Shoulder Pain

Evidence Reviewed as of before: 25-01-2013
Author(s)*: Tatiana Ogourtsova, MSc OT; Annabel McDermott, OT; Dr Nicol Korner-Bitensky, PhD OT
Table of contents

Introduction

Shoulder pain resulting from post-stroke hemiplegia can lead to significant disability. Shoulder pain can be caused by spasticity and by the position of the hemiplegic joint. Hemiplegic shoulder pain can limit patient’s recovery and functional independence post-stroke. Most rehabilitation strategies aim to prevent and/or improve pain through different therapeutic mediums (e.g. functional electrical stimulation, mirror therapy, etc.), as well to optimize hemiplegic shoulder positioning (e.g. use of positioning techniques, slings, etc.).

Clinician Information

Note: When reviewing the findings, it is important to note that they are always made according to randomized clinical trial (RCT) criteria – specifically as compared to a control group. To clarify, if a treatment is “effective” it implies that it is more effective than the control treatment to which it was compared. Non-randomized studies are no longer included when there is sufficient research to indicate strong evidence (level 1a) for an outcome.

The current module reviewed 24 studies (11 high quality RCTs, nine fair quality RCTs, one poor quality RCT and three non-randomized studies) that investigated interventions used in the treatment of shoulder pain following stroke. Interventions included Constraint-Induced Movement Therapy (CIMT), Therapeutic/Functional Electrical Stimulation (TES/FES), positioning and handling techniques, massage, mirror therapy, mental imagery, strapping, cryotherapy and aromatherapy acupressure and acupuncture.

Results indicated that FES is more effective than control interventions for managing shoulder pain and subluxation at different stages of stroke recovery, and can also benefit range of motion in patients with acute stroke. Mirror therapy was found to be more effective than comparison therapies in managing shoulder pain in patients with subacute and chronic stroke. Strapping was more effective than no strapping in patients with subacute stroke (but not acute stroke). An axilla handling technique was more effective than a distal hold technique for managing pain-free range of motion. Slow-stroke back massage benefited shoulder pain.

Other therapies such as positioning, CIMT and modified CIMT, mental imagery, cryotherapy and acupuncture were not more effective than comparison therapies in management of the hemiplegic shoulder following stroke.

Results Table

View results table

Outcomes

Acute stroke: Overhead pulleys

Shoulder pain
Not effective
2b

One quasi-experimental study (Kumar et al., 1990) investigated the effect of overhead pulleys on shoulder pain among patients with acute stroke. This quasi-experimental study assigned patients to receive 1 of 3 upper limb exercise programs: (i) overhead pulley exercises, (ii) passive range of motion (pROM) exercises, or (iii) active-assisted ROM exercises. There was a significant between-group difference in shoulder pain at rest and during passive range of motion (measurement tool not specified) at 4, 8, and 12 weeks, in favour of pROM exercises compared to overhead pulley exercises. No other between-group differences in shoulder pain were found.
Note: This study was insufficiently powered to provide strong conclusive results.

Conclusion: There is limited evidence (level 2b) from one quasi-experimental study that overhead pulley exercises are not more effective than comparison interventions (passive ROM exercises, active-assisted ROM exercises) for managing shoulder pain in patients with acute stroke. In fact, overhead pulley exercises were found to cause greater incidence of pain than passive range of motion exercises.
Note: Pulley exercises are NOT recommended for use with patients with flaccid muscle tone or limited movement in the upper limb or any concern for possible shoulder subluxation.

Shoulder subluxation
Not effective
2b

One quasi-experimental study (Kumar et al., 1990) investigated the effect of overhead pulleys on shoulder subluxation among patients with acute stroke. This quasi-experimental study assigned patients to receive 1 of 3 upper limb exercise programs: (i) overhead pulley exercises, (ii) passive range of motion (pROM) exercises, or (iii) active-assisted ROM exercises. There were no significant between-group differences in incidence of shoulder subluxation at 4, 8, and 12 weeks.
Note: This study was insufficiently powered to provide strong conclusive results.

Conclusion: There is limited (level 2b) evidence from one quasi-experimental study that overhead pulley exercises are not more effective than comparison interventions (passive or active-assisted ROM exercises) for the management of shoulder subluxation in patients with acute stroke.
Note: Pulley exercises are NOT recommended for use with patients with flaccid muscle tone or limited movement in the upper limb or any concern for possible shoulder subluxation.

Acute stroke: Functional Electrical Stimulation

Pain free Range of Motion
Not effective
2a

Two fair quality RCTs (Linn et al., 1999; Wang et al., 2002) investigated the effect of FES on pain-free shoulder ROM among patients with acute stroke.

The first fair quality RCT (Linn et al., 1999) randomized patients to receive FES and conventional rehabilitation or conventional rehabilitation alone. Stimulation was applied to the supraspinatus and posterior deltoid muscles for 2-4 hours/day for 4 weeks. No significant between-group differences in pain-free ROM on passive lateral shoulder rotation (measured by goniometer) were seen at post-treatment (4 weeks) or follow-up (3 months).

The second fair quality RCT (Wang et al., 2002) randomized patients with acute or chronic stroke to receive short FES (1 cycle of FES) or long FES (two cycles of FES interspersed with 6 weeks of conventional rehabilitation alone). Each cycle of FES comprised stimulation to the supraspinatus and deltoid muscles for 6 hours/day, 5 days/week for 6 weeks. Predictably, there were no significant between-group differences in pain-free ROM on passive external rotation at 6 weeks or 12 weeks among patients with acute stroke, at which points both groups had received the same intervention (6 weeks of FES followed by 6 weeks of no FES). There were also no significant between-group differences at 18 weeks, at which point only the second group had received a second cycle of FES (allowing comparison of FES vs. no FES).

Conclusion: There is limited evidence (level 2a) from two fair quality RCTs that FES is not more effective than conventional rehabilitation alone for improving pain-free range of motion in the affected shoulder of patients with acute stroke.

