Characterizing the trade-off between range of motion and stability of reverse total shoulder arthroplasty

Background: The trade-off between range of motion (ROM) and stability of reverse total shoulder arthroplasty (RSA) has long been hypothesized to exist but has not yet been well characterized. The goal of this study was to use design optimization techniques to obtain a Pareto curve, which quantifies the trade-off between 2 competing objectives and is defined by the performance of optimum designs that maximize one surgical outcome without sacrificing the other.

Methods: Multi-objective design optimization techniques were used; 4 design and surgical parameters including glenoid lateralization (GLat), neck-shaft angle (NSA), inferior offset of the center of rotation (COR_inf), and humerus lateralization (HLat) were tuned simultaneously. The ROM and stability, the objectives to be optimized, of any candidate design were characterized computationally using a combination of finite element models, musculoskeletal models, analytical equations, and surrogate models. Optimum designs and Pareto curves were determined separately for a standard cup depth and a shallow cup depth. The performance of the optimum designs, in terms of ROM and stability, was compared with a representative commercially available design.

Results: A Pareto curve was obtained for each cup depth, confirming there is a trade-off between ROM and stability of RSA. In comparison to the commercially available design (cup depth, 8.1 mm; GLat, 5 mm; NSA, 155°; COR_inf, 0 mm; HLat, 0 mm), the designs optimized for ROM offered 38.8% (cup depth, 6 mm; GLat, 16 mm; NSA, 163.6°; COR_inf, 4 mm; HLat, 0.6 mm) and 35.2% (cup depth, 8.1 mm; GLat, 16 mm; NSA, 160.5°; COR_inf, 4 mm; HLat, -0.2 mm) improvement in ROM. The designs optimized for stability (cup depth of 6 mm with GLat of 16 mm, NSA of 170°, COR_inf of 4 mm, and HLat of 3 mm and cup depth of 8.1 mm with GLat of 16 mm, NSA of 170°, COR_inf of 4 mm, and HLat of 3 mm) both offered 12.4% improvement in stability over the commercially available design. Designs in the toe region of the Pareto curve offered between 75% and 90% of the maximum possible improvement over the commercially available design for both objectives.

Conclusion: It was confirmed that a trade-off exists between ROM and stability of RSA, in which maximizing one outcome requires a sacrifice in the other. The relative gains and sacrifices in the competing outcomes when traversing the Pareto front could aid in understanding clinically optimum designs based on patient-specific needs.