Effects of Electrical Stimulation in Spastic Muscles After Stroke
Systematic Review and Meta-Analysis of Randomized Controlled Trials
Background and Purpose—Neuromuscular electric stimulation (NMES) has been used to reduce spasticity and improve range of motion in patients with stroke. However, contradictory results have been reported by clinical trials. A systematic review of randomized clinical trials was conducted to assess the effect of treatment with NMES with or without association to another therapy on spastic muscles after stroke compared with placebo or another intervention.
Methods—We searched the following electronic databases (from inception to February 2015): Medline (PubMed), EMBASE, Cochrane Central Register of Controlled Trials and Physiotherapy Evidence Database (PEDro). Two independent reviewers assessed the eligibility of studies based on predefined inclusion criteria (application of electric stimulation on the lower or upper extremities, regardless of NMES dosage, and comparison with a control group which was not exposed to electric stimulation), excluding studies with <3 days of intervention. The primary outcome extracted was spasticity, assessed by the Modified Ashworth Scale, and the secondary outcome extracted was range of motion, assessed by Goniometer.
Results—Of the total of 5066 titles, 29 randomized clinical trials were included with 940 subjects. NMES provided reductions in spasticity (−0.30 [95% confidence interval, −0.58 to −0.03], n=14 randomized clinical trials) and increase in range of motion when compared with control group (2.87 [95% confidence interval, 1.18–4.56], n=13 randomized clinical trials) after stroke.
Conclusions—NMES combined with other intervention modalities can be considered as a treatment option that provides improvements in spasticity and range of motion in patients after stroke.
Clinical Trial Registration Information—URL: http://www.crd.york.ac.uk/PROSPERO. Unique identifier: CRD42014008946.
Stroke is a leading cause of serious long-term disability in the United States. The American Heart Association estimated an overall stroke prevalence of 6.8 million Americans over the age of 20 years, accounting for 2.8% of the population, based on The National Health and Nutrition Examination Survey data from 2007 to 2010.1 The burden of stroke is a global problem that causes well-known long-term disabilities, and spasticity is one of them.2
Spasticity may be defined as a motor disorder characterized by velocity and acceleration-dependent increased resistance to passive muscle stretch and hyperactivity of stretch reflexes.3 The exact prevalence of spasticity is unknown. Recent studies showed that spasticity occurs in 20% to 30% of all stroke victims,4–6 and one recent study has reported contracture development in 50% of the cases 6 months after stroke.7
The pathophysiology of spasticity can occur as a result of abnormalities on different levels, including muscular and spinal properties, as well as supraspinal mechanisms.8 Traditional treatment modalities include use of an ankle-foot orthosis, physical therapy, systemic medications, tendon surgeries, and focal alcohol neurolysis. More recent treatment options include neuromuscular electric stimulation (NMES).9 It is one hypothesis that NMES induces specific plasticity of spinal chord pathways.10
However, although some randomized clinical trials (RCTs) showed the beneficial effects of NMES on the treatment of patients with spasticity after stroke,11,12 others did not find an additional reduction of spasticity with its usage when compared with a control group.13,14 A systematic review of the evidence would allow a more precise evaluation of its effectiveness and, if the benefits are proven, would aid in disseminating the use of NMES. Therefore, the aim of our study was to systematically review the effect of treatment with NMES on spastic muscles after stroke compared with placebo or another intervention.
Protocol and Registration
This systematic review was performed in accordance with the Cochrane Collaboration15 and the Preferred Reporting Items for Systematic Review and Meta-Analyses: the PRISMA Statement.16 The protocol of the study was registered at the International Prospective Register Of Systematic Reviews, PROSPERO, under the identification CRD42014008946 and can be integrally assessed online (http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42014008946#).
To be included in the review, the studies had to be RCTs which determined the effects of NMES or NMES combined with other treatment techniques on spasticity in stroke patients and had to have at least one intervention group and one comparison group. Studies that applied NMES on the lower or upper extremities were included, regardless of NMES dosage. The comparison group should not have been exposed to NMES in the same regimen as the NMES group, and the intensity of stimulation should not lead to visible or palpable contractions.
The exclusion criteria were having stroke survivors without muscular spasticity as participants, implementation of the intervention in regions other than feet or hands, and interventions with <3 days.
