Stroke often causes disability among elderly people. Hemiplegic shoulder pain (HSP) is one of the most common complications in patients after a stroke. The incidence of HSP is approximately 17–72% (1). HSP, which is related to depression and a poor quality of life, negatively affects functional recovery of the upper extremity and activities of daily living (ADLs) (2). The pathogenesis of HSP includes shoulder subluxation, adhesive capsulitis, bursitis, shoulder-hand syndrome, among others (3). However, the exact etiology of HSP remains unknown and many complicated factors are involved. Glenohumeral subluxation (GHS) may be considered a potential cause of shoulder pain development (4).

The prevalence of GHS was reported to be 15–81% (2). There is speculation that the peri-articular tissue of the shoulder may be overstretched because of malalignment of the joint. The capsule and ligaments contain high concentrations of pain receptors, which cause shoulder pain (5).

Factors contributing to joint malalignment include rotator cuff weakness, loose ligaments and capsule, and impingement between the humeral head and shoulder suture. GHS commonly occurs in the flaccid stage, which is characterized by areflexia and atonia (6). However, the relationship between HSP and subluxation is unclear. Magnetic resonance imaging (MRI), which is widely used in diagnosing and determining the pathologies of HSP, has been proven to be more advantageous than other imaging techniques because it can clearly show the details of the soft tissue (7).

Thus, the present investigation aimed to determine the differences between MRI findings in HSP patients with subluxation and non-subluxation after stroke, and to analyze the etiology of shoulder pain.

In this study, MRI of the affected shoulder was performed in patients with or without subluxation. Shoulder subluxation occurs mostly during the first 3 weeks in patients with HSP after a stroke (12). The incidence rate was reported to range from 32 to 81% (13), and the rate found in our study (51%) falls within this range.

The MRI scans demonstrated rotator cuff injury in 94% of the shoulders of patients in our study. However, in a study conducted by Dogun et al., 63.2% of patients (n = 68) with HSP were found to have a rotator cuff injury (7). The difference between their study and ours may be attributed to the different stroke onset durations: Their mean duration was 49 days, whereas ours was 85. Moreover, we found that the rotator cuff injury rate was high regardless of subluxation. In detail, supraspinatus injury rates were 92.6% in the GHS group and 76.9% in the nGHS group. Moreover, about half of patients had subscapularis injury. Contrastingly, the incidence of infraspinatus and teres minor injury was low. The main function of the supraspinatus is to abduct the shoulder, and it is an important posterior stabilizing structure of the shoulder. According to several studies (14, 15) supraspinatus injury is mainly caused by subacromial impingement. When the shoulder is frequently abducted and lifted upward, the supraspinatus tendon becomes easily impacted by the coracoacromial arch, resulting in edema, hyperemia, degeneration, and even tearing. Zhu et al. showed that when the shoulder was abducted 60°, it bore the greatest stress; additionally, it was beginning to rub between the supraspinatus tendon and the acromion and was thus most prone to injury and pain (16).

The functions of the subscapularis are to internally rotate the shoulder and dynamically stabilize the humeral head. Moreover, the subscapularis is an antagonist to the superior pull of the deltoid and it assists in abduction and adduction of the shoulder (17). External rotation and abduction can cause excessive strain injury of the subscapularis, which results in pain in the front of the shoulder (18). In addition, a frequent overhead throwing motion can cause coracoid impingement, which is closely associated with subscapularis injury. Therefore, rotator cuff injury, especially supraspinatus and subscapularis injury, is one of the causes of HSP. It is important for rehabilitation teams to not schedule exercises that involve moving the upper limbs above 60° with abduction action in patients with HSP. Furthermore, if the patient has a possible subscapular injury, external rotation of the shoulder should be avoided as much as possible.

Supraspinatus injury was recently hypothesized to result in compensation via greater force generation through the subscapularis, which could potentially hasten degeneration of the subscapularis (19). Accordingly, we could infer that there may be a correlation between supraspinatus and subscapularis injuries in patients with HSP, but this needs to be confirmed by a study with a large sample size.

