Effect of reverse shoulder arthroplasty design on deltoid and rotator cuff function in abduction

Reverse total shoulder arthroplasty (rTSA) designs have evolved from Grammont's original design due to the sequelae of postoperative problems that arose. However, the impact of rTSA design on shoulder muscle function is not fully understood. The purpose of this study was to calculate and compare deltoid and rotator cuff function during abduction in the native shoulder and in 3 generic rTSA designs: lateral glenosphere-medial humerus (LG/MH), medial glenosphere-medial humerus (MG/MH), and medial glenosphere-lateral humerus (MG/LH). A subject-specific musculoskeletal model of the native shoulder was created based on a generic model described in the literature. Biplane fluoroscopy was used to measure movements of the shoulder bones for one healthy adult male during abduction. Forward-dynamics simulation was combined with optimization theory to enforce the measured bone positions in the model and determine shoulder muscle forces. The 3 rTSA designs were then virtually implanted using components matched to the subject's bony anatomy, and the model simulations repeated to assess the impact of rTSA design on muscle function. Relative to native anatomy, all 3 rTSA designs increased the moment arm of the middle deltoid up to 120° of abduction, after which the moment arm fell sharply, and the function of the middle deltoid was then greatly diminished. In the native shoulder, the middle deltoid contributed >70% of the total elevation torque up to 90° of abduction and practically all the torque during terminal abduction, while the rotator cuff contributed minimally throughout. By comparison, in all 3 rTSA designs the middle deltoid contributed >70% of the total elevation torque up to 90° of abduction and just 9%-38% of the total elevation torque in terminal abduction. The infraspinatus and subscapularis compensated for the deficit in middle deltoid function by contributing up to 80% of the total elevation torque in terminal abduction. Peak compressive forces applied to the glenoid were comparable in the native shoulder and LG/MH (1.0 body weight, BW) and were 0.1 BW higher than that calculated for the MG designs. Shear forces applied to the glenoid were similar in the native shoulder and all 3 rTSA designs. Our results highlight the importance of retaining the infraspinatus and subscapularis in rTSA due to their ability to compensate for the deficit in deltoid torque during late abduction. Lateralizing the glenosphere increased the forces developed by the infraspinatus and subscapularis and hence the compressive force applied to the glenoid, whereas lateralizing the humerus with an onlay design had no discernible effect on deltoid and rotator cuff function.