Abstract
Each year, 795,000 people in the United States experience a stroke (Mozaffarian et al., 2015), which can disrupt almost every aspect of a person’s life, including the ability to participate in meaningful occupations such as self-care and leisure. Common deficits after stroke include impairments in sensory, vision, cognitive, communication, and motor functions (Nilsen & Geller, 2015). More than 50% of stroke survivors experience motor impairments lasting 6 mo or longer (Andringa et al., 2013a). Spasticity, a velocity-dependent increase in the tonic stretch reflex of a muscle (Lance, 1990), is one of the most common and disabling motor impairments after stroke. The exact prevalence of poststroke spasticity (PSS) is unknown; researchers have estimated rates from 20% to 92% (Li & Francisco, 2015; Sommerfeld et al., 2012).
Regardless of the variability in prevalence estimates, many stroke survivors experience PSS, and its effect on recovery is clear. If left untreated, PSS can cause muscle shortening, potentially resulting in contractures, joint deformities, and pain (Wissel et al., 2015), leading to decreased independence in all aspects of life. People who experience PSS have significantly lower quality of life compared with the general population (Zorowitz et al., 2013). In addition to physical and emotional complications, stroke survivors experience reduced engagement in and decreased ability to complete daily living tasks and paid employment (Zorowitz et al., 2013). Up to 78% of adults with PSS report decreased independence in self-care tasks because of their spasticity, placing them at higher risk for nursing home placement (Barros Galvão et al., 2014).
To address the negative effect of PSS on motor skills, occupational therapy practitioners commonly use stretching interventions, including range of motion (ROM) exercises, casting, and splinting, to prevent or reduce PSS. In one study, 85.5% of occupational therapists reported prescribing splints for their patients after stroke (Adrienne & Manigandan, 2011). However, little research has addressed the effectiveness of stretching interventions in decreasing PSS (Thibaut et al., 2013).
Lannin and Herbert (2003) completed a systematic review examining the effectiveness of splinting after stroke and concluded that the evidence neither supported nor refuted the effect on PSS. In a systematic review on the effects of stretching on spasticity, Bovend’Eerdt et al. (2008) found the evidence to be inconclusive, but the reviewed studies included adult participants with any type of spasticity and were not specific to the diagnosis of stroke or to the upper extremity. Finally, a systematic review by Katalinic and colleagues (2011) examined primarily the effectiveness of stretching to treat and prevent contractures and secondarily the effect of stretching on spasticity; they concluded that regular, consistent stretching had little or no effect on reducing spasticity.
The systematic reviews available to practitioners thus are older, are not specific to the diagnosis of stroke or to the upper extremity, and only briefly review the effect of spasticity on occupational performance; the evidence they provide is inconclusive. Additional appraisal of the research literature is needed to provide the most current evidence on the effectiveness of stretching interventions to reduce PSS in the upper extremity. Therefore, our research questions were as follows: What is the evidence for the effect of stretching interventions (including splinting) on reducing upper extremity spasticity and increasing hand function (e.g., grip strength, ROM) for people with PSS? What is the evidence for the effect of stretching interventions (including splinting) on improving engagement in functional tasks (e.g., self-care) for people with PSS?
Method
Design and Search Strategy
We completed a systematic review using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses methodology (Moher et al., 2009). The first and second authors (Lindsay Kerr and Vanessa D. Jewell) and a health science librarian developed the following key terms: adults after stroke, upper extremity stretching and splinting interventions, spasticity level, changes in range of motion, effect on activities of daily living abilities, changes in functional use of the hand, changes in motor control, and changes in fine motor coordination skills. Search terms were created using the population and intervention parameters, with additional terms defined using database headings and the synonyms of original key terms to ensure all relevant articles were included (Table 1). On January 9, 2017, the first author searched MEDLINE, CINAHL, OTseeker, AgeLine, and the Cochrane Library for articles published from 2004 to the date of the search.
Key Terms and Search Terms
Note. CVA & cerebrovascular accident; TIA & transient ischemic attack.
In an effort to locate as many articles as possible, we did not use search terms limiting the outcomes. Instead, we screened for the outcomes during the title and abstract review and full-text reviews to meet our inclusion criteria for outcome measures
Study Selection
The initial search results were compiled in RefWorks (Ex Libris, Jerusalem, Israel) for more efficient data management. The first author removed all duplicate articles and completed title and abstract screening. Articles were included if they provided Level I, II, or III evidence (Sackett, 1989); were written in English; were published in a peer-reviewed journal between January 2004 and January 2017; and had participants age 18 yr and older with chronic stroke (≥6 mo) and PSS. Articles were excluded if the study involved interventions focused solely on the lower body, did not include a measure of at least one outcome defined in Table 1, provided Level IV or V evidence, or had participants with a neurological diagnosis other than stroke or who were in the acute or subacute phase of stroke (<6 mo). The full text of articles marked as questionable for inclusion and articles that passed the title and abstract screening was retrieved, and the first author reviewed the articles on the basis of the inclusion and exclusion criteria. After the full-text review, the second author reviewed articles that remained questionable for inclusion, and the first and second authors reached a consensus on inclusion or exclusion.
