Multidisciplinary rehabilitation following botulinum toxin and other focal intramuscular treatment for post‐stroke spasticity

Abstract

Background

Spasticity may affect stroke survivors by contributing to activity limitations, caregiver burden, pain and reduced quality of life (QoL). Spasticity management guidelines recommend multidisciplinary (MD) rehabilitation programmes following botulinum toxin (BoNT) treatment for post‐stroke spasticity. However, the evidence base for the effectiveness of MD rehabilitation is unclear.

Objectives

To assess the effectiveness of MD rehabilitation, following BoNT and other focal intramuscular treatments such as phenol, in improving activity limitations and other outcomes in adults and children with post‐stroke spasticity. To explore what settings, types and intensities of rehabilitation programmes are effective.

Search methods

We searched the Cochrane Stroke Group Trials Register (February 2012), the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2011, Issue 12), MEDLINE (1948 to December 2011), EMBASE (1980 to January 2012), CINAHL (1982 to January 2012), AMED (1985 to January 2012), LILACS (1982 to September 2012), PEDro, REHABDATA and OpenGrey (September 2012). In an effort to identify further published, unpublished and ongoing trials we searched trials registries and reference lists, handsearched journals and contacted authors.

Selection criteria

We included randomised controlled trials (RCTs) that compared MD rehabilitation (delivered by two or more disciplines in conjunction with medical input) following BoNT and other focal intramuscular treatments for post‐stroke spasticity with placebo, routinely available local services, or lower levels of intervention; or studies that compared MD rehabilitation in different settings, of different types, or at different levels of intensity. We excluded RCTs that assessed the effectiveness of unidisciplinary therapy (for example physiotherapy only) or a single modality (for example stretching, casting, electrical stimulation or splinting only). The primary outcomes were validated measures of activity level (active and passive function) according to the World Health Organization's International Classification of Functioning, Disability and Health. Secondary outcomes included measures of symptoms, impairments, participation, QoL, impact on caregivers and adverse events.

Data collection and analysis

We independently selected the trials, extracted data, and assessed methodological quality using the Grades of Recommendation, Assessment, Development and Evaluation (GRADE). Due to the limited number of included studies, with clinical, methodological and statistical heterogeneity, quantitative meta‐analysis was not possible. Therefore, GRADE provided qualitative synthesis of 'best evidence'.

Main results

We included three RCTs involving 91 participants. All three studies scored 'low quality' on the methodological quality assessment, implying high risk of bias. All studies investigated various types and intensities of outpatient rehabilitation programmes following BoNT for upper limb spasticity in adults with chronic stroke. Rehabilitation programmes included: modified constraint‐induced movement therapy (mCIMT) compared with a neurodevelopmental therapy programme; task practice therapy with cyclic functional electrical stimulation (FES) compared with task practice therapy only; and occupational, manual therapy with dynamic elbow extension splinting compared with occupational therapy only. There was 'low quality' evidence for mCIMT improving upper limb motor function and spasticity in chronic stroke survivors with residual voluntary upper limb activity, up to six months, and 'very low quality' evidence for dynamic elbow splinting and occupational therapy reducing elbow range of movement at 14 weeks. Task practice therapy with cyclic FES did not improve upper limb function more than task practice therapy alone, only at 12 weeks. No studies addressed interventions in children and those with lower limb spasticity, or after other focal intramuscular treatments for spasticity.

Authors' conclusions

At best there was 'low level' evidence for the effectiveness of outpatient MD rehabilitation in improving active function and impairments following BoNT for upper limb spasticity in adults with chronic stroke. No trials explored the effect of MD rehabilitation on 'passive function' (caring for the affected limb), caregiver burden, or the individual's priority goals for treatment. The optimal types (modalities, therapy approaches, settings) and intensities of therapy for improving activity (active and passive function) in adults and children with post‐stroke spasticity, in the short and longer term, are unclear. Further research is required to build evidence in this area.

Plain language summary

Multidisciplinary rehabilitation programmes following treatment of spasticity after stroke

Stroke can cause muscle stiffness, spasms or tightness in the affected arm or leg, with pain and abnormal positioning of the limb. Consequently, there may be difficulties using the arm or leg in everyday activities or in caring for the affected limb. Treatments for spasticity may include botulinum toxin and other injected medications that paralyse the affected muscles. Following such injections, a multidisciplinary (MD) rehabilitation programme (usually delivered by two or more health professionals) is often employed. Interventions may include stretching, splinting, gait training, repetitive practice in using the arm for tasks, and orthotic prescription. Therapies are aimed at reducing spasticity to improve limb use or positioning, or to make it easier to care for the affected limb. The outcomes of such programmes focus on attainment of functional goals that are important to affected people in their everyday life. We included three relevant studies in the review, which investigated different types of MD rehabilitation interventions after botulinum toxin injections into the arms of 91 adults with previous stroke. There was low quality evidence for intensive forced use of the affected arm in improving spasticity, and very low quality evidence for elbow splinting with occupational therapy. We did not identify any studies of MD rehabilitation in children with post‐stroke spasticity or after other injected medications. The review findings are limited by the small number of studies that are methodologically flawed. More research is needed into what rehabilitation modalities and treatments are most effective for spasticity management following stroke.

Background

Description of the condition

Stroke and epidemiology

Stroke is the second leading cause of mortality and disease burden among adults aged 60 years and over (WHO 2003). The global burden of stroke is expected to rise due to the ageing population. Annually, 15 million people worldwide suffer a stroke. Of these, five million die and another five million are left permanently disabled, placing a burden on families and communities (Mackay 2004).

Spasticity is a common manifestation of neurological insults such as stroke. Whilst the incidence of spasticity is not known with certainty, it has been estimated to affect 38% of stroke survivors after 12 months (Watkins 2002). Direct costs for stroke survivors with spasticity have been found to be approximately four times those for stroke survivors without spasticity (Lundstrom 2010). The burden of post‐stroke spasticity is high in terms of treatment costs, quality of life (QoL) consequences, caregiver burden and the effects of comorbidities such as falls and fractures (Esquenazi 2011). Thus, the condition not only considerably impacts on the person and their family members but society as a whole due to the increased demand for health care.

Spasticity and the upper motor neurone syndrome

The primary feature of spasticity is hyperexcitability of muscle stretch reflexes (Lance 1980), although this has been debated (Pandyan 2005). Improved understanding of the complex pathophysiology and clinical interpretations of spasticity have lead to the broad definition of "a disordered sensori‐motor control, resulting from an upper motor neuron (UMN) lesion, presenting as intermittent or sustained involuntary activation of muscles" (Pandyan 2005). This definition suggests that 'spasticity' can be used as an umbrella term to describe the positive features of the UMN syndrome. Spasticity typically develops over a few weeks following stroke due to abnormal patterns of supraspinal descending drive (Gracies 2001). It often persists as a chronic neurological condition requiring long‐term management.

The interaction of spasticity with other features of the UMN syndrome is complex (Barnes 2001). The UMN syndrome is characterised by positive symptoms associated with muscle overactivity (such as spasticity, co‐contraction, associated reactions and clonus) and negative symptoms (such as weakness and fatigability).

Spasticity results in stiffness and abnormal posturing of the limb due to net imbalance of forces between agonist and antagonist muscles, affecting static joint position and dynamic limb movement (Sheean 1998). It may mask the return of selective movement in a paretic limb following stroke and, in some cases, relief of spasticity may facilitate the return of active movement. However, this cannot be guaranteed as underlying weakness will often persist, limiting the functional gains to be made (Ada 2006).

Adaptive features of the UMN syndrome may develop due to untreated spasticity and underlying muscle weakness (Ada 2006), including contractures and rheological changes of muscle, tendons and joints, further exacerbating limb positioning, movement and function. Elongating a spastic muscle may alter the viscoelastic and excitability properties thereby reducing muscle tone (Gracies 2001). However, the exact mechanisms remain uncertain. Stretching and positioning are used to maintain muscle length and prevent contractures in order to facilitate acquisition of normal postures.

Post‐stroke spasticity may be focal (affecting a localised part of a limb) or multifocal (affecting more than one part of a limb or limbs). The overall pattern of spasticity combined with other UMN symptoms and resulting problems may be variable; hence individualised, goal‐directed treatment is essential.

Spasticity is not always unwanted, as in some instances patients rely on spasticity to maintain function in an otherwise non‐functional limb. Yet, it should be treated when it interferes with activity or the ability to provide care to the stroke survivor, or causes pain or secondary complications. In addition to the broader effects of the UMN syndrome described above, other stroke‐related impairments such as neglect, fatigue and cognitive deficits can impact on function in conjunction with spasticity. Therefore, management is complex, requiring comprehensive multidisciplinary (MD) neurorehabilitation programmes. 

Impact of spasticity and the World Health Organization International Classification of Functioning, Disability and Health

Spasticity‐associated problems can be classified according to the World Health Organization (WHO) International Classification of Functioning, Disability and Health (ICF) (WHO 2001). The ICF is a framework used to describe health and disability. Understanding the impact of disease on a person at different levels facilitates the planning of individualised, goal‐directed and functionally orientated rehabilitation programmes. Goals for treatment of post‐stroke spasticity may relate to the following issues:

• 'impairments' (problems with body structures or physiological function) such as restricted joint range of movement, pain and involuntary movements, e.g. associated reactions and spasms;

• 'activity limitations' impacting on:

  •  'active' function (the execution of a functional task by the individual) such as reduced mobility, difficulty feeding or dressing limiting independence with self care, or

  •  'passive' function (provision of care to an affected limb) such as difficulty maintaining palmar hygiene or applying a splint or orthotic, and increased caregiver burden;

• 'restrictions in participation', which are problems limiting societal participation, such as engagement in work, family roles and leisure activities, or impacting on QoL.

Approaches to the management of spasticity

The two main approaches to the management of spasticity include pharmacological and physical interventions (Stevenson 2010). In clinical practice, these are often used in combination, but treatment is often piecemeal or not always provided as planned. One study reported that the majority of people with spasticity received at least one concomitant treatment following botulinum toxin (BoNT) injections, however physiotherapy and occupational therapy were not given as planned in a significant proportion of patients (Turner‐Stokes 2010a).

Oral anti‐spasmodic agents (for example baclofen) are commonly used for generalised spasticity. However, BoNT or other focal intramuscular treatments such as phenol are increasingly used for focal or multifocal post‐stroke spasticity. BoNT is a neurotoxin, derived from the bacterium Clostridium botulinum, which binds pre‐synaptically to the acetylcholine receptor at the neuromuscular junction, inhibiting release of acetylcholine. It is effective in reducing spasticity by causing temporary, focal muscle weakness, lasting three to four months (De Paiva 1999). Hence, it can provide a window of opportunity to maximise gains to be made during rehabilitation programmes. It can make it easier to stretch and lengthen muscles in order to prevent progression of contractures, and allow strengthening of antagonist muscles, which may improve selective movement control. BoNT has been shown to reduce muscle tone (Bhakta 2000; Elia 2009) and improve passive function, such as for hand hygiene and reducing caregiver burden (Bhakta 2000). However, the effect on active function remains unclear (Shaw 2011, Sheean 2001).

Recent guidelines and consensus statements on spasticity management advocate a holistic MD approach to rehabilitation to optimise the likelihood of treatment goals being achieved (Esquenazi 2010; Olver 2010; Turner‐Stokes 2009). This includes physicians trained in spasticity management and integrated allied health services to enable appropriate evaluation, goal setting, outcome measurement, treatment and follow‐up. However, recommendations are based on expert opinion rather than formal research evidence.

