The most prevalent dislocated articulation is the shoulder joint [1,2]. Most shoulder dislocations are anterior and traumatic [1,3]. Traumatic anterior shoulder instability (ASI) is associated with capsulolabroligamentous tears, glenoid rim fractures, and posterosuperior humeral head compression fractures [4]. Owens et al. [5] reported an association between the coracohumeral distance with ASI, while Lopez et al. [6] identified a link between coracoacromial arch with ASI. Many studies have investigated the relationship between glenoid version (GV) and ASI [1,3,7-9]. Aygün et al. [7] and Hohman et al. [8] found a significant increase in anteversion of the glenoid in patients with ASI. In contrast, Moroder et al. [3] and Peltz et al. [9] found no significant difference in the GV between patients with ASI and control group patients [1]. Hohmann et al. [8] showed that superior glenoid inclination (GI) was low in patients with ASI. Jacxsens et al. [1] found no significant relationship between recurrent shoulder instability and GI and GV in their comprehensive study of cadavers and three-dimensional image models. In the same study, the authors measured the coracoid pillar angle and found that these three angles were significantly higher in the ASI group. However, their measurements did not address the position of the coracoid apex relative to the glenoid center. Yaka et al. [10] defined sagittal central glenoid-coracoid angle (scGCA) and axial central glenoid-coracoid angle (acGCA) parameters that assess the position of the apex of the coracoid in relation to the center of the glenoid and showed that a coracoid process positioned inferior and lateral was associated with subscapularis tear. The relationship between coracoid and glenoid morphology and ASI remains controversial, and the position of the coracoid process apex relative to the glenoid and its association with ASI remains unclear.

This study investigated the relationship between the relative position of the coracoid apex to the glenoid and ASI. We hypothesized that the coracoid apex may be more superior and medial in ASI patients.

Ethical approval was obtained from the Necmettin Erbakan University Ethical Committee University Ethical Committee (No. 2022/3570). Informed consent was obtained from all patients for the use of their radiological images for scientific purposes, in accordance with the decision of the university ethics committee.

A total of 475 patients underwent shoulder arthroscopy in the Department of Orthopedics and Traumatology between 2017 and 2019. Inclusion criteria were patients who were radiographically proven to have anterior shoulder dislocation and underwent labrum repair with arthroscopic surgical treatment. Exclusion criteria were glenoid bone defects, bony Bankart lesions or other concomitant fractures, neuromuscular disease, epilepsy, previous upper extremity surgery and a history of more than one dislocation. A total of 72 ASI patients were included in this study. The control group included 72 patients who were age and sex matched with ASI patients, had no history of instability or subluxation, whose imaging and physical examinations did not suggest any specific shoulder pathology such as labral tear or rotator cuff disease, and who presented with shoulder pain.

Shoulder magnetic resonance (MR) images of the patients performed following standard procedures were retrospectively analyzed. In the sagittal plane, the sagittal central glenoid-coracoid angle, which evaluates the superior and anterior extension of the coracoid apex relative to the glenoid center, was measured as described by Yaka et al. (Fig. 1) [10]. scGCA measurement was performed on coronal and sagittal magnetic resonance imaging (MRI) slices that were referenced to each other. The apex of the coracoid process was marked on the coronal slice and the projection of the apex of the coracoid process was determined on the sagittal slice. The projection of the apex of the coracoid process on the sagittal slice was fixed on the screen and the image was advanced to the glenoid articular surface. At this stage, the glenoid center and the sagittal axis of the glenoid were determined. The angle between the line between the glenoid center and the projection of the apex of the coracoid process and the glenoid axis was determined as scGCA.

Similarly, in the axial plane, the axial central glenoid-coracoid angle, which evaluates the anterior and lateral extension of the coracoid apex relative to the glenoid center, was measured as described by Yaka et al. (Fig. 2) [10]. acGCA measurement was performed on axial and coronal MRI slices that were referenced to each other. The coracoid process apex determined on the axial slice was confirmed on the coronal slice. The coracoid process apex on the axial slice was advanced to the section where the glenoid center was located. Advancement to the level of the glenoid center was confirmed with a referred sagittal section. At the level of the glenoid center on the axial section, the angle between the line between the glenoid center and the coracoid process apex and the glenoid articular surface line was determined as acGCA.

