Experimental and FEA result of three-point bending of cadaver rabbit femora P. Pastila 1,* , N. Moritz 2 , L. Lassila 3 , M. Jokinen 3 , P.K. Vallittu 3 , H.T. Aro 2 and T. Mäntylä 1 1 Tampere University of Technology, Institute of Materials Science, Finland, pirjo.pastila@tut.fi 2 University of Turku, Orthopedic Research Center, Finland 3. University of Turku, Institute of Dentistry and Biomaterials Research, Finland Introduction FEA modeling is generally considered a valuable tool for implant development, provided the model contains essential material properties and is clinically relevant. One of the goals of our current research is to develop a mechano-biological model to simulate experimental in vivo conditions. As a first step three-point bending of cadaver rabbit femora, with and without implants, was performed and modeled with finite element method in order to select the most representative material property set from the literature. Strain is often used in mechano-biological models as driving signal 1,2 (or part of it) for biological processes. It is essential that the material model used in FEA results appropriate strain patterns if long term properties of bone-implant systems are to be simulated with mechano-biology and the models to be used for implant material development. Materials and methods Biomechanical testing was completed at University of Turku. The three-point bending rig had the span of 36 mm, the loading rate was 1 mm/min. Prior to the biomechanical tests bones were imaged with pQCT to create three- dimensional surface geometries that were further used in to simulate the actual experiments. Finite element analyses were completed at Tampere University of Technology with ABAQUS /CAE. Both static and quasi-static analyses were used. Quadratic tetrahedral elements (C3D10M 3 ) and the theory of large deformations were used in all models. The linear elastic material properties, both isotropic and transversely isotropic, were obtained from the literature. Three-point bending tests of empty, intact cadaver femora and osteonomy fixed with implant were completed. The purpose of experiments with intact bones was to select appropriate material properties for cadaver rabbit femora from those found in the literature. Six femora were tested and four of them were analysed with ABAQUS. In addition three femora were tested with strain gauge attached to the tensile surface below the loading point. Experimental strain patterns of the femora with the least deviation of the first principal strain from the bone long axis were compared with the computational results of the same geometry and with the computational results of one geometry from the other intact set. Three-point bending test of osteonomy fixed with titanium (Ti) and fiber reinforced composite (FRC) implants were completed in order to verify if the composite material would show biomechanically better behaviour than Ti. Intact femora were cut and fixed with implant and PMMA cement. Implants were 12 mm long and had diameter of 5 mm. Torsion tests of cadaver rabbit femora were performed to obtain the value of the shear modulus for a semi-empirical transversely isotropic material. Three femora were tested. Results and discussion It was shown that the isotropic linear elastic material model overestimates the stiffness of the bone in three-point bending. The simulated load needed to create 0.5 mm crosshead displacement was clearly higher with the isotropic material properties 4 than with the transversely isotropic 5,6 properties when the modulus E 3 in the direction of the bone long axis was equal to the isotropic Young’s modulus, Figure 1. Composite femur Transversely isotropic Isotropic Figure 1. The effect of anisotropy on simulated three point bending load, quasi-static analysis. To create equal displacement the model with isotropic material needed about 150 N higher force than transversely isotropic material with equal Young’s modulus in the direction of bone long axis. The comparison of the load level only suggested the transversely isotropic material set of the standardized composite femur 6 to match best with the experimental value ~225N of the intact femora without strain gauge. Simulations of the three- point bending of osteonomy fixed with Ti or FRC implants were completed with these properties for