Range of Motion
Effective
2a

One fair quality RCT (Chantraine et al., 1999) investigated the effect of FES on shoulder ROM among patients with stroke. This fair quality RCT randomized patients with shoulder subluxation following acute stroke or head injury* to receive FES and conventional rehabilitation or conventional rehabilitation alone (based on the Bobath approach). Stimulation was applied to the affected limb for 130-140 minutes/session for 5 weeks. There was a significant between-group difference in ROM on shoulder antepulsion and abduction at 6, 12, and 24 months post-intervention, in favour of FES compared to conventional rehabilitation alone.
* The majority of participants (101/120) presented with stroke.

Conclusion: There is limited evidence (level 2a) from one fair quality RCT that FES and conventional rehabilitation is more effective than comparison therapies (conventional rehabilitation alone) for improving shoulder ROM among patients with acute stroke and shoulder subluxation.

Shoulder pain (patients with subluxation)
Effective
2a

One fair quality RCT (Chantraine et al., 1999) investigated the effect of FES on shoulder ROM among patients with stroke. This fair quality RCT randomized patients with shoulder subluxation following acute stroke or head injury* to receive FES and conventional rehabilitation or conventional rehabilitation alone (based on the Bobath approach). Stimulation was applied to the affected limb for 130-140 minutes/session for 5 weeks. There was a significant between-group difference in ROM on shoulder antepulsion and abduction at 6, 12, and 24 months post-intervention, in favour of FES compared to conventional rehabilitation alone.
* The majority of participants (101/120) presented with stroke.

Conclusion: There is limited evidence (level 2a) from one fair quality RCT that FES and conventional rehabilitation is more effective than comparison therapies (conventional rehabilitation alone) for improving shoulder ROM among patients with acute stroke and shoulder subluxation.

Shoulder pain (patients without subluxation)
Not effective
1b

One high quality RCT (Church et al., 2006) and one fair quality RCT (Linn et al., 1999) investigated the effect of FES on shoulder pain among patients with acute stroke.

The high quality RCT (Church et al., 2006) randomized patients to receive surface neuromuscular stimulation (sNMES) or sham stimulation. Stimulation was applied to the supraspinatus and posterior deltoid muscles of the affected shoulder for 3 hours/day for 4 weeks. No significant between-group differences in shoulder pain (5-point severity scale and 10-point numerical rating scale) were seen at post-treatment (4 weeks) or follow-up (3 months).

The second fair quality RCT (Linn et al., 1999) randomized patients to receive FES and conventional rehabilitation or conventional rehabilitation alone. Stimulation was applied to the supraspinatus and posterior deltoid muscles for 2-4 hours/day for 4 weeks. No significant between-group differences in shoulder pain (measured by a five-point pain scale) were seen at post-treatment (4 weeks) or follow-up (3 months).

Conclusion: There is moderate evidence (level 1b) from one high quality RCT and one fair quality RCT that FES is not more effective than comparison therapies (sham stimulation or conventional rehabilitation alone) in the management of shoulder pain among patients with acute stroke without shoulder subluxation.

Shoulder subluxation (patients with subluxation)
Effective
2a

One fair quality RCT (Chantraine et al., 1999) investigated the effect of FES on shoulder subluxation among patients with stroke. This fair quality RCT randomized patients with shoulder subluxation following acute stroke or head injury* to receive FES and conventional rehabilitation or conventional rehabilitation alone (based on the Bobath approach). Stimulation was applied to the affected limb for 130-140 minutes/session for 5 weeks. There was a significant between-group difference in shoulder subluxation (measured by X-ray) at 6, 12, and 24 months post-intervention, in favour of FES compared to conventional rehabilitation alone.
* The majority of participants (101/120) presented with stroke.

Conclusion: There is limited evidence (level 2a) from one fair quality RCT that FES and conventional rehabilitation is more effective than comparison therapies (conventional rehabilitation alone) in the management of shoulder subluxation among patients with acute stroke and shoulder subluxation.

Shoulder subluxation (without subluxation/prevention)
Not effective
2a

Two fair quality RCTs (Linn et al., 1999; Wang et al., 2002) investigated the effect of FES on prevention or management of shoulder subluxation among patients with acute stroke.

The first fair quality RCT (Linn et al., 1999) randomized patients to receive FES and conventional rehabilitation or conventional rehabilitation alone. Stimulation was applied to the supraspinatus and posterior deltoid muscles for 2-4 hours/day for 4 weeks. No significant between-group differences in shoulder subluxation (measured by radiological exam) were seen at post-treatment (4 weeks) or follow-up (3 months).

The second fair quality RCT (Wang et al., 2002) randomized patients with acute or chronic stroke to receive short FES (1 cycle of FES) or long FES (two cycles of FES interspersed with 6 weeks of conventional rehabilitation alone). Each cycle of FES comprised stimulation to the supraspinatus and deltoid muscles for 6 hours/day, 5 days/week for 6 weeks. Predictably, there were no significant between-group differences in shoulder subluxation at 6 weeks or 12 weeks among patients with acute stroke, at which points both groups had received the same intervention (6 weeks of FES followed by 6 weeks of no FES). There were also no significant between-group differences at 18 weeks, at which point only the second group had received a second cycle of FES (allowing comparison of FES vs. no FES).

Conclusion: There is limited evidence (level 2a) from two fair quality RCTs that FES is not more effective than comparison therapies (conventional rehabilitation) for preventing shoulder subluxation in patients with acute stroke.

Acute stroke: Pain-free Range of Motion

Positioning
Conflicting
4

Two high quality RCTs (Ada et al., 2005; Gustafsson & McKenna, 2006) investigated the effect of static arm positioning on shoulder ROM among patients with acute stroke.

The first high quality RCT (Ada et al., 2005) randomized patients to a static positioning program or no positioning program. Static positioning was performed in lying and sitting for two 30-minute sessions per day, 5 days per week, for 4 weeks. Both groups received standard upper limb care and upper extremity exercises for 4 weeks. There was a significant between-group difference in passive ROM into external rotation and external rotation contractures at post-treatment (4 weeks), in favour of the positioning group compared to the control group. There were no significant between-group differences in passive ROM on shoulder flexion or shoulder flexion contractures at post-treatment.

