Está en la página 1de 53


Painful Hemiplegic Shoulder

Robert Teasell MD, Sanjit K. Bhogal MSc, Norine Foley MSc

Last Updated June 1, 2006

Table of Contents

11.1 CAUSES OF HEMIPLEGIC SHOULDER PAIN ..................................6

11.2 SHOULDER SUBLUXATION ..............................................................7

11.2.1 PATHOPHYSIOLOGY................................................................................................. 7
11.2.2 SCAPULAR ROTATION .............................................................................................. 9
11.2.3 PAIN IN SHOULDER SUBLUXATION .......................................................................... 10


SHOULDER PAIN (HSP)..................................................................13
11.3.1 SPASTIC MUSCLE IMBALANCE ................................................................................ 15
11.3.2 FROZEN OR CONTRACTED SHOULDER .................................................................... 17

11.4 ROTATOR CUFF DISORDERS .......................................................19


SHOULDER .....................................................................................20


11.6.1 POSITIONING OF THE HEMIPLEGIC SHOULDER ......................................................... 21
11.6.2 SLINGS AND OTHER AIDS....................................................................................... 22
11.6.3 STRAPPING THE HEMIPLEGIC SHOULDER ................................................................ 24
11.6.4 ACTIVE THERAPIES IN THE HEMIPLEGIC SHOULDER.................................................. 26
11.6.5 INJECTIONS IN THE HEMIPLEGIC SHOULDER............................................................. 29
11.6.7 SURGERY AS TREATMENT FOR MUSCLE IMBALANCE................................................. 35
11.6.8 MOTOR BLOCKS AS TREATMENT FOR MUSCLE IMBALANCE ....................................... 35
11.6.9 SUMMARY OF THE MANAGEMENT OF HEMIPLEGIC SHOULDER ................................... 37

11.7 SHOULDER HAND SYNDROME (SHS) ...........................................38

11.7.1 STAGES AND SYMPTOMS ....................................................................................... 38
11.7.2 PATHOPHYSIOLOGY............................................................................................... 39
11.7.3 DIAGNOSTIC TESTS ............................................................................................... 41
11.7.4 TREATMENT OF SHOULDER-HAND SYNDROME......................................................... 42
11.7.6 GRADED MOTOR IMAGERY AS A TREATMENT FOR CRPS1....................................... 44

11.8 SUMMARY.........................................................................................46

Key Points

Spasticity and hemiplegic shoulder pain are related.

Further research is needed before conclusions regarding positioning of the

hemiplegic shoulder can be made.

There is limited evidence that shoulder slings influence clinical outcomes.

Strapping the hemiplegic shoulder does not appear to improve upper limb
function, but may reduce pain.

Aggressive range of motion exercises (i.e. pullies) results in a markedly

increased incidence of painful shoulder; a gentler range of motion program is
preferred. Adding ultrasound treatments is not helpful while NSAIDs may be

Corticosteroid injections do not improve hemiplegic shoulder pain or range of


A potentially new treatment of the painful hemiplegic shoulder that requires

further research involves deinnervation of the subscapularis and pectoralis major

Oral corticosteroids appear to improve shoulder-hand syndrome for at least the

first 4 weeks.

A motor imagery program appears to improve shoulder-hand syndrome.

11. Painful Hemiplegic Shoulder

Shoulder pain resulting from hemiplegia is a common clinical consequence of a focal

cerebral insult resulting from a vascular lesion (ie. hemorrhagic or ischemic stroke).
The incidence of shoulder pain varies between studies, with estimates which range from
48% to 84% (Najenson et al. 1971, Poulin de Courval et al. 1990). Shoulder pain, by
itself, can result in significant disability (Najenson et al. 1971, Poduri 1993) and although
it can occur as early as 2 weeks post stroke, an onset time of 2-3 months post stroke is
more typical (Poduri 1993). In a recent prospective study, Gamble et al. (2002)
reported that 52/152 (34%) developed shoulder pain following stroke, 28% by two
weeks and 87% by two months. By 6 months, the pain had resolved in 80% of the

Shoulder pain can negatively affect rehabilitation outcomes as good shoulder function is
a prerequisite for successful transfers, maintaining balance, performing activities of daily
living and for effective hand function (Risk et al. 1984). Lo et al. (2005) catalogued the
different types of shoulder dysfunction based on both clinical and arthrographic findings
and reported that 16% of patients of a cohort of 32 patients with hemiplegic shoulder
pain within one-year of stroke had shoulder-hand syndrome, 4% had rotator cuff tears
and 50% suffered from frozen shoulder. 63% of patients had a single type of shoulder
dysfunction while 34% had two types.

11.1 Causes of Hemiplegic Shoulder Pain

Although many etiologies have been proposed for hemiplegic shoulder pain,
increasingly it appears to be a consequence of spasticity and the sustained hemiplegic
posture. Shoulder pain may be more common among patients with neglect following
stroke (Kaplan 1995). Possible sources of hemiplegic shoulder pain are listed in Table
11.1. Factors most frequently associated with shoulder pain are shoulder
(glenohumeral) subluxation
(Crossens-Sills and Schenkman Table 11.1 Potential Causes of Hemiplegic shoulder Pain
1985, Moskowitz et al. 1969b, Anatomical Site Mechanism
Savage and Robertson 1982, Muscle Rotator Cuff, Muscle Imbalance,
Shai et al. 1984), shoulder Subscapularis Spasticity, Pectoralis
contractures or restricted Spasticity
shoulder range of motion (Bloch Bone Humeral Fracture
and Bayer 1978, Braun et al.
1981, Fugl-Meyer et al. 1975, Joint Glenohumeral Subluxation
Crossens-Sills and Schenkman
1985, Hakuno et al. 1984, Risk Bursa Bursitis
et al. 1984) and spasticity, Tendon Tendonitis
particularly of the subscapularis
and pectoralis muscles (Braun et Joint Capsule Frozen or Contracted Shoulder
al. 1981, Caldwell et al. 1969, (Adhensive Capsulitis)

Other Shoulder-hand Syndrome

(Reflex Sympathetic Dystrophy)
Moskowitz 1969a, 1969b). Other suggested causes of shoulder pain include reflex
sympathetic dystrophy (Chu et al. 1981, Davis et al. 1977, Perrigot et al. 1975), or injury
to the rotator cuff musculotendinous unit (Najenson et al. 1971, Nepomuceno et al.
1974). The role of central post stroke pain in the etiology of shoulder pain is unclear
(Walsh 2001).

11.2 Shoulder Subluxation

11.2.1 Pathophysiology

Shoulder subluxation is best defined as changes in the mechanical integrity of the

glenohumeral joint causing a palpable gap between the acromion and humeral head.
The most reliable clinical measurement of the subacromial space used in clinical
research is calipers (Boyd 1992). The glenohumeral joint is multiaxial and has a range
of motion, which exceeds that of other joints in the body. To achieve this mobility the
glenohumeral joint must sacrifice stability. Stability is achieved through the rotator cuff,
a musculotendinous sleeve which maintains the humeral head in the glenoid fossa,
while at the same time allowing shoulder mobility. During the initial period following a
stroke the hemiplegic arm is flaccid or hypotonic. Therefore the shoulder musculature,
in particular the rotator cuff musculotendinous sleeve, cannot perform its function of
maintaining the humeral head in the glenoid fossa and there is a high risk of shoulder

Shoulder subluxation is a very common problem in hemiplegic patients. During the

initial flaccid stage of hemiplegia the involved extremity must be adequately supported
or the weight of the arm will result in shoulder subluxation. Improper positioning in bed,
lack of support while the patient is in the upright position or pulling on the hemiplegic
arm when transferring the patient all contribute to glenohumeral subluxation. Down and
lateral subluxation commonly occur secondary to prolonged downward pull on the arm
against which hypotonic muscles offer little resistance (Chaco and Wolf 1971). The
resulting mechanical effect is overstretching of the glenohumeral capsule (especially its
superior aspect) and flaccid supraspinatus and deltoid muscles (Basmajian and Bazant
1959, Shahani et al. 1981) (Figure 11.1).

Figure 11.1 A. Normal Shoulder. The humeral head is maintained in the glenoid fossa by the
supraspinatus muscle.

Figure 11.1 B. Shoulder Subluxation. During the initial phase of hemiplegia, the supraspinatus muscle
is flaccid. The weight of the unsupported arm can cause the humeral head to sublux downward out of the
glenoid fossa.

11.2.2 Scapular Rotation

There appear to be other factors playing a role in subluxation of the glenohumeral joint.
Basmajian and Bazant (1959) proposed that in the normal state, subluxation of the
humeral head was prevented by upward angulation of the glenoid fossa and the upper
part of the shoulder capsule, the coracohumeral ligament and supraspinatus muscle.
After a hemiplegic stroke they hypothesized that the upward angulation of the scapula
would be lost. Calliet (1980) added that in the flaccid stage, the scapula assumed a
depressed and downward rotated position, as the paretic serratus anterior and the
upper part of the trapezius muscles no longer support the scapula. The combination of
flaccid supportive musculature (in particular, the supraspinatus muscle) and a
downward rotated scapula was presumed to predispose the head of the humerus to
undergo inferior subluxation relative to the glenoid fossa.
Table 11.2 Scapular Rotation in the Hemiplegic Shoulder

Author, Year Method Outcome

Prevost et al. 50 patients who had experienced a The angle of abduction of the arm of the
1987 cerebrovascular accident, presenting with affected side was significantly greater
Canada right hemiplegia. Inferior subluxation of the than on the non-affected side, p<0.05, but
shoulder in hemiplegia was measured the relative abduction of the arm was on
using a tridimensional (3-D) x-ray the same order of magnitude for both
technique, giving true vertical distance sides. There was no significant
separating the apex of the humeral head relationship between the orientation of the
and the inferior margin of the glenoid scapula and the severity of the
cavity. Both shoulders were evaluated and subluxation. The abduction of the
the difference used as a measure of humerus was weakly (r=.24) related to the
subluxation. The measure was then subluxation, which partly explained the
compared to the orientation of the scapula weak association found between the
relative to the vertical and the abduction of relative abduction of the arm and the
the arm. subluxation.
Culham et al. 34 hemiplegic patients were divided into Scapula was significantly further from the
1995 high-tone and low-tone groups according midline and lower on the thorax on the
Canada to Ashworth scoring of muscle tone. Low affected side in the low-tone group.
No Score tone patients scored less than 4 and high Glenohumeral subluxation was
tone patients had a score of 4 or greater significantly greater in the low-tone group.
on the MAS. Scapular abduction angle was significantly
greater on the non-affected side in the
low-tone group. In the high-tone group, no
differences were found between the
affected and the non-affected side in
either the angular or linear measures.

Author, Year Method Outcome
Price et al. 30 patients, 6 months post stroke received 24 patients had no shoulder subluxation, 6
2001 a standardized clinical assessment, patients suffered from shoulder
UK whereby manual palpitation of the subluxation. Among all patients, the
No Score subacrimonial space was performed to average degree of scapular downward tilt
identify those with subluxation. The Upper was 10.04 for the unaffected side and
Limb Motricity Score (ULMS) was used to 10.46 for the affected side (p=ns). There
record arm strength. The degree of was no difference in the scapular
scapular downward tilt and dynamic downward tilt of the affected shoulder of
scapular lateral rotation were assessed by subjects with and without subluxation.
a Scapula Locator System device to
determine their association with

Prevost et al. (1987), using a 3-D x-ray technique, studied the movement of the scapula
and humerus in stroke patients. They studied 50 stroke patients comparing the affected
to the non-affected shoulder. They were able to demonstrate that there was a
difference between the affected and non-affected shoulders in terms of the vertical
position of the humerus (ie. degree of subluxation) in relation to the scapula. The
orientation of the glenoid fossa was also different; however, they found that with the
subluxed shoulder it was actually facing less downward. There was no significant
relationship noted between the orientation of the scapula and the severity of
subluxation. They concluded that the scapular position was not an important factor in
the occurrence of inferior subluxation in hemiplegia (Prevost et al. 1987). Culham et al.
(1995) reported that while patients with low-tone had significantly greater subluxation
compared to the high-tone group, (0.52 vs. 0.21) there was no correlation between the
amount of subluxation and the scapular abduction angle or the humeral abduction
angle. Price et al. (2001) compared patients with and without stroke (n=15) and reported
that subluxation in stroke patients was unrelated to scapular resting position. These
authors also reported that the normal scapula tilts downward to a greater degree found
in other studies.

11.2.3 Pain in Shoulder Subluxation

Shoulder subluxation may be associated with several conditions including: shoulder

pain (Crossens-Sills and Schenkman 1985, Moskowitz et al. 1969b, Savage and
Robertson 1982, Shai et al. 1984, Roy et al. 1994) and frozen shoulder or brachial
plexus traction injury (Kingery et al. 1993), although evidence for the latter is lacking
(Kingery et al. 1993). It has long been assumed that if not corrected; a pattern of
traction on the flaccid shoulder will result in pain, decreased range of motion and
contracture. However, not all patients with a subluxed hemiplegic shoulder experience
shoulder pain and it remains controversial as to whether it causes hemiplegic shoulder
pain (Fitzgerald-Finch and Gibson 1975, Moskowitz et al. 1969b, Shahani et al. 1981,
Bender and McKenna 2001). The failure to consistently report an association may be
due, in part to a failure to examine the contribution of other probable etiological factors
concurrently and to account for the chronicity of stroke since there is a correlation
between early signs of shoulder subluxation and the development of pain. Paci et al.

(2005) suggested that pain associated with subluxation is probably present later after
stroke since “fibrous changes or injury can occur in connective tissue of the ligaments
and joint capsule due to incorrect alignment between the humerus and the scapula”.
Although several studies have reported an association, others have not confirmed this
finding. Heterogeneity of patient characteristics and timing and method of assessment
(radiological vs. clinical examination) may account for the lack of consistency of
findings. (see Table 11.3)
Table 11.3 Studies which Support or Fail to Report an Association between Shoulder Subluxation
and Pain

Studies Supporting the Role of Shoulder Studies Which Fail to Support the Role of
Subluxation in Pain Shoulder Subluxation in Pain

Shai et al. 1984 Peszczynski & Rardin 1965

Van Ouwenaller et al. 1986 Bohannon 1988
Poulin de Courval et al. 1990 Van Langenberghe & Hogan 1988
Roy et al. 1994 Bohannon & Andrews 1990
Chantraine et al. 1999 Kumar et al. 1990
Lo et al. 2003 Arsenault et al. 1991
Aras et al. 2004 Joynt 1992
Zorowitz et al. 1996
Ikai et al. 1998

Individual Studies

A selection of studies examining the relationship between shoulder subluxation and pain
are presented in Table 11.4.

