A Hybrid Uniplanar Pedicle Screw System with a New Intermediate Screw for Minimally Invasive Spinal Fixation: A Finite Element Analysis

Purpose A hybrid pedicle screw system for minimally invasive spinal fixation was developed based on the uniplanar pedicle screw construct and a new intermediate screw. Its biomechanical performance was evaluated using finite element (FE) analysis. Methods A T12-L2 FE model was established to simulate the L1 vertebral compression fracture with Magerl classification A1.2. Six fixation models were developed to simulate the posterior pedicle screw fracture fixation, which were divided into two subgroups with different construct configurations: (1) six-monoaxial/uniplanar/polyaxial pedicle screw constructs and (2) four-monoaxial/uniplanar/polyaxial pedicle screw constructs with the new intermediate screw. After model validation, flexion, extension, lateral bending, and axial rotation with 7.5 Nm moments and preloading of 500 N vertical compression were applied to the FE models to compare the biomechanical performances of the six fixation models with maximum von Mises stress, range of motion, and maximum displacement of the vertebra. Results Under four loading scenarios, the maximum von Mises stresses were found to be at the roots of the upper or lower pedicle screws. In the cases of flexion, lateral bending, and axial rotation, the maximum von Mises stress of the uniplanar screw construct lay in between the monoaxial and polyaxial screw constructs in each subgroup. Considering lateral bending, the uniplanar screw construct enabled to lower the maximum von Mises stress than monoaxial and polyaxial pedicle screw constructs in each subgroup. Two subgroups showed comparable results of the maximum von Mises stress on the endplates, range of motion of T12-L1, and maximum displacement of T12 between the corresponding constructs with the new intermediate screw or not. Conclusions The observations shown in this study verified that the hybrid uniplanar pedicle screw system exhibited comparable biomechanical performance as compared with other posterior short-segment constructs. The potential advantage of this new fixation system may provide researchers and clinical practitioners an alternative for minimally invasive spinal fixation with vertebral augmentation.


Introduction
Spinal fixation with the posterior pedicle screw-rod system has been considered the mainstream treatment method for degenerative and traumatic spinal diseases. A number of in vitro, finite element (FE) and clinical studies have been performed to figure out the influencing factors for the fixation stability and modification or alternative methods to improve the biomechanical or functional outcomes, which included bone mechanical properties [1], disc arthroplasty [2,3], semirigid rod [4], and dynamic or hybrid stabilization device [5,6].
Regarding thoracic and lumbar vertebral fractures, minimally invasive short-segment pedicle screw instrumentation is an easy procedure and gaining momentum as a routine option. However, limited exposure increases the difficulty of coupling pedicle screws with rods, which will lead to a superficial position of longitudinal rods and decreased strength of the entire construct. Benefited by a decreased force arm, the polyaxial pedicle screw (PAPS) allows easier assembling pedicle screws with rods and provides theoretically better biomechanical strength. Nevertheless, prior studies demonstrated that PAPS tended to slippage through the universal joint, and its mechanical strength was always less than the monoaxial pedicle screw (MAPS) in the sagittal plane [7,8]. These phenomena allow a new uniplanar pedicle screw (UPPS) to be introduced.
As the head of UPPS that can only swivel freely in one specific plane while fixed in other directions, the freedom of movement of the screw's head in the axial plane of body can facilitate the pedicle screw-rod assembly, without sacrificing the stiffness of the construct in the sagittal plane, whereas only two studies performed the biomechanical investigation of UPPS for spinal fixation to date, which verified its biomechanical advantage over PAPS in the sagittal plane [7,8]. Meanwhile, biomechanical and clinical studies have suggested that an intermediate screw can provide substantially improved stability of posterior short-segment constructs in spinal fracture fixation [9,10]. Nevertheless, few studies incorporated an intermediate screw into the UPPS construct.
In addition to these modifications to the instrument design, for most thoracolumbar bursts or compression fractures, vertebral augmentation has been used as a means to mechanically support the fractured vertebra anteriorly. Nonetheless, most clinical trials utilized one-above and one-below 4-MAPS short-segment construct, of which some reported excellent reduction and lordosis maintenance [11,12], while others concluded with indifferent results [13]. There was limited evidence about the six-screw short-segment construct with additional vertebroplasty at the fracture level. Liao et al.'s finite element analysis (FEA) study verified that additional vertebroplasty could provide a stiffer 6-MAPS construct and less stress in the thoracolumbar burst fracture model [10]. Moreover, the 6-MAPS short-segment construct was also performed with vertebroplasty [12] and bone grafting [14] in the open thoracolumbar fracture operation. However, during operation, the intermediate screw needed to be removed after distraction and reduction, followed by vertebroplasty. The current pedicle screw constructs do not allow vertebral augmentation without interrupting the procedure of reduction and fixation of the spinal fracture.
In the present study, a hybrid UPPS fixation system was introduced. With this short-segment construct, a new intermediate screw (NIS) was developed to facilitate vertebroplasty and reduction maneuver. Different from the traditional parallel configuration of pedicle screws, this NIS is inserted in a more outside-in direction. This configuration allows bilateral intermediate screws to achieve the center of the injured vertebra, elevate the depressed endplate, and then maintain its reduction position more efficiently. To demonstrate the biomechanical performance of this hybrid UPPS system, an FEA was performed to compare the monoaxial, uniplanar, and polyaxial screw short-segment constructs with or without the NIS.

