The long-term risk of stroke over an expected lifetime has been calculated at between one in five for females and one in six for males in a group of middle-aged adults (Seshadri 2006). Stroke causes a greater range of disabilities and a greater disability impact than any other chronic disease (Adamson 2004). One of the disabling impairments following stroke is spasticity. The burden of care is higher in stroke patients who had spasticity when compared with stroke patients without spasticity with regards to treatment costs, quality of life, caregiver burden and effects of co-morbidities (Esquenazi 2011).
Description of the condition
Spasticity occurs as a result of damage to the descending tracts of the upper motor neuron system. Generally this involves damage to the corticoreticular pathways (parapyramidal pathways), tracts originating at the brain stem or reticular formation, or the reticulospinal and vestibulospinal tracts within the spinal cord (Burke 1988). Such damage to the upper motor neuron system results in a loss of inhibition which in turn manifests as abnormal muscle activity.
Spasticity is poorly defined in the literature (Malhotra 2009). For the purposes of this review, in order to ensure that no data are missed, spasticity will be defined as "disordered sensory-motor control, resulting from an upper motor neuron lesion, presenting as intermittent or sustained involuntary activation of muscles" (Pandyan 2005).
Studies have identified that spasticity is common with prevalence being between 30% and 80% of stroke survivors (Malhotra 2008; Watkins 2002). If spasticity is not treated effectively then it may result in increased pain and stiffness, decreased range of movement and altered posturing (Ada 2006). It may also interfere with functional recovery (Sorinola 2009). However, the complexity of differentiating between neural and intrinsic stiffness makes the correlation between function and spasticity difficult to identify (Mirbagheri 2011). In spite of this, many of the studies will have used a functional measure to primarily assess the efficacy of the intervention. As the aim of rehabilitation is to facilitate and optimise the recovery of function this review will primarily focus on whether current treatments of spasticity lead to improvements in function, despite there being a lack of evidence of a causal relationship between spasticity and function.
Description of the intervention
The treatment of spasticity can be complex and relies on a multidisciplinary approach (Barnes 1998). Treatment of spasticity has traditionally taken a stepped approach whereby the procedures considered conservative are used first and progressively more invasive procedures are then employed (Bogey 2004). Physiotherapeutic approaches are deemed the most conservative and are therefore used initially. This is reinforced by consensus guidelines (Royal College of Physicians 2009; Sheean 2010). However, physiotherapeutic interventions do not have a direct effect on abnormal muscle activity (there are some unproven claims that there may be an indirect effect) resulting from spasticity but this can only be achieved by pharmacological interventions.
Pharmacological interventions for stroke spasticity are generally termed anti-spasmodics and can be divided in to two groups: those that act systemically and those that act locally (Sheean 2006), with the locally acting treatments tending to be more invasive. The stepped approach advocates that systemically acting drugs are used initially (Gormley 1997). Such drugs include baclofen, tizanidine or dantrolene. If these are not successful in alleviating the problem then more locally acting drugs are then employed (Bogey 2004), such as injections of botulinum toxin to the muscles or alcohol or phenol to the peripheral nerves (Kocabas 2010). Should none of these treatment selections be effective then surgery is the final treatment option to be used. However, this is rarely used in the stroke population.
The use of physical therapeutic methods to manage spasticity is currently being systematically reviewed (Monaghan 2011) as is the use of botulinum toxin, which is the most widely used locally acting pharmacological therapy (Lyons 2007). The use of systemically acting drugs and other locally acting drugs, however, has not been systematically reviewed. This review will, therefore, focus on these systemically and locally acting drugs excluding those technologies already under review (i.e. botulinum toxin and physiotherapeutic interventions).