GV was measured using a line joining the anterior and posterior glenoid margins and intersecting a line drawn medially from the posterior glenoid neck to the scapular body on the axial image, as defined by Tétreault et al. [11]. GV was calculated by subtracting 90° from the angle formed by the glenoid surface and the scapular body. Retroversion was defined as a positive angle, and anteversion was defined as a negative angle (Fig. 3) [11]. GI was measured using a line intersecting the line of the scapular body showing the deepest point of the supraspinatus fossa on the coronal oblique image and joining the upper and lower glenoid margins, as defined by Maurer et al. [12]. GI was calculated by subtracting 90° from the angle formed by the glenoid surface and scapular body. The upward and downward glenoid angles were defined as positive and negative angles, respectively (Fig. 4) [12]. Critical shoulder angle (CSA) was defined as the angle between the lines joining the inferior and superior glenoid and connecting the inferolateral aspect of the acromion with the inferior aspect of the glenoid on true anterior posterior radiographs (Fig. 5) [13].

Data analysis was performed using the SPSS software version 22.0 (IBM Corp.). In addition to descriptive statistics and frequency analysis, Cronbach’s alpha test was used to evaluate the intra- and inter-class correlations between the measurements by the same observer in two separate sessions 2 weeks apart and the measurements of two different observers. The mean values and 95% CI of the four measurements were calculated. Post-hoc power calculations were performed using the G*Power software (version 3.1.9.4, Heinrich Heine University) taking alpha error as 0.05 with a two-tailed significance. The skewness of the data was checked using the Shapiro-Wilk test. Independent sample t-test and Mann-Whitney U-test were used to compare independent variables. Receiver operating characteristic curve analysis was performed to calculate best diagnostic cut-off thresholds of the CSA, GI, and GV values. A P<0.05 indicated statistical significance.

This study included 144 patients, 82 males and 62 females, with a mean age of 25.1±6.3 years (range, 18–40 years), including 72 in the ASI group and 72 in the control group. The demographic information of the patient groups is summarized in Table 1. The Cronbach’s alpha values were between 0.82 and 0.94 for all calculations, showing high within-observation reliability (Table 2). The ASI group had significantly lower scGCA values compared with the control group (32°±4.3° vs. 39.2°±7°, P=0.004, power: 0.99, effect size d=0.924). There was no significant difference in acGCA between the groups (20°±3.2° vs. 22.1°±5.8°, P=0.377, power: <0.80, effect size d=0.368). The GV values of the ASI group were significantly lower than those of the control group (2.2°±5.4° vs. 6.4°±5.4°, P=0.033, power: 0.99, effect size d=0.80). There was no significant difference in GI between the groups (13.7°±5.4° vs. 11.7°±4.5°, P=0.524, power: <0.80, effect size d=0.333). CSA showed no significant difference between the groups (33.6°±4.1° vs. 33.1°±4.5°, P=0.847, power: <0.80, effect size d=0) (Tables 3 and 4).

Scores of GV less than 3.1° showed a 73.3% sensitivity and 71.4% specificity for ASI (likelihood ratio: 2.56, area under the curve [AUC]: 0.745; 95% CI, 0.557–0.933). Scores of scGCA less than 34.5° showed an 85.7% sensitivity and 73.3% specificity for ASI (likelihood ratio: 3.21, AUC: 0.810; 95% CI, 0.692–0.998).

There was a significant correlation between scGCA and GV (P<0.001, r=0.505) (moderate correlation). Logistic regression analysis using scGCA, acGCA, GV, GI, and CSA parameters showed that there was a significant association only in scGCA (P=0.039, odds ratio: 1.28; 95% CI, 1.053–1.589). For each degree of decrease in scGCA, the probability of ASI increased 1.28-fold, independent of other factors. Logistic regression analysis showed no significant association between GV, acGCA, GI, and CSA and ASI (P=0.095, P=0.128, P=0.215, P=0.882, respectively).

The most important finding of this study is that low scGCA values, which indicate a superior position of the coracoid apex relative to the glenoid center, are associated with ASI independently of other parameters. scGCA showed an 85.7% sensitivity and 73.3% specificity for ASI at values below 34.5°.

