Abstract
Background Botulinum toxin A (BoNT-A) injections are increasingly used to treat muscle spasticity and are often complemented by adjunctive rehabilitation therapies; however, little is known about the effect of therapy after injection.
Purpose The aim of this study was to identify and summarize evidence on rehabilitation therapies used after BoNT-A injections to improve motor function in adults with neurological impairments.
Data Sources Searches were conducted in PubMed, EMBASE, Cochrane Central Register of Controlled Trials (CENTRAL), CINAHL, National Research Register, metaRegistry of Controlled Trials, PEDro, and OTseeker.
Study Selection Randomized and quasi-randomized controlled trials were considered for inclusion. Participants with neurological impairments received BoNT-A to treat focal spasticity in limbs, with rehabilitation interventions provided to experimental groups only. Primary outcome measures were joint mobility, function of the affected limb, and spasticity. Eleven studies with 234 participants, most of whom had stroke, were included in the review.
Data Extraction Two reviewers extracted study details and data.
Data Synthesis Methodological quality was rated using the PEDro scale. Both fixed-effects and random-effects models were used to calculate effect size.
Results Studies were of variable quality: 3 were poor (PEDro score 1 to 4), and 8 were moderate (PEDro score 6 to 7). No study investigated effects for longer than 24 weeks (6 months). Included trials presented 9 therapy types, including ergometer cycling, electrical stimulation, stretch (casting, splinting, taping, or manual or exercise-induced stretch), constraint-induced movement therapy, task-specific motor training, and exercise programs. Statistical findings suggest that combined therapy and BoNT-A is slightly more effective than BoNT-A alone.
Conclusion Evidence relating to impact of adjunct therapy is available, but the heterogeneity of studies limits the opportunity to demonstrate overall impact. Researchers need to consider the benefits of greater consistency in study approaches and measures so that meaningful evaluations of overall adjunct therapy effects can be made.
Spasticity is a common impairment after central nervous system damage. It was defined by Lance in 1980 as “a motor disorder characterized by a velocity-dependent increase in tonic stretch reflexes (muscle tone) with exaggerated tendon jerks, resulting from hyper excitability of the stretch reflex.”1(p485) This motor disorder can have a profound impact on health outcomes and health-related quality of life.2 It is associated with pain3 and often has a detrimental impact on function, self-esteem, and body image.4 General health can suffer, as spasticity leads to difficulties in hygiene and maintaining skin integrity in affected limbs, with consequent infection common.5 Current treatment options for spasticity are expensive and time-consuming, and their effectiveness is “less than satisfactory.”6 There is an urgent need to identify treatment that is safe, effective, and efficient.7
In the past decade, traditional rehabilitation for spasticity, such as occupational therapy and physical therapy, has been augmented by pharmacotherapy such as botulinum toxin A (BoNT-A), which has been shown to reduce spasticity.8 One of the limitations of pharmacotherapy is that its effect is temporary, and many patients with spasticity have long-term rehabilitation goals. Typical goals include improving performance of everyday activities, increasing range of motion (ROM) to improve limb position, and ensuring skin integrity.9–11
Following BoNT-A injection, rehabilitation therapies, including stretch, splinting, casting, strengthening exercises, and movement training, are often provided in the context of functional activities where normal movement patterns are encouraged.2,6,12,13 These therapies are thought to improve ROM and strength and, in turn, improve functional performance in patients with motor disorders following neurological impairment.2 Both BoNT-A and rehabilitation therapies have been shown, together and separately, to have an impact on health outcomes such as walking ability.14,15 To date, however, no study has examined the cumulative evidence regarding combined BoNT-A and rehabilitation therapies for treatment of spasticity in the upper or lower limbs of adult patients with neurological impairment. Given the cost and scale of providing therapies, and the breadth and complexity of health and quality-of-life problems involved, a review that provides a comprehensive summary of evidence relating to this problem and intervention is needed.
The aim of this systematic review was to identify and summarize the evidence on rehabilitation therapies used after BoNT-A injection to improve motor function in adults with neurological impairments compared with BoNT-A alone. The results of this systematic review will provide a starting point for developing evidence-based practice guidelines for the management of spasticity. The primary objective of this review was to determine whether BoNT-A injections when combined with rehabilitation are more effective in reducing spasticity compared with BoNT-A alone. The secondary objectives were to determine the effects on health-related quality of life, caregiver care time, and mortality and to report any adverse events.
