Periprosthetic femoral fractures can occur as a complication of total hip arthroplasty and are often challenging to treat as the mechanical scenario is influenced by the presence of the metal prosthesis within the bone. This research focuses on finding the optimum fixation for transverse, Vancouver type B1 periprosthetic fractures, stabilised using locking plates and secured using screws. The aim of this study was to experimentally validate a computer model of a human femur, develop that model to represent a periprosthetic femoral fracture fixation and show how the model could be used to indicate differences between plating techniques.

In the first development stage, both a laboratory model and a finite element model were developed to evaluate the mechanical behaviour of an intact composite femur under axial loading. Axial strains were recorded along the medial length of the femur in both cases and compared to provide validation for the computational model predications. The computational intact femur model was then modified to include a cemented total hip replacement, and further adapted to include a periprosthetic fracture stabilised using a locking plate, with unicortical screws above, and bicortical screws below the transverse fracture.

For the intact femur case, the experimental and computational strain patterns correlated well with an average difference of 16%. Following the inclusion of the stem, there was a reduction in the strain in the region of the prosthesis reducing by an average of 45%. There was also a large increase in bulk stiffness with the introduction of the prosthesis. When the fracture and plate fixation were included, there was little difference in the proximal strain where the stem dominated, and the strains in the distal region were found to be highly sensitive to the distribution of the screws.

The results of this study indicate that screw configuration is an important factor in periprosthetic fracture fixation. A laboratory model of the periprosthetic facture case is now under development to further validate the computational models and the two approaches will then be used to determine optimum fixation methods for a range of clinical scenarios.