Table 11.4 Pain and Subluxation in the Hemiplegic Shoulder

Author, Year Methods Main Outcome

Shai et al. 33 patients received at least a single radiograph There was a significant correlation
1984 early in their hospitalization. Clinical status was between abnormal radiologic
Israel assessed 4 to 11 months after stroke. findings early in the course of
No Score stroke and the development of
pain. 19/33 patients had evidence
of subluxation on radiograph and
17/33 had shoulder pain. Of those
with shoulder pain 14/17 (82%)
had subluxed shoulders.
Bohannon 30 patients admitted for inpatient rehabilitation as 24 patients had shoulder pain on
1988 average of 31 days following stroke. A variety of initial assessment and 27 on final
USA tests were performed at admission and discharge assessment. 30% of shoulders
No Score to assess their correlation with shoulder pain. were subluxed on initial
Subluxation was assessed by palpitation and assessment and 47% at final
observation (clearly subluxed vs. not clearly assessment. There was no
subluxed). statistical significant relationship
between pain and subluxation.
Bohannon & 28 consecutively admitted patients undergoing 70.8% of patients demonstrated

Author, Year Methods Main Outcome
Andrews rehabilitation for their first stroke who could follow enough shoulder pain to at least
1990 instructions, and were aware of the position of cause them to wince when their
USA their paretic limb in space were included. Paretic shoulders were rotated laterally
No Score shoulder subluxation and paretic shoulder pain 900. The SROMP of the paretic
were measured. Shoulder subluxation was side was measured as 64.50+28.80
measured while the patients sat on the edge of a and 64.60+28.90. A significant
mat table with their paretic upper extremity Pearson correlation (-77s,
dependent and the examiner used his thumb to p<0.001) was observed between
palpate the separation between the acromion the Ritchie Index and SROMP
and the head of the humerus. He then graded indicating that patients with higher
subluxation as none (0), minimal (1) or scores on the Ritchie Index had
substantial (2). Shoulder pain was measured fewer degrees of lateral rotation of
during slow lateral rotation of the joint while the the shoulder before pain was
patients were supine. All patients’ shoulders experienced.
were abducted about 45 and their elbows were
held at 90 with their forearms pronated with
measurements beginning from neutral shoulder
rotation. Patients’ responses were graded on a
3-point scale (Ritchie Index) of: no pain,
complaint of pain and wince, complaint of pain,
wince and withdrawal. SROMP measurements
taken with patients’ shoulders laterally rotated
until they first expressed pain in the shoulder, at
which point a fluid-filled gravity goniometer, was
applied and read.
Joynt 1992 A convenience sample of 97 patients suffering 49 patients with specific shoulder
USA from pain in the upper extremity was examined. pain were compared to 18 patients
No Score The interval from stroke onset to examination with pain, not localized to the
ranged from several days to a few years. 49 shoulder. Patients complaining of
patients had specific complaints of shoulder pain. shoulder pain did not exhibit
subluxation more frequently than
patients with general pain in the
affected extremity.
Wanklyn et 108 patients were assessed clinically, 3 times Subluxation was detected clinically
al. 1996 over a 6-month period following stroke. in 31 (29%) patients at hospital
UK Subluxation was assessed clinically and graded discharge and 27 (26%) at 26
No Score in finer-breadths palpable below the acrimonion weeks. Shoulder pain was not
process. associated with subluxation at 2/3
assessment points. The authors do
not provide details of the data.
Zorowitz et 20 stroke patients with shoulder pain, admitted to Shoulder pain after stroke was not
al. 1996 a rehabilitation hospital within 6 weeks of their correlated with age, vertical,
USA first stroke were studied. horizontal, or total asymmetry,
No Score shoulder flexion or abduction, or
Fugl-Meyer scores. However,
shoulder pain was strongly
correlated with degree of shoulder
external rotation
Ikai et al. 75 patients with shoulder subluxation were At rest, 10 patients reported pain.
1998 assessed for pain using a visual analogue scale During passive range of
Japan at rest and during passive range of movement. movement, 5 patients reported no
No Score The degree of pain was expressed as pain, 25 reported mild pain, 36
nonexistent (0), mild (1-3), moderate (4-7), or reported moderate pain and 9
severe (8-10) during passive movement. reported severe pain. Shoulder

Author, Year Methods Main Outcome
pain was not related to the degree
of shoulder subluxation.
Lo et al. 2003 32 consecutive patients with shoulder pain 14 (44%) of patients had clinically
Taiwan following stroke were assessed for shoulder diagnosed shoulder subluxation.
No Score subluxation, which was diagnosed by a gap of
more than one finger breadth between the
acrimonion and the head of the humeral bone on

Aras et al. 85 consecutive stroke patients admitted to one of 27 patients had glenohumeral joint
2004 the largest rehab facilities in Turkey were studied subluxation and reported shoulder
Turkey to identify the incidence of shoulder pain and the pain, compared to 5 patients with
No Score factors associated with it. the same finding, but without pain.

Conclusions Regarding Shoulder Subluxation Post-Stroke

Shoulder subluxation occurs early on in the hemiplegic arm due to flaccid

supporting shoulder musculature and is not a result of downward scapular

Shoulder subluxation may be a cause of shoulder pain; however, patients with

shoulder subluxation do not necessarily experience pain and not all cases of
hemiplegic shoulder pain suffer from subluxation.

Although it has not been established that shoulder subluxation is the primary
cause of hemiplegic shoulder pain it would still seem prudent to take care early
on with the hemiplegic upper extremity to avoid subluxation.

11.3 Spasticity, Contractures and Hemiplegic Shoulder Pain

Abnormal muscle tone, including spasticity may be directly related to HSP. Spasticity is
defined as a disorder of motor function characterized by a velocity-dependent increase
in resistance to passive stretch of muscles accompanied by hyperactive muscle stretch
reflexes and often associated with a clasp-knife phenomena. Spasticity is one
component of the upper motor neuron (UMN) syndrome and is the inevitable
accompaniment of hemiplegia and an incomplete motor recovery. Under normal
circumstances a delicate balance exists between facilitating and inhibiting influences
upon both alpha and gamma motor neurons, which together maintain appropriate
control of skeletal muscle length and strength of contraction at the spinal cord level.
After a stroke, input from one or more of the supraspinal suppressor areas will decrease
or stop entirely. The balance of control over the muscle tips in favour of facilitation and
spasticity results. Spasticity develops only if there is loss of input from both pyramidal

and extrapyramidal motor systems. Spasticity presents as increased tone and reflexes
on the involved side of the body.

Individual Studies

Table 11.5 Spasticity and Hemiplegic Shoulder Pain

Author, Year Methods Outcome

Bohannon et 50 patients with hemiplegia was Of the 50 patients reviewed, 72% had
al. 1986 secondary to cerebrovascular accident, shoulder pain. 20 had some pain while 16 had
USA whose unaffected shoulders severe pain. Three zero-order correlations
No score demonstrated normal and pain-free were significant: ROSER and shoulder pain
range of hemiplegia shoulder external (r=-0.061, p<0.001), time since onset of
rotation (ROSER, 900); able to hemiplegia and shoulder pain (r=0.45,
adequately follow instructions to allow p<0.01), and time since onset of hemiplegia
testing of all variables pertinent to the and ROSER (r=0.37, p<0.01). One-way
study. Information was retrieved from ANOVA demonstrated that time since onset of
patients’ records concerning their initial hemiplegia (F=8.28, p<0.001) and the ROSER
physical therapy evaluation. (F=18.44, p<0.001) were significantly different
Relationships between pain and other in patients with no pain, some pain, and
variables were determined. pronounced/severe pain.
Van 219 hemiplegia patients were followed 72% of patients had shoulder pain at least
Ouwenaller et for 1 year after their stroke. once during their recovery occurring most
al. 1986 Radiographic examinations were done often in patients having spasticity (85%) than
Switzerland for each patient. in patients which flaccidity (18%). Appearance
No score of spasticity was evident in 80% of patients
while 20% remained hypotonic.
Joynt 1992 97 patients were examined between 6-9 67 patients were diagnosed with a shoulder
USA months post stroke for evidence of problem. 49% of patients reported shoulder
No Score shoulder dysfunction and pain, based pain. Shoulder pain was unrelated to
on clinical examination. spasticity, assessed by resistance to rapid
Aras et al. The association between spasticity 54 patients had shoulder pain and 31 did not.
2004 measured by the Ashworth scale and There was no association between spasticity
Turkey shoulder pain was assessed in 85 and shoulder pain.
No Score consecutive stroke patients, who were
grouped by the presence or absence of
shoulder pain.

Van Ouwenaller et al. (1986) looked at various factors in 219 patients followed for one
year after a stroke and identified a much higher incidence of shoulder pain in spastic
(85%) than in flaccid (18%) hemiplegics. They identified spasticity as "the prime factor
and the one most frequently encountered in the genesis of shoulder pain in the
hemiplegic patient." They were unsure of the etiology of the subsequent shoulder pain.
Poulin de Courval et al. (1990) examined 94 hemiplegic subjects involved in a
rehabilitation program after stroke and reported that subjects with shoulder pain had
significantly more spasticity of the affected limb than those without pain. In contrast,
Bohannon et al. (1986) conducted a statistical analysis of 50 consecutive hemiplegic
patients (36 with shoulder pain) and asserted that "spasticity ... was unrelated to
shoulder pain." Joynt (1992) also supported this finding after examining 67 patients with

shoulder problems following stroke. Nevertheless, evidence for spasticity in particular
hypertonic muscle imbalance, as a cause of hemiplegic shoulder pain is growing.

11.3.1 Spastic Muscle Imbalance

Hemiplegia following stroke is characterized by typical posturing reflecting hypertonic

muscle patterns. Flexor tone predominates in the hemiplegic upper extremity and
results in scapular retraction and depression as well as internal rotation and adduction
of the shoulder. This posture is the consequence of ablation of higher centers and
subsequent release of motor groups from pyramidal and extrapyramidal control. In
stroke recovery, this "synergy pattern" of muscles is inevitable where recovery is
incomplete. One consequence of this is the development of spastic muscle imbalance
about the shoulder joint.

Clinically the internal rotators of the shoulder predominate after a stroke involving that
arm and external rotation is one of the last areas of shoulder function to recover.
Hence, during recovery motor units are not appropriately recruited or turned off; the
result is simultaneous co-contraction of agonist and antagonist muscles. A shortened
agonist in the synergy pattern becomes stronger and the constant tension of the agonist
can become painful. Stretching of these tightened spastic muscles causes more pain.
Shortened muscles inhibit movement, reduce range of motion, and prevent other
movements especially at the shoulder where external rotation of the humerus is
necessary for arm abduction greater than 90 degrees. Muscles that contribute to
spastic internal rotation/adduction of the shoulder include the subscapularis, pectoralis
major, teres major and latissimus dorsi muscles. However, two muscles in particular
have been implicated as most often being spastic leading to muscle imbalance. These
are the subscapularis and pectoralis major muscles.

Subscapularis Spasticity Disorder

The subscapularis muscle originates on the undersurface of the scapula and inserts on
the lesser tuberosity of the humerus as well as the capsule of the shoulder joint (Figure
11.2). It is a major internal rotator of the shoulder (Hollinshead and Jenkins 1981). The
subscapularis muscle also participates in arm abduction and extension from a flexed
position (Cole and Tobis 1990). In normal individuals, nerve impulses to the
subscapularis are inhibited during arm abduction; the muscle then relaxes and allows
the humerus to externally rotate, thus preventing impingement of the greater tuberosity
on the acromion (Codman 1934). As part of the typical flexor synergy pattern in spastic
hemiplegics, internal rotators, including the subscapularis muscle, are tonically active.
This limits shoulder abduction, flexion and external rotation.

Bohannon et al. (1986) found limitation of external rotation of the hemiplegic shoulder
was the factor which most correlated with hemiplegic shoulder pain. Zorowitz et al.
(1996) also found that limitation in shoulder external rotation correlated strongly with
pain. Hecht (1995) specifically linked this problem to the subscapularis muscle when he
noted, "The subscapularis muscle is the primary cause of shoulder pain in spastic
hemiplegia where external rotation is most limited. Although other muscles may

contribute to spasticity, pain and functional contracture, the subscapularis is the
keystone of the abnormal synergy pattern."

Figure 11.2 The Subscapularis Muscle. The subscapularis muscle is a major internal rotator of the
shoulder. As part of the typical flexor synergy pattern in spastic hemiplegics, the subscapularis is
tonically active limiting not only external rotation but also shoulder abduction and flexion.

The subscapularis spasticity disorder is characterized by motion being most limited and
pain being reproduced on external rotation. A tight band of spastic muscle is palpated
in the posterior axillary fold. In support of this, Inaba and Piorkowski (1972) reported
external rotation was the most painful and limited movement of the hemiplegic shoulder.

Pectoralis Spasticity Disorder

The pectoralis major muscle serves to forward flex, adduct and internally rotate the arm.
Hecht (1995) has reported on a subset of hemiplegic patients with greater limitations in
abduction (and flexion) than on external rotation. In these patients a spastic pectoralis
major muscle appears to be problematic. This disorder is characterized by motion
being most limited and pain produced on abduction. A tight band of spastic muscle can
be palpated in the anterior axillary fold (Hecht 1995). It is also noteworthy that the
pectoralis major muscle is a synergist of the subscapularis muscle.

Figure 11.3 The Pectoralis Major Muscle. The pectoralis major muscle serves to adduct, internally
rotate and forward flex the arm at the shoulder.