Method
Data Sources and Searches
The protocol for this review is available elsewhere.16
Search Strategy
Computerized databases were searched in May 2010 and again in August 2013. Databases included in the search were: PEDro, Cochrane Central Register of Controlled Trials (CENTRAL), PubMed, EMBASE, CINAHL, National Research Register, metaRegister of Controlled Trials, and OTseeker. Unpublished, non–peer-reviewed sources, such as conference abstracts, were not included. Reference lists of included studies also were reviewed. Authors of included studies were contacted for additional and unpublished data where necessary and possible.
Study Selection
Studies were eligible only if they were published in English.
Types of studies.
Randomized and quasi-randomized controlled trials were included in this review. Studies could use parallel-group, within-subjects, or crossover designs. Only the first arm of crossover trials was included in this systematic review.
Types of intervention.
Studies that involved any form of physical therapy or rehabilitation interventions provided to patients following BoNT-A injection versus studies that did not involve any form of physical therapy or rehabilitation and used BoNT-A injection only to treat upper or lower limb spasticity were included in the review.
Types of participants.
Studies that involved adolescents and adults (age 16 years and older) with a neurological impairment (central nervous system damage) and injection of BoNT-A up to 3 months prior to study commencement for spasticity management were included in the review.
Types of outcome measures.
Studies that included measures of motor function of the affected limb, ROM, tone, and spasticity were included in the review. Secondary outcomes of interest were health-related quality of life, caregiver burden, mortality, and adverse events. Measures of treatment effect were extracted for 2 time frames: immediate (ie, effects present immediately after therapy ceases or within 24 hours of removal of intervention) and long term (outcomes after effect of BoNT-A has worn off over 3 months but before the next injection).
Studies were excluded from the review if BoNT-A was not used in treatment of spasticity or a second publication of the same study presented the same results.
Two review authors (B.K. and N.A.L.) independently screened all search results and reviewed study characteristics and outcome data from included studies using a Cochrane data extraction form. A third author (A.C.) moderated disagreements.
Data Extraction and Quality Assessment
The following data were extracted: information about study design, inclusion and exclusion criteria, characteristics of the participants, details of the interventions, and results of the outcome measures used. Discrepancies were checked against the original data. One author (B.K.) entered data in RevMan (Review Manager) meta-analysis software (Copenhagen, Denmark, The Nordic Cochrane Centre, The Cochrane Collaboration).17
The risk of bias in each study was assessed using Cochrane risk of bias tables.18 Each study was rated on 8 domains: sequence generation; allocation concealment; blinding of participants, therapists, and outcome assessors; incomplete data; selective outcome reporting; and other potential threats to validity noted by the reviewers. Methodological quality was assessed by one reviewer (B.K.) using the PEDro scale. The PEDro scale has established reliability and provides a score out of 10.19 Studies that attain a PEDro score of 7 or greater are considered “high quality,” those with a PEDro score of 5 or 6 are considered “moderate quality,” and those with a PEDro score of 4 or less are considered “poor quality” in terms of study methods and susceptibility to bias.20 Adequacy of concealment was rated using the procedure outlined by Schulz et al.21
Data Synthesis and Analysis
Analysis-of-covariance–adjusted between-group means and standard deviations were extracted in preference to between-group differences in change scores, and between-group differences in change scores were extracted in preference to between-group differences in final scores. Where data were reported as medians and interquartile ranges, medians were used as a surrogate for means and standard deviations and were estimated as 80% of the interquartile range (studies that did not report an interquartile range were excluded from the meta-analysis). For continuous outcome measures, a pooled estimate of treatment effect was determined by calculating the mean difference and the corresponding 95% confidence interval (95% CI). For dichotomous outcome measures, a pooled estimate of treatment effect was calculated for each outcome measure using risk ratio where appropriate and the corresponding 95% CI. Time-to-event data were analyzed using the hazard ratio and 95% CIs as required. In crossover trials, only the first-arm data were used in the meta-analyses. In the event of missing, incomplete, or unclear data, the original investigators were contacted, and these data were requested. The study results were just described where insufficient data were provided for inclusion in the meta-analysis or where studies were not clinically similar to others.
Meta-analyses were considered if there were at least 2 homogeneous therapies, using a fixed-effects model. Where heterogeneity was reported (I2>30% along with a significant chi-square test result, P<.01),22 a random-effects model was used, provided that studies investigated the effect of similar interventions on similar populations and reported similar outcomes.23 Data were not pooled where heterogeneity was substantial using the random-effects model (I2>50%).
