Clinical outcomes of arthroscopic stabilization procedures in patients with shoulder dislocation combined with posttraumatic adhesive capsulitis

Abstract

Background

This study was conducted to analyze clinical outcomes after arthroscopic stabilization procedures in patients with shoulder instability and posttraumatic adhesive capsulitis.

Methods

Patients who underwent arthroscopic Bankart repair between 2019 and 2022 with limitation of forward flexion <120°, external rotation <30°, and internal rotation below L3 were included. Demographic data, arthroscopic findings, visual analog scale (VAS) score, American Shoulder and Elbow Surgeons (ASES) score, and Rowe score were evaluated.

Results

Nineteen (4.3%) of 442 patients were included in the study with a mean follow-up of 33.7±2.8 months. The mean age was 32.8±1.9 years, with a mean time of 6.6±1.8 months from the last episode of instability. At final follow-up, 17 of 19 patients (89.5%) recovered full ranges of motion. Forward flexion improved from 112.2°±9.4° to 170.8°±2.1° (P<0.001), internal rotation from 2.3±0.4 to 7.2±0.8 (P=0.001), and external rotation from 23.1°±1.3° to 61.6°±2.8° (P=0.019). The mean VAS score decreased from 3.5±0.5 to 0.4±0.2 (P<0.001). The ASES score improved from 71.7±2.4 to 92.7±2.2 (P<0.001), and the Rowe score increased from 71.5±3.2 to 97.4±1.0 (P<0.001). No complications or revision surgeries occurred.

Conclusions

Arthroscopic stabilization provided satisfactory functional recovery and range of motion in patients with anterior instability and posttraumatic adhesive capsulitis.

Level of evidence

Ⅳ.

INTRODUCTION

Instability and stiffness are extremely different conditions among the pathologies of the glenohumeral joint. The pathologic tissues involved are also different. Instability usually occurs due to a labral tear, whereas stiffness is caused by capsular inflammation [1,2]. Based on these pathologies, instability and stiffness tend not to be combined in the same patient. Patients with shoulder dislocation usually preserve their range of motion (ROM). Although a shoulder may demonstrate restricted motion because of pain or apprehension immediately after an episode of instability, most patients have restored ROM through rehabilitation before surgery [3,4]. Rehabilitation after an episode of shoulder dislocation is not demanding, because patients are often young, their compliance is good, and pain is not the main symptom of instability, all of which could be important obstacles to ROM exercise [3].

However, some patients demonstrate persistent limitation of shoulder motion after instability despite appropriate rehabilitation [5,6]. The management of shoulder instability with concomitant stiffness remains a clinical challenge. Restoration of ROM through structured rehabilitation is traditionally prioritized in these patients, as stiffness is regarded as a potential protective or compensatory mechanism against recurrent instability. Therefore, surgical stabilization is often delayed until motion improves. Proper healing of the capsulolabral structures requires a period of immobilization, but early rehabilitation is crucial to prevent stiffness. Furthermore, as capsulolabral repair is a joint stabilizing procedure, it may paradoxically contribute to the progression of stiffness [7].

There have been many studies on rotator cuff tear with concomitant stiffness [8]. Recent studies have reported that preoperative shoulder stiffness does not negatively affect the clinical outcomes of rotator cuff repair when treated with manipulation or arthroscopic capsular release [9]. There is a growing consensus regarding the effectiveness of one-stage arthroscopic repair for rotator cuff tears with concomitant shoulder stiffness [10]. However, there is a lack of studies on shoulder instability with concomitant stiffness.

The purpose of this study was to analyze demographic data and clinical outcomes after arthroscopic stabilization procedures in patients with shoulder instability combined with posttraumatic adhesive capsulitis. Our hypothesis was that in patients with shoulder instability and concomitant stiffness, achieving a stable joint through firm fixation would lead to good clinical outcomes in both joint stability and ROM.

METHODS

This study was carried out in accordance with the ethical standards outlined in the 1964 Declaration of Helsinki and relevant regulations of the U.S. Health Insurance Portability and Accountability Act. This study was approved by the Institutional Review Board of Ewha Womans University Seoul Hospital Board (No. SEUMC 2025-01-018-001). Written informed consent was obtained from all participants.