Hohman and Tetsworth [8] and Aygün et al. [7] found a significant increase in the anteversion of the glenoid in patients with ASI, whereas Moroder et al. [3] and Peltz et al. [9] found no significant difference in GV between ASI and control group patients [1]. Hohman and Tetsworth [8] found a significant relationship between GI and ASI, while Kıvrak and Ulusoy [14] found no significant relationship between GI and ASI. Jacxsens et al. [1] did not find any significant relationship between both GV and GI and ASI. In our study, a significant association was found between GV and ASI, but no significant association was found between GI and ASI. Together these results indicate that the relationship between ASI and GV and GI remains controversial in the literature.

Jacxsens et al. [1] examined the relationship between coracoacromial morphology and ASI by evaluating three-dimensional structured computerized tomography of 31 cadaveric scapulae as a control group and 54 scapulae of recurrent ASI patients with the help of a special software program. In this detailed study, the authors defined the coracoid pillar angle as the angle between the sagittal axis of the glenoid and the bony column (pillar) connecting the coracoid base to the glenoid. This angle was significantly smaller in patients with recurrent instability and showed a relationship between ASI and the coracoid process being indirectly superior. The finding that scGCA was significantly lower in the ASI group compared with the control group in our study supports the study of Jacxsens et al. [1], as it shows that the apex of the coracoid process is located superiorly. One of the differences between the study of Jacxsens et al. [1] and our study is that the coracoid pillar angle is the angle between the plane of the bony column connecting the coracoid base to the glenoid and the sagittal axis of the glenoid and is not a measurement based on the glenoid center, whereas in our study, scGCA evaluates the position of the coracoid apex relative to the glenoid center. Considering that the humeral head is centralized at the center of the glenoid in a stable glenohumeral joint, we believe that it would be more accurate to perform the measurement at the center of the glenoid.

The second difference between Jacxsens et al. [1] and our study is that our study evaluated the coracoid apex, not the coracoid column. The coracoid itself and the soft tissues adhering to the coracoid act as a barrier anterior to the humeral head, and the conjoint tendon acts as a seatbelt at the level of the greater tuberosity during instability, preventing instability [1]. When evaluating both the bony barrier property of the coracoid and the soft tissue barrier property formed by the conjoint tendon adhering to the lateral aspect of the coracoid, we believe that the apex of the coracoid should be evaluated when examining the relationship between the coracoid and ASI because the apex of the coracoid is the first area affected by ASI and the conjoint tendon adheres to the region near the apex of the coracoid, not the base of the coracoid. The third and perhaps most important difference of our study that of Jacxsens et al. [1] is that only MRI is sufficient for scGCA measurement to obtain information about the position of the coracoid in relation to the center of the glenoid, which makes it more likely to be used in clinical practice.

In Mann-Whitney U-tests, GV and scGCA were associated with ASI. However, in regression analysis with GV, GI, CSA, and the parameters we identified, including scGCA and acGCA, all of which are scapula-related and have various associations with instability, we found that only scGCA was associated with ASI independently of the other parameters. In the light of the obtained results, the fact that scGCA is associated with ASI independently of other parameters may be an explanation for the fact that the relationship between GV and ASI is still contradictory in the literature.

To evaluate scGCA, which is a new parameter, in a homogeneous patient group, we included patients without bone defects and non-recurrent dislocations. While the power analysis seems satisfactory, the low number of patients can be considered as a limitation as the ASI group consists of first dislocation patients. Whether scGCA is associated with recurrent instability may be a subject of further study. A second limitation is the difficulty of the measurement technique, because the sagittal and coronal sections must be viewed on the same screen with the sagittal and coronal sections referenced to each other during the detection of the coracoid apex in the sagittal plane. Another limitation is that the body mass index of the patients was not known and the effect of a superiorly positioned coracoid apex on dynamic shoulder biomechanics has not been determined.

Lower scGCA values, which indicate a more superior position of the coracoid apex relative to the glenoid center, are associated with ASI independently of other parameters. scGCA values below 34.5°, indicating a superiorly positioned coracoid apex, show an 85.7% sensitivity and 73.3% specificity for ASI.