How the intervention might work
Drugs act either systemically or peripherally. Systemic drugs act in a number of different ways but generally aim to inhibit neurotransmitter activity at one or more sites within the central nervous system. Sites of action include both pre and post-synaptic terminals of the spinal interneurons (at varying levels of the upper motor neuron pathway), alpha motor neurons and primary sensory afferent neurons. This inhibitory effect acts on the neurotransmitters throughout the central nervous system and can result in drowsiness amongst other side effects (Gracies 1997). Administration tends to be oral so doses are high to cross the blood brain barrier. In order to lessen these negative effects it is possible to introduce some drugs directly in to the cerebrospinal fluid via an intrathecal pump. One particular group of drugs inhibits force production at the level of the muscle; however, this effect can also be systemic.
Other drugs are given peripherally via injection directly to the nerve. This is a far more invasive procedure; however, the systemic side effects are fewer. The aim of such injections is to destroy conduction in the nerves or the motor neuron junction or to decrease muscle contraction.
An outline of the most regularly used drugs is provided. While not systemically acting, alcohol and phenol have also been included in Table 1.
Why it is important to do this review
Using antispasmodics that have systemic effects appears to remain pervasive in the clinical setting. There is evidence that global antispasmodics tend to cause drowsiness (Gracies 1997; Simpson 2009). This in turn can be detrimental to motor skill acquisition as they depress the nervous system (Sheean 2009; Willerslev-Olsen 2011). As these pharmacological therapies are started during rehabilitation where the main objective is to re-learn motor skills such drugs may not be appropriate. Therefore, there is a need to review the level of evidence on which these drugs are used.
The use of phenol and alcohol are locally acting drug treatments but their effectiveness have not been systematically reviewed in regards to post-stroke spasticity so these have been included in this systematic review.
The management of spasticity post stroke using physical therapeutic means is being systematically reviewed having been accepted as a protocol (Monaghan 2011), and a systematic review of botulinum toxin post stroke is currently being reviewed by editors (Lyons 2007). Therefore, these will not be included in this review.
It is not clear whether there is evidence to indicate that one drug is more effective than others in treating both the spasticity and prevention of secondary consequences and whether any effects translate in to functional improvements.
It is also important to identify how common adverse events are identified with the use of such drugs.
To determine if pharmacological interventions for spasticity are more effective than no intervention, normal practice or control at improving function following stroke.
- To determine if pharmacological interventions for spasticity after stroke are more effective than no intervention, normal practice or control at:
- preventing secondary complications such as pain and contractures;
- decreasing spasticity at an impairment level.
- To determine if global antispasmodic interventions are more effective than local treatments at improving function after stroke.
- To determine if early administration of pharmacological interventions for spasticity (before six months) are more effective than late administration (after six months) of pharmacological intervention at improving function after stroke.
- To determine the side effects of the use of pharmacological interventions against placebo.
- To determine whether there is a difference between using pharmacological interventions for spasticity compared with no intervention, normal practice or control at improving function of the arm or leg following stroke.
Criteria for considering studies for this review
Types of studies
We will only include randomised controlled trials (RCTs) in this review. Studies will have investigated any pharmacological therapy and not limited to the outlined drugs documented. Systemically acting drugs tend to have a long half life and will therefore carry a high risk of carry-over effect. We will only include cross-over designed studies if the results of the first period of the data are available. Some studies may combine complex combinations of pharmacological interventions with physical therapeutic interventions where there is no clear placebo or normal treatment group. In such cases it will be difficult to ascertain what part of the treatment provided the main effect and so we will not include these.
Types of participants
Participants will have had a stroke resulting in spasticity. There is evidence that spasticity is poorly defined and measured in the literature (Malhotra 2009). Excluding papers on invalid outcome measures is likely to result in very few papers to review. We therefore propose to include all trials that have explicitly stated that they aim to treat spasticity (we will obtain evidence for this by reading the appropriately selected papers to ascertain that this was the case). There are a variety of methods available to measure aspects of spasticity, some that are direct (neurophysiological methods such as H-reflex and electromyography) and others that are not (biomechanical measures such as stiffness or clinical scales such as the Ashworth scale or tone assessment scale). Most of the commonly used measures are indirect and confounded. We will exclude studies where no such outcome measures have been used.
