Background
Burst fractures account for approximately 20% of thoracolumbar fractures, and occur due to an axial loading force that results in failure to support the anterior and middle column [
1,
2]. Surgery is usually indicated when there is a severe deformity, and/or neurologic deficit. Whether anterior or posterior surgery is the most effective treatment for burst fractures remains controversial. Some authors advise anterior surgery to remove retropulsed fragments [
3,
4], but posterior surgery is popular because it is an easier approach and allows clearance of the spinal canal by ligamentotaxis.
One-above and one-below posterior short-segment instrumentation with fusion has been widely used for unstable thoracolumbar burst fractures for the past 3 decades [
5]. Pedicular instrumentation enables kyphotic correction, indirect reduction of canal encroachment, and early mobilization. However, this method has a high rate of implant failure, and early loss of reduction because of loss of anterior support [
6]. Over the past decade, some studies have demonstrated that augmentation of the fractured vertebra with absorbable bone cement could enhance fracture union and prevent implant failure. Liao et al. [
7] and Korovessis et al. [
8] demonstrated that injectable calcium sulfate cement or injectable calcium phosphate cement used as a transpedicular grafting material in thoracolumbar fractures could obtain clinical and radiographic results equal to autogenous cancellous bone graft.
In recent years, biomechanical and clinical studies have suggested that two additional screws inside the fractured vertebra could improve stability and provide better kyphotic correction. In a biomechanical study, Norton et al. [
9] showed that two additional screws in the fractured vertebra (a six-screw construct) of an unstable thoracolumbar burst fracture could increase the stiffness of the implant, and reduce stress on each pedicle screw, as compared to a four-screw construct. However, there have been no studies examining the effect of augmentation by a combination of screws and bone cement at the fractured vertebra, and comparing the effect of this fixation method with other types of posterior short-segment instrumentation for thoracolumbar burst fractures.
We hypothesize that posterior short-segment instrumentation with fractured vertebra augmentation by a combination of vertebroplasty and two additional screws can provide a stronger construct than other types of posterior short-segment instrumentation for thoracolumbar burst fractures. In the current study, we established a finite element (FE) model of thoracolumbar burst fractures, and four posterior short-segment fixation methods were tested. The purpose was to determine the most optimal methods for the treatment of thoracolumbar burst fractures.
Discussion
Surgical management of a thoracolumbar burst fracture varies according to many factors. Fracture morphology, neurologic status, and surgeon preference all play major roles in deciding the surgical approach. An anterior approach to the fracture site can directly decompress the neural elements, and the reconstruction ca be performed by simultaneous iliac bone graft and plating. Hitchon et al. [
15] and Sasso et al. [
16] reported that the anterior approach was superior to the posterior approach in its ability to maintain the kyphotic correction at final follow-up, but both approaches achieved similar in clinical results. A major disadvantage of anterior surgery is potential donor site morbidity from harvesting iliac tri-cortical bone graft. In addition, the cost of the anterior approach is generally greater than that of the posterior approach. We usually use a posterior approach for the repair of thoracolumbar burst fractures, which is one of the reasons that led to the current study.
Traditional one-above and one-below short-segment posterior instrumentation has the potential for early implant failure and re-kyphosis due to lack of anterior support of the defect inside the injured vertebra. Several techniques, such as reinforcement with additional screws at the fractured level, or augmentation with any kind of bone cement to fill the defect of the fractured vertebra, have been proposed to improve the stability of the posterior instrumentation construct, and thus prevent implant failure and enhance fracture union [
17‐
20]. However, no studies examined reinforcement of a fractured vertebra with a combination of bone cement with two additional fracture-level screws. Several biomechanical studies have suggested that reinforcement with fracture-level screws could improve the biomechanical stability of the construct [
21‐
23]. Clinical studies have also suggested that reinforcement with additional screws at the fractured level can provide better kyphotic correction, more effectively restore fractured vertebra height, and allow earlier ambulation for patients with thoracolumbar burst fractures [
24,
25]. On the contrary, authors have also claimed that a fractured vertebra augmented by absorbable bone cement followed by a posterior short-segment construct can provide satisfactory clinical results, with a low implant failure rate of 0% to 5% [
19,
20]. Xu et al. [
26] designed a FE model of a thoracolumbar burst fracture, and demonstrated that vertebroplasty inside the fractured vertebra can significantly reduce the stresses of the pedicle instrumentations and spine to prevent kyphotic correction loss and implant failure. In a clinical study, Liao et al. [
27] showed that two screws placed inside the fractured vertebra with short-segment instrumentation was associated with shorter surgical time and less implant failure as compared to augmentation with injectable calcium sulphate/phosphate cement following posterior short-segment instrumentation.
