972

REVIEW ARTICLE (META-ANALYSIS)

Modified Constraint-Induced Movement Therapy Versus Traditional Rehabilitation in Patients With Upper-Extremity Dysfunction After Stroke: A Systematic Review and Meta-Analysis Yue X. Shi, MD, Jin H. Tian, PhD, Ke H. Yang, PhD, Yue Zhao, PhD ABSTRACT. Shi YX, Tian JH, Yang KH, Zhao Y. Modified constraint-induced movement therapy versus traditional rehabilitation in patients with upper-extremity dysfunction after stroke: a systematic review and meta-analysis. Arch Phys Med Rehabil 2011;92:972-82. Objective: To compare the effectiveness of modified constraint-induced movement therapy (CIMT) with traditional rehabilitation (TR) therapy in patients with upper-extremity dysfunction after stroke. Date Sources: Computerized database searches and handsearches, as 2 main search strategies, were used to collect studies. A comprehensive search of PubMed, EMBASE, the Cochrane Library, the Chinese academic journals full-text database, the Chinese biomedical literature database, the Chinese scientific journals database, and Chinese medical association journals was conducted. Relevant journals and article reference lists were hand-searched. Meanwhile, we searched unpublished trials by using the System for Information on Gray Literature database. Study Selection: Randomized controlled trials (RCTs) only about modified CIMT versus TR for treatment of patients with upper-extremity dysfunction after stroke were identified in this systematic review. Participants included adults age over 18 years with a clinical diagnosis of stroke and met the inclusion criteria of modified CIMT. Date Extraction: Two reviewers extracted relevant information from included studies according to a date extraction form. The methodologic quality of the included studies was assessed using a quality-scoring instrument, which was a 5-point scale that included a description of randomization, double-blind structure, and withdrawals/dropouts. Data Synthesis: Thirteen RCTs involving 278 patients (modified CIMT/TR⫽143/135) were included. Meta-analysis showed that patients receiving modified CIMT showed higher scores for the Fugl Meyer Assessment (mean difference [MD]⫽7.8; 95% confidence interval [CI], 4.21–11.38), the Action Research Arm Test (MD⫽14.15; 95% CI, 10.71– 17.59), the FIM (MD⫽7.00; 95% CI, .75–13.26), and the Motor Activity Log: Amount of Use (MD⫽.78; 95% CI, .37– 1.19) and Quality of Use (MD⫽.84; 95% CI, .42–1.25) than

From School of Nursing, Tianjin Medical University, Tianjin (Shi, Zhao); Evidence-Based Medicine Center, School of Basic Medical Sciences, Lanzhou University, Lanzhou (Tian, Yang), China. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Correspondence to Yue Zhao, PhD, School of Nursing, Tianjin Medical University, Observatory Road, Heping District, Tianjin, China, e-mail: [email protected]. Reprints are not available from the author. 0003-9993/11/9206-00810$36.00/0 doi:10.1016/j.apmr.2010.12.036

Arch Phys Med Rehabil Vol 92, June 2011

patients in the TR group. In kinematic variables, patients receiving modified CIMT had a shorter reaction time and a higher percentage of movement time where peak velocity occurred than patients receiving TR (P⬍.05), while meta-analysis showed that there was no significant difference in normalized movement time (P⫽.99), normalized total displacement (P⫽.44), and normalized movement unit (P⫽.68). Conclusions: This systematic review provided fairly strong evidence that modified CIMT could reduce the level of disability, improve the ability to use the paretic upper extremity, and enhance spontaneity during movement time, but evidence is still limited about the effectiveness of modified CIMT in kinematic analysis. Key Words: Occupational therapy; Rehabilitation; Review; Stroke; Upper extremity. © 2011 by the American Congress of Rehabilitation Medicine TROKE IS A COMMON disease and is a leading cause of S death and disability in China. In the United States and Western Europe, stroke is the third leading cause of death and 1

the leading cause of long-term disability.2,3 Approximately 75% of stroke survivors are left with motor dysfunction. Although this motor dysfunction is improved to some extent after rehabilitation training, a large proportion (30 – 60%) of patients are left with persistent impairment of upper-extremity movement, and moderate to severe disability affects about 10%.4,5 Generally, functional recovery of upper-extremity function is more difficult than recovery of lower-extremity function, mainly because the patient with stroke and unilateral upperextremity dysfunction may progressively avoid using the more affected arm in favor of the nonparetic upper extremity.6 Therefore, functional recovery is affected, and a learned nonuse phenomenon is formed. To solve this problem, Taub and his colleagues proposed constraint-induced movement therapy (CIMT) in 1993, and it is suggested that CIMT may be used to overcome learned nonuse and induce cortical reorganization.6-8 CIMT is one of the most developed training approaches for motor restoration and is based on a theory of brain plasticity and cortical functional reorganization. Constraint and massed and repeated practice may correct the learned nonuse and then improve functional performance of the more affected upper List of Abbreviations CI CIMT MD RCTs SIGLE TR

confidence interval constraint-induced movement therapy mean difference randomized controlled trials System for Information on Gray Literature traditional rehabilitation

