EVALUATING THE EFFICACY OF FOAM BONE SUBSTITUTES IN REPLICATING REVERSE SHOULDER ARTHROPLASTY IMPLANT MICROMOTION

Reverse Shoulder Arthroplasty (RSA) implant fixation is assessed by studying baseplate micromotion as dictated in ASTM F2028 standard. This standard uses homogeneous foam material as a proxy for glenoid bone, but this material approximates cancellous bone while RSA primary fixation also relies on screw engagement with the glenoid vault's cortex. As such, it is unclear whether this accurately replicates RSA fixation. Therefore, the goal of this study was to compare RSA micromotion in foam substitutes and cadaveric specimens, to assess their similarity and inform changes to ASTM F2028. Four cadaveric shoulder specimens were implanted with angled RSA baseplate components (Perform Reversed Angled Baseplate, Stryker). Identical implants were installed in four 20 PCF foam blocks (Sawbones Inc). A worst-case fixation scenario was used: four 26 mm screws, consisting of three locking with 0° angulation and an inferior compression screw with 3° angulation. Cyclic loading and micromotion testing followed ASTM F2028 (Figure 1). LVDT sensors measured shear and normal micromotion at the inferior glenosphere aspect while a constant 430N compressive load was applied and a shear load was ramped between 350N in the inferior and superior directions. The sensors were attached to the glenosphere and a fixed reference plate level with the implant-bone interface. The baseplate-glenosphere construct then underwent 45° abduction-adduction cylcing while a 750N compressive load was applied. Baseplate micromotion was remeasured after 10k and 100k cycles. Displacements are reported as the total micromotion between the maximum inferiorly and superiorly directed shear loads. Micromotions were normalized to the average uncycled foam specimen values and two-tailed T-tests were performed to compare the foam and cadaveric results. Cycling primarily increased micromotion but the increase was non-monotonic (Table 1). There were no statistically significant differences in micromotion between the foam and cadaveric specimens prior to cycling for the normal and shear directions: 72.1±34.1% (Mean Difference ± Standard Error of Difference), p = 0.095 and 38.2±20.0%, p = 0.105, respectively. After 10k cycles, the difference in normal (134.4±81.5%, p = 0.199) and shear (89.7±42.0%, p = 0.126) micromotion was not statistically significant. After 100k cycles, the difference in normal micromotion was not significant (45.9±61.5%, p = 0.585), but the difference in shear was significant (74.6±19.5%, p = 0.025). The cadaveric specimens exhibited greater normal and shear micromotion at all timepoints (range: 38%-134%), but, due to the large variance across cadaveric specimen results, the difference was only significant in shear after 100k cycles. These findings may result from the lack of cortical engagement in this worst-case (i.e. minimum screw length and angulation) screw configuration, and the reduced contact between the implant and the cadaveric glenoid due to the need to maintain subchondral bone during bone preparation. These results indicate that foam may be an efficacious surrogate for worst-case fixation scenarios where screws may not be effectively engaging with dense cortical bone. However, further study is needed to determine the efficacy of foam as a bone surrogate in cases where screw lengths and angulation are appropriately matched to the specific anatomy to achieve cortical bone engagement. For any figures or tables, please contact authors directly.