Outcomes and retear rates after subacromial and glenohumeral injection of a chitosan-based thermosensitive anti-adhesive agent during arthroscopic rotator cuff repair: a propensity score-matched analysis

Article information

Clin Shoulder Elb. 2026;29(3):316-326

Publication date (electronic) : 2026 May 13

doi : https://doi.org/10.5397/cise.2025.01382

Hyungsuk Kim, Younsung Jung, Hyun Seok Song

Department of Orthopedic Surgery, Eunpyeong St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul, Korea

Correspondence: Hyun Seok Song Department of Orthopedic Surgery, Eunpyeong St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, 1021 Tongil-ro, Eunpyeong-gu, Seoul 03312, Korea Tel: +82-2-2030-4628, Fax: +82-2-2030-4629, E-mail: hssongmd@hanmail.net

Received 2025 November 29; Revised 2026 March 9; Accepted 2026 March 20.

Abstract

INTRODUCTION

Postoperative stiffness is a significant impediment to recovery following arthroscopic rotator cuff repair (RCR) [1-3]. This condition, resulting from a complex interplay of adhesions in the subacromial space and capsular contracture, directly affects rehabilitation and quality of life [4,5]. Variability in the incidence of postoperative stiffness, which ranges from 4.9% to 32.7%, highlights the influence of demographic, clinical, and surgical factors [6-8]. Several risk factors predispose individuals to joint stiffness after RCR, including diabetes, age ≤50 years [9], the presence of calcific tendinitis, and a limited range of motion (ROM) preoperatively [10]. Concomitant labral repair, repair of a partial articular-sided tendon avulsion, tears smaller than 3 cm, and single cuff tendon repair also affect the occurrence of postoperative stiffness [7,11,12].

Adhesions within the subacromial space that result from surgical procedures such as acromioplasty, bursectomy, decortication for footprint preparation, and cuff repair are believed to play a crucial role in postoperative stiffness [13]. Those procedures have the potential to induce bleeding, which in turn can accelerate the proliferation of fibroblasts and further the formation of fibrous scar tissue adhesions, culminating in the development of fibrous scar tissue [14,15].

Despite advances in arthroscopic techniques intended to minimize postoperative complications, the prevention of adhesion formation has not been systematically integrated into RCR procedures [16]. Currently, no prophylactic anti-adhesive agent or mechanical barrier is routinely incorporated into standard arthroscopic RCR, and postoperative management primarily relies on surgical technique and structured rehabilitation. Consequently, the exploration of anti-adhesive agents that could establish a biophysical barrier to adhesion formation might improve postoperative recovery in many surgical fields.

A variety of agents can be used to establish biophysical barriers against adhesion formation, including sodium hyaluronate (HA) [17], poloxamer [18], sodium alginate [19], oxidized regenerated cellulose [20], sodium carboxymethyl cellulose (CMC) [21], polyethylene glycol [22], Gore-Tex [23], medical chitosan [24], and collagen [25]. For example, the anti-adhesive properties of poloxamer (a biocompatible copolymer) are due to its thermosensitive sol–gel transition, which converts it from a liquid into a gel in response to body temperature, ensuring the maintenance of its anti-adhesive effect at the surgical site [26]. Similarly, sodium alginate, a natural polysaccharide, acts by forming a physical barrier on wound tissue, thereby inhibiting fibrosis at the surgical site during the crucial early phase of wound healing [24,27]. Positive results have been reported after breast cancer surgery and liver resection [19,28]. In two previous studies involving shoulder surgery, the gliding motion between the deltoid and greater tuberosity was better and the recovery of forward flexion was faster after sodium HA/CMC was injected into the subacromial space [29,30].

A thermosensitive combination of poloxamer 407/188, medical chitosan, and gelatin was tested in animal studies for its ability to transition from an injectable liquid at room temperature into a gel upon reaching body temperature, which allowed it to maintain its presence at the surgical site for an extended duration [18,31], ensuring localized action and minimizing inadvertent dispersion. Moreover, in a clinical study, subacromial injection of that agent after RCR lowered retear rates, further supporting its use in human shoulders [32].

After arthroscopic RCR, postoperative adhesions can be observed in the subacromial, subcoracoid, and subdeltoid spaces, accompanied by capsular contraction [29]. Comparative studies have examined histological changes in the glenohumeral joint capsules of patients with frozen shoulder and patients with Dupuytren's contracture of the hand. Researchers have noted a correlation between Dupuytren's contracture and frozen shoulder, with reported association rates reaching as high as 52% [33]. Furthermore, surgical observations often reveal an inflamed synovium marked by neo-angiogenesis and a thickened, stiffened capsule that has been attributed to fibrosis [11]. The subacromial space plays an important role in postoperative stiffness [30]. Arthroscopic RCR can induce bleeding [34], which can in turn accelerate the proliferation of fibroblasts and further the formation of fibrous scar tissue adhesions, culminating in the development of fibrous scar tissue in the subacromial space. Therefore, studies have focused on injecting anti-adhesive agents into the subacromial space [29,30].

