Biomechanical comparison of posterior short-segment fixation with or without intermediate screws for thoracolumbar burst fractures under normal and osteoporotic conditions: a finite element analysis
Article information
Abstract
Study Design
Finite element analysis.
Purpose
To investigate the biomechanical response of posterior short-segment fixation with or without intermediate screws at the index vertebra in osteoporotic thoracolumbar burst fractures using finite element analysis.
Overview of Literature
Spinal fixation in elderly patients with osteoporotic vertebral fractures is challenging because osteoporosis weakens the screw–bone interface, leading to screw loosening and loss of fracture reduction. Short segment fixation with intermediate screws has been proposed to reduce kyphosis recurrence and implant failure in unstable thoracolumbar fractures. However, the mechanisms by which intermediate screws enhance fixation strength in osteoporotic spines remain unclear.
Methods
Six finite element models of T12 burst fractures were developed to simulate short-segment stabilization under normal or osteoporotic bone conditions, with/without augmentation screws at the fractured vertebra. Spinal stiffness, implant stresses, and axial displacement/micromotion of the bony defect were measured and compared under mechanical loading.
Results
Osteoporotic models exhibited a greater range of motion (ROM) than normal bone. All six-screw constructs reduced ROM across all motions compared with traditional four-screw models. Osteoporotic fracture models gained greater benefit from intermediate screw augmentation at the fracture vertebra, which also lowered axial displacement/micromotion. In six-screw models, rod stress increased while pedicle screw stress decreased. Intermediate screws at fractured vertebrae produced similar changes in stress distribution across all fixation models, regardless of bone quality.
Conclusions
Our findings may facilitate implant selection for osteoporotic burst fractures, supporting the use of more rigid fixation six-screw constructs to reduce the risk of mechanical failure and postoperative re-collapse.
Introduction
Symptomatic osteoporotic vertebral thoracolumbar fractures are often indications for surgical intervention, yet the optimal approach remains controversial with no universally accepted algorithm [1,2]. In AO Spine type A and type B fractures characterized by severe comminution and marked kyphotic deformity, where corporeal reduction is crucial, posterior internal fixation can be a valuable option for achieving kyphosis correction and stabilization [3]. However, instrumentation for osteoporotic vertebral fractures remains challenging, primarily because the poor purchase of pedicle screws in osteoporotic vertebral bone increases the risk of hardware failure. To address this, six-screw short-segment posterior fixation, incorporating additional augmentation at the fractured vertebrae, has been proposed to reduce kyphosis recurrence and implant failure [4]. Nevertheless, the mechanisms and biomechanical effects by which intermediate screws at the fracture site enhance fixation strength in the osteoporotic spine remain poorly understood.
The purpose of this study was to evaluate and compare the immediate postoperative biomechanical performance of traditional four-screw fixation and six-screw stabilization with additional augmentation at the fractured vertebra under osteoporotic and normal bone conditions. The aim was to clarify the biomechanical efficacy and appropriate indications of posterior six-screw fixation for osteoporotic thoracolumbar fractures. To achieve this, finite element (FE) models representing normal, mild, and severe osteoporotic thoracolumbar spines were developed for both procedures in vertebral burst fractures and tested through simulations replicating ex vivo biomechanical experiments.
Materials and Methods
Subjects for biomechanics simulation
The procedure was approved by the Ethics Committee of PLA General Hospital and the patient provided written informed consent to participate in this study.DICOM files of three vertebrae and two intervertebral discs between T11 and L1 were obtained from a 29-year-old healthy man (61 kg and 175 cm) with no history of spinal injury, osteoporosis, or radiographic evidence of degeneration. Images were acquired using a 256-slice spiral computed tomography (CT) scanner with a slice thickness of 0.625 mm.
