RESUMEN
PURPOSE: To evaluate glenoid version and humeral subluxation on preoperative multiplanar imaging of patients who underwent surgery for posterior glenohumeral instability compared with a matched group of patients who had shoulder surgery for other pathology. METHODS: All patients over a 2-year period who underwent surgery for posterior instability had preoperative magnetic resonance (MR) imaging or MR arthrogram reviewed. Patients undergoing shoulder surgery for reasons other than instability were identified as a control group and matched by sex, laterality, and age. Measurement of glenoid version and percentage of humeral subluxation was performed by 2 reviewers after completing a tutorial. Reviewers were blinded to diagnosis and to whether or not the patients were in the experimental or control group. RESULTS: There were 41 patients in each group. The average glenoid version in the control group was 5.6° of retroversion (standard deviation [SD] 3.0), and the average humeral subluxation was 54% (SD 5.1%). In the experimental group, the average glenoid version was 8.1° of retroversion (SD 5.0). The average humeral subluxation in the experimental group was 56% (SD 6.8%). Student t test revealed a statistically significant difference in glenoid version (P = .009) but not humeral subluxation (P = .25). Intra- and inter-rater reliability was measured by the intraclass correlation coefficient and found to have an excellent Fleiss rating with regard to both measurements. CONCLUSIONS: Glenoid retroversion is significantly increased in patients with symptomatic posterior labral tears compared with a control group. However, there was no statistically significant difference between the groups with regard to posterior humeral subluxation and, therefore, is not a reliable indicator of the presence or absence of symptomatic posterior shoulder instability. LEVEL OF EVIDENCE: Level III, retrospective comparative study.