11.3.2 Frozen or Contracted Shoulder

A frozen or contracted shoulder (adhesive capsuluitis) is characterized clinically by

limitations in range of movement, with a pattern of restriction. This condition is a
frequently identified source of pain in the spastic hemiplegic shoulder (Bohannon et al.
1986, Eto et al. 1980, Fugl-Meyer et al. 1975, Grossens-Sills and Schenkman 1985,
Hakuno et al. 1984, Risk et al. 1984).
Individual Studies

Table 11.6 Evidence of Frozen Shoulder Post Stroke

Author, Year Methods Outcome

Hakuno et al. 77 patients with hemiplegia caused by Contractures/adhesions were found in
1984 cerebrovascular accidents were randomly paralyzed shoulders at a statistically
Japan selected from all hemiplegic patients treated significant higher rate (54.6%) than in the
at rehab centre. Cases with dementia and non-paralysed side (32.5%). The
verbal dysfunction were eliminated. occurrence rate of contrast leakage from
Paralysis affected the right side in 35 a capsule tear on the subscapular bursa
patients and the left side in 42 patients. In and the bicipital tendon sleeve was
35 cases the affected arm was dominant higher on the non-paralysed side than on
whereas 42 cases had paralysis in non- the paralysed side. It was suggested
dominant arm. Positive contrast arthography that capsular contracture due to
was preformed on both shoulders of all hemiplegia reduces capsular tearing
patients. An anterior approach for injection during arthrographic maneuvers.
of the joint with contrast material was

Author, Year Methods Outcome
employed. The needle was inserted directly
into the glenohumeral joint space under
fluoroscopic control. Anteroposterior
radiographs were made in internal and
external rotation.
Crossen-Sills 21 male patients all received standard 67% of the patients entered the rehab
& physical therapy treatment. Shoulder pain, centre with signs of shoulder pain. An
Schenkman range of motion and subluxation were additional 10% developed initial signs of
1985 assessed on admission, three weeks post shoulder pain by 3 weeks post-admission
USA admission and at discharge. and another 5% developed signs of pain
at time of discharge. Positive correlation
noted between loss shoulder range and
increase in pain and between subluxation
and pain. There was no correlation
between subluxation and range of
motion. Suggestion that pain began in
the acute cares facility and worsened
while in rehab.
Rizk et al. Study of 30 spastic hemiplegic (18 with left 23 patients had capsular constriction
1984 hemiplegia and 12 with right hemiplegia) typical of frozen shoulder (adhesive
USA patients with painful ipsilateral shoulders capsulitis). 7 patients had normal
meeting the following criteria: maximum arthrograms. None showed rotator cuff
passive range of motion (ROM) of 600 of capsular tears. Electromyography
abduction, 900 forward flexion, 150 external revealed electrical silence in the shoulder
rotation, 450 extension; any stress at the limit musculature at rest.
of motion produced severe shoulder, with no
improvement during the previous 2 weeks,
no history of recent trauma to the affected
shoulder during the previous 2 weeks, no
history of seizures or anticonvuslant
medications; no clinical signs suggesting
shoulder-hand syndrome, no bone disease
or polyarthritis or previous shoulder pain
before stroke onset. All patients had shoulder
arthrograms performed. Electromyographic
studies were done on the deltoid, triceps,
and biceps brachii muscles on the muscles
on the involved side.
Bohannon et 50 patients whose hemiplegia was Of the 50 patients reviewed, 72% had
al. 1986 secondary to a stroke, whose unaffected shoulder pain. 20 had some pain while
USA shoulders demonstrated normal and pain- 16 had severe pain. Three zero-order
No Score free range of hemiplegia shoulder external correlations were significant: ROSER
rotation (ROSER, 900); able to adequately and shoulder pain (r=-0.061, p<0.001);
follow instructions to allow testing of all time since onset of hemiplegia and
variables pertinent to the study. Information shoulder pain (r=0.45, p<0.01); and time
was retrieved from patients’ records since onset of hemiplegia and ROSER
concerning their initial physical therapy (r=0.37, p<0.01). One-way ANOVA
evaluation. Relationships between pain and demonstrated that time since onset of
other variables were determined. hemiplegia (F=8.28, p<0.001) and the
ROSER (F=18.44, p<0.001) were
significantly different in patients with no
pain, some pain, and pronounced/severe
Lo et al. 2003 32 consecutive patients with shoulder pain 16 (54%) of patients had rotator cuff

Author, Year Methods Outcome
Taiwan following stroke were assessed for shoulder tears diagnosed by arthrograpy.
No Score subluxation, which was diagnosed by a gap
of more than one finger breadth between the
acrimonion and the head of the humeral
bone on palpation.

In summary, while shoulder subluxation is not always associated with shoulder pain,
spasticity generally is. The problem of hemiplegic shoulder pain appears to be due to a
combination of spastic muscle imbalance and a frozen contracted shoulder. However,
overaggressive stretching of the shoulder through an aggressive stretching program
may simply aggravate pain (see Treatment), as it does not address the issue of spastic
muscle imbalance.

Conclusions Regarding Spastic Hemiplegic Shoulder

There is an association between spasticity and the development of hemiplegic

shoulder pain.

Spasticity and subsequent frozen shoulder are the most likely causes of
hemiplegic shoulder pain.

Spasticity and hemiplegic shoulder pain are related.

11.4 Rotator Cuff Disorders

Because shoulder pain is so often associated with rotator cuff disorder in a non-stroke
population it should not be surprising that it would be seen as a potentially common
cause of hemiplegic shoulder pain. However, Risk et al. (1984) failed to demonstrate
any evidence of rotator cuff tears on arthrography in 30 patients with hemiplegic
shoulder pain (Risk et al. 1984). A similar study (Nepomuceno and Miller 1974)
reported a 33% incidence of rotator cuff tears in painful shoulders after strokes.
Najenson et al. (1971) reported that 13 of 32 (40%) patients with severe paralysis of the
upper extremity were found to have a rupture of the rotator-cuff ligament based on
arthrographic findings. Partial tears of the rotator cuff musculature are common and it is
always difficult determining whether they were present premorbidly even in previously
asymptomatic patients. Joynt (1992) diagnosed 67 stroke patients as having
hemiplegic shoulder pain. 28 patients received a subacromial injection of 1% lidocaine;
approximately half obtained moderate or marked relief of pain and improved range of
motion. However, this provides only indirect evidence of rotator cuff disorders as a

possible cause of hemiplegic shoulder pain. Generally, hemiplegic shoulder pain is not
commonly associated with rotator cuff disorders.

11.5 Functional Impact of Painful Hemiplegic Shoulder

A painful hemiplegic shoulder can be very limiting. It has the potential to further add to
the disability seen with hemiplegia.

Table 11.7 Impact of Painful Hemiplegic Shoulder on Function

Author, Year Methods Outcomes

Wankyln et 108 post stroke patients were studied. 63.8% of all patients developed hemiplegic
al. 1996 Patients were over the age of 60, shoulder pain (HSP). HSP was associated with
UK about to be discharged home and had reduced shoulder shrug and reduced pinch grip.
No Score suffered stroke with persisting Patients who required assistance with transfer
disability defined as a Barthel Index were more likely to suffer with HSP. Significantly
score less than 20. more patients with Barthel Index scores less than
15 reported HSP compared to those with a score
between 15-20 at both discharge (59% vs. 25%)
and at 8 weeks (77% vs. 51%).
Aras et al. 85 consecutive stroke patients 54 patients (63.5%) had shoulder pain. Shoulder
2004 admitted to one of the largest rehab pain was reported more frequently among patients
Turkey facilities in Turkey were studied to with reflex sympathetic dystrophy, lower motor
No Score identify the incidence of shoulder functional level of shoulder and hand, subluxation,
pain. and limitation of external rotation and flexion of
shoulder. Age was also associated with the
development of shoulder pain. No relationships
were found between shoulder pain and sex, time
since onset of disease, hemiplegic side,
pathogenesis, spasticity, neglect, and thalamic
pain, or extension of hospital stay.
Roy et al. 76 patients suffering from a first Shoulder pain on movement was associated with
1995 stroke were studied. Separate increased LOS, poorer performance on ADL, arm
UK regression equations were created for function and arm power. Shoulder pain was a
No Score the outcomes of length of hospital statistically significant predictor of arm function.
stay (LOS), performance of ADL
(Barthel Index), arm function
(Frenchay arm function) and arm
power (Motricity Index)

Wanklyn et al. (1996) and Roy et al. (1995, 1996) both demonstrated an association
between hemiplegic shoulder pain and poor functional outcomes. However, a cause
and effect relationship has not yet been established.

Conclusions Regarding Functional Impact of Hemiplegic Shoulder Pain

The development of painful hemiplegic shoulder is associated with severe

strokes and poorer functional outcome.

11.6 Management of the Painful Hemiplegic Shoulder
Management of the painful hemiplegic shoulder, once the condition has developed, is
difficult and response to treatment is frequently unsatisfactory (Risk et al. 1984). The
best treatment approach has not been definitely established, in part, due to the
uncertainty of the etiology of the pain. As a result, a wide variety of treatments have
been used, with varying degrees of success (Snels et al. 2002). Ideally, measures
should be taken immediately following stroke to minimize the potential for the
development of shoulder pain. Early passive shoulder range of motion, and supporting
and protecting the involved shoulder, in the initial flaccid stage are regarded as
important steps to reduce the development of shoulder pain.

11.6.1 Positioning of the Hemiplegic Shoulder

The muscles around the hemiplegic shoulder are often paralyzed, initially with flaccid
tone and later with associated spasticity. Careful positioning of the shoulder serves to
minimize subluxation and later contractures as well as possibly promote recovery, while
poor positioning may adversely affect symmetry, balance and body image.
As cited by Gilmore et al. (2004), Davies (2000) suggests that through careful and
correct positioning, the development of shoulder pain can be prevented. Bender and
Mckenna (2001) have noted that a primary goal of early stroke management is to
prevent the development of hypertonicitiy (Johnstone 1992) and to discourage
inefficient patterns (Bobath 1990). Bender and McKenna (2001) noted that the
“recommended position for the upper limb is towards abduction, external rotation and
flexion of the shoulder,” however, from Carr and Kenny (1992) review, Bender and
McKenna cite that “most popular theories failed to yield consensus for exact degrees of
the positioning.”

Individual Studies

Table 11.8 Positioning of the Shoulder in Stroke Patients

Author, Year Methods Outcomes

PEDro Score
Carr and Review of the literature. General agreement about postures with the
Kenney 1992 shoulder protracted, the arm brought forward,
No Score the spine straight and fingers extended. Still
controversy regarding height of the arm in the
forward position.
Dean et al. 23 patients were randomized to receive Changes in active and passive range of motion
2000 an experimental therapy or to a control were not significant between the groups with
Australia group. Subjects in both groups the level of pain remaining unchanged.
5 (RCT) participated in a multidisciplinary
rehabilitation program and participated in
active training of reaching and
manipulation tasks. The experimental
group received prolonged positioning to
the affected shoulder each day, five days

Author, Year Methods Outcomes
PEDro Score
a week for six days (positioning).
Ada et al. 36 stroke patients were randomized to Positioning the shoulder in maximal external
2005 an intervention or a control condition. rotation (position 1) significantly reduced the
Australia Patients in the experimental group development of contractures, compared to the
6 (RCT) received two, 30-minute sessions of control group. In position 2 (where patients sat
sustained shoulder positioning. Patients with the affected arm resting on a table with the
in both groups received 10 minutes of shoulder at 900, for 30 minutes daily), did not
shoulder exercises and routine upper prevent the development of contractures.
limb care. The treatment was provided
for 4 weeks. Assessments of contracture
were taken at weeks 2 and 6 after

Conclusions Regarding Positioning of the Hemiplegic Shoulder

There is consensus (Level 3) opinion that proper positioning of the hemiplegic

shoulder helps to avoid subluxation. However, there is conflicting (Level 4)
evidence that prolonged positioning does not influence active and passive range
of motion or level of pain.

Further research is needed before conclusions regarding positioning of the

hemiplegic shoulder can be made.

11.6.2 Slings and Other Aids

Arm slings are often used in the initial stages following a stroke to support the affected
arm. However, their use is controversial and they can have disadvantages in that they
encourage flexor synergies, inhibit arm swing, contribute to contracture formation and
decrease body image causing the patient to further avoid using that arm. However, a
sling remains the best method of supporting the flaccid hemiplegic arm while the patient
is standing or transferring. Ada et al (2005) conducted a systematic Cochrane review
evaluating the benefit of shoulder slings and supports, and concluded that there is
insufficient evidence that these devices reduce or prevent shoulder subluxation
following a stroke. The review included only four RCTs (Ancliffe et al. 1992, Griffin et al.
2003 [unpublished data], Hanger et al. 2000 and Hurd et al. 1974). The results are
presented in Table 11.9.

Table 11.9 Results from Systematic Review (Ada et al. 2005)

Outcome Studies Included & Intervention Peto Odds ratio (95% CI) or
Weighted Mean Difference
(WMD) (95% CI)
Proportion of patients with Hurd et al. 1974 – hemisling OR 8.7 (1.1, 67.1)
pain at follow up (Favours no slings)
Number of days pain was Ancliffe et al. 1992 -strapping WMD 14 days (9.7, 17.8)
delayed with treatment Griffin et al. 2003 (Favours slings)
Pain scores on VAS Hanger et al. 2000 – strapping WMD 0.83 (-1.46, 3.12)
(10 cm scale) (No difference)
Motor Assessment sub scores Hanger et al 2000. -strapping WMD 0.8 (-1.5, 3.1)
(0-18) (No difference)
Proportion of patients with Hurd et al 1974- sling OR 1.00 (0.1, 9.3)
contractures (No difference)
Range of shoulder external Hanger et al. 2000 - strapping WMD –1.4 degrees (-10.9, 8.10)
rotation at end of follow up (No difference)

As tone returns to the shoulder muscles, the risk of shoulder subluxation decreases and
slings can then be withdrawn. Slings tend to hold the limb in a poor position, which may
accentuate the adduction and internal rotation posture and may contribute to shortening
of tonically active muscles. The best method to support the shoulder has yet to be
determined. In the absence of empirical evidence of their efficacy, many devices are
available and in common use, including a variety of slings and lapboards.