MOVEMENT THERAPY VERSUS TRADITIONAL REHABILITATION POSTSTROKE, Shi

extremity.6-9 CIMT involves massed and intensive practice with the more affected upper extremity and includes 2 components: use of the unaffected upper extremity is restrained during 90% of waking hours, and at the same time, the more affected upper extremity receives repeated and intensive training for 6 hours or more a day. By this means, the use of the more affected arm may be increased, and learned nonuse may be overcome.7,9-12 CIMT has been widely used in patients with chronic, subacute, and acute stroke with motor impairment of the unilateral upper extremity.8,13-15 In a recent systematic review, the available evidence about the effects of CIMT demonstrated that compared with other traditional rehabilitative techniques, CIMT could improve functional performance and increase the usage of the more affected upper extremity.16,17 However, an increased amount of practice task and longer restraint time may be dangerous for patients during the treatment period. In addition, patients may have difficulty with full compliance for this prolonged practice session; thus, the clinical feasibility of CIMT has been questioned.18 It has been confirmed that only 32% of patients comply with the CIMT restriction schedule.11,19 According to the points mentioned, Page and colleagues9,12,19-21 designed a modified CIMT that shortens both the intensive training session of the paretic upper extremity (30min/d–2h/d) and the restraint time of the nonparetic upper extremity (ⱕ6h/d). Outcome measures for assessing the effects of modified CIMT include both clinical evaluation and kinematic analysis. Scales are used to assess clinical measures, such as the Fugl Meyer Assessment,22 the Action Research Arm Test,23,24 and the FIM,25 and to evaluate the level of disability, impairment, and so forth, but these measures may have subjective aspects. Thus, more objective outcome measures are necessary to provide an accurate analysis of the results.22 Kinematic analysis has been used to assess motor performance in recent years and can provide objective and quantitative measures to understand the movement mechanism of the upper extremity.26,27 Whether modified CIMT is a more effective rehabilitation strategy than traditional rehabilitation (TR) for improving motor function in adults with hemiparalysis after stroke is not well known. The aim of this study was to determine the effectiveness of modified CIMT from different perspectives (not only clinical measures but also kinematic measures) in order to provide the best clinical evidence. METHODS Search Strategy Searches were conducted by 2 authors. Computerized database searches and hand-searches were used as the 2 main search strategies to collect studies. We searched for “(stroke OR cerebral stroke OR cerebrovascular disorder OR brain vascular accident OR cerebral infarction OR brain injury OR brain hemorrhage OR brain ischemia OR hemiplegia OR paresis OR paretic) AND (constraint induced movement therapy OR constraint-induced movement therapy OR modified constraint-induced therapy OR forced use) AND (upper extremity OR upper limb)” in PubMed, EMBASE, the Cochrane Library, the Chinese academic journals full-text database, the Chinese biomedical literature database, the Chinese scientific journals database, and Chinese medical association journals. All computerized database searches were conducted from inception to April 2010. The search method incorporated National Library of Medicine Medical Subject Heading terms with a text words search. The detailed search strategy for each electronic database is available from the authors. We hand-searched relevant

973

journals and article reference lists for additional eligible studies. We included unpublished trials by retrieving them from the System for Information on Gray Literature (SIGLE) database to reduce publication bias. Inclusion Criteria We selected all publications and screened studies if they met the following criteria: 1. Randomized controlled trials (RCTs) only about modified CIMT versus TR for treatment of patients with upper-extremity motor dysfunction after stroke were included. No restrictions were made regarding the language or nationality of publication. 2. Participants included adults age over 18 years with a clinical diagnosis of stroke and met the inclusion criteria of modified CIMT (ability to extend at least 10° at the interphalangeal and metacarpophalangeal joints and 20° at the wrist; a score ⱖ69 on the Modified Mini-Mental State Examination; the more affected arm nonuse defined as a score of ⬍2.5 on the Motor Activity Log, etc). 3. For interventions, intensive training session of the paretic upper extremity ranged from 30 minutes to 3 hours a day, and restraint time of the nonparetic upper extremity was less than 6 hours a day in the modified CIMT group. In the TR group, the patients received only TR without any restraint of the nonparetic upper extremity. 4. The trials reported scores for at least 1 measure of upper-limb function. Studies mixed with other interventions were excluded. Outcome Measures Outcome measures were made up of clinical variables and kinematic variables. The Fugl Meyer Assessment, Action Research Arm Test, FIM, Motor Activity Log including Amount of Use and Quality of Use, Wolf Motor Function Test, and Stroke Impact Scales were adopted as clinical variables. Kinematic variables used were as follows: normalized movement time, reaction time, normalized movement unit, normalized total displacement, percentage of movement time where peak velocity occurred, and peak velocity. Selection of Literature Both authors according to inclusion criteria independently reviewed the title and abstract and excluded the studies that did not meet the inclusion criteria obviously. A full text of any published article that potentially met the inclusion criteria was obtained to confirm. Any disagreements during the selection course were resolved by discussion with a third reviewer. Data Extraction All relevant information about the studies was collected in data extraction form, which included study characteristics (author, year of publication, country of origin), sample size, patient demographics, content and frequency of intervention, treatment time, outcome measures, and methodologic quality of included studies. We e-mailed the original authors to inquire about deficient information in these included studies. Methodologic Quality Assessment The methodologic quality assessment was completed by 2 independent authors. The criterion for methodologic quality was assessed using a quality-scoring instrument developed by Jadad et al,28 which is a 5-point scale including a description of randomization, double-blind structure, and withdrawals/dropArch Phys Med Rehabil Vol 92, June 2011

974

MOVEMENT THERAPY VERSUS TRADITIONAL REHABILITATION POSTSTROKE, Shi

Fig 1. Selection of studies. Abbreviation: UE, upper extremity.

outs. In addition, we assessed the comparability of the baseline and allocation concealment of the included studies. Statistical Analysis Review Manager Software (RevMan5.0)a was used for meta-analysis. Heterogeneity among the studies was evaluated by using the chi-squared test and the I2 statistic. The fixedeffects model was used if the heterogeneity test showed no statistical significance (I2⬍50%; P⬎0.1). Otherwise we adopted the random-effects model. A sensitivity analysis was conducted if heterogeneity existed among the studies. All of variables in the included studies were continuous data, so we used the mean difference (MD) and 95% confidence interval (CI) to analyze the effect size of the studies. We considered P less than .05 statistically significant. We describe the results from the original studies if items of the scale were not fully reported or outcome measures were reported in only 1 study. RESULTS Study Selection A total of 294 references were identified in our screening. Among these, 55 duplicates were excluded by Endnote software,b and 224 citations were excluded after initial screening Arch Phys Med Rehabil Vol 92, June 2011

because they were reviews, they were case reports, or subjects and interventions were not related to our topic. Fifteen full-text articles were maintained for further screening. Four studies were related to our study after searching the reference lists of the 15 full-text articles. Of 19 studies, 6 of them14,26,29-32 were excluded for the following reasons: restraint time of the unaffected upper extremity was more than 6 hours,14,26,29,30 the unaffected upper extremity was restrained in both the modified CIMT and the TR groups,31 and treatment time of the more affected upper extremity was more than 3 hours and without any restraint of the unaffected upper extremity between the groups.32 Finally, 13 RCTs9,10,12,20,25,33-40 were included in our systematic review (fig 1). Study Description This systematic review identified 13 RCTs9,10,12,20,25,33-40 from 4 countries (United States,9,12,20,38-40 Taiwan, China,10,25,35,36 Saudi Arabia,37 Sweden33,34) involving 278 patients (modified CIMT/TR⫽143/135). All included trials were published between 2000 and 2009. Table 1 shows the basic information of the studies in this systematic review. Table 2 includes the quality assessment of the studies. Table 3 lists the description of interventions in both groups.