Materials and Methods
2.1. Instrument Design. The geometric 3D model of UPPS was drawn using the Solidworks software (Dassault Systèmes, Concord, MA, USA). The UPPS is barrel designed, with the outer and inner diameter of 6 mm and 4.1 mm, respectively. The screw-tulip joint of our UPPS allows about ±30°angular motion of the screw in the axial plane of the body (perpendicular to the plane in which the slot of the gasket is located), which is greater than Ye et al.'s UPPS of ±25° [7]. During manufacture, the spot welding technique was used to minimize the mobility of the screw-tulip joint in other directions. Unlike the UPPS, the NIS was designed based on the USS® cannulated schanz screw. Our modification is that the threads of the NIS are located at the anterior and posterior one-third shank, with a smooth part in the middle. Both the UPPS and NIS have lateral fenestrations in the distal portion of the thread for bone cement or graft injection. The outer and inner diameters of the NIS are 6 mm and 5.2 mm, respectively. Besides, PAPS and MAPS (VIPER® System, DePuy Spine, Inc.) were also used in this study for comparison between four types of constructs in FEA. The thread type and pitch of the PAPS, UPPS, and MAPS are single lead and 3 mm, whereas those of the NIS are double lead and 4 mm. All four types of screws used in this study were shown in Figure 1.  Figure 2 and Table 1.
The material properties of the T12-L2 FE model were assigned based on previous studies [10,15] (Table 2). The cortical bone had a thickness of 1 mm. The posterior elements included the pedicle, spinal process, transverse process, and articular process. The interfaces between the facet joints were set as frictionless surface-to-surface contact. The interfaces between the screw and bone were set to be bonded. The contact between the head and tulip of UPPS and PAPS was set as surface-to-surface contact with a friction coefficient of 0.3. According to the attachment positions, the spinal ligaments were established and defined as nonlinear hyperplastic material using Combine 39, which only allowed for tension deformation without the compression behavior. The mechanical behaviors of the spinal ligaments were based on nonlinear stress-strain curves [16].
The bone and intervertebral disc were meshed with Solid 187 (Figure 3). Due to a variety of material properties of the spine FE model and methods of the FE modeling, a mesh sensitivity test was performed to verify the developed FE model of T12-L2. The maximum von Mises stress was used to evaluate the mesh convergence. Four T12-L2 models were generated with mesh sizes of 0.5 mm, 1 mm, 2 mm, and 3 mm for 2 BioMed Research International all structures, respectively. After the lower surface of L2 was fixed in all directions, a pure moment of 7.5 Nm was applied on the upper surface of T12. The maximum von Mises stresses of models with mesh size of 1 mm, 2 mm, and 3 mm were calculated and compared with that with a mesh size of 0.5 mm. When the difference was less than 5%, the mesh was considered convergent. Considering the burden and precision of calculation, the mesh size of 1 mm was chosen in further analysis. The percentage of error was 2.70% in this situation.