Literature searches were conducted in the following electronic databases (from the inception to February 2015): MEDLINE (accessed by PubMed), EMBASE, Cochrane Central Register of Controlled Trials (Cochrane CENTRAL), and Physiotherapy Evidence Database (PEDro). The search terms used individually or combined included stroke, electric stimulation, and a string of words previously proposed, which yielded a high sensibility in the search for randomized controlled trials.17 To enhance the sensitivity of the search, words related to the outcomes of interest were not included. There were no language restrictions in the strategy, but non-English studies were included in the review only when translation was available. The references included in the published articles identified in these searches were used as an additional source to identify other clinical trials. The complete search strategy used for the PubMed database is shown in Table 1. The terms were adjusted to fit the requirements of each electronic database.
Study Selection and Data Extraction
Two reviewers (C. Stein, C.G. Fritsch) separately and independently screened the titles and abstracts of studies identified in the initial searches. A standard screening checklist based on the eligibility criteria was used for each study. Studies that did not meet the criteria according to the titles or abstracts were excluded. Full text versions of the remaining studies, including those potentially eligible and uncertain, were retrieved independently for a second review by the 2 reviewers to determine the eligibility. Disagreements regarding study eligibility were discussed between reviewers. When consensus was not reached, a third reviewer (C. Robinson) arbitrated. For studies without sufficient information to evaluate the eligibility, the authors were contacted via e-mail to obtain clarifications. The studies with insufficient information after this contact were excluded, or the procedures for estimation of missing data15 were performed when possible. Studies with >1 publication reporting the results from the same population were excluded, and the publication with the largest sample size was chosen. Abstracts published in academic conferences were evaluated by case, and the authors were contacted for details when necessary. The reviewers were not blinded to the authors and institutions of the studies undergoing review.
The following data were extracted from included studies: methodological design, number of subjects, comparison groups, intervention protocol, and results of the outcomes. The primary outcome extracted was spasticity, assessed by Modified Ashworth Scale, and the secondary outcome extracted was range of motion, assessed by Goniometer. Two review authors (C. Stein, C.G. Fritsch) separately and independently extracted the data, and disagreements regarding the data extraction were solved by discussion. When consensus was not reached, a third author (C. Robinson) arbitrated. When data were missing for synthesis or study quality assessment, we contacted the study authors via e-mail at least 2 times. The study was excluded if data were still insufficient after this process.
Risk of Bias Assessment
Two review authors (C. Stein, C.G. Fritsch) independently assessed the risk of bias of the included studies by considering the items established in the Cochrane Collaboration’s tool (27) for assessing risk of bias within and across randomized trials: adequate sequence generation, allocation concealment, blinding of patients and investigators, blinding of outcome assessors, description of losses and exclusions, and intention-to-treat analysis. Studies without a clear description of these items were considered as unclear or not reporting the item. Given the small number of included studies, it was considered inappropriate to present publication bias through funnel plot.
After data extraction, if the outcome measurements could not be transformed in a common numeric scale for quantitative synthesis, a descriptive synthesis was performed. For quantitative synthesis, pooled-effect estimates were obtained by comparing the change from baseline to study end for each group. Regarding the continuous outcomes, if the unit of measurement was consistent across trials, the results were presented as the weighted mean difference with 95% confidence intervals (95% CIs), and if the unit of measurement was inconsistent, the results were expressed as the standard mean difference with 95% CI. Calculations were performed using the random effects method, given the heterogeneity of outcome measurements (I2>0). A P value ≤0.05 was considered statistically significant. The statistical heterogeneity of the treatment effects among studies was assessed using Cochran’s Q test and the inconsistency I2 test, in which values above 25% and 50% were considered as indicatives of moderate and high heterogeneity, respectively. To reduce the statistical heterogeneity, a sensitivity analysis was performed considering the type of comparison (control or placebo) group. All analyses were conducted using Review Manager, version 5.2.
Description of Studies
From 5066 potentially relevant citations retrieved from electronic databases and reference lists searches, 29 RCTs11–14,18–42 met the inclusion criteria and were included in the systematic review, providing data on 940 subjects. Six additional studies were found from the reference list of published articles. The 29 included studies reported data on spasticity, but only 14 where included in the meta-analysis. Fifteen studies were not included in the meta-analysis for spasticity because it was not possible to obtain clarifications about the missing data on the publication18,26,30–33,40,41 and because other scales were used to assess spasticity.19,34–39 The 13 studies which assessed range of motion were included in our second meta-analysis. Figure 1 shows the flow diagram of the included studies, and Table I in the online-only Data Supplement summarizes the characteristics of these studies (online-only Data Supplement).
