Abstract
INTRODUCTION
Reverse shoulder arthroplasty (RSA) provides an effective alternative to anatomic shoulder replacements for individuals with cuff tear arthropathy, but certain osteoarthritic glenoid deformities make it challenging to achieve sufficient long term fixation. To compensate for bone loss, increase available bone stock, and lateralize the glenohumeral joint center of rotation, bony increased offset RSA (BIO-RSA) uses a cancellous autograft for baseplate augmentation that is harvested prior to humeral head resection. The motivations for this computational study are twofold: finite element (FE) studies of BIO-RSA are absent from the literature, and guidance in the literature on screw orientations that achieve optimal fixation varies. This study computationally evaluates how screw configuration affects BIO-RSA graft micromotion relative to the implant baseplate and glenoid.
METHODS
A senior shoulder specialist (GSA) selected a scapula with a Walch Type B2 deformity from patient CT scans. DICOM images were converted to a 3D model, which underwent simulated BIO-RSA with three screw configurations: 2 divergent superior & inferior locking screws with 2 convergent anterior & posterior compression screws (SILS); 2 convergent anterior & posterior locking screws and 2 superior & inferior compression screws parallel to the baseplate central peg (APLS); and 2 divergent superior & inferior locking screws and 2 divergent anterior & posterior compression screws (AD). The scapula was assigned heterogeneous bone material properties based on the DICOM images’ Hounsfield unit (HU) values, and other components were assigned homogenous properties. Models were then imported into an FE program for analysis. Anterior-posterior and superior-inferior point loads and a lateral-medial distributed load simulated physiologic loading. Micromotion data between the RSA baseplate and bone graft as well as between the bone graft and glenoid were sub-divided into four quadrants.
RESULTS
In all but 1 quadrant, APLS performed the worst with the graft having an average micromotion of 347.1µm & 355.9 µm relative to the glenoid and baseplate, respectively. The SILS configuration ranked second, having 211.2 µm & 274.4 µm relative to the glenoid and baseplate. AD performed best, allowing 247.4 µm & 225.4 µm of graft micromotion relative to the glenoid and baseplate.
DISCUSSION
Both APLS and SILS techniques are described in the literature for BIO-RSA fixation; however, the data indicate that AD is superior in its ability to reduce graft micromotion, and thus some revision to common practices may be necessary. While these micromotion data are larger than data in the extant RSA literature, there are several factors that account for this. First, to properly model the difference between locking and compression screws, we simulated friction between the compression screw heads and baseplate rather than a tied constraint as done in other studies, resulting in larger micromotion. Second, the trabecular bone graft is at greater risk of deforming than metallic spacers used when studying micromotion with glenosphere lateralization, increasing graft deflection magnitude. Future work will investigate the effects of various BIO-RSA variables.
For any figures or tables, please contact authors directly.
