unit tested) between each visit, we will use a stable M-wave prior to Hmax and H80, determined during visit 1.³⁶

An electroencephalogram will be performed using a connected helmet 10–20 system (Neuronaute, Bioserenity, Paris) to record the cerebral electrical activity in real time, once a week during an LMV session for all enrolled participants.

During all LMV sessions and every 5 min, participants will evaluate their sensations of movement illusions and vibrations by orally answering two questions.

Data collection and management

All the data are collected anonymously by the investigators of the study. Health data will be deidentified, and the corresponding variables will be kept secret in a coded file. This document was created before the database construction. The study data will be entered into a paper and an electronic case report form during the study, with only the patient’s code appearing. Only the inclusion ranking will be kept when extracting the data. Only the investigators will have access to the dataset.

Data analysis

The main criteria will be realised in the intention to treat analysis, using the hypothesis of maximal bias. The percentage of patients developing spasticity within 6 weeks will be described and compared between groups, with a χ² test or a Fisher’s exact test.

The other complementary quantitative analyses will be done using the Student t-test to compare the two groups or the Mann-Whitney test. For intragroup (time effect) comparisons, we use either a repeated measure analysis of variance (ANOVA) (parametric distribution) or a Friedman ANOVA (in case of non-parametric distribution). The qualitative variables will be analysed by a χ² test or a Fisher exact test for the standard test conditions of application.

For research of predictive markers of spasticity, the evolution of different neurophysiological parameters will be linked with the measure of spasticity over several time points in both groups, using a multivariable logistic model.

The data analysis will be done with the R statistical package. For all tests, the degree of statistical significance will be estimated at alpha=0.05. A Bonferroni correction will be used when necessary.

Adverse events

Tolerance to the intervention was good in other studies using similar vibration devices. No adverse events have been reported. However, information about events are collected throughout the study in the patient file. The investigator could temporarily or permanently withdraw the participant from the study, or the patient may withdraw on their own at any time of the study.

Patient and public involvement

There is no patient or public involvement to declare in the design, or conduct, or reporting, or dissemination plans of this trial.

Ethics and dissemination

Monitoring and auditing

The CHU Dijon is the legal sponsor of this study. Several monitoring visits with dedicated personnel are planned in Dijon and Chalon centres. A report on data collection will be done, and the investigator must confirm all the issues reported.

Ethics and dissemination

This study was recorded on the European clinical register database (EUDRAct) with the number 2021-A03219-32 and in a clinical trial under the number NCT05315726. The research ethics committee (CPP Ile de France IV) provided agreement for this study, according to the Helsinki Declaration for biomedical research. All participants will provide an intentional verbal informed consent before their inclusion in this protocol. According to the French regulation authorities, this study is a category 2, at minimal risks and constraints, and requires a free, enlightened and express (ie, verbal or written) consent of all participants.

Any relevant protocol or information note revisions to the actual version (V.4, July 2023) will be communicated to the ethics committee, trial registries and any other relevant authorities.

We will ensure that the results of this trial will be published in peer-reviewed literature and presented at several scientific conferences to deliver and disseminate information to clinicians and impact their practice for improved patient’s care. We will share and collaborate with other research teams on reasonable request for access to study data, as described in online supplemental file.

Discussion

This randomised controlled trial aims to test the effects of a 6-week LMV protocol on upper-limb spasticity in a subacute stroke population.

Despite the growing use of LMV in clinical settings, including in neurological rehabilitation programmes, there is a lack of understanding of their effects both when used chronically and on their neurophysiological influence. Several studies have already demonstrated the clear temporary benefit of one LMV session on spasticity, both on contracted and relaxed muscles.^{25 37 38} This can be partly explained by the post-vibration H-reflex reduction of the vibrated muscle,¹⁸ as post-stroke spasticity is associated with an increased H-reflex.^{23 39 40} Functional benefits, measured by clinical scales (ie, Fugl-Meyer, Wolf motor function test), often accompanied this spasticity reduction in these previous studies.^{25 27 38} In this trial, we expect to measure the progressive effects of LMV during several weeks and their long-term impact on spasticity and function of the paretic upper limb. Our first hypothesis is that the interventional group would have a lower prevalence of spasticity on the elbow and wrist flexors than the control group and that this benefit would remain in time. Second, we expect that LMV at 80 Hz would enhance neuromuscular plasticity in this population, potentially improving upper-limb function. Alongside, we will evaluate for a pathophysiological approach the evolution of the excitability at the cortical and spinal level to better understand the neurophysiological impact of vibrations and to better adapt LMV modalities for further protocols.

Because paresis has been identified as a predictive marker for developing post-stroke spasticity,^{8 11} the population in this trial is selected with upper-limb paresis and poor function. Therefore, most of them could suffer from hypotony and are not able to contract their wrist and elbow flexors; we choose to perform vibrations on relaxed muscles, both for feasibility with this population and because it has been shown that LMV induces illusion only when the muscle is relaxed and hidden from the visual field.¹⁶ Besides, the cortical activity induced by LMV has been shown to enhance when illusions occur in comparison with no illusion conditions, as previously shown by functional MRI^{41 42} or electroencephalography.⁴³ This cortical activity, mainly in the sensorimotor cortex, shares similar structures as when a real movement is performed.⁴² We would therefore expect that the interventional group would experience illusions and would have more cortical activity during and after vibrations sessions, hypothetically equally improving corticospinal plasticity.

We recognise several limitations in this study. One is that the experimenters are not blinded. We have tried to bound this problem with very standardised procedures during experiments and vibration sessions and by randomising participants only after visit 1. Also, despite two clinical measures (ie, MAS and FMA-UE) subject to the evaluator subjectivity, the other methods (ie, isokinetic dynamometer, EEG and neurophysiological evaluations) are quantitative and treated with automatised mathematical techniques. We also acknowledge that the quantity of rehabilitation may vary between participants; this is why we will track this time in minutes during all the study for all participants. Finally, the inclusion and exclusion criteria do not discriminate the stroke location and severity, which may lead to some heterogeneity in the studied population. Participants are selected with similar upper-limb functional impairments, but the evaluation of CST integrity (ie, TMS) would allow us to mitigate this result and use it as a predictive biomarker in our participants. Indeed, it has been shown that patients with a non-functional CST have poorer predictive upper-limb motor recovery and are more likely to develop spasticity, compared with patients with an intact tract.^{34 44}

Throughout this study, we want to expand the knowledge of the effects of LMV to optimise their use in clinical settings. These vibration modalities possibly will become a strong complement to physical therapy to promote plasticity during recovery following a stroke.

Trial status

Recruitment started in June 2022 and will end in June 2025. The trial will end 7.5 months after the last inclusion, in January 2026.

Ethics statements

Patient consent for publication

Not applicable.