In 22 studies,11,13,14,18–36 the stimulation frequencies with NMES ranged from 18 to 50 Hz and pulse duration from 0.1 to 0.4 ms. In 3 studies12,37,38 the stimulation frequencies with NMES ranged from 80 to 100 Hz and pulse duration from 0.1 to 0.3 ms. Four studies39–42 did not describe the stimulation characteristics. The time of intervention the studies was averaged 3038.75 minutes. The intervention occurred in an outpatient environment, except for one study that performed a domiciliary approach.31
Risk of Bias
93% of the studies presented adequate sequence generation and 86% described losses to follow-up and exclusions, showing a low risk of bias for these analyses. When considering the assessment of the outcomes, 62% reported blinding of the investigators, with a moderate risk of bias. Therefore, 28% reported adequate allocation concealment, 14% reported blinding, and 24% performed intention-to-treat analyses, showing high risk of bias for these characteristics (Table 2).
Effects of Interventions
Fourteen studies11–14,20–25,27–29,41 (n=383) evaluated spasticity using the Modified Ashworth Scale and were included in the meta-analysis. NMES or NMES combined with other treatment techniques was associated with reductions in spasticity when compared with control group (−0.30 [95% CI, −0.58 to −0.03; I2 81%]; Figure 2A). Fifteen studies were not included in the meta-analysis for spasticity because it was not possible to obtain clarifications about the missing data on the publication18,26,30–33,40,41 and because other scales were used to assess spasticity.19,34–39
To investigate possible differences between studies, a sensitivity analysis related to type of intervention and the site of application of NMES was performed. Twelve studies11–14,20,22,24,25,27–29,41 that applied NMES combined with other treatment technique interventions showed significant reductions in spasticity compared with control (−0.35 [95% CI, −0.63 to −0.07; I2 80%]; Figure 2B). The other 2 studies21,23 applied only NMES and showed no significant reductions in spasticity (0.13 [95% CI, −1.53 to 1.78; I2 92%]; Figure 2B). Five studies that used NMES on the leg11,12,14,28,29 showed significant reduction in spasticity compared with control (−0.78 [95% CI, −1.02 to −0.54; I2 48%]; Figure 2C), whereas six studies which used NMES on the wrist13,21–23,25,41 showed no significant reduction in spasticity (0.12 [95% CI, −0.41 to 0.64; I 2 81%]; Figure 2C). Four other trials applied NMES on the elbow20,22,25,42 and reported no significant reduction in spasticity (−0.39 [95% CI, −0.89 to 0.11; I2 54%]; Figure 2C).
We performed a sensitivity analysis to evaluate possible differences in RCTs that comprised botulinum toxin combined with NMES compared with control group. Three studies had this characteristic and in 2 of them the control group intervention was botulinum toxin,24,25 whereas the other was physiotherapy. The 3 studies14,24,25 showed no significant improvements on spasticity (−0.39 [95% CI, −0.97 to 0.19; I2 47%)]; Figure 2D) with the addition of NMES in comparison with the control groups.
The studies that were not included in the spasticity meta-analysis reported different results. In 8 studies,19,26,32–35,37,38 there was an improvement in the spasticity on the intervention group when compared with control group (P<0,05). In 2 RCTs,40,41 there was a tendency of reduction in spasticity on the intervention group when compared with control group. In 5 studies,18,30,31,36,39 there was no differences between intervention and control groups (P>0,05).
Range of Motion
Range of motion was evaluated with the use of goniometer in 13 trials11,12,21–23,28,29,34,36–39,42 (n=447), and they were all included in meta-analysis. NMES or NMES combined with other treatment technique interventions was associated with an increase in range of motion when compared with the control group (2.87 [95% CI, 1.18–4.56; I2 60%]; Figure 3A).