In this study, we compared how injecting a thermosensitive medical chitosan-based anti-adhesive agent into two critical spaces, the glenohumeral joint and subacromial space, affects stiffness. Our goal was to determine the role played by injection location and analyze the differential effects on postoperative shoulder function after arthroscopic RCR. The primary outcome was postoperative ROM recovery assessed longitudinally after surgery using forward flexion, external rotation, and internal rotation. Postoperative stiffness was defined using previously published ROM-based thresholds; however, we focus on ROM trajectories and clinical outcomes, rather than a binary stiffness endpoint. We also compared clinical outcomes using visual analog scale (VAS) scores for pain and published scores for shoulder function. We hypothesized that the injection site would influence postoperative ROM recovery and thereby postoperative stiffness, expecting glenohumeral injection to be more effective than subacromial injection, without affecting rotator cuff healing after arthroscopic RCR.

METHODS

The study protocol was approved by Institutional Review Board of Eunpyeong St. Mary’s Hospital (No. PIRB-20230824-027), and the requirement for informed consent was waived due to the retrospective study design.

Propensity Score-Matched Cohort

The 45 patients of the GH group were matched 1:1 with patients in the SA group, based on propensity scores with a caliper of 0.2 times the standard deviation (SD). Propensity scores were calculated via a multivariate logistic regression that focused on patient-related factors that influence rotator cuff healing [35]: age, sex, dominant hand, smoking status, level of sports activity, type of occupation, presence of diabetes mellitus (DM), and tear size. Sports activity was categorized into contact sports, overhead activities with objects, fitness or swimming, and none. Specific overhead activities with objects were sports such as baseball, badminton, and tennis. Occupation was categorized based on demand as high, medium, or low. Positions involving heavy manual labor with significant shoulder use were considered high demand; those with less shoulder engagement were considered medium demand; and sedentary roles or unemployment were considered low demand [36]. Patients were classified as having DM if they had been diagnosed by a physician or were taking oral or injectable antihyperglycemic medications. Propensity score matching resulted in 37 matched patients in each group. After excluding patients without a minimum 1-year postoperative magnetic resonance imaging (MRI) follow-up, 28 patients in the GH group and 28 patients in the SA group were included in the final analysis (Fig. 1).

Fig. 1.

Study flow diagram. RCR: rotator cuff repair, GH group: glenohumeral joint injection of thermosensitive anti-adhesive agent, SA group: subacromial space injection of the agent.

Surgical Procedures

All surgeries were performed under general anesthesia with the patient sitting in the beach chair position. The glenohumeral joint was examined through a standard posterior arthroscopic portal. Subscapularis and labral lesions were managed via debridement or repair, and rotator interval release was performed. Biceps tenotomy or tenodesis was performed based on the patient’s age and daily activities. After examining articular lesions on the supraspinatus, the arthroscope was moved to the subacromial space. Acromioplasty was performed if a large sharp subacromial bony spur was seen following bursectomy. To evaluate the extent of the tear, a shaver was used to remove the frayed, torn end. The footprint of the greater tuberosity was debrided and decorticated using a shaver and burr.

For the medial row, double- or triple-loaded suture anchors (Healicoil Regenesorb; Smith & Nephew) were positioned slightly lateral to the edge of the exposed footprint. The number of medial anchors (one to three) was determined by the size of the tear and tear configuration. After the medial row repair, with the knots tied using the Revo knot technique, two lateral row anchors (BioComposite SwiveLock; Arthrex) were used to secure the tendon over the footprint while ensuring that the sutures were tensioned appropriately. After the repair, as much saline as possible was removed using the suction function of the shaver device. To inject the anti-adhesive agent (Mediclore; CGBio Co.) into the glenohumeral joint, the needle was inserted through the anterior portal while being viewed from the standard posterior portal (Fig. 2A). To inject the agent into the subacromial space, the needle was inserted through the lateral portal while being viewed from the posterolateral portal (Fig. 2B).

Fig. 2.

(A) Injection of a thermosensitive anti-adhesive agent into the glenohumeral joint after arthroscopic rotator cuff repair, viewed from the standard posterior portal. (B) Injection of the agent into the subacromial space after arthroscopic rotator cuff repair, viewed from the posterolateral viewing portal.