Data Collection and Analysis
Data collected included level of evidence, study design, number of participants, average age of participants, percentage of participants by gender, description of intervention and control conditions, outcome measures used, and results related to the effectiveness of the intervention. For data analysis, studies with a similar intervention type (e.g., stretching devices) were grouped to establish the effectiveness of the intervention type in reducing PSS. The U.S. Preventive Services Task Force (2018) guidelines, which rate the strength of evidence as strong, moderate, or low, were used to evaluate the strength of the evidence by intervention type.
To assess the risk of bias of each study, we used the Modified Downs and Black Checklist (Downs & Black, 1998) because of its ability to assess the methodological quality of both the randomized and nonrandomized studies that were included in the review. The checklist consists of 26 items allocated among five distinct domains. The first and second authors completed the Modified Downs and Black Checklist independently, discussed the results, and came to a consensus for all articles.
Results
The initial search process identified 3,019 articles. After screening, 58 full-text articles were reviewed, of which 11 met the inclusion criteria (see Figure 1). These studies, which include 6 Level I studies, 0 Level II studies, and 5 Level III studies, are summarized in Table A.1.
Flow of articles through the stages of the systematic review.
Risk of Bias
We examined the risk of bias in all included studies (Table A.2). All 6 Level I studies determined their groups by randomization. Only 2 of the 6 Level I studies used group allocation concealment; 4 were unclear regarding concealment. The 5 Level III studies had high risk of selection and performance bias, and the nature of the intervention likely prevented blinding of participants and personnel. Only 2 studies blinded the clinician responsible for outcome measures. Most studies used objective measurements, lowering the risk of detection bias. Risk of attrition bias was generally low. One study was rated as having a high risk of attrition bias because outcome measures were administered to the intervention and control groups at different times.
Stretching Interventions
Analysis of themes identified four primary intervention types: (1) static splinting, (2) dynamic splinting, (3) manual stretching, and (4) stretching devices.
Static Splinting
Low strength of evidence from 2 studies (1 Level I, 1 Level III) was found regarding the effectiveness of static splinting in reducing spasticity. Basaran et al. (2012, Level I) found no significant differences in spasticity, measured using the Modified Ashworth Scale (MAS; lower scores indicate less spasticity), between the intervention and control groups after 5 wk; no follow-up (i.e., assessment after the end of the intervention) was completed. Fujiwara and colleagues (2004, Level III) found that participants showed significant decreases in MAS scores after 8 wk but provided no follow-up data. Basaran et al. and Fujiwara et al. used similar splint-wearing times of 10 hr overnight and 8 hr during the day, respectively. In addition to splinting, Fujiwara et al. provided one or two occupational or physical therapy sessions a week, and Basaran et al. provided a home exercise program.
Moderate strength of evidence from 3 studies (2 Level I, 1 Level III) was found for the effect of static splinting on hand function. Basaran et al. (2012, Level I) found no significant differences between the intervention and control groups in wrist passive range of motion (PROM) after 5 wk of splint use 10 hr overnight. Fayez and Sayed (2013, Level I) found that PROM and active range of motion (AROM) increased significantly immediately after 1 hr of wearing a static splint; no other measures were completed. Fujiwara et al. (2004, Level III) reported significant increases in AROM for shoulder flexion and finger extension after 6–8 hr of splint wearing during the day for 8 wk. Only Fayez and Sayed examined the effect of static splinting on grip strength, which increased significantly after 1 hr of static splint use.
Two studies (1 Level I, 1 Level III) provide moderate strength of evidence for the effect of static splinting on improving functional tasks. Garros and colleagues (2010, Level III) examined performance of functional tasks and engagement in activities of daily living (ADLs). Participants scored significantly higher on the Box and Block Test after the 3-mo intervention compared with baseline, indicating an increase in hand use; no follow-up results were provided. Garros et al. also used the Canadian Occupational Performance Measure (COPM) to examine ADL engagement, and participants scored significantly higher at 3 mo than at baseline, indicating increased self-perceived performance and satisfaction with performance of ADLs.