Individual Studies

Table 11.10 Slings and Other Aids in Hemiplegic Shoulders

Author, Year Methods Outcomes

PEDro Score
Hurd et al. 14 patients were alternately Of the 7 patients without slings, 5 had no pain, while 2
1974 assigned to be treated with a had little pain. Of the 7 patients treated with slings, 6
USA sling or without a sling, assessed had little pain, while 1 had no pain.
No Score 2 to 3 weeks and 3 to 7 months
post stroke.
Moodie et al. Series of radiographs Shoulder roll and Hook-Hemi Harness did not reduce
1986 (anterior/posterior view) of 10 subluxation to the same extent as the other 3 devices.
Canada patients’ affected and unaffected Subluxation was reduced within 20% of the correct
No Score limbs in order to permit alignment in 8 out of 10 patients when treated with the
comparisons for degree of sling; 6 of the 10 treated with the arm trough, and 7 of
subluxation and the 5 aids to be the 10 patients treated lap tray. Suggested that the
evaluated were then applied to sling, trough, and lap tray reduced the mean
the patients’ affected arm and an subluxation to within .56 cm of normal control while the
A/P view was taken of each: roll and hook tended to under-correct the subluxation.
conventional sling; shoulder roll;
Hook-Hemi Harness; arm trough;
and plexiglass lap tray.
Williams et al. Radiographs were taken of 26 There was no significant difference in the reduction of
1988 hemiplegic patients with inferior subluxation between the two types of shoulder
Canada subluxated shoulders with two supports. However, there were significant differences in
No Score different supports-the Bobath subluxation between measurements of the unsupported

Author, Year Methods Outcomes
PEDro Score
shoulder roll and the Henderson affected shoulder and the unaffected shoulder and
shoulder sling. Radiographs of between measurements of the unsupported affected
the unsupported affected shoulder and the supported affected shoulder using
shoulder were compared with both supportive devices.
radiographs of the same shoulder
with each support applied.
Radiographs of the unaffected
shoulder were used as a
comparison in determining the
amount of subluxation.
Brooke et al. Three different shoulder supports Harris hemi-sling improved correction of the
1991 were applied to 10 patients by subluxation with mean vertical distance of 37.8mm vs
USA their occupational and physical 38.5mm compared to the uninvolved shoulder while the
No Score therapists: Harris hemi-sling, the mean difference between Harris and Bobath sling was
Bobath sling and the arm trough 5.5(2.9) mm, in favour of the Harris sling. For
or lapboard. horizontal measurement, mean difference between
Harris and Bobath slings was 8.3(6.3) mm, in favour of
the Harris sling.
Zorowitz et An occupational therapist applied The single-strap hemisling corrected vertical
al. 1995 each shoulder support to each of displacement, while the Roylan and Bobath roll
USA 20 patients in the following order: significantly reduced vertical displacement. The
No Score (1) single-strap hemisling; (2) Bobath roll and the Cavalier support produced a
Rolyan humeral cuff sling; (3) significant lateral displacement of the humeral head of
Bobath roll; and (4) Cavalier the affected shoulder compared with the unaffected
support. shoulder. The Roylan humeral cuff sling significantly
decreased the total subluxation asymmetry.

Conclusions Regarding Slings in Hemiplegic Shoulder

There is limited (Level 2) evidence that shoulder slings prevent subluxation

associated with hemiplegic shoulder pain, although there is also limited (Level 2)
evidence that one device or method is no better than another.

There is limited evidence that shoulder slings influence clinical outcomes.

11.6.3 Strapping the Hemiplegic Shoulder

Strapping the hemiplegic shoulder is used as a method to prevent or reduce the severity
of shoulder subluxation and may provide some sensory stimulation. There are three
different forms of strapping the hemiplegic shoulder, which have been described

Ancliffe 1992: 5-cm wide lightweight adhesive tape (Fixomull Stretch), “the first length of
tape was applied to the shoulder half way along the length of the clavicle, continued

across the deltoid muscle in a diagonal direction... the tape was terminated
approximately one-quarter of the way of the along the spine of the scapula. A second
length of tape was applied in the same direction as the first but 2 cm below. A small
length of tape was applied over the shoulder to secure the ends”.

Morin & Bravo 1997: “A 10 cm-wide Elastoplast adhesive bandage was applied under
tension from the forearm under the olecranon laterally to the top of the shoulder. Two
other 7.5 cm-wide bandages were applied from the olecranon under the forearm to the
forearm to the top of the shoulder, with one passing anteriorly over the clavicle and the
other posteriorly covering the spine of the scapula. No free space was left between the

Hanger et al. 2000: Three lengths of nonstretch Elastoplast Sports tape were used.
“The two main supporting tapes were applied first. Both were applied using a lifting
action, starting 5 cm above the elbow, and moving up the arm front and back, crossing
at the top of the shoulder. The posterior arm tape was then anchored down past the
clavicle whereas the tape from the anterior aspect of the arm came across the shoulder
and down past the spine of the scapula. They were both supported at the lower end by
a short tape to prevent them peeling off”.

Individual Studies

Table 11.11 Strapping the Hemiplegic Shoulder

Author, Year Methods Outcomes

PEDro Score
Ancliffe 1992 A pilot study of 8 patients who were Patients in the strapping group experienced
Australia assigned to receive strapping of the a significantly longer pain free period than
No Score shoulder applied by one physiotherapist the patients who were not strapped (21 vs.
and changed every 3 to 4 days as 5.5 days). However, all patients in the
needed to the hemiplegic side or to strapping group eventually did experience
receive no strapping. Treatment began pain. The longest pain-free period was 25
within 48 hours of admission to hospital. days.
Hanger et al. 98 patients were randomized to have No significant differences were found
2000 their affected shoulder strapped for 6 between groups on measure of pain, range
New Zealand weeks in addition to standard of movement and on functional outcome
7 (RCT) physiotherapy or to receive standard measures. There was a trend for less pain at
physiotherapy only 15 days following 6 weeks and improved functional upper limb
stroke. function for the strapping group.
Griffin & 33 patients at risk of developing One person in the TS group developed
Bernhardt hemiplegic shoulder pain were shoulder pain compared to 5 persons in the
2006 randomized to therapeutic shoulder other 2 groups. Subjects in the TS group
Australia (TS) strapping, sham shoulder (SS) reported 26 pain-free days, compared to 19
7 (RCT) strapping or to a no strapping (control) days for SS subjects and 16 days for
group 10 days post stroke. Lightweight subjects in the control group. The difference
adhesive tape was used and strapping was statistically significant for the
continued for over 4 weeks. Primary comparison of TS and control group. There
outcome was the number of pain free were no differences between groups on any
days. Secondary outcomes included of the secondary outcomes.
range of movement, Motor Assessment

Author, Year Methods Outcomes
PEDro Score
scale scores and Modified Ashworth

Conclusions Regarding Strapping the Hemiplegic Shoulder

There is conflicting (Level 4) evidence that strapping the hemiplegic shoulder

reduces the development of pain. There is moderate (Level 1b) evidence that
strapping does not improve upper limb function or range of motion.

Strapping the hemiplegic shoulder does not appear to improve upper limb
function, but may reduce pain.

11.6.4 Active Therapies in the Hemiplegic Shoulder

The association of spasticity, muscle imbalance and a frozen shoulder with shoulder
pain suggests that a therapeutic approach designed to improve range of motion of the
hemiplegic shoulder will improve pain.

Individual Studies

Table 11.12 Active Therapies in Hemiplegic Shoulder

Author, Year Methods Outcomes

PEDro Score
Inaba et al. 33 patients with hemiplegia who experienced No significant differences between the
1972 shoulder pain in the range of 0-90 degrees of groups were observed in measures of
USA flexion or abduction of the arm after stroke ROM.
7 (RCT) were treated. Patients were randomly
assigned to 1 of 3 groups: Range of motion
(ROM) exercises and positioning group;
ROM exercises and ultrasound; or ROM
exercises and mock ultrasound. All patients
received ROM exercises for 4 weeks and
given a minimum of 15 treatments.
Kumar et al. 28 patients were assigned to receive a Significant difference in the incidence of
1990 rehabilitation program of range of motion by pain reported between the groups.
USA therapist (ROMT) once a day, 5 days a Shoulder pain was more common in the
5 (quasi- week; or a rehabilitation program with use of overhead pulley (63%) group than in the
randomized skate board once a day, 5 days a week; or a ROMT group (8%). ROM was
controlled rehabilitation program with use of overhead significantly reduced in those patients
trial) pulley once a day, 5 days a week while an who developed shoulder pain when
inpatient on a stroke rehabilitation unit. compared to those who did not develop

Author, Year Methods Outcomes
PEDro Score
shoulder pain motion abduction, forward
flexion, internal rotation and external
rotation. Shoulder subluxation was found
in 46% of all patients with no significant
difference between treatment groups.
Patridge et al. 65 patients were randomized to receive A greater proportion of patients treated
1990 cryotherapy or Bobath therapy daily for five by the Bobath method reported no pain
UK days and then after at the therapist’s or only occasional pain on exit of the
5 (RCT) discretion for a total of four additional weeks study compared to those treated by the
and assessed by a blinded investigator. cryotherapy method.
Poduri et al. Patients with stroke experiencing shoulder A significantly greater proportion of
1993 pain after completing outpatient therapy were patients receiving the treatment drug
USA studied. One group of patients received prior to therapy experienced pain relief.
No Score either a nonsteroidal anti-inflammatory drugs Flexion, abduction and functional
(Ibuprofen 400-800g tid, and Sulindac, 150 recovery were significantly greater in
mg bid.) taken 30 to 60 minutes prior to those patients who were taking the
occupational therapy. A second group of nonsteriodal anti-inflammatory drug
patients received only occupational therapy before therapy.
consisting of range of motion, active assistive
and strengthening exercises and activities of
daily living training.
Tyson & 22 stroke patients with consequential Mean shoulder flexion for the axilla hold
Chissim weakness of the arm instructed to hold the was 115.2 degrees and 97.7 degrees for
2002 hemiplegic shoulder at: (1) an axilla hold the distal hold (p < 0.001).
UK involving shoulder support and (2) a distal
4 (RCT) hold without shoulder support. Each hold
was repeated to obtain three measurements
with order of testing randomized.
Lynch et al. 35 stroke patients with significant upper There were no between group
2005 motor impairment were randomized to a differences in changed scores between
USA control group (n=16), which received self- groups on any of the outcome measures
6 RCT range of motion exercises under the assessed (joint stability, Modified
supervision of a physiotherapist or to the Ashworth scale, Fugyl-Meyer (pain and
experimental group (n=19) of continuous self-care FIM scores).
passive motion treatments with the use of a
device (25 min sessions, 5 days/week until
discharge). All patients received
rehabilitation therapies for 3.5 hours per day.
Gustafson & 34 patients with upper extremity hemiparesis There were no significant between group
McKenna admitted within 100 days of stroke were differences reported for any of the
2006 randomized to a participated in a programme outcomes. Over time, all participants
Australia of two static positional stretches, each held reported decreased range of motion,
6 (RCT) for 20 minutes, once daily or to a control motor recovery and functional
condition where the affected arm was independence. The control group
supported when seated in bed. Outcome reported a decrease in pain while the
measures included: pain-free passive range treatment group reported an increase.
of motion, shoulder pain (Ritchie Articular
Index), Motor recovery (Motor Assessment
Scale) and functional independence
(modified BI) measured at hospital admission
and discharge.


Inaba et al. (1972) in a “good” (PEDro = 7) study found no significant differences in the
outcomes of patients who received: ROM exercises and positioning, ROM exercises
and ultrasound or ROM exercises and mock ultrasound. Kumar et al. (1990) found that
overhead pullies caused dramatically higher levels of shoulder pain than more
restrained ROM exercises. Although there were no statistically significant differences in
change scores between the control and the experimental group, Lynch et al. (2005)
reported a trend towards improvement in the area of shoulder joint stability associated
with continuous passive motion using the OrthoLogic Danniflex600 shoulder CPM
system. A programme of positional static stretches was not only ineffective in reducing
loss of range of motion into external rotation, but was also associated with increasing
levels of pain (Gustafson & McKenna 2006). Counter to previous research, subjects in
this study continued to improve functionally, despite a loss of range of motion and
increasing pain, reported by subjects in the treatment group.

Shoulder Pain in Hemiplegia: The Role of Exercise (Kumar et al. 1990)

To assess the occurrence of pain in patients in patients undergoing rehabilitation of

hemiplegia, 28 patients were assigned to one of three exercise programs commonly used in
the rehabilitation of hemiplegia: 1) range of motion by the therapist (ROMT), 2) skate board on
a table (SB) and 3) overhead pulley (OP).

Pain Experience of Groups

% in Pain


Comparing the number of patients who developed pain in each group, there were dramatically
more patients in the OP group experiencing pain after rehabilitation compared to the other two
groups. Patients in the ROMT group experienced the least amount of pain after rehabilitation.

Partridge et al. (1990) found that treatments using Bobath therapy resulted in
significantly less pain than cryotherapy. The general message that emerges from these
three studies is that an active ROM exercise approach is preferable to more passive
modalities but an overly aggressive approach (i.e. overhead pullies) resulted in a very
high incidence of hemiplegic shoulder pain when compared to a gentler approach.

Conclusions Regarding Active Therapies in the Hemiplegic Shoulder

There is moderate (Level 1b) evidence that aggressive range of motion therapies,
using overhead pullies results in increased rates of shoulder pain.

There is moderate (Level 1b) evidence that Bobath therapy for the hemiplegic
shoulder is associated with greater pain reduction than passive cryotherapy
(application of local cold therapy).

There is moderate (Level 1b) evidence that gentle exercises to improve range of
motion are the preferred approach. There is moderate (Level 1b) evidence that
adding ultrasound therapy to range of motion exercises does not change

There is limited (Level 2) evidence that providing an oral nonsteriodal anti-

inflammatory drug leads to less pain, improved range of motion and improved
functional recovery in stroke patients with shoulder pain receiving occupational

There is moderate (Level 1b) evidence that static positional stretches performed
daily during rehabilitation are associated with increasing pain and decreasing
range of motion.

Aggressive range of motion exercises (i.e. pullies) results in a markedly increased

incidence of painful shoulder; a gentler range of motion program is preferred.
Adding ultrasound treatments is not helpful while NSAIDs may be helpful.

11.6.5 Injections in the Hemiplegic Shoulder

Corticosteroid and/or local anesthetic injections are commonly used for shoulder pain, in
particular rotator cuff tendinopathies. It is not surprising that there would be interest in
this mode of treatment in hemiplegic shoulder pain.

Individual Studies

Table 11.13 Injections in Hemiplegic Shoulder

Author, Year Methods Outcomes

PEDro Score
Joynt 1992 A retrospective study following 67 stroke 50% of the 28 patients who underwent
USA patients with hemiplegic shoulder pain. a subacromial injection of local
No Score Twenty-eight patients received a subacromial anaesthetic obtained moderate or
injection of 1% lidocaine. marked pain relief and improved ROM.
Bhakta et al. 17 patients received a single course of Shoulder pain improved in 6 of 9

Author, Year Methods Outcomes
PEDro Score
1996 intramuscular injection of botulinum toxin to patients. Pain resolved completely in
UK biceps brachii, flexor digitorum profundus, two patients. Benefit lasted for up to 11
No Score flexor digitorum superficialis and flexor carpi months.
ulnaris an average of 1.5 years post stroke.
Dosages of 400-1,000 MU of Dysport or 400
MU of Botox were administered.
Dekker et al. 9 patients with a presence of shoulder pain in Intra-articular injections of
1997 the paretic arm with disturbances of sleep and triamcinolone acetonide demonstrated
Netherlands with the presence of ROM restriction of a significant reduction in pain with
No Score external rotation partook in a multiple-baseline highly significant effect in 5 of the 9
(AB) study of tramicinolone acetonide. The patients. Range of motion improved in
treatment condition (phase B) was 4 weeks 4 out of 7 patients but improvement did
long, during which three intra-articular not reach statistical significance at the
injections of triamcinolone acetonide were group level.
administered at day 1, 8 and 22.
Snels et al. 35 patients demonstrating hemiplegia after No significant improvement was
2000 stroke and pain in hemiplegic shoulder greater observed for any of the primary
Netherlands than 4 on the visual analogue scale (VAS) outcome measures with triamicinolone
8 (RCT) with limitation of passive external rotation of acetonide treatment.
the hemiplegic shoulder were randomized to
receiver either three injections with
triamicinolone acetonide (40 mg Kenacrot A-
40 in 1ml) or three placebo injections (1 ml
saline solution).


