2.1 Patient cohort

This retrospective study was conducted in accordance with the Declaration of Helsinki and received approval from the ethics committees of the participating institutions (TOYOTA Memorial Hospital, Toyota, Japan and Asahi University Hospital, Gifu, Japan). Informed consent was obtained from all patients.

Between January 2018 and December 2024, a total of 42 shoulders in 40 patients underwent the ASBB procedure for traumatic anterior shoulder instability. All procedures were performed by a single surgeon (T.S.) at two institutions. The decision to perform surgery was based on the following criteria⁶: Competitive collision sports athletes who wished to continue active participation in sports,⁷ Presence of a severe glenoid bone defect (>20 % of the glenoid surface as measured on preoperative computed tomography (CT) images⁵),⁶ previous failure of anterior stability repair, and⁷ patient request. Of the initial cohort, 38 shoulders in 36 patients met the study inclusion criteria, having undergone shoulder CT preoperatively, immediately postoperatively, and at 3 and 6 months postoperatively.

2.2 Surgical treatment and perioperative management

All procedures were performed arthroscopically with the patient in the beach chair position. Five portals were utilized: posterior, anterior, anterolateral, anterosuperior (coracoid portal), and inferomedial (pectoralis major portal; PM portal). The PM portal was positioned 3 to 4 finger widths inferior to the coracoid. The torn capsule-labral complex was released and mobilized from the anteroinferior glenoid neck, which was then abraded using a burr to create a bleeding bony surface. A guide pin was inserted through the subscapularis muscle into the anterior glenoid neck via the PM portal. The pin was positioned anteriorly below the glenoid equator, 3 mm medial to the glenoid surface, and angled slightly downward relative to the glenoid equator. A screw hole was then predrilled from anterior to posterior at the glenoid neck. The coracoid process was osteotomized approximately 1 cm from its tip using the coracoid portal, along with the conjoined tendon. The bone graft was temporarily extracted through the coracoid portal, and a 4.0-mm half-threaded cannulated cancellous screw with a washer (Meira) was inserted centrally in the graft to secure it in a standing position on the anterior glenoid neck. Using the guide pin, the coracoid graft with the screw was then reintroduced into the body and guided to glenoid neck through the PM portal. Once the coracoid was fixed, an arthroscopic Bankart repair was performed using three to five suture anchors (SutureFix 2.7; Smith & Nephew, ICONIX 1.4; Stryker), placed from the 5:30 to 2 o'clock position on the right shoulder. Finally, the glenohumeral joint was arthroscopically inspected, and the location of the transferred coracoid graft was confirmed using anteroposterior and axial plain radiographs.

Postoperatively, the shoulder was immobilized in internal rotation using a shoulder brace for 4 weeks. After brace removal, patients began range-of-motion exercises with a physiotherapist. Weight training commenced at 12 weeks, followed by a strengthening program. Return to sports activities was permitted between 3 and 6 months postoperatively. CT evaluation was performed at 3 and 6 months postoperatively to assess bony healing of the coracoid graft.

2.3 Data collection

The following patient data were collected from medical records: age, sex, injured side, dominant side, height, body mass index (BMI), previous ABR, smoking status (nonsmoker or current smoker), and preoperative Rowe score. Additionally, intra-articular lesions identified arthroscopically, such as superior labrum anterior and posterior (SLAP) lesions, capsular tears, and humeral avulsions of the glenohumeral ligament (HAGL), were recorded.

All patients underwent preoperative and postoperative CT examinations with the Aquilion scanner (Canon Medical Systems Corporation). Images were acquired at a slice thickness of 1.0 mm and reconstructed into three-dimensional models of the glenoid. Osseous glenoid defect size was determined using three-dimensional CT images.⁵ Postoperative CT examinations were obtained on the first day after surgery for all patients, and the following parameters were assessed:

Transferred coracoid graft length: Measured from its tip to the glenoid neck.

Graft depth: Evaluated to determine whether the graft reached the contralateral bone cortex or not.