We will include participants irrespective of gender and older than 18 years of age. Some studies may have included participants with a variety of diagnoses (e.g. traumatic brain injuries or multiple sclerosis as well as stroke). In such cases we will contact the study authors to request stroke-specific data. In the case where study authors are not able to provide stroke-specific data we will still include the studies if the proportion of participants with stroke is greater than 80%. We will not include studies where the proportion of stroke patients is less than 80%.
Types of interventions
We will include any pharmacological intervention that aims to reduce spasticity regardless of dose or mode of delivery.
The pharmacological intervention will be compared with placebo, normal practice or no intervention. We will also include trials that compare two forms of anti-spasticity intervention as long as at least one intervention is pharmacological.
In cases where a trial methodology assesses varying doses of a particular trial drug against a placebo for optimal efficacy we will combine the results of the varying doses.
Types of outcome measures
There are wide variations in the types of outcome measures used to assess the efficacy of anti-spasticity treatments. Many studies use a variety of measures that attempt to assess disability at the level of impairment, activity and participation. As elucidated already, the main aim of rehabilitation is to achieve functionally relevant activity that transfers in to usual activities of daily living. Such scales tend to be general and not dependent on specific areas of the body. However, other activity scales are more localised. We have decided that such specific outcome measures should be divided in to those assessing upper limb and those assessing the lower limb.
To assess the primary research objective the use of an outcome measure to assess functional ability during activities of daily living will need to be present. These might include: the Barthel Index, the Functional Independence Measure or Stroke Impact Scale.
We will investigate functional movement scales that measure activity at the upper limb separately to those that measure activity at the lower limb. Functional activity scales of the upper limb might include Action Research Arm Test (ARAT), appropriate sub-sections of motor assessment scale or Wolf Motor Function Test. Functional activity scales of the lower limb might include Functional Ambulatory Classification or Functional Gait Assessment.
Other measures are timed measures of specific activities. Again such temporal measures can be divided in to upper limb and lower limb measures. Upper limb temporal measures might include the 9 hole peg test or Jebsen Hand function test. Lower limb temporal measures might include the timed 10m walk or Get up and Go.
To measure secondary consequences of spasticity we will investigate specific objective measures such as pain and range of movement.
To assess the efficacy at an impairment level a measure of spasticity will need to be used. This will ideally be measured through neurophysiological methods such as electromyographic (EMG) activity or the Hoffmann (H) reflex activity. Many trials use indirect measures of stiffness such as the Modified Ashworth Scale or tone assessment scale, which are not necessarily measures of spasticity. However, as such measures are common we will also include these indirect measures.
We will identify any reported adverse events.
Search methods for identification of studies
See the 'Specialized register' section in the Cochrane Stroke Group module. We will search for trials in all languages and arrange translation of relevant papers published in languages other than English.
We will search the Cochrane Stroke Group Trials Register and the following electronic bibliographic databases and trials registers:
- The Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, latest issue);
- MEDLINE (from 1946) (Appendix 1);
- EMBASE (from 1980);
- CINAHL (from 1982);
- AMED (from 1985);
- Physiotherapy Evidence Database (PEDro) (www.pedro.org.au/);
- REHABDATA (www.naric.com/research/rehab/);
- Center for International Rehabilitation Research Information and Exchange (CIRRIE) (http://cirrie.buffalo.edu/);
- 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/).
We developed the MEDLINE search strategy with the help of the Cochrane Stroke Group Trials Search Co-ordinator and we will adapt this for the other databases.
Searching other resources
In order to identify further published, unpublished and ongoing trials, we will:
- search the reference lists from published reviews and trials identified by the above methods;
- contact relevant pharmaceutical companies;
- contact authors and researchers in the field;
- use Science Citation Index Cited Reference Search for forward tracking of important articles.
