There are approximately 12 000 new cases of spinal cord injury (SCI) per year in the United States.1 Depending on the level of SCI, manual wheelchairs are chosen to be the primary mode of locomotion. Shoulder pain is highly prevalent in manual wheelchair users. Jain et al2 reported shoulder pain to be present in 38% to 67% of manual wheelchair users with SCI. Furthermore, the incidence and severity of shoulder pain have been shown to increase with time postinjury. Shoulder pain in this population is attributed to chronic overuse, muscle imbalance, postural changes, repetitive trauma, and acute physiological changes directly from injury.3 The upper limb sustains repetitive weight-bearing loads through daily activities, which can lead to trauma at the shoulder joint, often resulting in pain. Morrow et al4 reported that joint intersegmental forces and moments vary with different activities, the highest joint forces were associated with pressure relief, ramp propulsion, and the start of propulsion. Because the upper extremities are used as the primary mode of force production for locomotion and transfers, individuals with SCI who use manual wheelchairs are unable to avoid using their upper extremities in the presence of pain.
Studies have shown that surgery is typically considered as an option to alleviate shoulder pain in this population only as a last resort; among other reasons the strict protocols postsurgery are not feasible for individuals who wish to maintain independence.5 While patient education regarding activity modification, mechanisms of injury, and self-management of shoulder pain remains the standard of care for individuals with SCI, previous studies have shown clinically significant improvement of shoulder pain with strengthening and stretching interventions in persons without disabilities.6 Thus, researchers have been investigating the feasibility, safety, and effectiveness of exercise interventions for shoulder pain in manual wheelchair users with SCI. The objective of this review was to evaluate current literature on the effectiveness of exercise programs on the reduction of shoulder pain in manual wheelchair users with SCI.
The design of this systematic review was developed using the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). The PRISMA statement includes a 27-item checklist designed to be used for reporting systematic reviews.7 The protocol for this systematic review is described later.
We conducted a systematic computerized search of the literature in PubMed, CINAHL, Web of Science, and EMBASE for articles published between 1966 and January 22, 2014. After completion of this literature search, but before submitting this manuscript for publication, we identified another study that was published after January 2014. We included this study in this systematic review. We used the key words related to SCI, manual wheelchairs, and shoulder pain. Reference lists of selected publications were checked to identify additional relevant publications that were not in the computerized search. Two independent reviewers screened titles and abstracts of all identified citations to select all relevant articles. The full text of an article was retrieved if the article passed the eligibility screening.
An article was eligible if it was published in the English language, peer-reviewed journal and met the following criteria: study population consisting of adult (age ≥ 18 years) individuals with SCI and who are manual wheelchair users with shoulder pain; exercise intervention for management of shoulder pain; prospective study design with or without a comparator group; and use of the Wheelchair Users' Shoulder Pain Index (WUSPI) as an outcome measure. All criteria were independently applied by 2 reviewers to the full text of the articles that passed the title and abstract screening. Any disagreement between the 2 independent reviewers was resolved by discussion.
Outcome Measure Assessed
The WUSPI is a 15-question, self-reported measure of the intensity of shoulder pain during 15 different activities, encompassing transfers, activities of daily living, and wheelchair mobility. The questionnaire utilizes a series of visual analog scales consisting of 10-cm lines anchored by “no pain” and “worst pain ever experienced,” with a maximum total score of 150, in which a higher score indicates higher pain levels.8 The WUSPI is a reliable and valid outcome measure for measurement of shoulder pain in wheelchair users.9 In cases where 1 or more items were not applicable or not regularly performed by a subject, the studies utilized a Performance-Corrected WUSPI (PC-WUSPI) score. The PC-WUSPI score is calculated by dividing the raw total WUSPI score by the number of activities performed and then multiplying by 15.8 Although PC-WUSPI scores were used in a number of the included studies, the psychometric properties of the PC-WUSPI have not been described.8
Data abstraction was performed by 1 reviewer and checked for accuracy and completeness by a second reviewer. Data abstracted from the articles included the number and characteristics of subjects, different groups and interventions applied, duration of intervention, and mean WUSPI scores with standard deviations at baseline and postintervention. Change scores for the WUSPI were also calculated for baseline and postintervention differences.
