DOI: https://doi.org/10.21203/rs.3.rs-2091689/v1
Previous biomechanical studies have compared posterior lumbar interbody fusion (PLIF) and transforaminal lumbar interbody fusion (TLIF), however the cages used in TLIF/PLIF surgery are different. Therefore, comparing the two surgical procedures with the same fusion cages is more reflective of the real differences between the two procedures. This study was to compare the biomechanical effects between PLIF and TLIF with placing two straight-shaped cages using finite element analysis.
A previously validated intact L3-L5 lumbar spine finite element model was modified to simulate two straight-shaped cages PLIF and two straight-shaped cages TLIF. A moment of 7.5 N-m with a compressive preload of 400 N was applied on the L3 vertebra to test the range of motion (ROM) and stress.
There were no significant differences in the ROM between PLIF and TLIF with less than 1 degree for all loading cases. We found that stress of cage, and stress of L4 endplate were high in PLIF, and stress of pedicle screw was high in TLIF. Similar bone graft stress was found in PLIF and TLIF.
The biomechanical result indicated that both TLIF and PLIF can acquire similar ROM and bone graft stress. PLIF increases the stress of cage and L4 inferior endplate, and pedicle screw stress was high in the TLIF model. The findings of our study need to be verified through further clinical studies that explore differences between the two methods.
Lumbar interbody fusion has been approved to be an effective treatment for patients with lumbar degenerative disease. Instrumented transforaminal lumbar interbody fusion (TLIF) are clinically used to obtain a solid arthrodesis so as to treat degenerative spinal diseases, which was revealed as an alternative to posterior lumbar interbody fusion (PLIF) and these two are the most widely used ones.1 Comparing to PLIF, TLIF has a more lateral approach to lumbar disc, and only unilateral facetectomy is needed. No need to retract nerve root in TLIF, so nerve root injury is reduced when exposing disc with a lateral approach. The safety and efficacy of TLIF surgery have been assessed by clinical and biomechanical studies.2-5 In the instrumented PLIF/TLIF surgery, two straight-shaped cages has been widely recommended as a standard treatment to treat lumbar degenerative disease, although there are other cages in the surgery.6
Finite element analysis has been widely applied to evaluate the biomechanical characteristics of using different cages in different lumbar fusion surgeries.7 To our knowledge there are some biomechanical literatures comparing PLIF and TLIF, however the cages used in TLIF/PLIF surgery are different3. Therefore, comparing the two surgical procedures with the same fusion cages is more reflective of the real differences between the two procedures. It remains to be determined whether the placement of two straight-shaped cages could achieve an acceptable levels of stability, as well as similar levels of cage migration, cage subsidence, mechanical failure and fusion ability in TLIF/PLIF surgery. In this study, biomechanical effects between placing two straight-shaped cages in TLIF/PLIF were compared to provide information for guiding clinical decision-making, preventing of possible postoperative complications, and improving the surgery method.
The intact L3-L5 lumbar spine and finite element models (FEMs) using two types of surgical techniques, two straight-shaped cages TLIF and two straight-shaped cages PLIF, were simulated.
FEM of the intact lumbar spine. A previously constructed non-linear FE model of the intact L3–L5 lumbar spine was used for this study, which was validated in our previous study.3 This intact L3–L5 FE model included vertebral bodies, posterior bony elements, intervertebral discs and lumbar spinal ligaments. Three vertebrae, two intervertebral discs and seven ligaments were included in the intact model. Each vertebrae was composed of a 1.0 mm cortical shell (including endplate) surrounding a cancellous inner core. Each intervertebral disc was composed of a nucleus pulposus surrounded by an annulus ground substance comprising six fiber layers. The nucleus pulposus was modeled as a linearly elastic fluid element. The annulus fibrosus of the disc was designed to be isotropic and hyperelastic. The elastic strength of fiber stiffness decreased proportionally from the outside layer to the inside layer.8 The major ligaments of the lumbar region, including anterior longitudinal, posterior longitudinal, flavum, supraspinous, interspinous, intertransverse and capsular ligaments were incorporated in this model and were active in tension only. The material properties used in the model are listed in Table 1. The interactions between facet joints were considered as surface-to-surface contact elements with a friction coefficient of 0.1 and gaps of 0.5 mm.9