Trials investigating non‐pharmacological interventions for post‐stroke spasticity have focused on the use of BoNT injections with or without single treatment modalities, such as casting (Farina 2008), ankle taping (Reiter 1998) or electrical stimulation (Baricich 2008; Hesse 1998). Trials of BoNT effectiveness often allow concomitant 'routine' therapy which is not controlled for and varies amongst centres (McCrory 2009). Additionally, therapy programmes are rarely described in any detail. The optimal timing, type, duration and intensity of therapy remains poorly defined. Similarly, the relationship between the types of rehabilitation therapies provided and patient outcomes is not known. Hence, whilst rehabilitation practices may be routine, they are often based on trial‐and‐error as they are difficult to characterise and standardise (De Jong 2004). 

Description of the intervention

Following BoNT or other focal intramuscular treatment, MD rehabilitation involves the provision of a co‐ordinated programme by a specialised team of health professionals, delivered by two or more disciplines (nursing, physiotherapy, occupational therapy, orthotists and others). As described in Stiens 2002, necessary elements of a MD rehabilitation programme are:

• an individualised, patient‐centred plan formulated by the patient and rehabilitation team;

• goals derived and prioritised through a MD process, where goals are specific, measurable, achievable, realistic and timely (SMART) (Wade 2010);

• patient participation, required to achieve the goals with improvement in the patient’s personal potential as a result;

• outcomes demonstrating improvement in one or more domains of the ICF. 

Unidisciplinary therapy (for example physiotherapy only) or single‐treatment modalities (such as casting and functional electrical stimulation) used in isolation do not comply with the definition of MD rehabilitation, therefore we excluded randomised controlled trials (RCTs) or other studies exploring such treatments. 

How the intervention might work

Post‐stroke spasticity can manifest in multiple ways, with detrimental effects on function, QoL and caregiver burden (Bhakta 2000; Esquenazi 2001). MD rehabilitation utilises a co‐ordinated, goal‐directed treatment approach to address these issues, including allied health and medical input. It encompasses multiple therapeutic interventions aimed at improving patient experience at the level of impairment, activity or participation, and educating patients and caregivers in the ongoing self‐management of spasticity. BoNT or other focal intramuscular treatments provide a window of opportunity to facilitate gains to be made during rehabilitation and, in this context, they may be considered to be an adjunct to a MD rehabilitation programme rather than vice versa (Esquenazi 2010).

Why it is important to do this review

The effectiveness of MD rehabilitation in neurological conditions such as multiple sclerosis (Khan 2007), acquired brain injury (Turner‐Stokes 2005) and stroke (Langhorne 2011) has been proven. However, its effectiveness in managing post‐stroke spasticity following BoNT or other focal intramuscular treatment has not been determined.

The general consensus is that optimal treatment of post‐stroke spasticity following BoNT requires a comprehensive MD rehabilitation programme (Olver 2010; Sheean 2010). However, these programmes are not always implemented and delivered in practice.

Other Cochrane reviews are focused on the effectiveness of BoNT injections (Lyons 2008) or individual physical interventions (Monaghan 2011) for spasticity management in stroke survivors. To date, no systematic review has evaluated the evidence for the effectiveness of MD rehabilitation following BoNT or other focal intramuscular treatment for post‐stroke spasticity. The optimal intensity, type, and setting of rehabilitation programmes and the effects on outcomes are unclear.

Objectives

To assess the effectiveness of MD rehabilitation, following BoNT and other focal intramuscular treatments such as phenol, in improving activity limitations and other outcomes in adults and children with post‐stroke spasticity. Specific questions addressed by this review were as follows.

• Does co‐ordinated MD rehabilitation achieve better outcomes than the absence of such services in persons with post‐stroke spasticity?

• What types of rehabilitation programmes are effective, and in which setting?

• Does a greater intensity (time, expertise, or both) of rehabilitation lead to better outcomes?

Methods

Criteria for considering studies for this review

Types of studies

We included RCTs that assessed the effectiveness of MD rehabilitation programmes following BoNT or other focal intramuscular treatment for upper limb or lower limb post‐stroke spasticity, or both, with either routinely available local services or lower levels of intervention; or studies that compared MD rehabilitation programmes in different settings, of different types or at different intensities.  

Types of participants

All children (less than 18 years) and adults (18 years or over) with a confirmed diagnosis of stroke who had upper or lower limb spasticity, or both. A diagnosis of stroke fulfils the clinical criteria of the World Health Organization (WHO) of 'rapidly developed clinical signs of focal (or global) disturbances of cerebral function, lasting more than 24 hours or leading to death, with no other apparent cause than of vascular origin' (WHO 1989), with or without confirmation by a computed tomography (CT) scan or magnetic resonance imaging (MRI). A diagnosis of stroke encompassed ischaemic or haemorrhagic stroke (including subarachnoid, intraventricular or intracerebral haemorrhage).

We did not include studies of participants with conditions other than stroke unless stroke‐specific data were provided separately or more than 75% of participants had a diagnosis of stroke. Where the proportion of the study population with stroke was < 75%, we would contact study authors for data relating to stroke participants only.

Types of interventions

In this review, we defined MD rehabilitation as any co‐ordinated therapy programme delivered by two or more disciplines (such as occupational therapy, physiotherapy, exercise physiology, orthotics, other allied health and nursing) in conjunction with medical input (neurologist or rehabilitation medicine physician) and aiming to achieve patient‐centred goals related to optimising activity and participation as defined by the ICF (WHO 2001). BoNT injections were administered using individualised or standardised injection protocols in both the control and intervention groups.

Rehabilitation programmes may be delivered in:

• outpatient or day treatment settings, which may be located within private or public hospitals, community rehabilitation centres or specialist rehabilitation centres;

• home‐based settings, in the patients' own homes and local community;

• inpatient rehabilitation settings where care is delivered 24 hours per day, including specialised medical rehabilitation units or hospital wards.

Rehabilitation programmes are individualised, thus the therapy provided can be variable and the actual content of MD care may vary from patient to patient. Therefore, we included any study that stated or implied MD care or rehabilitation provided it satisfied the definition, as stated above, and compared it with a type of control situation.

Control situations were:

• no treatment;

• placebo or sham;

• other interventions, including a lower level or different types of intervention such as 'routinely available local services' (e.g. medical care or physiotherapy only), 'minimal intervention' (such as 'information only'), waiting list conditions, or interventions given in different settings and at lower intensity of intervention.

We excluded RCTs that assessed the effectiveness of a unidisciplinary therapy (for example physiotherapy only) or a single intervention (for example stretching, casting, electrical stimulation or splinting only).

Types of outcome measures

We expected diverse outcomes given the varied presentations of spasticity‐related problems and goals of treatment related to stroke severity.

Primary outcomes

Primary outcomes reflected the level of activity limitation according to the ICF (WHO 2001) and included:

• passive function (e.g. Leeds Arm Spasticity Impact Scale (Bhakta 1996); Disability Assessment Scale (Brashear 2002); Arm Activity measure (Turner‐Stokes 2010b));

• active function of the upper limb (e.g. Motor Activity Log (MAL) (Van der Lee 2004); or Action Research Arm Test (ARAT) (Lyle 1981));

• active function of the lower limb with such mobility measures including tests of walking speed, balance and gait pattern (e.g. Timed Up And Go (TUAG) (Podsiadlo 1991); 10 m walk test (Green 2002)).

We included the measure of achievement of intended goals for treatment, for example goal attainment scaling (Kiresuk 1968) or other measure of goal achievement. The assimilation of individualised goals for treatment is increasingly used as an overall measure of outcome for trials of MD rehabilitation.

Secondary outcomes

Secondary outcomes were measures of:

• symptoms and impairments, e.g. pain (measured by verbal scores or visual analogue scales, etc), spasm frequency, joint range of movement, involuntary movements, measures of spasticity or tone such as the Modified Ashworth Scale (MAS) (Bohannon 1987) or Tardieu scale (Mehrholz 2005);

• restriction in participation and impact on caregivers, e.g. QoL measures (such as WHOQoL‐BREF (Murphy 2000)), reduction of caregiver strain and burden (e.g. Caregiver Strain Index (Robinson 1983)). 

We considered adverse events that may have resulted from the intervention. We defined serious adverse effects as events that were life‐threatening or required prolonged hospitalisation. 

Timing of outcome measures

We divided outcome time points into short term (up to three months) and long term (greater than three months).

Search methods for identification of studies

See the 'Specialized Register' section in the Cochrane Stroke Group module. We searched for trials in all languages.

Electronic searches

The Cochrane Stroke Group Trials Register was last searched by the Managing Editor on 8 February 2012. In addition, we searched the following electronic bibliographic databases:

• Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2011, Issue 12) (Appendix 1);

• MEDLINE (1948 to December 2011) (Appendix 2);

• EMBASE (1980 to January 2012) (Appendix 3);

• CINAHL (1982 to January 2012) (Appendix 4);

• AMED (1985 to January 2012) (Appendix 5);

• LILACS (Latin American & Caribbean Health Sciences Literature) at http://search.bvsalud.org/regional/index.php (1982 to September 2012) (Appendix 6);

• Physiotherapy Evidence Database (PEDro) at www.pedro.org.au/ (September 2012) (Appendix 7);

• REHABDATA at www.naric.com/research/rehab/ (September 2012) (Appendix 8).

The MEDLINE search strategy was developed with the help of the Cochrane Stroke Group Trials Search Co‐ordinator and adapted for the other databases.

We also searched the following ongoing trials registers:

• ClinicalTrials.gov (www.clinicaltrials.gov/);

• EU Clinical Trials Register (www.clinicaltrialsregister.eu);

• Stroke Trials Registry (www.strokecenter.org/trials/);

• Current Controlled Trials (www.controlled‐trials.com);

• WHO International Clinical Trials Registry Platform (http://apps.who.int/trialsearch/);

• Australian New Zealand Clinical Trials Registry (www.anzctr.org.au).

Searching other resources

In an attempt to identify further published, unpublished and ongoing trials we:

• searched the reference lists of all retrieved articles, texts and other reviews on the topic;

• searched journals related to spasticity research and treatment not already searched on behalf of The Cochrane Collaboration. Handsearching of relevant journals included Archives of Physical Medicine and Rehabilitation (December 1995 to January 2012) using the search engine ScienceDirect, and Journal of Rehabilitation Medicine (January 2001 to January 2012);

• used the PubMed related articles feature;

• used the Science Citation Index Cited Reference Search for forward tracking of important articles;

• searched Open Grey (formerly SIGLE) (System for Information on Grey Literature in Europe) at www.opengrey.eu/;

• contacted authors, researchers and experts in the field.

Data collection and analysis

Selection of studies

Two review authors (MD, SM) independently screened all titles and abstracts of records identified from the searches of the electronic databases and excluded obviously irrelevant studies. We obtained the full texts of the remaining articles and assessed them for inclusion and appropriateness based on the previously defined inclusion criteria for eligibility.

Once we had obtained all potentially appropriate studies, two review authors (MD, FK) independently evaluated each study for inclusion. Where they could not obtain a consensus about the possible inclusion or exclusion of any individual study, they made a final consensus decision by discussion with a third author (LTS). We did not mask studies regarding the name(s) of the author(s), institution(s) or publication source at any level of this review.

Data extraction and management

Review authors independently extracted the data (MD, LTS, CB) from each study that meet the inclusion criteria. We have individually summarised all studies that met the inclusion criteria in RevMan 5.1 (RevMan 2011) to include the following information:

• publication details;

• study design, study setting, inclusion and exclusion criteria, method of allocation, risk of bias;

• patient population, e.g. age, sex, type of stroke, site of spasticity;

• details of interventions;

• outcome measures;

• withdrawals, length and method of follow‐up, and number of participants followed up.

