Biomechanical comparison of posterior intermediate screw fixation techniques with hybrid monoaxial and polyaxial pedicle screws in the treatment of thoracolumbar burst fracture: a finite element study

A finite element model including 7 vertebrae and 6 discs between T9 and L3 of the spine obtained from 64 spiral computed tomography (CT) images of a 40-year-old healthy male (65 kg and 175 cm) without a history of spinal injury, osteoporosis, and radiographic evidence of degeneration was reconstructed and analyzed using finite element analysis software [6, 22, 23]. The CT images were scanned and imported into Mimics 10.0 (Materialise, Belgium). The surface model was then exported into Rapidform 2006 (INUS, Korea) to generate and enhance the quality of the solid model. Eventually, the model was imported into Abaqus 6.9 (Simulia) for meshing. Each vertebral body consisted of cortical bone and cancellous bone, and each vertebral disc was composed of nucleus pulposus, annulus fibrosus, and endplates. Posterior elements were built separately from the vertebral bodies. Based on a Boolean operation, the lower half of the T12 segment was resected, and the structure of the posterior part was reserved to establish a finite element model of an unstable thoracolumbar fracture. Surface-to-surface contact was defined between articulation facets. We have built the intact normal spine model and fractured spine model. The intact spine model without implants had a total of 20,924 nodes and 72,055 elements which including 48,099 tetrahedron elements, 5212 hexahedral elements, 1236 spar elements, and 17,508 shell elements (Fig. 1). We have used a truss element to replace the ligament, and the thickness of the shell element was 0.4 mm.

This was a prospective study to assess the biomechanical characteristics of different posterior intermediate screw fixation techniques with hybrid Mps and Pps used in thoracolumbar burst fracture model. Fixation models were described as MMM, PPP, PMM, MPM, MMP, MPP, PMP, and PPM (Figs. 2 and 3) which may be used in the clinical practice. Surface-to-surface contact was defined between articulation facets. The element types, material properties, ligamentary cross-sectional area, and implants are shown in our previous study [6].

The screw diameter was 6 mm, and the screw length was 45 mm. The pedicle screws in the current study included Mps and Pps. The constraint was defined between polyaxial pedicle screw heads and shafts. However, a load limitation was defined. Surface-to-surface contact was defined between polyaxial pedicle screw heads and shafts. The screw tilt (the maximal deviation of the long axis of the screw away from perpendicular to the longitudinal rod) was 25°, the static torque was 8 Nm which meant that the polyaxial pedicle screw heads will move relative to the shafts when the torque between the heads and shafts reached 8 Nm. These parameters are referred to as the polyaxial pedicle screw of Sofamor. The top surface of T9 was applied by a pure moment of 10 Nm combined with a pre-compressive load of 150 N, the inferior endplate of L3 was constrained in all degrees of freedom (Fig. 4). To validate the rationalities of the models, including model simplification, material properties, boundary conditions, and loads, a moment of 10 Nm and a compressive load of 150 N were applied to the reference point. The range of motion (ROM) among different models was compared in our previous study [6]. There is little difference between the models. Therefore, the models in the present study are effective for further analyses.

We measured the ROM of the intact spine model T9–L3 under flexion, extension, left/right lateral bending, and left/right axial rotation and then applied ROM displacement loading to the four fixation models. The redistributed ROM of the T11–L1 segment, the largest maximal VMS of the pedicle screws and rods, and IDPs of the adjacent segment under displacement loading were evaluated. The procedure was approved by the ethics committee of Xinqiao Hospital, and the patients provided written informed consent to participate in this study.

We used SPSS 15.0 software (SPSS Inc., Illinois, USA) to perform all statistical analyses, and P < 0.05 was considered significant (two-tailed). The independent sample t test was used to compare the means.

The fixation models presented with a decreased ROM than the intact normal spine model (Table 1). The redistributed ROM of the MMM model in flexion, extension, and axial rotation was the smallest. The redistributed ROM of the fixation models with Pps fixed at the lowest segment was twice of the other fixation models in flexion and extension (Fig. 5). There were significant differences between the fixation models with Pps fixed at the lowest segment or not in the flexion (8.0 ± 0.1°, 3.5 ± 0.9°, P = 0.002) and extension (6.7 ± 0.1°, 3.1 ± 0.8°, P = 0.003), no significant differences in the axial rotation (4.7 ± 0.7°, 3.1 ± 1.3°, P = 0.073) and lateral bending (3.3 ± 0.3°, 2.6 ± 0.5°, P = 0.058).

The largest and smallest value of maximal VMS of a pedicle screw was 382.6 MPa in the PMP model and 136.9 MPa in the PPP model, respectively (Table 2). The largest value of maximal VMS of a pedicle screw was located at the lowest pedicle screws when Mps are fixed at the lowest segment. The largest and smallest value of maximal VMS of the rod was 439.9 MPa in the MMM model and 341.7 MPa in the PPP model, respectively. The largest value of maximal VMS of the rods was decreased when more Pps are fixed at the models (Table 2), but there were no significant differences between the fixation models with two Pps fixed and models with four Pps fixed (429.2 ± 10.3, 409.8 ± 15.5, P = 0.145).

Maximal IDPs of the adjacent segment was observed in the lateral bending. Maximal IDPs of the upper adjacent segments were all larger than those of the lower adjacent segments (Table 3). The maximal IDPs of the fixation model with Mps fixed at the lowest segment were larger than the other models in flexion and extension (Fig. 6). With regard to the upper adjacent segments, there were significant differences between the fixation models with Mps fixed at the lowest segment in the flexion (1.9 ± 0.1, 1.3 ± 0.1, P = 0.000) and extension (2.2 ± 0.1, 1.8 ± 0.1, P = 0.001), no significant differences in the axial rotation (1.3 ± 0.2, 1.2 ± 0.1, P = 0.235) and lateral bending (2.5 ± 0.3, 2.4 ± 0.3, P = 0.902). With regard to the lower adjacent segments, there were significant differences between the fixation models with Mps fixed at the lowest segment or not in the flexion (0.7 ± 0.1, 0.4 ± 0.1, P = 0.000) and extension (1.0 ± 0.2, 0.6 ± 0.1, P = 0.017), no significant differences in the axial rotation (0.8 ± 0.1, 0.9 ± 0.2, P = 0.072) and lateral bending (1.5 ± 0.1, 1.5 ± 0.1, P = 1.000).