Data collection and analysis
CL and one other review author will independently screen the titles and abstracts of the records obtained from the electronic searches and exclude obviously irrelevant studies. We will then obtain the full text of the remaining studies and the same two authors will select studies that potentially meet the review inclusion criteria.
Selection of studies
CL and a further review author will independently consider all intervention studies and identify those that are RCTs. The review authors will resolve any disagreements at this time, in the first instance, through discussion and if necessary with ADP as mediator.
As recommended in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (Moher 2009), we will include a flow diagram outlining the phases for selection of studies included in the review using the PRISMA template available in the Review Manager software, RevMan 5.2 (RevMan 2012).
Data extraction and management
Having identified all randomised controlled trials, CL and another review author will then independently assess each study and extract relevant details using a data extraction form (Appendix 2) based on a modified version of the van Tulder pro-forma previously advocated for assessment of bias by the Cochrane Back Review Group (van Tulder 1997). This pro-forma is designed to assess risk of bias but also contains items that relate to quality and precision of reporting rather than internal validity.
Assessment of risk of bias in included studies
CL and another review author will independently assess each study for risk of bias using the domains of the Cochrane Collaboration's tool for assessing risk of bias in included studies (Higgins 2011). A judgement on whether there was a low risk, high risk or unclear risk of bias in two specific domains will be made first. These are 'risk of selection bias' (random sequence generation and allocation concealment) and to assess for 'performance and detection bias' (appropriate blinding procedures to blind participants, personnel and assessors). We will then investigate studies deemed to have a low risk of bias in these two domains for attrition bias and reporting bias (outcome data completeness). We will assess the completeness of data for each individual outcome. The results from both review authors will then be reviewed and discrepancies will be resolved through discussion. If consensus is not possible the discussion will be expanded to include a mediator (ADP).
Measures of treatment effect
By entering details from each study in to RevMan 2012, we will determine statistical analysis of the treatment effect. The outcomes to be studied use a variety of data. We will analyse dichotomous data as odds ratios (ORs) with 95% confidence interval (CIs). In the context of this review it is expected that pain and the presence of side effects will be presented as dichotomous data. We will also recode measures of stiffness and tone such as the Ashworth Scale into dichotomous data identifying if there is an improvement or not. This is because there are many variations on this scale system and it is not possible to compare such scales via continuous data. There is also no clinically significant cut-off point available to use for these scales.
We will analyse continuous data in two ways depending on the measures used in studies. Examples of continuous data that we will analyse in this way are the H-reflex or Barthel Index.
For separate studies where the same measure was used, we will use the mean difference (MD) with 95% CI to calculate treatment effect. For separate studies that used different scales to measure the same outcome we will employ the use of standardised mean difference (SMD) with 95% CI. An example of this may be using the Functional Independence Measure and the Barthel Index to measure functional outcome.
Because some patients will have had more than one adverse event and some patients may have had none, we will analyse adverse events as risk ratio (RR) with 95% CI.
In cases where an outcome cannot be summarised in such a way we will tabulate the results.
Unit of analysis issues
Due to the high risk of a carry-over effect in crossover trials of systemic medications, we will only use studies where the results of just the first period data is available (i.e. a baseline measure and then a further measure just before cross over).
Dealing with missing data
In the case of missing data we will contact the original investigators to request the information. It will be possible to calculate some missing data using calculations of provided data (as might be the case in missing standard deviations).
In the case that a study simply does not report a designated outcome we will not include the study in the analyses of the outcome.
Assessment of heterogeneity
It is likely that the studies will be clinically heterogeneous with variation in interventions and outcomes being observed. Such differences may be possible to spot visually using the forest plot. We will use statistical methods to assess heterogeneity in the first place. Using the Chi
This review will employ methods to quantify the inconsistency and impact on the meta-analysis using the I
Assessment of reporting biases
By using the search strategy outlined above to include and analyse all positive or negative available data in any language reporting bias will be kept to a minimum. Any unpublished trials will be subject to detailed review prior to inclusion. To assess for outcome reporting bias we will tabulate the stated outcome measures of studies and identify if the results are reported for each outcome measure as either; fully, not fully or missing. We will use a funnel plot of the standard error of the intervention effect to identify asymmetry.