Where insufficient data and methodological details were available, we contacted study authors to obtain further information and clarification (Lai 2009; Sun 2010; Weber 2010).

Assessment of risk of bias in included studies

Three authors (MD, LTS, CB) independently assessed the methodological quality of the included studies during data extraction, using the Cochrane 'Risk of bias' tool according to the Cochrane Handbook for Systematic Reviews of Interventions Chapter 8.5 (Higgins 2011). We assessed the following domains: random sequence generation and allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessors (detection bias), attrition bias, reporting bias, and other sources of bias. We scored each domain as 'low risk of bias', 'high risk of bias', or 'unclear risk of bias'.

We considered studies to be of high methodological quality ('high quality studies') if the risk of bias for domain items was low. We rated studies to be of low methodological quality ('low quality studies') if there was unclear or high risk of bias for one or more domains (Table 1). The reviews authors resolved any disagreements or lack of consensus through discussion with a fourth review author (FK).

1. Levels of quality of individual studies.

Judgement of risk of bias
Risk of bias of all domains low	High methodological quality = 'high ‐quality study'
Unclear or high risk of bias for one or more domains	Low methodological quality = 'low‐quality study'
High risk of bias for most domains	Very low methodological quality = 'very low‐quality study'

Measures of treatment effect

It was not possible to pool data for quantitative analysis due to the small number of included studies, with significant heterogeneity in type and delivery of interventions, outcome measures, study designs (assessment time points and duration of follow‐up) and participant characteristics. Therefore, we presented a qualitative synthesis of the 'best evidence' based on the GRADE levels of evidence (Table 2), taking into consideration the five factors that impact on quality of evidence (Table 3) according to the Cochrane Handbook for Systematic Reviews of Interventions Chapter 12.2 (Higgins 2011). We presented the results of individual studies in Table 4.

2. Levels of quality of a body of evidence in the GRADE approach.

Underlying methodology	Quality rating
Randomised trials or double‐upgraded observational studies	High
Downgraded randomised trials or upgraded observational studies	Moderate
Double‐downgraded randomised trials or observational studies	Low
Triple‐downgraded randomised trials or downgraded observational studies or case series/case reports	Very low

3. Factors that may decrease the quality level of a body of evidence.

1. Limitations in the design and implementation of available studies suggesting high likelihood of bias

2. Indirectness of evidence (indirect population, intervention, control, outcomes)

3. Unexplained heterogeneity or inconsistency of results (including problems with subgroup analyses)

4. Imprecision of results (wide confidence intervals)

5. High probability of publication bias

4. Description of results of included studies.

Lai 2009
Participants	N = 36 (N = 30 in analysis)
Summary of Results	Greater improvement in active range of movement elbow extension at 14 weeks in intervention compared with control group. MAS scores for elbow flexors showed comparable changes of 9.3% and 8.6% in intervention and control groups respectively Statistical analysis: trends and mean changes in outcomes were assessed using Excel tables
Results in favour of the intervention	Active range of movement mean % change; 33.5 (29.6)# versus 18.7 (48.7)# in control group #Mean (SD)
Authors' conclusions	This study confirmed the efficacy of BoNT in tone management and occupational therapy in contracture reduction, and showed the value of dynamic splinting in maintaining gains in range of motion
Sun 2010
Participants	N = 32 (N = 29 in analysis)
Summary of Results	MAS improved in both groups at 4 weeks and 3 months post injection, with no between‐group differences. Median MAS score change was ‐2 in all cases, except at the elbow in the control group at 3 months where it was ‐1.5. At 6 months there was a significant reduction in MAS at elbow (P value = 0.004), wrist (P value = 0.003), and fingers (P value < 0.001) in the intervention group (median MAS score change of ‐1). Benefit persisted only in the wrist flexors in the control group (P value = 0.014) Intervention group had statistically significant improvements compared with control group at: 6 months in MAS (elbow (P value = 0.019), wrist (P value =0.019) and fingers (P value < 0.001)); and 3 and 6 months on the ARAT (P value = 0.012 and P value < 0.001) and MAL (AOU (P value < 0.001 and P value < 0.001), QOM (P value = 0.007 and P value < 0.001)). Both groups improved on the ARAT at 4 weeks, with no between‐group differences. Patient satisfaction was high in the intervention group at 3 and 6 months and declined in both groups at 6 months Statistical analysis: Mann‐Whitney U tests, Chi2 tests or Fisher exact tests for baseline comparisons, Wilcoxon signed rank tests for within‐group change from baseline, Mann‐Whitney U test for between‐group comparisons, P value < 0.05 were statistically significant
Results in favour of the intervention	MAS at 6 months; greater reduction in MAS score for elbow (0.7, 95% CI 0.1 to 1.3; P value = 0.19), wrist (0.7, 95% CI 0.2 to 1.2; P value = 0.19), and fingers (1.2, 95% CI 0.9 to 1.5. P value < 0.001), median MAS score change = ‐1 MAL AOU at 3 months (1.1 ± 0.5 versus 0.1 ± 0.2; P value < 0.001) and 6 months (1.2 ± 0.5 versus 0.1 ± 0.2; P value < 0.001) and QOM at 3 months (0.9 ± 0.6 versus 0.3 ± 0.2; P value < 0.007) and 6 months (1.0 ± 0.5 versus 0.1 ± 0.1; P value < 0.001) ARAT at 3 months (7.3 ± 5.0 versus 3.1 ± 2.6; P value = 0.012) and 6 months (7.9 ± 5.2 versus 1.2 ± 1.7; P value < 0.001) Patient satisfaction at 3 and 6 months (93.3% and 86.7% versus 78.6% and 64.3%)
Authors' conclusions	Combined BoNT and mCIMT produced significantly greater improvements in spasticity and upper extremity motor function than BoNT and conventional rehabilitation in chronic stroke patients with upper extremity spasticity, with benefits lasting up to 6 months. The combined therapy resulted in high patient satisfaction with no serious adverse events
Weber 2010
Participants	N = 23 (all included in analysis)
Summary of Results	No significant differences between intervention and control groups for any outcome variable over time. Positive but insignificant trends toward improvement for MAL‐O, MAL‐SR and ARAT for both groups from baseline to week 6, and MAL‐SR for intervention group from 6 to 12 weeks. Control group had an insignificant trend toward deterioration in MAL‐SR from 6 to 12 weeks The entire cohort had significant improvements in MAL‐O (0.41 (0.24 to 0.58)), MAL‐SR (0.64 (0.32 to 0.95)) and ARAT (6.39 (3.38 to 9.4)) from baseline to week 6. Improvements in MAL‐SR (0.65 (0.33 to 0.97)) and ARAT (6.17 (2.31 to 10.03)) were sustained to week 12. Results were mean difference (95% CI)
Results in favour of the intervention	Nil
Author’s conclusions	BoNT and task practice therapy combined are effective in improving upper limb motor function and reducing spasticity in patients with chronic spastic hemiparesis. However, the cyclic FES protocol used in this study did not increase the gains achieved

AOU: amount of use
 ARAT: Action Research Arm Test
 BoNT: botulinum toxin
 FES: functional electrical stimulation
 MAL: Motor Activity Log
 MAL‐O: Motor Activity Log‐Observation
 MAL‐SR: Motor Activity Log‐Self Report
 MAS: Modified Ashworth Scale
 mCIMT: modified constraint‐induced movement therapy
 QOM: quality of movement
 SD: standard deviation

Unit of analysis issues

We analysed the methodological quality of included RCTs using GRADE, as described in Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Dealing with missing data

We contacted study authors to obtain missing data. Weber 2010 did not report outcomes for the Motor Assessment Scale (MAS). On enquiry, the authors stated that "the journal requested that the MAS not be reported", but they did not provide data for this measure.

None of the included studies had fatal flaws (withdrawals of more than 40% of patients, total or nearly total non‐adherence to the protocol, or very poor or non‐adjusted comparability in the baseline criteria).

Assessment of heterogeneity

We assessed clinical heterogeneity by examining the characteristics of the studies, the similarity between the types of participants, the interventions and the outcomes, as specified in the criteria for included studies. We reported trial strengths and limitations.

We had planned to quantify statistical heterogeneity between studies using the I² statistic, where an I² greater than 50% indicates substantial inconsistency, as described in the Cochrane Handbook for Systematic Reviews of interventions (Higgins 2011). However, it was not possible to pool data and so no assessment of heterogeneity was required.

Assessment of reporting biases

We minimised publication bias by sourcing unpublished data where possible. Where data were not reported in full for certain outcomes, we contacted study authors to request the full data set or the reason for not publishing the data (Weber 2010).

Data synthesis

Data synthesis was not possible.

Subgroup analysis and investigation of heterogeneity

Due to the small number of included studies it was not possible to perform a subgroup analysis for:

• type of stroke and location;

• site of spasticity, i.e. upper or lower limb, or both;

• age (children, adults less than 65 years of age versus 65 years of age or older);

• type of rehabilitation programme (e.g. outpatient, home‐based);

• intensity of treatment (high intensity, low intensity); and

• time of treatment following stroke (acute, less than six weeks; subacute, six weeks to six months; and chronic, more than six months).

Factors considered in heterogeneity included: setting, type and intensity of MD rehabilitation programmes.

Sensitivity analysis

Due to the small number of included studies it was not possible to perform a sensitivity analysis to determine whether the overall results would be the same if we analysed studies above different methodological cut‐off points.

Results

Description of studies

See Characteristics of included studies; Characteristics of excluded studies.

Results of the search

Electronic and manual searches yielded a total of 877 titles and abstracts after removing duplicates. Of these, we selected 33 for closer scrutiny resulting in three studies being included based on the review criteria. We identified two potentially relevant ongoing trials (Graham 2009; Lannin 2011). See Figure 1.

1.

Study flow diagram.

The main reasons for exclusion were:

• RCTs where BoNT versus placebo was compared, with both groups receiving the same therapy programme;

• RCTs where the intervention being compared was not consistent with the definition of MD rehabilitation (see Description of the intervention) e.g. unidisciplinary or single modality treatment only.

Included studies

See Characteristics of included studies.

All three included RCTs investigated MD rehabilitation programmes in the ambulatory setting, in adults with chronic stroke following BoNT for upper limb spasticity (Lai 2009; Sun 2010; Weber 2010). They were single‐centre trials, conducted in the USA (Lai 2009; Weber 2010) and Taiwan (Sun 2010), with small sample sizes and underpowered.

Participants

There were a total of 91 adults, with 82 included in the analyses. Weber 2010 included people with stroke as well as traumatic brain injury (TBI), with 5/13 of the control group comprising TBI participants (22% of the study population). We included the study as the intervention group consisted only of stroke patients. The five TBI patients in the control group and compared to the intervention group were matched for cognitive deficits and other neurological impairments.

Inclusion criteria for the studies varied (see Characteristics of included studies) and included chronic stroke (greater than six months (Lai 2009; Sun 2010) or one year (Weber 2010) following event) and moderate or severe upper limb spasticity at specified joints (MAS ≥ 2 (Lai 2009; Weber 2010) or MAS ≥ 3 (Sun 2010)). Two studies included participants with voluntary upper limb motor activity (Sun 2010; Weber 2010), using different criteria for inclusion. Sun 2010 based inclusion on minimal motor criteria (10 ° active extension at metacarpophalyngeal and interphalyngeal joints and 20 ° at the wrist). Weber 2010 used the Chedoke McMaster Assessment of hand impairment (Gowland 1993) and the ability to do at least one of three upper limb tasks. Lai 2009 included participants with range of movement deficits greater than 24% in elbow extension, with no criteria reflecting whether they had functional or non‐functional upper limbs.