| Material properties | Young's Modulus(E:MPa) | Poisson Ratio(µ) |
|---|---|---|
| Cortical bone | 12000 | 0.3 |
| Cancellous bone | 100 | 0.2 |
| Posterior bone | 3500 | 0.25 |
| Nucleus | 1 | 0.499 |
| Anulus ground substance | 4.2 | 0.45 |
| Anulus fibers | 357.5–550 | 0.3 |
| Endplate | 3000 | 0.25 |
| Ligament | ||
| Anterior longitudinal | 20 | 0.3 |
| Posterior longitudinal | 20 | 0.3 |
| Transverse | 59 | 0.3 |
| Ligamentum flavum | 19.5 | 0.3 |
| Interspinous | 12 | 0.3 |
| Supraspinous | 15 | 0.3 |
| Facet capsular | 32.9 | 0.3 |
| Spinal instrumentation (titanium alloy) | 110000 | 0.28 |
| Spinal cage | 3600 | 0.25 |
FEM of the instrumented lumbar interbody fusion. The intact L3-L5 FE model was modified to simulate PLIF and TLIF at the L4–L5 level, with two straight-shaped cages and adding posterior fixation with four pedicle screws (55 mm long, 6.5 mm diameter) and two connecting rods (45 mm long, 6 mm diameter). Laminectomy and partial discectomy were performed on the L4-5 segment in the PLIF model, and unilateral superior and inferior articular processes of 1 facet joint were resected at the level of the L4-5 segment in the TLIF model. The screws and rods, bone-screw interfaces were simulated by assigning these interaction to be bonded. The simulated surgical procedure included the resection of left facetectomy in TLIF, laminectomy in PLIF, the disc and cartilaginous endplates at L4–L5. Two 25 mm long O.I.C. (Stryker, South Allendale, NJ) cages (polyether ether ketone; PEEK) packed with autologous cancellous bone grafts were placed at L4-L5 intervertebral space through distraction of the segment (Figs. 1 and 2). The cage and vertebral body interface was simulated by surface-to-surface contact, modeling the early stage of postoperative TLIF surgery. These contact surfaces transmitted compression force but not tension. A higher coefficient of friction (0.8) was set at the cage-vertebra interface, which can mimic small teeth on the cage that could prevent cage slipage.9
Boundary and loading conditions. For all FE models, the bottom of the L5 vertebra was constrained. An axial preload of 400 N was imposed on the superior surface of L3 to simulate the weight of the human upper body.9 A moment of 7.5 Nm was imposed on superior endplate of the L3 vertebra to simulate six different physiological motions, including flexion, extension, lateral bending, and axial rotation motion modes.
Range of Motion. There was a significant reduction in ROM at the level of fusion (L4–L5) for TLIF and PLIF models compared to the intact model. There was a large difference in flexion, with a change in ROM of -89.59%, and − 96.47% for the L4-5 segment for the TLIF and PLIF models, respectively, compared to the 5.38° obtained for the intact model. The difference between the six motion modes was less than 1 degree in ROM between the two models (Fig. 3). The findings suggested that TLIF and PLIF models had similar levels of biomechanical stability.
Cage Stress. We found that the maximal von Mises stress exerted on the cage-endplate interface of the cage was high in the PLIF model, except under r-rotation (Fig. 4). The maximum stress of the cage was 116.7 MPa under flexion in the PLIF model, and the minimum stress was 41.39 MPa in the TLIF model under extension. The greatest difference in cage stress was 116.7 MPa in the PLIF model and 79.35 MPa in the TLIF model under flexion. The distribution contour of von Mises stress of the cage under flexion and extension are shown in Fig. 5.
Bone Graft Stress. In extension, l-rotation and l-bending, the maximal von Mises stress in the bone graft was high in the TLIF model (Fig. 6). In flexion, the PLIF model induced the highest stress in the bone graft (4.62 MPa). In extension, the PLIF model induced the lowest stress in the bone graft (0.65 MPa). In extension, there was the greatest difference in bone graft stress in the PLIF model and the TLIF model (0.65 MPa vs. 3.63 MPa).
Pedicle Screw Stress. Pedicle screw stress was was high in the TLIF model, except under extension and l-bending (Fig. 7). The largest pedicle screw stress was 190.8 MPa in the TLIF model under flexion, and the minimal pedicle screw stress was 86.2 MPa in the PLIF model under r-rotation. In flexion, there was the greatest difference in pedicle screw stress in the PLIF model and the TLIF model (114.7 MPa vs. 190.8 MPa).
L4 inferior endplate stress. Higher stresses were noted at the L4 inferior endplate under the conditions of flexion, extension, r-rotation, and lateral bending with PLIF model compared with TLIF model (Fig. 8). The largest endplate stress was 37.04 MPa in the PLIF model under flexion, and the minimal endplate stress was 16.46 MPa in the TLIF model under extension. The distribution contour of von Mises stress of L4 lower endplate under flexion and extension are shown in Fig. 9.
