The Critical Shoulder Angle is Predictive of Rotator Cuff Tears and Shoulder Osteoarthritis and is Better Assessed with Radiographs over MRI

OBJECTIVES: The critical shoulder angle (CSA) (Moor et al., Annual Meeting AAOS 2013) is measured as the angle between glenoid and lateral border of the acromion and has been associated with glenohumeral osteoarthritis (OA) as well as rotator cuff tears (RTC). The purpose of this study was to determine the reproducibility of the CSA for the detection of patients with RTC or OA by x-ray and to determine whether the CSA can also be measured with magnetic resonance imaging (MRI). METHODS: In this IRB approved study, a pre-hoc power analysis was performed to determine sample size. The CSA was measured by x-ray and MRI in 3 age-matched groups of 10 patients each, consisting of a non-traumatic full-thickness RCT group, a non-traumatic and non-inflammatory OA group, and a non-RCT/non-OA-pathology group. All patients underwent routine x-rays with a standardized technique and had standard shoulder MRI’s in a 3.0T magnet. All pathologies were determined from pre-op imaging studies and were confirmed at surgery. Subjects were retrieved from a prospectively collected data registry. Three independent observers, one radiologist and 2 orthopedic surgeons, performed the measurements. CSA angles were measured using Stryker OfficePACS Power 4.1 Express Edition (Kalamozoo, MI). RESULTS: The inter-observer correlation coefficients of the measurements done by x ray and MRI were 0.95 and 0.83, respectively. There were differences in the mean CSA’s among the three groups, measured by x-ray (ANOVA between RCT, OA, non-RCT/non-OA: p <0.001). MRI showed greater variability in the CSA measurements among observers and across pathologies. Radiographs were more predictive of pathology than MRI. For example, patients with CSAs greater than 35° had a 79.3% chance to be part of the RCT group when measured by x-ray, whereas the chance was only 69.0% using MRI. Those with angles less than 30° by x-ray had an 82.6% probability to be part of the OA group, whereas the probability dropped to 47.3% when the measurement was done by MRI. When the study groups were compared, no significant differences were seen in the mean CSAs by x-ray and MRI in the RCT and non-RCT/non-OA pathology group, 37.3±2.6° versus 36.4±1.9° (p = 0.24) and 32.7±2.5° versus 31.8±2.7 (p = 0.20), respectively. Whereas a significant difference was observed the OA group by x-ray 28.7±2.2° versus 31.3±4.4 by MRI (p = 0.01), although the trend remained for a lower CSA in OA. CONCLUSION: The CSA measured by x-ray showed excellent inter-observer agreement and less variability than measurements done by MRI, especially in patients with OA. The CSA was also highly diagnostic for predicting RCT’s and OA of the shoulder. This study furthermore confirms the work of Moor and Gerber and suggests that there may be biomechanical and anthropomorphic components that underlie these shoulder disorders. CSA may also have effects on clinical outcome after surgical treatment. Furthermore, novel surgical procedures might be able to affect the CSA to hasten or prevent progression of RCT’s or OA.