Patients who underwent arthroscopic Bankart repair for shoulder instability at a single institution between January 2019 and December 2022 were enrolled retrospectively. Surgical indications included patients who complained of instability symptoms in daily activities with a positive apprehension test, and patients who could not recover full ROM due to pain and instability despite rehabilitation. The inclusion criteria were as follows: (1) patients who had histories of anterior shoulder dislocation or subluxation, (2) patients who showed limited shoulder motion despite 3 months of rehabilitation primarily focused on ROM recovery since their last dislocation episode, with limitation defined as any of the following: passive forward flexion <120°, external rotation <30°, and internal rotation below the L3 vertebral level under anesthesia [11], (3) patients who underwent preoperative imaging via magnetic resonance imaging (MRI) or magnetic resonance arthrography, and computed tomography arthrography (CTA) with postoperative follow-up imaging at 1 year, and (4) outpatient follow-up for a minimum of 2 years. Patients with brachial plexus injury, axillary nerve palsy, osteoarthritis, history of previous surgery, multidirectional instability, or who underwent Latarjet procedures or were lost to follow-up were excluded.

All patients underwent conservative treatment before surgery. The preoperative rehabilitation protocol was applied uniformly to all patients, regardless of age, severity of instability, or associated lesions. The patients underwent shoulder immobilization with a shoulder sling for 2 weeks immediately after an instability episode. After 2 weeks, the sling was removed, and active and passive ROM rehabilitation exercises for shoulder flexion, abduction, external rotation, and internal rotation were performed for 2 months. Patient data were collected retrospectively by reviewing electronic medical records and operative reports.

Clinical Evaluations

Patients were surveyed for personal information including age, body mass index, conservative treatment duration, history related to injury, number of instability episodes, presenting symptoms, dominant arm, and participation in sports. Preoperative physical examinations included the active ROM of the shoulder, Beighton score for evaluation of generalized laxity, anterior apprehension test, sulcus test, posterior jerk test, Kim’s test, and O’Brien test. A Beighton score of 4 or higher indicated generalized laxity. Passive ROM was assessed under anesthesia and compared with the contralateral shoulder to exclude limitation of motion caused by pain or fear due to apprehension. Shoulder ROM was determined using a goniometer in relationship to the long axis of the thorax. Motion was recorded for forward flexion, external rotation with the arm at the side, and internal rotation at the back in a seated position using a goniometer. The examiner guided the patient’s hand upward along the alignment of the spinous processes and documented the highest spinous process that could be reached with the tip of the thumb. The values were converted into consecutive numerical categories—the sacrum was labeled as 0, L5 to L1 were assigned numbers from 1 to 5, and T12 to T5 from 6 to 13.

For functional improvement, patients were assessed preoperatively and postoperatively at 3, 6, and 12 months, and at the last follow-up visit using the visual analog scale (VAS) for pain, Rowe score, and American Shoulder and Elbow Surgeons (ASES) score. Active ROM was also evaluated at 3, 6, and 12 months, and at the last follow-up visit postoperatively. All functional assessments were evaluated by a physician assistant with 9 years of experience in orthopedic surgery who was not involved in this study.

Radiologic Evaluations

Preoperative CT scans were used to evaluate the size of the glenoid defect and Hill-Sachs lesion. The glenoid bone defect was assessed as a percentage using the en face view of three-dimensional CT, following the method described by Sugaya [12]. Hill-Sachs lesion size was quantified as a percentage by measuring the width and depth of the humeral head defect relative to the humeral head diameter on axial and coronal views, and the location was measured as a Hill-Sachs angle according to the method described by Cho et al. [13]. The on-track and off-track lesions were evaluated using the method described by Itoi [14]. Postoperative CTA was performed at the 1-year follow-up to evaluate labral healing [15].

Capsule elasticity and inflammation were assessed during arthroscopic surgery with neutral arm position. Capsular elasticity was assessed by gently probing the middle and inferior capsule intraoperatively. The severity of inflammation was classified into mild and severe based on the coverage of inflammatory tissue on the capsule. Mild inflammation was defined as partial coverage of inflammatory tissues (Fig. 1). Severe inflammation was defined as full coverage of the capsule of inflammatory tissues (Fig. 2). Two orthopedic surgeons independently evaluated the images to determine interobserver reliability. After the initial evaluation, one orthopedic surgeon relabeled patients with blinded processing and shuffled the order for a second classification for intraobserver reliability.

Fig. 1.