The methodological quality of each study was assessed individually by reviewers using the checklist proposed by Downs and Black.10 This checklist assesses the quality of both randomized and nonrandomized control trials specifically related to intervention, potential confounders, and the outcome. The checklist uses 27 criteria across 5 sections: (1) Reporting, (2) External Validity, (3) Internal Validity-Bias, (4) Internal Validity-Confounding (Selection Bias), and (5) Power.10 Disagreement was resolved by discussion. Following individual assessment of the studies and discussion for any discrepancies, we used professional judgment and the definitions provided by the Agency for Healthcare Research and Quality to decide upon a qualitative description of good, fair, or poor quality for each study.11 The definitions of quality are reported as follows: A “good” study was defined as having the least bias, and results are considered valid. A good study has a clear description of the population, setting, interventions, and comparison groups; uses a valid approach to allocate participants to alternative treatments; has a low dropout rate; and uses appropriate means to prevent bias, measure outcomes, and analyze and report results. A “fair” study may be missing information or may not have used a valid approach to participant selection, or allocation to intervention, or outcome measurement, thereby introducing susceptibility to bias. As the fair-quality category is broad, studies with this rating vary in their strengths and weaknesses. A “poor” rating indicates a high potential for significant bias.
Selection of Studies
We identified 229 titles through database and reference searches (Figure 1). We assessed 16 full-text articles, of which 6 met our inclusion criteria.3,12–16 We later identified and included a seventh eligible study published after we completed our systematic search of the literature.17 These 7 studies involved a combined total of 158 manual wheelchair users. Three studies were randomized control trials,3,12,13 while the other 4 were cohort studies.14–17 All 7 studies utilized exercise interventions to manage shoulder pain. Four studies3,13,15,17 utilized stretching and strengthening home exercise programs (HEMs) targeting the muscles of the shoulder girdle. One study12 utilized electromyographic (EMG) biofeedback in addition to stretching and strengthening targeting the muscles of the shoulder girdle. One study14 utilized arm ergometry in addition to strengthening targeting the shoulder girdle, and another study16 utilized double poling ergometry. Norrbrink et al16 used a custom-built seated double-poling ergometer modified for persons with motor impairments in the lower extremities. Double poling ergometry is an activity that simulates the motion performed by a cross-country skier using the upper extremities to propel both poles simultaneously in the posterior direction. This form of training has been used in other research with subjects with SCI and depicted by Lindberg et al18 in a photographic form. A description of the included studies' participants, interventions, comparators, and outcome measures is found in Table 1. Two studies3,13 were rated as “good” quality and 5 studies12,14–17 were rated as “fair” quality. Curtis et al3 was graded as good quality because the study included an experimental group and a no-intervention control group in a randomized controlled trial (RCT) design. The study by Middaugh et al12 was given a fair quality grade despite randomizing participants into 1 of 2 treatment groups, because of lacking a true control group without intervention. The study by Mulroy et al13 was graded as good quality because it was an RCT that randomized symptomatic individuals in both experimental and control groups. The study by Nawoczenski et al15 was graded as fair quality on the basis of the fact that it was a clinical trial that included an asymptomatic control group. The remaining studies were ranked as fair quality because of cohort study designs that lacked control groups.
