de Jong et al: Combined arm positioning and NMES poststroke

Combined arm stretch positioning and neuromuscular

electrical stimulation during rehabilitation does not improve
range of motion, shoulder pain or function in patients after
stroke: a randomised trial
1,2 2,3 4 5
Lex D de Jong , Pieter U Dijkstra , Johan Gerritsen , Alexander CH Geurts and Klaas
2
Postema
1
School of Physiotherapy, Hanze University of Applied Sciences, Groningen, 2Department of Rehabilitation Medicine, University of Groningen,
University Medical Center Groningen, Groningen, 3Department of Oral and Maxillofacial Surgery, University of Groningen, University Medical Center
Groningen, 4ViaReva, Center for Rehabilitation, Apeldoorn, 5Department of Rehabilitation, Radboud University Nijmegen Medical Center, Nijmegen The
Netherlands

Question: Does static stretch positioning combined with simultaneous neuromuscular electrical stimulation (NMES) in
the subacute phase after stroke have beneficial effects on basic arm body functions and activities? Design: Multicentre
randomised trial with concealed allocation, assessor blinding, and intention-to-treat analysis. Participants: Forty-six people in
the subacute phase after stroke with severe arm motor deficits (initial Fugl-Meyer Assessment arm score ) 18). Intervention:
In addition to conventional stroke rehabilitation, participants in the experimental group received arm stretch positioning
combined with motor amplitude NMES for two 45-minute sessions a day, five days a week, for eight weeks. Control
participants received sham arm positioning (ie, no stretch) and sham NMES (ie, transcutaneous electrical nerve stimulation
with no motor effect) to the forearm only, at a similar frequency and duration. Outcome measures: The primary outcome
measures were passive range of arm motion and the presence of pain in the hemiplegic shoulder. Secondary outcome
measures were severity of shoulder pain, restrictions in performance of activities of daily living, hypertonia, spasticity, motor
control and shoulder subluxation. Outcomes were assessed at baseline, mid-treatment, at the end of the treatment period (8
weeks) and at follow-up (20 weeks). Results: Multilevel regression analysis showed no significant group effects nor significant
time × group interactions on any of the passive range of arm motions. The relative risk of shoulder pain in the experimental
group was non-significant at 1.44 (95% CI 0.80 to 2.62). Conclusion: In people with poor arm motor control in the subacute
phase after stroke, static stretch positioning combined with simultaneous NMES has no statistically significant effects on range
of motion, shoulder pain, basic arm function, or activities of daily living. 5SJBM SFHJTUSBUJPO: NTR1748. <EF +POH -%
%JKLTUSB 16
(FSSJUTFO +
(FVSUT "$)
1PTUFNB , ﾵ $PNCJOFE BSN
stretch positioning and neuromuscular electrical stimulation during rehabilitation does not improve range of motion, TIPVMEFS QBJO PS
GVODUJPO JO QBUJFOUT BGUFS TUSPLF B SBOEPNJTFE USJBM ﾤ Journal of Physiotherapy ﾧ o ﾧ>

Keywords: Stroke, Upper extremity, Muscle stretching exercises, Electrical stimulation, Activities of daily living,
Randomized controlled trial

Introduction (Coupar et al 2010) or constraint-induced movement therapy

(Sirtori et al 2009). However, these interventions are not
Annually, 15 million people worldwide suffer a stroke suitable for people with severe motor deficits because they
(Mackay and Mensah 2004). About 77–81% of stroke require ‘active’ residual arm motor capacity. For these people
survivors show a motor deficit of the extremities (Barker ‘passive’ interventions may be needed to prevent secondary
and Mullooly 1997). In almost 66% of patients with an impairments and optimise long-term handling
initial paralysis, the affected arm remains inactive and
immobilised due to a lack of return of motor function after
six months (Sunderland et al 1989, Wade et al 1983). Over What is already known on this topic: Contracture of
time, the central nervous system as well as muscle tissue muscles in the arm after stroke is common. Stretch
of the arm adapt to this state of inactivity, often resulting alone does not typically produce clinically important
in residual impairments such as hypertonia (de Jong et al reductions in contracture in people with neurological
2011, van Kuijk et al 2007), spasticity (O’Dwyer et al conditions. Hypertonia may limit the application of
1996) or contractures (Kwah et al 2012, O’Dwyer et al stretch and therefore its potential benefits.
1996, Pandyan et al 2003). In turn, these secondary What this study adds: In people with poor arm motor
impairments are associated with hemiplegic shoulder pain control after stroke, static arm positioning to stretch
(Aras et al 2004, Roosink et al 2011) and restrictions in muscles prone to contracture combined with
neuromuscular stimulation of the antagonist muscles
performance of activities of daily living (Lindgren et al did not have significant benefits with respect to range
2007, Lundström et al 2008). of motion, shoulder pain, performance of activities of
daily living, hypertonia, spasticity, motor control or
Several interventions improve arm function after stroke and shoulder subluxation.
prevent secondary impairments, eg, bilateral arm training
Journal of Physiotherapy 2013 Vol. 59 – ©
Australian Physiotherapy Association 2013 . Open access under CC BY-NC-ND license. 245
Research