RESULTS

Patient Demographics

Table 1 outlines the patient demographics and relevant baseline factors. After excluding patients without 1 year of postoperative follow-up MRI, 28 matched patients per group were included in the final analysis. The baseline demographic and clinical characteristics did not differ significantly between the groups (Table 1). Covariate balance was additionally assessed using SMDs, which are summarized in Table 1, with a Love plot provided in Fig. 3. After matching, covariate balance was improved overall; however, residual imbalance remained in some sparse categorical levels, particularly in the sports activity and tear size categories, likely due to small cell counts. Patients in the GH group had 3 large tears and 7 medium ones. Patients in the SA group had 1 large tear and 12 medium ones. No significant statistical differences were noted in tear size between the groups (P=0.833). The mean follow-up period for all patients was 14.2±4.7 months (range, 12–30 months).

Table 1.

Patient demographics

Fig. 3.

Love plot demonstrating covariate balance in the matched analytic cohort. Absolute standardized mean differences are shown for each baseline covariate, with the dashed vertical line indicating the threshold of 0.1.

Clinical Assessment

Preoperative VAS scores for pain, functional scores, and ROM did not differ significantly between the groups. The VAS scores for pain and all of the functional scores had improved 1 year postoperatively (Fig. 4). At the final visit, the VAS score for pain was significantly lower in the SA group (1.04±0.69 for GH group vs. 0.78±1.21 for SA group, P=0.038), but the difference did not meet the criteria for the MCID. However, none of the functional scores differed significantly between the groups at the final visit (Table 2). Compared with the preoperative ROM, FF, IR, abduction, and ABIR in the GH group and side-ER, abduction, and ABIR in the SA group improved significantly at the last follow-up visit (P<0.05) (Table 3). No differences in ROM were found between the groups 2, 3, or 6 months postoperatively or at the final follow-up. No patient in either group met the predefined criteria for postoperative shoulder stiffness during follow-up.

Fig. 4.

Comparison of visual analog scale (VAS) pain scores between the groups according to the time after rotator cuff repair. GH group: glenohumeral joint injection of thermosensitive anti-adhesive agent, SA group: subacromial space injection of the agent. ^*P<0.05.

Table 2.

Between-group comparison of functional outcomes preoperatively and at the final visit

Table 3.

Pre- and postoperative functional outcomes

DISCUSSION

The key finding of this study is that neither ROM nor the retear rate differed between patients who received a glenohumeral injection and those who received a subacromial injection of a thermosensitive medical chitosan-based anti-adhesive agent, but the SA group reported lower VAS pain scores at the last follow-up. Understanding the underlying biological processes that lead to postoperative adhesions is important. Postoperative adhesions result from the proliferation of fibroblasts that develop into fibrotic scar tissue [15,43]. A potential pathway involves the transportation of fibroblasts, which play a pivotal role in the development of fibrosis, from adjacent tissues to the surgical site via the circulation [44]. Consequently, targeting fibroblast proliferation and hemostasis is vital for mitigating postoperative adhesions. Using biophysical barriers to limit fibroblast adherence at the surgical site has shown efficacy in preventing fibrosis [45].

After arthroscopic RCR, adhesions commonly stem from the disruption of tissue planes during surgery, leading to contractures around the joints, abnormal adhesions across movement interfaces, and reduced mobility of muscle–tendon units. Symptoms can include pain and weakness. Distinct forms of stiffness identified after RCR include intra-articular capsular contracture, other contractures or adhesions that affect the gliding tendons (such as the rotator cuff or biceps), and adhesions at the extra-articular glenohumeral motion interface [11].

The occurrence of shoulder stiffness following RCR is a major concern because it can substantially affect clinical outcomes. Several authors have suggested criteria for shoulder stiffness, such as passive FF less than 100° and ER less than 30° [46-48]. Oh et al. [40] defined shoulder stiffness as passive FF <120°, side-ER <30°, or IR at the back lower than L3. This condition, with a reported incidence ranging from 4.9% to 32.7% of all repairs, can lead to decreased patient satisfaction and hinder functional recovery by limiting motion and causing pain [6,7].

Factors such as the details of the tear, postoperative immobilization, adhesions in the glenohumeral joint, capsular contracture, and patient-specific comorbidities such as diabetes all contribute to postoperative shoulder stiffness [11]. Given the difficulty of treating postoperative stiffness, researchers continue to explore strategies that can reduce the likelihood of stiffness. Trenerry et al. [49] highlighted the extended recovery period required for patients experiencing postoperative stiffness, underscoring the importance of properly managing RCR patients.

Interest has been growing in the use of anti-adhesive agents as a preventive strategy to mitigate postoperative stiffness and improve overall surgical outcomes. HA, a key component of the extracellular matrix, is a commonly used anti-adhesive agent. This non-sulfated glycosaminoglycan plays a crucial role in regulating cell growth and movement, while also mitigating inflammation and fibrosis during the initial stages of wound healing [50-53]. Leveraging those properties, HA has been applied in abdominal [54], spinal [55], and tendon surgery [56] to curtail postoperative adhesions, and the outcomes have been encouraging [27].