Arthroscopic findings of the right shoulder from the anterosuperior viewing portal. (A) Mild capsular inflammation was observed with a Bankart lesion. (B) Removal of capsular inflammation and Bankart repair were performed.

Fig. 2.

Arthroscopic findings of the right shoulder from the anterosuperior viewing portal. (A) Severe capsular inflammation was observed with a Bankart lesion and chondral defect. (B) Removal of capsular inflammation and Bankart repair with coverage of chondral defect were performed.

Surgical Technique

All surgeries were performed by a senior orthopedic surgeon. Arthroscopic procedures were performed with the patient in the lateral decubitus position under interscalene block and general anesthesia. All suture anchors were inserted into the glenoid, and a curved suture hook was used through the posterior portal to pass through the labrum and adjacent capsule at a position inferior to the anchor insertion site. Simultaneous labral repairs were performed using sliding locking knots. After completing the sutures, an intraoperative physical examination was performed to assess engagement of the Hill-Sachs lesion. Dynamic assessment of Hill-Sachs engagement was performed under anesthesia using controlled posterior translation and external rotation within the available ROM, and engagement was defined as mechanical contact between the Hill-Sachs lesion and the glenoid rim. If engagement was observed, the Remplissage procedure was performed. In cases with severe capsular inflammation with loss of elasticity, rotator interval closure was performed using two strands of sutures to augment anterior stability and reduce residual capsular redundancy. As procedures for adhesive capsulitis such as manipulation under anesthesia (MUA) or capsular release were not performed, surgery was exclusively focused on shoulder stability.

Rehabilitation

All patients followed a uniform postoperative rehabilitation protocol. Immobilization was maintained for 4 weeks with an abduction brace. Passive flexion ROM exercises were initiated 4 weeks after surgery after the weaning of the brace. Internal and external rotations were initiated at 6 weeks postoperatively. Shoulder muscle strengthening exercises were started at 12 weeks after surgery. Patients were allowed to return to sports activities 6 months after surgery. All patients received training to perform rehabilitation exercises for self-administration that was rreinforced through face-to-face instruction during outpatient visits, with progress systematically monitored during outpatient follow-up visits.

Statistical Analysis

All statistical analyses were conducted using IBM SPSS version 26.0.0 (IBM Corp.). The chi-square test, Fisher’s exact test, Student t-test, and Mann-Whitney U-test were used for analyses. A paired t-test was used to compare differences between the preoperative and final follow-up VAS scores for pain, Rowe scores, and ASES scores, with the significance level set at P<0.05.

RESULTS

A total of 442 patients who underwent arthroscopic Bankart repair were enrolled. Among them, 32 patients met the inclusion criteria. Two patients with neurological symptoms, two with osteoarthritis, seven with a history of previous surgery, and two lost to follow-up were excluded. Finally, 19 patients (4.3%) were included with a mean follow-up duration of 33.7±2.8 months. The mean number of instability episodes was 5.7±1.3. There were no patients with underlying medical diseases. Demographic data for included patients are summarized in Table 1.

Table 1.

Demographic data of patients with shoulder instability concomitant with posttraumatic adhesive capsulitis

Variable	Patients (n=19)
Sex (male:female)	14:5
Mean age (yr)	32.8±1.9 (17–47)
Time from last instability to surgery (mo)	6.6±1.8 (3–10)
Instability history
First-time dislocation	7 (36.9)
Recurrent dislocation	9 (47.4)
Recurrent subluxation	3 (15.8)
Mechanism of last instability episode
High-energy trauma	13 (68.4)
Low-energy trauma	6 (31.6)
Recreational sports participants	16 (84.2)

Values are presented as mean±standard deviation (range) or number (%).

On radiologic evaluation, the average glenoid defect was 14.4%±1.9%. Regarding Hill-Sachs lesions, the axial cut revealed a mean width of 35.2%±1.9% and a mean depth of 14.1%±1.5%. The coronal cut showed a mean width of 28.0%±1.9% and a mean depth of 11.0%±1.3%. The mean Hill-Sachs angle was 38.5°±2.0°. An off-track lesion was found in one patient. All patients showed loss of elasticity of the capsule. Mild capsular inflammation was found in 8 patients, and severe inflammation in 11 patients. The mean number of anchors used was 5.0±0.4. Additional rotator interval closure was performed in 11 patients with severe capsular inflammation. One patient underwent the Remplissage procedure for an engaging Hill-Sachs lesion as determined by intraoperative dynamic assessment.

