Lateralizing the center of rotation in reverse shoulder arthroplasty has been the subject of renewed interest due to complications associated with medialized center of rotation implants. Benefits of lateralization include: increased joint stability, decreased incidence of scapular notching, increased range of motion, and cosmetic appeal. However, lateralization may be associated with increased risk of glenoid loosening, which may result from the increased shear forces and the bending stresses that manifest at the bone-implant interface. To address glenoid loosening in reverse implants with lateralized joint centers, recent studies have focused on testing and improving implant fixation. However, these studies use loads derived from literature specific to subjects with normal anatomy. The aim of this study is to characterize how joint center lateralization affects the loading in reverse shoulder arthroplasty.

Using an established computational shoulder model that describes the geometry of a commercial reverse prosthesis (DELTA® III, DePuy), motion in abduction, scapular plane elevation, and forward flexion was simulated. The simulations were run for five progressively lateralized centers of rotation: −5, 0, +5, +10, and +15 mm (Figure 1). The model was modified to simulate a full thickness rotator cuff tear, where all cuff musculature except Teres Minor were excluded, to reflect the clinical indication for reverse shoulder arthroplasty on cuff tear arthropathy patients. To analyze the joint contact forces, the resultant glenohumeral force was decomposed into compression, anterior-posterior shear, and superior-inferior shear on the glenoid.

Joint center lateralization was found to affect the glenohumeral joint contact forces and glenoid loads increased by up to 18% when the center was lateralized from −5 mm to +15 mm. Compressive forces were found to be more sensitive to lateralization in abduction, while changes in shear forces were more affected in forward flexion and scapular plane abduction. On average, the superior shear component showed the largest increases due to lateralization (up to a 21% increase), while the anterior-posterior shear component showed larger changes than those of compression, except in the most lateralized center position (Figure 2).

The higher joint loads in the lateralized joint centers reflect a shortening of the Deltoid muscle moment arms (Figure 3), since the muscle needs to exert more force to provide the desired motions. The additional shear forces generated by the lateralization may increase the risk of the ‘rocking-horse’ effect. Together with the lateralized joint center, this creates an additional bending stress at the bone-implant interface that puts the implant at further risk of loosening (Figure 1). Current studies on implant fixation tend to use loads in compression and superior shear that exceed the forces seen in this study but have not investigated anterior-posterior shear loads. Our data support that loading in anterior-posterior direction can be significant. Using inappropriate loads to design fixation may result in excessive loss of bone stock and/or unforeseen implant loosening. The implication is that future studies may be performed using this more relevant data set to navigate the tradeoff between fixation and bone conservation.