The second high quality RCT (Gustafsson & McKenna, 2006) randomized patients to receive a static positional stretching program or no stretching program. The stretching program comprised stretches, use of pillows in supine and an Otto Bock modular arm rest support in sitting for two 20-minute sessions per day for 4 weeks, whereas the control group used locally fabricated cushion supports in sitting and lying. No significant between-group difference in pain-free range of motion into external rotation was seen at post-treatment (4 weeks).

Conclusion: There is conflicting evidence (level 4) about the effect of static arm positioning on shoulder ROM among patients with acute stroke. While one high quality RCT that found a significant between-group difference in contractures and passive ROM on external rotation (but not flexion) following a positioning program, another high quality RCT that found no significant between-group difference in pain-free ROM into external rotation following a static positional stretching program.
Note: Both groups performed the positioning program in supine and sitting for a similar time frame (20-30mins/position/day, 5 days/week for 4 weeks). Differences in results between studies may relate to differences in positions used: the group that saw significant between-group differences in passive ROM and contractures in external rotation position the arm at 45 degrees shoulder abduction (vs. 90 degrees shoulder abduction) in supine and at 90 degrees shoulder flexion (vs. 90 degrees shoulder abduction) in sitting. Further, this study provided both groups with upper extremity exercises for 10 minutes/day (vs. 30 minutes/day).

Strapping
Not effective
1b

One high quality RCT (Hanger et al., 2000) investigated the effect of hemiplegic shoulder strapping on pain-free ROM among patients with acute stroke. This high quality RCT randomized patients to receive strapping and conventional rehabilitation or conventional rehabilitation alone. Strapping of the affected shoulder was performed for 6 weeks, until discharge or until patients were able to achieve 90° active shoulder abduction against gravity with elbow flexed. There were no significant between-group differences in pain-free shoulder lateral ROM at post-treatment (6 weeks) or follow-up (14 weeks).

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that shoulder strapping is not more effective than conventional rehabilitation for improving pain-free ROM among patients with acute stroke.

Acute stroke: Shoulder pain

Constraint-Induced Movement Therapy (CIMT)
Not effective
1b

One high quality RCT (Dromerick et al., 2009) examined the effect of CIMT on shoulder pain among patients with acute stroke. This high quality RCT randomized patients to receive ‘standard’ CIMT (shaping therapy for 2 hours/day and restraint 6 hours/day), ‘intensive’ CIMT (shaping therapy for 3 hours/day and restraint 90% of waking hours) or conventional occupational therapy comprising ADL training and bilateral exercises. There were no significant differences in shoulder pain (Wong-Baker Faces Scale) between any group at post-treatment (14 days) or follow-up (90 days).
Note: The terms ‘standard’ and ‘intensive’ CIMT were defined by the authors of this study.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that CIMT is not more effective than conventional rehabilitation in the management of shoulder pain among patients with acute stroke.
Note:
This high quality RCT also found no difference in pain between different intensities of modified CIMT.

Positioning
Not effective
1a

Two high quality RCTs (Ada et al., 2005; Gustafsson & McKenna, 2006) investigated the effect of static arm positioning on shoulder pain among patients with acute stroke.

The first high quality RCT (Ada et al., 2005) randomized patients to a static positioning program or no positioning program. Static positioning was performed in lying and sitting for two 30-minute sessions per day, 5 days per week, for 4 weeks. Both groups received standard upper limb care and upper extremity exercises for 4 weeks. There were no significant between-group differences in shoulder pain (visual analogue scale) at post-treatment (4 weeks).

The second high quality RCT (Gustafsson & McKenna, 2006) randomized patients to receive a static positional stretching program or no stretching program. The stretching program comprised stretches, use of pillows in supine and an Otto Bock modular arm rest support in sitting for two 20-minute sessions per day for 4 weeks, whereas the control group used locally fabricated cushion supports in sitting and lying. There were no significant between-group differences in shoulder pain on movement (Ritchie Articular Index) or pain at rest (VAS) at post-treatment (4 weeks).

Conclusion: There is strong evidence (level 1a) from two high quality RCTs that static positioning during supine lying and sitting is not more effective than control interventions (no positioning program, standard positioning) in the management of shoulder pain at movement or at rest in patients with acute stroke.

Strapping
Not effective
1b

One high quality RCT (Hanger et al., 2000) and one poor quality RCT (Ancliffe, 1992) investigated the effect of hemiplegic shoulder strapping on shoulder pain among patients with acute stroke.

The high quality RCT (Hanger et al., 2000) randomized patients to receive strapping and conventional rehabilitation or conventional rehabilitation alone. Strapping of the affected shoulder was performed for 6 weeks, until discharge or until patients were able to achieve 90° active shoulder abduction against gravity with elbow flexed. There were no significant between-group differences in shoulder pain (VAS) at post-treatment (6 weeks) or follow-up (14 weeks).

The poor quality RCT (Ancliffe, 1992) assigned patients to receive shoulder strapping or no strapping (duration of intervention not specified). There was a significant between-group difference in the mean number of days before the onset of pain (Ritchie Articular Index), in favour of the strapping group compared to the control group.
Note: This study set the statistical significance at p=0.01 to compensate for a small sample size.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that shoulder strapping is not more effective than conventional rehabilitation in the management of shoulder pain among patients with acute stroke.
Note: However, a poor quality RCT found that strapping the hemiplegic shoulder resulted in greater number of days before the onset of shoulder pain.

Subacute stroke: Positioning

Shoulder abduction Range of Motion
Conflicting
4

Two high quality RCTs (Dean et al., 2000; de Jong et al., 2006) investigated the effect of static arm positioning on shoulder abduction ROM among patients with subacute stroke.

The first high quality RCT (Dean et al., 2000) randomized patients to receive a prolonged positioning program or conventional multidisciplinary rehabilitation only. The positioning program was performed in supine and sitting positions for 60 minutes per day for 6 weeks. There were no significant between group differences in active shoulder abduction or passive external rotation ROM (measured by standard goniometer and a gravity goniometer) at post-treatment (6 weeks).
Note: Authors noted that this study was insufficiently powered to provide conclusive results.