One RCT (Snels et al. 2000), of “good” quality (PEDro = 8), failed to show a benefit of
corticosteriod injections. There is insufficient evidence to recommend this mode of
treatment and one trial casting doubt on its efficacy. A single uncontrolled study by
Bhakta et al. (1996) evaluated the use of botulinum toxin in the treatment of shoulder
pain in an uncontrolled study. The majority of patients responded to treatment.

Conclusions Regarding Injections in the Hemiplegic Shoulder

There is moderate (Level 1b) evidence, based on one “good” RCT that
corticosteroid injections do not improve shoulder pain or range of motion in
patients with hemiplegia.

There is limited (Level 2) evidence that botulinum toxin can reduce pain in the
hemiplegic shoulder.

Corticosteroid injections do not improve hemiplegic shoulder pain or range of


11.6.6 Functional Electrical Stimulation (FES) in the Hemiplegic Shoulder

As cited by Gresham et al. (1995), the U.S. AHCPR Post Stroke Rehabilitation
Guidelines defines FES as “bursts of electrical stimulation applied to the nerves or
muscles affected by the stroke, with the goal of strengthening muscle contraction and
improving motor control.” The supraspintus and posterior deltoid muscles are most
likely to be treated as they are important muscles in maintaining the correct alignment of
the glenohumeral joint (Paci et al. 2005). Theoretically, FES should help to compensate
or facilitate flaccid shoulder muscles, which in turn should reduce the risk of shoulder
subluxation. The ideal intensity of treatment is thought to be 6 hours daily, five days a
week for 6 weeks. FES is performed at frequencies of between 35 to 50 Hz (Paci et al.

Price & Pandyan (2001) conducted a systematic review of all forms of electrical
stimulation (ES) used in the prevention and treatment of post stroke shoulder pain. The
included studies and the results are presented in Tables 11.14(a) and 11.14(b). The
authors concluded that there was insufficient evidence from which to draw conclusions.
There was evidence that FES, in addition to conventional therapy, improves function but
is not superior for preventing pain.

Table 11.14(a) Studies included in the Systematic Review authored by Price & Pandyan (2001)

Study Intervention Length of Treatment

Faghri et al. 1994 FES vs. no sham treatment 6 weeks
Leandri et al. 1990 Sham treatment vs. high intensity TENS vs. 4 weeks
low intensity TENS
Linn et al. 1999 No sham treatment vs. electrical stimulation 4 weeks
(not FES or TENS)
Sonde et al. 1998 No sham treatment vs. low frequency TENS 3 months

Table 11.14(b) Results From Studies Evaluating Any form of ES in the Treatment and Prevention
of Shoulder Pain

Outcome Significant Odds Ratio* or Weighted (or

Result (Y/N) Standardized) Mean Difference
and 95% CI
New cases of shoulder pain No 0.64 (0.19, 2.14) *
Pain intensity rating change from baseline No 0.13 (-1.00, 1.25)
Passive humeral lateral rotation Yes 9.17 (1.43, 16.9)
Motor score change from baseline No 0.24 (-0.14, 0.62)
Glenohumeral subluxation compared to baseline Yes -1.13 (-1.66 , -0.60)
Spasticity score change from baseline No 0.05 (-0.28,0.371)

Ada &Foongchomcheay (2002) also conducted a meta-analysis to examine the effect of

electrical stimulation on shoulder subluxation following stroke. This review included the
results from 6 RCTs (Baker & Parker 1986, Faghri et al. 1994, Kobayashi et al. 1999,
Linn et al. 1999, Wang et al. 2000). The results, presented in Table 11.15 suggest that
early treatment following stroke helps to prevent the development of hemiplegic
shoulder while later treatment helps to reduce pain, in addition to conventional therapy.

Table 11.15. Pooled Results from Ada & Foongchomcheay (2002)

Outcome Intervention Significant Result Weighted Mean

(Y/N) Difference and 95% CI
Shoulder subluxation (mm) Early ES + CT* Yes 6.5 (4.4, 8.6)
vs. early CT
Shoulder subluxation (mm) Late ES + CT vs. No 1.9 (-2.3, 6.1)
Late CT
Function (Bobath assessment Early ES + CT vs. Yes 18.6 (0.4, 36.7)
chart, Motor assessment scale early CT
and Fugl-Meyer) expressed as a
Function (Bobath assessment Early ES + CT vs. No 14.4 (-5.4, 34.2)
chart, Motor assessment scale early CT
and Fugl-Meyer) expressed as a
Pain (range of motion-degrees) Early ES + CT vs. No 3.7 (-1.2, 8.6)
early CT
Pain (Visual analogue scale-cm) Late ES + CT vs. Yes 1.6 (0.1, 3.0)
Late CT
* CT= conventional therapy; ES= electrical stimulation

Individual Studies

Eleven studies specifically evaluated the effects of FES on the treatment of shoulder
pain. (Table 11.16).

Table 11.16 FES in Hemiplegic Shoulder

Author, Year Methods Outcomes

PEDro Score
Baker & 63 patients with a minimum of 5 mm of At six weeks, the mean subluxation of the study
Parker 1986 shoulder subluxation in their involved upper group was significantly less compared to the
USA extremity were randomized to a treatment or control (8.6 vs. 13.3). Three- month
4 RCT control group. Patients in the treatment radiographs demonstrated that patients in the
group received neurosmuscular electrical treatment group had lost an average of 1-2 mm,
stimulation (NMES) for 5 weeks, while which had been achieved during the study
patients in the control group used period. The authors did not demonstrate a
conventional hemi-slings or wheelchair arm causal relationship between subluxation and
supports. shoulder pain.
Faghri et al. 26 patients were randomized to receive After treatment, the FES group showed a
1994 either functional electrical stimulation (FES) significant increase in arm function, tone and
USA in which two flaccid/paralyzed shoulder EMG activity compared to control patients.
4 (RCT) muscles (supraspinatus and posterior
deltoid) were induced to contract repetitively
up to 6 hours daily for six days in addition to
conventional therapy (FES) or to receive
conventional therapy (control).
Faghri & 26 patients received conventional physical Weekly evaluation of arm and shoulder muscle
Rodgers therapy with 13 patients randomized to the function (range of motion), tone, and posterior
1997 treatment group, which received additional deltoid muscle EMG activity showed significant
USA functional neuro-muscular stimulation (FNS) improvement in the experimental group when
4 (RCT) therapy for 6 weeks. compared to the control group over 6 weeks.

Author, Year Methods Outcomes
PEDro Score
Advantage of treatment group was maintained
6 weeks after termination of the FNS.
Chantraine et 115 patients were alternately assigned to Significant motor recovery was noted in favour
al. 1999 receive traditional Bobath treatment in of FES treatment at three months and was
Switzerland addition to functional electrical stimulation maintained at 24 months. Significant reduction
4 (quasi- (FES) for 5 weeks or to receive traditional also noted in pain in favour of FES treatment at
randomized Bobath treatment for 5 weeks. three months and again maintained at 24
controlled months. Significant reduction in shoulder
trial) subluxation in favour of FES treatment was
noted at three months and maintained 24
months post-treatment.
Kobayshi et 17 patients were randomized to receive Difference in subluxation in group S, 2.8(3.6)
al. 1999 therapeutic electrical stimulation (TES) for mm and group D, 2.8(2.5) mm was significantly
Japan 15 minutes twice a day to either the greater than that of the control –1(2.8)mm
5 (RCT) suprasupinatus muscle (group S) or middle under the stress test. The mean abduction
deltoid muscle (group D) in conjunction with force tended to increase in group S and was
conventional therapy, or to receive significantly greater in group D.
conventional therapy only.
Linn et al. 40 patients were randomly assigned to a The treatment group had significantly less
1999 control or treatment group. Patients in the subluxation and pain after the treatment period,
Scotland treatment group received electrical but at the end of the follow-up period there were
6 (RCT) stimulation (ES) 4 times daily, ranging from no significant differences between the 2 groups.
30 minutes in week one to 60 minutes by
week 4. Patients in both groups received
daily occupational and physical therapy.
Treatments lasted for four weeks.
Assessments of shoulder subluxation, pain,
and motor control. were carried out at 4
and 12 weeks after stroke
Wang et al. 32 inpatient and outpatient rehabilitation The experimental subgroup of short duration
2000 patients with hemiplegia were assigned to showed significant improvements in reducing
Taiwan one of two groups based on the duration of subluxation as indicated by x-ray compared
5 (RCT) hemiplegia: the short and the long duration. with the control subgroup of short duration after
Subjects in each group were randomly the first FES treatment. The same effect was
assigned to either a control subgroup or an not shown for the experimental subgroup of
experimental subgroup. Subjects in the long duration. The second FES treatment
experimental subgroups were treated in a program only resulted in an insignificant change
type A-B-A study design, which consisted of of shoulder subluxation for both the short- and
an FES training (A), routine therapy or long-duration subgroups.
regular daily activity without FES training
(B), and another FES training (A). Each
period lasted for 6 wk. FES training
program, consisted of five sessions/week.
Yu et al. 2001 8 patients participated in six weeks of At end of treatment, there was a significant
USA percutaneous intramuscular electric improvement of shoulder subluxation, pain,
No Score stimulation (per-NMES). shoulder pain-free rotation and in FIM scores.
Further improvements were noted at the 3
month follow up in subluxation, pain, Fugl-
Meyer, shoulder pain-free external rotation and
in FIM scores.
Yu et al. 2001 10 hemiplegic stroke patients with at least 1 Pain scores were significantly lower for perc-
USA fingerbreadth of glenohumeral subluxation NMES than trans-NMES as assessed by the

Author, Year Methods Outcomes
PEDro Score
6 (RCT) received 3 randomly ordered pairs of VAS and the MPQ.
neuromuscular stimulation (NMES) to the
suprasinatus and poterior deltoid muscles of
the subluxated shoulder. The stimulation
types were percutaneous-NMES (perc-
NMES) and transcutaneous-NMES (trans-
NMES). After each stimulation pain was
evaluated with the visual analog scale
(VAS) and the McGill Pain Questionnaire
Renzenbrink 15 stroke survivors with chronic (> six A significant reduction in pain was found on the
& Ijerman months) hemiplegia and a therapy-resistant Brief Pain Inventory. Pain reduction was still
2004 painful shoulder with subluxation were present at six months follow-up. All domains, in
Netherlands studied. Shoulder subluxation was indicated particular bodily pain, of the SF-36 showed
No Score by at least 1/2 fingerbreadth of improvement in the short term. After six months
glenohumeral separation on palpation. of follow-up, bodily pain was still strongly and
Patients received 6 hours of Percutaneous significantly reduced, whereas social
Neuromuscular electrical stimulation (P- functioning and role physical demonstrated a
NMES) per day for a total of six weeks. nonsignificant improvement of more than 10%
compared with baseline.
Yu et al. 7 site, single-blinded, randomized clinical The main outcome measure was the Pain
2004 trial. 61 chronic stroke survivors with outcome measure Brief Pain Inventory question
USA shoulder pain and subluxation volunteered 12 (BPI 12), an 11-point numeric rating scale,
7 (RCT) to be randomized to receive intramuscular assessed at the end of treatment, and at 3 and
neuromuscular electric stimulation (NMES) 6 months post treatment. Post stroke shoulder
to the supraspinatus, posterior deltoid, pain was significantly less in the NMES group
middle deltoid, and trapezius for 6 hours a compared to control at 3 and 6 months
day for 6 weeks or to the control condition following treatment.
of a cuff-type sling for 6 weeks. Main
outcome measure Brief Pain Inventory
Chae et al. 12 month follow up from Yu et al. 2004. 22 patients in the intervention group and 21
2005 Treatment success was defined as a patients received 12-month assessments. The
USA minimum 2-point reduction in Brief Pain electrical stimulation group exhibited a
7 (RCT) Inventory question 12 at all post treatment significantly higher success rate than controls
assessments. Secondary measures (63% vs. 21%). Repeated-measure analysis of
included pain-related quality of life (Brief variance revealed significant treatment effects
Pain Inventory question 23), subluxation, on post treatment Brief Pain Inventory
motor impairment, range of motion, questions 12 and 23. There were no other
spasticity, and activity limitation. significant between group differences.


All of the RCTs reviewed reported a benefit associated with FES treatment, although
there was variability in the outcomes assessed: range of motion, muscle tone, EMG
activity, shoulder subluxation, shoulder pain and muscle function. The results suggest
that FES can reduce pain in the affected shoulder and also improve upper extremity
function. Percutaneously placed devices may improve treatment compliance.

Conclusions Regarding FES in the Hemiplegic Shoulder

There is strong (Level 1a) evidence that functional electrical stimulation improves
a number of clinical outcomes associated with the hemiplegic shoulder.
Significant improvements have been reported for muscle function, tone, EMG
activity, pain, subluxation and range of motion. The improvement was
maintained for up to 24 months follow-up.

Functional electrical stimulation improves a number of hemiplegic shoulder

clinical outcomes.

11.6.7 Surgery as Treatment for Muscle Imbalance

Given that spastic muscle imbalance has been identified as a cause of hemiplegic
shoulder pain, treatment designed to reverse that imbalance could potentially relieve
hemiplegic shoulder pain.

Table 11.17 Surgery for Muscle Imbalance in the Hemiplegic Shoulder

Author, Year Methods Outcomes

PEDro Score
Braun et al. Surgical procedure on 13 patients to release the 10 of the 13 patients operated on
1971 insertion of the major muscle causing internal regained 90 degrees of passive
USA rotation and adduction of the shoulder. Gradual abduction and 20 degrees of
No Score increase in range of motion was obtained by external rotation within 2 months
completing an intensive post-operative exercise following surgery. However, 6
therapy program involving passive range of months later, all of the patients
motion, positioning of the shoulder, abduction were complaining of pain and
while supine, through the use of reciprocal pulley discomfort.