Rehabilitation interventions

All included trials compared outpatient MD rehabilitation programmes in the intervention group with an active control situation, as described in Types of interventions. The types of rehabilitation programmes were diverse, based on various therapy approaches and combinations of physical modalities. Lai 2009 included 36 participants (30 completers) who had occupational, manual therapy combined with dynamic elbow splinting (Elbow Extension Dynasplint® (EED)) or occupational therapy alone. Sun 2010 included 32 participants (29 completers) comparing modified constraint induced movement therapy (mCIMT) (high intensity training of the affected upper limb whilst restraining the non‐affected upper limb) with neurodevelopmental therapy. Weber 2010 included 23 participants (18 completers) and compared task practice therapy, incorporating occupational therapy sessions and a home exercise programme, with cyclic functional electrical stimulation (FES) to facilitate grasp and release versus task practice therapy only.

Therapy protocols, for the control and intervention groups, differed in all three studies. Variables included: programme duration; amount and frequency of therapist contact; time spent doing a home programme or other activity outside of the programme; and intensity (defined as additional total time (hours) spent doing the rehabilitation programme (therapy sessions plus home exercises or intervention) when comparing intervention to control group (Peurala 2011)). Two studies investigated high versus lower intensity programmes, with intervention groups receiving a total of more than 670 hours (Lai 2009) and 490 hours (Sun 2010) compared with controls. All participants in Lai 2009 received two hours of occupational therapy weekly for 16 weeks. The intervention group additionally had education in using the EED (worn six to eight hours daily during sleep) and fortnightly visits to adjust the device. In Sun 2010 all participants had one hour of occupational therapy and one hour of physiotherapy thrice weekly for three months. The mCIMT group had a higher intensity of upper limb training during restraining of the non‐affected limb for at least five hours per day. In Weber 2010 the control group received six one‐hour sessions and the intervention group received seven one‐hour sessions of occupational therapy over 12 weeks. All participants were required to do a one‐hour daily task practice home exercise programme, during which the intervention group also wore the cyclic FES device. The frequency of therapy sessions was high (more than two sessions per week) in Sun 2010 and lower (less than or equal to two sessions per week) in Lai 2009 and Weber 2010.

Although the studies provided generic descriptions of the therapy programmes, there was little detail or systematic analysis (conceptualising, measuring, counting) of interventions (De Jong 2004), such as quantification of time spent on specific therapeutic modalities (Lai 2009; Sun 2010). Whilst the home exercise programme (task‐specific practice) was clearly described in Weber 2010, interventions and treatments administered during the occupational therapy sessions were not mentioned. See Characteristics of included studies for descriptions of the interventions.

Standardised protocols for upper limb BoNT injections (muscles, doses) were used in two studies (Lai 2009; Sun 2010) and one study used individualised protocols (Weber 2010). Two studies mentioned use of electromyography for needle localisation (Sun 2010; Weber 2010).

Outcome measures

A measure of activity limitation, the primary outcome for this review, was used in two studies that is active upper limb function measured by the MAL and ARAT (Sun 2010; Weber 2010). Although both studies used the MAL, the versions and administration varied. The MAL, quality of upper limb use during activities of daily living, was assessed observationally rather than self‐reported in Weber 2010. Whereas the MAL was the primary outcome measure in Weber 2010, an impairment measure (MAS) was the primary outcome in Sun 2010. Lai 2009 used impairment‐based outcome measures only that is mean per cent change in active range of movement and MAS score at the elbow. Sun 2010 was the only study to assess patient satisfaction. None of the studies considered the impact on caregivers nor used a measure of passive function or goal achievement for example GAS (which have been shown to be more sensitive measures of outcome in this context) (Turner‐Stokes 2010c). Different outcome time points were used across the studies: Lai 2009 14 weeks; Sun 2010 one, three and six months; and Weber 2010 six and 12 weeks.

None of the included studies determined the minimal clinically important difference for the primary outcome measure. Power calculations were only provided in Sun 2010. Lai 2009 stated that "The population of patients who completed this study (n = 30) was not adequate to power statistical analysis of variance". Sun 2010 required 16 participants in each group to power the study but only had 14 and 15 participants in the control and intervention groups respectively. The authors of Weber 2010 stated that "the group sizes in the present study were small (n = 13 and n = 10 in the no‐FES and FES groups, respectively). A larger sample size may have been more effective in detecting group differences between the FES and no‐FES treatments".

Excluded studies

We excluded 28 studies (see Characteristics of excluded studies) for the following reasons:

• the intervention was not consistent with the definition of a MD rehabilitation programme (n = 16);

• the comparator was BoNT versus placebo with both groups receiving the same rehabilitation programmes (n = 12).

Risk of bias in included studies

See Characteristics of included studies for risk of bias assessment. Figure 2 shows a risk of bias summary of the review authors' judgements about each risk of bias item for each included study and Figure 3 shows the risk of bias graph of review authors' judgements about each risk of bias item presented as percentages across all included studies. All three studies had a high risk of bias and were graded as 'low quality' based on the criteria (Table 1).

2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Allocation

As random sequence generation and allocation concealment were often not described, we sought clarification from study authors. Randomisation methods were adequate in all studies (Lai 2009; Sun 2010; Weber 2010). However, allocation concealment was not provided by Weber 2010 and Lai 2009 as the treating therapist and physician, respectively, had knowledge of the upcoming assignment.

Blinding

In all three included studies there was a high risk of performance bias (participants and therapists were unblinded) and low risk of detection bias (outcome assessors blinded as clarified by authors). However, outcome assessors were not questioned to assess whether blinding was maintained throughout the follow‐up period. In Sun 2010 the therapy sessions were conducted at different times in an attempt to reduce the risk of performance bias. The injecting physician was blinded in Weber 2010 only.

Incomplete outcome data

Attrition rates were moderate to high: 16.7% (Lai 2009), 21.7% (Weber 2010), but only 9.4% in Sun 2010. Reasons for dropouts were: decline in health status (n = 2), personal factors (n = 2), or non‐attendance at follow‐up (n = 1) (Weber 2010); or inability to attend therapy sessions due to practical reasons (n = 2) and a car accident (n = 1) (Sun 2010). Six participants were excluded in Lai 2009 due to "non‐compliance with scheduled therapy sessions", which was not defined, and the time points of participant withdrawal were unclear. Only Weber 2010 performed intention‐to‐treat (ITT) analysis and compared baseline differences between completers and dropouts, with no significant differences reported. It is unknown whether dropouts or exclusions would have significantly impacted on the outcomes in Lai 2009 and Sun 2010.

Selective reporting

All study outcomes were reported, except the MAS at follow‐up in Weber 2010.

Other potential sources of bias

All studies were single‐centre, with small sample sizes, and were reported to be underpowered. The sample size calculation was described in Sun 2010 only, with 16 participants required in each group to achieve 90% power to detect differences between the groups of MAS ≥ 1 with standard deviation (SD) ≤ 0.9 and accounting for 5% to 10% dropouts. The authors’ conclusions must be considered in the light of the studies being underpowered and having other high risks of bias.

Lai 2009 did not report the baseline data for time since stroke and other stroke characteristics (such as type and location). In Weber 2010 the baseline data between the groups were not entirely comparable. The control group were significantly younger (P = 0.03) and had 39% of participants with a diagnosis of TBI (versus 0% in the intervention group) as stratified randomisation by aetiology was not performed. Subgroup analysis for TBI versus stroke was not provided in the control group. The authors rationalised that the differences would not influence findings as there were similar baseline variables, cognitive function and programme adherence for TBI and stroke participants, and age was controlled for in the analysis.

Weber 2010 reported no significant differences between the two groups for any outcome variable at any time point. Hence, the authors proceeded to present collapsed data for the entire cohort to show the benefit of task practice training for improved upper limb function. However, the confounder in this instance was the unknown influence of BoNT versus task practice therapy on outcomes, as there was no BoNT‐only group as a comparator.

One study had a long‐term follow‐up period of six months (three months after completing the rehabilitation programme) (Sun 2010). The other studies had shorter follow‐up periods of 12 weeks (at the time of completion of the rehabilitation programme) in Weber 2010 and 14 weeks (two weeks prior to completion of the therapy programme) in Lai 2009, so longer‐term impacts could not be ascertained.

The results in Lai 2009 had large standard deviations indicating that the data included a wide range of values (see below).

Effects of interventions

See Table 4 for the description of the results of the included studies.

A meta‐analysis was not possible due to the clinical, methodological, and statistical heterogeneity of the included studies.

In the three studies, the mean age ranged from 49.1 years (Lai 2009) to 58.7 years (Sun 2010) in the intervention groups and 41.2 years (Weber 2010) to 61.5 years (Sun 2010) in the control groups. The lower control group mean age in Weber 2010 may have been due to inclusion of TBI participants, however this was unclear. The study populations had chronic stroke, with the mean duration following the cerebral event being 2.9 years in both groups in Sun 2010 and 4.3 and 9.7 years in the control and intervention groups respectively in Weber 2010; this was unreported in Lai 2009. Reporting of stroke characteristics varied (Sun 2010; Weber 2010). Stroke characteristics were unreported in Lai 2009. Baseline ARAT scores for the intervention groups were 32.1 ± 12.7 and 19.5 ± 13.9, and for the control groups they were 29.0 ± 14.1 and 25.8 ± 15.5 in Sun 2010 and Weber 2010 respectively. Results are the mean ± standard deviation (SD).

Lai 2009 reported that occupational therapy in conjunction with dynamic elbow splinting (EED) resulted in a greater improvement in active range of movement elbow extension compared with the control group (occupational therapy only) at 14 weeks (33.5% (29.6%) versus 18.7% (48.7%) respectively; mean (SD)). The MAS (elbow flexors) score change was comparable in both groups. There were no measures of activity level (active or passive upper limb function), goal attainment, or patient or caregiver perspective. The authors concluded that BoNT is effective with tone management and occupational therapy in contracture reduction, and there is 'value of dynamic splinting in maintaining gains in range of motion'. 

BoNT injection followed by mCIMT was shown to improve spasticity and upper limb motor function in chronic stroke patients with residual voluntary upper limb motor activity, more than BoNT followed by a neurodevelopmental therapy programme, with benefits maintained up to six months (Sun 2010). Both groups had significant improvements in the MAS at four weeks and three months without between‐group differences. However, the mCIMT group showed significantly greater improvements in elbow, wrist, and finger spasticity (P = 0.019, P = 0.019, and P < 0.001, respectively) at six months. Compared with the control group, the mCIMT group had greater improvements on the: MAL (amount of use (AOU) and quality of movement (QOM)) at three months (1.1 ± 0.5 versus 0.1 ± 0.2; P < 0.001 and 0.9 ± 0.6 versus 0.3 ± 0.2; P < 0.007) and six months (1.2 ± 0.5 versus 0.1 ± 0.2; P < 0.001 and 1.0 ± 0.5 versus 0.1 ± 0.1; P < 0.001), and ARAT at three months (7.3 ± 5.0 versus 3.1 ± 2.6; P = 0.012) and six months (7.9 ± 5.2 versus 1.2 ± 1.7; P < 0.001). Patient satisfaction was high following treatment with BoNT and mCIMT up to three months (93.3% versus 78.6% in the control group) but was not sustained beyond this (86.7% versus 64.3% in the control group at six months).