The second high quality RCT (de Jong et al., 2006) randomized patients to receive an upper limb positioning program and conventional rehabilitation or conventional rehabilitation alone. The positioning program was performed in supine for 60-minutes per day, 5 days per week for 5 weeks. At post-treatment (5 weeks) there was a significant between-group difference in passive ROM on shoulder abduction (measured by fluid-filled goniometer), in favour of positioning compared to conventional rehabilitation alone. There were no significant between-group differences in pROM on shoulder external rotation and flexion at post-treatment.

Conclusion: There is conflicting evidence (level 4) between one high quality RCT that found static positioning is not more effective than conventional rehabilitation alone for improving shoulder abduction active ROM, and a second high quality RCT that found static positioning is more effective than conventional rehabilitation alone for improving shoulder abduction passive ROM.
Note: Both studies used a similar supine position (maximum shoulder abduction, external rotation, elbow extension). Interestingly, the group that saw no between-group differences in shoulder abduction active ROM used two positions (lying and sitting) with the arm in 90 degrees shoulder abduction, whereas the group that saw a significant between-group difference in shoulder abduction passive ROM used only the lying supine position and provided a slightly shorter positioning program (5 weeks vs. 6 weeks).

Shoulder external rotation passive Range of Motion
Not effective
1a

Two high quality RCTs (Dean et al., 2000; de Jong et al., 2006) investigated the effect of static arm positioning on shoulder external rotation passive ROM among patients with subacute stroke.

The first high quality RCT (Dean et al., 2000) randomized patients to receive a prolonged positioning program or conventional multidisciplinary rehabilitation only. The positioning program was performed in supine and sitting positions for 60 minutes per day for 6 weeks. There were no significant between group differences in active shoulder abduction or passive external rotation ROM (measured by standard goniometer and a gravity goniometer) at post-treatment (6 weeks).
Note: Authors noted that this study was insufficiently powered to provide conclusive results.

The second high quality RCT (de Jong et al., 2006) randomized patients to receive an upper limb positioning program and conventional rehabilitation or conventional rehabilitation alone. The positioning program was performed in supine for 60-minutes per day, 5 days per week for 5 weeks. At post-treatment (5 weeks) there was a significant between-group difference in passive ROM on shoulder abduction (measured by fluid-filled goniometer), in favour of positioning compared to conventional rehabilitation alone. There were no significant between-group differences in pROM on shoulder external rotation and flexion at post-treatment.

Conclusion: There is strong evidence from (level 1a) two high quality RCTs that static positioning of the hemiplegic arm is not more effective than conventional rehabilitation for improving shoulder external rotation passive ROM among patients with subacute stroke.
Note:
While comparable results from these 2 high quality RCTS results in a strong level of evidence, it is important to note that both studies used small sample sizes and as such may not have been sufficiently powered to find conclusive results.

Shoulder flexion passive Range of Motion
Not effective
1b

One high quality RCT (de Jong et al., 2006) investigated the effect of static arm positioning on shoulder flexion passive ROM among patients with subacute stroke. This high quality RCT randomized patients to receive an upper limb positioning program and conventional rehabilitation or conventional rehabilitation alone. The positioning program was performed in supine for 60-minutes per day, 5 days per week for 5 weeks. At post-treatment (5 weeks) there was a significant between-group difference in passive ROM on shoulder abduction (measured by fluid-filled goniometer), in favour of positioning compared to conventional rehabilitation alone. There were no significant between-group differences in pROM on shoulder external rotation and flexion at post-treatment.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that static positioning of the hemiplegic arm is not more effective than conventional rehabilitation for improving shoulder flexion passive ROM among patients with subacute stroke.

Shoulder pain at rest and on movement
Not effective
1a

Two high quality RCTs (Dean et al., 2000; de Jong et al., 2006) investigated the effect of static arm positioning on shoulder pain at rest and on movement among patients with subacute stroke.

The first high quality RCT (Dean et al., 2000) randomized patients to receive a prolonged positioning program or conventional multidisciplinary rehabilitation only. The positioning program was performed in supine and sitting positions for 60 minutes per day for 6 weeks (supine with shoulder in maximal comfortable abduction and external rotation and elbow flexion; supine with 90° shoulder abduction, maximal tolerable external rotation and elbow flexion; sitting with 90° shoulder flexion, elbow and wrist extension and web space stretch). There were no significant between group differences in shoulder pain at rest or while dressing (Visual Analogue Scale), at post-treatment (6 weeks).
Note: As indicated by the authors, the study was insufficiently powered to provide conclusive results.

The second high quality RCT (de Jong et al., 2006) randomized patients to receive an upper limb positioning program and conventional rehabilitation or conventional rehabilitation alone. The positioning program was performed in supine with maximum comfortable shoulder abduction, external rotation, elbow extension and forearm supination for 60-minutes per day, 5 days per week for 5 weeks. There was no significant between-group difference in pain at end of range of passive movements (shoulder abduction, external rotation, shoulder flexion, elbow extension, forearm supination) at 5 weeks (post treatment),

Conclusion: There is strong evidence (level 1a) from two high quality RCTs that static positioning of the hemiplegic arm is not more effective than conventional rehabilitation for managing shoulder pain at rest and on movement in patients with subacute stroke.
Note:
While comparable results from these two high quality RCTS results in a strong level of evidence, it is important to note that both studies used small sample sizes and as such may not have been sufficiently powered to find conclusive results.

Subacute stroke: Shoulder pain

Acupuncture
Insufficient evidence
5

One single-subject case study (Lindfield et al., 2002) investigated the effect of acupuncture on shoulder pain among one patient with subacute stroke. This single-subject case report assigned a patient to receive 5 acupuncture treatments to the affected side. The patient reported reduced pain (Visual Analogue Scale) and reduced use of pain relief at post-treatment (15 days).
Note: No statistical analyses of data were reported.