Conclusions Regarding Surgery as Treatment for Hemiplegic Shoulder Pain

There is limited (Level 2) evidence that surgically resecting the subscapularis and
pectoralis muscle tendons improves pain and range of motion in stroke patients
with a painful hemiplegic shoulder. Further research is needed to confirm these

11.6.8 Motor Blocks as Treatment For Muscle Imbalance

As discussed previously, subscapularis spasticity is characterized by shoulder range of

motion being most limited with pain being reproduced on external rotation. This
appears to correlate well with hemiplegic shoulder pain that is now thought to be a

consequence of spastic muscle imbalance about the shoulder in many cases.
Pectoralis muscle spasticity, characterized by limitation of range and pain on shoulder
abduction, is seen to a lesser extent, causing a similar muscle imbalance. Motor blocks
for spastic muscle imbalance offers the ability to redress that imbalance and relieve
hemiplegic shoulder pain.

Individual Studies

Table 11.18 Motor Block for Muscle Imbalance in Hemiplegic Shoulder

Author, Year Methods Outcomes

PEDro Score
Hecht 1992 A retrospective study of 13 patients with Immediate and significant improvement in
No Score spastic hemiplegia, limited ROM and ROM observed in abduction, flexion and
painful shoulder and with duration of external rotation. Relief of pain was noted
hemiplegia for 2 to 13 months. Patients with previously painful movements.
received percutaneous phenol nerve
blocks to subscapularis muscle
Hecht 1995 Prospective study of 20 patients 85% benefited from subscapularis block,
No Score receiving botulinum toxin muscle blocks and 55% benefited from pectotalis major
to the subscapular and pectoralis major block and 45% showed improved active
musculature. ROM.
Bhakta et al. 17 patients received a single course of Shoulder pain improved in 6 of 9 patients
1996 intramuscular botulinum toxin to biceps with shoulder pain.
UK brachii, flexor digitorum profundus, flexor
No Score digitorum superficialis and flexor carpi


Three small cohort studies examining deinnervating specific muscles, in particular the
subscapularis and pectoralis major muscles, improved ROM and pain. This is a
promising line of research that nevertheless requires a RCT to demonstrate its efficacy
as a viable treatment before definitive conclusions can be drawn.

Conclusions Regarding Motor Block for Muscle Imbalance

There is limited (Level 2) evidence that motor blocks of the subscapular and
pectoralis muscles can be used to treat muscle imbalance, pain and decreased
range of motion of the hemiplegic shoulder, although this new treatment requires
further research.

A potentially new treatment of the painful hemiplegic shoulder that requires

further research involves deinnervation of the subscapularis and pectoralis major

11.6.9 Summary of the Management of Hemiplegic Shoulder

Despite the high prevalence of patients suffering from painful hemiplegic shoulders, the
evidence for effective treatment is underwhelming. There is strong (Level 1a) evidence
supporting functional electrical stimulation and moderate (Level 1b) evidence supporting
an active therapy-oriented approach. There is also moderate (Level 1b) evidence that
overaggressive therapies using pullies substantially increases pain when compared to
gentler range of motion therapy approaches. There is insufficient evidence that
positioning of the shoulder, shoulder strapping, local corticosteroid injections and adding
local ultrasound therapy do not prevent subluxation, decrease pain or increase

There have been no RCTs conducted on the use of slings, motor blocks for spastic
muscle imbalance or providing NSAID medications prior to therapy, although there is
limited (Level 2) evidence of a benefit for all three of these treatment approaches.
There is consensus (Level 3) opinion that prevention and avoidance of overaggressive
therapy is important. Those individuals caring for the stroke patient, particularly early
on, should be aware of the potential for shoulder injury. The shoulder should be
carefully positioned and supported against gravity while sitting or standing. Range of
motion exercises should not carry the shoulder beyond 90 degrees of flexion and
abduction unless there is upward rotation of the scapula and external rotation of the
humeral head (Gresham et al. 1995).

Table 11.19 Summary of RCTs for Management of

Hemiplegic Shoulder
Author, Year PEDro Score n Outcomes
Positioning of the Shoulder
Dean et al. 2000 5 23 -
Ada et al. 2005 6 36 +/-
Slings and Other Aids
No RCTS - - -
Hanger et al. 2000 7 98 -
Griffin & Bernhardt 2006 7 32 +
Active Therapies
Inaba et al. 1972 7 33 -
Patridge et al. 1990 5 65 +
Kumar et al. 1990 5 28 -
(for aggressive
Lynch et al. 2005 6 35 -
Gustafson & McKenna 2006 6 34 -
Snels et al. 2000 8 35 -
Functional Electrical Stimulation
Baker et al. 1986 4 63 +
Faghri et al. 1994 4 26 +
Faghri & Rogers 1997 4 26 +
Linn et al. 1999 6 40 +/-

Table 11.19 Summary of RCTs for Management of
Hemiplegic Shoulder
Kobayshi et al. 1999 5 17 +
Chantraine et al. 1999 4 115 +
Wang et al. 2000 5 32 +
Yu et al. 2001 6 10 +
Yu et al. 2004 & Chae et al. 2005 7 61 +
Surgery for Muscle Imbalance
No RCTs - - o
Motor Block for Muscle Imbalance
No RCTs - - o

11.7 Shoulder Hand Syndrome (SHS)

11.7.1 Stages and Symptoms

Shoulder hand syndrome, also known as reflex sympathetic dystrophy (RSD) and
complex regional pain syndrome (type 1) is characterized by numerous peripheral and
central nervous system changes. Peripheral changes include vasomotor tone with
associated hand pain and swelling, exquisite tenderness or hyperaesthesia, protective
immobility, trophic skin changes and vasomotor instability of the involved upper
extremity. Central changes include a disruption of sensory cortical processing,
disinhibition of the motor cortex and disrupted body schema (Moseley et al. 2004). Iwata
et al. (2002) empirically describe 3 stages of RSD (Table 11.20).

Table 11.20 Stages and Characteristics of SHS

Stage Characteristics
I Persistent pain, described as burning, or aching and aggravated by movement
The extremity is edematous, warm and hyperesthetic
Lasting 3-6 months
2 Early dystrophic changes in the limb present
Atrophy of the muscle and skin
Vasospam with hyperhydrosis
3 Soft-tissue dystrophy
Contractures which produce “frozen shoulder”
Pain and vasomotor changes are infrequent

Shoulder hand syndrome generally presents initially with pain in the shoulder followed
by a painful, edematous hand and wrist. There is frequently decreased range of motion
at the shoulder and hand while the elbow joint is spared (Davis et al. 1977). Passive
flexion of the wrist, MCP and PIP joints is painful and limited due to edema over the
dorsum of the fingers. As time progresses, the extensor tendons become elevated and
the collateral ligaments shorten. If untreated it has long been thought that shoulder
hand syndrome eventually progresses to a dry, cold, bluish and atrophied hand.
However, experience would suggest that in most cases the pain and often the edema
subsides spontaneously after a few weeks.

Shoulder hand syndrome is often regarded as a form of sympathetically mediated pain
involving the hemiplegic upper extremity. The relationship between the sympathetic
nervous system and pain remains hypothetical and has yet to be proven. Shoulder hand
syndrome develops in about one in four hemiplegics. It is associated with involvement
of the premotor region and spasticity in the involved upper extremity. Diagnosis can be
made clinically, with metacarpal phalangeal joint tenderness to compression the most
consistent sign. While recovery is largely spontaneous, conditions which persists for
greater than 6 months are often difficult to treat.

11.7.2 Pathophysiology

Shoulder hand syndrome has been associated with lesions of the pre-motor area of the
brain. The etiology of shoulder hand syndrome is unknown; the sympathetic nervous
system has often been implicated largely because of the associated vasomotor
changes. Theoretical peripheral and central etiologies have been proposed. Peripheral
etiological theories postulate a role for trauma to the peripheral nerves. One of these
theories postulates ephaptic conduction between efferent sympathetic nerves and
afferent somatic nerves with the latter depolarization being perceived as pain.
Numerous central etiological theories have also been proposed. For instance, it has
been postulated that there is a disruption of autonomic nervous control from higher CNS
centres, which directly affects the internuncial pool of the spinal cord leading to
decreased inhibition of the sympathetic neurons of the lateral horn. Pain, either from
contractures or shoulder subluxation, may stimulate the internuncial pool of the spinal
cord resulting in an abnormal sympathetic response. A link between the abnormal
sympathetic nervous system and pain has also been postulated but never proven.

Geurts et al. (2000) systematically reviewed the etiology and treatments of post stroke
hand oedema and shoulder-hand syndrome. The authors identified 5 etiological studies
and 6 therapeutic studies. The authors evaluated the studies based on 11
methodological criteria and by standardized effect sizes. Based on their systematic
review of the literature, the authors concluded that the shoulder was involved in only
half the cases with all of the cases characterized by painful swelling of the wrist and
hand, thereby suggesting a “wrist-hand syndrome” in half the cases. Furthermore, they
noted that the hand edema was not a lymphoedema and that shoulder hand syndrome
usually coincided with increased arterial blood flow.

Iwata et al. (2002) suggested that SHS might be due to paresis following stroke,
mediated by disruption of homeostasis and the balance between intracellular and
extracellular fluid. Three possible mechanisms include: i) an increase in capillary blood
pressure, caused by a decrease in peripheral venous return and lymph flow; ii) a drop in
the colloidal osmotic pressure in the early stages of stroke due to an acute phase
response; iii) enhanced permeability of capillary walls which may result from synovial
inflammation, brought about by rough management of the affected arm and hand.

11.7.2 Incidence

Table 11.21 Incidence of Shoulder-Hand Syndrome

Author, Year Methods Outcome

PEDro Score
Davis et al. 540 rehabilitation inpatients with 68 (12.5%) patients were diagnosed with
1977 hemiplegia were evaluated with respect to RSD. Care was used to distinguish these
USA side of hemiplegia, dates of onset of patients from those with other shoulder
No Score hemiplegia and of pain, age, sex, pathologic conditions and pain syndromes.
handedness, sensory losses, associated
medical diseases and treatment response.
All patients’ charts were reviewed by a
team composed of authors
Eto et al. 1980 7 autopsy cases of hemiplegia shoulder- One of the cases showed an isolated brain
Japan hand syndrome from patients that had been lesion in the premotor area due to a
No Score admitted to the University of Tokyo Hospital metastasis from malignant melanoma. Four
or an affiliated hospital. Observed over a 6 other cases with cerebral infarction and one
year period, autopsy cases of hemiplegia with glioblastoma of the cortex in the area
shoulder-hand syndrome. Their clinical and supplied by the middle cerebral artery. The
post-mortem records were available in seventh case showed a hemorrhagic
determining the cerebral localization of the cerebral lesion in the lentiform nucleus.
syndrome and of unilateral longstanding The most common overlap area in the 6 of
autonomic dysfunction following the 7 cases was located in the premotor
hemiplegia. region including the anterior part of the
motor region. The shoulder-hand syndrome
following hemiplegia always develops on
the side contralateral to the brain lesion.
Tepperman et 85 consecutive patients admitted to a 21 patients (25%) exhibited radionuclide
al. 1984 rehabilitation unit with post-stroke evidence of RSD based on delayed scan
Canada hemiplegia. All had suffered a hemispheric criteria of increased uptake in the
No Score stroke of thrombotic, embolic or hemiplegic wrist, metacarpal-phalangeal
hypertensive hemorrhagic etiology as (MCP) and interphalangeal (IP) joints. 8 of
determined by CT brain scan. Patients the scan-positive RSD patients presented a
were assessed prospectively for low soft tissue blood flow pattern identical to
radionuclide and clinical features of reflex the non-RSD hemiplegic patients while the
sympathetic dystrophy (RSD). Each patient remaining 13 showed a high flow pattern.
underwent a three-phase bone scan within MCP tenderness to compression proved to
72 hours of admission. Delayed images of be the most valuable clinical sign of RSD,
both wrists and hands as well as total body with a predictive value, sensitivity, and
were obtained 2 to 4 hours after injection. specificity rates of 100%, 85.7% and 100%
Van 219 hemiplegic patients of whom 44 had The RSD syndrome was present in only
Ouwenaller et flaccid paralysis and 175 evolved rapidly 23% of all cases but was seen more often in
al. 1986 toward a spastic state. Stroke was the spastic patients, 27% vs. 7%.
Switzerland most frequent cause (79%). Spasticity was
No score diagnosed on the basis of an increase of
the myotatic reflex. The presence of
subluxation was sought. Criteria for
diagnosing the RSD syndrome were
involvement of both the hand and shoulder
and presence of the usual characteristic
symptoms. Each shoulder was tested for
an isolated tendon lesion and for several

Author, Year Methods Outcome
PEDro Score
contributing factors.
Kondo et al. 152 stroke patients admitted to a The incidences of RSD were 15/81 (18.5%)
2001 rehabilitation unit and followed for approx. for patients receiving the protocol and 23/71
Japan 200 days were monitored for the (32.4%) among patients who did not.
No Score development of RSD, assessed clinically by
a physician. Half of the patients were
treated with a protocol to prevent RSD,
consisting of passive ROM exercises,
performed by therapists and restrictions on
passive movement by patients. The
remaining patients received standard
inpatient rehabilitation.

While the incidence of RSD appears to range between 12-32%, Petchkrua et al. (2000),
suggested that the incidence of RSD is over-estimated and the results from previous
studies were obtained before patients routinely received early intensive inpatient
rehabilitation. At admission to hospital and once a week until discharge, patients
admitted to an acute rehabilitation facility were evaluated for shoulder pain, decreased
passive range of motion of the shoulder, wrist/hand pain, edema, and skin changes. If
three of these five criteria were positive, patients underwent a triple-phase bone scan
(TPBS). Bone scan findings consistent with CRPS type 1 were taken as confirming the
diagnosis. Of 64 subjects, 13 underwent bone scans, with only one (1.56%) positive
result. The authors noted it was possible that patients were discharged before they
developed symptoms of RSD. Patients from a more recent study (Kondo et al. 2001)
who received standard multidisciplinary rehabilitation had a much higher incidence of
RSD (34%).

Conclusions Regarding the Incidence of RSD Post Stroke

The incidence of RSD post stroke ranges form 12-34% and may be influenced by
the timing as well as the type of assessment.