975

MOVEMENT THERAPY VERSUS TRADITIONAL REHABILITATION POSTSTROKE, Shi Table 1: Basic Information of Included Studies First Author, Year, Country

Age (y), Mean ⫾ SD

Sample Size (N), (mCIMT/TR) (n/n)

Hammer AM, 2009, Sweden33 Hammer AM, 2009, Sweden34 Page SJ, 2008, US9

mCIMT: 66.3⫾10.3 TR: 60.4⫾11.1 mCIMT: 66.3⫾10.3 TR: 60.4⫾11.1 57.9⫾8.4

30 (15/15)

Lin KC, 2007, Taiwan, China10 Wu CY, 2007, Taiwan, China25 Wu CY, 2007, Taiwan, China35

mCIMT: 57.1⫾18.3 TR: 58.7⫾15.1 mCIMT: 54.6⫾8.6 TR: 53.3⫾6.3 mCIMT: 71.4⫾6.4 TR: 71.9⫾6.8

32 (17/15)

Wu CY, 2007, Taiwan, China36

mCIMT: 59.9⫾11.2 TR: 56.8⫾12.9

Page SJ, 2005, US20

Time Since Stroke (mo)* Range, Mean ⫾ SD

Baseline Comparability

Outcome Measures

1–6, mCIMT: 2.6⫾1.5 TR: 2.3⫾1.2 1–6, mCIMT: 2.6⫾1.5 TR: 2.3⫾1.2 20–60, 39.8†

Yes

Motor Activity Log

Yes

13–26, mCIMT: 15.97⫾3.46 TR: 16.61⫾2.89 12–36, mCIMT: 18.53⫾6.92 TR: 17.61⫾7.55 0.5–31, mCIMT: 6.70⫾8.99 TR: 8.32⫾7.89

Yes

47 (24/23)

3wk–37, mCIMT: 2.51⫾9.64 TR:11.98⫾11.72

Yes

mCIMT: 58.6⫾6.4 TR: 62.2⫾10.3

10 (5/5)

2-9d, mCIMT: 4.00⫾1.58d TR: 4.80⫾3.03d

Unreported

Page SJ, 2004, US12

mCIMT: 54.5⫾12.8 TR: 60.8⫾13.6

11 (7/4)

Unclear

Atteya AA, 2004, Saudi Arabia37

mCIMT: 55.0⫾2.8 TR: 52.0⫾4.2

4 (2/2)

14-74 mo, mCIMT: 25.43⫾6.53 TR: 38.00⫾23.95 2.3–5.8, mCIMT: 5.60⫾0.28 TR: 3.95⫾2.33

Page SJ, 2002, US38

mCIMT: 59.9⫾11.2 TR: 56.8⫾12.9

9 (4/5)

1–6, mCIMT: 5.00⫾0.82 TR: 4.90⫾0.89

Unclear

Page SJ, 2001, US39

mCIMT: 55.0⫾4.2 TR: 52.0⫾5.7

4 (2/2)

2–5.8, mCIMT: 5.65⫾0.21 TR: 3.75⫾2.47

Unreported

Dromerick AW, 2000, US40

mCIMT: 61.5⫾11.7 TR: 71.4⫾5.3

20 (11/9)

4–14d, 6⫾2.6d‡

Yes

Fugl Meyer Assessment, Action Research Arm Test Fugl Meyer Assessment, Action Research Arm Test, Motor Activity Log Motor Activity Log, Kinematic Analysis Variables, FIM Motor Activity Log, Kinematic Analysis Variables, FIM Fugl Meyer Assessment, Motor Activity Log, FIM, Stroke Impact Scales Fugl Meyer Assessment, Motor Activity Log, Kinematic Analysis Variables Fugl Meyer Assessment, Action Research Arm Test, Motor Activity Log Fugl Meyer Assessment, Action Research Arm Test, Motor Activity Log Fugl Meyer Assessment, Action Research Arm Test, Motor Activity Log, Wolf Motor Function Test Fugl Meyer Assessment, Action Research Arm Test, Motor Activity Log Fugl Meyer Assessment, Action Research Arm Test, Motor Activity Log, Wolf Motor Function Test Action Research Arm Test, FIM, Barthel Index

30 (15/15) 25 (13/12)

30 (15/15) 26 (13/13)

Yes

Yes Yes

Unreported

Abbreviation: mCIMT, modified CIMT. *Unit is months unless otherwise specified. † SD was not available from original article. ‡ Subgroup information was not available from the original research.

Baseline Characteristics Thirteen RCTs involved 278 patients with an age range from 31 to 83 years. Participants across all trials had a clinical diagnosis of the upper-limb hemiparesis after stroke. The time since onset of stroke varied from 2 days to 60 months. Two trials20,40 included patients in the acute stage (⬍14d), 533-34,37-39 in the subacute stage (⬍6mo), 49,10,12,25 in the chronic stage (⬎12mo), and 235,36 in both the subacute and the chronic stages. As for clinical variables, 9 trials9,12,20,34,35-39 used the Fugl Meyer Assessment to evaluate the arm motor impairment. The Action Research Arm Test and the Wolf Motor Function Test were used in 89,12,20,34,37-40 and 237,39 trials, respectively, to assess the arm motor function. Eleven trials9,10,12,20,25,29,35-39 adopted the Motor Activity Log to assess the arm motor function reported by patients and caregivers. The FIM and the Barthel Index were used in 4 trials10,25,35,40 and 1 trial,40 respectively, to evaluate the independence of patients in daily living. One trial35 used the Stroke Impact Scales to measure the quality of life of patients with stroke. All of these scales have been shown to have a high level of reliability and validity.22-25,41-43 Three

trials10,25,36 used Kinematic Analysis Variables to explore the spatiotemporal characteristics of movement (see table 1). Methodologic Quality Assessment The randomization procedures appeared adequate in 9 studies,9,10,12,20,33-36,40 and 4 studies25,37-39 were unclear. Only 1 study10 depicted the allocation concealment, and the others were unclear. Triple-blinding was performed by 1 trial,10 and 4 trials9,25,36,39 adopted double-blinding. Evaluators were blind in 5 trials,12,20,35,38,40 while 3 trials33,34,37 did not report any form of blinding at all. Three studies10,33,34 reported posttreatment withdrawals, and intention to treat was performed in 2 of them.33,34 Nine studies9,10,12,25,35,36,39,40 were assigned a Jadad28 score of ⱖ4, whereas 4 studies33,34,37,38 received a score of 3 because they did not report blinding or describe sequence generation. Two included trials12,38 reported comparability of baseline characteristics partially, while 3 trials20,37,39 did not report the comparability of baseline characteristics (see table 2). Arch Phys Med Rehabil Vol 92, June 2011