Summary of Individual Studies
The duration of intervention, change score for each intervention group, and whether a statistical significance was found for each of the included articles are summarized in Table 2
Across the 7 studies that met the inclusion criteria, all intervention groups demonstrated a reduction in shoulder pain as determined by a decrease in WUSPI score. Three studies3,13,15 performed between-group comparisons: Curtis et al3 compared an HEP to no intervention—those who received the HEP indicated a WUSPI change score of −8.3, while those receiving no intervention indicated a −0.4 WUSPI change score; Mulroy et al13 compared an HEP to an attention control group—the HEP group WUSPI change score was −36.3 while the attention control group actually increased by 0.2; finally, Nawoczenski et al15 compared an HEP to an asymptomatic control group—the HEP group WUSPI change score was −22.85, while the asymptomatic control group change score was +2.01. The between-group comparisons for Mulroy et al13 and Nawoczenski et al15 were statistically significant, but the P values for Curtis et al3 for main effect of treatment group and interaction between group and time were not significant. Middaugh et al12 reported WUSPI change scores associated with both an HEP group and a group receiving an HEP coupled with EMG biofeedback (HEP+EMG). The HEP WUSPI change score was −11.9 (P = 0.42) while the HEP+EMG was −37.0 (P = 0.02). The HEP+EMG change score was significant, but the change score for the HEP alone was not.
Several authors have reported outcomes associated with the study interventions with no between-group comparisons.14,16,17 Nash et al14 employed a circuit resistance training program that produced a WUSPI change score of −26.8 (statistically significant). Norrbrink et al16 utilized a program consisting solely of double-poling upper body ergometry, producing a change score of −19.0, but the statistical significance was not reported for this study. Van Straaten et al17 utilized an HEP with videoconferencing for feedback and progression during the intervention period, producing a WUSPI change score of −10.3 (P = 0.007).
In reviewing the literature available for manual wheelchair users with SCI, we set out to evaluate the effectiveness of exercise programs that have been developed to manage shoulder pain. Although the studies identified were heterogenous in regard to type of study and interventions investigated, each intervention was effectively and safely implemented and resulted in a decreased score on the WUSPI. The articles reviewed consistently demonstrated a reduction in WUSPI scores following an intervention period based on the studies' parameters. The reduction in WUSPI score exceeded the minimal detectable change of 5.10,9 indicating a clinically important change in shoulder pain. There were no adverse events reported in any of the studies that were directly attributable to the exercise intervention.
It is important to consider differences among the studies with regard to the time periods over which the exercise interventions were applied. The durations of study interventions ranged from 8 weeks12,15 to 6 months.3 Because of heterogeneity in interventions applied, baseline WUSPI scores of the participants, and comparators across the 7 studies, it is not possible to conclude whether a longer intervention time frame is more beneficial in reducing pain scores. To determine the most effective duration of exercise intervention, a future randomized controlled trial comparing different intervention periods is needed, utilizing an identical intervention across the different time groups and controlling for differences in baseline shoulder pain.
Across the 7 studies reviewed, there were both similarities and differences in exercise prescription. Of the 6 studies that evaluated stretching interventions, 5 utilized anterior shoulder joint or pectoralis stretching,3,12,13,15,17 3 utilized biceps stretching,3,12,15 3 utilized upper trapezius stretching,12,13,15 3 utilized posterior capsule stretching,13,15,17 and 1 utilized cervical rotation stretching.12 There were 7 studies that evaluated strengthening interventions. Two studies utilized arm ergometry as a form of endurance strengthening intervention,14,16 with 1 of these 2 alternating arm ergometry with resistance training. The other 5 studies provided strengthening exercises without incorporating arm ergometry.3,12,13,15,17 Of these studies, all five utilized a shoulder external rotation exercise,3,12,13,15,17 4 utilized scapular retraction,3,12,13,17 3 studies utilized shoulder adduction exercises,3,12,13 1 utilized a shoulder extension exercise.12 One study utilized a shoulder elevation in the scaption plane exercise,13 1 utilized a middle and lower trap,15 and 2 utilized a serratus anterior exercise.15,17 Two studies utilized EMG biofeedback in the intervention.12,15 One study12 used EMG biofeedback in the HEP + EMG biofeedback intervention group before introducing the HEP to train the participant to utilize the correct muscles and demonstrated improvement in the effectiveness of the exercises. The other study15 utilizing EMG biofeedback introduced it early as a technique to train the muscles and again at follow-up if needed for each participant. Based on the common exercises that improve the WUSPI and pain in this population, a standardized shoulder program could be created.