Many studies have reported of satisfactory outcomes in patients with degenerative spinal diseases who have undergone lumbar interbody fusion.10,11 Many biomechanical studies have proven that TLIF or PLIF with one cage can acquire effective spinal stability similar to that achieved using two cages, however, one cage TLIF has the limitations of leading to a slight increase in screw stress and cage-endplate interface comparing to two cages TLIF.12,13 Previous studies have compared PLIF and TLIF, however the cages in the two methods were different, and different cages can lead to inaccurate results.3,14 In the current study, we first used the finite-element method to analyze differences between the PLIF and TLIF techniques with the same two cages.
In our study, instrumented TLIF and PLIF were found to be able to achieve effective and similar spinal segment stability, compared with the intact spine model. We must pay attention to these results because insufficient stabilization and increased motion can elevate implant stresses that may lead to mechanical failure and a lack of fusion under repetitive physiological loading.
Cage dislodgement has been reported to be a complication of transforaminal lumbar interbody fusion, especially when spinal construct stability is not sufficient.15 Aoki et al16 conducted a clinical study and found that TLIF performed using a bullet-shaped cage could increase the cage dislodgement, since the geometrical structure of a bullet-shaped cage is less able to prevent migration than a box-shaped cage. In our study, we found the maximal cage stress produced by PLIF was high, except under r-rotation, increasing the possibility of cage dislodgement and cage collapse.
In our study, we found that the maximal stress in the bone graft in TLIF was high in extension, l-rotation and l-bending. TLIF may offer less stress-shielding between intervertebral bodies, resulting in the fusion bone receiving greater stress. The stress in the bone may be helpful for the fusion process in accordance with Wolff’s law. Zhang et al.17 conducted a finite element study on instrumented PLIF using one spacer and two spacers. They found that one spacer PLIF resulted in better bone formation than two spacer PLIF under physiological loading.
Screw breakage has been reported in 12.4% of the patients treated with lumbar interbody fusion.18 In our study, TLIF produced greater pedicle screw stress than PLIF except under extension and l-bending. This may be related to the TLIF procedure with complete resection of one side of the articular processes. The TLIF model placed more stress on the pedicle screw, which may explain why the breakage of pedicle screw often occurred in this model. Under flexion motion, we found the largest pedicle screw stress to be induced by TLIF. The results of our study indicated that patients should exercise caution in lumbar flexion movements when treated with TLIF surgery in order to avoid pedicle screw breakage.
Cage subsidence is a complication arising from spinal interbody fusion surgery, which can result in a decrease of normal intervertebral height, and can cause nerve oppression.19 Stress on the L4 lower endplate was higher in PLIF. The largest stress on the endplate was found to be exerted in the PLIF model under flexion loading. This indicated that cage subsidence could easily occur as a result of PLIF under flexion motion.
There are some limitations in our study. First, the material properties used for the finite element analysis of the lumbar spine did not exactly model lumbar tissue. For example spinal ligaments show nonlinear behavior, viscoelasticity of intervertebral disc, and spinal orthotropic characteristics. Second, our study did not simulate spinal muscle contraction, and perhaps it can induce certain errors that affect the final outcome. Third, our study did not consider differences in the mineral density of the lumbar spine between patients, such as patients with osteoporotic spinal bone and other positions and sizes of spacers. Therefore, further precise modelling studies need to be performed in future to obtain accurate conclusions.
The biomechanical result indicated that both TLIF and PLIF can acquire similar ROM and bone graft stress. PLIF increases the stress of cage and L4 inferior endplate, and pedicle screw stress was high in the TLIF model. The findings of our study need to be verified through further clinical studies that explore differences between the two methods.
.
Ethics approval and consent to participate
The present study was approved by the Ethics Committee of Yongding Hospital of Suzhou. Informed consent obtained from each participant was written. All protocols are carried out in accordance with relevant guidelines and regulations.
Consent for publication
Not applicable.
Competing interests
The authors declare that they have no conflict of interest.
Funding
Not applicable.
Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Readers can access the data and material supporting the conclusions of the study by contacting Hao Xu at [email protected].
Authors' contributions
Hao Xu designed the study. Hao Xu and Junjie Wu conduct the experiment. Yanwen Hu wrote the main manuscript text and analyzed the data.
Acknowledgements
We thank Chen Bo, MD (Traumatology and Orthopaedics Institute of Shanghai Jiaotong University), for his help in the experimental studies.