and assistive use of the affected arm. It is also important to Method

elicit muscle activity if at all possible, and to improve arm
function. To prevent the loss of passive range of joint Design
motion as a result of contracture of at-risk muscles in the A multicentre, assessor-blinded, randomised controlled trial
shoulder (eg, internal rotators, adductors) and forearm (eg, was conducted. After inclusion, participants were randomised
pronators, wrist and finger flexors) in particular, the in blocks of four (2:2 allocation ratio) in two strata (Fugl-
application of arm stretch positioning alongside regular Meyer Assessment arm score 0–11 points and 12–18 points)
physiotherapy was deemed important (Ada and Canning at each treatment centre. Opaque, sealed envelopes
1990), especially because contractures are associated with containing details of group allocation were prepared by the
shoulder pain (Aras et al 2004, de Jong et al 2007, main co-ordinator (LDdJ) before trial commencement. After a
Wanklyn et al 1996). However, in general, passive stretch local trial co-ordinator had determined eligibility and
does not produce clinically important changes in joint obtained a patient’s consent, the main co-ordinator was
range of motion, pain, spasticity, or activity limitations contacted by phone. He instructed an independent person to
(Katalinic et al 2011). One explanation for the lack of draw an envelope blindfolded and to communicate the result
effect of passive stretch of the shoulder muscles could be back to the local trial co-ordinator. The local trial co-
the inadequate duration of stretch, with clinical trials using ordinator then made arrangements for the baseline
a dose of 20 or 30 minutes only (Borisova and Bohannon measurement after which the allocated intervention was
2009). However, it is questionable whether stretch of the initiated. Mid-treatment, end-treatment, and follow-up
shoulder muscles for much more than 60 minutes per day measurements took place at 4, 8, and 20 weeks after baseline
during intensive rehabilitation programs is feasible (Turton measurement by two independent assessors
and Britton 2005). (physiotherapists), who were unaware of group allocation and
not involved in the treatment of participants. To keep the
People with severe motor deficits after stroke have a assessors blinded, participants were reminded before each
higher risk of developing increased resistance to passive measurement not to reveal the nature of their treatment.
muscle stretch (hypertonia) and spasticity of the muscles Participants were considered to be unaware of group
responsible for an antigravity posture (de Jong et al 2011, allocation because they were informed about the existence of
Kwah et al 2012, Urban et al 2010). These muscles are two intervention groups but not about the study hypothesis.
also at risk of developing contracture. As a result, the The participants’ and assessors’ beliefs regarding allocation
passive range of the hemiplegic shoulder (exteral rotation, were checked at the eight-week (ie, end of treatment)
flexion and abduction), elbow (extension), forearm assessment using a three-point nominal scale (I suspect
(supination) and wrist (extension) can become restricted. allocation to experimental/control group, I have no clue of
group allocation). All investigators, staff, and participants
Stretching hypertonic muscles is difficult when they are
were kept blinded with regard to the outcome measurements.
not sufficiently relaxed. Cyclic neuromuscular electrical
stimulation (NMES) (Chae et al 2008), another example of
a ‘passive’ intervention, can not only be used to improve Participants
pain-free range of passive humeral lateral rotation (Price
and Pandyan 2000), but also to reduce muscle resistance Between August 2008 and September 2010, consecutive
(King 1996) and glenohumeral subluxation (Pomeroy et al newly admitted patients on the neurological units of three
2006, Price and Pandyan 2000). From these results we rehabilitation centres in the Netherlands (Beetsterzwaag,
hypothesised that NMES of selected arm muscles opposite Doorn, and Zwolle) were approached for participation.
to muscles that are prone to the development of spasticity Willing patients were initially screened by a physician for
and contracture might facilitate static arm stretching both the following inclusion criteria: first-ever or recurrent
through reciprocal inhibition (‘relaxation’) of antagonist stroke (except subarachnoid haemorrhages) between two
muscles (Alfieri 1982, Dewald et al 1996, Fujiwara et al and eight weeks poststroke; age > 18 years; paralysis or
2009) and the imposed (cyclic) stretch caused by motor severe paresis of the affected arm scoring 1–3 on the
amplitude NMES. Consequently, static arm stretch recovery stages of Brunnstrom (1970); and no planned
positioning combined with NMES could potentially result date of discharge within four weeks. Subsequently, a local
in larger improvements of arm passive range of motion trial co-ordinator excluded patients with: contraindications
and less (severe) shoulder pain compared to NMES or for electrical stimulation (eg, metal implants, cardiac
static stretching alone. From these hypotheses we pacemaker); pre-existing impairments of the affected arm
developed the following research questions: (pre-existing contracture was not an exclusion criterion);
1. Does eight weeks of combined static arm stretch severe cognitive deficits and/or severe language
positioning with simultaneous NMES prevent the comprehension difficulties, defined as < 3/4 correct verbal
loss of shoulder passive range of motion and the responses and/or < 3 correct visual graphic rating scale
occurrence of shoulder pain more than sham stretch scores on the AbilityQ (Turner-Stokes and Rusconi 2003);
positioning with simultaneous sham NMES (ie, and moderate to good arm motor control (> 18 points on
transcutaneous electrical stimulation, TENS) in the the Fugl-Meyer Assessment arm score).
subacute phase of stroke?
2. Does the experimental intervention have any additional
Interventions
effects on timing and severity of shoulder pain, All participants received multidisciplinary stroke
restrictions in daily basic arm activities, resistance to rehabilitation, ie, daily training in activities of daily living by
passive stretch (hypertonia) and spasticity, arm motor rehabilitation nurses, occupational therapists,
control, and the degree of shoulder subluxation? physiotherapists, and speech therapists. These interventions
were not standardised, but generally administered in a way
that was consistent with the recommendations of

246 Journal of Physiotherapy 2013 Vol. 59 – © Australian Physiotherapy Association 2013 . Open access under CC BY-NC-ND license.
 de Jong et al: Combined arm positioning and NMES poststroke

A B C

Figure 1. Experimental and control arm muscle stretch positions and electrode placements. (a) The intervention used by
experimental group participants with sufficient shoulder external rotation to achieve the position. (b) The intervention used by
experimental group participants with insufficient shoulder external rotation. (c) The control (ie, sham) intervention.