Suat et al. (2011, Level I) examined the effect of static splinting on balance, functional reach, and ambulation tasks, all vital components of safe and successful performance of daily activities. Intervention participants wore their splint at least 2 hr/day for 6 mo. On the L Test of Functional Mobility and the Timed Up and Go (TUG), the intervention group scored significantly higher at 4 and 6 mo. Score changes on the Functional Reach Test were significant only at 6 mo, and no significant changes were seen for the Berg Balance Test. In addition, a significant between-groups difference on the TUG was found at 6 mo favoring the intervention group (Suat et al., 2011).
Dynamic Splinting
Low strength of evidence from 1 Level III study was found for the effect of dynamic splinting on reducing spasticity. Andringa et al. (2013b) found no significant changes in MAS scores after 3 and 6 mo of dynamic splint use, and no follow-up was completed. However, of two participants who had previously received Botox injections every 3 mo, one received no injections during the 6-mo intervention period, and the other received only one dose for one of three previously treated muscles.
Three studies (1 Level I, 2 Level III) contributed moderate strength of evidence that dynamic splints can increase hand function. Fayez and Sayed (2013, Level I) identified significant increases in wrist PROM and AROM and grip strength immediately after 1 hr of dynamic splint wearing compared with static splint wearing but completed no other measurements. Andringa et al. (2013b, Level III) noted no significant differences in wrist PROM after 3 mo of intervention but found significant increases after 6 mo of intervention. Chang and Lai (2015, Level III) found significant increases in wrist and finger strength after 1 and 3 mo of intervention but did not include follow-up data.
Low strength of evidence from 1 Level III study was found for the effect of dynamic splinting on functional outcomes. Chang and Lai (2015) found no improvements on the Fugl-Meyer Assessment at 1 and 3 mo compared with baseline measurements.
Manual Stretching
We found 1 manual stretching study that, although it did not measure an effect on PSS, provided moderate strength of evidence for effectiveness in increasing hand function and improving functional task performance. Tseng et al. (2007, Level I) investigated the impact of two manual stretching interventions on ROM and ADL performance for people with chronic PSS. Intervention Group 1 received nursing supervision while completing a stretching protocol on their own, whereas Intervention Group 2 completed the same stretching protocol with physical assistance by a registered nurse; the control group received no intervention. After 4 wk, both intervention groups showed significantly increased ROM, with Group 2 showing a significantly greater increase than Group 1; the control group had decreased ROM. Regarding functional tasks, both intervention groups showed significantly better postintervention FIM® ADL scores than the control group, with no significant difference between the two groups. No follow-up data were provided (Tseng et al., 2007).
Stretching Devices
Stretching devices are resting hand splints with a movable finger piece affixed to a frame with attachments that hold the stretch for various time increments. Strong strength of evidence from 2 Level I studies was found indicating that stretching devices are effective in reducing spasticity. Jung et al. (2011) and Kim et al. (2013) used similar devices but differing protocols in terms of type of stretch, length of stretch, number of sessions per week, and weeks of intervention provided. Despite study differences, participants in both studies showed significant decreases in MAS scores from baseline to the end of intervention (3 wk for Jung et al., 4 wk for Kim et al.). At 1-wk follow-up, Jung et al. found that MAS scores had increased compared with immediately postintervention but were still significantly lower than at baseline. Kim et al. did not provide follow-up data.
Low strength of evidence from 1 Level III study was found for the effect of stretching devices on increasing hand function. Triandafilou et al. (2011) investigated the effect of a stretching device (extension glove) on grip and pinch strength. Participants engaged in one 30-min session each of prolonged finger stretching, repetitive finger stretching, and rest (control session); sessions were separated by at least 1 wk. Measurements were taken before and immediately after each session, with no follow-up measurements. Grip strength showed nonsignificant improvements after both stretching sessions. Lateral pinch strength did not change significantly, and it decreased slightly after prolonged stretching. Performance of functional tasks, tested using components of the Graded Wolf Motor Function Test (e.g., lifting a washcloth, lifting a pen), improved more after stretching than after rest, but this difference did not reach significance.
Discussion
The results of this review provide occupational therapy practitioners with the latest evidence to guide the use of stretching and splinting interventions to reduce PSS, contributing to the understanding of the effect of stretching and splinting on increasing hand function and improving engagement in functional tasks. For reducing spasticity, this review found low strength of evidence for the use of static and dynamic splinting and strong strength of evidence for the use of stretching devices; the single article on manual stretching we found did not include reduced spasticity as an outcome measure. For increasing hand function, moderate strength of evidence was found to support the use of static splinting, dynamic splinting, and manual stretching, and low strength of evidence was found for the use of stretching devices. For improving functional tasks, moderate strength of evidence was found to support the use of static splinting, dynamic splinting, and manual stretching, and low strength of evidence was found for the use of stretching devices.