Finite element model of an intact normal thoracolumbar spine
The FE model of the intact thoracolumbar spine used in this study was developed and validated in our previous study [5]. CT images of intact thoracolumbar spine were imported into Mimics 21.0 (Materialise NV, Leuven, Belgium) to generate three-dimensional vertebral models of the T11–L1 region. These were then exported in STL format to Geomagic Studio 14.0 (3D Systems Inc., Rock Hill, SC, USA) for reverse engineering and fitted to a NURBS surface. The NURBS surface was subsequently imported into Hypermesh 2019 (Altair Engineering Inc., Troy, MI, USA) in IGES format for meshing, and the corresponding paraspinal ligaments and intervertebral discs were modeled.
The vertebral body was composed of cortical bone, cancellous bone, and endplate, with cortical bone and endplate thickness set at 1 mm. Intervertebral discs consisted of nucleus pulposus and annulus fibrosus, the latter being composed of annulus matrix and collagen fibers [6]. Collagen fibers of the annulus and major spinal ligaments were modeled as tension-only truss elements [7]. Facet joint articulation was simulated as a frictionless surface-to-surface contact with a minimum gap of 0.2 mm [8]. Mesh convergence was verified by refining the element size; the optimal size for the vertebral body was 1 mm. The final intact T11–L1 model contained 536,959 elements and 135,938 nodes (Fig. 1).
(A) T11–L1 three-dimensional finite element model of thoracolumbar segment. (B) A section of finite element model. (C) Finite element mesh of intervertebral disc. (D) Finite element mesh of nucleus pulposus. (E) Finite element mesh of annulus substance. (F) Finite element mesh of annulus collagen fiber.
Material properties
Normal material parameters of the thoracolumbar model were assigned according to previous studies [5]. Osteoporosis was modeled by reducing the elastic modulus of cancellous bone, cortical bone, and the vertebral endplate, while leaving other material parameters unchanged [9]. Material properties for normal bone mass, mild osteoporosis, and severe osteoporosis were sourced from prior biomechanical research [9]. Material properties and element types are summarized in Table 1.
Material properties and element types of spine tissues used in the finite element model
Establishment of fracture and fixation models
The fractured vertebra model was created in SolidWorks 2019 (Dassault Systèmes SolidWorks Corp., Waltham, MA, USA). Pedicle screws and rods were also modeled in SolidWorks. Conventional pedicle screws and connecting rods measured 6.0 mm and 5.0 mm in diameter, with a total screw length of 45 mm. Intermediate screws measured 5.5 mm in diameter and 30 mm in length. The following six fixation models were constructed using 4-screw or 6-screw constructs, respectively, based on the T12 burst fracture FE model (Fig. 2): (1) NO4: four-pedicle screw fixation for non-osteoporotic T12 burst fracture (normal bone mass); (2) NO6: six-pedicle screw stabilization with intermediate pedicle screws at the fractured vertebra for non-osteoporotic T12 burst fracture; (3) MO4: four-pedicle screw fixation for T12 burst fracture with mild osteoporosis; (4) MO6: six-pedicle screw stabilization with intermediate pedicle screws at the fractured vertebra for T12 burst fracture with mild osteoporosis; (5) SO4: four-pedicle screw fixation for T12 burst fracture with severe osteoporosis; and (6) SO6: six-pedicle screw fixation with intermediate pedicle screws at the fractured vertebra for T12 burst fracture with severe osteoporosis.
(A) Lateral projection of conventional four-pedicle screw fixation for T12 burst fracture. (B) Anterior-posterior projection of conventional four-pedicle screw fixation for T12 burst fracture. (C) Lateral projection of six-pedicle screw stabilization with intermediate pedicle screws at the fractured vertebrae for T12 burst fracture. (D) Anterior-posterior projection of six-pedicle screw stabilization with intermediate pedicle screws at the fractured vertebrae for T12 burst fracture.
Finite element analysis
A general static analysis method was used to solve the FE models. All degrees of freedom at the inferior surface of L1 were constrained. Because the screw and vertebral body are not fully bonded after surgery, a friction coefficient of 0.2 was applied between the pedicle screw and the vertebral body. To simulate loading of the internal fixation model, a combined load of 350 N and 7.5 Nm was applied to the superior surface of T11, as described in previous studies. FE analysis focused on three measurements: (1) range of motion (ROM) of T11–L1 to assess spinal stiffness; (2) von Mises stress in the implants and corresponding stress distribution (nephograms) to estimate load borne by the fixation devices; and (3) axial displacement/micromotion at the bony defect to assess the risk of correction loss.