the Dutch stroke guidelines (Van Peppen et al 2004). the same two trained assessors. Every effort was made to
Participants were requested to undergo the additional motivate participants to undergo all planned measurements
allocated treatment twice daily for 45 minutes on even after withdrawal from the study.
weekdays for 8 weeks. Participants from the experimental
group received arm stretch positioning (presented in Passive range of shoulder external rotation, flexion and
Figures 1a and 1b) with simultaneous four-channel motor abduction, elbow extension, forearm supination, wrist
amplitude NMES. Participants from the control group extension with extended and flexed fingers were assessed
received a sham stretch positioning procedure (presented because these movements often develop restrictions in range
in Figure 1c) with simultaneous sham conventional TENS as a result of imposed immobility, with muscle contractures
with minimal sensory sensation by using a similar causing a typical flexion posture of the hemiplegic arm. The
treatment protocol, electrical stimulator and electrode (entire) ShoulderQ was administered in participants who
placement (but on the forearm only) as the experimental indicated that they had shoulder pain. This questionnaire
group. A detailed description of the experimental and assesses timing and severity of pain by means of eight verbal
control group procedures can be found in Appendix 1 (see questions and three vertical visual graphic rating scales. We
the eAddenda for Appendix 1). were primarily interested in the answer to the (verbal)
question How severe is your shoulder pain overall? (1= mild,
Treatment was planned to result in 60 hours of positioning 2 = moderate, 3 = severe, 4 = extremely severe) and pain
and 51 hours of NMES/TENS. All procedures were severity measured at rest, on movement, and at night using the
performed by the local trial coordinator or instructed 10-cm vertical visual graphic rating scales. The ShoulderQ is
nursing staff. Nursing staff monitored compliance to the sensitive (Turner-Stokes and Jackson 2006) and responsive to
intervention by logging each session on a record sheet, change in pain experience (Turner-Stokes and Rusconi 2003).
which was always kept in the vicinity of the participant’s Performance of basic functional activities of daily life
bed. During the first 8 weeks of the trial, prescription of involving the passive arm was assessed using the Leeds
pain and spasticity medication as well as content of Adult/Arm Spasticity Impact Scale (Ashford et al 2008).
physical and occupational therapy sessions for the arm Using this semi-structured interview, participants were asked
were also monitored. to indicate whether they or their carer(s) experienced
difficulty performing 12 different tasks involving the
Outcome measures hemiplegic arm (cleaning the palm/ elbow/armpits, cutting
The primary outcome measures were passive range of arm fingernails, putting the arm through a sleeve/in a glove,
motion and pain in the hemiplegic shoulder. All goniometric rolling over in bed, doing exercises, balancing while
assessments were performed by two observers using a fluid- standing/walking, and holding objects). The scores on the
a separate items (1 point = no difficulty, 0 = difficulty or activity
filled goniometer . Inter-observer reliability of this technique
was high (de Jong et al 2012). The presence of shoulder pain not yet performed) were summed, divided by the total number
was checked using the first (yes/no) question of the of items performed and multiplied by 100, resulting in a
ShoulderQ (Turner-Stokes and Jackson 2006). The secondary summary score (0 = severe disability, 100 = no disability).
outcome measures were timing and severity of poststroke Hypertonia and spasticity of the shoulder internal rotators,
shoulder pain, performance of real-life passive and basic daily elbow flexors, and long finger flexors were assessed using a
active arm activities, hypertonia and spasticity, arm motor detailed version (Morris 2002) of the Tardieu Scale (Held and
control and shoulder subluxation. All measurements were Pierrot-Deseilligny 1969). The Tardieu Scale can differentiate
carried out in the same fixed order by spasticity from

Journal of Physiotherapy 2013 Vol. 59 – © Australian Physiotherapy Association 2013 . Open access under CC BY-NC-ND license. 247
Research

Screened for eligibility (n = 260)

Excluded after initial screening (n = 180)a Excluded after inclusion testing (n = 32)
t OP JTDIBFNJDIBFNPSSIBHJD TUSPLF O - t unable to fill out, read or understand the AbilityQ (n = 9)
t > 8 weeks post stroke (n = 28) t Fugl-Meyer Assessment Arm score > 18 points (n = 14)
t Brunnstrom’s recovery stage * 4 (n = 169) t contraindications to electrical stimulation (n = 1)
t pre-existing arm impairments (n = 24) t other reasons (n = 8)
t planned date of discharge too soon (n = 64)
t SFGVTFEVOBCMF UP QBSUJDJQBUF O -
t other (n = 8)b
t VOLOPXONJTTJOH EBUB O - ﾧ

Measured passive range of motion, shoulder pain, restrictions in performance of activities

Week 0 PG EBJMZ MJGF
IZQFSUPOJBTQBTUJDJUZ
NPUPS DPOUSPM BOE TIPVMEFS TVCMVYBUJPO
(n = 24) Randomised (n = 24)
(n = 23) Analysedc (n = 23)

Experimental group Control group t

Discontinued Lost to follow-up t
t multidisciplinary stroke multidisciplinary stroke
intervention readmission to
rehabilitation rehabilitation t sham arm
t shoulder
t static arm positioning positioning hospitald (n = 1)
pain (n = 2) Discontinued
and NMES 90 min per and TENS 90 min per
intervention
day, 5 days per week day, 5 days per week
t discharge (n = 1)

Measured passive range of motion, shoulder pain, restrictions in performance of activities

Week 4 PG EBJMZ MJGF
IZQFSUPOJBTQBTUJDJUZ
NPUPS DPOUSPM BOE TIPVMEFS TVCMVYBUJPO
(n = 21) Received prescribed intervention (n = 21)
(n = 23) Analysed (n = 21)e

Experimental group Control group t

Lost to follow-up Discontinued
t multidisciplinary stroke multidisciplinary stroke
t death (n = 1) t intervention
rehabilitation rehabilitation t sham arm
increased t forearm pain (n = 1)
t static arm positioning positioning
shoulder pain and NMES 90 min per and TENS 90 min per t recurrent stroke
(n = 1) day, 5 days per week day, 5 days per week (n = 2)
Discontinued t discharge (n = 1)
intervention
t discharge (n = 1)

Measured passive range of motion, shoulder pain, restrictions in performance of activities

Week 8 PG EBJMZ MJGF
IZQFSUPOJBTQBTUJDJUZ
NPUPS DPOUSPM BOE TIPVMEFS TVCMVYBUJPO
(n = 18) Received prescribed intervention (n = 17)
(n = 21) Analysed (n = 21)f
Experimental group Control group
Lost to follow-up
t no intervention t no intervention
t severe shoulder
subluxation (n = 1)
t did not attend
(n = 3)

Measured passive range of motion, shoulder pain, restrictions in performance of activities of

Week 20 EBJMZ MJGF
IZQFSUPOJBTQBTUJDJUZ
NPUPS DPOUSPM BOE TIPVMEFS TVCMVYBUJPO
(n = 17) Analysed (n = 22)

Figure 2. Design and flow of participants through the trial. aAll reasons for exclusion are listed where patients were ineligible
for multiple reasons. bIncluding multiple sclerosis, Alzheimer’s disease, locked-in syndrome, recurrent stroke, and participation
in another trial. NMES = neuromuscular electrical stimulation. cOne participant from each group dropped out after
randomisation but before receiving any intervention. dUnrelated to stroke. eOne participant missed the Week 4 assessment
due to poor weather. fOne participant missed the Week 8 assessment due to recurrent stroke but was subsequently available
for the Week 20 follow-up assessment.