976

MOVEMENT THERAPY VERSUS TRADITIONAL REHABILITATION POSTSTROKE, Shi Table 2: Methodologic Quality of Included Studies

First Author, Year

Sequence Generation

Allocation Concealment

Hammer AM, 200933 Hammer AM, 200934 Page SJ, 20089 Lin KC, 200710

Draw lot Draw lot Computer-generated Random numbers table

No No No Envelope

Wu CY, 200725 Wu CY, 200735 Wu CY, 200736 Page SJ, 200520 Page SJ, 200412 Atteya AA, 200437 Page SJ, 200238 Page SJ, 200139

Unclear Random numbers table Random numbers table Random numbers table Computer-generated Unclear Unclear Unclear

No No No No No No No No

Dromerick AW, 200040

Random numbers table

No

Withdrawals (n) (mCIMT/TR)

Intention to Treat

Score (/5)

3/1 3/1 No 0/2

Yes Yes Unmentioned Unclear

3 3 5 4

assessor assessor, blind

No No No No No No No No

Unmentioned Unmentioned Unmentioned Unmentioned Unmentioned Unmentioned Unmentioned Unmentioned

4 4 5 4 4 3 3 4

assessor

No

Unmentioned

4

Blinding

No No Blind outcome Blind outcome therapist Blind outcome Blind outcome Blind outcome Blind outcome Blind outcome No Blind outcome Blind outcome therapist Blind outcome

assessor and therapist assessor, subject and assessor, blind subject assessor assessor and subject assessor assessor

Abbreviation: mCIMT, modified CIMT.

Interventions We included only the trials in which modified CIMT was compared with TR. The content of interventions was different in the included trials but mainly focused on physiotherapy, occupational therapy, neurodevelopmental treatment, neuromuscular facilitation, and daily living retraining. Total treatment time of the paretic upper limb, restraint time of the nonparetic upper limb, and frequency of practice sessions varied across these trials. The total treatment time of 3 trials33,34,40 lasted for 2 weeks, 4 trials10,25,35,36 for 3 weeks, and 6 trials9,12,20,37-39 for 10 weeks. In 6 trials,9,12,20,37-39 the restraint time of the nonparetic arm lasted for 5 hours a day, and in 7 trials,10,25,33-36,40 it lasted for 6 hours. The prescribed practice session of 4 trials9,12,20,37 was 30min/d for 3d/wk. Two trials38,39 were executed 1 hour daily for 3d/wk. Five trials10,25,35,36,40 were executed 2h/d, and 2 trials33,34 were performed 3h/d for 5d/wk (see table 1). Effect of Modified CIMT: Clinical Evaluation Modified CIMT versus TR— effect on arm motor impairment level. Six RCTs9,20,35-37,39 with 116 participants reported the Fugl Meyer Assessment. The fixed-effects model was used in these studies because the heterogeneity was small among them (P⫽.34; I2⫽12%). In pooled analysis, the Fugl Meyer Assessment score was higher in the modified CIMT group than in the TR group, with an MD of 7.8 (95% CI, 4.21–11.38; P⬍.001) (fig 2). Modified CIMT versus TR—effect on arm motor function. Five RCTs9,20,37,39,40 with 63 participants reported the Action Research Arm Test. The fixed-effects model was used in 5 studies because the heterogeneity was small among them (P⫽.29; I2⫽20%). The results of meta-analysis showed that the score of the Action Research Arm Test in the modified CIMT group was higher than in the TR group, with an MD of 14.15 (95% CI, 10.71–17.59; P⬍.001) (fig 3). The data for the Wolf Motor Function Test were reported in only 2 trials,37,39 and the results showed substantial improvements between pretesting and posttesting in the modified CIMT group alone. The results could not be pooled because of the small sample size (only 8 patients in 2 studies) and missing data on some items of the Wolf Motor Function Test. Arch Phys Med Rehabil Vol 92, June 2011

Modified CIMT versus TR— effect on focal disability level. Three RCTs10,25,35 with 88 participants reported on the FIM. The heterogeneity test showed an I2 of 0% among the studies. The results of meta-analysis revealed that there was a statistical significance between the groups with an MD of 7 (95% CI, .75–13.26; P⫽.03) (fig 4). The data from the Barthel Index were available in only 1 study, by Dromerick et al.40 The results showed there was no statistical significance between the groups in this measure (t⫽1.14; P⫽.27). Modified CIMT versus TR— effect on perceived arm motor function. The detailed data from 6 RCTs10,20,25,33,35,36 with 173 participants used the Motor Activity Log to assess the perceived arm motor function. The Motor Activity Log contains an Amount of Use scale and a Quality of Use scale. The random-effects model was used because there was heterogeneity among the studies (Amount of Use, P⬍.001, I2⫽90%; Quality of Use, P⫽.004, I2⫽71%). In pooled analysis, the score on the Motor Activity Log in the modified CIMT group was higher than in the TR group (Amount of Use, MD⫽1.09, 95% CI, .26 –1.91, P⫽.01; Quality of Use, MD⫽1.02, 95% CI, .55–1.49, P⬍.001) (fig 5). In the study by Page et al,20 10 patients received repetitive and concentrated training for ⱕ30min/d for 10 weeks, which was greater than the time of training in 5 studies10,25,33,35,36 and this may be the reason for the better treatment result in this trial compared with the others. In addition, this trial had a smaller sample size than others. Therefore, we excluded this trial to perform sensitivity analysis. After sensitivity analysis, the heterogeneity was reduced (Amount of Use, P⫽.22, I2⫽31%; Quality of Use, P⫽.18, I2⫽36%). There was still statistical significance between the groups (Amount of Use, MD⫽.78, 95% CI, .37–1.19, P⬍.001; Quality of Use, MD⫽.84, 95% CI, .42–1.25, P⬍.001) (fig 6). Modified CIMT versus TR— effect on quality of life. The data from the Stroke Impact Scales were assessed in 1 study,35 so the results could not be combined. The results of this trial demonstrated that modified CIMT could improve the quality of life of patients with stroke, especially in the physical domains (t⫽1.14; P⫽.27).