In the current study, the four fixation models all demonstrated decreased ROM in all directions as compared with the intact spine model. ROM in flexion, extension, axial rotation, and lateral bending was the smallest in the S-I-C fixation model, followed by the S-I and S-L-C fixation models; the greatest ROM was seen in the S-L fixation model. The S-I-C and the S-I fixation model both contained six screws in the construct; the S-L-C and the S-L fixation only had four screws in the construct. Reinforcement with two additional screws at the fractured level can enhance biomechanical stability; and these results were similar to a previous FE study by Li et al. [
28]. Two additional screws inside the fractured vertebra was also stronger than when bone cement was used inside the fractured level (S-I fixation model versus S-L-C fixation model). This result was similar to the clinical results demonstrated by Liao et al. [
27].
The pedicle screw is considered to be the weakest point of the posterior fixation, and implant failure is usually caused by screw loosening or screw root breakage [
29]. It has also been shown that the pedicle screws still continue to bear most of the load after fusion [
30,
31]. To prevent early screw breakage, and to prolong the life of the screw, it is necessary to prevent the load on the screw from reaching the fatigue load. For this reason, we evaluated the maximum von Mises stress on the root of the pedicle screws in the four fixation models. A lower stress distribution around the root of the pedicle screw means the less chance of pedicle screw breakage. The maximum von Mises stress of the pedicle screws in the S-L group was higher than that in the other three fixation groups. In the S-I-C group, the maximum von Mises stress of the pedicle screw in all directions was the lowest. These results indicate that using two additional screws inside the fractured vertebra can reduce the stress on the pedicle screw root. Further augmentation with bone cement at the fractured level can further reduce the stress around the pedicle screw root, and further decrease the probability of screw failure. In addition, we found the pedicle screw root under lateral bending and axial rotation load sustained higher von Mises stresses, as finding similar to that of Xu et al. [
32].
Injectable bone substitutes such as calcium sulfate, calcium phosphate cement, and hydroxyapatite cement are widely used for filling bone defects. These bone substitutes not only have osteoconductive ability to promote bone union, but also provide initial mechanical support. Evaniew et al. [
33] used calcium sulfate/phosphate cement to manage patients with bone defects after curettage of primary tumor. Orsini et al. [
34] demonstrated that calcium sulfate cement could promote new bone formation in a rabbit model of bone defects. Furthermore, Xu et al. [
26] used a FE model to show that cement augmentation of a fractured vertebra could decrease the von Mises stress on the rods by 50% and on the screws by 40%. In the current study, however, the initial stability provided by bone cement inside the fractured vertebra did not achieve the stability provided by two additional screws. Nevertheless, an advantage of bone cements is that they can stimulate bone healing inside the vertebra, which screws cannot provide. Therefore, a combination of bone cement and two additional screws inside the fractured vertebra along with short-segment instrumentation may be an ideal surgical method for thoracolumbar burst fractures because provides greater initial stability, and also stimulates bone growth.
There are limitations of this study that should be considered. First, the FE spine models were reconstructed from data of a single patient, and thus is not representative of various ages or different sexes. Second, most cases of thoracolumbar burst fractures are associated with injury to the upper endplate and/or adjacent intervertebral disc. This was not addressed in the current models. Removal of the upper end-plate of the vertebrae in the FE model may better reflect the real injury pattern of a burst fracture. However, the main purpose of this study was to compare different fixation techniques, not different injuries. Third, because the current FE study only provided comparison data for the stability of different fixation methods, and the material representations of the biological structures were assumed to be linearly elastic, biomechanical studies of cadavers are necessary to validate the results for clinical practice. If the results are validated in cadaver studies, clinical cohort studies would be warranted.