To test the mechanical stability of spinal constructs in relation to various lordotic angles, the biomechanical study was performed on a synthetic model simulating the spinal motion segment, which is composed of two ultrahigh molecular weight polyethylene (UHMWPE) blocks modified from Cunningham et al [22]. In the experiment, there was an anterior wedged cage placed between the UHMWPE blocks. One group used monoaxial screws linked with pre-bent rods of various angles to form the spinal constructs, and the other group adapted polyaxial screws to form the constructs.
Experiment: The synthetic model was posteriorly instrumented with four titanium polyaxial or monoaxial pedicle screws (6.0-mm diameter, 40-mm length; Mathys Co, Bettlach, Switzerland), and connected with two rods (5.0-mm diameter) to build the testing constructs. The torque of insertion for each pedicle screw was set at 4.34 Nm, and the design of polyaxial screws was allowed ± 25° of motion for the screw-rod mounting. For each construct, the rod contours were set at four different lordotic configurations (0°, 7°, 14°, and 21°) for the alignment between the screws. A rod of 7°, 14°, or 21° lordotic configurations was pre-bent each based on the curvature, which was formed between the lines of two screw insertion points and the center of a circle (Figure 1). A single SynCage (12° wedged, 18-mm height and 22-mm length; Mathys Co, Bettlach, Switzerland) was positioned centrally (20 mm from the anterior midline and lateral edges of the UHMWPE block). The serrated surfaces above and below the SynCage offering high friction on the interface ensured no sliding of the cage on the block. Once the UHMWPE blocks, as the vertebral endplates, contacted with the upper and lower edges of posterior margin of the SynCage, we tightened nuts for the screw-rod mounting. The torque of coupling the pre-bent rod onto the monoaxial or polyaxial screw head was set at 9.81 Nm. In the monoaxial screws group, the technique of screw-rod mounting was performed along the axis of screw body with a rod adapting to the screw head. In the polyaxial screws group, the technique of screw-rod mounting was performed with a screw holder perpendicular to the rod, while securing the screw head tightly. Consistent biplanar interpedicular distances (40 mm in the coronal plane and 38 mm in the sagittal plane) and a uniform lever-arm distance (40 mm) between the point of anterior load application and the center of the posterior rods were maintained. The segmental lordosis of two blocks was rechecked as 0°, 7°, 14°, or 21° on each construct before compression, flexion, lateral bending and torsion tests. If the lordotic anlgle lost during testing, we applied the construct again.
A uniaxial strain gauge (KFG-1-120-C1-11L1M2R, KYOWA Electronic Instruments Co., Tokyo, Japan) was attached to the posterior part of the cage, and another strain gauge was affixed to the middle of left connecting rod to measure the surface strains of cage and rod separately during the subsequent mechanical testing (Figure 1). During compression and flexion, the surface strain data were obtained from the implants with the sagittally aligned gauges. During lateral bending, the data were obtained from the implants with the coronally aligned gauges [23]. The surface strains on the bent rod and the SynCage were correlated with the applied axial loads. For example, under compression and flexion, we calculated a higher strain data obtained from the posterior surface in the bent rods than in a straight rod, which determined the bending stress of the whole contoured rod during testing. Two prescale Fuji super low pressure films, with a sensitivity range from 0.6 to 2.5 MPa, (Fuji Photo Film Co., Tokyo, Japan) were placed between the contact surfaces of the SynCage and the two UHMWPE blocks to measure the contact area of the cage. The film became red in the compressed area. The red area was scanned with a scanner (HP Scanjet 3570C, HP Co., Palo Alto, CA) to calculate the contact area. The contact ratio was then obtained by dividing the original area of the film with the contact area (Figure 2).
There were four testing steps: compression, flexion, left lateral bending and torsion that were similar to the protocols used by Pavlov et al [21]. For each construct set at each of the four lordotic configurations (0°, 7°, 14°, and 21°), a compressive load of 0–300 N was applied at a displacement rate of 25 mm/min using a testing machine (Q test 10, MTS system Co., Eden Prairie, MN). In the flexion test, a 5 Nm bending moment was loaded 43.5 mm anterior to the center of rotation to make the construct flexed. In the left lateral bending test, a 5 Nm bending moment was loaded to make the construct bend to the left. In the torsion test, 5 Nm of torque was loaded to twist the construct counterclockwise. A cross-link was added transversely across 2 rods to ensure the fixation during torsion. An electric goniometer SG65 (Biometrics Ltd., Gwent, UK) was used to detect the angular motion occurring in the mechanical testing. Every testing step was performed for 8 cycles and the first 3 cycles served as conditioning cycles. A mean of data retrieved from the following 5 cycles and 3 repeated testing experiments were employed for further analysis. Inputs concerning the construct stiffness, surface strain gauges, interface contact ratio performed on the MTS machine was synchronized through a multi-channel signal-conditioning amplifier (InstruNet, GWI, Somerville, MA), and all the data were collected to a personal computer. Compression stiffness of the construct was computed as a ratio of applied load (in Newtons) to linear deformation of the construct (in millimeters). Other rotational stiffness was presented as a ratio of applied torques (in Nm) to linear rotational displacement (in degrees). The surface strain (in microstrain) was recorded at peak load during the fourth loading cycle. Each group involving monoaxial or polyaxial pedicle screws linked with different rod contours has 3 data points for the statistical analysis.
Statistics
All data were shown as the mean ± standard deviation. Multiple-factor analyses of variance (ANOVA) on the construct stiffness, rod strain, cage strain, and percentage of contact area were conducted to find differences existed among the two instrumentation methods (monoaxial screws with a cage support, and polyaxial screws with a cage support) and the four lordotic rod contours (0°, 7°, 14°, and 21°). The statistical difference between the monoaxial and polyaxial screws groups was calculated by independent t-test. If the p value was less than 0.05, it was defined as significant difference. Statistical analyses were performed with SPSS, Version 10.0 (SPSS, Inc., Chicago, IL).
