It is estimated that up to 1% of the world’s population are affected by ankle arthritis. Current surgical treatment options for end stage arthritis include total ankle replacement (TAR). The success rate of this treatment is lower than equivalent joint preservation procedures in the hip and knee, with survivorship rates reported between 81% and 93% at 5 year followup. Patients requiring treatment tend to be younger and more active due to the etiology of ankle arthritis which is commonly related to previous trauma.
Despite clinical need, research into the failure mechanisms of TARs is limited. In the hip and knee, extensive research has successfully used finite element analysis (FEA) to study aspects of failure including contact pressures and fixation methods. The research focus of this project was to develop a range of finite element models of the TAR to explore the effects of implant geometry, bone quality and implant positioning on its performance.
Finite element models were created using CT images of five cadaveric ankle scans. In initial modelling, five positions in the gait cycle were used to quantify the affect of loading. Five different clinically relevant tibial fixation conditions were also modelled, from a fully fixed tibial component to a completely loose one. Minimum principal strains (maximum compressive) were highest at maximum loading magnitude, with flexion angles having little effect on the resultant strains. An unfixed tibial implant produced higher bone strains in all models. Variations between models showed bone material properties had the largest ffect on the resultant strain distributions.
A combined experimental and computational study was undertaken using distal tibial samples to determine a relationship between CT greyscale and Young’s Modulus in order to validate bone material properties. The computational methodology was also modified to ensure the process was applicable to all bone samples, including those lacking bone marrow. Good correlation between experimental and computational stiffness was observed. The resultant bone material properties were found to be less stiff than initial material property literature suggested.
Finally, combining all previous findings, three clinically relevant studies were undertaken under the guidance of a local surgeon. These three studies were based on implant alignment, implant sizing and tibial implant design. Changes in implant alignment of up to 10° did not alter the magnitude of peak strains seen in the tibia, however the location of these peak strains changed around the implant. Changes in implant sizing affected the strain distribution around the tibial resection surface, with higher strains seen using smaller implants. Subsidence of these implants may be more likely than larger implants where seating on the cortex was possible. Changes in tibial implant design produced the largest changes in strain distributions compared to the other two studies.
Differences between models (patients) were more pronounced. Bone quality was found to be more important than any of the tibial implant changes studied. Higher compressive strains were consistently found in models with poor bone quality, indicating patient selection is the most important determinant in implant outcomes. Through these studies, bone quality was shown to influence resultant bone strains more than implant size, alignment and design.