Model Validation.
Before FEA was performed on the spinal fixation model, the FE model of the intact spine was validated with range of motion (ROM) of T12-L1 and L1-L2, respectively. The inferior surface of the inferior vertebra was fixed in all directions. A reference point was created on the superior surface of the T12 and L1 vertebrae for force and moment application, respectively. Flexion, extension, right lateral bending, and right axial rotation with pure 7.5 Nm moment were applied to T12-L1 and L1-L2, respectively. The ROM of two segmental units was calculated at the endpoint of loading.

Fixation Model.
Following the previously established methods [7,17,18], a cuneiform osteotomy was applied to the L1 vertebra to construct the vertebral compression fracture model with Magerl classification A1.2 [19]. Six types of short-segment pedicle screw constructs were established to simulate the thoracolumbar vertebral compression fracture fixation, including three six-screw constructs with 211,632 elements and 373,160 nodes on average: 3 BioMed Research International to the reference point of the T12 vertebra [20]. Under different types of loading, the maximum von Mises stress on the fixation construct and the endplates of T12 and L2, the ROM of T12-L2, and the maximum displacement of T12 were recorded and compared between six groups.

FE Model Validation.
The ROMs of the intact T12-L1 model were shown in Table 3. These findings were similar to those of previous studies [21,22].
3.2. ROM of T12-L2 and Maximum Displacement of T12. As shown in Table 4, the ROM of T12-L2 and the maximum displacement of T12 under flexion loading were similar between the six fixation groups. Under the other three loading types, the ROM and maximum displacement were the smallest in the MAPS groups, followed by the uniplanar and polyaxial groups in sequence, irrespective of the type of the intermediate screw.

Von Mises
Stress on the Pedicle Screw System. Disregarding the sporadic stress concentration point caused by the screw-tulip mobile joint of UPPS, which was out of the scope of this study [23], the maximum von Mises stresses on the instrument were mainly located at the roots of the upper and lower pedicle screws under four types of loading, according to the stress nephogram of the FE model (Figure 4; Supplementary Material (available here)). Maximum von Mises stresses of three six-screw constructs and their NIS counterparts were seen at similar positions (Table 5).
Under flexion loading, the maximum von Mises stresses were seen on the upper screws of the monoaxial and uniplanar screw constructs and the lower screws of the polyaxial screw constructs. Moreover, the maximum stress was the lowest in the 6-PAPS construct, followed by the 6-UPPS and 6-MAPS construct in sequence, and the NIS subgroup displayed the same varying rule.
In the case of extension, the maximum von Mises stresses were shown on the lower screws in all six con-struct groups. The maximum stress was the lowest in the 6-MAPS construct, followed by the 6-UPPS and 6-PAPS construct in sequence, and the NIS subgroup displayed similar varying rule.
For the condition of lateral bending, the maximum von Mises stresses were found at the lower screws in uniplanar and polyaxial screw constructs but the upper screws in monoaxial screw constructs. Besides, the varying rule was opposite to the flexion condition.
The right axial rotation movement resulted in similar changing trends like the flexion loading, except for the pair of 4-PAPS/2-NIS and 6-PAPS. The possible reason might be that the fixed NIS altered the force-bearing condition of the polyaxial screw constructs and caused stress concentration.

Von Mises
Stress on the Lower Endplate of T12 and the Upper Endplate of L2. Table 6 shows that the maximum von Mises stresses on the two endplates were almost the

