Accepted Nov 7, 2019; Epub ahead of print Nov 19, 2019

J Rehabil Med 2020; 52: jrm00005

Correspondence address: Patricia Mills, GF Strong Rehabilitation Centre, 4255 Laurel St, Vancouver, BC, V5Z 2G9, Canada. E-mail: patricia.mills@vch.ca

Spasticity is a common and problematic consequence of neurological conditions affecting the brain and spinal cord. It is characterized by intermittent or sustained involuntary muscle activation that can limit function and quality of life. Intramuscular injection with botulinum toxin is a useful treatment in such patients in order to weaken the spastic muscle. This study reviewed the published evidence for the use of casting after botulinum toxin injection for limb spasticity in adults. Casting of a limb after botulinum toxin injection shows promise in improving function in patients with spasticity, although further research is needed to determine the best
method to use.

Spasticity, a sensori-motor disorder characterized by intermittent or sustained involuntary muscle activation, is a common and potentially problematic consequence of upper motor neurone disorders (1). A recent systematic review of 17 observational studies revealed that the presence of spasticity is associated with worse health status in patients with chronic neurological conditions (2). Spasticity may also reduce quality of life and heighten economic burden (2, 3).

There are various treatment approaches for spasticity, which differ based on the pattern of increased muscle tone, patient characteristics and functional goals. Botulinum toxin (BoNT) is an effective pharmacological treatment for focal muscle over-activity in a wide range of neurological conditions (4, 5). There is high-level evidence to suggest that adjunct therapies may improve outcomes after BoNT injection (6, 7). However, at present there are no systematic reviews in the literature describing the current evidence and protocols used for casting as an adjunct therapy to BoNT injection.

In spastic muscle, the number of serial sarcomeres becomes reduced, which may contribute to more rapid development of contractures. Immobilization in the lengthened position may increase the length and number of serial sarcomeres (8). Casting with a non-removable external device theoretically provides a prolonged stretch of the muscle-tendon complex to prevent or correct soft-tissue contractures associated with spasticity. In addition to the soft-tissue changes that occur with prolonged muscle stretching, casting may also affect the neural circuits contributing to spasticity. Casts provide firm, consistent pressure distributed across the skin in a total-contact fashion. This may reduce the sensory input from cutaneous, muscle and joint receptors that contribute to spasticity (9, 10). Furthermore, prolonged stretch of the muscle-tendon complex may contribute to decreased alpha motor excitability, which could also improve spasticity (9).

Casting is an attractive treatment option, as it can be done with relatively inexpensive and accessible materials, such as plaster, it does not require a significant amount of time for application, and can be applied either once or several times (serial casting). Potential downsides include soft-tissue injury and pain, requiring prompt removal of the cast. Since the cast cannot be applied or removed by the patient, frequent appointments may be required depending on the number of cast changes. This can be particularly inconvenient for patients with mobility impairments. Casting protocols vary widely between clinicians and it is not yet understood which approach yields the best outcomes. The systematic review by Mills et al. demonstrated that, at the time of publication, there were a limited number of randomized controlled trials (RCT) using casting following injections (6). It highlighted the need to explore this topic in greater depth, including understanding what treatment protocols have been used across all studies (not just RCTs) and with what degree of success.

The primary objective of this study is to synthesize the current evidence for casting in adults as an adjunct treatment following BoNT injection for limb spasticity resulting from various neurological conditions. The secondary objective of this study is to present casting protocols and outcome measures used in the literature for the purpose of informing future research in this area.

Systematic review

This review was prepared and reported with reference to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (11). Details of the protocol for this systematic review were registered on PROSPERO (CRD42017073098).

A systematic search strategy, developed by the College of Physicians and Surgeons of British Columbia librarians, was conducted to identify relevant studies published between 1990 and August 2018 using electronic databases MEDLINE, EMBASE, CINAHL and Cochrane Central Register of Controlled Trials. Reference lists of systematic reviews and included studies were reviewed to expand the dataset. The search was restricted to the English language. The search strategy was based on 3 key concepts: spasticity, BoNT, and casting. An example of the search strategy, as applied in MEDLINE, is shown in Appendix SI¹.

