Text Size:    +   -

Management of Shoulder Pain & Complex Regional Pain Syndrome (CRPS) following Stroke

February 2016

The Canadian Stroke Best Practice Recommendations for Stroke Rehabilitation, 5th Edition (2015) is published in the International Journal of Stroke (IJS) and available freely online. To access the specific recommendations for Management of Shoulder Pain & Complex Regional Pain Syndrome (CRPS) following Stroke, and all other sections of the Stroke Rehabilitation recommendations, please click on this URL which will take you to the recommendations online in the IJS: http://journals.sagepub.com/doi/pdf/10.1177/1747493016643553

For the French version of these recommendations, open the appendix at this link :  http://wso.sagepub.com/content/suppl/2016/04/18/1747493016643553.DC1/Stroke_Rehabilitation_2015_IJS_Manuscript_FINAL_FRENCH.pdf

All other supporting information, including performance measures, implementation resources, evidence summaries and references, remain available through www.strokebestpractices.ca, and not through the IJS.  Please click on the appropriate sections on our website below for this additional content.



The incidence of shoulder pain following a stroke is high. As many as 72 percent of adult stroke patients report at least one episode of shoulder pain within the first year after stroke. Shoulder pain may inhibit patient participation in rehabilitation activities, contribute to poor functional recovery and can also mask improvement of movement and function. Hemiplegic shoulder pain may contribute to depression and sleeplessness and reduce quality of life.


System Implications

To achieve timely and appropriate assessment and management of shoulder pain the organization requires:

  • Organized stroke care, including stroke rehabilitation units with a critical mass of trained interprofessional staff during the rehabilitation period following stroke.
  • Equipment for proper limb positioning (e.g. pillows, arm troughs).

To achieve timely and appropriate assessment and management of shoulder pain the organization should provide:

  • Initial assessment of active or passive upper extremity range of motion of shoulder, based on Chedoke-McMaster Stroke Assessment score and assessment of external rotation performed by clinicians experienced in stroke rehabilitation.
  • Timely access to specialized, interprofessional stroke rehabilitation services for the management of shoulder pain.
  • Timely access to appropriate rehabilitation therapy intensity/ treatment modalities for management or reduction of shoulder pain in stroke survivors.
  • Long-term rehabilitation services in nursing and continuing care facilities, and in outpatient and community programs.
  • Physicians trained in stroke care and, where needed, intra-articular shoulder injections and botulinum toxin injections.


Performance Measures

  1. Proportion of stroke patients who experience shoulder pain in acute care hospital, inpatient rehabilitation and following discharge into the community (NRS tool has a self-report question about pain on admission/discharge).
  2. Length of stay during acute care hospitalization and inpatient rehabilitation for patients experiencing shoulder pain (versus patients not experiencing shoulder pain).
  3. Proportion of stroke patients who report shoulder pain at three-month and six-month follow-up.
  4. Pain intensity rating change, from baseline to defined measurement periods.
  5. Motor score change, from baseline to defined measurement periods.
  6. Range of shoulder external rotation before and after treatment for shoulder pain.
  7. Proportion of patients with restricted range of motion related to shoulder pain.

Measurement notes

  • Performance measure 4: Standardized rating scales should be used for assessment of pain levels and motor scores.
  • Some data will require survey or chart audit. The quality of documentation related to shoulder pain by healthcare professionals will affect the quality and ability to report some of these performance measures.
  • Audit tools at a local level may be helpful in collecting shoulder pain data on patients who experience shoulder pain.


Implementation Resources and Knowledge Transfer Tools

Health Care Provider Information

Patient Information


Summary of the Evidence, Evidence Tables and References

Evidence Table 5.3: Management of Shoulder Pain & Complex Regional Pain Syndrome (CRPS) following Stroke

The use of supportive slings and supports has been shown to reduce the amount of subluxation (evident upon radiographic examination) and may also help to reduce hemiplegic shoulder pain. A Cochrane review authored by Ada et al. (2005) included the results from 4 RCTs evaluating the use of strapping (n=3) and hemisling (n=1). All patients were in the acute phase of stroke (less than 4 weeks) with a flaccid arm with no history of shoulder pain. The number of pain-free days associated with treatment was significantly greater; (mean difference: 13.6 days, 95% CI 9.7 to 17.8, p<0.0001); however, the results from only two studies were included in the pooled result. Among two RCTs that examined the use of strapping, specifically to prevent the development of shoulder pain, the results were conflicting (Hanger et al. 2000, Griffin & Bernhardt 2006). In a recent RCT, the use of the tri-pull method of taping paired with conventional therapy (experimental group) was compared to a sham taping with therapy; results demonstrated a significant reduction in pain among the experimental group (Pandian et al. 2013). A recent meta-analysis, including the results from five RCTs, reported that shoulder positioning programs were not effective in preventing or reducing the range of motion loss in the shoulders’ external rotation (Borisova & Bohannon 2009).