Transferred Coracoid Graft Positioning:

Horizontal position: Measured on axial CT images as the distance between the transferred coracoid graft and the glenoid surface (mm) (Fig. 1(a))

Close

Fig. 1 The transferred coracoid graft positioning and the fixation screw angulation.

Export to PPT

Vertical position: Measured on three-dimensional CT images as the angle between the glenoid equator and the center of the coracoid graft (°) (Fig. 1(b))

Fixation screw angulation:

Axial angle: Measured on axial CT images as the angle between the screw insertion direction and the glenoid surface (°) (Fig. 1(c))

Sagittal angle: Measured on three-dimensional CT images as the angle between the screw insertion direction and the glenoid equator (°) (Fig. 1(d))

Bone graft union was evaluated at 3 and 6 months postoperatively using sagittal and axial CT images. Bone union was defined as the continuity of trabecular bone between the transferred coracoid graft and the glenoid.⁸ At the 3-month follow-up, patients were categorized into two groups: Group 3m-C: bone union obtained, Group 3m-N: bone union not achieved, similarly, at the 6-month follow-up: Group 6m-C: bone union obtained and Group 6m-N: bone union not achieved.

The association between bone union of the transferred coracoid graft and patient demographics, as well as immediate postoperative CT measurements, was analyzed at 3 and 6 months postoperatively. Additionally, complications—including hardware-related issues (loosening, migration, or breakage), reoperation, and infections—were recorded (Group 3m-F).

2.4 Statistical analysis

All statistical analyses were performed using the EZR software program (version 1.40, Jichi Medical University, Tochigi, Japan) with the significance level set at p ≤ 0.05.⁹ The normality of the data distribution was checked for equality of variance. The Mann-Whitney U test was used to compare age, height, body mass index, preoperative Rowe score, preoperative osseous glenoid defect, the length of the coracoid graft, the position of the transferred coracoid graft (Horizontal position, Vertical position), and the screw angulation (Axial angle, Sagittal angle). Fisher's exact test was used to detect differences within the following categorical data; sex, dominant side, history of previous ABR, current smoker, arthroscopic findings, and the screw depth. Receiver operating characteristic (ROC) analysis was performed to assess the predictive ability of the position of the transferred coracoid graft and the screw angulation by calculating the area under curve (AUC) and to determine the cutoff value by the maximum Youden index.

3.1 Patient characteristics

Of the 42shoulders (40 patients), 38 shoulders (36 patients) with a median age of 19 years at the time of surgery met the study criteria. Table 1 presents the characteristics and preoperative findings of the patients.

Among the cohort, 7 shoulders (18.4 %) underwent revision procedures following failed ABR. Additionally, 30 patients participated in competitive sports, including 22 in rugby, 2 in sumo wrestling, 2 in baseball, 1 in soccer, 1 in handball, 1 in basketball, and 1 in judo. None of the patients were smokers.

3.2 Immediately postoperative CT measurements

Evaluations based on immediate postoperative CT findings are presented in Table 2. In all shoulders, the transferred coracoid graft was positioned within 5 mm of the glenoid surface in the axial plane and placed below the glenoid equator on 3DCT. Additionally, the fixing screw was consistently inserted in a medial direction relative to the glenoid surface in the axial plane and in a proximal direction relative to the glenoid equator on 3DCT. Furthermore, in all cases, the transferred coracoid graft was in direct contact with the glenoid neck.

Shoulders were categorized based on whether bone union of the transferred coracoid graft was obtained at 3 and 6 months postoperatively (Fig. 2). As illustrated in Fig. 2, a detailed flowchart presents the postoperative outcomes of the 38 shoulders. The bone union rate was 63.2 % at 3 months postoperatively and increased to 81.6 % at 6 months postoperatively.

Close

Fig. 2 Flowchart of the participants included in the study.

Export to PPT

This flowchart outlines the postoperative outcomes of 34 shoulders who underwent the modified arthroscopic Bankart-Bristow procedure between January 2018 and December 2024. The chart tracks whether the transferred coracoid graft achieved bone union or not.

