Poster 137: Do Sex-Based Differences in Native Glenoid Anatomy Predispose to Shoulder Instability? A Radiographic Analysis

OBJECTIVES: The most common five-year age group that experiences a shoulder dislocation is 15 to 19 years. Certain glenoid dimensions, including radius of curvature, depth, and retroversion have been identified as native anatomic characteristics that contribute to instability risk. In addition, diminished glenoid surface area in the setting of bone loss is a critical risk factor for recurrent instability and failure of soft tissue-based repair techniques. However, less is known about how native glenoid width and surface area impact risk of instability in young patients. Further, incidence of instability among young male patients is far greater than females, though no anatomic factors other than soft tissue laxity have been evaluated as contributing factors for instability. Therefore the objectives of this study were to compare the following dimensions between young patients with a documented instability event to matched controls: glenoid width, humeral head width, glenoid surface area, glenoid index (ratio of glenoid height to width), and glenohumeral mismatch ratio (ratio of humeral head width to glenoid width). In addition, we evaluated for sex-based differences across measurements. METHODS: A prospectively maintained database was queried for patients under the age of 21 years who underwent magnetic resonance imaging (MRI) for shoulder instability. Patients with prior shoulder surgery, bony Bankart, or glenoid bone loss were excluded. Patients were matched by sex and age to control patients. Two blinded independent raters measured glenoid height and width and humeral head width on sagittal MRI (Figure 1). The humeral width was measured as the maximum humeral anteroposterior diameter perpendicular to the long axis of the humerus. The long axis of the humerus was established by a line drawn through the centers of two perfect circles created tangent to both outer humeral cortices. The glenoid measurements were performed on the MRI slice that displayed the largest glenoid surface area en face. The glenoid width was measured as the maximum width of the glenoid on this slice perpendicular to its long axis (Figure 1). Glenoid surface area, glenoid index (ratio of glenoid height to width), and glenohumeral mismatch ratio (ratio of humeral head width to glenoid width) were calculated. Intra-observer reliability was assessed based on remeasurement of 50 MRIs by each reviewer and showed ICCs of 0.84, 0.81, and 0.75 for glenoid height, glenoid width, and humeral head width, respectively. Intraclass Correlation Coefficients (ICC) were 0.79, 0.66, and 0.72 for glenoid height, glenoid width, and humeral head width, respectively. RESULTS: A total of 107 patients with an instability episode were age (± 1 year) and sex-matched in a 1:1 manner to 107 controls. Of the total cohort of 214 patients, 150 (70.1%) were male and 64 (29.9%) were female. There were no significant differences in height (p = 0.241), weight (p = 0.167), or BMI (p = 0.341) between the instability and control cohorts (Table 1). Among the entire cohort, there was no difference in glenoid height, glenoid width, glenoid surface area, humeral head width, or glenoid index. Patients in the instability group showed a greater glenohumeral mismatch ratio, indicating that those with instability had smaller glenoids relative to the width of their humeri (p = 0.029). When evaluating the male cohort, there were no significant differences in glenoid height, humeral head width, or glenoid index between patients with instability and controls. Those with history of instability demonstrated smaller glenoid width by 3.5% (p = 0.017), smaller glenoid surface area by 5.2% (p = 0.015), and a greater glenohumeral mismatch ratio (p = 0.027; Table 2). Male patients with instability in the highest quartile glenohumeral mismatch ratio had a dislocation rate of 62.2% compared to 39.5% in the lowest cohort (p = 0.001). On multivariable regression analysis that included patient height and glenoid width, there was a trend toward greater risk of dislocation in those with the smaller glenoid widths (OR 1.18, 95% CI 1.00-1.39, p = 0.053). Similarly, when controlling for height, there was a trend toward greater risk of dislocation in those with higher glenohumeral mismatch ratio (OR 2.45, 95% CI 0.96-6.23, p = 0.061). Among females, there were no significant differences in glenoid height, width, surface area, glenoid index, or glenohumeral mismatch ratio between patients with instability and controls (Table 3). There were no significant differences in instability rates between the highest and lowest quartiles of glenohumeral mismatch (62.5% vs. 43.8%, p = 0.288). On multivariable regression analysis that controlled for patient height, there were no differences in glenoid width (p = 0.556) or glenohumeral mismatch ratio (p = 0.336). CONCLUSIONS: Compared to controls, males with instability had a smaller glenoid surface area compared to controls by 1.74 cm(2), corresponding to a deficit of 5.2%, as well as glenoids that were proportionally smaller relative to their humeri. In contrast, bony glenohumeral morphology was not a significant risk factor for instability among females. This sheds light on the potential of sex-based differences in underlying glenohumeral and capsulolabral anatomy that predisposes to instability. It further provides a potential rationale as to why young patients have significant risk of recurrent instability after soft-tissue stabilization. Male patients who experience instability at a young age may have an anatomic predisposition towards instability based on smaller glenoid size in relation to the humerus. In essence, there may be baseline anatomic "bone loss" inherent in their native anatomy, which may have implications on treatment decision-making.