Conclusion: There is insufficient evidence (level 5) regarding the use of acupuncture to manage shoulder pain in patients with subacute stroke. However, a single-subject case study reported reduced shoulder pain and use of pain relief after acupuncture.

Aromatherapy
Effective
1b

One high quality RCT (Shin & Lee, 2007) investigated the effect of aromatherapy acupressure on hemiplegic shoulder pain among patients with subacute stroke. This high quality RCT randomized patients to receive aromatherapy acupressure using lavender, rosemary, and peppermint or standard (‘dry’) acupressure. All patients received acupressure to the affected shoulder for 20 minutes, twice daily for 2 weeks. At post-treatment (2 weeks) there was a significant between-group difference in shoulder pain (measured by an 8-point verbal pain rating scale), in favour of aromatherapy acupressure compared to dry acupressure.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that aromatherapy acupressure is more effective than ‘dry’ acupressure in the management of hemiplegic shoulder pain among patients with subacute stroke.

Mirror Therapy
Effective
1b

One high quality RCT (Cacchio et al., 2009a) investigated the effect of mirror therapy on upper limb pain among patients with subacute stroke. This high quality RCT randomized patients with Complex Regional Pain Syndrome type 1 to receive mirror therapy and conventional rehabilitation or conventional rehabilitation alone. Significant between-group differences in pain at rest, on shoulder flexion (Visual Analogue Scale), were seen at 1 week (post-treatment) and 6-months (follow-up), in favour of mirror therapy and conventional rehabilitation compared to conventional rehabilitation alone.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that mirror therapy with conventional rehabilitation is more effective than conventional rehabilitation alone in the management of shoulder pain among patients with subacute stroke and Complex Regional Pain Syndrome type 1.

Subacute stroke: Strapping

Muscle Tone
Not effective
1b

One high quality RCT (Griffin & Bernhardt, 2006) investigated the effect of shoulder strapping on muscle tone among patients with subacute stroke. This high quality RCT randomized patients at risk of developing hemiplegic shoulder pain to receive therapeutic strapping of the affected shoulder and standard care, placebo strapping and standard care or standard care alone. There were no significant between-group differences in muscle tone (Modified Ashworth Scale), at 4 weeks (post-treatment).

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that therapeutic strapping of the affected shoulder is not more effective than comparison interventions (placebo strapping, usual care) for management of hemiplegic shoulder muscle tone among patients with subacute stroke.

Range of Motion
Not effective
1b

One high quality RCT (Griffin & Bernhardt, 2006) investigated the effect of shoulder strapping on ROM among patients with subacute stroke. This high quality RCT randomized patients at risk of developing hemiplegic shoulder pain to receive therapeutic strapping of the affected shoulder and standard care, placebo strapping and standard care or standard care alone. There were no significant between-group differences in ROM on shoulder flexion, abduction or external rotation at 4 weeks (post-treatment).

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that therapeutic strapping of the affected shoulder is not more effective than comparison interventions (placebo strapping, usual care) for improving range of motion among patients with subacute stroke.

Shoulder pain
Effective
1b

One high quality RCT (Griffin & Bernhardt, 2006) investigated the effect of shoulder strapping on shoulder pain among patients with subacute stroke. This high quality RCT randomized patients at risk of developing hemiplegic shoulder pain to receive therapeutic strapping of the affected shoulder and standard care, placebo strapping and standard care or standard care alone. There was a significant between-group difference in the mean number of pain-free days (Ritchie Articular Index) over a period of 4 weeks, in favour of therapeutic strapping compared to standard care alone. There were no significant differences in pain between therapeutic strapping and placebo strapping.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that therapeutic strapping of the affected shoulder is more effective than usual care (but not placebo strapping) in the management of shoulder pain among patients with subacute stroke.

Chronic stroke: Mental imagery

Shoulder pain
Not effective
2a

One fair quality RCT (Cacchio et al., 2009b) investigated the effect of mental imagery on shoulder pain among patients with chronic stroke. This fair quality RCT randomized patients with Complex Regional Pain Syndrome type 1 to a mental imagery group, an active mirror group, or a covered mirror group. Interventions were performed for 30 minutes daily for 4 weeks. At post-treatment (4 weeks) there were significant between-group differences in pain (Visual Analogue Scale), in favour of the active mirror group compared to the covered mirror and mental imagery groups.
Note: After the 4-week treatment period 12 participants crossed over to the active mirror group from the covered mirror and mental imagery groups. A significant reduction in pain was reported among participants who moved to the active mirror group.

Conclusion: There is limited evidence (level 2a) from one fair quality RCT that mental imagery is not more effective than comparison interventions (mirror therapy, covered mirror therapy) in the management of shoulder pain among patients with chronic stroke and Complex Regional Pain Syndrome type 1.

Chronic stroke: Mirror Therapy

Shoulder pain
Effective
2a

One fair quality RCT (Cacchio et al., 2009b) investigated the effect of mirror therapy on shoulder pain among patients with chronic stroke. This fair quality RCT randomized patients with Complex Regional Pain Syndrome type 1 to an active mirror group, a covered mirror group or a mental imagery group. Interventions were performed for 30 minutes daily for 4 weeks. At post-treatment (4 weeks) there were significant between-group differences in pain (Visual Analogue Scale), in favour of the active mirror group compared to the covered mirror and mental imagery groups.
Note: After the 4-week treatment period 12 participants crossed over to the active mirror group from the covered mirror and mental imagery groups. A significant reduction in pain was reported among participants who moved to the active mirror group.

Conclusion: There is limited evidence (level 2a) from one fair quality RCT that mirror therapy is more effective than comparison interventions (covered mirror therapy, mental imagery) in the management of shoulder pain among patients with chronic stroke and Complex Regional Pain Syndrome type 1.