This review highlights the scarcity of research on stretching and splinting interventions, despite their common use with people who have experienced a stroke. Only 11 articles met the inclusion criterion of publication in the 13-yr span from 2004 to 2017. In addition, this review highlights the mixed nature of the strength of evidence currently available, which is low or moderate for most interventions and strong for only one intervention.
This review supports the assertion that stretching and splinting cannot be provided as stand-alone interventions and should be used in conjunction with other hands-on interventions provided by occupational therapy practitioners. For example, Fujiwara et al. (2004) found significantly reduced spasticity after 8 wk of static splinting, yet Basaran et al. (2012) found no significant differences after 5 wk. Although the length of the intervention may have affected these outcomes, Fujiwara et al. included one or two sessions of physical or occupational therapy each week, whereas Basaran et al. included a home exercise program; the hands-on therapy sessions may have improved Fujiwara et al.’s results. Likewise, Tseng et al. (2007) found that participants who received hands-on assistance from a nurse had significantly better ROM than those who self-administered the intervention with supervision.
This review suggests that dynamic splints may have a more beneficial effect than static splints in reducing spasticity. Fayez and Sayed (2013) found that dynamic splints were significantly better than static splinting at increasing ROM. Andringa et al. (2013b) indicated that dynamic splints were better than static splints in controlling spasticity on the basis of wearer opinion. Although these findings are not conclusive because they are based on limited data, the possibility of an advantage to dynamic splinting merits further inquiry.
The intervention periods in the studies in this review ranged widely, from 30 min to 6 mo. At the systematic review level, there is no clear indication that the longer interventions were better than the shorter ones. For example, Fayez and Sayed (2013) had participants wear a static splint for one 60-min session and saw significant improvements in ROM and grip strength, yet Basaran et al. (2012) had participants wear a static splint 10 hr every night for 5 wk and found no significant changes in ROM. At the individual study level, some studies found that longer interventions were more effective. For example, Andringa et al. (2013b) found that changes in PROM were not significant at 3 mo but were significant after 6 mo of dynamic splint use. Chang and Lai (2015) also used dynamic splinting and found significant improvements in finger and wrist strength at 1 mo and 3 mo compared with baseline and at 3 mo compared with 1 mo.
Of the 11 articles included in this review, only Jung et al. (2011) completed a follow-up assessment that allows insight into the effectiveness of a stretching device intervention once the device was no longer used. The 1-wk postintervention assessment showed a significant decrease in spasticity compared with baseline but an increase compared with immediately postintervention. Jung et al. concluded that their results suggest a need for continuous use of the stretching device to maintain reduced PSS.
Finally, attention to the effect of stretching and splinting on functional tasks varied among the included studies. Triandafilou et al. (2011) noted an increase in speed of picking up a washcloth and a pen after use of a stretching device, but the change was not significant compared with the control group. Chang and Lai (2015) found no significant improvement on the Fugl-Meyer Assessment, a performance-based measure, after 3 or 6 mo of dynamic splint use. Tseng et al. (2007) found significant improvement in FIM ADL scores after manual stretching interventions compared with a control group. Garros et al. (2010) noted increased ADL abilities after 3 mo of static splint use, as shown by increased COPM scores compared with baseline.
Limitations
Moderate strength of evidence was found for the use of stretching devices, low strength of evidence for the use of static splinting, and inconclusive evidence for the use of manual stretching and dynamic splinting for the treatment of PSS. These results should be interpreted with caution, however, because of limitations in the individual studies reviewed. Six of the 11 studies provided Level I evidence; the remainder provided Level III evidence. Blinding of participants was limited, and in most of the randomized controlled trials blinding of personnel was not attempted. Many studies had small sample sizes.
Intervention protocols were inconsistent across the studies. Many participants received a stretching intervention along with a home exercise program or traditional therapy, making it unclear whether the improvements were a result of the stretching only or of the combination of stretching and additional therapy. Several articles did not fully explain the methods and interventions in a way that would enable the studies to be easily duplicated.
The variety of outcome measures used makes the studies difficult to compare. The majority of the studies used objective measures; however, one study used retrospective self-reporting, a method open to bias. Only one study (Jung et al., 2011) used a follow-up measurement, at only 1 wk postintervention, so no insight into long-term effects can be gained from these studies.