Using HA following RCR to prevent postoperative shoulder stiffness has also shown promise. Oh et al. [30] reported that injecting sodium HA/CMC into the subacromial space after arthroscopic RCR promoted faster recovery of FF 2 weeks postoperatively. In that prospective randomized comparison study, FF was superior in patients who received an injection of HA/CMC (146.5°±17.3° vs. 137.7°±24.7° in the control group). No side effects or complications from using HA/CMC were reported. In another prospective cohort study, Jeong et al. [29] reported superior gliding motion in patients who received an HA/CMC injection in the subacromial space after arthroscopic RCR. Although that study found no differences in functional outcomes or ROM, the gliding motion between the deltoid muscle and greater tuberosity, as evaluated by ultrasonography, was better in the HA/CMC group: poor gliding motion was observed in 2.5% of the injected patients and 15% of the control group 8 weeks postoperatively.

Lee et al. [13] compared the types and doses of three anti-adhesive agents injected into the subacromial space after arthroscopic RCR (3 mL poloxamer/sodium alginate–based anti-adhesive agent; 3 mL sodium HA-based anti-adhesive agent; and 1.5 mL sodium HA-based anti-adhesive agent; with no injection as the control). All of the ROM measurements, pain VAS scores, and functional scores were improved irrespective of the specific anti-adhesive agent used or its type or dosage. Furthermore, no clinical differences were found according to the type or dosage of an anti-adhesive agent, nor between the control group and the groups that received an anti-adhesive agent.

We used a thermosensitive poloxamer 407/188 anti-adhesive agent containing medical chitosan and gelatin. This formulation undergoes a sol-gel transition at body temperature, which promotes localized retention and potentially provides a temporary physical barrier during the early postoperative period. Chitosan has been reported to suppress fibroblast growth and collagen fiber formation [57] and to influence hemostasis, potentially reducing fibrin deposition [58]. It has also been reported to modulate profibrotic signaling pathways such as transforming growth factor beta, which has led to interest in using chitosan-based materials as anti-adhesive agents [58]. Gelatin, the other ingredient in Mediclore, strongly adheres to wound sites. It has been suggested that gelatin not only promotes hemostasis but also encourages epithelial and endothelial cell proliferation, while potentially inhibiting the proliferation of fibroblasts [58]. From a rabbit rotator cuff model, Kim et al. [59] reported reduced fibrosis in the subacromial space after the application of a medical chitosan-based agent, along with greater biomechanical strength at 8 weeks. Kim et al. [32] reported a lower retear rate after subacromial injection of this agent following arthroscopic RCR, although functional scores and ROM did not differ between the groups. That report also discussed systemic associations, such as serum lipid profiles, in relation to rotator cuff pathology [1,60,61]. Our results are consistent with those findings, with postoperative improvement in ROM and functional outcomes in both groups and retear rates of 7.1% in the GH group and 10.7% in the SA group. Because we did not directly quantify postoperative adhesions or fibrosis, these mechanistic considerations should be regarded as background rationale. Our conclusions are based on measured clinical outcomes.

To the best of our knowledge, this study was the first to compare the clinical outcomes of injecting an anti-adhesive agent at different locations. The SA group had a lower VAS score than the GH group at final follow-up. However, the difference, which was about 0.3, did not represent the minimal clinically important difference (MCID) in patients with rotator cuff disease, which is reported to be 1.4 [62]. Therefore, the statistically significant difference in pain should be interpreted with caution, and our findings do not support clinically meaningful superiority of one injection site over the other.

This study has some limitations. First, despite the use of propensity score matching, this study was retrospective, which carries inherent limitations. Second, the relatively small sample size might have affected our results. Third, the 1-year follow-up was relatively short. Given that the effects of anti-adhesive agents are predominantly in the early stages after surgery and we observed no discrepancy in any of the functional outcomes, it might be necessary to extend the follow-up to confirm our findings. Fourth, although all patients followed the same rehabilitation protocol, patient compliance with the rehabilitation program was not assessed. Variation in patient compliance might have affected the findings related to the restoration of ROM and functional outcomes. Fifth, balance diagnostics were primarily assessed within the final matched analytic cohort, and the full covariate distribution of the pre-matching population could not be comprehensively evaluated. Sixth, because multiple outcomes and time points were evaluated, statistically significant findings with small effect sizes should be interpreted cautiously.

CONCLUSIONS

Neither ROM nor the retear rate differed between patients treated with a glenohumeral injection and those treated with a subacromial injection of a thermosensitive medical chitosan-based anti-adhesive agent after RCR. Although the SA group had a slightly lower VAS score for pain at the last follow-up, this difference did not represent the MCID. Overall, the injection site did not appear to alter postoperative ROM recovery.

Article information Continued