A previously validated virtual shoulder model was implanted with RSA components that varied by humeral implant type (inset/onset), glenosphere diameter (30, 36, and 42 mm), glenosphere placement (inferior/neutral), glenosphere center-of-rotation offset (0, 5, and 10 mm), humeral neck-shaft angle (130¡ã and 150¡ã), and humeral offset (zero, five, and ten mm). Motion was simulated in all technique combinations until the point of impingement in abduction, flexion/extension (F/E), and internal/external rotation (IR/ER). Regression analysis was used to rank combinations based on motion.
Of 216 possible study combinations, 126 constructs (58 % ) demonstrated no arm-at-side impingement and were included for analysis. Models with the largest motion in abduction, F/E, and IR/ER, respectively, were inset-42-inferior-10-150-zero (107¡ã), inset-36-inferior-10-130-five (146¡ã), and inset-42-inferior-10-130-ten (121¡ã). Humeral neck-shaft angle, glenosphere center-of-rotation offset, glenosphere placement, and glenosphere diameter had a significant effect on motion in all planes tested. Of these variables, humeral neck-shaft angle was most predictive of a change in abduction and F/E motion, whereas glenosphere placement was most predictive of a change in IR/ER motion.
Higher glenosphere center-of-rotation offsets led to an increase in motion in all planes. To maximize motion in abduction, a valgus humeral component should be selected; to maximize F/E, a varus humeral component should be selected; and, to maximize IR/ER, the glenosphere should be placed inferiorly.