Results
ROM of the internal fixation models
Global ROM values for the six fixation models under flexion, axial rotation, and lateral bending are shown in Fig. 3. The greatest coupled motions were observed during anterior flexion for all constructs. Compared with the normal bone mass model, global ROM increased across all loading conditions in the osteoporotic models, indicating that bone quality strongly influences the rigidity of posterior spinal constructs. ROM from T11 to L1 rose markedly in the severe osteoporosis models. Relative to conventional four-screw fixation, all six-screw stabilization constructs exhibited greater stiffness, particularly during anterior flexion. ROM during anterior flexion decreased by 2.02°, 2.25°, and 2.52° in NO6, MO6, and SO6, respectively, compared with NO4, MO4, and SO4, respectively, suggesting that augmentation with intermediate screws at the fracture vertebra provided additional stability even in impaired bone quality. However, no significant differences in ROM reduction were noted between the four- and six-screw constructs during extension, lateral bending, or axial rotation.
Range of motion (ROM) of T11–L1 segment for all investigated fixation models under the flexion, axial rotation, and lateral bending loading.
Maximum von Mises stresses on screws and rods
Maximum von Mises stress values for the pedicle screws and rods in all fixation models are presented in Figs. 4 and 5. The highest stresses on both rods and screws occurred in severe osteoporosis models, followed by mild osteoporosis and normal bone mass models, across all six loading positions. Flexion produced the largest mechanical stresses in all implanted models, ranging from 267.6 to 667.6 MPa for screws and 245.7 to 1,010.6 MPa for rods. Comparison of four-screw and six-screw constructs showed that rod stresses increased in six-screw models, whereas pedicle screw stresses decreased. Maximum von Mises stresses (MMS) on rods in osteoporotic bone were only mildly elevated compared with normal bone (8.53% to 11.09%), and osteoporosis severity did not significantly affect overall construct stress. Intermediate screws at the fractured vertebra altered stress distribution similarly in all fixation models regardless of bone quality. In the six-screw fixation models, MMS on T11 screws decreased by approximately 10% (10.08%–13.83%) during flexing loading. Conversely, MMS on L1 screws decreased substantially, especially during anterior flexion. NO6, MO6, and SO6 generated 45.65%, 41.17%, and 36.76% less stress on L1 screws under flexion compared with NO4, MO4, and SO4, respectively. Stress nephograms of pedicle screws are shown in Fig. 6 and demonstrate that maximum stresses were concentrated at the screw neck. Stress nephogram of rods on conventional four-screw and six-screw fixation constructs for all models under flexion load are shown in Fig. 7. The distribution analysis showed that high stresses were mainly concentrated on the rod segments between the upper and intermediate screws in six-screw constructs, whereas stresses were well distributed along the connecting rods between the upper and lower screws in conventional four-screw constructs.
The maximum von Mises stress of pedicle screws during all motions.
The maximum von Mises stress of rods during all motions.
Von Mises stress distribution on the screws for all fixation models during flexion. (A) Four-pedicle screw fixation for non-osteoporotic T12 burst fracture with normal bone mass. (B) Four-pedicle screw fixation for T12 burst fracture with mild osteoporosis. (C) Four-pedicle screw fixation for T12 burst fracture with severe osteoporosis. (D) Six-pedicle screw stabilization with intermediate pedicle screws at the fractured vertebrae for non-osteoporotic T12 burst fracture. (E) Six-pedicle screw stabilization with intermediate pedicle screws at the fractured vertebrae for T12 burst fracture in the context of mild osteoporosis. (F) Six-pedicle screw fixation with intermediate pedicle screws at the fractured vertebrae for T12 burst fracture in the context of severe osteoporosis.