Regarding shoulder ROM, 17 patients (89.5%) demonstrated ROM restoration at final follow-up (Table 2). Forward flexion (P<0.001) and internal rotation (P=0.003) showed statistically significant improvement from 3 months postoperatively. External rotation initially decreased at 3 months postoperatively (P=0.232), but gradually recovered to contralateral side ranges at final follow-up (P=0.019).

Table 2.

Serial changes of postoperative shoulder range of motion in patients with shoulder instability concomitant with posttraumatic adhesive capsulitis

	Preoperative	Postoperative 3 mo	Postoperative 6 mo	Last follow-up
Forward flexion (°)	112.2±9.4	140.6±6.4 (P=0.007)	164.4±3.4 (P<0.001)	170.8±2.1 (P<0.001)
External rotation (°)	23.1±1.3	22.6±2.9 (P=0.232)	40.3±3.4 (P=0.007)	61.6±2.8 (P=0.019)
Internal rotation	2.3±0.4	3.9±0.4 (P=0.003)	6.8±0.8 (P=0.001)	7.2±0.8 (P=0.001)

Values are presented as mean±standard deviation. Internal rotation was recorded at the highest spinous process reached by the patient’s thumb, with values assigned sequentially from sacrum.

A statistically significant improvement in shoulder function was found postoperatively. The VAS score for pain significantly decreased, from 3.5±0.5 preoperatively to 0.4±0.2 postoperatively (P<0.001). The ASES score improved from 71.7±2.4 preoperatively to 92.7±2.2 postoperatively (P<0.001). The Rowe score increased from 71.5±3.2 preoperatively to 97.4±1.0 postoperatively (P<0.001). At 1-year follow-up CTA, one patient (5.3%) showed a labral retear. The rest showed healed labrum without capsular tear. One patient showed a positive apprehension test postoperatively. No patient underwent revision surgeries. No complications related to surgeries including neurologic or vascular complications were observed.

DISCUSSION

The most important finding in this study was that satisfactory clinical outcomes and restoration of ROM were obtained after arthroscopic stabilization procedures in patients who had shoulder instability with concomitant posttraumatic adhesive capsulitis without additional procedures for stiffness. Simultaneous glenohumeral instability and stiffness is a rare condition. In this study, 4.3% of patients who had arthroscopic stabilization procedures showed limitation of shoulder motion compared with contralateral side, despite preoperative rehabilitation. Although the number of patients included in this study was small and we were unable to conduct direct comparisons with patients who had instability with normal shoulder motion, patients with instability and stiffness showed different clinical characteristics. The mean age of patients included in this study was older than their 20s, which is the age range in which the incidence of shoulder dislocation typically peaks [16,17].

Gordins et al. [18] reported that Bankart lesions are generally associated with low-energy trauma in 91.8% of patients. However, in patients with instability with stiffness, the proportion of patients who experienced high-energy trauma leading to their last instability episode was relatively high at 68.4%. Limitations of motion could develop more frequently after shoulder dislocation in older patients suffering from high-energy trauma. All patients had mild to moderate Hill-Sachs lesions, but the size and off-track status of the Hill-Sachs lesions were not correlated with posttraumatic stiffness. When compared to the measurement of Hill-Sachs lesion severity on CT, our findings were closer to those of non-engaging rather than engaging Hill-Sachs lesions [13].

Pathological findings of adhesive capsulitis tissue have revealed persistent synovial inflammatory changes with active fibroblast proliferation on the capsule [19]. Our arthroscopic findings revealed that all patients had loss of capsular elasticity, and 57.9% had severe inflammation in the capsule. Nagata et al. [5] suggested that posttraumatic adhesive capsulitis associated with anterior instability differs pathologically from primary adhesive capsulitis with respect to rotator cuff involvement or soft tissue contracture, which may be linked to poor prognosis. However, no rotator cuff tendon or other soft tissue involvement was found in this study.

Severe capsular inflammation and contracture were related to lack of rehabilitation. Failure of appropriate rehabilitation often results from pain and apprehension following the last dislocation from high-energy trauma. In this study, the VAS score for pain was higher, and all patients showed positive apprehension tests, indicating persistent symptomatic instability rather than isolated motion restriction. It is possible that in patients with instability after significant trauma, inappropriate rehabilitation due to pain and apprehension developed since the last episode contributes to the development of adhesive capsulitis [4,6,20].