Chronic stroke: Functional Electrical Stimulation

Muscle tone
Insufficient evidence
5

One fair quality RCT (Kobayashi et al., 1999) investigated the effect of FES on hemiplegic shoulder muscle tone among patients with chronic stroke. This fair quality RCT randomized patients with shoulder subluxation to receive FES to the supraspinatus muscle (FES-S), FES to the deltoid muscle (FES-D) or no FES. All patients received conventional physical therapy that included neuromuscular facilitation, joint mobilization and muscle stretching. FES was applied at 20Hz for 5-15 minutes, 2 times/day, 5 days/week for 6 weeks. There were no significant within-group changes in muscle tone of the pectoralis major muscle (Modified Ashworth Scale) at post-treatment (6 weeks).
Note: Between-group analysis of spasticity was not performed.

Conclusion: There is insufficient evidence (level 5) regarding the effectiveness of FES compared to other interventions for muscle tone in patients with chronic stroke. However, one fair quality RCT found no significant change in muscle tone following FES to the supraspinatus or deltoid muscles of patients with chronic stroke.

Pain-free Range of Motion
Not effective
2a

One fair quality RCT (Wang et al., 2002) investigated the effect of FES on pain-free ROM among patients with stroke. This fair quality RCT randomized patients with acute or chronic stroke to receive short FES (1 cycle of FES) or long FES (two cycles of FES interspersed with 6 weeks of conventional rehabilitation alone). Each cycle of FES comprised stimulation to the supraspinatus and deltoid muscles for 6 hours/day, 5 days/week for 6 weeks. Predictably, there were no significant between-group differences in pain-free ROM on passive external rotation at 6 or 12 weeks among patients with chronic stroke, at which points both groups had received the same intervention (6 weeks of FES followed by 6 weeks of no FES). There were also no significant between-group differences at 18 weeks, at which point only the second group had received a second cycle of FES (allowing comparison of FES vs. no FES).

Conclusion: There is limited evidence (level 2a) from one fair quality RCT that FES is not more effective than comparison interventions (no FES) for improving pain-free shoulder ROM among patients with chronic stroke.

Rotator cuff tear
Insufficient evidence
5

One fair quality RCT (Kobayashi et al., 1999) investigated the effect of FES on rotator cuff tear among patients with chronic stroke. This fair quality RCT randomized patients with shoulder subluxation to receive FES to the supraspinatus muscle (FES-S), FES to the deltoid muscle (FES-D) or no FES. All patients received conventional physical therapy that included neuromuscular facilitation, joint mobilization and muscle stretching. FES was applied at 20Hz for 5-15 minutes, 2 times/day, 5 days/week for 6 weeks. There were no significant changes in rotator cuff tear (detected by Magnetic Resonance Imaging) within any group at post-treatment (6 weeks).
Note: Between-group analyses were not performed.

Conclusion: There is insufficient evidence (level 5) regarding the effectiveness of FES on rotator cuff tear among patients with chronic stroke. However, one fair quality RCT found that FES to the supraspinatus or deltoid muscles resulted in no significant change in rotator cuff tear among patients with chronic stroke.

Shoulder pain
Insufficient evidence
5

One fair quality RCT (Kobayashi et al., 1999) investigated the effect of FES on shoulder pain among patients with chronic stroke. This fair quality RCT randomized patients with shoulder subluxation to receive FES to the supraspinatus muscle (FES-S), FES to the deltoid muscle (FES-D) or no FES. All patients received conventional physical therapy that included neuromuscular facilitation, joint mobilization and muscle stretching. FES was applied at 20Hz for 5-15 minutes, 2 times/day, 5 days/week for 6 weeks. There were no significant improvements in pain on active abduction (Visual Analogue Scale) within any group at post-treatment (6 weeks).
Note: Between-group analyses of pain outcomes were not performed.

Conclusion: There is insufficient evidence (level 5) regarding the effectiveness of FES compared to other interventions in the management of shoulder pain, however one fair quality RCT reported no significant change in shoulder pain following FES among patients with chronic stroke and shoulder subluxation.

Shoulder strength
Insufficient evidence
5

One fair quality RCT (Kobayashi et al., 1999) investigated the effect of FES on shoulder strength among patients with chronic stroke. The fair quality RCT randomized patients with shoulder subluxation to receive FES to the supraspinatus muscle (FES-S), FES to the deltoid muscle (FES-D) or no FES. All patients received conventional physical therapy that included neuromuscular facilitation, joint mobilization and muscle stretching. FES was applied at 20Hz for 5-15 minutes, 2 times/day, 5 days/week for 6 weeks. The group that received FES-D demonstrated a significant improvement in maximum active abduction force at post-treatment. The other groups showed no significant improvement in maximum active abduction force.
Note: between-group analyses were not performed.

Conclusion: There is insufficient evidence (level 5) regarding the effectiveness of FES compared to other interventions for shoulder strength in patients with chronic stroke. However, one fair quality RCT found a significant improvement in abduction force following FES to the deltoid muscle among patients with chronic stroke and shoulder subluxation.

Shoulder subluxation
Conflicting
4

Two fair quality RCTs (Kobayashi et al., 1999; Wang et al., 2002) investigated the effect of FES on degree of shoulder subluxation in patients with chronic stroke.

The first fair quality RCT (Kobayashi et al., 1999) randomized patients with shoulder subluxation to receive FES to the supraspinatus muscle (FES-S), FES to the deltoid muscle (FES-D) or no FES. All patients received conventional physical therapy that included neuromuscular facilitation, joint mobilization and muscle stretching. FES was applied at 20Hz for 5-15 minutes, 2 times/day, 5 days/week for 6 weeks. At post-treatment there was a significant difference in shoulder subluxation (X-ray), in favour of both FES groups compared to the control group.

The second fair quality RCT (Wang et al., 2002) randomized patients with acute or chronic stroke to receive short FES (1 cycle of FES) or long FES (two cycles of FES interspersed with 6 weeks of conventional rehabilitation alone). Each cycle of FES comprised stimulation to the supraspinatus and deltoid muscles for 6 hours/day, 5 days/week for 6 weeks. Predictably, there were no significant between-group differences in shoulder subluxation at 6 or 12 weeks among patients with chronic stroke, at which points both groups had received the same intervention (6 weeks of FES followed by 6 weeks of no FES). There were also no significant between-group differences at 18 weeks, at which point only the second group had received a second cycle of FES (allowing comparison of FES vs. no FES).