Despite the variability among studies noted earlier, based on some commonalities among studies, general recommendations for exercises can be forwarded for this population, from which further RCTs might be developed. The evidence reviewed indicates a number of potential combinations of strengthening exercises targeting the posterior shoulder musculature for power and endurance, and stretching targeting the muscles of the anterior shoulder joint structures. Three studies utilized unilateral, shoulder external rotation with the elbow at 90° of flexion, and the shoulder in a neutral position.3,12,13 This exercise was to be performed using a resistance band. Two studies utilized simultaneous, bilateral external rotation with a resistance band with the elbows at 90° of flexion and neutral positioning of the shoulders.15,17 These exercises specifically targeted the external rotator as part of the posterior shoulder musculature. These strengthening exercises and stretching, possibly in combination with EMG biofeedback training and arm ergometry, could be used to create a program targeting the management and treatment of shoulder pain in this population. Furthermore, there may be value in understanding whether a home-based exercise program could be effective for improving strength and reducing pain.19
There is concern about exacerbating shoulder pain with exercises in manual wheelchair users with SCI. Overuse injuries may result from inappropriate strengthening programs. Although no explanation was given for the increased pain scores at the 2-month time period in Curtis et al,3 this is the only study that included exercise on a daily basis. As such, we cannot ignore the possibility that increased intensity of exercise may have contributed to increased pain. To effectively implement a new strengthening and stretching exercise program for manual wheelchair users, customization of the program to suit the unique needs of each individual is necessary. This is particularly important with regard to the amount of resistance, the number of sets and repetitions, rest time, and modification of mechanics. This could be achieved by developing a standardized set of exercises from which a therapist selects the exercises appropriate to a specific individual. For example, in the studies by Mulroy et al13 and Nash et al,14 each participant was given a different level of resistance to utilize in strengthening exercises based on initial evaluation of individual repetition maximum. In Nawoczenski et al,15 follow-up visits with a physical therapist after initiation of the exercise program were used to ensure correct form and progress resistance, intensity, and duration for each participant. These 3 studies demonstrated the highest WUSPI change scores.13–15 In addition, Van Straaten et al17 provided modifications in position, resistance (isometric vs isotonic), and the number of repetitions; exercises were then advanced through videoconference feedback throughout the intervention time frame. This highlights the importance of and need for further research for evaluating baseline exercise capabilities in manual wheelchair users with SCI, identifying the most critical individual exercises for improvement of shoulder pain, and tailoring exercise programs specifically to suit individual needs and prevent pain exacerbation.
Limitations of the Review
Our systematic review is not without limitations. We excluded non-English language publications, so there may be informative studies that we did not identify. Our selection of the WUSPI as the primary outcome of interest may have decreased our opportunity to identify other relevant research related to the management and treatment of shoulder pain that did not use the WUSPI. We only considered studies that used exercise programs as an intervention of interest, but there are other interventions such as wheelchair propulsion and positioning strategies, acupuncture, and surgical options that may play a role in the treatment of wheelchair-associated shoulder pain among individuals with SCI.
