I want to congratulate Nathaniel Price from University of Florida for winning the Research & Design award at BMES2012. The BMES 2012 annual meeting in Atlanta Georgia had over 5,000 attendees.  I had the pleasure of meeting Nathaniel and his professor Dr. Kim. 

The Effects of Cortical Thickness, Bone Strength, and Screw Length on Rigid Sternal Fixation Stability

N.B. Price1, N. H. Kim1, B. Wilcox2, and B. Hatcher21

University of Florida, Gainesville, FL, 2Biomet, Jacksonville, FL

Introduction: Bony instability following median sternotomy procedures increases the risk for sternal wound infection[1].Sternal wound infection and sternal dehiscence cause significantly increased morbidity and mortality [1]. Biomechanicalstudies of sternal closure techniques are currently limited by the wide variability in sternum size, strength, and density [2].Synthesized bone models provide a greater degree of reproducibility but currently lack the ability to represent the bi-material,cortical and cancellous, composition of natural bone. Finite element analysis (FEA) allows for the consistent and reproducible study of a variety of interrelated material and geometric variables. The objective of this study was to develop amodel for evaluating the mechanical performance of two sternal closure techniques, rigid plate fixation and wire cerclage. AFEA design study was performed to determine the effects of screw length, cortical thickness, and bone strength when using arigid fixation closure technique. The study analyzed the sternal separation and stresses that developed in the anterior cortical,cancellous, and posterior cortical. Lateral distraction and rostral-caudal (longitudinal) shear load cases corresponding tocoughing (350N) and lifting a 25lb weight were analyzed.

Materials and Methods: A two-level modeling approach was utilized in the design study. A global model consisting of two sternum halves was constructed, based on a patient’s CT scan, to determine forces surrounding the plate or wires and sternal separation. A local model consisting of a detailed bone model was constructed to determine stresses that developed in the plate, screws and bone.

 In the global model, half of the sternum was fixed, and a lateral distraction load of 350 N was applied to the opposite half. In the second loading case, a rostral-caudal (longitudinal) shear load of 175 N was applied to the sternum. The plate was modeled as grade IV pure titanium with an elastic modulus of 104.1 GPa, and the screw connections were modeled by tying the screw locations on the plate to the sternum. The global model consisted of two sternum halves, two X-shaped plates, and one L-shaped plate as shown in Fig. 1.

In the local model, lateral distraction and rostral-caudal shear loading cases were analyzed using the loads determined from the global model. Fixed displacements were applied until reaction forces matched the desired applied load. Screws were modeled as cylinders with a rough contact condition between the bone and screws. A tie constraint was applied between the plate and screws to simulate locking screws. In order to simulate patients with variable bone quality, three different elastic moduli were tested (6, 12 and 25 GPa for cortical bone; 0.04, 1.1 and 2.2 GPa for cancellous bone). The cortical thickness was adjusted to 0.50, 0.75 and 1.00 mm while the bone thickness was held constant at 12 mm. The screw length was also adjusted to investigate the effects of longer screws that engaged the posterior cortex.

Results and Discussion: In the global model, sternal separation from wire fixation was seen 48-75 times more than with rigid plate fixation (0.317mm vs. 0.0042-0.0066mm). Due to the low contact area between the wires and sternum, the stresses were concentrated on the wires, with a low distribution of load into the bone. Stress contour plots of rigid plate fixation revealed areas of high stress concentration at the anterior cortex and the screw neck. The stress in the anterior cortical bone was much higher than in the cancellous or posterior cortical bone, indicating that the majority of the load is carried by the anterior cortex. In addition, stress contour plots indicated the majority of the load in rigid plate fixation was carried by the four inner screws as shown in Fig 3. The plate design was improved by shifting the bridge locations of the plate to create a more uniform load distribution to the bone as shown in Fig. 4. In the longitudinal distraction loading case, the improved load distribution decreased the stress in the plate, the stress in the anterior cortex, and the gap in the sternum.

Increasing cortical thickness decreased stress in the plate, increased stress in the anterior cortex, and decreased the sternum gap. Increasing the bone strength decreased the stress in the plate, increased stress in the anterior cortex, and decreased the sternum gap. For the parameters tested, the screw length did not have a significant effect on the bone and plate stresses or the sternal separation.  In all scenarios with rigid plate fixation, there was less than 0.1mm of separation between the sternal halves, indicating good stabilization. Compared to lateral distraction, the shear loading in the rostral-caudal direction resulted in lower stresses and less sternal separation.

Screw length had minimal impact on sternal separation in this study. However, clinical experience shows posterior cortical purchase improves mechanical fixation. It is hypothesized that this result may be related to the modeling assumption that the screw locking mechanism essentially welds the screws to the plate, when in reality, the screw is able to rotate slightly. Future work includes mechanical testing utilizing dual density foams to determine if the screws are able to rotate under loading.

In order to achieve absolute stability a compressive preload is necessary. Compression of the fracture maintains close contact when the applied load is less than the compressive force and results in frictional forces that resist sliding displacement. Future work may include the design of a pre-bent plate which, when screwed to bone, applies a compressive force.

Conclusions: Rigid plate fixation resulted in less sternal separation and improved load distribution to the bone than wire cerclage. With rigid plate fixation, bone strength and cortical thickness had the greatest impact on the load distribution and sternal separation. FEA analysis revealed the majority of the load from rigid fixation is carried by the anterior cortex, which demonstrates the importance of using a bone model that accurately represents the bi-material composition of bone. This study demonstrates how FEA can be used to design a more stable rigid plate fixation by optimizing the load distribution to the bone. The model developed allows for an evaluation of rigid plate configurations to be performed in order to design rigid fixation systems with minimal sternal separation.  Future work to compare these results with bench testing is warranted.

References:

     [1]            Schimmer C, et al. Ann Thorac Surg. 2008 12;86(6):1897-904.

     [2]            Cohen, David J, et al. Ann Thorac Surg. 73.2 (2002): 563-8.