The purpose of this study is to report the clinical and radiological outcomes of patients undergoing primary or revision reverse total shoulder arthroplasty using custom 3D printed components to manage severe glenoid bone loss with a minimum of 2-year follow-up. After ethical approval (reference: 17/YH/0318), patients were identified and invited to participate in this observational study. Inclusion criteria included: 1) severe glenoid bone loss necessitating the need for custom implants; 2) patients with definitive glenoid and humeral components implanted more than 2 years prior; 3) ability to comply with patient reported outcome questionnaires. After seeking consent, included patients underwent clinical assessment utilising the Oxford Shoulder Score (OSS), Constant-Murley score, American Shoulder and Elbow Society Score (ASES), and quick Disabilities of the Arm, Shoulder, and Hand Score (quickDASH). Radiographic assessment included AP and axial projections. Patients were invited to attend a CT scan to confirm osseointegration. Statistical analysis utilised included descriptive statistics (mean and standard deviation) and paired t test for parametric data. 3 patients had revision surgery prior to the 2-year follow-up. Of these, 2/3 retained their custom glenoid components. 4 patients declined to participate. 5 patients were deceased at the time of commencement of the study. 21 patients were included in this analysis. The mean follow-up was 36.1 months from surgery (range 22–60.2 months). OSS improved from a mean 16 (SD 9.1) to 36 (SD 11.5) (p < 0.001). Constant-Murley score improved from mean 9 (SD 9.2) to 50 (SD 16.4) (p < 0.001). QuickDASH improved from mean 67 (SD 24) to 26 (SD 27.2) (p = 0.004). ASES improved from mean 28 (SD 24.8) to 70 (SD 23.9) (p = 0.007). Radiographic evaluation demonstrated good osseointegration in all 21 included patients. The utility of custom 3D-printed components for managing severe glenoid bone loss in primary and revision reverse total shoulder arthroplasty yields significant clinical improvements in this complex patient cohort

Introduction. Stemless shoulder implants have recently gained increasing popularity. Advantages include an anatomic reconstruction of the humerus with preservation of bone stock for upcoming revisions. Several implant designs have been introduced over the last years. However, only few studies evaluated the impact of the varying designs on the load transfer and bone remodeling. The aim of this study was to compare the differences between two stemless shoulder implant designs using the micro finite element (µFE) method. Materials and Methods. Two cadaveric human humeri (low and high bone mineral density) were scanned with a resolution of 82µm by high resolution peripheral quantitative computer tomography (HR-pQCT). Images were processed to allow virtual implantation of two types of reverse-engineered stemless humeral implants (Implant 1: Eclipse, Arthrex, with fenestrated cage screw and Implant 2: Simpliciti, Tornier, with three fins). The resulting images were converted to µFE models consisting of up to 78 million hexahedral elements with isotropic elastic properties based on the literature. These models were subjected to two loading conditions (medial and along the central implant axis) and solved for internal stresses with a parallel solver (parFE, ETH Zurich) on a Linux Cluster. The bone tissue stresses were analysed according to four subregions (dividing plane: sagittal and frontal) at two depths starting from the bone-implant surface and the distal region ending distally from the tip of Implant 1 (proximal, distal). Results. Medial loads produced higher bone tissue stresses when loading was applied along the implant axis. This was more prominent in the lower density bone, causing more than 3 times higher stresses in the highest region for both implants. Bone tissue stresses were also shown to be higher in the low density specimen, especially in the distal zone. The maximum bone tissue stress ratio for low/high density bone reached 4.4 below Implant 1 and 2.2 below Implant 2, occurring both with a medially-directed load. For both implants, the highest bone tissue stresses were predicted in the distal region than in the proximal region, with larger distal-to-proximal stress ratios below Implant 1 than Implant 2 (3.8 and 1.7, respectively). Discussion. Our µFE analyses show that the implant anchorage design clearly influences load transfer to the periprosthetic bone. The long fenestrated cage screw of Implant 1 showed more direct distal stress transfer, which may lead to stress shielding in the proximal region, in a larger extent than Implant 2 which tends to distribute loads more evenly. Furthermore, periprosthetic bone quality appears to be an important factor for load transfer, causing dramatic changes due to different loading condition and implant geometry. These findings will help further improve anchorage design for stemless humeral heads in order to minimize bone remodeling and the long-term fixation of these implants