Following BoNT, cyclic FES in addition to task practice therapy did not improve upper limb motor function and spasticity more than task practice therapy only, up to 12 weeks (Weber 2010). When examining outcomes for the entire cohort, there were significant improvements in upper limb activity: MAL‐Observation 0.41 (95% CI 0.24 to 0.58), MAL‐Self Report 0.64 (95% CI 0.32 to 0.95), ARAT 6.39 (95% CI 3.38 to 9.4) from baseline to week six, MAL‐Self Report 0.65 (95% CI 0.33 to 0.97), and ARAT 6.17 (95% CI 2.31 to 10.03) to week 12. Results are mean difference (95% CI).

No significant adverse events were reported in any of the studies.

All three studies were of 'low' methodological quality (high risk of bias). Therefore, using GRADE methodology for best evidence synthesis (Table 2) there was:

• 'low quality' evidence that high intensity training of the affected limb with mCIMT, compared with a lower intensity neurodevelopmental therapy programme, improved spasticity or tone (MAS) and active upper limb function (ARAT and MAL) and achieved high satisfaction in persons with residual motor function, with benefits maintained up to six months (Sun 2010);

• 'very low quality’ evidence that a higher intensity programme of occupational therapy with additional EED assisted in maintaining active range of movement at the elbow in the short‐term, compared with occupational therapy only (Lai 2009);

• no evidence that task practice therapy with cyclic FES was superior to task practice therapy only in improving spasticity or tone (MAS) and upper limb motor function (MAL‐Observation, ARAT, and MAL‐Self Report) in people with residual motor function, at 12 weeks (Weber 2010).

Discussion

Summary of main results

This review evaluated the effectiveness of MD rehabilitation following BoNT, or other focal intramuscular treatment, in improving the activity level (primary outcome) and other outcomes (symptoms and impairments, participation restriction, caregiver burden and QoL) in adults and children with post‐stroke spasticity. 

For chronic stroke survivors treated with BoNT for upper limb spasticity, there is: low quality evidence that mCIMT improves upper limb function in the long‐term. There is very low quality evidence that occupational therapy with additional dynamic elbow splinting (EED) improves elbow tone and contractures. There is no evidence that task practice therapy with cyclic FES provides additional short‐term benefit compared with task practice therapy alone. The quality of the existing evidence is reduced by the methodological weakness of the included studies, which were underpowered and had a high risk of bias.

There was no available evidence for benefits of MD rehabilitation programmes compared with the absence of such services in different settings (inpatient), for lower limb spasticity, in children with stroke, or after focal intramuscular treatments for spasticity other than BoNT. Although results for intensity of therapy were presented wherever possible, it was difficult to categorise studies based on this criterion only. Subgroup analysis for type of stroke and location, and acute versus chronic stroke was not possible. There was no evidence for the benefit of MD rehabilitation interventions following BoNT for post‐stroke spasticity on passive function, community participation, goal achievement, caregiver burden and QoL.

Overall completeness and applicability of evidence

Despite some evidence for MD rehabilitation interventions for upper limb spasticity in adults with chronic stroke, many aspects of MD care for people with post‐stroke spasticity remain unproven.

It was not possible to determine the evidence for MD interventions for lower limb spasticity in this review as available studies included single treatment modalities only, such as taping, casting, or electrical stimulation (Baricich 2008; Hesse 1998; Johnson 2004). Evidence for unidisciplinary interventions is presented elsewhere (Monaghan 2011). There were no identified studies that included MD interventions in children or those with an acute stroke.

The emerging evidence for the effectiveness of BoNT in stroke is in improving passive function and achieving individual goals (Turner‐Stokes 2010c). However, none of the included studies addressed these particular outcomes. Two of the studies (Sun 2010; Weber 2010) addressed active function in the minority subset of patients with residual upper limb function, for which this is a realistic goal.

The optimal intensity, duration and frequency of therapy that should be provided are unclear. Although high intensity mCIMT was considered superior to a lower intensity neurodevelopmental therapy programme for those with voluntary upper limb activity (Sun 2010), similar benefits could be achieved with less therapist contact time if patients are motivated. Whilst other studies have shown benefits of CIMT in stroke survivors without significant spasticity, the programmes varied in terms of duration and frequency of therapy (Peurala 2011). Further studies are needed to determine optimal protocols for CIMT in stroke survivors with spasticity.

Applying the evidence in clinical practice is challenging. Firstly, evidence for ideal patient selection criteria for spasticity management is not available. Stroke survivors with upper limb spasticity have varied clinical presentations, whereas the study populations were restricted by the strict inclusion criteria for example the degree of residual upper limb motor activity and chronicity of stroke. This may limit the relevance of the evidence to an individual patient and, conversely, the evidence cannot be generalised to the heterogeneous stroke population. Furthermore, it is challenging to replicate the study therapy protocols in real‐life due to the variability of available resources, need for specialised equipment for example the elbow extension Dynasplint® device, therapist time and expertise, consistency of therapy programmes (intensity, duration, modalities) and other confounders (therapist‐patient interaction).

Performing high quality RCTs when investigating complex interventions such as MD spasticity management is challenging in the real world, for the reasons described below. Whilst high quality RCTs are needed, other study designs, such as the use of clinical practice improvement methodology, may assist in building on the current evidence. Data collected prospectively and retrospectively in the routine clinical setting can be used to identify the relative contributions of specific interventions and therapies (the 'black box') to rehabilitation outcomes whilst accounting for contributing patient and environmental factors (De Jong 2004; Gassaway 2005). This may assist clinicians in translating and generalising study findings to clinical practice. A better understanding of the processes of care in rehabilitation for spasticity management may then improve practice of care and ultimately patient outcomes.

Despite the gaps in the literature we believe that development of integrated MD rehabilitation teams, to provide a long‐term, comprehensive continuum of care to stroke survivors affected by spasticity, is clinically important for improving outcomes. This may be facilitated by training clinicians in setting individualised SMART goals, appropriate evaluation, and assessment of meaningful outcomes for patients and their caregivers and that assess impact at the levels of the ICF, for example using GAS. Use of item banks of standardised goals for spasticity management may be helpful where there is less expertise in goal setting (Tennant 2007).

Quality of the evidence

The quality of the evidence was limited as there were only three heterogeneous, single‐centre trials of low methodological quality, prohibiting pooling of data for quantitative meta‐analysis.

The limitations affecting the quality of the evidence in this review include:

• inconsistent terminology and definitions for 'spasticity', given that it is only one element of the UMN syndrome. In practice, rehabilitation interventions target various patterns of muscle overactivity, weakness, and contractures. When investigating 'spasticity' there is ambiguity in what is being treated and measured, causing difficulty in the interpretation and application of evidence (Bakheit 2011; Malhotra 2009). Sun 2010 and Weber 2010 did not define spasticity, and Lai 2009 used multiple terms such as 'hypertonicity', 'tone', and 'spasticity' with interventions aiming to reduce 'contracture';

• inadequate allocation concealment (Lai 2009; Weber 2010);

• high risk of performance bias due to non‐blinding of treating therapists and participants (all studies). Additionally, patients may have revealed information about the intervention to outcome assessors; however this was not addressed to reduce the risk of detection bias. In rehabilitation trials blinding of participants and personnel can be particularly challenging;

• small sample sizes and underpowered studies. Recruitment from single‐centre trials with strict inclusion and exclusion criteria can be limiting;

• the influence of significant attrition rates (16.7% in Lai 2009 and 21.7% in Weber 2010), non‐compliance affecting retention of participants and follow‐up, excluding dropouts from analyses and outcome reporting;

• differences in baseline characteristics between groups (Weber 2010) including age and etiology. The authors rationalised that the differences would not influence findings as there were similar baseline variables, cognitive function and programme adherence for TBI and stroke participants, and age was controlled for in the analysis;

• difficulty controlling for personal factors, which influence patient‐therapist interaction, compliance, and delivery of therapy thus impacting on outcomes. These include patient motivation and self‐efficacy, and activity level outside of therapy programmes, which were not assessed in any of the studies;  

• selection of inappropriate outcome measures that do not show change following therapy or correlate with meaningful outcomes for patients and their caregivers. Impairment measures, e.g. MAS and range of movement, or other standardised measures such as ARAT do not necessarily translate into improved function or benefits for patients and caregivers, which may be better captured through measures of passive function or individual goal attainment.

Potential biases in the review process

The review authors accept that there may have been a degree of:

• selection bias from the literature search;

• publication bias, if trials have not been published due to small study populations and negative results, or effects of treatments have been exaggerated in published trials.

Although Weber 2010 included participants with TBI in the control group, without subanalysis for stroke versus TBI, excluding this study would have resulted in bias in the review as 78% of the cohort had a diagnosis of stroke.

Agreements and disagreements with other studies or reviews

This review found insufficient evidence for the optimal type and intensity of MD rehabilitation programmes following BoNT for upper and lower limb spasticity, consistent with international consensus statements (Olver 2010; Sheean 2010). Thus, recommendations advocating integrated MD rehabilitation programmes following focal spasticity management are based on expert opinion only.

The evidence for CIMT after BoNT for post‐stroke upper limb spasticity, in improving active upper limb function (Sun 2010), is also supported by a cohort study (Levy 2007). This study showed benefits of two weeks of CIMT compared with a home exercise programme in patients at more than 90 days following stroke. However, unlike Sun 2010, the benefits diminished by 24 weeks. Other studies have shown that CIMT improves arm motor function in stroke survivors with benefits persisting up to one year (Wolf 2006), so longer follow‐up is warranted. However, evidence for improvement in active upper limb function after BoNT is variable, being supported by a few studies (Rousseaux 2002; Slawek 2005) but not by others (Elia 2009; Shaw 2011; Sheean 2001).

Authors' conclusions

Implications for practice.

There was limited evidence found in this review, which is matched by a lack of detail in recent spasticity management guidelines for recommending optimal multidisciplinary (MD) therapy programmes after BoNT. Hence clinical practices are based on a level of professional judgement and expertise with trial and error to determine which therapies are effective in an individual. Further trials are required to facilitate evidence‐based practice.

Implications for research.

Robust clinical trials are required to investigate the optimal timing, types (combinations of therapy approaches and modalities) and intensities (frequency, amount and duration of therapy) of MD rehabilitation programmes that improve activity and participation after BoNT and other focal intramuscular treatments for post‐stroke spasticity.

Future trials may consider:

• better methodological quality including larger sample sizes and multicentre designs with internationally agreed data sets;

• longer follow‐up (beyond six months) to determine the benefits of an intervention, as there may be a time lag between improvement in function and change in spasticity (Francis 2004), and whether the effects are maintained;

• development of appropriate outcome measures or consensus on a bank of measures that assess the 'activity (passive and active function) and participation' and 'environmental factors' domains of the ICF, which translate into real‐world functional abilities rather than focusing on impairments only. Other areas include addressing personal factors in the ICF that influence outcomes of rehabilitation interventions, and using appropriate patient‐centred outcomes with standardised measures to provide a more holistic picture;

• evaluation of the perspectives of patients and caregivers, and cost‐effectiveness of rehabilitation programmes;

• investigating optimal and effective mCIMT protocols;

• investigating the contribution of individual components of rehabilitation programmes to outcomes, individually and combined, in order to explore the 'black box' of rehabilitation.