Eligibility criteria

For this review, the study inclusion criteria were: (i) full-text studies of any design using a casting protocol following BoNT injection for spasticity management on any proportion of the participants; (ii) adult participants over the age of 18 years with limb spasticity from a neurological condition; (iii) clinical outcomes including spasticity, range of motion, pain, function and adverse events; and (iv) English language studies. As per the Lannin et al. systematic review of upper extremity casting in patients with central nervous system motor disorders, casts were defined as any non-removable, external device made from plaster or casting tape applied with the intention of modifying the structural or functional characteristics of the neuromuscular system (12).

Study selection

Two reviewers independently reviewed titles and abstracts of studies to determine eligibility for inclusion. Disagreement was resolved through consensus and, if necessary, by third-party resolution. Studies that clearly failed to meet the inclusion criteria were not reviewed further. Those that could not be excluded were retrieved and reviewed in full-text by the 2 reviewers. In all instances, differences of opinion were resolved by discussion. Studies that met criteria were retrieved and reviewed in detail.

Data collection

Data were extracted independently from all included studies and in duplicate into Excel spreadsheets, with the templates adapted from the Cochrane Collaboration (13). Data included a description of participants, intervention protocols, outcome measures and results. When BoNT injection or casting protocol information was missing or not available, the authors were contacted for additional information. Contact was made via email at 2 time-points separated by 2 weeks; contacted authors had 2 weeks to respond to each email, for a total of 4 weeks of response time following the first email contact. If no casting protocol information was available following paper abstraction or email contact, the study was excluded. If at least partial casting protocol information was available, the study was included. It was determined a priori by the study authors that casting protocols would be considered completely reported (100%) if the study methods described: timing of casting application post BoNT injection, casting material, position of patient during casting, casting angle, total duration of casting, frequency of cast change. Time that it took to cast per casting session was also documented when available, but did not contribute to the percentage of reporting calculation.

Risk of bias assessment

Two reviewers independently assessed the methodological quality of the included studies. Differences in scores were resolved by a third party. Quality assessment was performed using the Physiotherapy Evidence Database (PEDro) scoring system for RCTs and a modified Downs and Black tool for non-RCTs (14–16). The PEDro scale is composed of 11 yes or no quality items, 10 of which are used to calculate the final PEDro score (0–10). The modified Downs and Black scale consists of 27 1-point questions and 1 2-point question, resulting in a final score ranging from 0 to 28 (16). For both tools, higher scores are indicative of greater methodological quality. To simplify interpretation of results, studies scoring 9 or 10 on the PEDro scale or ≥ 24 on the Downs and Black scale are considered methodologically to be of excellent quality; scores of 6–8 on the PEDro scale or 20–23 on the Downs and Black scale are considered of good quality; scores of 4–5 on the PEDro scale or 15–19 on the Downs and Black scale are of fair quality; and scores < 4 on the PEDro scale or ≤ 14 on the Downs and Black scale are considered of poor quality (16). The level of evidence for study results was evaluated using a 5-level scale (simplified form of Sackett), where level 1 (the highest level of evidence) is an RCT with a PEDro score ≥ 6; level 2 is an RCT with a PEDro score ≤ 5, a non-randomized prospective-controlled study, or a cohort study; level 3 is a case-control study; level 4 is a pre- and post-test or a case series; and level 5 is an observational report or case report with only a single subject (14–16).

Outcome measures

Outcome measures from the studies included in this review were listed and classified according to the International Classification of Functioning, Disability and Health (ICF) domains. The ICF was published by the World Health Organization (WHO) in May 2001 to provide a common international language for describing health and disability in clinical and research settings (17). The ICF framework classifies function into 3 domains: body structure and function, activity, and participation. Outcome measures that did not fit within a domain were classified as “other”.