Von Mises stress distribution on rods for all fixation models during flexion. (A) Four-pedicle screw fixation for non-osteoporotic T12 burst fracture with normal bone mass. (B) Four-pedicle screw fixation for T12 burst fracture with mild osteoporosis. (C) Four-pedicle screw fixation for T12 burst fracture with severe osteoporosis. (D) Six-pedicle screw stabilization with intermediate pedicle screws at the fractured vertebrae for non-osteoporotic T12 burst fracture. (E) Six-pedicle screw stabilization with intermediate pedicle screws at the fractured vertebrae for T12 burst fracture in the context of mild osteoporosis. (F) Six-pedicle screw fixation with intermediate pedicle screws at the fractured vertebrae for T12 burst fracture in the context of severe osteoporosis.
Postoperative axial displacement/micromotion of the bony defect in T12
Fig. 8 shows the maximum axial displacement at T12 in all fixation constructs under different spinal motions. Flexion resulted in the greatest maximum vertebral height reduction (absolute value), with an average decrease of 7.95 mm at T12, whereas extension resulted in the smallest height reduction (average 2.02 mm). Maximum axial displacement or micromotion of the bony defect during all motions was largest in the severe osteoporotic burst fracture models in each group, followed by mild osteoporosis and normal bone mass models. Comparison between four-screw and six-screw constructs showed that intermediate screws at the fractured vertebra significantly reduced axial displacement/micromotion of T12 in all fixation constructs during all motions. During flexion, the mean axial displacement at the bony defect was 1.6 mm, 1.8 mm, and 2.0 mm in NO6, MO6, and SO6, respectively, compared with higher values in NO4, MO4, and SO4, respectively.
The axial micro-motion of the bony defect in the fractured vertebrae for all investigated fixation models under the flexion, axial rotation, and lateral bending loading.
Discussion
Mechanical failure is more likely in patients with low bone density, including rod breakage, screw loosening, pullout, or migration [10]. Although percutaneous insertion of cannulated pedicle screws causes less muscle damage and blood loss than open surgery, it is biomechanically less stable and may result in early implant failure. Reported screw loosening rates range from <0.6% to 15% in non-osteoporotic patients and up to 60% in osteoporotic individuals [11]. Beyond implant failure, controversy persists regarding whether conventional four-screw short-segment instrumentation provides sufficient stability to maintain reduction, raising concerns of progressive kyphotic deformity [12,13]. Alanay et al. [14] and Wang et al. [15] reported that four-screw short-segment transpedicular instrumentation for type A thoracolumbar fractures carries a high risk of correction loss due to impaired bone resistance and absence of anterior support.
In recent decades, various strategies have been developed to enhance pedicle screw fixation in osteoporotic vertebral fractures, including lengthening of posterior instrumentation [16], use of expandable pedicle screws [17], bone cement augmentation [18], insertion of hydroxyapatite granules into the screw hole [19], and placement of intermediate screws at the fractured vertebra [20,21]. Each technique has advantages and drawbacks, and ongoing research aims to minimize complications related to osteoporosis.
Currently, polymethyl methacrylate cement (PMMA) augmentation is the most frequently reported method for improving screw stability [22,23]. However, PMMA augmented screw procedures, just like vertebroplasty and kyphoplasty, carry the risk of adverse events, the most frequent being cement leakage and the potentially fatal complication of pulmonary cement embolism [24,25]. Expandable pedicle screws can improve screw–bone purchase but may damage vertebral bone during screw removal, and medial or inferior expansion can injure the spinal cord or nerve roots [26].
Placement of intermediate screws at the fractured vertebra is another common technique to improve the fixation of the four-screw construct when the pedicles are intact and the screws can be inserted safely. Augmentation at the fracture level theoretically increases construct stiffness, reducing instrumentation failure, screw pullout, and postoperative loss of correction, while preserving adjacent segment motion and maintaining sagittal alignment comparable to long-segment stabilization [20,21,27]. Nevertheless, the mechanism by which six-screw systems enhance anchoring of screw-rod constructs in osteoporotic spines is not fully understood. Moreover, pedicle screw placement at the fractured vertebra is not entirely a benign, harmless procedure, especially in elderly patients with osteoporosis. Therefore, clarifying the strongest advantages, appropriate indications, and biomechanical effects of six-screw short-segment fixation are key research imperatives.