Appendices

Appendix 1. Cochrane Central Register of Controlled Trials (CENTRAL)

#1.MeSH descriptor Cerebrovascular Disorders explode all trees
 #2.stroke or poststroke or post‐stroke or cerebrovasc* or brain vasc* or cerebral vasc* or cva* or apoplex* or SAH
 #3.brain* or cerebr* or cerebell* or intracran* or intracerebral
 #4.ischemi* or ischaemi* or infarct* or thrombo* or emboli* or occlus*
 #5.(#3 AND #4)
 #6.brain* or cerebr* or cerebell* or intracerebral or intracranial or subarachnoid
 #7.haemorrhage* or hemorrhage* or haematoma* or hematoma* or bleed*
 #8.(#6 AND #7)
 #9.hemipleg* or hemipar* or paresis or paretic
 #10.MeSH descriptor Hemiplegia, this term only
 #11.MeSH descriptor Paresis explode all trees
 #12.(#1 OR #2 OR #5 OR #8 OR #9 OR #10 OR #11)
 #13.MeSH descriptor Muscle Spasticity, this term only
 #14.MeSH descriptor Muscle Hypertonia, this term only
 #15.MeSH descriptor Muscle Rigidity, this term only
 #16.MeSH descriptor Muscle Tonus, this term only
 #17.MeSH descriptor Spasm, this term only
 #18.MeSH descriptor Dystonia, this term only
 #19.MeSH descriptor Paraparesis, Spastic, this term only
 #20.spastic* or high tone
 #21.muscle*
 #22.spasm or spasms or rigid* or tone or tonus or hyperton* or hypermyoton* or dyston* or contracture*
 #23.(#21 AND #22)
 #24.(#13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #23)
 #25.MeSH descriptor Botulinum Toxins explode all trees
 #26.MeSH descriptor Phenols explode all trees
 #27.MeSH descriptor Ethanol, this term only
 #28.MeSH descriptor Injections, Intramuscular, this term only
 #29.MeSH descriptor Nerve Block, this term only
 #30.botulinum* or botulin or botox or BoNT* or BTX* or dysport or xeomin or myobloc or neurobloc or oculinum or onabotulinum* or abobotulinum* or incobotulinum* or rimabotulinum*
 #31.phenol or ethanol or alcohol or nerve block* or motor point block*
 #32.intramuscular
 #33.injection* or treatment* or medication* or neurolysis
 #34.(#32 AND #33)
 #35.(#25 OR #26 OR #27 OR #28 OR #29 OR #30 OR #31 OR #34)
 #36.(#12 AND #24 AND #35)

Appendix 2. MEDLINE (Ovid)

1. cerebrovascular disorders/ or exp basal ganglia cerebrovascular disease/ or exp brain ischemia/ or exp carotid artery diseases/ or exp intracranial arterial diseases/ or exp "intracranial embolism and thrombosis"/ or exp intracranial hemorrhages/ or stroke/ or exp brain infarction/ or vertebral artery dissection/
 2. (stroke or poststroke or post‐stroke or cerebrovasc$ or brain vasc$ or cerebral vasc$ or cva$ or apoplex$ or SAH).tw.
 3. ((brain$ or cerebr$ or cerebell$ or intracran$ or intracerebral) adj5 (isch?emi$ or infarct$ or thrombo$ or emboli$ or occlus$)).tw.
 4. ((brain$ or cerebr$ or cerebell$ or intracerebral or intracranial or subarachnoid) adj5 (haemorrhage$ or hemorrhage$ or haematoma$ or hematoma$ or bleed$)).tw.
 5. hemiplegia/ or exp paresis/
 6. (hemipleg$ or hemipar$ or paresis or paretic).tw.
 7. 1 or 2 or 3 or 4 or 5 or 6
 8. muscle spasticity/ or muscle hypertonia/ or muscle rigidity/ or muscle tonus/
 9. spasm/ or dystonia/ or paraparesis, spastic/
 10. (spastic$ or high tone).tw.
 11. (muscle$ adj5 (spasm or spasms or rigid$ or tone or tonus or hyperton$ or hypermyoton$ or dyston$ or contracture$)).tw.
 12. 8 or 9 or 10 or 11
 13. exp botulinum toxins/ or exp phenols/ or ethanol/ or injections, intramuscular/ or nerve block/
 14. (botulinum$ or botulin or botox or BoNT$ or BTX$ or dysport or xeomin or myobloc or neurobloc or oculinum or onabotulinum$ or abobotulinum$ or incobotulinum$ or rimabotulinum$).tw.
 15. (phenol or ethanol or alcohol or nerve block$ or motor point block$).tw.
 16. (intramuscular adj3 (injection$ or treatment$ or medication$ or neurolysis)).tw.
 17. 13 or 14 or 15 or 16
 18. 7 and 12 and 17
 19. cerebral palsy/ or cerebral palsy.tw.
 20. 18 not 19
 21. Randomized Controlled Trials as Topic/
 22. random allocation/
 23. Controlled Clinical Trials as Topic/
 24. control groups/
 25. clinical trials as topic/
 26. double‐blind method/
 27. single‐blind method/
 28. Placebos/
 29. placebo effect/
 30. Multicenter Studies as Topic/
 31. Therapies, Investigational/
 32. Research Design/
 33. Program Evaluation/
 34. evaluation studies as topic/
 35. randomized controlled trial.pt.
 36. controlled clinical trial.pt.
 37. clinical trial.pt.
 38. multicenter study.pt.
 39. (evaluation studies or comparative study).pt.
 40. random$.tw.
 41. (controlled adj5 (trial$ or stud$)).tw.
 42. (clinical$ adj5 trial$).tw.
 43. ((control or treatment or experiment$ or intervention) adj5 (group$ or subject$ or patient$)).tw.
 44. (quasi‐random$ or quasi random$ or pseudo‐random$ or pseudo random$).tw.
 45. ((multicenter or multicentre or therapeutic) adj5 (trial$ or stud$)).tw.
 46. ((control or experiment$ or conservative) adj5 (treatment or therapy or procedure or manage$)).tw.
 47. ((singl$ or doubl$ or tripl$ or trebl$) adj5 (blind$ or mask$)).tw.
 48. versus.tw.
 49. placebo$.tw.
 50. sham.tw.
 51. (assign$ or alternate or allocat$).tw.
 52. controls.tw.
 53. or/21‐52
 54. 20 and 53
 55. exp animals/ not humans.sh.
 56. 54 not 55

Appendix 3. EMBASE (Ovid)

1. cerebrovascular disease/ or basal ganglion hemorrhage/ or cerebral artery disease/ or cerebrovascular accident/ or stroke/ or exp carotid artery disease/ or exp brain hematoma/ or exp brain hemorrhage/ or exp brain infarction/ or exp brain ischemia/ or exp intracranial aneurysm/ or exp occlusive cerebrovascular disease/
 2. stroke unit/ or stroke patient/
 3. (stroke or poststroke or post‐stroke or cerebrovasc$ or brain vasc$ or cerebral vasc$ or cva$ or apoplex$ or SAH).tw.
 4. ((brain$ or cerebr$ or cerebell$ or intracran$ or intracerebral) adj5 (isch?emi$ or infarct$ or thrombo$ or emboli$ or occlus$)).tw.
 5. ((brain$ or cerebr$ or cerebell$ or intracerebral or intracranial or subarachnoid) adj5 (haemorrhage$ or hemorrhage$ or haematoma$ or hematoma$ or bleed$)).tw.
 6. hemiplegia/ or paresis/
 7. (hemipleg$ or hemipar$ or paresis or paretic).tw.
 8. 1 or 2 or 3 or 4 or 5 or 6 or 7
 9. muscle hypertonia/ or muscle rigidity/ or spastic paresis/ or spasticity/
 10. muscle spasm/ or muscle tone/ or dystonia/ or paraplegia/
 11. (spastic$ or high tone).tw.
 12. (muscle$ adj5 (spasm or spasms or rigid$ or tone or tonus or hyperton$ or hypermyoton$ or dyston$)).tw.
 13. 9 or 10 or 11 or 12
 14. botulinum toxin/ or botulinum toxin a/ or botulinum toxin b/ or botulinum toxin e/ or botulinum toxin f/
 15. phenol/ or alcohol/ or intramuscular drug administration/ or exp nerve block/
 16. (botulinum$ or botulin or botox or BoNT$ or BTX$ or dysport or xeomin or myobloc or neurobloc or oculinum or onabotulinum$ or abobotulinum$ or incobotulinum$ or rimabotulinum$).tw.
 17. (phenol or ethanol or alcohol or nerve block$ or motor point block$).tw.
 18. (intramuscular adj3 (injection$ or treatment$ or medication$ or neurolysis)).tw.
 19. 14 or 15 or 16 or 17 or 18
 20. 8 and 13 and 19
 21. cerebral palsy/ or cerebral palsy.tw.
 22. 20 not 21
 23. Randomized Controlled Trial/
 24. Randomization/
 25. Controlled Study/
 26. control group/
 27. clinical trial/
 28. Double Blind Procedure/
 29. Single Blind Procedure/ or triple blind procedure/
 30. placebo/
 31. Multicenter Study/
 32. experimental design/ or experimental study/ or quasi experimental study/
 33. experimental therapy/
 34. evaluation/ or "evaluation and follow up"/ or evaluation research/ or clinical evaluation/
 35. "types of study"/
 36. Comparative Study/
 37. random$.tw.
 38. (controlled adj5 (trial$ or stud$)).tw.
 39. (clinical$ adj5 trial$).tw.
 40. ((control or treatment or experiment$ or intervention) adj5 (group$ or subject$ or patient$)).tw.
 41. (quasi‐random$ or quasi random$ or pseudo‐random$ or pseudo random$).tw.
 42. ((multicenter or multicentre or therapeutic) adj5 (trial$ or stud$)).tw.
 43. ((control or experiment$ or conservative) adj5 (treatment or therapy or procedure or manage$)).tw.
 44. ((singl$ or doubl$ or tripl$ or trebl$) adj5 (blind$ or mask$)).tw.
 45. versus.tw.
 46. placebo$.tw.
 47. sham.tw.
 48. (assign$ or alternate or allocat$).tw.
 49. controls.tw.
 50. or/23‐49
 51. 22 and 50

Appendix 4. CINAHL (EBSCO)