3.3 The postoperative status of transferred coracoid graft

We compared the postoperative outcomes between Group 3m-C (n = 24) and Group 3m-N (n = 11) (Table 3a) as well as between Group 6m-C (n = 31) and Group 6m-N (n = 4) (Table 3b). No significant differences were observed between these groups in terms of patient characteristics, including age, sex, dominant side, height, body mass index (BMI), history of previous ABR, or osseous glenoid defect size.

However, sagittal screw angulation relative to the glenoid equator was significantly smaller in Group 3m-C and Group 6m-C compared to Group 3m-N and Group 6m-N (both p < 0.001). Receiver operating characteristic (ROC) curve analysis identified that a sagittal screw angulation of less than 16° at 3 months postoperatively (sensitivity: 91 %, specificity: 88 %) and less than 18° at 6 months postoperatively (sensitivity: 100 %, specificity: 90 %) was predictive factor of bone union.

Conversely, no significant differences were found in other parameters, including coracoid graft length, screw depth, horizontal and vertical positions of the transferred coracoid graft, or axial screw angulation.

3.4 Postoperative complication

There were hardware-related issues, the transferred coracoid graft was broken in 2 shoulders and migrated in 1 shoulder at 3 months postoperatively (Group 3m-F). None of the patients experienced reoperation and postoperative infections.

4.1 Clinical implications of sagittal screw angulation

A reduced sagittal screw angulation likely promotes optimal contact between the coracoid graft and the glenoid neck, minimizing micromotion and facilitating biological integration. The screw trajectory may also influence the force distribution across the graft-bone interface. The conjoint tendon attaches to the apex of the coracoid process.¹⁰ As the conjoined tendon applies anterior and inferior traction to the graft, a steeper sagittal angle may predispose the fixation to shear stress and micro-instability, impairing bone healing (Fig. 3). Conversely, screws inserted with minimal sagittal angulation may resist such forces more effectively, enhancing primary stability.^13,14

Close

Fig. 3 illustrates the effect of different sagittal screw angulations on the fixation stability of the transferred coracoid graft. (a) shows a screw inserted with minimal sagittal angulation relative to the glenoid surface, while (b) depicts a screw inserted with excessive sagittal angulation. The directional force vectors indicate how screw angulation might influence graft stability and compression at the graft-bone interface, through the sling effect exerted by the conjoined tendon. Consequently, the over-angulated screw (b) is subjected to greater forward tension compared to the minimally angulated screw (a), which might impact bone union.

Export to PPT

Our findings are supported by ROC curve analysis, which demonstrated excellent sensitivity and specificity at the identified cutoff values. This provides a clinically applicable guideline for surgeons to optimize fixation strategy during ASBB procedures.

4.2 Comparison with previous literature

Previous studies on coracoid transfer procedures have primarily assessed bony healing of the transferred coracoid graft at least one year postoperatively.^5,7,8,15,16 To our knowledge, few studies have examined the association between the position of the transferred coracoid graft and bone union. Regarding postoperative recurrence, the ideal graft position has been reported as below the glenoid equator in the vertical plane and flush with the glenoid rim in the axial plane.^17,18 These findings are widely reflected in surgical practice, as was the case in this study, where no significant differences were observed in coracoid graft position concerning postoperative bone union.

Previous studies have reported that an excessively angulated axial screw (>25° relative to the glenoid surface) is associated with poor bone graft fixation and healing at least six months postoperatively.⁷ In contrast, a cadaveric study found that screw divergence did not compromise biomechanical stability until it exceeded 30°.¹⁹ In this study, no significant differences were observed in axial screw angulation, making its optimal angle for graft fixation still debatable.

4.3 Limitations and future directions

The retrospective design and limited sample size of this study constrain the generalizability of our findings. Although CT-based evaluation offers objective metrics, measurement error or interobserver variability may affect screw angle analysis. Future prospective, multicenter studies with standardized imaging protocols are warranted to confirm these thresholds and further evaluate the long-term clinical outcomes.

Additional biomechanical studies would also be valuable to quantify the forces at the graft-bone interface across varying screw angles, particularly under dynamic loading conditions relevant to return-to-play protocols.