Results
Study Selection and Characteristics
The search identified 1,314 references; of these, 516 were duplicates (Fig. 1). Twelve studies were identified as potentially eligible. On review of the full articles, 11 studies were included, and 1 study was excluded (Tab. 1). Six studies, with a total of 129 participants, investigated the adjunctive effect of BoNT-A and rehabilitation therapy in the upper limb. Four studies, with a total of 61 participants, investigated the adjunctive effects of therapy and BoNT-A in the lower limb. One study, with a total of 38 participants, included upper and lower limbs; this study's data were included in the analyses of both groups (Tab. 1).
Flow diagram: physical therapies as an adjunct to botulinum toxin-A injection of the upper or lower limb in adults following neurological impairment. Adapted from Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009;6:e1000097.
Study Detailsa
All study participants experienced spasticity following a neurological impairment. Ten studies included participants with spasticity following a stroke,13,15,24–32 and 1 study included participants with spasticity due to multiple sclerosis.32 Rehabilitation therapies were applied following BoNT-A across all studies to participants in the experimental groups.13,15,24–31 Rehabilitation interventions in the studies were categorized for data analysis into 4 therapy intervention groups: stretch, electrical stimulation, constraint-induced movement therapy (CIMT), and exercise-based therapy. Stretch interventions (casting, splinting, taping, and manual or exercise-induced stretch) were the most common and were used in 5 studies.15,25,28,30,32 The dose of stretch applied across these studies was variable (stretch was applied for a cumulative length of time ranging from 40 minutes to 4 weeks).15,25,28,30,32 Three studies investigated use of electrical stimulation.26,27,31 Settings and frequency of electrical stimulation applied were not clearly described in the studies. Dosage varied across studies (electrical stimulation was applied for a cumulative length of time of 4.5 hours, 9 hours, and 84 hours).26,27,31 Two studies13,29 investigated the effects of CIMT. Both studies had significant variation in their protocols of modified CIMT, with the study by Levy et al29 using a therapy-based constraint protocol for 2 weeks administered over a cumulative total of 70 hours. The study by Sun et al13 used a higher-dosage protocol that incorporated a therapy-based constraint protocol in addition to a constraint protocol that extended into participants' waking hours for 12 weeks with a cumulative dosage of 540 hours. The study by Diserens et al24 used an exercise therapy approach involving repetitive arm cycling 30 minutes per session 3 times per week for 2 weeks with a cumulative dosage of 3 hours.
Spasticity was an outcome measure in 7 studies,13,15,27,28,30,32 joint mobility or position was an outcome measure in 4 studies,24,27,28,30 motor function was an outcome measure in 6 studies,13,15,25,26,30,31 and passive caregiver tasks were an outcome measure in 1 study.27 Of the 11 studies included in the review, 1 study32 used a quality-of-life measure in which participants rated their satisfaction with relief from spasticity on a visual analog scale.
Adverse Events
Adverse events were reported in 3 studies,15,26,30 and all adverse events were related to the BoNT-A injections only. Karadag-Saygi et al15 and Reiter et al30 reported no serious adverse effects, but 2 participants in each study reported transient pain at the injection site. Hesse et al26 reported bladder paresis in 1 participant following injection with BoNT-A, and this participant also experienced a grand mal seizure 3 weeks following BoNT-A injection. There were no adverse events reported in any study as a result of rehabilitation therapy interventions.
Quality of Evidence
The risk of bias in the 11 studies was variable (Tabs. 2 and 3). Three of the studies (27%) failed to use adequate methods of randomization, 10 studies (91%) failed to conceal allocation, and 6 studies (55%) failed to adequately blind outcome assessors. All studies failed to blind therapists administering interventions. Results from all studies were included in the main analysis regardless of quality.
Methodological Quality Summary (Risk of Bias)a
PEDro Scale Study Detailsa
Effect of Therapy on Spasticity Following BoNT-A
Five studies15,25,28,30,32 investigated the immediate effects of stretch interventions following BoNT-A on spasticity (Fig. 2) as measured using the modified Ashworth scale. The overall effect was a standard mean difference (SMD) of −0.48 favoring the group that received stretch interventions (95% CI=−0.85, −0.11); no statistical heterogeneity was identified (I2=0%, P=.55; test for overall effect, P=.01). The long-term effects of stretch-based interventions on spasticity following BoNT-A were investigated by pooling these same 5 studies15,25,28,30,32 (Fig. 3). Substantial statistical heterogeneity (I2=75%, P=.003) was detected when using a random-effects model; therefore, meta-analysis is not presented for long-term stretch outcomes.