11.7.3 Diagnostic Tests

Several approaches to diagnose RSD have been used. Routine radiographs of the
involved upper extremity may demonstrate a patchy, periarticular demineralization
(Sudek's atrophy) as early as 3-6 months after the onset of clinical signs. The most
sensitive diagnostic test is the technetium diphosphonate bone scan which
demonstrates increased periarticular uptake (mostly at the shoulder and wrist) in the
affected upper extremity. Bone scan abnormalities appear earlier than the x-ray
changes. Tepperman et al. (1984) found 25% of hemiplegic patients demonstrated
evidence of reflex sympathetic dystrophy in the involved upper extremity although only
two-thirds went on to develop the clinical syndrome. Temporary resolution of symptoms
with sympathetic blockade is considered diagnostic despite potential difficulties with the
technique in terms of diagnostic validity. Thermography, in controlled studies, has failed

to consistently diagnose reflex sympathetic dystrophy and is not considered a valid test.
However, Kozin et al. (1981) suggested that that clinical measurements such as grip
strength, tenderness and ring size were more accurate diagnostic indicator of RSD.
Iwata et al. (2002) have suggested that a ratio of the circumference of the middle finger
(affected:unaffected) greater than 1.06 at four weeks post stroke was predictive of RSD

Conclusions Regarding Shoulder-Hand Syndrome

Shoulder hand syndrome is a painful clinical entity, which is not understood from
a pathophysiological basis. The diagnosis is made clinically. Most cases appear
to improve with time.

11.7.4 Treatment of Shoulder-Hand Syndrome

Prevention of shoulder problems and aggressive early treatment are recommended to

prevent the development of a non-functional painful upper extremity. The various
treatment options are outlined in Table 11.22. Therapy consists of vigorous
physiotherapy with a focus on range of motion exercises. A one to two week course of
high dose corticosteroids and/or sympathetic blocks either in the form of stellate
ganglion blocks or guanethedine local venous blocks may be tried in persistent
disabling cases. A surgical sympathectomy may be considered if stellate ganglion
sympathetic blocks are consistently effective but symptoms recur. However, there is no
evidence that surgical sympathectomy alter outcomes. There is no definitive
therapeutic intervention for reflex sympathetic dystrophy, as reflected by the large
number of suggested treatments. Shoulder hand syndrome, which presents for greater
than 6 months without appropriate treatment, has a poor prognosis (Lieberman 1986).

Table 11.22 Potential Treatments for Shoulder Hand Syndrome

Prevention Splints
• Extremely early ROM exercises • Resting splint of hand and wrist
• Avoid shoulder subluxation (controversial)

Exercise Medication
• Prevention and treatment of upper • Analgesics
extremity contractures • NSAIDs
• Active exercise if possible • High dose oral corticosteroids (10 day
• Frequent passive ROM course and then taper)

Modalities Injections
• Interferential deep heat • Stellate ganglion sympathetic block
• Heat/cold modalities especially contrast • Guanethedine bier block
• Hand desensitization program Surgical
• Transcutaneous electrical nerve stimulation • Sympathectomy

11.7.5 Pharmacological Treatment of Shoulder-Hand Syndrome

Three studies evaluated drug therapies for the treatment of SHS associated with stroke.
The results are presented in table 11.23.

Table 11.23 Corticosteroid Treatment of Shoulder-Hand Syndrome

Author, Year Methods Outcome

PEDro Score
Davis et al. A retrospective study of 68 patients All patients became pain- free, when
1977 suffering from hemiparesis and SHS subjected to passive stretching of the
New York resulting from stroke. All patients received involved joints, within 3 weeks. Six patients
No Score 16 mg of an oral steroid (triamcinolone experienced a relapse of their pain, which
diacetate) daily for 14 to 21 days. Patients resolved during a second course of
self-described their pain. treatment.
Braus et al. 36 hemiplegic patients secondary to a No significant improvement was noted in
1994 stroke of the middle cerebral artery and shoulder-hand syndrome in the placebo
Germany exhibited definite shoulder-hand syndrome group after 4 weeks at which time all patients
5 (RCT) were studied. Patients were randomized switched over the corticosteroid protocol.
to orally receive either 8 mg Patients receiving the corticosteroid
methylprednisolone or a placebo over 4 treatment demonstrated significant
weeks. All patients received daily physical improvement in shoulder-hand syndrome
therapy. For patients in the placebo group, that was maintained at 6 months. 31 of the
if no improvement was noted in shoulder- 36 patients became almost symptom free
hand syndrome then they were given 4 within 10 days of treatment with low dose
weeks of corticosteroid treatment as per oral corticosteroids.
the experimental group.
Hamamci et A controlled trial of 41 hemiplegic patients By the end of the 4th week patients receiving
al.1996 with grades 1-2 of SHS. All patients calcitonin had significantly lower median pain
Turkey received conventional physical therapy as scores compared to the controls (1 vs. 5).
No Score part of inpatient rehabilitation. 25 patients Patients receiving calcitonin treatments also
received 1 x 100 IU salmon calcitonin daily reported less tenderness and improved
for 4 weeks and 16 patients received a range of motion. There was no difference in
saline injection as a placebo. Pain was hand edema between groups.
measured on a 7-point visual analogue
scale at the beginning and end of


The study by Braus et al. (1994) was the only RCT examining a treatment for shoulder-
hand syndrome. Oral corticosteroids improved SHS for at least 4 weeks. Despite a
limited number of trials, a review by Geurts et al. (2000) concluded that oral
corticosteroids were the most effective treatment for SHS. While a single controlled trial
found that calcitonin treatment effectively treated pain associated with SHS, it is not
widely used clinically.

Conclusions Regarding Oral Corticosteroids in SHS

There is moderate (Level 1b) evidence that oral corticosteroids improves

shoulder-hand syndrome for at least the first 4 weeks.

There is limited (Level 2) evidence that calcitonin improves pain associated with
SHS following stroke.

Oral corticosteroids appear to improve shoulder-hand syndrome for at least the

first 4 weeks.

11.7.6 Graded Motor Imagery as a Treatment for CRPS1

Motor Imagery has been suggested as an alternative therapy to conventional medical

management of SHS and involves activation of cortical networks without initially
involving movement of the affected limb. A version of this strategy (mirror therapy) has
been used successfully for patients suffering from phantom pain. There may be an
association between CRPS and a neglect-like condition, such that patients may need to
focus their attention to move the affected limb (Moseley et al. 2004). A single RCT
evaluated the effectiveness of this treatment following stroke.

Table 11.24 Studies Evaluating Motor Imagery Post Stroke

Author, Year Methods Outcome

PEDro Score
Moseley et al. 13 chronic SHS patients were randomly There was a significant improvement in
2004 allocated to a motor imagery program Neuropathic pain scale scores associated
Australia (MIP) or to ongoing management (usually with MIP treatment which persisted at 6
6 (RCT) pharmacological). The MIP consisted of weeks. The NNT to experience a > 50%
two weeks each of a hand laterality reduction in pain was 3.
recognition task, imagined hand
movements and mirror therapy, each two
weeks in duration. After 12 weeks, the
control group was crossed-over to MIP.

In the first phase of the treatment hand laterality recognition, avoidance of activation of
the primary motor cortex was achieved by only initiating activation in the pre-motor
cortices. In the second stage, patients were asked to imagine their own hand placed in
the same position as a picture selected from 28 pictures chosen at random. In the final
stage, pictures of the unaffected hand were placed into a cardboard mirror box. Patients
were asked to adopt the posture in the picture (n=20) times with both hands, but to
discontinue if they experienced pain.

Moseley et al. (2004) reported that treatment with MIP was more effective than ongoing
medical management of CRPS1. Patients experienced significant reductions in pain and
swelling associated with treatment, which persisted for at least 6 weeks. The authors
also noted that 6 weeks after completing the MIP program, approximately 50% of
patients no longer fulfilled the diagnostic criteria for CRPS1.

Conclusions Regarding Graded Motor Imagery

There is moderate (Level 1b) evidence that a modified imagery program can
reduce pain associated with shoulder-hand syndrome.

A motor imagery program appears to improve shoulder-hand syndrome.

11.8 Summary
1. Shoulder subluxation occurs early following a stroke.

2. Hemiplegic shoulder pain is associated with shoulder subluxation and

spasticity, but not with scapular rotation.

3. There appears to be an important role for the subscapularis muscle and to a

lesser extent pectoralis major musculature, which develop greater tonic
activity on the hemiparetic side with subsequent muscle imbalance about the

4. The development of painful hemiplegic shoulder is associated with severe

strokes and poorer functional outcome.

5. There is moderate (Level 1b) evidence that prolonged positioning does not
negatively influence shoulder range of motion or pain.

6. There is limited (Level 2) evidence that shoulder slings prevent subluxation

associated with hemiplegic shoulder pain, although there is also limited (Level
2) evidence that one device or method is better than another.

7. There is conflicting (Level 4) evidence that strapping the hemiplegic shoulder

reduces the development of pain. There is moderate (Level 1b) evidence that
strapping the hemiplegic shoulder does not improve range of motion or upper
limb function.

8. There is moderate (Level 1b) evidence that the use of overhead pullies results
in surprisingly high levels of hemiplegic shoulder pain and should be avoided.

9. There is moderate (Level 1b) evidence that a gentle range of motion program
by a therapist results in less hemiplegic shoulder pain.

10. There is moderate (Level 1b) evidence that corticosteroid injections do not
improve shoulder pain or range of motion in hemiplegic patients. There is
limited (Level 2) evidence that oral non-steroidal anti-inflammatories can
reduce pain during therapy sessions. There is limited (Level 2) evidence that
botulinum toxin can reduce pain in the hemiplegic shoulder.

11. There is strong (Level 1a) evidence that functional electrical stimulation
improves muscle function, pain, subluxation and range of motion of the
hemiplegic shoulder.

12. There is limited (Level 2) evidence that surgically resecting subscapularis and
pectoralis tendons improves outcomes in stroke patients with painful
hemiplegic shoulder.

13. There is limited (Level 2) evidence that motor blocks of the suprascapular and
pectoralis muscles treat muscle imbalance, pain and decreased range of
motion of the hemiplegic shoulder.

14. Shoulder hand syndrome is a poorly understood clinical entity. Most cases
improve with time.

15. There is moderate (Level 1b) evidence that oral corticosteroids improves
shoulder hand syndrome for at least the first 4 weeks.

16. There is moderate (Level 1b) evidence that a modified imagery program can
reduce pain associated with shoulder-hand syndrome. There is limited (Level
2) evidence that calcitonin improves pain associated with SHS following

Ada L, Foongchomcheay A. Efficacy of electrical stimulation in preventing or reducing subluxation of the
shoulder after stroke: a meta-analysis. Aust J Physiother 2002; 48(4):257-267.

Ada L, Goddard E, McCully J, Stavrinos T, Bampton J. Thirty minutes of positioning reduces the
development of shoulder external rotation contracture after stroke: A randomized controlled trial. Arch
Phys Med Rehabil 2005;86:230-234.

Ada L, Foongchomcheay A, Canning C. Supportive devices for preventing and treating subluxation of the
shoulder after stroke. Cochrane Database Syst Rev 2005;CD003863.

Ancliffe J. Strapping the shoulder in patients following a cerebrovascular accident (CVA): A pilot study.
Australian Journal of Physiotherapy 1992;38:37-39.

Aras MD, Gokkaya NK, Comert D, Kaya A, Cakci A. Shoulder pain in hemiplegia: results from a national
rehabilitation hospital in Turkey. Am J Phys Med Rehabil 2004;83:713-719

Arsenault AB, Silodean M, Duil E, Riley E. Clinical significance of the V-shaped space in the subluxed
shoulder of hemiplegics. Stroke 1991;22:867-871.

Bhakta BB, Cozens JA, Bamford JM, Chamberalin M. Use of botulinum toxin in stroke patients with
severe upper limb spasticity. J Neurol Neurosurg Psychiatr 1996; 61:30-35.

Baker LL, Parker K. Neuromuscular electrical stimulation of the muscles surrounding the shoulder. Phys
Ther 1986;66:1930-1937.

Basmajian JV, Bazant FJ. Factors preventing downward dislocation of adducted shoulder joint:
electromyographic and morphological study. J Bone Joint Surg (Am) 1959;41:1182-1186.

Bender L, Mckenna K. Hemiplegic shoulder pain: Defining the problem and its management. Disability
and Rehabilitation 2001;23(16):698-705.

Bloch R, Bayer N. Prognosis in stroke. Cl Orthop Related Res 1978;131:10-14.

Bohannon RW. Relationship between shoulder pain and selected variables in patients with hemiplegia.
Clinical Rehabilitation 1988; 2: 111-117.

Bohannon RW, Andrews AW. Shoulder subluxation and pain in stroke patients. American J Occup
Therapy 1990; 44(6):507-509.

Bohannon RW, Larkin PA, Smith MB, Horton MG. Shoulder pain in hemiplegia: statistical relationship with
five variables. Arch Phys Med Rehabil 1986;67:514-516.

Boyd EA. Clinical measures of shoulder subluxation: their reliability. Can J Public Health 1992;83 Suppl

Braun RM, West F, Mooney V, Nickel RL, Roper B, Caldwell C. Surgical treatment of the painful shoulder
contracture in the stroke patient. J Bone Joint Surg (Am) 1971;53:1307-1312.

Brooke MM, de Lateur BJ, Diana-Rigby GC, Questad KA. Shoulder subluxation in hemiplegia: effects of
three different supports. Arch Phys Med Rehabil 1991;72:582-586.

Cailliet R. The shoulder in hemiplegia. Philadelphia, PA, F.A. Davis Co., 1980.

Caldwell CB, Wilson DJ, Brown RM. Evaluation and treatment of the upper extremity in the hemiplegic
stroke patient. Clin Orthop 1969;63:69-93.

Carr EK, Kenney FD. Positioning of the stroke patient: a review of the literature. Int J Nurs Stud
Chaco J, Wolf E. Subluxation of the glenohumeral joint in hemiplegia. Am J Phys Med 1971; 50:139-

Chae J, Yu DT, Walker ME, et al. Intramuscular Electrical Stimulation for Hemiplegic Shoulder Pain: A 12-
Month Follow-Up of a Multiple-Center, Randomized Clinical Trial. Am J Phys Med Rehabil 2005;84:832-

Chantraine A, Baribeault A, Uebelhart D, Gremion G. Shoulder pain and dysfunction in hemiplegia:

effects of functional electrical stimulation. Arch Phys Med Rehabil 1999;80:328-331.

Chu Ds, Petrillo C, Davis SW, Eichberg R. Shoulder-hand syndrome: Importance of early diagnosis and
treatment. J Am Ger Soc 1981; 29:58-60.

Codman EA. The Shoulder. Boston, Thomas Todd Co, 1934.

Cole TM, Tobis JS. Measurement of musculoskeletal function. In: Kottke FJ, Lehmann JF (eds).
Krusen's Handbook of Physical Medicine and Rehabilitation. Philadelphia, Saunders, p.25, 1990.

Crossens-Sills J, Schenkman M. Analysis of shoulder pain, range of motion, and subluxation in patients
with hemiplegia. Phys Ther 1985; 65:731.

Culham EG, Noce RR, Bagg SD. Shoulder complex position and glenohumeral subluxation in hemiplegia.
Arch Phys Med Rehabil 1995;76:857-64.