Statistical analysis

Because of the paucity of studies for the primary objective and the differences in outcome measured used, a formal meta-analysis was not feasible. Therefore, the results of this review are presented in a narrative form. Description of pre-casting interventions, casting protocols and study results are shown in Table I. Description of outcome measures subcategorized by the ICF domains are presented in Table II. Levels of evidence are summarized in Table III.

Search strategy

Fig. 1 shows the flow of study selection. A total of 108 studies were identified through the electronic database search. After the removal of duplicate studies and screening of articles based on title and abstract, 10 studies were included for full review. Three of these studies did not meet the inclusion criteria. Two of these studies did not include casting protocol details and were excluded after telephone or email follow-up (23, 24).

Risk of bias assessment

The PEDro and Downs and Black scores for all included studies are recorded in Table I. PEDro values ranged between 7 and 8; Downs and Black scores ranged between 5 and 17.

Population

Characteristics of included studies are summarized in Table I. The 5 studies included a total of 98 participants. The number of participants ranged from 1 to 42 per study.

Casting protocols

Casting protocols varied widely between studies; all were on casting of the lower limb. Of the 5 studies included for full review, all had some degree of missing data. Completeness of reporting of casting methodology ranged from as low as 50% (3/6) to as high as 83% (5/6). Thus, all authors were contacted for missing data, with 3 of the 5 authors responding with information via email. Information acquired from authors is denoted in italics in Table I.

Outcome measures

A total of 10 distinctive outcome measures were used within the studies (see Table II). In all, 6 measurements were classified in the ICF Body Structure and Function domain and 4 in the Activity domain. None of the outcome measurements were classified in the Participation domain, and none of the studies used an outcome measurement for quality of life. The most commonly used outcome measure was ankle passive range of motion, which was used in all studies.

Study results and levels of evidence

Study results are summarized in Table I. There were no RCTs that compared BoNT alone with BoNT and casting. In general, use of casting and BoNT improved spasticity outcomes compared with baseline status. Carda et al. (18) compared 3 adjunct therapies (casting, taping and stretching) after BoNT injection. Verplancke et al. also compared 3 groups: casting post BoNT injection, casting post saline injection, physical therapy only. In Singer et al., 5 of the 10 participants were enrolled in an RCT to examine the effect of serial casting plus BoNT vs placebo injection. The other 5 received casting only. Results were combined (pre-post study methodology) as the RCT did not show between group differences. Adverse events as a result of casting were reported in 4 studies, providing an adverse event profile in 58 participants (18–21). A total of 16 adverse events were reported for the 58 participants, all of which were related to soft-tissue injury. Only one treatment discontinuation due to pain as a result of casting was reported. Levels of evidence are summarized in Table III.

This is the first systematic review that has been conducted with an in-depth look at casting as an adjunct to BoNT injections for limb spasticity in the adult population. The included studies provide important insights into the use of casting for treatment of lower limb spasticity that can be applied to clinical practice and guide future research in this area.

It is important to note that currently there are no studies published that address whether casting improves spasticity outcomes when used as an adjunct compared with BoNT injection alone; the 3 prospective controlled studies on casting and BoNT lacked a control group with BoNT only, which is needed in order to answer this clinical question. It is also notable that there were no studies identified on casting of the upper limb. It would be worthwhile for clinicians to report their experiences with the upper limb, even as case reports, as this would be a valuable contribution to the literature.

Currently, for the lower limb, there is level 1b evidence that casting as an adjunct is more effective than stretching and taping, and that casting after either BoNT or saline injections is better than physical therapy alone. Interestingly, with respect to adverse events captured by Verplanke et al. (19), casting with BoNT injection resulted in less severe soft-tissue injury (mainly skin discolouration) compared with casting with saline injection (mainly partial skin breakdown). This is presumably as a result of the decrease in muscle tone from the BoNT resulting in less generation of pressure of the limb against the cast. These preliminary results suggest that casting probably is useful as an adjunct to BoNT injection for improving outcomes in the treatment of lower limb spasticity, especially passive range of motion and Modified Ashworth Scale. If there is a goal of minimizing the risk of casting-related soft-tissue injury, then the cast should be administered in conjunction with BoNT injection; however, further research is required to confirm these findings.