Discussion
As known to all, lack of adequate anterior column support is an important reason for rekyphosis and early implant failure after the traditional posterior pedicle screw fixation [10]. To increase the stiffness of the entire instrument, several technical modifications have been made in the posterior pedicle screw construct, including crosslink, hook [24], and intermediate screw [25] or vertebral augmentation at the fracture level [10]. In Liao et al.'s study [10], the crosslink could relieve the stress of the whole construct under a rotation load, which cannot be used in the minimally invasive procedure. The intermediate screw has been verified to be an efficient reinforcer to the biomechanical stability of the whole construct. Baaj [26] and increase the stiffness of the construct [27]. Wang et al.'s in vitro test showed that the intermediate screw of the 6-MAPS construct could decrease more ROM of the burst fractured spine than the 6-PAPS construct [28]. The biomechanical superiority of the intermediate screw was also demonstrated in other FEA studies [10,15,17].  [30].
Until now, there were limited FEA studies on the UPPS system and hardly any intermediate screw used in the UPPS short-segment construct either. To facilitate vertebral augmentation, we designed a new intermediate screw with a new screw configuration. Our FEA results showed that the biomechanical performance of the UPPS short-segment construct with the NIS was acceptable by comparing with 6-UPPS or 4-MAPS/2-NIS. Our results were not entirely the same as the previous FEA studies because of inconsistent conditions and parameters. We did not see the maximum stress on the rod like Wang et al. [30]. The possible reason might be the preloading vertical compression with 500 N in our study that was much greater than 150 N in theirs [30]. However, the positions of maximum stresses on six-screw constructs were comparable to their four-screw counterparts with the NIS, which also indirectly validated our established FE model. Table 2 shows that the maximum von Mises stresses on the NIS were less than their counterparts in the six-screw constructs with the parallel configuration. The possible reason might be that the NIS was deviated from the plane on which the upper and lower pedicle screws were located. Thus, with different types of loading, the NIS accepted less stress than that in the parallel configuration, especially under flexion and extension. This difference became less significant in the PAPS groups because of its screw-tulip universal joint. However, under axial rotation, the maximum von Mises stress was extremely large on the NIS in the 4-PAPS/2-NIS construct. The reason might be that the NIS was rigidly connected with the rod rather than the mobile joints of the upper and lower PAPS. Under flexion, extension, and lateral bending, this connection difference between the NIS and upper/lower PAPS was not so significant; however, under the axial rotation moment,  One potential advantage of our new construct is that the smooth middle one-third shank of the NIS allowed increased ROM of the screw within the injured vertebra, which could conveniently elevate the endplate of the injured vertebra and strengthen the fracture fixation simultaneously. Another advantage is the combination of vertebroplasty and the intermediate screw. The conventional short-segment construct does not allow additional vertebroplasty without interfering its standard procedure. The current solutions include a short-segment construct with vertebroplasty but without the intermediate screw [13] or partial vertebroplasty with a short intermediate screw [10,31]. In the studies on the six-screw construct combined with bone grafting [14] or vertebroplasty [12], after the reduction maneuver, the intermediate screw   BioMed Research International needed to be removed and reinserted at the interval between the augmentation and fixation steps, which interrupted the normal operation process and extended the operative time. Korovessis et al. performed minimally invasive 6-MAPS short-segment fixation with kyphoplasty in A2/A3 lumbar fractures [31]. However, the short intermediate screws could only be inserted partly into the fractured vertebra after kyphoplasty. Neither of the two alternatives can give full play to both technologies. Our study had several limitations. First, although there existed several potential advantages, including vertebroplasty and the elevation and augmentation of the depressed endplate with the help of the NIS, we did not prove it in the present study due to the study design and space limitation. The application in minimally invasive spinal fixation with vertebral augmentation should be conducted in further simulated and clinical studies. Second, the FE model was based on a healthy male. How our system performs in the vertebral osteoporotic compression fracture is still needed to be determined. Third, our FE model was simplified without considering all the influencing factors. For example, the endplates were considered a part of the cortical bone rather than an independent structure. In vitro experiments on cadaveric specimen should be performed to verify the results achieved in the present study.

Conclusions
The observations shown in this study verified that the developed hybrid uniplanar pedicle screw system exhibited comparable biomechanical performance than the current posterior short-segment constructs. Considering the biomechanical performance and superiority of vertebral augmentation and operation maneuvers, this new fixation system may provide researchers and clinical practitioners an important alternative for minimally invasive spinal fixation in patients with thoracolumbar vertebral compression fracture.

Data Availability
The data of the FE model validation, ROM of T12-L2 and maximum displacement of T12, and von Mises stress on the pedicle screw system and endplate were used to support the findings of this study that are included within the article.