Conclusion: There is conflicting evidence (level 4) about the effect of FES on the degree of shoulder subluxation in patients with chronic stroke. While the first fair quality RCT found a significant between-group difference in shoulder subluxation following a FES to the supraspinatus and the deltoid muscles, a second fair quality RCT found no significant difference in in shoulder subluxation following two different cycles of FES.

Stage of stroke not specified to one period: Pain-free Range of Motion

Axilla handling technique vs. distal hold technique
Effective
2b

One quasi-experimental study (Tyson & Chissim, 2002) investigated the effect of handling techniques on shoulder pain-free ROM among patients with acute to chronic stroke. This quasi-experimental study assigned patients with stroke (time since stroke 1 to 89 weeks, average time since stroke 21.3 weeks) to receive handling using an axilla hold technique and a distal hold technique, in random order. There was a significant difference in pain-free ROM on shoulder flexion (bubble goniometer) during the axilla hold compared to the distal hold.

Conclusion: There is limited evidence (level 2b) from one quasi-experimental study that an axilla hold is more effective than a distal hold in managing pain-free range of motion on shoulder flexion in patients with acute to chronic stroke.

Functional Electrical Stimulation
Not effective
2a

One fair quality RCT (Koyuncu et al., 2010) investigated the effect of functional electrical stimulation (FES) on shoulder pain-free ROM among patients with subacute and chronic stroke. This fair quality RCT randomized patients to receive FES and conventional rehabilitation or conventional rehabilitation alone. FES was applied to the supraspinatus and posterior deltoid muscles of the affected shoulder 5 times per day for 1 hour daily for 4 weeks. There were no significant between-group differences in pain on passive or active range of motion (measured by Visual Analogue Scale) at 4 weeks (post-treatment).

Conclusion: There is limited evidence (level 2a) from one fair quality RCTs that FES is not more effective than conventional rehabilitation for improving pain-free shoulder ROM among patients with subacute and chronic stroke.

Stage of stroke not specified to one period: Shoulder pain

Constraint-Induced Movement Therapy
Not effective
1b

One high quality RCT (Underwood et al., 2006) examined the effect of CIMT on shoulder pain among patients with subacute or chronic stroke. This high quality RCT randomized participants from the EXCITE trial to receive immediate CIMT or delayed CIMT (1 year after enrollment in the study). At post-treatment (2 weeks) there were no significant differences in pain (Fugl-Meyer Assessment joint pain subtest) between the group who had received CIMT and the group who had not yet received CIMT.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that CIMT is not more effective than control therapy (no CIMT*) in the management of shoulder pain among patients with subacute or chronic stroke.
*
The study used a ‘delayed CIMT’ control group, who received CIMT approximately 12 months later than the intervention group. Given that the control group had not yet received CIMT at 2 weeks, the control group is considered to have received ‘no CIMT’.

Massage
Not effective
2a

One fair quality RCT (Partridge et al., 1990) examined the effect of cryotherapy on hemiplegic shoulder pain among patients with acute to chronic stroke. This fair quality RCT randomized patients to receive cryotherapy and upper limb active range of motion (ROM) exercises, or physiotherapy using the Bobath approach daily for four weeks. Cryotherapy comprised ice application to the shoulder for 10 minutes per day. There were no significant between-group differences in pain on movement or at rest (measured by a 6-point scale), or in patients’ affective response to pain (measured using a 4-point scale). There was a significant between-group difference in frequency of pain (measured using a 5-point scale), with participants in the cryotherapy group reporting more frequent pain than those in the Bobath group.
Note: Authors report that study limitations include lack of homogeneity in the underlying pathologies of the condition, shoulder trauma occurring during the study and potential effects of the general physiotherapy treatments that were provided to all patients.

Conclusion: There is limited evidence (level 2a) from one fair quality RCT that cryotherapy is not more effective than comparison interventions (physiotherapy using the Bobath approach) in the management of hemiplegic shoulder pain among patients with acute to chronic stroke. In fact, cryotherapy was found to result in greater frequency of pain than the comparison intervention.

Functional Electrical Stimulation
Not effective
2a

One fair quality RCT (Koyuncu et al., 2010) investigated the effect of functional electrical stimulation (FES) on shoulder pain among patients with subacute and chronic stroke. This fair quality RCT randomised patients to receive FES and conventional rehabilitation or conventional rehabilitation alone. FES was applied to the supraspinatus and posterior deltoid muscles of the affected shoulder 5 times/day for 1 hour/day for 4 weeks. There were no significant between-group differences in pain at rest or pain on passive and active ROM on shoulder flexion or abduction (measured by Visual Analogue Scale) at 4 weeks (post-treatment).

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that FES is not more effective than conventional rehabilitation alone in the management of shoulder pain among patients with subacute and chronic stroke.

Massage
Effective
1b

One high quality RCT (Mok & Woo, 2004) examined the effect of massage on shoulder pain in elderly patients with stroke. This high quality RCT randomized patients with stroke (stage unspecified) to receive slow-stroke back massage (SSBM) and routine nursing care or routine nursing care alone. SSBM was provided for 10 minutes over seven consecutive sessions. At post-treatment (7 days) and follow-up (10 days) there was a significant between-group difference in shoulder pain (Visual Analogue Scale), in favour of SSBM compared to routine nursing care alone.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that massage is more effective than routine nursing care alone for improving shoulder pain in elderly patients with stroke.

Stage of stroke not specified to one period: Shoulder subluxation

Functional Electrical Stimulation
Effective
2a

One fair quality RCT (Koyuncu et al., 2010) investigated the effect of functional electrical stimulation (FES) on shoulder subluxation among patients with subacute and chronic stroke. This fair quality RCT randomised patients to receive FES and conventional rehabilitation or conventional rehabilitation alone. FES was applied to the supraspinatus and posterior deltoid muscles of the affected shoulder 5 times/day for 1 hour for 4 weeks. There was a significant between-group difference in shoulder subluxation (X-rays and Van Langenberghe Classification) at 4 weeks (post-treatment), in favour of FES compared to conventional rehabilitation alone.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that FES to the hemiplegic shoulder is more effective than conventional rehabilitation alone in the management of shoulder subluxation among patients with subacute and chronic stroke.