248 Journal of Physiotherapy 2013 Vol. 59 – © Australian Physiotherapy Association 2013 . Open access under CC BY-NC-ND license.
 de Jong et al: Combined arm positioning and NMES poststroke

5BCMF ﾤ Baseline characteristics of participants and datasets, we aimed to recruit at least 20 participants per
centres. group.
Characteristic Exp Con All participants minus two premature dropouts were analysed
(n = 23) (n = 23) as randomised (intention-to-treat). Arm passive range of
Age (yr), mean (SD) 56.6 (14.2) 58.4 (9.6) motion was analysed using a multilevel regression analysis.
Time post-stroke at 43.7 (13.3) 43.3 (15.5) As main factors time (baseline, 4, 8, and 20 weeks), group
baseline (days), allocation (2 groups) and time × group interaction were
mean (SD) explored using the –2log-likelihood criterion for model fit, as
MMSEa, median 27 (23 to 28.25) 28 (26 to 29.5) well as random effects of intercept and slope. For
(IQR) completeness, this analysis was repeated using the data of the
Gender, n males 15 (65) 12 (52) participants including the two premature dropouts (n
(%) = 48) using the last observation carried forward approach.
Stroke type, n (%) Nominal outcome measures (presence of hypertonia/
ICVA 19 (83) 18 (78) spasticity and subluxation) at eight weeks were analysed
HCVA 4 (17) 5 (22) using a Chi-square test. Ordinal outcome measures (Fugl-
Affected 12 (52) 8 (35) Meyer Assessment, Leeds Adult/Arm Spasticity Impact
hemisphere, n right Scale, ShoulderQ) were first analysed for time effects
(%) within subjects using the Friedman test. If differences over
Aphasia, n (%) 5 (22) 6 (26) time (from baseline to follow-up) were found, these were
further explored using the Wilcoxon signed-rank test with
Initial FMA arm
Bonferroni-Hochberg correction (Norman and Streiner
score, n (%)
2000). Between-group differences were analysed using a
0–11 points 19 (83) 17 (74)
Mann-Whitney U test only at 8 weeks to avoid multiple
12–18 points 4 (17) 6 (26) testing.
Centres,
participants Results
treated, n (%)
Beetsterzwaag 7 (30) 8 (35) Flow of participants through the trial
Doorn 4 (17) 4 (17) The flow of participants through the trial is presented in
Zwolle 12 (52) 11 (48) Figure 2. Forty-eight patients met all eligibility criteria.
Exp = experimental group, Con = control group, FMA = Fugl- One participant from the experimental group (a 68-year-
Meyer Assessment arm score, HCVA = haemorrhagic old female with a right-sided ischaemic stroke who
cerebrovascular accident, ICVA = ischaemic cerebrovascular regretted participation) and one from the control group (a
accident, MMSE = Mini Mental State Examination. aNot
administered in subjects with aphasia.
62-year old male with a left-sided ischaemic stroke who
was rehospitalised due to acute liver and kidney failure)
dropped out the day after baseline measurement and before
receiving any intervention. These participants were not
contracture (Haugh et al 2006, Patrick and Ada 2006) and included in the analyses because their data were missing
has fair to excellent test-retest reliability and inter-observer due to unavailability for further measurements.
reliability (Paulis et al 2011). The mean angular velocity of
the Tardieu Scale’s fast movement was standardised (see Of the 11 patients who were lost to follow-up or
the eAddenda for Appendix 2). Muscle reaction quality discontinued their prescribed intervention during the 8-
scores * 2 were considered to be clinically relevant week treatment period, four (36%) complained of pain.
hypertonia. Spasticity was deemed present if the angle of Baseline characteristics of the 46 participants analysed are
catch was present and occurred earlier in range than the shown in Table 1. Twenty-two participants (51%, n = 43)
maximal muscle length after slow stretching (ie, spasticity had no clue as to which group they were allocated, but 17
angle > 0 degs). Arm motor control was assessed using the participants (40%) were correct in their belief regarding
66-point arm section of the Fugl-Meyer Assessment (Fugl- allocation. The three participants who were lost to follow-
Meyer et al 1975, Gladstone et al 2002). Shoulder inferior up before 8 weeks did not provide data about allocation
subluxation was diagnosed by palpation (Bohannon and beliefs. The two assessors had no clue regarding group
Andrews 1990) in finger breadths (< ½, < 1, * 1, > 1½) allocation in 67% and 72% of the cases. They were correct
and considered present if it was one category higher than in their belief regarding allocation in 9 (21%) and 4 (9%)
on the nonaffected side. of the participants, respectively.

Data analysis Co-interventions and compliance with trial

method
Sample size calculation was based on a reliably assessable
change in passive shoulder external rotation range of motion In the experimental group more participants were prescribed
of * 17 degs (de Jong et al 2012). The clinically relevant pain and spasticity medication, as presented in Table 2. They
difference between the experimental and control intervention also received slightly more conventional therapy for the arm
was therefore set at a minimum of 20 deg. The standard and adhered less to the prescribed intervention protocol.
deviation was considered to be 21.5 deg (Ada et al 2005). Overall, compliance in the experimental group was 68%
Alpha was set at 5% (two-sided), beta at 80%. Thus, the (stretch positioning) and 67% (NMES), compared to 78%
required number of participants in each group was 18. (sham positioning) and 75% (TENS) in the control group.
Anticipating a 10% drop-out rate and requiring 36 complete Non-compliance was mainly caused by drop-out and

Journal of Physiotherapy 2013 Vol. 59 – © Australian Physiotherapy Association 2013 . Open access under CC BY-NC-ND license. 249
Research

5BCMF . Mean (SD) or number of participants (%) for co-interventions and compliance to the intervention protocol during
the eight-week intervention period and mean difference (MD) or percentage risk difference (RD) between groups, with 95%
confidence intervals (95% CI).