977

MOVEMENT THERAPY VERSUS TRADITIONAL REHABILITATION POSTSTROKE, Shi Table 3: Description of Treatment Strategy First Author, Year

mCIMT Content of Interventions With the Affected Arm

Hammer AM, 200933 Hammer AM, 200934

RT

TR Content of Interventions

Ongoing rehabilitation programmer included PT and OT Facilitation; improvement of strength and endurance; skilled task training and ADL; range of motion Shaping techniques

6h

3hr/d, 5d/wk

2wk

3h/d, 5d/wk

2wk

Lin KC, 200710

ADL such as writing, combing, flipping cards, etc.

6h

Wu CY, 200725

Shaping techniques; normalization of muscle tone

6h

Wu CY, 200735

Shaping and adaptive, repetitive task practice; normalization of muscle tone Shaping techniques

6h

Page SJ, 200520

Shaping techniques; range of motion

5h

Page SJ, 200412

OT: shaping techniques, compensatory techniques; PT: stretching to facilitate ADL OT⫹PT: PNF emphasis on ADL task; compensatory techniques; 2 functional tasks on WMFT OT: function task; shaping techniques; PT: stretching, balance activities, gait training OT⫹PT: PNF emphasis on ADL tasks; compensatory techniques; 2 functional tasks on WMFT OT focused on ADLs

5h

Ongoing rehabilitation programmer included PT and OT Facilitation; improvement of strength and endurance; skilled task training and ADL; range of motion PNF focused on functional task and stretching; compensatory techniques Strength, balance, fine motor dexterity training; functional task practice; stretching/weightbearing NDT focused on balance training, stretching, weight-bearing; finemotor dexterity training NDT, stretching, weight bearing; fine motor dexterity; compensatory techniques NDT focused on functional task; stretching, weight-bearing; fine dexterity training Stretching; weight bearing; dexterity exercise; compensatory techniques PNF, compensatory techniques

5h

Page SJ, 20089

Wu CY, 200736

Atteya AA, 200437 Page SJ, 200238 Page SJ, 200139 Dromerick AW, 200040

6h

5h

6h

TT

30min/d, 3d/wk

TTT

10wk

2h/d, 5d/wk

3wk

2h/d, 5d/wk

3wk

2h/d, 5d/wk

3wk

2h/d, 5d/wk

3wk

30min/d, 3d/wk

10wk

30min/d, 3d/wk

10wk

OT⫹PT: PNF emphasis on ADL task; compensatory techniques

30min/d, 3d/wk

10wk

5h

PNF during the majority of PT and OT; compensatory techniques

1h/d, 3d/wk

10wk

5h

OT⫹PT: PNF emphasis on ADL task; compensatory techniques

1h/d, 3d/wk

10wk

6h

OT: including compensatory techniques; strength; range of motion; traditional positioning

2h/d, 5d/wk

2wk

Abbreviations: ADL, activity of daily living; mCIMT, modified CIMT; NDT, neurodevelopmental treatment; OT, occupational therapy; PNF, proprioceptive neuromuscular facilitation; PT, physical therapy; RT, restrained time; TT, training time; TTT, total treatment time; WMFT, Wolf Motor Function Test.

Effect of Modified CIMT: Kinematic Analysis In evaluating kinematic variables, 3 RCTs10,25,36 reported the data of normalized movement time and percentage of movement time where peak velocity occurs, while reaction time, normalized movement unit, and normalized total displacement were assessed in 2 RCTs,10,36 and 1 trial36 evaluated peak velocity. The forest plot from said article10,25,36 showed the significant difference between the groups was only in the percentage of movement time where peak velocity occurs10,25,36 and in the reaction time10,36 (percentage of movement time where peak velocity occurs, MD⫽7.5, 95% CI, 1.94 –13.05, P⫽.008; reaction time, MD⫽–.23, 95% CI, ⫺.38 to ⫺.08, P⫽.003). There was no statistical significance in 3 of the other measures (table 4). The result of peak velocity from 1 original trial36 revealed that there was no statistical significance between the groups (P⫽.062).

DISCUSSION We performed a comprehensive search according to published recommendations and identified 13 RCTs9,10,12,20,25,33-40 including 278 participants in total. The effectiveness of modified CIMT and TR for upper-extremity impairment was assessed in our study by using clinical variables and kinematic variables. The collective results of 13 RCTs showed that there was statistical significance between the groups in clinical variables, but no significant difference was observed in kinematic variables except for 2 measures. It seems that modified CIMT has some specific advantages such as reducing the level of disability, increasing the use of the more affected upper limb in daily living, and enhancing the motor automaticity of the more affected arm during movement. It is possible that repeated affected limb use can bring about cortical reorganization and then improve motor function and overcome the learned nonArch Phys Med Rehabil Vol 92, June 2011

978

MOVEMENT THERAPY VERSUS TRADITIONAL REHABILITATION POSTSTROKE, Shi

Fig 2. Clinical evaluation: Fugl Meyer Assessment for mCIMT group and TR group. Abbreviations: IV, inverse variance; mCIMT, modified CIMT.

use,12,20 even though the duration of both the practice task of the paretic upper extremity and the restraint time of the nonparetic upper extremity were shorter than the traditional CIMT. The evidence from Sirtori et al16 demonstrated the prescribed duration of the practice task affected the success of therapy. When total practice time was equal to or less than 30 hours, a marginally significant improvement of motor function of the paretic upper extremity was shown, but not when practice time was extended to over 30 hours.16 A survey from a case report suggested that 80% or more patients preferred shorter practice sessions lasting for a longer period to a large amount of training in a short period.39,44 All of these findings showed the merit of modified CIMT. Total treatment time of 3 RCTs33,34,40 included in this systematic review was 2 weeks, and the results from these 3 original studies revealed no significant difference between the groups in the Fugl Meyer Assessment, the Action Research Arm Test, the FIM, or the Motor Activity Log, but meta-analysis showed that betweengroup difference was in favor of the modified CIMT group. Furthermore, it is commonly believed that natural motor recovery occurs within 6 months after stroke. Seven trials20,33,34,37-40 in our study included participants with onset of stroke occurring less than 6 months before. A subgroup analysis of the course of disease and the amount of practice could not be conducted because of the limited number of trials and insufficient data in some trials. We were unable to define which subgroup of participants benefited most from modified CIMT and which prescribed practice task had the best results for patients. It would be desirable to determine the most effective amount of practice task time and the role of modified CIMT in

Fig 3. Clinical evaluation: Action Research Arm Test for mCIMT group and TR group. Abbreviations: IV, inverse variance; mCIMT, modified CIMT.

