*4.3.3. Interfragmentary movement in acetabular fracture osteosynthesis measured with PU- foam based synthetic bones*

192 Polyurethane

fixation.

shown on the far right column.

Data clouds obtained from laser scanning PU-

foam pelvis

clouds, which accurately described the shapes of the fragment and the fractured pelvis, were obtained (Figure 17). The fractured pelvis model was developed from our previous FE model of the pelvis, which was generated from CT scans of the synthetic pelvis used in the experiment (Shim et al., 2010) . Our elements had inhomogeneous location dependent material properties despite large element size and different material properties were assigned to solid and cellular polyurethane foams which mimic cortical and cancellous bone properties separately. The loading and boundary condition that mimics the mechanical experiment setup were employed. The FE models of the screws were not generated explicitly. Instead tied contact was used to model the bond between the fractured pelvis and the fragment from the screws. The locations of the screws on the fragment FE model were identified first from the laser scanned data. Then, the tied contact condition that ensures a perfect bond between slave and master faces was imposed on the identified faces to simulate the bond that screws provide when connecting the fragment with the bone. The rest of fragment faces were modeled with frictional contact (μ=0.4)(Gordon et al., 1989). The predicted interfragmentary movements from the FE model under the same loading and boundary conditions as the experiment were then compared with the experimental value to test our hypothesis. Once tested, then, the screw positions were varied by changing tied contact faces to simulate all possible screw positions in order to identify the positions that achieved the most stable

**Figure 17.** The far left column shows clouds of data points obtained from laser scanning. The center column shows the meshes for the fragment and fractured pelvis that were generated by geometric fitting to laser scanned data points. The red faces on the fragment mesh indicate where tied contact conditions were imposed in order to simulate the support provided by the screws. The final mesh is

Model generated from data points by geometric fitting

Model for fracture reduced pelvis

The overall amount of displacement between the pelvis and the fragment was relatively small and the main direction of the fragment movement was in the lateral posterior direction (in body directions). The average displacement was around 0.4 – 0.9 mm for both screw and plate fixations. The plates gave higher stability especially in the horizontal and lateral directions (Figure 18). However, screw fixations also gave good stability of less than 1mm on average in all directions. Therefore, considering the fact that the maximum load of our experiment was higher than normally allowed weight bearing (around 20kg after the surgery for 3 months), the stability of screw fixation was sufficient for the cyclic loading condition used.

**Figure 18.** Comparison of interfragmentary movement between plate and screw fixation