Our MRI scans also demonstrated subacromial-subdeltoid bursa effusion in 71.74% of shoulders in patients with HSP, which is consistent with previous reports (53–80.9%) (7, 20, 21). Besides, we found a high subcoracoid bursa effusion and subscapular bursa-cavity effusion rate, even though this was not statistically significant between the groups. It has been suggested that subcoracoid bursa effusion and subscapular bursa-cavity effusion, except subacromial-subdeltoid bursa effusion, may cause HSP.

In a study of 42 fresh cadaveric shoulders, it was found that the function of the subscapularis and subcoracoid bursae is to manage friction of the superficial fibers against the scapular neck, humeral head, and coracoid process (22). Thus, it remains to be explored whether friction of these bone structures is increased after biomechanical changes in hemiplegic patients, resulting in fluid accumulation in the bursa.

Thanks to MRI, BME has been detected in the humeral head of numerous patients with HSP: BME was mostly reported in the femoral head and knee (23, 24). In recent studies, instances of BME were observed in the foot, ankle, wrist, or other bones (25, 26). There are two perspectives about the mechanism of BME: (1) secondary BME, secondary to infection, trauma, or arthritis, is caused by an external force acting on the cancellous bone, resulting in a microfracture of the trabecula bone that increases permeability and rupture of the local capillary, adding to the exosmosis of cell fluid and vascular perfusion; and (2) physiological BME is caused by a long-term external force or change of the normal load of the bone, resulting in bone marrow hyperemia and excessive perfusion of the capillaries. In our study, BME occurred more in patients with subluxation which may be due to (1) rotator cuff weakness and change in gravity that caused secondary BME and (2) excessive passive movement or incorrect posture that caused physiological BME in flaccid paralysis period. Because of its self-limiting course (24), BME could improve after the paralysis period. Thus, the potential evolution and mechanism of BME in the shoulder requires further study.

Long bicipital tendon-glenoid labrum injury occurred more commonly in the GHS group. It is caused by repeated contraction of the long head of the bicipital tendon or trauma to the humeral head with repeated external rotation and abduction movements (27). Thus, when rehabilitating patients with HSP, especially those with subluxation, physicians should avoid such movements.

We usually divide the labrum into eight directions: anterior, posterior, superior, inferior, anterior superior, anterior inferior, posterior superior, and posterior inferior. We also briefly differentiated the exact location of the simple labrum injury observed herein. MRI findings showed that simple labrum injury was more common in the GHS group than in the nGHS group, and it included extensive anterior superior, posterior superior, anterior inferior, anterior, and posterior injury. Our finding that extensive injury in the anterior superior was the most common injury type differs from findings of shoulder imaging in cases of sports injury (27). This discrepancy implies that the biological stress produced by passive motion is different from that of active motion.

The limitations of this study include the small sample size, absence of MRI images of contralateral healthy shoulders, and retrospective design. A larger scale, controlled clinical trial with a longer-term, longitudinal follow-up is warranted.

On the basis of our MRI findings, we found a high frequency of bursa effusion, rotator cuff injury, and long head of the biceps tendon injury in patients with HSP. Compared to patients with HSP in the non-subluxation group, we found that patients with HSP in the subluxation group were more prone to BME, long bicipital tendon-glenoid labrum injury, and glenoid labrum injury. Thus, in patients with HSP, we recommend that physicians avoid moving the upper limbs above 60° with abduction action. If patients have subluxation, physicians and therapists should plan to reduce the external rotation and abduction movements to prevent rotator cuff injury and long bicipital tendon-glenoid labrum injury during rehabilitation. Patients with subluxation should be moved gently if they require passive actions. Moreover, they should be educated about protecting their shoulders during ADLs. Further studies with a larger sample size are needed to confirm our findings.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

The studies involving human participants were reviewed and approved by No. S2019 230 01. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

Z-SJ, X-TZ, and L-NZ designed the experiment. H-MX and LX read and checked the MRI images. NW and RW assessed the clinical information of the patients. H-MX wrote the draft of the manuscript. All authors discussed the results, commented on the manuscript, and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.