Arch Phys Med Rehabil Vol 92, June 2011

different stages of stroke with the aim of maximizing therapeutic effect while minimizing adverse reactions. The contents of intervention may be another important factor associated with a better prognosis of patients with upper-limb dysfunction after stroke. As shown in table 3, the content of interventions was very similar in both study groups except for limiting the nonparetic arm in the studies by Hammer and Lindmark.33,34 In contrast, there were large differences in the practice content between the groups in each of the remaining trials. Furthermore, the results from the 2 studies by Hammer and Lindmark33,34 showed that there was no statistical significance in the Fugl Meyer Assessment, the Action Research Arm Test, and the Motor Activity Log between the groups, which were also contrasted with the other included trials. However, the limitations of these 2 studies33,34 must be considered. For instance, these 2 studies were not blinded, and total treatment time was only 2 weeks. Future research would be needed to enhance the methodologic quality and unify the content and intensity of intervention between the groups as much as possible. Kinematic analysis is, indeed, an objective and effective means for quantifying the functionality of upper limbs and evaluating the efficacy of modified CIMT intervention. Kinematic analysis is mainly used for exploring the spatiotemporal characteristics of movement.10,27,45 In kinematic variables, efficiency in movement execution may be characterized by the reaction time and normalized movement time; normalized total displacement and normalized movement unit represent the movement smoothness; peak velocity indicates impulse at movement initiation; percentage of movement time where peak

MOVEMENT THERAPY VERSUS TRADITIONAL REHABILITATION POSTSTROKE, Shi

979

Fig 4. Clinical evaluation: FIM for mCIMT group and TR group. Abbreviations: IV, inverse variance; mCIMT, modified CIMT.

velocity occurs reflects motor control strategy.10,25,27 Three studies10,31,33 reported kinematic variables. Meta-analysis showed statistical significance only in reaction time and percentage of movement time where peak velocity occurred between the groups. These findings indicated that modified CIMT may induce the spontaneity and decrease the online error correction during movement time.10,25,26 These results were consistent with the original studies10,25 included in our review. Meta-analysis showed that there was no statistical significance in other kinematic variables between the groups. Wu et al25 used kinematic variables to estimate the motor characteristics of the bilateral and unilateral upper extremities, respectively. They reported that each kinematic variable of the bilateral upper extremity showed statistical significance between the groups, while there was no statistical significance when the more affected upper extremity executed the movement. Kinematic analysis applied for evaluating the functional recovery of the more affected upper extremity is still in its early stage.

Different studies have different methods to evaluate the kinematic variables. In addition, a small amount of available literature used this outcome measure. Hence, caution is needed in interpreting the results. Higher risk, poor compliance, and lower quality of life were mentioned as flaws of the traditional CIMT in some included studies,9,20,36-38 so the feasibility of traditional CIMT has been questioned in recent years. As we know, modified CIMT was developed on the basis of overcoming shortcomings of the traditional CIMT; however, only 1 trial35 reported a measure of health-related quality of life, and no RCTs reported the safety and compliance in the duration of modified CIMT. It is therefore unknown whether modified CIMT has an effect on these factors. Future studies might refine attention to these aspects. Study Limitations The systematic reviews from Bonaiuti et al17 and Hakkennes and Keating43 reported that clinical homogeneity was poor

Fig 5. Clinical evaluation: Motor Activity Log (AOU and QOU) for modified CIMT group and TR group. Abbreviations: AOU, amount of use; IV, inverse variance; mCIMT, modified CIMT; QOU, quality of use.

Arch Phys Med Rehabil Vol 92, June 2011

980

MOVEMENT THERAPY VERSUS TRADITIONAL REHABILITATION POSTSTROKE, Shi

Fig 6. Sensitivity analysis of Motor Activity Log (AOU and QOU) for mCIMT group and TR group. Abbreviations: AOU, amount of use; IV, inverse variance; mCIMT, modified CIMT; QOU, quality of use.

among their included studies. Thirteen RCTs only about comparing modified CIMT with TR were included in this systematic review, which displayed good clinical homogeneity and high methodologic quality, but as is the case with all systematic reviews, several caveats with this systematic review should be considered. 1. There were small sample sizes within trials and a small number of trials in our study. These characteristics may result from the specific inclusion criteria of modified CIMT. There was insufficiency of raw numeric information in some studies—for example, SD was not available, and charts were used to describe the results. We

contacted the authors, but no reply was obtained, so we could not perform subgroup analysis (eg, according to prescribed practice session a day, time since stroke). Moreover, care must be taken in the interpretation of the statistical heterogeneity. The heterogeneity tests of this study might have low sensitivity because of small sample size of the included studies and limited trials.46 All of these affected the strength of evidence. 2. We did not limit the language in the process of literature retrieval, but only English-language trials were identified according to inclusion criteria, and this might have resulted in language bias. Furthermore, this systematic

Table 4: Meta-Analysis of Outcome Measures Variables

FMA ARA FIM MAL: AOU QOU MAL: AOU QOU PPV NMT RT NMU NTD

No. of Studies

6 5 3 6 6 5 5 3 3 2 2 2

Sample Size (mCIMT/TR)

Heterogeneity Test

Model Selection

116 (59/57) 63 (33/30) 88 (45/43) 173 (87/86)

I ⫽12%; P⫽.34 I2⫽20%; P⫽.29 I2⫽0%; P⫽.85 I2⫽90%; P⬍.01 I2⫽71%; P⬍.01 I2⫽31%; P⫽.22 I2⫽36%; P⫽.18 I2⫽36%; P⫽.21 I2⫽0%; P⫽.55 I2⫽0%; P⫽.52 I2⫽0%; P⫽.55 I2⫽82%; P⫽.02