There remains controversy regarding whether preoperative ROM exercises should be prioritized or when surgical intervention should be performed. Nonoperative treatment for stiffness requires significant time, and the results can be unpredictable and dependent on patient compliance [21]. In addition, inadequate rehabilitation in the presence of persistent instability may predispose patients to recurrent instability episodes, which could contribute to additional capsulolabral injury, glenoid bone defects, and prolonged recovery with stiffness. Shoulder dislocation with concomitant stiffness has a pathology similar to that of anterior cruciate ligament (ACL) injury, as both conditions involve the coexistence of instability and stiffness of the joint. Previous studies suggested that delaying reconstructive surgery for at least 3 weeks from the time of an acute ACL injury will result in a significantly decreased incidence of arthrofibrosis [22]. However, recent studies indicate that the optimal timing of ACL reconstruction does not affect postoperative stiffness when modern surgical techniques and accelerated rehabilitation protocols are applied [23]. Delaying surgery for prolonged preoperative rehabilitation may lead to additional capsulolabral injury, persistent instability, and decreased patient satisfaction. Although few studies have reported optimal surgical timing in patients with shoulder instability combined with stiffness, our findings suggest that satisfactory outcomes can be achieved without delaying surgery due to the presence of stiffness.

MUA for posttraumatic adhesive capsulitis after upper limb fractures has been reported to have favorable clinical outcomes and ROM restoration [24]. Levy et al. [25] reported significant restoration of ROM and pain relief immediately after arthroscopic capsular release. However, some surgeons prefer arthroscopic capsular release to MUA, due to the risk of complications including humeral fractures, glenohumeral dislocation, and osteochondral defect [2,26]. Owing to the paradoxical characteristics of shoulder instability and adhesive capsulitis, we did not consider MUA or capsular release before arthroscopic stabilization procedures. Instead, as the pathology of adhesive capsulitis is known to be related to inflammation and formation of extensive scar tissue, we focused on achieving a stable joint through firm fixation and emphasized on intensive postoperative rehabilitation. Our patients, including 11 who underwent rotator interval closure, showed recovery in forward flexion, external rotation and internal rotation, with significant improvements in functional outcomes at last follow-up. However, ROM recovery, especially forward flexion, is slower compared with improvements in pain and functional outcomes. External rotation ranges even decreased up to 3 months postoperatively. In our study, only one patient showed labral retear on 1-year follow-up CTA, which is a low incidence considering the extent of labral injury, including pan-labral tears [27]. These findings suggest that in cases of concomitant posttraumatic adhesive capsulitis, satisfactory outcomes can be achieved by focusing on labral firm fixation without performing MUA or capsular release.

Limitations

First, the definition of adhesive capsulitis remains unclear, and there is no established consensus regarding the extent of ROM deficit required for the diagnosis [28,29]. Although multiple studies have proposed different criteria, significant subjectivity persists, highlighting the need for further research to standardize the definition and diagnostic parameters. Second, the number of patients enrolled was small. Posttraumatic adhesive capsulitis following shoulder dislocation is an uncommon condition, making it difficult to accumulate a large dataset. Therefore, this study was conducted as a case series without a control group, which may limit the reliability and generalizability of the findings. A third limitation is variability in patient compliance with postoperative rehabilitation. Differences in adherence to prescribed rehabilitation protocols may introduce bias in ROM recovery, affecting the interpretation of clinical outcomes. The lack of an objective measure to monitor rehabilitation compliance adds another layer of uncertainty. Additionally, we assessed capsular inflammation and elasticity intraoperatively without using standardized quantitative MRI measurements. Although MRI findings such as capsular thickening have been described in the literature, there is no widely accepted grading system for objectively quantifying posttraumatic capsular stiffness. Therefore, some degree of subjectivity in assessment was unavoidable. Finally, the follow-up duration in this study may not be sufficient to capture long-term outcomes. The possibility of late-onset stiffness, recurrent instability, or degenerative changes beyond the study period remains unknown.

CONCLUSIONS

Arthroscopic stabilization provided satisfactory shoulder functional outcomes and improvement in ROM in patients with anterior instability combined with posttraumatic adhesive capsulitis. Favorable clinical outcomes were achieved without concomitant surgical procedures for stiffness.