A Computational Model to Predict In Vivo Kinetics in Implanted and Non-Implanted Shoulders

The purpose of this study was to develop and implement a computational model designed to input in vivo kinematic and predict in vivo forces and torques for the shoulder, elbow, and wrist in normal, rotator cuff-deficient (RCD), reverse shoulder arthroplasty (RSA) and total shoulder arthroplasty (TSA) shoulder subjects. Twenty subjects, divided evenly amongst the four shoulder types, performed a box-lift activity while under fluoroscopic surveillance. Three dimensional (3D) in vivo kinematics was determined for the subjects using implant models and bone models created from CT (computed tomography) scans in a 2D-to-3D registration process. The kinematics were used as input for an inverse dynamics mathematical model, and the subject-specific kinetics were derived. Average resultant shoulder forces were 78.3N (range: 70.4N to 117N, SD: 5.213), 102N (range: 90.2N to 180.2N, SD: 12.339), 94.9N (range: 84.9N to 149N, SD: 10.02), and 92.5N (range: 87.984N to 95.370N, SD: 1.848), for normal, RCD, RSA, and TSA subjects, respectively. Average resultant shoulder torques were 23.6Nm (range: 8.32Nm to 73.7Nm, SD: 11.227), 29.6Nm (range: 22.892Nm to 71.377Nm, SD: 7.581), 27.2Nm (range: 19.961Nm to 59.352Nm, SD: 6.664), 20.3Nm (range: 11.700Nm to 31.409Nm, SD: 6.496), for normal, RCD, RSA, and TSA shoulders, respectively. This study revealed that RCD subjects exhibited a decreased ROM (range of motion) of the humeral head with respect to the glenoid, as compared to the other groups. This study also showed that subjects having a rotator cuff-deficient shoulder and/or a replaced shoulder tend to use compensatory motions to perform the task of lifting a box, and, as a result, they experience greater forces at the glenohumeral joint. Paradoxically, the RCD subjects experienced the highest joint forces and torques among the different shoulder types.