Fixed Fixed Fixed Random

163 (82/81) 109 (56/53) 109 (56/53) 79 (41/38) 79 (41/38) 77 (39/38)

2

Fixed Fixed Fixed Fixed Fixed Random

MD (95% CI)

P

7.80 (4.21 to 11.38) 14.15 (10.71 to 17.59) 7.00 (0.75 to 13.26) 1.09 (0.26 to 1.91) 1.02 (0.55 to 1.45) 0.78 (0.37 to 1.19) 0.79 (0.42 to 1.25) 7.50 (1.94 to 13.05) 0.00 (⫺0.01 to 0.01) ⫺0.23 (⫺0.38 to ⫺0.08) 0.10 (⫺0.37 to 0.57) ⫺0.17 (⫺0.60 to 0.26)

.000 .000 .030 .010 .000 .000 .000 .008 .990 .003 .680 .440

Abbreviations: AOU, Amount of Use; ARA, Action Research Arm Test; FMA, Fugl Meyer Assessment; MAL, Motor Activity Log; mCIMT, Modified CIMT; NMT, normalized movement time; NMU, normalized movement unit; NTD, normalized total displacement; PPV, percentage of movement time where peak velocity occurs; QOU, Quality of Use; RT, reaction time.

Arch Phys Med Rehabil Vol 92, June 2011

MOVEMENT THERAPY VERSUS TRADITIONAL REHABILITATION POSTSTROKE, Shi

review may be biased because the literature search failed to identify all relevant studies (positive results are more likely to be published than negative results; larger studies are more likely to report positive results than smaller studies). Moreover, there were still shortcomings of methodologic quality in the included studies. Four trials25,37-39 failed to describe the randomization process; allocation concealment was unclear in most studies; there was partial use of blinding. Selection bias, performance bias, and measuring bias may be caused by these factors. However, in order to reduce the likelihood of bias, we conducted a well defined, comprehensive literature search by consulting relevant reviews and performed data extraction and quality assessment with 2 independent authors. In addition, we searched unpublished studies through tracking the SIGLE database in order to reduce publication bias as far as possible. 3. Three trials20,37,39 did not report the comparability of the baseline, and 2 partially reported12,38 the baseline. We calculated the baseline from available data (eg, age, sex, time since stroke, side of affected arm) in these trials, and no significant difference was found between the groups. CONCLUSIONS Modified CIMT is a feasible alternative intervention for patients with upper-extremity dysfunction after stroke because the current study revealed that compared with TR, modified CIMT could reduce the level of disability, improve the ability to use the paretic upper extremity, and increase the use of the paretic upper limb in daily living. However, evidence is still limited about the effectiveness of modified CIMT in kinematic analysis because of the paucity of RCTs. Multicenter and well designed RCTs are necessary to overcome the shortcomings of the current trials (eg, small sample size, poor methodologic quality). Comprehensive and valid outcome measures should be monitored, which include not only clinical and kinematic variables but also compliance of patients for modified CIMT and health-related quality of life of patients with stroke. Acknowledgments: We acknowledge Camden Garrett, MA, Professor, Department of Foreign Language, Tianjin Medical University, for writing assistance. 1. 2. 3. 4. 5.

6.

7.

8.

9.

References Hachinski V. Stroke in Chinese. Stroke 2006;37:1347. Michael KM, Shaughnessy M. Stroke prevention and management in older adults. J Cardiovasc Nurs 2006;21:21-6. Frizzell JP. Acute stroke: pathophysiology, diagnosis, and treatment. AACN Clin Issues 2005;16:421-40. Zhu YL. Neurological rehabilitation. Beijing: People’s Military Medical Publisher; 2003. p 325. Broeks JG, Lankhorst GJ, Rumping K, Prevo AJH. The long term outcome of arm function after stroke: results of a follow up study. Disabil Rehabil 1999;21:357-64. Taub E, Crago JE. Uswatte G. Constraint-induced movement therapy: a new approach to treatment in physical rehabilitation. Rehabil Psychol 1998,43:152-70. Van der Lee JH, Wagenaar RC, Lankhorst GJ, Vogelaar TW, Deville WL, Bouter LM. Forced use of the upper extremity in chronic stroke patients: results from a single-blind randomized clinical trial. Stroke 1999;30:2369-75. Taub E, Miller NE, Novack TA, et al. Technique to improve chronic motor deficit after stroke. Arch Phys Med Rehabil 1993; 74:347-54. Page SJ, Levine P, Leonard A, Szaflarski JP, Kissela BM. Modified constraint-induced therapy in chronic stroke: results of a

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22. 23.

24.

25.

26.

27.

28.

