Shoulder arthoplasty has increased in the last years and its main goal is to relieve pain and restore function. Shoulder prosthesis enters in the market without any type of pre-clinical tests. Within this paper we present study experimental and computational tests as pre-clinical testing to evaluate total shoulder arthoplasty performance. An in vitro experimental simulator was designed to characterize experimentally the intact and implanted shoulder glenoid articulation. Fourth generation Sawbones® composite left humerus and scapula were used and the cartilage was replicated with silicone for the intact articulation (figure 1). In the intact experimental articulation we considered the inferior glenohumeral ligament as an elastic band with equivalent mechanical properties. For the implanted shoulder, the Comprehensive® Total Shoulder System (Biomet®) with a modular Hybrid® glenoid base and Regenerex® central post was considered (figure 2). The prostheses were implanted by an experienced surgeon and clinical results from orthopedic registers were collected. The system structures were placed to simulate 90º in abduction, including the following muscle forces: Deltoideus 300N, Infraspinatus 120N, Supraspinatus 90N and Subscapularis 225N. The finite element model was created with tetrahedral linear elements with linear elastic and isotropic material for the humerus in figure 3, (Young's modulus for cortical bone − 16.5 GPa; trabecular bone − 124 MPa). Anisotropic behavior was considered for the scapula model (E11 = 342.1 MPa, E22 = 212.8 MPa, E33 = 194.4 MPa). The shoulder prosthesis was of polyethylene with 1GPa and titanium with 110 GPa. The Poisson's ratio was 0.3 in all material, except for polyethylene where we assumed a value of 0.4. A long-term post-operative condition was simulated.Introduction
Materials and methods
Total shoulder arthroplasty is a well-tested procedure that offers pain relief and restores the joint function. However, failure rate is still high, and glenoid loosening is pointed as the main reason in orthopedic registers. In order to understand the principles of failure, the principal strain distributions after implantation with Comprehensive® Total Shoulder System of Biomet® were experimental and numerically studied to predict bone behavior. Fourth generation composite left humerus and scapula from Sawbones® were used. These were implanted with Comprehensive® Total Shoulder System (Biomet®) with a modular Hybrid® glenoid base and Regenerex® glenoid and placed in situ by an experienced surgeon. The structures were placed in order to simulate 90º abduction, including principal muscular actions. Muscle forces used were as follows: Deltoideus 300N, Infraspinatus 120N, Supraspinatus 90N, Subscapularis 225N. All bone structures were modeled considering cortical and the trabecular bone of the scapula. The components of prosthesis were placed in the same positions than those in the in vitro models. Geometries were meshed with tetrahedral linear elements, with material properties as follows: Elastic modulus of cortical bone equal to 16 GPa, elastic modulus of trabecular bone equal to 0.155 GPa, polyethylene equal to 1GPa and titanium equal to 110 GPa. The assumed Poisson's ratio was 0.3 in all except for polyethylene where we assumed a value of 0.4. The prosthesis was considered as glued to the adjacent bone. The finite element model was composed of 336 024 elements. At the glenoid cavity, the major influence of the strain distributions was observed at the posterior-superior region, in both cortical and trabecular bone structures. The system presents critical region around holes of fixation in glenoid component. At the trabecular bone, the maximum principal strains at the posterior-superior region ranged from 2250 µε to 3000 µε. While at the cortical bone, the maximum principal strains were 300 µε to 400 µε. The results observed evidence some critical regions of concern and the effect of implant in the bone strains mainly at the posterior-superior region of the glenoid cavity is pronounced. This indicates that this region is more affected by the implant if bone remodeling is a concern and it is due to the strain-shielding effect, which has been connected with loosening of the glenoid component.
