Acromial morphology is not associated with rotator cuff tearing or repair healing

Background: The purposes of this study were to determine whether acromial morphology (1) could be measured accurately on magnetic resonance images (MRIs) as compared to computed tomographs (CTs) as a gold standard, (2) could be measured reliably on MRIs, (3) differed between patients with rotator cuff tears (RCTs) and those without evidence of RCTs or glenohumeral osteoarthritis, and (4) differed between patients with rotator cuff repairs (RCRs) that healed and those that did not.

Methods: This is a retrospective comparative study. We measured coronal, axial, and sagittal acromial tilt; acromial width, acromial anterior and posterior coverage, and glenoid version and inclination on MRI corrected into the plane of the glenoid. We determined accuracy by comparison with CT via intraclass correlation coefficients (ICCs). To determine reliability, these same measurements were made on MRI by 2 observers and ICCs calculated. We compared these measurements between patients with a full-thickness RCT and patients aged >50 years without evidence of an RCT or glenohumeral osteoarthritis. We then compared these measurements between those patients with healed RCRs and those with a retorn rotator cuff on MRI. In this portion, we only included patients with both a preoperative MRI and a postoperative MRI at least 1 year from RCR. Only those patients without tendon defects on postoperative MRIs were considered to be healed. In these patients, we also radiographically measured the critical shoulder angle.

Results: In a validation cohort of 30 patients with MRI and CT, all ICCs were greater than 0.86. In these patients, the inter-rater ICCs of the MRI measurements were >0.53. In our RCT group of 110 patients, there was greater acromial width [mean difference (95% confidence interval) = 0.1 (0, 0.2) mm, P = .012] and significantly less sagittal acromial tilt [9° (5°-12°), P < .001] than in our comparison group of 107 patients. A total of 110 RCRs were included. Postoperative MRI scans were obtained at a mean follow-up of 24.2 ± 15.8 months, showing 84 patients (76%) had healed RCRs. Aside from acromial width, which was 0.2 mm different and thus did not have clinical significance, there was no association between healing and any of the measured morphologic characteristics. Patients with healed repairs had significantly smaller tears in terms of both width (P < .001) and retraction (P < .001).

Conclusion: Although the acromion is wider in RCTs, the difference of 0.1 mm likely has no clinical significance. The acromion is more steeply sloped from posteroinferior to anterosuperior in those with RCTs. These findings call into question subacromial impingement due to native acromial morphology as a cause of rotator cuff tearing. Acromial morphology, critical shoulder angle, and glenoid inclination were not associated with healing after RCR. This study does not support lateral acromioplasty.