Additional limitations to this review include small sample sizes of the primary studies; a limited number of eligible studies; a heterogenous mix of study designs and differences in reporting raw data, change scores, and statistical significance of outcomes. Baseline pain scores of participants varied across the eligible studies, which may have influenced the relative degree of pain reduction that could be achieved. For example, in Curtis et al,3 the presence of pain was not an inclusion criterion and only 50% of the participants had shoulder pain at the start of the study. Furthermore, Nawoczenski et al15 had an asymptomatic control group while the treatment group had levels of shoulder pain at baseline that were in the low to moderate severity range. Another limitation pertains to study participants: 3 articles included some participants who did not have SCI but had a different reason for using a manual wheelchair, these included 2 participants with amputation, 2 participants with cerebral palsy, 3 participants with postpolio syndrome, 1 participant with multiple sclerosis, and 1 participant with spina bifida.3,15,17
Of the studies that included only participants with SCI, the participants had various levels of injury. Within 1 study some participants may have had full innervation and use of the upper extremities while other participants did not. The prevalence and intensity of shoulder pain with performance of functional activities are significantly higher in those with tetraplegia than in those with paraplegia, and the report of bilateral shoulder pain is higher in those with tetraplegia as well.20,21 This is to be expected, considering the greater involvement of the upper extremities in those with tetraplegia. The pathology underlying shoulder pain is likely to be different among persons with paraplegia and persons with tetraplegia, considering the impairments of muscle balance caused by paralysis, decreased range of motion, and spasticity present in those with tetraplegia. In the study by Silfverskiold and Waters,21 more than 75% of persons with tetraplegia had shoulder pain in the first 6 months after SCI. In addition, in participants with shoulder pain that was initially associated with spasticity, the pain persisted at the time of follow-up.21 In 3 studies, the PC-WUSPI was used in an attempt to correct for activities that participants could not perform because of level of injury3,12,15; however, it is a limitation of our review that we were unable to compare differences among participants with paraplegia versus participants with tetraplegia.
It has been reported that subjects with tetraplegia also decrease their performance of functional activities with advancement of age resulting in functional disability.20 Performance of activities that provoke pain results in less frequent performance of those activities. This may account for differences in PC-WUSPI scores reported with advancing age, a time when a greater duration of wheelchair use and exacerbation of shoulder pain and dysfunction is more likely to occur, resulting in less frequent performance of activities that provoke pain.20 While both people with paraplegia and those with tetraplegia seem likely to benefit from a strengthening and stretching program, ability to perform some of the exercises may be limited in persons with tetraplegia. It is unclear whether both SCI groups would have the same response to an exercise program.
Finally, the methodological quality of the studies available for review varied greatly. The current paucity of high-quality randomized controlled trials and the lack of consistency in reporting outcomes across studies preclude meta-analysis to calculated pooled effect sizes. Study participants demonstrated variability in improved pain ratings, which makes it difficult to determine magnitude of response to intervention. Regarding long-term outcomes, none of the included articles provided more than 6-month intervention3 or more than 6-month follow-up,12,17 which makes it difficult to predict effectiveness of interventions beyond this time frame. Many unanswered questions remain to be considered in future clinical trials.
Current literature supports the use of various exercise programs to reduce shoulder pain and increase daily function over time in manual wheelchair users with SCI, with the caveat that the existing literature reports on findings from exercise programs that differ across published studies. The studies included in this review support the effectiveness of various therapeutic interventions on the management and treatment of shoulder pain, as demonstrated by a clinically meaningful reduction in self-reported WUSPI scores. Exercise is a feasible, conservative, and therapeutic intervention for the treatment of shoulder pain in our population of interest. Additional high-quality studies are needed to further evaluate and differentiate techniques for the reduction of shoulder pain. Further studies should also focus on determining the most efficient and effective time frame of intervention with the potential to develop a standardized set of exercises for initial consideration that can be customized to specific needs of the individual wheelchair user.
1. National Spinal Cord Injury
Statistical Center. Spinal cord injury
facts and figures at a glance. www.uab.edu/nscisc
. Published February 2013. Accessed July 3, 2014.
2. Jain NB, Higgins LD, Katz JN, Garshick E. Association of shoulder pain with the use of mobility devices in persons with chronic spinal cord injury
. PMR. 2010;2(10):896–900. doi:10.1016/j.pmrj.2010.05.004.
3. Curtis KA, Tyner TM, Zachary L, et al. Effect of a standard exercise protocol on shoulder pain in long-term wheelchair users. Spinal Cord. 1999;37(6):421–429.
4. Morrow MMB, Hurd WJ, Kaufman KR, An K-N. Shoulder demands in manual wheelchair users across a spectrum of activities. J Electromyogr Kinesiol. 2010;20(1):61–67. doi:10.1016/j.jelekin.2009.02.001.