References

Ada L., Goddard E., McCully J., Stavrino, T., & Bampton J. (2005). Thirty minutes of positioning reduces the development of shoulder external rotation contracture after stroke: a randomized controlled trial. Archives of Physical Medicine and Rehabilitation, 86(2): 230-34.

Ancliffe, J. (1992). Strapping the shoulder in patients following a cerebrovascular accident (CVA): a pilot study. Australian Journal of Physiotherapy, 38, 37-41.

Baker, L.L. & Parker, K. (1986). Neuromuscular electrical stimulation of the muscles surrounding the shoulder. Physical Therapy, 66(12), 1930-37.

Cacchio, A., De Blasis, E., De Blasis, V., Santilli, V., & Spacca, G. (2009a). Mirror therapy in complex regional pain syndrome type 1 of the upper limb in stroke patients. Neurorehabilitation and Neural Repair, 23, 792-9.

Cacchio, A., De Blasis, E., Necozione, S., di Orio, F., & Santilli, V. (2009b). Mirror therapy for chronic complex regional pain syndrome type 1 and stroke. The New England Journal of Medicine, 361(6), 634-6.

Chantraine A., Baribeault A., Uebelhart D., & Gremion, G. (1999). Shoulder pain and dysfunction in hemiplegia: effects of functional electrical stimulation. Archives of Physical Medicine and Rehabilitation, 80(3), 328-31.

Church C., Price C., Pandyan A.D., Huntley S., Curless R., Rodgers H. (2006). Randomized controlled trial to evaluate the effect of surface neuromuscular electrical stimulation to the shoulder after acute stroke. Stroke, 37(12), 2995-3001.

Dean C.M., Mackey F.H., & Katrak P. (2000). Examination of shoulder positioning after stroke: a randomized controlled pilot trial. Australian Journal of Physiotherapy, 46,35-40.

de Jong, L.D., Nieuwboer, A., & Aufdemkampe, G. (2006). Contracture preventive positioning of the hemiplegic arm in subacute stroke patients: a pilot randomized controlled trial. Clinical Rehabilitation, 20, 656-67.

Dromerick A.W, Lang C.E, Birkenmeier, R.L., Wagner J.M., Miller J.P., Videen T.O., Powers W.J., Wolf S.L., & Edwards D.F. (2009). Very Early Constraint-Induced Movement during Stroke Rehabilitation (VECTORS): a single-center RCT. Neurology, 73(3), 195-201.

Griffin, A. & Bernhardt, J. (2006). Strapping the hemiplegic shoulder prevents development of pain during rehabilitation: a randomized controlled trial. Clinical Rehabilitation, 20, 287-95.

Gustafsson, L. & McKenna, K. (2006). A programme of static positional stretches does not reduce hemiplegic shoulder pain or maintain shoulder range of motion – a randomized controlled trial. Clinical Rehabilitation, 20, 277-86.

Hanger, H.C., Whitewood, P., Brown, G., Ball, M.C., Harper, J., Cox, R., & Sainsbury, R. (2000). A randomized controlled trial of strapping to prevent post-stroke shoulder pain. Clinical Rehabilitation, 14, 370-80.

Lindfield, H. (2002). Chronic shoulder pain in stroke. Are we missing the acupoint? Physiotherapy Research International, 7(1) 44-50.

Linn, S. L., Granat, M. H., & Lees, K. R. (1999). Prevention of shoulder subluxation after stroke with electrical stimulation. Stroke, 30(5), 963-8.

Mok, E. & Woo, C.P. (2004). The effects of slow-stroke back massage on anxiety and shoulder pain in elderly stroke patients. Complementary Therapies in Nursing & Midwifery, 10, 209-16.

Partridge, C.J., Edwards, S.M., Mee, R., & van Langenberghe, H.(1990). Hemiplegic shoulder pain: a study of two methods of physiotherapy treatment. Clinical Rehabilitation, 4, 43-9.

Kobayashi, H., Onishi, H., Ihashi, K., Yagi, R., & Handa, Y. (1999). Reduction in subluxation and improved muscle function of the hemiplegic shoulder joint after therapeutic electrical stimulation. Journal of Electromyography and Kinesiology, 9, 327-36.

Koyuncu, E., NakipoÄŸlu-Yüzer, G.F., DoÄŸan, A., & Özgirgin, N. (2010). The effectiveness of functional electrical stimulation for the treatment of shoulder subluxation and shoulder pain in hemiplegic patients: a randomized controlled trial. Disability and Rehabilitation, 32(7), 560-6.

Kumar, R., Metter, E.J., Mehta, A.J., & Chew, T. (1900). Shoulder pain in hemiplegia: the role of exercise. American Journal of Physical Medicine and Rehabilitation, 69(4), 205-8.

Shin, B-C. & Lee, M.S. (2007). Effects of aromatherapy acupressure on hemiplegic shoulder pain and motor power in stroke patients: a pilot study. The Journal of Alternative and Complimentary Medicine, 13(2), 247-51.

Tyson, S.F. & Chissim, C. (2002). The immediate effect of handling technique on range of movement in the hemiplegic shoulder. Clinical Rehabilitation, 16, 137-40.

Underwood, J., Clark, P.C., Blanton, S., Aycock, D.M., & Wolf, S.L. (2006). Pain, fatigue, and intensity of practice in people with stroke who are receiving constraint-induced movement therapy. Physical Therapy, 86(9), 1241-50.

Wang, R.Y., Yang, Y.R., Tsai, M.W., Wang, W.T.J., & Chan, R.C. (2002). Effects of functional electrical stimulation on upper limb motor function and shoulder range of motion in hemiplegic patients. American Journal of Physical Medicine and Rehabilitation, 81, 283-90. 

Survey