Our simulation demonstrated that global ROM increased in all states of motion in the osteoporosis models compared with the normal bone mass model, indicating that osteoporosis negatively affects the insertion torque of pedicle screws. Liao [28] reported a 3%–25% increase in implant ROM under osteoporotic conditions compared with normal bone. In the present study, global ROM of the MO4 and SO4 models increased by 47% and 20%, respectively, compared with non-osteoporotic models. The application of intermediate screws reduced ROM in extension, lateral bending, and axial rotation across all models. The greatest reduction in global ROM (2.52°) was observed in the six-screw stabilization for severely osteoporotic models, indicating that intermediate screws in poorer bone quality bear more load than in burst fractures. These findings suggest that softer osteoporotic vertebrae may benefit more from rigid fixation with six-screw constructs.
In the present study, maximal von Mises stress on screws decreased during nearly all states of motion in the six-screw stabilization models. Intermediate screws at the fractured vertebra carried part of the transmitted load, thereby redistributing von-Mises stress in the fixation construct, especially during flexion. Compared with four-screw constructs, screw stress in the six-screw models decreased by approximately 10% (10.08%–13.83%) at T11 and by approximately 40% (36.76%–45.65%) at L1. Augmentation screws at the fracture level thus reduce screw stress, lowering the risk of breakage and loosening, particularly at the level below the fracture. Interestingly, our findings extend previous work by showing that osteoporosis severity has no significant influence on construct stress and that all six-screw fixation models reduced von Mises stress on screws regardless of bone quality [28,29]. Stress distribution analysis confirmed that the maximal pedicle screw stress was concentrated at the root of the screws under all loading conditions, consistent with clinical observations that most screw fractures occur at the neck.
Consistent with increased pedicle screw stress, osteoporosis also negatively affected axial displacement or micromotion at the injured vertebra, which may contribute to postoperative re-collapse. Comparison of constructs revealed significantly lower axial displacement/micromotion in the six-screw models during all motions. These results indicate that re-collapse of injured vertebrae may be prevented in osteoporotic and unstable fractures treated with six-screw fixation constructs. For osteoporotic burst fractures, bilateral intermediate screws should therefore be strongly considered to reduce early post-traumatic kyphosis and loss of reduction rather than relying on conventional four-screw fixation.
Some limitations of this study should be noted. First, fracture morphologies and material properties in the FE model were simplified and may not fully reflect clinical reality, potentially altering the biomechanical responses of tissues and constructs. Second, although FE analysis offers many advantages for biomechanical assessment, the computational models developed in this study did not include spinal musculature, which contributes to spinal stability and stiffness during motion.
Conclusions
Augmentation with intermediate screws at the fractured vertebra provides stiffer fixation even under osteoporotic conditions. Osteoporotic burst fractures appear to benefit particularly from rigid stabilization with six-screw constructs. These findings may help guide implant selection for osteoporotic burst fractures by supporting the use of six-screw constructs to reduce the risk of mechanical failure and postoperative re-collapse.
Key Points
Intermediate-screw augmentation at the fractured vertebra provides stiffer fixation than conventional four-screw models.
In osteoporotic fracture models, intermediate-screw augmentation at the injured vertebra conferred greater biomechanical advantage than in normal-bone models.
Our findings may help guide implant selection for osteoporotic burst fractures: six-screw constructs provide markedly stiffer fixation, lowering the risk of mechanical failure and late vertebral recollapse.
Notes
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Author Contributions
Conceptualization: CX, XMZ, CZ. Methodology: CX, XMZ, CZ, HJC, XDB. Investigation: CX, XMZ, CZ, HJC, XDB. Data curation: CS, FG. Formal analysis: CS, FG, CX, XMZ. Writing–original draft: CX, XMZ. visualization: CS, FG. Writing–review & editing: CZ, XDB. Project administration: CX, HJC. Supervision: CZ, XDB. Validation: CZ, XDB. Resources: CZ, XDB. Final approval of the manuscript: all authors.