Abstract
Finite element modelling is being extensively used to evaluate the biomechanical behaviour of fractured bone treated with fixation devices. Appropriate modelling of the bone-implant interface is key to quality biomechanical prediction.
The present study considers this interface modelling in the context of locking plates. A majority of previous studies assume the interface to be represented by a tied constraint or a fully bonded interface. Many other studies incorporate a frictional interface but ignore screw threads. This study compares the various interface modelling strategies. An interface with screw threads explicitly included is also considered.
The study finds that interface modelling has significant impact on both the global and local behaviour. Globally, the load-deflection behaviour shows considerable difference depending on the interface model. Locally, the stress-strain environment within the bone close to the screws is significantly altered.
The results show that the widely used tie constraint can overestimate stiffness of a construct which must be correctly predicted to avoid non-union or periprosthetic re-fracture, especially in osteoporotic bone. In addition, the predictions of screw loosening, bone damage and stress shielding are very different when screw threads are included in the model.