This review demonstrates that there is wide variation in casting protocols of the lower limb post BoNT injection, and highlights the need for studies that lead to the development of a standardized casting protocol. This standardized protocol can then be uniformly applied across future studies and in clinical practice. Use of a standardized casting protocol will allow future meta-analyses (comparison and pooling) of data across studies, thus increasing power and confidence in study results. Notably, studies are needed to determine whether better outcomes result with: immediate vs delayed cast application; keeping the knee flexed vs extended during casting; starting with the ankle at neutral vs at maximum dorsiflexion; and shorter (e.g. 1 week) vs longer (e.g. 12 weeks) cast application.

Given that the method of BoNT injection has been shown to affect spasticity-related outcomes (5), future studies should standardize BoNT injection protocols including muscle localization techniques and BoNT dilutions. Which muscle is being injected needs to be considered; for example, use of ultrasound has been shown to be superior to other methods of injection localization (e.g. anatomical localization, electrical stimulation) when injecting the soleus and gastrocnemius muscles (5).

It is also important to investigate pre-casting protocols that optimize intervention outcomes and improve study results. For example, Xu et al. (22). used a peroneal nerve block to determine available foot and ankle range prior to BoNT injection and casting. Clinicians are increasingly using nerve blocks as part of spasticity treatment algorithms to help select who would be most likely to benefit from what treatment, and determining available range prior to casting would allow for sub-analysis of patient groups according to pre-intervention range of motion (25).

Ideally future studies will also standardize the use of outcome measures. In the lower limb, we suggest that, at minimum, passive range of motion and Modified Ashworth Scale be used as outcome measures within the Body Structure and Function domain. The Modified Ashworth Scale is widely used by spasticity clinicians, and has moderate intra- and inter-rater reliability (26). There will have to be consensus on which Activity domain outcome measure should be included; the 10-m walk test is easier to administer in terms of training and time; however, it may not be sufficiently sensitive to detect differences compared with other measures, such as the 6-min walk test, as demonstrated by Carda et al. (18). These outcomes should be measured at multiple time-points post-intervention (e.g. 3 weeks and 3 months) to determine both short- and long-term effects. Consideration should also be given to including an outcome measure within the Participation domain of the ICF, such as Goal Attainment Scaling. With this tool, patient-specific goals, such as returning to work or recreational pursuits, could be measured. Lastly, future studies should include quality of life outcomes, as these are often most important to the patient. There also needs to be systematic collection and reporting of the burden on the patient and the payer in the form of time to cast, costs related to casting and costs related to adverse events, as these factors are important when fully assessing the risk to benefit ratio of an intervention. For example, skin breakdown could require significant interventions (e.g. pain medication, surgery, time off recreational pursuits or work) that can increase direct and indirect costs. Currently it appears as though casting is a relatively well-tolerated intervention, although care needs to be taken to monitor the skin for potential injury.

Conclusion

Casting as an adjunct to BoNT injection appears to improve spasticity-related outcomes compared with other adjunct therapies, and probably results in less significant soft tissue injury when used following BoNT injection compared with a stand-alone intervention. Currently there are no studies that address whether casting in addition to BoNT is more effective compared with BoNT alone. Addressing this clinical question is important given the extra time, costs and potential adverse events that can be incurred with casting. Casting protocols and outcome measures reported in the literature vary widely and need to be standardized for future studies. Results of this systematic review can be used to inform the development of an international consensus on casting protocols and outcome measures so as to increase the quality, power and confidence in results of future studies for the purpose of determining best practice guidelines for casting in the setting of spasticity.

The authors have no conflicts of interest to declare.