Forest plot comparing the immediate effects (between 2 and 4 weeks) of stretch interventions following botulinum toxin-A (BoNT-A) versus BoNT-A alone on spasticity. IV=inverse variance, 95% CI=95% confidence interval.
Forest plot of comparison of the long-term effects (post-12 weeks) of stretch interventions on spasticity following botulinum toxin-A versus no stretch interventions. IV=inverse variance, 95% CI=95% confidence interval.
The use of electrical stimulation following BoNT-A was examined in only one study of 12 participants.27 This study reviewed the immediate effects of electrical stimulation plus BoNT-A on reducing spasticity and reported an SMD of −0.15 (95% CI=−1.28, 0.98; test for overall effect, P=.79) favoring the electrical stimulation intervention group. At 12 weeks postinjection, the SMD was −0.43 (95% CI=−1.58, 0.72; test for overall effect, P=.47), again favoring the electrical stimulation intervention group.27
Two studies, with a total of 36 participants, investigated the effect of CIMT after BoNT-A.13,29 There was a small reduction in spasticity following the immediate effects of therapy (SMD=0.05, 95% CI=−0.77, 0.68; test for overall effect, P=.90). Long-term effects showed an SMD of −0.88 (95% CI=−1.65, −0.11; test for overall effect, P=.02).
Effect of Therapy on ROM Following BoNT-A
Two studies investigated the effect of a combined approach of therapy and BoNT-A on passive ROM.15,30 These studies reviewed the effect of stretch-based interventions on dorsiflexion of the ankle. Pooled immediate effects produced an SMD of −0.15 favoring the group that also received stretch (95% CI=−0.81, −0.51); statistical heterogeneity testing demonstrated an I2 of 73%, but with a nonsignificant chi-square value (3.71, P=.65). Long-term effects of a combined therapy approach produced an SMD of −0.35 (95% CI=−1.01, 0.30), showing a small benefit of stretch plus BoNT-A, although statistical heterogeneity was present (I2=67%, χ2=3.06, P=.29).
Two other studies investigated the effects of therapy on passive ROM27 and active ROM28 of the upper limb. These studies used different outcome measures and, therefore, could not be combined and are not included in the pooled analysis. One study27 investigated the impact of combined electrical stimulation and BoNT-A on resting limb position, reporting an SMD of −0.61 (95% CI=−1.78, 0.56). The other study28 investigated the effect of stretch interventions on active ROM, reporting an immediate effect of stretch on active ROM (SMD=0.36, 95% CI=−0.36, 1.08; test for overall effects, P=.33) and a long-term effect (SMD=0.35, 95% CI −0.37, 1.07; test of overall effects, P=.34).
Effect of Therapy on Motor Function Following BoNT-A
The effect of therapy on motor function was investigated in 6 studies.13,15,25,26,30,31 Therapies included electrical stimulation, CIMT, and stretch. These studies used 14 different outcome measures; therefore, only 2 studies were sufficiently homogeneous to permit pooled analysis.15,30 These 2 studies examined the effect of stretch therapies on step length and gait speed (m/s).15,30 The immediate effects of stretch on step length were an SMD of −0.13 (95% CI=−0.77, 0.52) and a P value of .7 for test for overall effect, indicating a beneficial effect (statistical heterogeneity testing, I2=18%). The immediate effect of stretch on gait speed was an SMD of −0.20 (95% CI=−0.44, 0.84; test for overall effect, P=.54) favoring the control group; there was no statistical heterogeneity (I2=0%). These same studies15,30 identified long-term effects from stretch-based therapy following BoNT-A that resulted in an SMD of −0.11 (95% CI=−0.75, 0.53) for step length (χ2=0.76, I2=0%, P=.73) and an SMD of −0.18 (95% CI=−0.46, 0.82) for gait speed (χ2=0.18, I2=0%, P=.59), both of which favored the experimental group, although the SMDs were small for both analyses (≤0.2).
Effect of Therapy on Quality of Life Following BoNT-A
Only one study32 measured the effect of rehabilitation therapies after BoNT-A on quality of life. Participants evaluated their satisfaction regarding relief from spasticity using a visual analog scale after each period of combined therapy (BoNT-A plus therapy). The immediate effect of perceived symptoms indicated an SMD of −0.25 (95% CI=−0.89, −0.39; test for overall effect, P=.44) favoring the experimental group. Long-term effects demonstrated an SMD of 1.79 (95% CI=−1.03, 2.56) favoring the control group; however, the SMD size was small.