5. Popowitz RL, Zvijac JE, Uribe JW, Hechtman KS, Schürhoff MR, Green JB. Rotator cuff repair in spinal cord injury
patients. J Shoulder Elb Surg Am Shoulder Elb Surg Al. 2003;12(4):327–332. doi:10.1016/mse.2003.S1058274603000351.
6. Escobedo EM, Hunter JC, Hollister MC, Patten RM, Goldstein B. MR imaging of rotator cuff tears in individuals with paraplegia. Am J Roentgenol. 1997;168(4):919–923.
7. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg Lond Engl. 2010;8(5):336–341. doi:10.1016/j.ijsu.2010.02.007.
8. Rehab Measures—Wheelchair User's Shoulder Pain Index. http://rehabmeasures.org/Lists/RehabMeasures/DispForm.aspx?ID=1075
. Accessed July 3, 2014.
9. Curtis KA, Roach KE, Applegate EB, et al. Development of the Wheelchair User's Shoulder Pain Index (WUSPI). Paraplegia. 1995;33(5):290–293. doi:10.1038/sc.1995.65.
10. Downs SH, Black N. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. J Epidemiol Community Health. 1998;52(6):377–384.
11. Home | AHRQ Effective Health Care Program. http://www.effectivehealthcare.ahrq.gov/
. Accessed July 3, 2014.
12. Middaugh S, Thomas KJ, Jackson T, Smith AR, McFall TL, Klingmueller J. EMG biofeedback and exercise for treatment of cervical and shoulder pain in individuals with a spinal cord injury
: a pilot study. Top Spinal Cord Inj Rehabil. 2013;19(4):311–323. doi:10.1310/sci1904-311.
13. Mulroy SJ, Thompson L, Kemp B, et al. Strengthening and optimal movements for painful shoulders (STOMPS) in chronic spinal cord injury
: a randomized controlled trial. Phys Ther. 2011;91(3):305–324. doi:10.2522/ptj.20100182.
14. Nash MS, van de Ven I, van Elk N, Johnson BM. Effects of circuit resistance training on fitness attributes and upper-extremity pain in middle-aged men with paraplegia. Arch Phys Med Rehabil. 2007;88(1):70–75. doi:10.1016/j.apmr.2006.10.003.
15. Nawoczenski DA, Ritter-Soronen JM, Wilson CM, Howe BA, Ludewig PM. Clinical trial of exercise for shoulder pain in chronic spinal injury. Phys Ther. 2006;86(12):1604–1618. doi:10.2522/ptj.20060001.
16. Norrbrink C, Lindberg T, Wahman K, Bjerkefors A. Effects of an exercise programme on musculoskeletal and neuropathic pain after spinal cord injury
—results from a seated double-poling ergometer study. Spinal Cord. 2012;50(6):457–461. doi:10.1038/sc.2011.160.
17. Van Straaten MG, Cloud BA, Morrow MM, Ludewig PM, Zhao KD. Effectiveness of home exercise on pain, function, and strength of manual wheelchair users with spinal cord injury
: a high-dose shoulder program with telerehabilitation. Arch Phys Med Rehabil. 2014;95(10):1810–1817.e2. doi:10.1016/j.apmr.2014.05.004.
18. Lindberg T, Arndt A, Norrbrink C, Wahman K, Bjerkefors A. Effects of seated double-poling ergometer training on aerobic and mechanical power in individuals with spinal cord injury
. J Rehabil Med. 2012;44(10):893–898. doi:10.2340/16501977-1038.
19. Sasso E, Backus D. Home-based circuit resistance training to overcome barriers to exercise for people with spinal cord injury
: a case study. J Neurol Phys Ther. 2013;37(2):65–71. doi:10.1097/NPT.0b013e31829247a9.
20. Curtis KA, Drysdale GA, Lanza RD, Kolber M, Vitolo RS, West R. Shoulder pain in wheelchair users with tetraplegia and paraplegia. Arch Phys Med Rehabil. 1999;80(4):453–457.
21. Silfverskiold J, Waters RL. Shoulder pain and functional disability in spinal cord injury
patients. Clin Orthop. 1991;272:141–145.