Davis SW, Pestrillo CR, Eischberg RD, Chu DS. Shoulder-hand syndrome in a hemiplegic population: a
5-year retrospective study. Arch Phys Med Rehab 1977;58:353-355.

Dean CM, Mackey FH, Katrak P. Examination of shoulder positioning after stroke: A randomised
controlled pilot trial. Australian Journal of Physiotherapy 2000;46:35-40.

Dekker JH, Wagenaar RC, Lankhorst GJ, de Jong BA. The painful hemiplegic shoulder: effects of intra-
articular triamcinolone acetonide. Am J Phys Med Rehabil 1997;76:43-48.

Eto F, Yoshikawa M, Ueda S, Hirai S. Post-hemiplegic shoulder-hand syndrome with special reference to
related cerebral localization. J Am Ger Soc 1980;28:13-17.

Faghri PD, Rodgers MM, Glaser RM, Bors JG, Ho C, Akuthota P. The effects of functional electrical
stimulation on shoulder subluxation, arm function recovery, and shoulder pain in hemiplegic stroke
patients. Arch Phys Med Rehabil 1994;75:73-79.

Faghri PD. The effects of neuromuscular stimulation-induced muscle contraction versus elevation on
hand edema in CVA patients. J Hand Ther 1997;10:29-34.

Fitzgerald-Finch OP, Gibson II: Subluxation of shoulder in hemiplegia. Age Ageing 1975;4:16-18.

Fugl-Meyer AR, Jaasko L, Leyman I, Olsson S, Steglind S. Post-stroke hemiplegic patient: I. Method for
evaluation of physical performance. Scand J Rehab Med 1975;7:13-31.

Gamble GE, Barberan E, Laasch HU, Bowsher D, Tyrrell PJ, Jones AK. Poststroke shoulder pain: a
prospective study of the association and risk factors in 152 patients from a consecutive cohort of 205
patients presenting with stroke. Eur J Pain 2002;6:467-474.

Geurts A, Visschers B, van Limbeek J, Ribbers G. Systematic review of aetiology and treatment of post-
stroke hand oedema and shoulder-hand syndrome. Scan J Rehab Med 2000;32:4-10.

Griffin A, Bernhardt J. Strapping the hemiplegic shoulder prevents development of pain during
rehabilitation: a randomized controlled trial. Clin Rehabil 2006; 20(4):287-295.

Griffin A, Bernhardt J. Strapping the hemiplegic shoulder prevents development of pain during
rehabilitation: a randomized controlled trial. Clin Rehabil 2006; 20(4):287-295.

Gustafsson L, McKenna K. A programme of static positional stretches does not reduce hemiplegic
shoulder pain or maintain shoulder range of motion--a randomized controlled trial. Clin Rehabil 2006;

Hakuno A, Sashika H, Ohkawa T, Itoh R. Arthrographic findings in hemiplegic shoulders. Arch Phys
Med Rehab 1984; 65:706-711.

Hamamci N, Dursun E, Ural C, Cakci A. Calcitonin treatment in reflex sympathetic dystrophy: a

preliminary study. Br J Clin Pract 1996; 50(7):373-375.

Hanger HC, Whitewood P, Brown G, Ball MC, Harper J, Cox R, Sainsbury R. A randomized controlled
trial of strapping to prevent post-stroke shoulder pain. Clin Rehabil 2000;14:370-380.

Hecht JS. Subscapular nerve block in the painful hemplegic shoulder. Arch Phys Med Rehabil

Hecht JS. The role of spasticity in hemiplegic shoulder pain and what to do about it. 57th Annual
Assembly of American Academy of Physical Medicine and Rehabilitation, Orlando, Florida, November 17,
1995, pp. 248-255.

Hollinshead WH, Jenkins DB. Functional anatomy of the limbs and back. Philadelphia, Saunders p.105,

Hurd MM, Farrell KH, Waylonis GW. Shoulder sling for hemiplegia: friend or foe? Arch Phys Med Rehabil

Ikai T, Tei K, Yoshida K, Miyano S, Yonemoto K. Evaluation and treatment of shoulder subluxation in
hemiplegia: relationship between subluxation and pain. Am J Phys Med Rehabil 1998; 77(5):421-426.

Inaba MK, Piorkowski M. Ultrasound in treatment of painful shoulder in patients with hemiplegia. J Phys
Ther 1972;52:737-741.

Iwata M, Kondo I, Sato Y, Satoh K, Soma M, Bar-Or O. Prediction of reflex sympathetic dystrophy in
hemiplegia by evaluation of hand edema. Arch Phys Med Rehabil 2002; 83(10):1428-1431.

Janig W, Baron R. Complex regional pain syndrome is a disease of the central nervous system. Clin
Auton Res 2002;12:150-164.

Johnstone M. The stroke patient: Principles of rehabilitation. Churchill Livingstone, Edinburgh, 1982.

Joynt RL. The source of shoulder pain in hemiplegia. Arch Phys Med Rehabil 1992;73:409-413.

Kaplan M. Hemiplegic shoulder pain--early prevention and rehabilitation. West J Med. 1995

Kaplan PE, Meredith J, Taft G, Betts HB. Stroke and brachial plexus injury: A difficult problem. Arch
Phys Med Rehab 1977;38:415-18.

Kingery WS, Date ES, Bobobo CR. The absence of brachial plexus injury in stroke. AM J Phys Med
Rehab 1993; 72:127-175.

Kobayashi H, Onishi H, Ihashi K, Yagi R, Handa Y. Reduction in subluxation and improved muscle
function of the hemiplegic shoulder joint after therapeutic electrical stimulation. J Electromyogr Kinesiol

Kondo I, Hosokawa K, Soma M, Iwata M, Maltais D. Protocol to prevent shoulder-hand syndrome after
stroke. Arch Phys Med Rehabil 2001; 82(11):1619-1623.

Kozin F, McCarty D, Sims J. Genant H. Reflex sympathetic dystrophy: I. Clinical and histological
studies: evidence for bilaterality, response to corticosteroids and articular involvement. Am J Med

Kozin F, Genant HK, Bekerman C, McCarty DJ. Reflex sympathetic dystrophy syndrome: II.
Roentgenographic and scintigraphic evidence of bilaterality and of periarticular accentuation. Am J Med

Kozin F, Ryan LM, Carrera GF, Soin JS, Wortman RL. Reflex sympathetic dystrophy syndrome (RSDS):
III Scintigraphic studies, further evidence for the therapeutic efficacy of systemic corticosteroids, and
proposed diagnostic criteria. Am J Med 1981;70 23-30.

Kozin F. Two unique shoulder disorders: Adhesive capsulitis and reflex sympathetic dystrophy
syndrome. Postgrad Med 1983;73(5):207-216.

Kumar R, Metter EJ, Mehta AJ, Chew T. Shoulder pain in hemiplegia. The role of exercise. Am J Phys
Med Rehabil 1990; 69(4):205-208.

Lieberman JS. Hemiplegia: Rehabilitation of the upper extremity. In: Kaplan PE, Cerullo LJ (eds.):
Stroke Rehabilitation. Stoneman, MA, Butterworth Publishers, p.95-117, 1986.

Linn SL, Granat MH, Lees KR. Prevention of shoulder subluxation after stroke with electrical stimulation.
Stroke 1999;30:963-968.

Lo SF, Chen SY, Lin HC, Jim YF, Meng NH, Kao MJ. Arthrographic and clinical findings in patients with
hemiplegic shoulder pain. Arch Phys Med Rehabil 2003;84:1786-1791.

Lynch D, Ferraro M, Krol J, Trudell CM, Christos P, Volpe BT. Continuous passive motion improves
shoulder joint integrity following stroke. Clin Rehabil 2005;19:594-599.

Moodie NB, Brisbin J, Morgan AM. Subluxation of the glenohumeral joint in hemiplegia: Evaluation of
supportive devices. Physiotherapy Canada 1986;38:151-157.

Morin L, Bravo G. Strapping the hemiplegic shoulder: A radiographic evaluation of its efficacy to reduce
subluxation. Physiotherapy Canada 1997; 49:103-108.

Moseley GL. Graded motor imagery is effective for long-standing complex regional pain syndrome: a
randomised controlled trial. Pain 2004;108:192-198.

Moskowitz E. Complications in rehabilitation of hemiplegic patients. Med Clin North Am 1969;53 541-

Moskowitz E, Goodman CR, Smith E, Balthazar E, Mellins HZ. Hemiplegic shoulder. NY State J Med
1969; 69:548-550(b).

Najenson T, Yacubovich E, Pikelini S. Rotator cuff injury in hemiplegic patients. Scand J Rehab Med
1971; 3:131-137.

Nepomuceno CS, Miller JM III. Shoulder arthrography in hemiplegic patients. Arch Phys Med Rehab
1974; 55:49-51.

Paci M, Nannetti L, Rinaldi LA. Glenohumeral subluxation in hemiplegia: An overview. J Rehabil Res Dev
2005; 42(4):557-568.

Partridge CJ, Edwards SM, Mee R, Van Langenberghe HV. Hemiplegic shouldar pain: a study of two
methods of phsyiotherapy treatment. Clinical Rehabilitation 1990;4:43-49.

Patterson JR, Zabransky R, Grabois M, Ferro P. Evaluation of the effectivenss of sling orthoses for the
correction of glenhummeral subluxation in the hemiplegic. Archives Phys Med Rehabil 1984;65:635.

Perrigot M, Bussel B, Pierrot Deseilligny E, Held JP. L'epaule de l'hemiplegique. Ann Med Phys

Poduri KR. Shoulder pain in stroke patients and its effects on rehabilitation. J Stroke Cerebrovasc Dis

Post MW, Visser-Meily JM, Boomkamp-Koppen HG, Prevo AJ. Assessment of oedema in stroke patients:
comparison of visual inspection by therapists and volumetric assessment. Disabil Rehabil 2003;25:1265-

Poulin de Courval L, Barsauskas A, Berenbaum B, Dehaut F, Dussault R, Fontaine FS, Labrecque R,

Leclerc C, Giroux F. Painful shoulder in the hemiplegic and unilateral neglect. Arch Phys Med Rehabil.
1990 Aug;71(9):673-6.

Prevost R, Arsenault AB, Dutil E, Drouin G. Rotation of the scapular and shoulder subluxation in
hemiplegia. Arch Phys Med Rehab 1987;68:786-790.

Price CI, Pandyan AD. Electrical stimulation for preventing and treating post-stroke shoulder pain: a
systematic Cochrane review. Clin Rehabil 2001;15:5-19

Rizk TE, Christopher RP, Pinals RS, Salazar JE, Higgins C. Arthrographic studies in painful hemiplegic
shoulders. Arch Phys Med Rehab 1984;65:254-55.

Rommel O, Malin JP, Zenz M, Janig W. Quantitative sensory testing, neurophysiological and
psychological examination in patients with complex regional pain syndrome and hemisensory deficits.
Pain 2001;93:279-293.

Roper TA, Redford S, Tallis RC. Intermittent compression for the treatment of the oedematous hand in
hemiplegic stroke: a randomized controlled trial. Age Ageing 1999;28:9-13.

Roy CW, Sands M, Hill L, Harrison, Marshall S. The effect of shoulder pain on outcome of acute
hemiplegia. Clin Rehabil 1995;9:21-27.

Roy CW, Sands MR, Hill LD. Shoulder pain in acutely admitted hemiplegics. Clinical Rehabilitation

Renzenbrink GJ, IJzerman MJ. Percutaneous neuromuscular electrical stimulation (P-NMES) for treating
shoulder pain in chronic hemiplegia. Effects on shoulder pain and quality of life. Clin Rehabil

Savage R, Robertson L. Relationship between adult hemiplegic shoulder pain and depression.
Physiotherapy Can 1982;34:86-90.

Shahani BT, Kelly EB, Glasser S. Hemiplegic shoulder subluxation. Arch Phys Med Rehab 1981;62:519.

Shai G, Ring H, Costeff H, Solzi P. Glenohumeral malalignment in hemiplegic shoulder. Scan J Rehab
Med 1984; 16:133-136.

Snels IA, Beckerman H, Twisk JW, Dekker JH, Peter DK, Koppe PA, Lankhorst GJ, Bouter LM. Effect of
triamcinolone acetonide injections on hemiplegic shoulder pain : A randomized clinical trial. Stroke

Steinbrocker D. Shoulder-hand syndrome: present perspective. Arch Phys Med Rehab 1969;49:388-
Tahmoush AJ. Causalgia: redefinition as a clinical pain syndrome. Pain 1981;10:187.

Tepperman PS, Greyson ND, Hilbert L, et al. Reflex sympathetic dystrophy in hemiplegia. Arch Phys
Med Rehab 1984;65(8):442-447.

Tyson SF and Chissim C. The immediate effect of handling technique on range of movement in the
hemiplegic shoulder. Clinical Rehabilitation 2002;16:137-140.

Van Ouwenwaller C, Laplace PM, Chantraine A. Paniful shoulder in hemiplegia. Arch Phys Med Rehabil
1986; 67:23-26.

Walsh K. Management of shoulder pain in patients with stroke. Postgrad Med J 2001;77:645-649.

Wang RY, Chan RC, Tsai MW. Functional electrical stimulation on chronic and acute hemiplegic shoulder
subluxation. Am J Phys Med Rehabil 2000;79:385-390.

Wanklyn P, Forester A, Young J. Hemiplegics houlder pain (HSP): natural history and investigation of
associated features. Disability and Rehabilitation 1996;18:497-501.

Williams R, Taffs L, Minuk T. Evaluation of two support methods for the subluxated shoulder of
hemiplegic patients. Phys Ther 1988;68:1209-1214.

Yu DT, Chae J, Walker ME, Fang ZP. Percutaneous intramuscular neuromuscular electric stimulation for
the treatment of shoulder subluxation and pain in patients with chronic hemiplegia: a pilot study. Arch
Phys Med Rehabil 2001;82:20-25.

Yu DT, Chae J, Walker ME, Hart RL, Petroski GF. Comparing stimulation-induced pain during
percutaneous (intramuscular_ and transcutaneous neuromuscular electric stimulation for treatment
shoulder subluxation in hemiplegia. Arch Phys Med Rehabil 2001;82:756-60.

Zorowitz RD, Idank D, Ikai T, Hughes MB, Johnston MV. Shoulder subluxation after stroke: A
comparison of four supports. Arch Phys Med Rehabil 1995;76:763-771.

Zorowitz RD, Hughes MB, Idank D, Ikai T, Johnston V. Shoulder pain and subluxation after stroke:
Correlation or coincidence? American Journal of Occupational Therapy 1996;50:194-201.