Discussion
This is the first systematic review to evaluate the effects of rehabilitation therapies when used in combination with BoNT-A to treat spasticity in adults with neurological impairment. Eleven studies of varied methodological quality were included in the review.
Five pooled studies provided marginal evidence for additional reduction of spasticity when rehabilitation therapies were used with BoNT-A injection (Figs. 2 and 3). This finding was evident in both immediate and follow-up results (maximum of 3 months). However, this finding is limited by the small number of studies and the heterogeneity of those included. There were few studies where BoNT-A alone was used as the primary intervention condition in a control group. It is possible that the paucity of such studies is because the mainstay of spasticity management has been inpatient rehabilitation therapy, and the use of BoNT-A is relatively new. Studies in which BoNT-A is provided without any therapy interventions at all are thus uncommon.
Three studies15,26,30 reported adverse events specific to receiving BoNT-A and not the therapy programs received. The types of rehabilitation therapies used in the studies included motorized ergometer cycling, electrical stimulation, stretch interventions (casting, splinting, taping, manual or exercise-induced stretch), CIMT, task-specific motor training, and exercise programs. These may not represent all rehabilitation therapies provided after BoNT-A injection in clinical practice.
Comparisons of therapy type and impact with and without BoNT-A were challenging because of the wide variety of rehabilitation therapies used, the varied doses, and the disparate number and types of outcome measures reported in the studies. As a result, we are unable to provide guidance regarding the most effective adjunct therapy intervention to use in conjunction with BoNT-A. Out of all of the therapies reviewed, there is preliminary evidence to suggest that therapies such as stretch and CIMT used adjunctively to BoNT-A may be more effective than the use of BoNT-A alone. It is important to acknowledge that uncertainty remains around how effective these therapies are with the 95% CI sometimes spanning zero.
There is an urgent need for large-scale, rigorous clinical trials that investigate the relative efficacy of therapy types as independent or combined interventions with BoNT-A injection. Until such studies are conducted using therapy approaches, designs, and measures that enable synthesis of findings, it will remain difficult to define evidence-based spasticity management in rehabilitation.
A surprising finding of the present review was that no study measured activities of daily living, even though a common and important goal for patients is to improve their ability to participate and achieve independence in meaningful life roles.9–11 Apart from patient goals, some measurement of activities of daily living function is typical of institutional reporting or, in some cases, eligibility for subsidized funding of BoNT-A, making the absence of such measures in the studies perplexing. Outcomes such as gait speed and step length that could be considered proxy indicators of functional movement were measured in 2 studies15,30 and demonstrated a small, additional benefit for the treatment groups.
In conclusion, our findings are equivocal because the reviewed studies were few in number, of variable quality, inconsistent in time points assessed and measures used, and diverse in analytic approaches. Our findings suggest that the combination of stretch therapies and BoNT-A may be statistically more effective than the use of BoNT-A alone; for other outcomes, this association was less clear. Adverse events were rarely reported. This review challenges future researchers to fill evidence gaps. Future research studies need to increase quality through: (1) use of rigorous designs that include intention-to-treat analyses, adequately sized samples, and blinded outcome assessment; (2) reliable and valid outcome measures for short-term and long-term impacts, including those that relate to activities of daily living and function; (3) treatment conditions that include single and multiple rehabilitation intervention types so that efficacy of different therapies can be compared in the BoNT-A adjunct context; (4) precise control and recording of BoNT-A doses in analysis; and (5) careful consideration of measurement time points and endpoints so that meaningful pooling of study results can occur. An expanded high-quality research base will provide the necessary foundation for evidence-based practice that can improve quality of life, demonstrate clinical impact, and show economic benefit of a combined injection-therapy approach.
Footnotes
All authors provided concept/idea/research design. Ms Kinnear, Dr Lannin, Dr Cusick, and Dr Harvey provided writing. Ms Kinnear and Dr Lannin provided data collection. Ms Kinnear, Dr Lannin, and Dr Cusick provided data analysis. Ms Kinnear provided project management and administrative support. Dr Rawicki provided consultation (including review of the manuscript before submission).
The study was conducted as part of Ms Kinnear's PhD program.
The study was presented as an abstract at Occupational Therapy Australia, 24th National Conference and Exhibition; June 29–July 1, 2011; Broadbeach, Australia; and orally presented at the Stroke Society of Australasia Conference; August 29–31, 2012; Sydney, Australia.
- Received September 11, 2013.
- Accepted July 4, 2014.
- © 2014 American Physical Therapy Association
References
Issue highlights