Outcome Groups Difference between groups (95% CI)

Exp Con
(n = 23) (n = 23)
Prescription of pain medication, n (%) 16 (73)a 11 (48) RD 25% (–4% to 50%)
Prescription of spasticity medication, n (%) 5 (23)a 2 (9) RD 14% (–8% to 36%)
Upper limb occupational therapy (hr), mean (SD) 5 (4)a 4 (4)a MD 1 (–2 to 3)
Upper limb physiotherapy (hr), mean (SD) 3 (5) 2 (3)a MD 1 (–2 to 3)
Total of positioning (hr), mean (SD) 41 (17)a 47 (16) MD –6 (–15 to 4)
Total of electrical stimulation (hr), mean (SD) 34 (16)a 38 (14) MD –4 (–13 to 5)
Exp = experimental group, Con = control group. aData missing for one participant.

early weekend leaves. All mentioned differences between Discussion

the groups were not statistically significant.
To our knowledge this is the first study to analyse the effects
Effect of intervention of a daily arm stretch positioning procedure combined with
All primary and secondary outcome measures are simultaneous NMES in patients with a poor prognosis for
presented in Tables 3, 4 and 5. Individual participant data functional recovery in the subacute phase after stroke. The 8-
are presented in Table 6 (see eAddenda for Tables 4, 5 and week high-intensity multimodal intervention did not result in
6). Except for elbow extension and the control any significant differences in arm passive range of motion
participants’ wrist extension with extended fingers, both (contractures), shoulder pain, basic arm activities,
groups showed reductions in mean passive range of hypertonia/spasticity, arm motor control or shoulder
motion of all joints (Table 3). The multilevel regression subluxation compared to a control group receiving a similar
analysis identified significant time effects for the three amount of sham positioning combined with TENS in addition
shoulder movements and for forearm supination. There to conventional rehabilitation.
was no significant group effect nor a significant time ×
Previous attempts to maintain hemiplegic arm joint range of
group interaction. A random intercept model fitted the data motion using static muscle stretching procedures could not
best (–2log-likelihood criterion). At end-treatment, the
prevent considerable loss of shoulder passive range of motion
mean between-group difference for passive shoulder (Ada et al 2005, Gustafsson and McKenna 2006, de Jong et al
external rotation was 13 deg (95% CI 1 to 24).
2006, Turton and Britton 2005). Our participants showed
At baseline, 37% of all participants (ie, 17/46) reported similar reductions in mean passive range of motion across
shoulder pain, as presented in Table 4 (see eAddenda for most arm joints. Overall, there were no significant differences
Table 4). At 8 weeks, this percentage was 52% (ie, 22/42) in passive range of motion between the two groups. At
with a relative risk of shoulder pain in the experimental baseline (on average, six weeks post-stroke), 37% of the
group of 1.44 (95% CI 0.80 to 2.62), but no significant participants reported (shoulder) pain. During the intervention
2 period, the prevalence increased to 52% and decreased to 36%
difference between the groups (r = 1.53, p = 0.217). At
three months later. These findings are in line with reports that
follow-up 36% (ie, 13/39) of all participants had shoulder post-stroke shoulder pain is common, affecting 22–64% of
pain. At 8 weeks, participants with shoulder pain showed cases, particularly patients with poor arm function (Aras et al
no significant between-group differences in their responses 2004, Gamble et al 2002, Lindgren et al 2007). Overall, pain
to the verbal question as well as in the visual graphic severity also increased, particularly on movement and at
rating scale scores on movement and at night. Overall, the night. This adverse effect was also noted in other trials
pain scores showed inconsistent patterns which hindered
(Gustafsson and McKenna 2006, Turton and Britton 2005).
within- and between-group comparisons of those with
Although there were no significant between-group differences
shoulder pain only. There were no significant between-
regarding shoulder pain, worrisome observations were that in
group differences on the Leeds Adult/Arm Spasticity
the experimental group some participants reported that they
Impact Scale, the Modified Tardieu Scale, the Fugl-Meyer
considered the intervention to be very arduous, pain and
Assessment arm score, and the subluxation scores at end-
spasticity medication were prescribed more frequently, and
treatment, as presented in Table 5 (see eAddenda for Table
protocol compliance was lower. Combined with the finding
5). It is of note that all participants with clinically relevant
that shoulder pain was more likely to occur in participants in
hypertonia also demonstrated a spasticity angle > 0 deg
the experimental group than in the control group (relative risk
and that Tardieu Scale scores for the internal rotators could
1.44), these findings may indicate that for some participants
not be obtained in a large number of participants because
the experimental procedure was not well tolerated.
they had very limited (< 70 deg) total shoulder external
rotation range. The overall prevalence of subluxation
decreased from baseline (61%) to follow-up (31%). During the eight weeks of intervention our participants
showed increased Leeds Adult/Arm Spasticity Impact
Scale sum scores and Fugl-Meyer Assessment arm motor
scores – changes that were probably not clinically relevant

250 Journal of Physiotherapy 2013 Vol. 59 – © Australian Physiotherapy Association 2013 . Open access under CC BY-NC-ND license.
 5BCMF ﾵ. Mean (SD) for passive range of motion in degrees for each group, mean (SD) difference within groups, and mean (95% CI) difference between groups. The multi-level
Journal of Physiotherapy 2013 Vol. 59 – © Australian Physiotherapy Association 2013 . Open access under CC BY-NC-ND license.

regression analysis identified significant time effects for the three shoulder movements and for forearm supination. There was no significant group effect nor a significant group x
time interaction. A random intercept results in the best fit for the data (–2log-likelihood criterion).