Ada and Foongchomcheay (2002) conducted a meta-analysis to examine the effect of electrical stimulation on shoulder subluxation following stroke. Participants with subluxation or shoulder muscle paralysis in both the acute and chronic stages of stroke, from seven RCTs were included. The results suggested that early treatment, starting with electrical stimulation for 2 hours per day increasing to between 4 and 6 hours per day, in addition to conventional therapy helps to prevent the development of hemiplegic shoulder while later treatment helps to reduce pain. A systematic review of 14 studies conducted by Mathieson et al. (2014) found that the use of functional electrical stimulation (FES) plus imagery (mirror therapy or mental imagery) was the most effective treatment compared to passive and active assisted therapy, while usual care plus FES was also beneficial. In a RCT by Manigandan et al. (2014), participants received either electrical stimulation to the supraspinatus and posterior deltoid plus physio- and occupational therapy for 5 weeks (group 1), or electrical stimulation to the supraspinatus, posterior deltoid and long head of the bicep plus physio- and occupational therapy for 5 weeks (group 2). The authors found that group 2 improved significantly compared to group 1 in the reduction of shoulder subluxation, improvement of passive pain free external rotation, and improvement in range of active shoulder abduction ROM. Church et al. (2006) randomized 176 patients to receive active or sham surface FES treatments in addition to conventional therapy, for four weeks following acute stroke. There was no significant difference in prevalence of pain between groups post intervention. Koyuncu et al. (2010) also reported no differences in shoulder pain of all patients during resting, passive range of motion or active range of motion following 20 sessions of surface FES in addition to inpatient rehabilitation, compared with patients who did not receive electrical stimulation treatments. An RCT by de Jong et al. (2013) compared the effects of arm stretch positioning combined with motor amplitude NMES in relation to sham arm positioning with sham NMES. No significant differences in shoulder pain between the control and experimental group were observed at 8 weeks.

There is evidence that treatment with botulinum toxin type a (BTX-A) may help to improve hemiplegic shoulder pain, but the results from systematic reviews and RCTs are not consistent. A Cochrane review (Singh & Fitzgerald 2010) examined the efficacy of the use of BTX-A toxin in the treatment of shoulder pain. Six RCTs were included, five of which included patients with post-stroke shoulder pain. Treatment with BTX-A was associated with reductions in pain at 3 and 6 months, but not at 1 month following injection. De Boer et al (2008) randomized 22 patients, an average of 6 months following stroke with significant shoulder pain to receive a single injection of 100 U Botox or placebo to the subscapularis muscle. In addition, all patients received some form of physical therapy. While pain scores improved in both groups over time, there was no significant difference at 12 weeks following treatment, nor was there significant improvement between groups in degree of humeral external rotation.

Intra-articular corticosteroids injections may also help to improve symptoms of shoulder pain. Rah et al. (2012) randomized 58 patients with chronic shoulder pain (at least 3/10 on a Visual Analog Scale (VAS) to receive a single subacromial injection of 40 mg triamcinolone acetonide or lidocaine (control condition). All patients participated in a standardized exercise program. There was significant reduction in the average shoulder pain level at day and night, measured on a 10 cm VAS at 8 weeks associated with steroid injection. In contrast, Snels et al. (2000) reported that in 37 patients with hemiplegic shoulder pain (≥ 4 on a 0 to 10 VAS) randomized to receive three injections (1-2 weeks apart) of 40 mg triamcinolone acetonide or placebo, active treatment was not associated with improvements in pain scores three weeks later. Dogan et al. (2013) found that compared to traditional rehabilitation alone, the addition of intra-articular steroid, and intra-articular steroid plus hydraulic distention significantly improved range of motion immediately after treatment and at 1 month follow-up. Both steroid groups had significant improvements on VAS score at rest and during activity but the group which received steroid plus hydraulic distention were significantly more effective than only the intra-articular steroid injection and therapy.

There is no definitive therapeutic intervention for complex regional pain syndrome (CRPS). Although a wide variety of preventative measures and treatments have been used including exercise, heat, contrast baths, hand desensitization programs, splints, medications, and surgical options, there is little evidence that many of the commonly-used treatments are effective. Although physiotherapy is regarded as the cornerstone of integrated treatment, no controlled trials have been conducted to evaluate its effect in preventing the development of CRPS. There is some evidence that a two-week, tapering dose of 32 or 40 mg of oral corticosteroids is more effective than either NSAIDS or placebo in improving symptoms of CRPS (Bruas et al. 1994, Kalita et al. 2006). An overview conducted by O’Connel et al. (2013) evaluated 19 studies that used a variety of interventions to treat pain, disability, and CRPS. The authors found moderate quality evidence that intravenous regional blockade with guanethidine is not effective in CRPS and is associated with adverse events. Low quality evidence was found for biphosphates, calcitonin or daily IV of ketamine for the treatment of pain compared to a placebo. Both motor imagery and mirror therapy may be effective for the treatment of pain compared to a control condition. There is some evidence that local anaesthetic sympathetic blockade, physiotherapy, and occupational therapy are not effective for CRPS.