The aim of this review is to establish whether there is evidence of an effect. This type of review makes a meta-analysis appropriate. It is expected that this review will identify studies that use a variety of drugs to achieve the same outcome though measured in various ways. If the pooled data from these studies demonstrate statistical homogeneity then we will employ a fixed-effect meta-analysis. However, it is more likely that we will need to use a random-effects model. If these results are divergent then we will report the most conservative outcome (Deeks 2011).
In order to investigate the primary objective, we will pool studies using continuous data. As described in the Measures of treatment effect section, for separate studies where the same measure was used, we will use MD with 95% CI to calculate treatment effect; for separate studies that used different scales to measure the same outcome we will employ SMD with 95% CI.
We will pool the timed functional tests using SMD with 95% CI.
In order to investigate the secondary objectives of spasticity (and pain and secondary consequences of spasticity), we will pool studies using dichotomous outcome measures together. We will calculate an OR using the Mantel-Haenszel statistical method with 95% CI to assess treatment effect.
We will pool studies that used EMG or H-reflex to assess spasticity and we will use MD with 95% CI to calculate treatment effect.
In order to determine if global antispasmodic interventions are more effective than local treatments at improving function after stroke, we will pool studies using the same drug to study the effect of particular drugs. We will then pool the results of drugs to combine those that are locally- or globally-acting drugs.
In order to determine the rate of side effects we will pool the RRs of each trial using the Mantel-Haenszel method with 95% CI.
Subgroup analysis and investigation of heterogeneity
If deemed appropriate, we will conduct subgroup analysis to investigate if the treatment effects on the primary outcome (functional recovery) vary in sub-populations. We will conduct subgroup analyses for the following groups.
- Time since stroke onset: we will explore the effect of beginning the intervention at varying times since stroke. The sub-groups will be those where the intervention begins within one month post stroke, within one to three months post stroke and those that begin between three to six months post stroke. We will use the mean time from stroke to intervention to identify participants that will qualify for inclusion.
- Treatments focused specifically on arm function versus specifically on leg function.
- To determine if there is a difference between the side-effects in drugs that were globally acting or locally acting.
We will undertake a sensitivity analysis to assess the difference between using RR and OR for the dichotomised outcomes.
The review authors would like to acknowledge the staff at the UK Cochrane Centre in Oxford for providing training and advice during the writing of this protocol.
Appendix 1. MEDLINE search strategy
- 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/
- (stroke or poststroke or post-stroke or cerebrovasc$ or brain vasc$ or cerebral vasc$ or cva$ or apoplex$ or SAH).tw.
- ((brain$ or cerebr$ or cerebell$ or intracran$ or intracerebral) adj5 (isch?emi$ or infarct$ or thrombo$ or emboli$ or occlus$)).tw.
- ((brain$ or cerebr$ or cerebell$ or intracerebral or intracranial or subarachnoid) adj5 (haemorrhage$ or hemorrhage$ or haematoma$ or hematoma$ or bleed$)).tw.
- hemiplegia/ or exp paresis/ or brain injuries/ or brain injury, chronic/ or brain damage, chronic/
- (hemipleg$ or hemipar$ or paresis or paretic or acquired brain injur$).tw.
- 1 or 2 or 3 or 4 or 5 or 6
- muscle spasticity/ or muscle hypertonia/ or muscle rigidity/ or muscle tonus/
- spasm/ or dystonia/ or paraparesis, spastic/
- (spastic$ or high tone).tw.
- (muscle$ adj5 (spasm or spasms or rigid$ or tone or tonus or hyperton$ or hypermyoton$ or dyston$)).tw.
