Reverse shoulder arthroplasty (RSA) is a joint replacement procedure used mainly to treat patients with severe shoulder osteoarthritis combined with massive rotator cuff tears. It involves reversing the ‘ball and socket’ orientation of the glenohumeral joint in the shoulder. While RSA has been largely successful in treating pain and improving function in these patients, complication rates remain high. Many of these complications, including joint instability and scapular notching (excessive bone wear), are caused or exacerbated by impingement of the humerus or the humeral component against the scapula. Adduction deficit refers to a patient’s inability to fully adduct the arm due to impingement. Minimizing adduction deficit may improve RSA patients’ functional outcomes.

An existing mechanical shoulder simulator was further developed to model glenohumeral range of motion in RSA. The three heads of the deltoid were modelled using polyethylene cable and electric linear actuators with inline load cells. RSA components were implanted in Sawbones® scapula and humerus bone models. The scapula was fixed in the frame of the simulator. Triads of optical tracking markers were attached to the humerus and simulator frame and used to track segment motion.

A data analysis technique was developed to determine when joint impingement occurred. The convex and concave surfaces of the glenoid and humeral components were digitized, and a leastsquares sphere fit was used to find their centres. The distance between these centres was then calculated during passive abduction and adduction of the humerus, and labeled d_GH. Impingement onset was defined as the point where d_GH was five standard deviations above its baseline value, indicating that the components were no longer concentric. This technique was used to determine the effects of humeral neck-shaft angle, socket depth, glenosphere diameter and eccentricity on range of motion and adduction deficit. A retentive humeral cup depth increased adduction deficit by 14 degrees and reduced range of motion by 26 degrees. A decreased neck-shaft angle reduced adduction deficit by 10 degrees but had little effect on overall range of motion. Diameter and eccentricity had no effect on either measure.