Outcome Groups Difference within groups Difference between groups

Week 0 Week 4 Week 8 Week 20 Week 4 Week 8 Week 20 Week 4 Week 8 Week 20
minus minus minus minus minus minus
Week 0 Week 0 Week 0 Week 0 Week 0 Week 0
Exp Con Exp Con Exp Con Exp Con Exp Con Exp Con Exp Con Exp minus Exp minus Exp
(n = 23) (n = 23) (n = 23) (n = 21) (n = 21) (n = 21) (n = 17) (n = 22) Con Con minus
Con
Shoulder 29 34 20 19 18 11 20 21 –9 –14 –10 –23 –5 –13 5 13 8
external (20) (19) (28) (21) (23) (24) (29) (25) (17) (14) (15) (21) (23) (21) (–5 to 14) (1 to 24) (–7 to 22)
rotation
Shoulder 130 122 111 104 107 100 107 103 –18 –15 –22 –22 –16 –18 –3 0 2
flexion (33) (29) (37) (22) (37) (20) (36) (20) (24) (18) (26) (30) (31) (27) (–16 to 10) (–17 to 18) (–17 to 21)
Shoulder 110 93 93 71 92 66 84 72 –17 –17 –18 –27 –18 –20 0 9 2

abduction (48) (41) (51) (32) (51) (27) (46) (27) (41) (21) (48) (34) (49) (33) (–20 to 20) (–17 to 35) (–24 to 29)
Elbow 3 3 2 5 3 5 6 2 –1 1 0 2 2 –1 –2 –2 3

extensiona (8) (7) (9) (7) (10) (7) (12) (12) (6) (5) (8) (7) (8) (11) (–5 to 2) (–7 to 3) (–4 to 9)

de Jong et al: Combined arm positioning and NMES poststroke

Forearm 77 78 68 68 67 69 59 67 –8 –9 –10 –9 –15 –12 1 –1 –3
supination (13) (11) (16) (15) (17) (12) (16) (16) (12) (17) (12) (12) (18) (14) (–8 to 10) (–8 to 7) (–13 to 7)
Wrist 58 54 55 47 56 54b 54 59 –3 –5 –2 0b –2 6 2 –3 –8
extension I (18) (17) (20) (14) (20) (16) (20) (14) (11) (12) (15) (16) (20) (19) (–5 to 9) (–12 to 7) (–21 to 5)
Wrist 66 60 59b 53 62 57 60 63 –6 –6 –4 –3 –4 3 0 –1 –7

extension II (12) (14) (17) (13) (18) (15) (20) (15) (9) (8) (11) (14) (16) (15) (–5 to 5) (–9 to 6) (–17 to 4)
a
Exp = experimental group, Con = control group, I = wrist extension with extended fingers, II = wrist extension with flexed fingers. Elbow extension values indicate deviation from the neutral
position, ie, degrees of elbow flexor contracture with negative values representing hyperextension. bData missing for one participant.
251
Research

and caused by a mix of spontaneous post-stroke recovery of intervention will not have a clinically relevant impact in
function, learned capacity to use compensatory movement this subgroup of patients either.
strategies of the nonaffected arm and/or increased
involvement of the carer. Overall, the prevalence of elbow Research to date suggests that it is not possible to control or
flexor hypertonia and spasticity jointly increased up to 55% at overcome (the emergence of) contractures and hypertonia
the end of the treatment period, roughly corresponding to using the current static arm muscle stretching procedures.
three months post-stroke for our participants. These results are Similarly, NMES of the antagonists of the muscles prone to
in concordance with previous work (de Jong et al 2011, van shortening does not seem to provide additional benefits either.
Kuijk et al 2007, Urban et al 2010). The unexpected high We therefore argue that these techniques should be
prevalence of hypertonia and spasticity (62%) and a discontinued in the treatment of patients with a poor prognosis
decreasing prevalence of shoulder subluxation (31%) at for functional recovery. In this subgroup of patients it is
follow-up in our sample may be explained by the fact that becoming an increasingly difficult challenge to find effective
patients with relatively poor arm motor control have a higher treatments that can prevent the development of the most
risk of developing hypertonia (de Jong et al 2011). common residual impairments such as contractures,
hypertonia, and spasticity and its associated secondary
Although we performed an intention-to-treat analysis (ie, problems such as shoulder pain and restrictions in
using any available data from all randomised subjects), we did performance of daily life activities. Further research is
not use forward imputation of missing data representing a required to investigate what renders these interventions
clinical variable (eg, shoulder passive range of motion) that is ineffective. The efficacy of other approaches, such as
worsening over time (de Jong et al 2007), as this might transcranial magnetic stimulation, NMES of the muscles
increase the chance of a Type I error. However, for prone to shortening (Goldspink et al 1991), or other
completeness, this stricter intention-to-treat analysis using the combinations of techniques, could also be investigated. Q
data of all randomised subjects (n = 48) was performed. This
analysis was similar in outcome to the original analysis but
revealed an additional time effect of wrist extension with a
flexed fingers. A per protocol analysis would also have Footnotes: MIE Medical Research Ltd, Leeds, UK.
b
resulted in similar results because no patients crossed over to STIWELL-med4, Otto Bock HealthCare, Germany.
the other group. We also refrained from performing a
eAddenda: Table 4, 5, 6 (individual patient data) and
sensitivity analysis based on compliance because meaningful
conclusions could not be drawn from the resulting limited
Appendix 1 and 2.
sample sizes. We furthermore acknowledge that the Leeds Ethics: The study was approved by the Medical Ethics
Adult/Arm Spasticity Impact Scale lacks psychometric Committee of the University Medical Center Groningen.
evaluation and our method to standardise the Tardieu Scale’s All participants gave written informed consent prior to
stretch velocity (V3) using a metronome was not validated participation.
and tested for reliability. Therefore, our data regarding basic
arm activities, hypertonia, and spasticity should be interpreted Support: This study was financially supported by Fonds
with caution. Finally, because overall compliance to both NutsOhra [SNO-T-0702-72] and Stichting Beatrixoord
protocols was only about 70%, an underestimation of the Noord-Nederland.
treatment effect may also have occurred. Nevertheless, the
combined administration of 43 hours of static stretching and Acknowledgements: We thank the assessors Ank Mollema
36 hours of NMES was more than administered during any and Marian Stegink (De Vogellanden, Zwolle), the local
previous trial (Borisova and Bohannon 2009). trial co-ordinators Marijke Wiersma and Siepie Zonderland
(Revalidatie Friesland, Beetsterzwaag), Astrid Kokkeler
and Dorien Nijenhuis (MRC Aardenburg, Doorn), Alinda
A recent study produced inconclusive evidence about the Gjaltema and Femke Dekker (De Vogellanden, Zwolle)
effectiveness of a combined intervention of electrical and the participants, physicians, physio- and occupational
stimulation in conjunction with prolonged muscle stretch therapists and nursing staff involved in the trial.
(using a splint) to treat and prevent wrist contracture (Leung
et al 2012). Similarly, our results also showed no added Competing interests: Otto Bock Healthcare provided
benefit of electrical stimulation during static stretching of the electrical stimulators free of charge. None of the sponsors
shoulder and arm. The results of these multimodal approaches had any involvement in study design, data collection and
to the problem of post-stroke arm contracture development are analysis, decision to publish, or preparation of the
in line with the conclusion of a review (Katalinic et al 2011) manuscript.
that static stretch positioning procedures have little, if any,
short or long term effects on muscle contracture (treatment Correspondence: Lex D de Jong, Hanze University of
effect ) 3 deg), pain, spasticity, or activity limitations. Applied Sciences, School of Physiotherapy, Eyssoniusplein
Although pooled data from studies investigating the effects of 18, 9714 CE Groningen, The Netherlands. Email: l.d.de.
electrical stimulation suggested some treatment effects on jong@pl.hanze.nl
functional motor ability (Pomeroy et al 2006) and pain-free
range of passive humeral lateral rotation in patients with References
residual arm motor capacity (Price and Pandyan 2000), we Ada L, Canning C (1990) Anticipating and avoiding muscle
found no such results in our sample of patients without shortening. In Ada L, Canning C (Eds) Key Issues in
residual arm motor capacity. As the combined procedure did /FVSPMPHJDBM 1IZTJPUIFSBQZ ﾤ 0YGPSE
not result in any meaningful treatment effects, it suggests that #VUUFSXPSUI Heinemann, pp 219–236.
application of muscle stretching or NMES alone as a Ada L, Goddard E, McCully J, Stavrinos T, Bampton J (2005)
monotherapeutic Thirty minutes of positioning reduces the development