S37 .S33 not S36
 S36 .S34 or S35
 S35 .TI cerebral palsy OR AB cerebral palsy
 S34 .(MH "Cerebral Palsy")
 S33 .S22 and S32
 S32 .S23 or S24 or S25 or S26 or S27 or S28 or S29 or S30 or S31
 S31 .AB intramuscular AND AB ( injection* or treatment* or medication* or neurolysis )
 S30 .TI intramuscular AND TI ( injection* or treatment* or medication* or neurolysis )
 S29 .TI ( phenol or ethanol or alcohol or nerve block* or motor point block* ) OR AB ( phenol or ethanol or alcohol or nerve block* or motor point block* )
 S28 .TI ( botulinum* or botulin or botox or BoNT* or BTX* or dysport or xeomin or myobloc or neurobloc or oculinum or onabotulinum* or abobotulinum* or incobotulinum* or rimabotulinum* ) OR AB ( botulinum* or botulin or botox or BoNT* or BTX* or dysport or xeomin or myobloc or neurobloc or oculinum or onabotulinum* or abobotulinum* or incobotulinum* or rimabotulinum* )
 S27 .(MH "Nerve Block")
 S26 .(MH "Injections, Intramuscular+")
 S25 .(MH "Ethanol")
 S24 .(MH "Phenols")
 S23 .(MH "Botulinum Toxins")
 S22 .S12 and S21
 S21 .S13 or S14 or S15 or S16 or S17 or S20
 S20 .S18 and S19
 S19 .TI ( spasm or spasms or rigid* or tone or tonus or hyperton* or hypermyoton* or dyston* ) or AB ( spasm or spasms or rigid* or tone or tonus or hyperton* or hypermyoton* or dyston* )
 S18 .TI muscle* or AB muscle*
 S17 .TI ( spastic* or high tone ) or AB ( spastic* or high tone )
 S16 .(MH "gait disorders, neurologic")
 S15 .(MH "dystonia")
 S14 .(MH "spasm")
 S13 .(MH "Muscle Spasticity") OR (MH "Muscle Hypertonia") OR (MH "Muscle Tonus")
 S12 .S1 or S2 or S5 or S8 or S9 or S10 or S11
 S11 .(MH "Brain Injuries")
 S10 .TI ( hemipleg* or hemipar* or paresis or paretic ) or AB ( hemipleg* or hemipar* or paresis or paretic )
 S9 .(MH "Hemiplegia")
 S8 .S6 and S7
 S7 .TI ( haemorrhage* or hemorrhage* or haematoma* or hematoma* or bleed* ) or AB ( haemorrhage* or hemorrhage* or haematoma* or hematoma* or bleed* )
 S6 .TI ( brain* or cerebr* or cerebell* or intracerebral or intracranial or subarachnoid ) or AB ( brain* or cerebr* or cerebell* or intracerebral or intracranial or subarachnoid )
 S5 .S3 and S4
 S4 .TI ( ischemi* or ischaemi* or infarct* or thrombo* or emboli* or occlus* ) or AB ( ischemi* or ischaemi* or infarct* or thrombo* or emboli* or occlus* )
 S3 .TI ( brain* or cerebr* or cerebell* or intracran* or intracerebral ) or AB ( brain* or cerebr* or cerebell* or intracran* or intracerebral )
 S2 .TI ( stroke or poststroke or post‐stroke or cerebrovasc* or brain vasc* or cerebral vasc or cva or apoplex or SAH ) or AB ( stroke or poststroke or post‐stroke or cerebrovasc* or brain vasc* or cerebral vasc or cva or apoplex or SAH )
 S1 .(MH "Cerebrovascular Disorders+") or (MH "stroke patients") or (MH "stroke units")

Appendix 5. AMED (Ovid)

1. cerebrovascular disorders/ or cerebral hemorrhage/ or cerebral infarction/ or cerebral ischemia/ or cerebrovascular accident/ or stroke/
 2. (stroke or poststroke or post‐stroke or cerebrovasc$ or brain vasc$ or cerebral vasc$ or cva$ or apoplex$ or SAH).tw.
 3. ((brain$ or cerebr$ or cerebell$ or intracran$ or intracerebral) adj5 (isch?emi$ or infarct$ or thrombo$ or emboli$ or occlus$)).tw.
 4. ((brain$ or cerebr$ or cerebell$ or intracerebral or intracranial or subarachnoid) adj5 (haemorrhage$ or hemorrhage$ or haematoma$ or hematoma$ or bleed$)).tw.
 5. (hemipleg$ or hemipar$ or paresis or paretic).tw.
 6. hemiplegia/
 7. 1 or 2 or 3 or 4 or 5 or 6
 8. exp muscle spasticity/ or muscle hypertonia/ or muscle tonus/
 9. spasm/ or dystonia/
 10. (spastic$ or high tone).tw.
 11. (muscle$ adj5 (spasm or spasms or rigid$ or tone or tonus or hyperton$ or hypermyoton$ or dyston$ or contracture$)).tw.
 12. 8 or 9 or 10 or 11
 13. botulinum toxins/ or phenols/ or alcohol ethyl/ or nerve block/
 14. (botulinum$ or botulin or botox or BoNT$ or BTX$ or dysport or xeomin or myobloc or neurobloc or oculinum or onabotulinum$ or abobotulinum$ or incobotulinum$ or rimabotulinum$).tw.
 15. (phenol or ethanol or alcohol or nerve block$ or motor point block$).tw.
 16. (intramuscular adj3 (injection$ or treatment$ or medication$ or neurolysis)).tw.
 17. 13 or 14 or 15 or 16
 18. 7 and 12 and 17

Appendix 6. LILACS

(MH:"cerebrovascular disorders" or MH:"basal ganglia cerebrovascular disease" or MH:"brain ischemia" or MH:"carotid artery diseases" or MH:"intracranial arterial diseases" or MH:"intracranial embolism and thrombosis" or MH: "intracranial hemorrhages" or MH:"stroke" or MH:"brain infarction" or MH:"vertebral artery dissection" or TW:"stroke" or TW:"poststroke" or TW:"post‐stroke" or TW:cerebrovasc$ or TW:brain vasc$ or TW:cerebral vasc$ or TW:cva$ or TW:apoplex$ or TW:"SAH" or TW:cerebr$ or TW:cerebell$ or TW:intracran$ or TW:intracerebral or TW:"subarachnoid" or MH:"hemiplegia" or MH:"paresis" or TW:hemipleg$ or TW:hemipar$ or TW:paresis or TW:paretic)
 
 AND
 
 (MH:"muscle spasticity" or MH:"muscle hypertonia" or MH:"muscle rigidity" or MH:"muscle tonus" or MH:"spasm" or MH:"dystonia" or MH:"paraparesis, spastic" or TW:spastic$ or TW:"high tone" or MH:spasm$ or TW:rigid$ or TW:"tone" or TW:"tonus" or TW:hyperton$ or TW:hypermyoton$ or TW:dyston$ or TW:contracture$)
 
 AND
 
 (MH:"botulinum toxins" or MH:"phenols" or MH:"ethanol" or MH:"injections, intramuscular" or MH:"nerve block" or TW:botulinum$ or TW:"botulin" or TW:"botox" or TW:BoNT$ or TW:BTX$ or TW:"dysport" or TW:"xeomin" or TW:"myobloc" or TW:"neurobloc" or TW:"oculinum" or TW:onabotulinum$ or TW:abobotulinum$ or TW:incobotulinum$ or TW:rimabotulinum$ or TW:"phenol" or TW:"ethanol" or TW:"alcohol" or TW:nerve block$ or TW:motor point block$ or TW:intramuscular or TW:injection$ or TW:treatment$ or TW:medication$ or TW:"neurolysis")

Appendix 7. PEDro

Search completed in advanced search using the abstract and title field using the following terms:
 
 spasticity ethanol
 spasticity phenol
 spasticity rimabotulinum*
 spasticity incobotulinum*
 spasticity abobotulinum*
 spasticity onabotulinum*
 spasticity oculinum
 spasticity neurobloc
 spasticity myobloc
 spasticity xeomin
 spasticity dysport
 spasticity BTX*
 spasticity BoNT*
 spasticity botox
 spasticity botulin
 spasticity botulinum*

Appendix 8. REHABDATA

Botulinum* botulin botox BoNT* BTX* dysport xeomin myobloc neurobloc oculinum onabotulinum* abobotulinum* incobotulinum* rimabotulinum* phenol ethanol
 
 "nerve block", "motor point block", "intramuscular injection"

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Lai 2009.

Methods	Single‐blind pilot RCT
Participants	36 adults (18 to 75 years) ≥ 6 months after stroke with spasticity (MAS ≥ 2) of elbow flexors and range of movement deficit of > 24% in elbow extension 30 completed the study (6 excluded due to non‐compliance) Intervention: N = 15; age = 49.1 (4)* years; % female = 53.3% Control: N = 15; age = 55.6 (5)* years; % female = 33.3% *Mean (SD)
Interventions	All groups: BoNT (standard injections into biceps, brachialis and brachioradialis), occupational, manual therapy (2 x 1‐hour sessions weekly for 16 weeks) for moist heat, patient education and re‐evaluation of symptoms, joint mobilisation, passive and active ROM, proprio‐neural facilitation, and therapeutic exercise Intervention group: adjunctive dynamic elbow splinting with EED worn for 6 to 8 hours during sleep, and education in use of EED with change in tension prescribed twice a month
Outcomes	Impairments: mean % change in active range of movement elbow extension and MAS elbow flexors Time points: before and 14 weeks after injection
Notes	No outcomes of activity limitations (active or passive upper limb function) used Physicians monitored adherence to wearing time of EED on a monthly basis
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation schedule (correspondence from author)
Allocation concealment (selection bias)	High risk	Sealed envelopes numbered sequentially, opened by injecting physicians (correspondence from author)
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unblinded participants, therapists and injecting physicians
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinded outcome assessor (correspondence from author)
Incomplete outcome data (attrition bias) All outcomes	High risk	High attrition rate 16.7% (6/36), exclusions due to non‐compliance with occupational therapy sessions with no justification provided No ITT analysis or comparison of baseline characteristics between withdrawals and completers
Selective reporting (reporting bias)	Low risk	All outcomes reported
Other bias	High risk	Single centre Participants were volunteers and recruitment process was not clearly explained Small sample size and underpowered; sample size calculation not reported Baseline data for time following stroke and stroke characteristics not reported Rate of adherence to occupational therapy sessions and EED use not reported Inadequate follow‐up period (2 weeks prior to completion of therapy protocol) Results with large SDs

Sun 2010.

Methods	Single‐blind RCT
Participants	32 adults (18 to 80 years) ≥ 1 year after stroke with severe spasticity (MAS ≥ 3) in elbow, wrist or finger flexors and ≥10° active interphalangeal and metacarpal phalangeal extension and 20° active wrist extension (minimal motor criteria)  29 completed the study (3 dropouts) Intervention: N = 15; age = 58.7 (9.9)* years; % female = 20%; time since stroke = 2.9 (1.5) years; infarction = 80% Control: N = 14; age = 61.5 (9.4)* years; % female= 21.4%; time since stroke = 2.9 (1.3) years; infarction = 78.6% *Mean (SD)
Interventions	All groups: BoNT (1000 units Dysport®, standard injections into elbow, wrist and finger flexors). Physiotherapy and occupational therapy (2 hours, 3 times per week for 3 months) starting the day following injections Intervention group: modified CIMT with non‐affected UL restrained for ≥ 5 hours per day Therapy included massed practice, shaping (individualised task selection, graduated tasks difficulty and complexity, positive verbal feedback and physical assistance with movements), behavioural contract (activities to be done with restrain on and situations in which to remove it), and daily treatment diary. Participants encouraged to use the affected upper limb during home activities Control group: NDT focusing on normalising tone and movement patterns, proximal upper limb control, restoration of stance, gait, dexterity and stamina training exercises. 40% of therapy time focused on upper limb exercises
Outcomes	Primary outcome: impairment: MAS Secondary outcomes: activity limitation: MAL amount of use (AOU) and quality of movement (QOM) (questionnaire for patient self report), ARAT Other: patient’s global satisfaction with treatment (7‐point categorical scale from completely satisfied to completely dissatisfied) and adverse events Time points: before injection, 1, 3 and 6 months  Exceptions: MAL not assessed at 1 month; patient satisfaction recorded at 3 and 6 months, and baseline ARAT performed twice, 4 weeks apart with averaged score used
Notes	Baseline data comparable Therapy adherence rates, assessed with daily exercise diaries, were high (93% and 87% for intervention and control groups respectively)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Central web‐based randomisation; block randomisation in groups of 4
Allocation concealment (selection bias)	Low risk	Sealed opaque envelopes
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unblinded participants and therapists. Therapy sessions conducted at different times to avoid contact between participants. Unclear whether injecting physician was blinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinded outcome assessor. MAL and patient global satisfaction associated with high risk of bias as participants unblinded
Incomplete outcome data (attrition bias) All outcomes	High risk	Low attrition rate 9.4% (3/32). Reasons for dropouts given: 1 in intervention group relocated, 2 in control group due to transportation and traffic accident. No ITT analysis or comparison of baseline characteristics between withdrawals and completers
Selective reporting (reporting bias)	Low risk	All outcomes reported
Other bias	High risk	Single centre Small sample size and underpowered (16 participants required in each group to achieve 90% power to detect differences between the groups of MAS ≥ 1 with SD ≤ 0.9 and accounting for 5% to 10% dropouts)

Weber 2010.

