ABSTRACT
Due to the increasing number of periprosthetic femoral fractures (PFF), the optimisation of implant design gains importance. For the presented research a validated, subject specific finite element model of a human femur with an inlying total hip stem was used to compare the influence of different geometrical implant parameters on the development of PFF. The heterogeneous bone tissue was modelled on the basis of computed tomography scans. A ductile damage model with element deletion was applied to simulate bone fracture in a load case re-enacting a stumbling scenario. The results were compared in terms of fracture load, subsidence and fracture pattern to analyse the influence of friction at the implant-bone interface, implant size and stem length. The results showed that higher friction coefficients lead to an increase of fracture load. Also, the usage of an oversized implant has a negligible effect while an undersized implant reduces the fracture load by 48.9% for the investigated femur. Lastly, a higher fracture load was reached with an elongated stem, but the bending and change in fracture path indicate a more distal force transmission and subsequent stress shielding in the proximal femur.
Subject(s)
Femoral Fractures , Humans , Finite Element Analysis , Femur , Lower Extremity , Bone and BonesABSTRACT
Periprosthetic femoral fractures (PFF) around total hip replacements are one of the biggest challenges for orthopaedic surgeons. To understand the risk factors and formation of these fractures, the development of a reliable finite element (FE) model incorporating bone failure is essential. Due to the anisotropic and complex hierarchical structure of bone, the mechanical behaviour under large strains is difficult to predict. In this study, a state-of-the-art subject specific FE modelling technique for bone is utilised to generate and investigate PFF. A bilinear constitutive law is applied to bone tissue in subject specific FE models of five human femurs which are virtually implanted with a straight hip stem to numerically analyse PFF. The material parameters of the models are expressed as a function of bone ash density and mapped node wise to the FE mesh. In this way the subject specific, heterogeneous structure of bone is mimicked. For material mapping of the parameters, computed tomography (CT) images of the original fresh-frozen femurs are used. Periprosthetic fractures are generated by deleting elements on the basis of a critical plastic strain failure criterion. The models are analysed under physiological and clinically relevant conditions in two different load cases re-enacting stumbling and a sideways fall on the hip. The results of the analyses are quantified with experimental data from previous work. With regard to fracture pattern, stiffness and failure load the simulations of the load case stumbling delivered the most stable and accurate results. In general, mapping of material properties was found to be an appropriate way to reproduce PFF with finite element models.