252 Journal of Physiotherapy 2013 Vol. 59 – © Australian Physiotherapy Association 2013 . Open access under CC BY-NC-ND license.
 de Jong et al: Combined arm positioning and NMES poststroke

of shoulder external rotation contracture after stroke: a Gladstone DJ, Danells CJ, Black SE (2002) The Fugl-Meyer
randomized controlled trial. Archives of Physical Medicine & assessment of motor recovery after stroke: a critical review
Rehabilitation 86: 230–234. of its measurement properties. Neurorehabilitation & Neural
Alfieri V (1982) Electrical treatment of spasticity. Reflex tonic Repair 16: 232–240.
activity in hemiplegic patients and selected specific Goldspink DF, Easton J, Winterburn SK, Williams PE,
electrostimulation. Scandinavian Journal of Rehabilitation Goldspink GE (1991) The role of passive stretch and
Medicine 14: 177–182. repetitive electrical stimulation in preventing skeletal muscle
Aras MD, Gokkaya NK, Comert D, Kaya A, Cakci A (2004) atrophy while reprogramming gene expression to improve
fatigue resistance. Journal of Cardiac Surgery 6: 218–224.
Shoulder pain in hemiplegia: results from a national
rehabilitation hospital in Turkey. American Journal of Gustafsson L, McKenna K (2006) A programme of static
Physical Medicine & Rehabilitation 83: 713–719. positional stretches does not reduce hemiplegic shoulder
Ashford S, Slade M, Malaprade F, Turner-Stokes L (2008) pain or maintain shoulder range of motion – a randomized
Evaluation of functional outcome measures for the controlled trial. Clinical Rehabilitation 20: 277–286.
hemiparetic upper limb: a systematic review. Journal of Haugh AB, Pandyan AD, Johnson GR (2006) A systematic
Rehabilitation Medicine 40: 787–795. review of the Tardieu Scale for the measurement of
Barker WH, Mullooly JP (1997) Stroke in a defined elderly spasticity. Disability & Rehabilitation 28: 899–907.
population, 1967–1985. A less lethal and disabling but no Held J, Pierrot-Deseilligny E (1969) Le bilan moteur central. In
less common disease. Stroke 28: 284–290. Rééducation motrice des affections neurologiques. Paris:
Bohannon RW, Andrews AW (1990) Shoulder subluxation and Libraire Balliere, pp 31–42.
pain in stroke patients. American Journal of Occupational Katalinic OM, Harvey LA, Herbert RD (2011) Effectiveness of
Therapy 44: 507–509. stretch for the treatment and prevention of contractures in
Borisova Y, Bohannon RW (2009) Positioning to prevent or people with neurological conditions: a systematic review.
reduce shoulder range of motion impairments after stroke: a Physical Therapy 91: 11–24.
meta-analysis. Clinical Rehabilitation 23: 681–686. King TI, II (1996) The effect of neuromuscular electrical
Brunnstrom S (1970) Movement therapy in hemiplegia: a stimulation in reducing tone. American Journal of
neurophysiological approach. Hagerstown: Harper and Row. Occupational Therapy 50: 62–64.
Kwah LK, Harvey LA, Diong JH, Herbert RD (2012) Half of the
Chae J, Sheffler L, Knutson J (2008) Neuromuscular electrical adults who present to hospital with stroke develop at least
stimulation for motor restoration in hemiplegia. Topics in one contracture within six months: an observational study.
Stroke Rehabilitation 15: 412–426. Journal of Physiotherapy 58: 41–47.

