Mechanical and adaptive behaviour of bone in relation to hip replacement: A study of bone remodelling and bone grafting

Abstract

This thesis, consisting of an introduction and four separate papers, gathers considerations on bone behaviour in relation to total hip replacement. Aspects related to primary hip replacement, principally adaptive bone remodelling, are addressed in Paper I from a clinical point of view and in Paper II from a mechanical point of view. Morsellised bone applied in secondary hip replacement is studied in Paper III, where its recoil is modelled as viscoelastic, while Paper IV questions the validity of a solid model for this material.

The relationship between preoperative bone stock and relative postoperative change of bone amount was investigated in Paper I. Younger patients with custom uncemented femoral implants who had high preoperative bone stock had more bone loss than those with low preoperative bone stock. This unexpected result around the hip is nevertheless an accepted result in knee replacement. Also a new graphic interpretation of the paired variations of bone mineral density and projected bone area showed that bone tends to remodel after surgery to reach a lower density and a higher volume.

The main purpose of Paper II was to connect mechanical stimulus to the remodelling observed in the same patients as in Paper I. Bone remodelling was simulated individually and compared with the clinical measurements in the corresponding patient. An additional modelling of a hypothesised fading memory of the bone was implemented to an established set of equations connecting mechanical stimulus to remodelling. Comparisons at a global level of simulated and clinical results showed that simulations have a good predictive value but are not quantitatively correct prior to statistical processing. The observed discrepancy suggested an improvement of the material modelling.

The recoil behaviour of morsellised bone is of great clinical relevance for the primary stability of revision implants. The aim of Paper III was threefold: derive from experiments clinically relevant material parameters, use these to discriminate the effect of pre-treatment of the bone grafts on their recoil properties, and compare these outcomes to loading properties. The experimental unloading was a model using a linear viscoelastic solid model from which three parameters were derived describing the swelling retardation, the swelling speed and the amount of swelling. They allowed the identification of significant effects of water content and particle size on the recoil of morsellised bone. Two of the parameters correlated to loading properties.

The protocol used in Paper III investigates only part of the behaviour of morsellised bone. A different geometry and load modus was studied in Paper IV. Impacted morsellised bone in a cavity mimicking the femoral canal was loaded axially and with torsion. The experiment was modelled with finite elements using the same material modelling as in Paper III, a linear viscoelastic solid model. Though the simulation captured the gross features of the response of bone grafts to loading, it did not achieve displacements as large as in the experiments. This suggested that the pulverulent behaviour of morsellised bone dominated in this load case, allowing it to flow under load, which indicates that fluid viscoelasticity could be a better model for bone grafts.