Expérimentation et modélisation détaillée de la colonne vertébrale pour étudier le rôle de facteurs anatomiques et biomécaniques sur les traumatismes rachidiens

The thesis was divided into five objectives. The first objective was to perform experimental tests to collect data on the biomechanical failure behavior of the spine subjected to dynamic compression. Normalized force-displacement curves, videos and QCT images allowed characterizing failure parameters (force, displacement and energy at failure) and fracture patterns. Specimens with substantial vertebral body osteophytes presented higher stiffness (2.7 times on average), force (1.8 times), displacement (5.2 times) and energy (1.6 times) at failure than other segments. The presence of osteophytes significantly influenced the location, pattern and type of fracture. Vertebral osteophytes provided to the underlying vertebra a protective mechanism against severe compression fractures. The VTD was a good predictor of the dynamic force and energy at failure for specimens without substantial osteophytes.

The second objective was the refinement of the Spine Model for Safety and Surgery (SM2S) to improve its biofidelity during dynamic loading. SM2S is an anatomically realistic 3D FEM of the thoracic and lumbar human spine created in joint collaboration between the Laboratory of Biomechanics and Applications of IFSTTAR/Aix-Marseille University and École Polytechnique de Montréal for dynamics (impacts, virtual trauma) and quasi-statics (implants biomechanics) applications. The L2-L3 functional spinal unit (FSU) of SM2S was developed by Marwan El- Rich during his post-doctoral work.

The third objective was to validate the SM2S model using experimental data measured under dynamic loading conditions. Compression, flexion, extension, anterior and posterior shear loads were applied to different spinal segments (T8-T10, T11-L1 and/or T12-L5), at displacement rates ranging from 0.1 to 4 m/s. Simulation results (force-displacement curves, failure parameters, type and location of spinal injuries) were compared to the results of experimental tests carried out during the thesis (ref. objective 1) or taken directly from literature (Demotropoulos et al., 1999; Demotropoulos et al., 1998; Kifune et al., 1995). The model-predicted results agreed well with the average experimental data, thus supporting the relevance and validity of the SM2S model.

The fourth objective was to exploit the SM2S model to investigate the influence of the load rate on spinal injuries (Q2). Simulations performed previously for the validation of the lumbar spine model (T12-L5) in compression, flexion-extension and shear, at slow (0.1 m/s) and fast (1 or 4 m/s) displacement rates, were run up to failure. Simulation results were then used to compare the initiation sites of the injuries. For most loading cases, spinal injuries occurred at the L1 and L4 levels. However, various initiation sites were observed according to the load rate, thus confirming the key-role of this parameter on the onset of spinal trauma.

Finally, the fifth objective was to exploit the SM2S model to investigate the influence of the sagittal profile of the spine on spinal injuries sustained in two different injury scenarios. The first scenario represents accidents for which a loading primarily oriented in compression is responsible for the spinal injury. The second scenario represents a frontal car crash for which a flexion-distraction mechanism is responsible for the spinal injury. Results showed that the spinal curvature and the tilt angle of the specimen have a significant influence on the bone fracture. This influence is modulated by the mobility of the segment at the time of the impact.

In the second scenario, an original method that combined experimental measurements with computer assisted design (CATIA v5 software) was developed to characterize the sacral slope and the lumbar lordosis of 34 sagittal profiles in automobile driving posture. These profiles were transferred to the T8-Sacrum spinal segment of the SM2S model. Loading and boundary conditions that simulate a frontal car crash were applied on each profile, with and without the weight of the upper trunk. Different types and locations of spinal injuries were observed for profiles with high lordosis or low kyphosis. These profiles were infrequently adopted by car users, thus suggesting that the sagittal profile of the spine has a limited influence on the nature of spinal injuries in car accidents that involve flexion-distraction mechanisms. (Abstract shortened by UMI.)