981

single-blinded randomized controlled trial. Phys Ther 2008;88:333-40. Lin KC, Wu CY, Wei TH, Gung C, Lee CY, Liu JS. Effects of modified constraint induced movement therapy on reach-to-grasp movements and functional performance after chronic stroke: a randomized controlled study. Clin Rehabil 2007;21:1075-86. Sunderland A, Tuke A. Neuroplasticity, learning and recovery after stroke: a critical evaluation of constraint-induced therapy. Neuropsychol Rehabil 2005;15:81-96. Page SJ, Sisto S, Levine P, McGrath RE. Efficacy of modified constraint-induced movement therapy in chronic stroke: a singleblinded randomized controlled trial. Arch Phys Med Rehabil 2004;85:14-8. Suputtitada A, Suwanwela NC, Tumvitee S. Effectiveness of constraint-induced movement therapy in chronic stroke patients. Med Assoc Thai 2004;87:1482-90. Boake C, Noser EA, Ro T, et al. Constraint-induced movement therapy during early stroke rehabilitation. Neurorehabil Neural Repair 2007;21:14-24. Dahl AE, Askim T, Stock R, Langørgen E, Lydersen S, Indredavik B. Short and long-term outcome of constraint-induced movement therapy after stroke: a randomized controlled feasibility trial. Clin Rehabil 2008;22:436-47. Sirtori1 V, Corbetta1 D, Moja L, Gatti R. Constraint-induced movement therapy for upper extremities in stroke patients. Cochrane Database Syst Rev 2009;(4):1-35. Bonaiuti D, Rebasti L, Sioli P. Constraint-induced movement therapy: a systematic review of randomised controlled trials. Eura Medicophys 2007;43:139-42. Morris DM, Uswatte G, Crago JE, Cook EW, Taub E. The reliability of the wolf motor function test for assessing upper extremity function after stroke. Arch Phys Med Rehabil 2001;82: 750-5. Page SJ, Levine P, Sisto S, Bond Q, Johnston MV. Stroke patients’ and therapists’ opinions of constraint-induced movement therapy. Clin Rehabil 2002;16:55-60. Page SJ, Levine P, Leonard AC. Modified constraint-induced therapy in acute stroke: a randomized controlled pilot study. Neurorehabil Neural Repair 2005;19:27-32. Page SJ, Sisto S, Johnston MV, Levine P, Hughes M. Modified constraint-induced therapy in stroke: a case study. Arch Phys Med Rehabil 2002;83:286-90. Bi S, Ji SR, Gu Y, Weng CS. Study of validity of motor status scale. Chin J Rehabil Theory Pract 2007;13:114-6. Lang CE, Wagner JM, Dromerick AW, Edwards DF. Measurement of upper extremity function early after stroke: properties of the action research arm test. Arch Phys Med Rehabil 2006;87: 1605-10. Hsieh CL, Hsueh IP, Chiang FM, Lin PH. Inter-rater reliability and validity of the Action Research Arm Test in stroke patients. Age Ageing 1998;27:107-13. Wu CY, Lin KC, Chen HC, Chen IH, Hong WH. Effects of modified constraint-induced movement therapy on movement kinematics and daily function in patients with stroke: a kinematic study of motor control mechanisms. Neurorehabil Neural Repair 2007;21:460-6. Caimmi M, Carda S, Giovanzana C, et al. Using kinematic analysis to evaluate constraint-induced movement therapy in chronic stroke patients. Neurorehabil Neural Repair 2008;22:31-9. Wu CY, Trombly CA, Lin KC, Tickle-Degnen L. A kinematic study of contextual effects on reaching performance in persons with and without stroke: influences of object availability. Arch Phys Med Rehabil 2000;81:95-101. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of repots of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17:1-12. Arch Phys Med Rehabil Vol 92, June 2011

982

MOVEMENT THERAPY VERSUS TRADITIONAL REHABILITATION POSTSTROKE, Shi

29. Brogårdh C, Flansbjer UB, Lexell J. What is the long-term benefit of constraint-induced movement therapy? a four-year follow-up. Clin Rehabil 2009;23:418-23. 30. Fritz SL, Light KE, Patterson TS, Behrman AL, Davis SB. Active finger extension predicts outcomes after constraint-induced movement therapy for individuals with hemiparesis after stroke. Stroke 2005;36:1172-7. 31. Lin KC, Wu CY, Liu JS, Chen YT, Hsu CJ. Constraint-induced therapy versus dose-matched control intervention to improve motor ability, basic/extended daily functions, and quality of life in stroke. Neurorehabil Neural Repair 2009;23:160-5. 32. Yen JG, Wang RY, Chen HH, Hong CT. Effectiveness of modified constraint-induced movement therapy on upper limb function in stroke subjects. Acta Neurol Taiwan 2005;14:16-20. 33. Hammer AM, Lindmark B. Is forced use of the paretic upper limb beneficial? a randomized pilot study during subacute post-stroke recovery. Swed Clin Rehabil 2009;23:424-33. 34. Hammer AM, Lindmark B. Effects of forced use on arm function in the subacute phase after stroke: a randomized, clinical pilot study. Phys Ther 2009;89:526-39. 35. Wu CY, Chen CL, Tsai WC, Lin KC, Chou SH. A randomized controlled trial of modified constraint-induced movement therapy for elderly stroke survivors: changes in motor impairment, daily functioning, and quality of life. Arch Phys Med Rehabil 2007;88: 273-8. 36. Wu CY, Chen CL, Tang SF, Lin KC, Huang YY. Kinematic and clinical analyses of upper-extremity movements after constraintinduced movement therapy in patients with stroke: a randomized controlled trial. Arch Phys Med Rehabil 2007;88:964-70. 37. Atteya AA. Effects of modified constraint induced movement therapy on upper limb function in subacute stroke patients. Neurosciences 2004;9:24-9. 38. Page SJ, Sisto S, Johnston MV, Levin P. Modified constraintinduced therapy after subacute stroke: a preliminary study. Neurorehabil Neural Repair 2002;16:290-5.

Arch Phys Med Rehabil Vol 92, June 2011

39. Page SJ, Sisto SA, Levine P, Johnston MV. Modified constraint induced therapy: a randomized feasibility and efficacy study. J Rehabil Res Dev 2001;38:583-90. 40. Dromerick AW, Edwards DF, Hahn M. Does the application of constraint induced movement therapy during acute rehabilitation reduce arm impairment after ischemic stroke? Stroke 2000;31: 2984-8. 41. Ferraro M, Demaio JH, Krol J, et al. Assessing the motor status score: a scale for the evaluation of upper limb motor outcomes in patients after stroke. Neurorehabil Neural Repair 2002;16:283-9. 42. Hakim RM, Kelly SJ, Grant-Beuttler M, Healy B, Krempasky J, Moore S. Case report: a modified constraint-induced therapy (mCIT) program for the upper extremity of a person with chronic stroke. Physiother Theory Pract 2005;21:243-56. 43. Hakkennes S, Keating JL. Constraint-induced movement therapy following stroke: a systematic review of randomized controlled trials. Aust J Physiother 2005;51:221-31. 44. Blanton S, Wolf SL. An application of upper extremity constraintinduced movement therapy in a patient with subacute stroke. Phys Ther 1999;79:847-53. 45. Zhang YY, Cai WZ, Zhong ZY, et al. The effects of modified constraint-induced therapy and bilateral training for stroke patients: kinematic analysis. Rehabil Med Taiwan 2009;37:19-30. 46. Higgins JPT, Green S, editors. Cochrane handbook for systematic reviews of interventions version 5.0.2 [updated September 2009].The Cochrane Collaboration. 2008. Available at: http:// www.cochrane-handbook.org. Accessed May 20, 2010. Suppliers a. Review Manager Software (RevMan5.0); Cochrane Collaboration. Available at: http://ims.cochrane.org/RevMan. b. Endnote software. Available at: http://www2.piaodown.com/down/ soft/19233.htm.