In this grant, we continue to investigate the multiaxial failure behavior of human trabecular bone. The clinical motivation is to improve mechanistic understanding of osteoporotic hip fractures. So far, we have described the failure behavior of trabecular bone for the human proximal femur, tibia, and vertebral body (n=214, each with a high-resolution three-dimensional reconstruction) for habitual type loading. We found that the strains at which the bone fails are relatively independent of volume fraction, but can differ across sites. The difference between tensile and compressive failure strains is greater in the higher density bone of the femoral neck. These results suggest that micro-structural bending of individual trabeculae dominates failure mechanisms at low density, but becomes less important at high density. Our work on axial-shear mutliaxial testing demonstrated that engineering type multiaxial failure criteria-expressed in units of strain-can predict the experimental behavior very well despite variations in volume fraction and anatomic site, with mean errors between experiment and criterion of less than 1 percent. These findings provide the foundation for more advanced research on multiaxial failure behavior. Given the difficulty of performing the requisite multiaxial experiments on relatively scarce human tissue, we propose to combine sophisticated computer simulations of the micro-mechanical failure mechanisms of trabecular bone with selected multiaxial experiments to develop a complete multiaxial failure criterion for femoral (neck and trochanter) trabecular bone. Specifically, we hypothesize that multiaxial failure of femoral trabecular bone is determined by critical strain levels and is independent of volume fraction. High- resolution micro-mechanical computer models, using 3D images obtained to date, will be used to simulate multiaxial behavior. These models will be first calibrated against our existing experimental data and then used to predict a wide variety of states of multiaxial failure. The predictions will be validated by new multiaxial experiments (n=100). The validated criterion will then be analyzed for a dependence on volume fraction. To address clinical significance, we also hypothesize that this multiaxial behavior is an independent determinant of whole bone femoral strength. To test this, anatomically detailed finite element models will be developed of the whole proximal femur using computed tomography and magnetic resonance scans. Simulations will be done with and without inclusion of the multiaxial failure criterion, and these predictions will be compared against experimental data for 20 cadaver bones. This work should provide substantial insight into mechanisms of whole bone and trabecular bone failure under traumatic conditions.
Showing the most recent 10 out of 48 publications