Implications for Occupational Therapy Research
There is a clear need for additional research on stretching interventions for people with PSS to strengthen the available evidence. Areas of research needed include the following:
Definition of wearing protocols for static and dynamic splints
Continued examination of the effectiveness of manual stretching as an intervention
Further exploration of the use of stretching devices as an intervention
Use of occupation-based and occupation-focused outcomes to better understand the effect PSS has on engagement in meaningful occupations
Exploration of new and alternative interventions for potential use.
Implications for Occupational Therapy Practice
The findings of this systematic review have the following implications for occupational therapy practice with people with PSS:
For static or dynamic splinting interventions, some patients may need extended use (up to 6 mo) before improvements in PSS are noted.
Patients may experience better comfort and tolerance with dynamic splints than with static splints.
Stretching and splinting should not be used as stand-alone interventions. The evidence shows that better outcomes can be obtained when these interventions are used in conjunction with interaction with a skilled occupational therapy practitioner.
Stretching and splinting need to be ongoing for maximum benefit.
Stretching and splinting interventions alone may not improve patients’ ADL abilities, so practitioners may need to also use a modification or adaptation approach to intervention to increase ADL independence.
Practitioners should monitor additional research that further supports or refutes the use of stretching and splinting interventions as it comes available.
Conclusion
This systematic review offers insight into interventions to address PSS and describes the low to moderate strength of evidence for the effectiveness of stretching interventions aimed at reducing spasticity, increasing hand function, and improving engagement in functional tasks for adults with PSS. The studies included in the review varied in terms of level of evidence, intervention protocols, number of participants, and outcome measures. To reduce spasticity, moderate strength of evidence supports the use of static splinting, dynamic splinting, and stretching devices, but the best protocol is yet to be determined, and the long-term effects of stretching interventions remain unclear. To increase hand function, moderate strength of evidence supports the use of static splinting and dynamic splinting, but low strength of evidence was found for the use of manual stretching and stretching devices. To improve functional tasks, low strength of evidence was found for the use of static splinting, dynamic splinting, and manual stretching; engagement in functional tasks was not addressed in the stretching device studies.
To assist occupational therapy practitioners in providing better patient care, additional research is needed that focuses on defining protocols and dosage recommendations for stretching and splinting interventions. For example, what the best daily wearing schedule is and how long a splint must be worn to be effective still need to be determined. In addition, stretching devices seem to be promising, but more research is required to determine effectiveness. Finally, the one study of manual stretching in this review supported its use, but without additional studies it is unknown whether manual stretching is truly effective.
Footnotes
Appendix
Risk-of-Bias Table for Included Intervention Studies
| Citation | Selection Bias | Performance Bias: Blinding of Participants and Personnel | Detection Bias: Blinding of Outcome Assessment: Self-Reported Outcomes | Attrition Bias: Incomplete Outcome Data | Reporting Bias: Selective Reporting | |
| Random Sequence Generation | Allocation Concealment | |||||
| Andringa et al. (2013) | – | – | – | + | + | + |
| Basaran et al. (2012) | + | + | – | + | + | + |
| Chang & Lai (2015) | – | – | – | + | + | + |
| Fayez & Sayed (2013) | + | ? | – | + | + | + |
| Fujiwara et al. (2004) | – | – | – | + | + | + |
| Garros et al. (2010) | – | – | – | – | + | + |
| Jung et al. (2011) | + | ? | – | + | – | + |
| Kim et al. (2013) | + | ? | + | + | + | + |
| Suat et al. (2011) | + | + | – | + | + | + |
| Triandafilou et al. (2011) | – | – | – | + | + | + |
| Tseng et al. (2007) | + | ? | + | + | + | + |
Note. Categories for risk of bias: + = low risk of bias; ? = unclear risk of bias; – = high risk of bias. Risk-of-bias table format adapted from “Assessing Risk of Bias in Included Studies,” by J. P. T. Higgins, D. G. Altman, and J. A. C. Sterne, in Cochrane Handbook for Systematic Reviews of Interventions (Version 5.1.0), by J. P. T. Higgins and S. Green (Eds.), March 2011. Retrieved from https://handbook-5-1.cochrane.org/. Copyright © 2011 by The Cochrane Collaboration.
Copyright © 2020 by the American Occupational Therapy Association. This table may be freely reproduced for personal use in clinical or educational settings as long as the source is cited. All other uses require written permission from the American Occupational Therapy Association. To apply, visit www.copyright.com.
Suggested citation: Kerr, L., Jewell, V. D., & Jensen, L. (2020). Stretching interventions for poststroke spasticity, hand function, and functional tasks: A systematic review (Table A.2). American Journal of Occupational Therapy, 74, 7405205050. https://doi.org/10.5014/ajot.2020.029454
*
Indicates studies that were included in the systematic review.