Sensitivity analysis was performed to investigate possible differences between studies toward type of intervention and site of application of NMES. Eleven studies11,12,22,28,29,34,36–39,42 that applied NMES combined with other treatment techniques showed significant increase in range of motion when compared with control (2.73 [95% CI 1.07–4.39; I2 62%]; Figure 3B). The other 2 studies21,23 that used only NMES showed no significant improvement in range of motion (6.93 [95% CI, −9.31 to 23.16; I 2 62%]; Figure 3B). Five RCTs applied NMES on the leg combined with other treatment techniques interventions11,12,28,34,38 and reported a significant increase in range of motion when compared with control (3.13 [95% CI 0.61–5.64; I 2 77%]; Figure 3C). Seven trials applied NMES or NMES combined with other interventions on the wrist21,23,36,38,39,42 and showed no significant improvement in range of motion (0.46 [95% CI −2.28 to 3.21; I 2 60%]; Figure 3C). Three other trials applied NMES combined with other interventions on the elbow22,39,42 and showed significant effects on range of motion (4.57 [95% CI 0.57–8.57; I 2 0%]; Figure 3C).
This systematic review with meta-analysis showed that NMES combined with other treatment techniques is an additional treatment option that provides improvements in spasticity and increase in range of motion in stroke patients. To the best of our knowledge, this is the first systematic review with meta-analysis evaluating the effectiveness of treatments with NMES on spastic muscle after stroke compared with placebo or another intervention.
The use of NMES in the rehabilitation of patients with neurological diseases has increased in recent years.43 Its effects on the reduction of spasticity may be explained by its actions on increasing Ib fiber activation via mechanisms that facilitate the Renshaw cell recurrent inhibition, on antagonist reciprocal inhibition, and on increasing cutaneous sensory stimuli.44,45
The application of NMES combined with other interventions was associated with reductions on spasticity and improvements in range of motion when compared with a control group. As spasticity may impair functional activities in stroke patients, it is necessary to control it before applying any motor control therapeutic protocol.12 From our findings, the application of NMES to reduce spasticity in these patients can be recommended. It would lead to a bigger benefit from motor control programmes and a better improvement in the functional activity. The usage of NMES could not inhibit the use of the unaffected hand, but it could enhance and facilitate the patient to use the affected hand for day-to-day tasks, resulting in improvements in range of motion.22
Additionally, when we performed sensitivity analysis related to the site of application of NMES, applications on wrists showed no effects on both spasticity and range of motion and on elbows showed no effects only on spasticity. These results may be because of differences between trials. For example, participant groups varied between studies in time after stroke (ranged from 1.5 months to >12 months), time of treatment, degree of spasticity, and degree of functional deficit. It is possible that time after stroke, time of treatment, degree of spasticity, and degree of ability to voluntarily contract a muscle might affect response to electric stimulation. Another source of variation was the different conventional therapies used as comparator treatments in included trials, which were Bobath,12 Active leg cycling,19 SMART Arm,20 Conventional Occupational Therapy,22 Botulinium Toxin A,24 and Stretching with proprioceptive neuromuscular facilitation technique.37 Furthermore, the different outcome measures, which were not always possible to combine, resulted in some units of analysis containing data from only one study.
Moreover, chronic tissue changes because of immobilization, such as atrophy, loss of sarcomeres, muscle conversion to connective tissue, and decreased resting length of the muscle, may compromise the success of NMES.46 In addition, loss of motor units in the paretic arm, which might be caused by secondary trans-synaptic degeneration, could compromise effective NMES performance.47
When we analyze Figure 2A, we observe that the studies11,12,21,28,29 that are favorable to the usage of NMES or NMES combined with other treatment techniques have a greater weight in determining the final result. In relation to range of motion (Figure 3A), 3 studies11,12,21 that have a greater weight have a favorable result for the intervention, whereas the remaining studies did not favor any group.
In the subanalysis with the 3 studies concerning botulinum toxin combined NMES showed no significant reduction in spasticity.14,24,25 This can be explained by the small number of studies included in the subanalysis and the insufficiency of the sample size to demonstrate efficacy.
Concerning the stimulation parameters to be used, there are still discussions in the literature. There is scope of variation not only across studies but also across interindividuals to evoke optimal muscle contraction and to avoid discomfort, pain, and skin irritations.48 In this review, the usage of NMES with a frequency between 30 and 50 Hz and a pulse width between 0.1 and 0.5 ms for 30 minutes 5 times per week for 3 to 4 weeks were associated with successful results.
A limitation of this systematic review and meta-analysis is that most of the studies retrieved showed some biases. Two studies no properly described the generation of a random sequence33,38 and only 8 studies12,14,18,20,34,36,37,39 clearly described allocation concealment. Moreover, only 4 studies23,24,31,39 reported the patient blinding. Nonpharmacological studies that involve physical activity interventions have several limitations compared with drug RCTs. These limitations can involve learning curves, standardization of interventions, lack of blinding, and cointervention. The lack of blinding could be considered a major limitation of RCTs included in this review, but, as discussed earlier, this is an inherent limitation of nonpharmacological studies.