Coupar F, Pollock A, van Wijck F, Morris J, Langhorne P (2010) Leung J, Harvey LA, Moseley AM, Tse C, Bryant J, Wyndham
Simultaneous bilateral training for improving arm function S, et al (2012) Electrical stimulation and splinting were not
clearly more effective than splinting alone for contracture
after stroke. Cochrane Database of Systematic Reviews
management after acquired brain injury: a randomised trial.
CD006432.
Journal of Physiotherapy 58: 231–240.
de Jong LD, Dijkstra PU, Stewart RE, Postema K (2012)
Lindgren I, Jonsson AC, Norrving B, Lindgren A (2007)
Repeated measurements of arm joint passive range of
Shoulder pain after stroke: a prospective population-based
motion after stroke: interobserver reliability and sources of
study. Stroke 38: 343–348.
variation. Physical Therapy 92: 1027–1035.
Lundström E, Terent A, Borg J (2008) Prevalence of disabling
de Jong LD, Hoonhorst MH, Stuive I, Dijkstra PU (2011) Arm
spasticity 1 year after first-ever stroke. European Journal of
motor control as predictor for hypertonia after stroke: a
Neurology 15: 533–539.
prospective cohort study. Archives of Physical Medicine &
Rehabilitation 92: 1411–1417. Mackay J, Mensah G (2004) The atlas of heart disease and
de Jong LD, Nieuwboer A, Aufdemkampe G (2006) stroke. Geneva: World Health Organization.
Contracture preventive positioning of the hemiplegic arm in Morris S (2002) Ashworth and Tardieu Scales: their clinical
subacute stroke patients: a pilot randomized controlled trial. relevance for measuring spasticity in adult and paediatric
Clinical Rehabilitation 20: 656–667. and neurological populations. Physical Therapy Reviews 7:
de Jong LD, Nieuwboer A, Aufdemkampe G (2007) The 53–62.
hemiplegic arm: interrater reliability and concurrent validity of Norman G, Streiner D (2000) Biostatistics: the bare essentials.
passive range of motion measurements. Disability & Toronto: Dekker Inc.
Rehabilitation 29: 1442–1448. O’Dwyer NJ, Ada L, Neilson PD (1996) Spasticity and muscle
Dewald JP, Given JD, Rymer WZ (1996) Long-lasting reductions of contracture following stroke. Brain 119: 1737–1749.
spasticity induced by skin electrical stimulation. IEEE Pandyan AD, Cameron M, Powell J, Stott DJ, Granat MH
Transactions on Rehabilitation Engineering 4: 231–242. (2003) Contractures in the post-stroke wrist: a pilot study of
Fugl-Meyer AR, Jaasko L, Leyman I, Olsson S, Steglind S its time course of development and its association with
(1975) The post-stroke hemiplegic patient. 1. A method for upper limb recovery. Clinical Rehabilitation 17: 88–95.
evaluation of physical performance. Scandinavian Journal of Patrick E, Ada L (2006) The Tardieu Scale differentiates
Rehabilitation Medicine 7: 13–31. contracture from spasticity whereas the Ashworth Scale is
Fujiwara T, Kasashima Y, Honaga K, Muraoka Y, Tsuji T, Osu confounded by it. Clinical Rehabilitation 20: 173–182.
R, et al (2009) Motor improvement and corticospinal Paulis WD, Horemans HL, Brouwer BS, Stam HJ (2011)
modulation induced by hybrid assistive neuromuscular Excellent test-retest and inter-rater reliability for Tardieu
dynamic stimulation (HANDS) therapy in patients with Scale measurements with inertial sensors in elbow flexors of
chronic stroke. Neurorehabilitation & Neural Repair 23: 125– stroke patients. Gait & Posture 33: 185–189.
132.
Pomeroy VM, King L, Pollock A, Baily-Hallam A, Langhorne P
(2006) Electrostimulation for promoting recovery of
Gamble GE, Barberan E, Laasch HU, Bowsher D, Tyrrell PJ,
movement or functional ability after stroke. Cochrane
Jones AK (2002) Poststroke shoulder pain: a prospective
Database of Systematic Reviews CD003241.
study of the association and risk factors in 152 patients from
a consecutive cohort of 205 patients presenting with stroke. Price CI, Pandyan AD (2000) Electrical stimulation for
European Journal of Pain 6: 467–474.

Journal of Physiotherapy 2013 Vol. 59 – © Australian Physiotherapy Association 2013 . Open access under CC BY-NC-ND license. 253
Research

preventing and treating post-stroke shoulder pain. Cochrane Turton AJ, Britton E (2005) A pilot randomized controlled trial of
Database of Systematic Reviews CD001698. a daily muscle stretch regime to prevent contractures in the
Roosink M, Renzenbrink GJ, Buitenweg JR, Van Dongen RT, arm after stroke. Clinical Rehabilitation 19: 600–612.
Geurts AC, IJzerman MJ (2011) Persistent shoulder pain in Urban PP, Wolf T, Uebele M, Marx JJ, Vogt T, Stoeter P, et al
the first 6 months after stroke: results of a prospective cohort (2010) Occurrence and clinical predictors of spasticity after
study. Archives of Physical Medicine & Rehabilitation 92: ischemic stroke. Stroke 41: 2016–2020.
1139–1145. van Kuijk AA, Hendricks HT, Pasman JW, Kremer BH, Geurts
Sirtori V, Corbetta D, Moja L, Gatti R (2009) Constraint-induced AC (2007) Are clinical characteristics associated with upper-
movement therapy for upper extremities in stroke patients. extremity hypertonia in severe ischaemic supratentorial
Cochrane Database of Systematic Reviews CD004433. stroke? Journal of Rehabilitation Medicine 39: 33–37.
Sunderland A, Tinson D, Bradley L, Hewer RL (1989) Arm Van Peppen R, Kwakkel G, Harmeling-van der Wel B, Kollen B,
function after stroke. An evaluation of grip strength as a Hobbelen J, Buurke J, et al (2004) KNGF Clinical Practice
measure of recovery and a prognostic indicator. Journal of Guideline for physical therapy in patients with stroke. Review
Neurology, Neurosurgery and Psychiatry 52: 1267–1272. of the evidence. Nederlands Tijdschrift voor Fysiotherapie
Turner-Stokes L, Jackson D (2006) Assessment of shoulder 114: suppl.
pain in hemiplegia: sensitivity of the ShoulderQ. Disability & Wade DT, Langton-Hewer R, Wood VA, Skilbeck CE, Ismail
Rehabilitation 28: 389-395. HM (1983) The hemiplegic arm after stroke: measurement
Turner-Stokes L, Rusconi S (2003) Screening for ability to and recovery. Journal of Neurology, Neurosurgery and
complete a questionnaire: a preliminary evaluation of the Psychiatry 46: 521–524.
AbilityQ and ShoulderQ for assessing shoulder pain in stroke Wanklyn P, Forster A, Young J (1996) Hemiplegic shoulder pain
patients. Clinical Rehabilitation 17: 150–157. (HSP): natural history and investigation of associated
features. Disability & Rehabilitation 18: 497–501.

254 Journal of Physiotherapy 2013 Vol. 59 – © Australian Physiotherapy Association 2013 . Open access under CC BY-NC-ND license.