Methods	Single‐blind pilot RCT
Participants	23 adults with ≥ 6 months unilateral spastic hemiparesis (MAS ≥ 2 for wrist or finger flexors) due to stroke or TBI and moderate‐severe hand impairment based on Chedoke‐McMaster Assessment ≥ 2 with ability to do at least 1 of the following stage‐3 tasks: active wrist extension greater than half range; active finger/wrist flexion greater than half range; or actively touch thumb to index finger when the hand was placed in supination with thumb fully extended Exclusion criteria: no voluntary motion or severe fixed joint contracture of the affected arm 18 completed the study Intervention: N = 8; age = 54.0 (10.3)* years; % female = 70%; time since cerebral event = 9.7 (8.6)* years; cortical insult = 100% Control: N = 10; age = 41.2 (14.2)* years; % female = 62%; time since cerebral event = 4.3 (2.5)* years; cortical insult = 92.3% TBI participants: 39% of control group, 0% of intervention group * Mean (SD)
Interventions	All groups: BoNT (using individualised injections). Task practice therapy: 6 x 1‐hour visits with the occupational therapist; 1‐hour home‐based daily task practice (4‐5 individualised functional tasks) for 12 weeks using a standardised protocol without constraining the unimpaired arm Intervention group: H200 FES device (Bioness®) with forearm‐wrist‐hand orthosis, worn during daily task practice activities, to activate flexor/extensor muscles producing grasp/release at an individualised intensity 1 additional hour of occupational therapy to fit and train in use of FES device Subsequent visits to check functioning and use of FES device and make adjustments
Outcomes	Primary outcome: activity limitation: How Well Scale of MAL‐Observation (MAL‐O) for observational assessment of upper limb function during activities of daily living. Investigators standardised administration and scoring of each MAL item through observation and trained the blinded assessor to interrater reliability (intra‐class correlation coefficient 0.97) with the occupational therapist researcher (ERS) Secondary outcomes: activity limitation: ARAT, MAL‐Self‐Report (MAL‐SR) Time points: baseline (2 weeks prior) and 6 and 12 weeks after injection
Notes	All TBI participants randomised to control group (5/13) Control group were significantly younger (P value = 0.03) which may be related to TBI participants Adherence to the task practice programme, assessed through home diary records and additional electronic log in the intervention group, was high in both groups
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random sequence of numbers; block randomisation in groups of 4
Allocation concealment (selection bias)	High risk	Author's response: "master list of participants' group assignment was kept by the treating therapist and used to determine whether or not to use the Bioness device."
Blinding of participants and personnel (performance bias) All outcomes	High risk	Unblinded participants and therapists. Blinded injecting physician
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinded outcome assessor for primary outcome (MAL‐O) and ARAT MAL‐SR associated with high risk of bias as participants unblinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	High attrition rate (21.7%). Provided reasons for dropouts (n = 5): 3 in control group (discontinued intervention), 2 in intervention group (discontinued intervention and lost to follow‐up) ITT analysis (n = 23) No significant baseline differences between dropouts and completers
Selective reporting (reporting bias)	Unclear risk	MAS reported at baseline but not at follow‐up. All other outcomes reported.
Other bias	High risk	Single‐centre study Convenience sample Small sample size and underpowered Groups unmatched: all TBI participants randomised to control group (5/13) which was significantly younger (P value = 0.03). Age controlled for in analyses. No subgroup analysis for TBI versus stroke participants in control group Authors' conclusions related to results for entire cohort as there were no significant outcome differences between control and intervention groups

ARAT: Action Research Arm Test
 BoNT: botulinum toxin
 CIMT: constraint‐induced movement therapy
 EED: Elbow Extension Dynasplint®
 FES: functional electrical stimulation
 ITT: intention‐to‐treat
 MAL: Motor Activity Log
 MAS: Modified Ashworth Scale
 NDT: neurodevelopmental treatment
 RCT: randomised controlled trial
 ROM: range of motion
 SD: standard deviation
 TBI: traumatic brain injury
 UL: upper limb

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Baricich 2008	Intervention not MD (electrical stimulation, stretching or taping)
Bayram 2006	Intervention not MD (short‐term electrical stimulation)
Borg 2011	Comparator BoNT versus placebo, rehabilitation programme (standard care) in both groups
Carda 2011	Intervention not MD (casting, stretching or taping)
Childers 2004	Comparator 3 doses of BoNT, splinting and physical therapy in both groups
Clark 2011	Comparator BoNT A versus placebo, upper limb rehabilitation programme in both groups
Clarke 2003	Intervention not MD (functional electrical stimulation)
Cui 2006	Comparator BoNT A + rehabilitation therapy versus rehabilitation therapy
Duarte 2011	Intervention not MD (electrical stimulation)
Farina 2008	Intervention not MD (casting)
Guo 2006	Comparator BoNT A versus placebo, rehabilitation programme in both groups
Hesse 1995	Intervention not MD (short‐term electrical stimulation)
Hesse 1998	Intervention not MD (physiotherapy only with short‐term electrical stimulation)
Johnson 2004	Intervention not MD (physiotherapy only with electrical stimulation)
Karadag‐Saygi 2010	Intervention not MD (kinesiotaping versus sham taping)
Kent 2006	Intervention not MD (physiotherapy only and resting splint/orthotic)
Lagalla 1997	Intervention not MD (forearm splint)
Marco 2007	Comparator BoNT versus placebo, inpatient rehabilitation programme and transcutaneous electrical nerve stimulation in both groups
Mawson 2007	Intervention not MD (BoNT + physiotherapy versus BoNT + stretching advice versus placebo + physiotherapy)
McCrory 2009	Comparator BoNT versus placebo, physiotherapy and occupational therapy as per 'routine practice' in both groups
Meythaler 2009	Comparator BoNT versus placebo, occupational therapy programme both groups
Pieber 2011	Intervention not MD (FES only), participants were children with mixed neurological diagnoses
Pisano 2002	Intervention not MD (physiotherapy only with short term electrical stimulation)
Reiter 1998	Intervention not MD (ankle taping)
Shaw 2011	Comparator BoNT versus placebo, upper limb therapy programme in both groups
Sheikh 2006	Comparator BoNT versus placebo, NDT programme in both groups
Sisto 2001	Intervention not MD (acupressure)
van Wijck 2006	Intervention not MD (task specific therapy programme was physiotherapy only)
Werner 2011	Comparator BoNT versus placebo, inpatient rehabilitation programme in both groups
Wolf 2007	Comparator BoNT versus placebo, rehabilitation programme (modalities, repetitive task practice, strengthening and functional activities) in both groups

BoNT: botulinum toxin
 FES: functional electrical stimulation
 MD: multidisciplinary
 NDT: neurodevelopmental treatment

Characteristics of ongoing studies [ordered by study ID]

Graham 2009.

Trial name or title	Efficacy and safety study of botulinum neurotoxin A with rehabilitation versus botulinum neurotoxin A alone in treatment of post‐stroke spasticity
Methods	RCT
Participants	Adults ≥18 years Stroke (ischaemic or haemorrhagic) > 6 months prior Upper limb focal spasticity at elbow, wrist, fingers and thumb, MAS ≥ 3 at wrist and/or fingers Functional impairment secondary to spasticity e.g. difficulty with hygiene, dressing, posture or pain
Interventions	BoNT with rehabilitation therapy for the duration of the study (for up to 2 injections of BoNT A) versus BoNT only
Outcomes	Primary outcome measure: Fugl‐Meyer upper extremity score Secondary outcome measures: length of time to meet re‐injection criteria, number of participants that do not meet re‐injection criteria prior to completion of the study
Starting date	January 2009
Contact information	Principal Investigator: Glenn D Graham, New Mexico VA Health Care System Albuquerque, New Mexico, USA
Notes	Status: recruiting

Lannin 2011.

Trial name or title	The effectiveness of best practice therapy after Botulinum Toxin A injections for adults diagnosed with neurological impairment and onset of spasticity
Methods	RCT
Participants	Adults ≥18 years Neurological injury (including stroke and brain injury) ≥ 1 month prior Spasticity in at least 1 limb
Interventions	Group A: best practice therapy provided by an occupational or physical therapist including evidence‐based protocols for casting, electrical stimulation, task‐specific movement training and home practise. 14 sessions (1 hour per session) of best practice therapy is provided over 8 weeks (2 weeks of casting (1 session per week), followed by 2 weeks of 3 x weekly therapy, then 2 weeks of 2 x weekly therapy, and finishing with 2 weeks of 1 x weekly therapy) Group B: best practise therapy plus botulinum toxin injections Group C: botulinum toxin injection only (i.e. without best practice therapy)
Outcomes	Primary outcome measure: Goal Attainment Scaling T‐score change score Secondary outcome measures: Box and Block Test change score (hand dexterity), 6 metre walk test change score, Tardieu Scale
Starting date	September 2010
Contact information	Natasha Lannin, Alfred Health Clinical School Level 4 The Alfred Centre 99 Commercial Road Prahran Victoria 3181, Australia. Email: n.lannin@latrobe.edu.au
Notes	Status: recruitment completed

BoNT: botulinum toxin
 MAS: Modified Ashworth Scale
 RCT: randomised controlled trial

Differences between protocol and review

We did not include studies involving participants with conditions other than stroke unless stroke‐specific data were provided separately or more than 75% of participants had a diagnosis of stroke. Where the proportion of the study population with stroke was < 75% we contacted study authors for data on stroke participants only.

Due to the lack of clinical homogeneity, and the variability of methods and available data between the included studies, quantitative meta‐analysis was not possible. Thus, it was not possible to use a fixed‐effect model and the I² statistic for heterogeneity to assess outcome data for compatibility with the assumption of a uniform risk ratio (P value > 0.10), perform visual inspection of forest plots, or conduct statistical analysis as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Subgroup analysis for the following subgroups was not possible:

• type of stroke and location;

• site of spasticity, i.e. upper or lower limb, or both;

• age (children, adults less than 65 years of age versus 65 years of age or older);

• type of rehabilitation programme (e.g. outpatient, home‐based);

• intensity of treatment (high intensity, low intensity); and

• time of treatment following stroke (acute, less than six weeks; subacute, six weeks to six months; and chronic, more than six months).

Contributions of authors

Marina Demetrios (MD), Lynne‐Turner Stokes (LTS) and Fary Khan (FK) were involved in writing the protocol and review, incorporating comments from the reviewers. Caroline Brand (CB) and Shane McSweeney (SM) were involved at the review stage.

Sources of support

Internal sources

• Department of Rehabilitation Medicine, Royal Melbourne Hospital, Australia.

External sources

• No sources of support supplied

Declarations of interest

The authors are therapists or clinicians in the field of Physical and Medical Rehabilitation with a clinical interest in management of spasticity.

Lynne Turner‐Stokes has received honoraria from Ipsen Ltd and from Allergan on a number of occasions over the last few years for lecturing and running training courses and workshops – particularly with respect to the use of goal attainment scaling in spasticity. She has been involved with the development of clinical guidelines and consensus statements regarding spasticity management that were sponsored variously by Ipsen Ltd and Allergan. She has received consultancy fees for advising on research projects and for data analysis, and has held research grants from Ipsen Ltd to conduct studies relating to spasticity management. She has also published papers and presented abstracts at conferences reporting trials and other studies relating to the use of botulinum toxin for spasticity. Some of her work in this area may have been eligible for inclusion in the review.

New