In addition to that, 11 studies13,14,26,28–30,32,38,40–42 did not describe the blinding of the assessors, 4 studies26,32,38,41 did not report drop-outs and exclusions that occurred during the treatment period, and only 7 studies12,13,18,20,24,36,39 used the intention to treat analysis. Therefore, sensitivity analyses were limited because of lack of methodological quality of the included studies and the small number of studies and participants.
In the meta-analysis of studies involving NMES or NMES combined with other treatment techniques versus control group, the criteria for spasticity assessment were different. Some studies used the 6-level Modified Ashworth Scale instead of the regularly used 5-level Ashworth Scale. Although the Ashworth Scale is a widely used measure of spasticity, it may not be sensitive enough to detect minimum changes. A modification of the scale has been created and adds an additional intermediate grade (1+).49
NMES combined with other intervention modalities is a treatment option that provides improvements in spasticity and range of motion in stroke patients. This data provides support for further NMES use as an additional therapy technique, but the conduction of large scale and high-quality RCTs is needed to establish its true efficacy.
C. Stein elaborated the systematic review protocol and the search strategy, performed the search strategy, the analysis process, and the redaction of the article, as well as its final revision. C.G. Fritsch performed the search strategy, the studies selection, the data extraction, gave intellectual support for the analysis process, the redaction of the article, and its final revision. C.C. Robinson gave intellectual support for the search strategy construction, arbitrated in the final consensus of included articles and risk of bias assessment, gave intellectual support for the analysis process, the redaction of the article, and its final revision. G. Sbruzzi gave intellectual support for the systematic review protocol elaboration, the search strategy construction, the analysis process, the redaction of the article, and performed its final revision. R.D.M. Plentz gave intellectual support for the systematic review protocol, the search strategy construction, the analysis process, the redaction of the article, and performed its final revision.
Sources of Funding
C. Stein and C. Robinson are supported by the the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) by Universidade Federal de Ciências da Saúde de Porto Alegre (UFCSPA). C. Gassen and G. Sbruzzi did not have sponsor for this study. R.D.M. Plentz is supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq). No sponsor or funding participated in the study design, data analysis and interpretation, article writing, and dissemination of results.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.115.009633/-/DC1.
- Received April 2, 2015.
- Revision received June 11, 2015.
- Accepted June 15, 2015.
- © 2015 American Heart Association, Inc.
- Go AS,
- Mozaffarian D,
- Roger VL,
- Benjamin EJ,
- Berry JD,
- Borden WB,
- et al
- Mesci N,
- Ozdemir F,
- Kabayel DD,
- Tokuc B.
- Bakhtiary AH,
- Fatemy E.
- Mangold S,
- Schuster C,
- Keller T,
- Zimmermann-Schlatter A,
- Ettlin T.
- Higgins JPT,
- Sally G.
- Robinson KA,
- Dickersin K.
- Bauer P,
- Krewer C,
- Golaszewski S,
- Koenig E,
- Müller F.
- Barker RN,
- Brauer SG,
- Carson RG.
- Boyaci A,
- Topuz O,
- Alkan H,
- Ozgen M,
- Sarsan A,
- Yildiz N,
- et al
- Chan MK,
- Tong RK,
- Chung KY.
- Hesse S,
- Reiter F,
- Konrad M,
- Jahnke MT.
- Mesci N,
- Özdemir F,
- Kabayel DD,
- Tokuç B.
- Mokrusch T.
- Cheng JS,
- Yang YR,
- Cheng SJ,
- Lin PY,
- Wang RY.
- Yan T,
- Hui-Chan CW,
- Li LS.
- Malhotra S,
- Rosewilliam S,
- Hermens H,
- Roffe C,
- Jones P,
- Pandyan AD.
- Heckmann J,
- Mokrusch T,
- Krockel A,
- Warnke S,
- von Stockert T,
- Neundorfer B.
- de Jong LD,
- Dijkstra PU,
- Gerritsen J,
- Geurts AC,
- Postema K.
- Quandt F,
- Hummel FC.
- Bohannon RW,
- Smith MB.