- 8 or 9 or 10 or 11
- exp muscle relaxants, central/
- cannabis/ or exp cannabinoids/
- exp phenols/ or ethanol/ or injections, intramuscular/ or nerve block/ or neuromuscular blockade/ or injections, spinal/ or intrathecal.tw
- (baclofen or carisoprodol or chlormezanone or chlorphenesin or chlorzoxazone or cyclobenzaprine or dantrolene or dimethothiazine or diazepam or eperisone or gabapentin or idrocilamide or ketazolam or medazepam or mephenesin or meprobamate or methocarbamol or nefopam or orphenadrine or quinine or tetrazepam or tizanidine or tolperisone or zoxazolamine or cannabis or cannabinoid* or sativex or GW-1000 or phenol or alcohol or ethanol).tw,nm.
- (antispastic$ adj5 (agent$ or drug$ or medication)).tw.
- ((nerve or neuromuscular or motor point) adj5 block$).tw.
- (intramuscular adj3 (drug$ or injection$ or treatment$ or medication$ or neurolysis)).tw.
- (drug therapy or drug effects).fs.
- 7 and 12 and 21
- cerebral palsy/ or cerebral palsy.tw.
- 22 not 23
- exp animals/ not humans.sh.
- 24 not 25
- Randomized Controlled Trials as Topic/
- random allocation/
- Controlled Clinical Trials as Topic/
- control groups/
- clinical trials as topic/ or clinical trials, phase i as topic/ or clinical trials, phase ii as topic/ or clinical trials, phase iii as topic/ or clinical trials, phase iv as topic/
- double-blind method/
- single-blind method/
- placebo effect/
- Therapies, Investigational/
- Drug Evaluation/
- Research Design/
- randomized controlled trial.pt.
- controlled clinical trial.pt.
- (clinical trial or clinical trial phase i or clinical trial phase ii or clinical trial phase iii or clinical trial phase iv).pt.
- (random$ or RCT$).tw.
- (controlled adj5 (trial$ or stud$)).tw.
- (clinical$ adj5 trial$).tw.
- ((control or treatment or experiment$ or intervention) adj5 (group$ or subject$ or patient$)).tw.
- (quasi-random$ or quasi random$ or pseudo-random$ or pseudo random$).tw.
- ((control or experiment$ or conservative) adj5 (treatment or therapy or procedure or manage$)).tw.
- ((singl$ or doubl$ or tripl$ or trebl$) adj5 (blind$ or mask$)).tw.
- (assign$ or allocat$).tw.
- 26 and 52
Appendix 2. Data extraction form for included trials
- Study ID
- Contact details
- Title of study
- Authors of study
- Country of study
- Objective stated
- Inclusion criteria
- Exclusion criteria
- Definition of stroke
- Type of stroke
- Time from stroke to inclusion as criterion
- Definition of spasticity
- Time post stroke
- Laterality of stroke
- Type of study
- Treatment allocation/method of randomisation
- Number of interventions
- Pharmaceutical treatment
- Placebo treatment
- Dosage of trial drug
- Co-interventions (comparable/avoided)
- Durations of trial
- Outcome definition
- Primary outcome measure
- Secondary Outcome measures
- If measure is a scale include upper and lower limits and whether high or low score is positive
- Outcomes and time points collected
- Intervention effect on the outcome measures chosen for the study
- Short term assessment
- Long term assessment
- Sample size
- Withdrawal/drop-out rate
- Missing participants
- Compliance to treatment reported
- Summary data for each intervention group. 2 × 2 table for dichotomous data; means and SDs for continuous data
- Estimate of effect with confidence interval and P value
- Subgroup analyses
- Intention to treat
Contributions of authors
Professor Anand Pandyan and Cameron Lindsay conceived and designed the protocol for the review.
Declarations of interest
Both authors have previously received unconditional funding from both Allergan and Ipsen drug companies. In accordance with The Cochrane Collaboration's policy on commercial sponsorship no money was received from a commercial source to fund the development of this review.
Sources of support
- Cameron Lindsay, UK.Keele University Studentship
- No sources of support supplied