Abstract
Category:
Ankle Arthritis
Introduction/Purpose:
Loosening of the tibial component is the primary failure mode in total ankle arthroplasty (TAA). The mechanics of the tibial component loosening has not been fully elucidated. Clinically observed radiolucency and cyst formation in the periprosthetic bone may be associated with unfavorable load sharing at and adjacent to the tibial bone-implant interface contributory to implant loosening. However, no study has fully investigated the load transfer from the tibial component to the bone under multiaxial loads in the ankle. The objective of this study was to utilize subject-specific finite element (FE) models to investigate the load transfer through tibial bone-implant interface, as well as periprosthetic bone strains under simulated multiaxial loads.
Methods:
Bone-implant FE models were developed from CT datasets of three cadaveric specimens that underwent TAA using a modern fixed-bearing tibial implant (a cobalt-chrome tray with a polyethylene bearing, Salto Talaris, Integra LifeSciences). Implant placement was estimated from the post-operative CT scans. Bone was modeled as isotropic elastic material with inhomogeneous Young’s modulus (determined from CT Hounsfield units) and a uniform Poisson’s ratio of 0.3. The tibial tray (Young’s modulus: 200,000 MPa, Poisson’s ratio: 0.3) and the polyethylene bearing (Young’s modulus: 600 MPa, Poisson’s ratio: 0.4) were modeled as isotropic elastic. A 100-N compressive force, a 300-N anterior force, and a 3-Nm moment were applied to two literature based loading regions on the surface of the polyethylene bearing. The proximal tibia was fixed in all directions. The bone-implant contact was modeled as frictional with a coefficient of 0.7, whereas the polyethylene bearing was bonded to the tray.
Results:
Along the long axis of the tibia, load was transferred to the bone primarily through the flat bone-contacting base of the tibial tray and the cylindrical top of the keel, little amount of load was transferred to the bone between those two features (Fig. 1A). Low strain was observed in bone regions medial and lateral to the keel of the tibial tray, where bone cysts were often observed clinically (Fig. 1A). On average, approximated 70% of load was transferred through the anterior aspect of the tibial tray at the flat bone-contacting base, which corresponded to the relatively high bone strain adjacent to the implant edge in the anterior bone-implant interface (Fig. 1B).
Conclusion:
Our results demonstrated a two-step load transfer pattern along the long axis of the tibia, revealing regions with low bone strain peripheral to the keel indicative to stress shielding. Those regions were consistent with the locations of bone cysts observed clinically, which may be explained by the stress shielding associated remodeling of bone. These findings could also describe the mechanism of implant loosening and failure. Future studies may use our model to simulate more loading scenarios, as well as different implant placement and design, to identify means to optimize load transfer to the bone and prevent stress shielding.