The objective of this research was to better understand the behavior of trabecular bone in response to altered strain states. The specific purpose was to computationally simulate trabecular bone adaptation, using idealized trabecular bone models and an exanple of in vivo trabecular bone adaptation around an implant. A trabecular bone remodeling theory that quantifies Wolff's law of trabecular bone remodeling (Cowin et al., 1992) was incorporated into a remodeling finite element program, KFEM3D (Hart, 1983). The basis of the theory is that under normal conditions in trabecular bone, the fabric tensor, which is a measure of the trabecular bone orientation, is aligned with the stress and strain tensors. When the strain distribution is altered, however, (e.g., with the introduction of an inplant), the trabeculae reorient over time so that the fabric, stress, and strain tensors realign. The computational implementation of this theory was first tested with idealized models to simulate trabecular reorientation when a new stress state was applied to a section of bone.
An animal experiment was conducted to provide the data for an in vivo exanple of trabecular bane adaptation. Cylindrical titanium hydroxyapatite-coated implants were inserted into the distal femur of two mongrel dogs. A finite element model of the normal distal canine femur was constructed, using geometry data from microradiographs taken of ground bone sections. An imaging system was set up to map stereological measurements of trabecular orientation and density that were incorporated into the finite element models. Appropriate loading and boundary conditions were applied, and the normal canine femur equilibrium stresses and strains were calculated. The iitplant was then added to the normal canine finite element model and the modified RFEM3D program was run to simulate the trabecular reorientation, which was ccnpared to the experimental reorientation observed from the microradiographs of the linb with the implant.
This new implementation of the trabecular remodeling theory was shown to effectively simulate theoretical exanples of trabecular bone adaptation and also appears to be valid for in vivo trabecular adaptation.