Changes of adjacent segment biomechanics after anterior cervical interbody fusion with different profile design plate: single- versus double-level

Abstract Low-profile angle-stable spacer Zero-P is claimed to reduce the morbidity associated with traditional plate and cage construct (PCC). Both Zero-P and PCC could achieve comparable mid- and long-term clinical and radiological outcomes in anterior cervical discectomy and fusion (ACDF). It is not clear whether Zero-P can reduce the incidence of adjacent segment degeneration (ASD), especially in multi-segmental fusion. This study aimed to test the effect of fusion level with Zero-P versus with PCC on adjacent-segment biomechanics in ACDF. A three-dimensional finite element (FE) model of an intact C2–T1 segment was built and validated. Six single- or double-level instrumented conditions were modeled from this intact FE model using Zero-P or the standard PCC. The biomechanical responses of adjacent segments at the cephalad and caudal levels of the operation level were assessed in terms of range of motion (ROM), stresses in the endplate and disc, loads in the facets. When comparing the increase of adjacent-segment motion in single-level PCC fusion versus Zero-P fusion, a significantly larger increase was found in double-level fusion condition. The fold changes of PCC versus Zero-P of intradiscal and endplate stress, and facet load at adjacent levels in the double-level fusion spine were significantly larger than that in the single-level fusion spine during the sagittal, the transverse, and the frontal plane motion. The increased value of biomechanical features was greater at above segment than that at below. The fold changes of PCC versus Zero-P at adjacent segment were most notable in flexion and extension movement. Low-profile device could decrease adjacent segment biomechanical burden compared to traditional PCC in ACDF, especially in double-level surgery. Zero-P could be a good alternative for traditional PCC in ACDF. Further clinical/in vivo studies will be necessary to explore the approaches selected for this study is warranted.


Introduction
Anterior cervical discectomy and fusion (ACDF) has been a well-established and successful treatment for cervical spondylotic myelopathy and radiculopathy since it was firstly descripted by Smith and Robinson (1958). After that, much advancement has been made in surgical technique and prosthesis options. Plate and cage construct (PCC) is often used to achieve stability and promote fusion in ACDF procedure. In order to address some problems resulted from plate design and additional anterior dissection, zero-profile anchored spacer (Zero-P) devices have been developed (Scholz et al. 2009). The Zero-P devices have been shown to provide biomechanical stability comparable to PCC in both single-and double-level ACDF Reis et al. 2014;Stein et al. 2014).
With the increase of the number of cervical spine fusion surgery annually, adjacent segment disease (ASD) has become a major concern after cervical fusion surgery (Hashimoto et al. 2019). There are many factors have been implicated in the development of ASD although the etiology is unclear. ASD is not only driven by the natural history but also by the changes of mechanical environment in adjacent segment (Harrod et al. 2012). The anterior disc and bilateral facet joints form the central path for the transmission of the loads along the spine. Load sharing of the spine occurs anteriorly through the disc and posteriorly through the facet joints. Altered biomechanical environment resulted from the fused motion segment possibly play roles in the development of ASD.
Compared with single-level ACDF group, there is a trend for a greater rate of ASD in the double-level ACDF group (Zigler et al. 2016). There is increased adjacent-segment motion at the adjacent levels after a double-level compared with a single-level ACDF using PCC construct (Prasarn et al. 2012). A meta-analysis has demonstrated that, compared with the PCC group, the Zero-p group had a significantly reduced incidence of ASD (Sun et al. 2018). Over the 2-year follow-up, ACDF with traditional PCC showed a higher incidence of ASD than ACDF with stand-alone cages in double-level ACDF patients (Ji et al. 2015).
To our knowledge, there is no study specifically examining the effect of fusion level on adjacent-segment biomechanics of Zero-P versus PCC in ACDF. This study aimed to compare the biomechanical effects between double-level and single-level fusion with Zero-P versus PCC. We hypothesized that increased number of operated segment will lead to obviously more greater biomechanical changes at adjacent segments in PCC fusion model than that in Zero-P fusion condition. A three-dimensional computational study was carried out to examine these effects.

Model development and validation
A nonlinear three-dimensional finite element (FE) model of cervical spine segments (C2 À T1) was developed. The bony geometry was derived from computed tomography (CT) scan images of a healthy 38-year-old male. The male subject was scanned with the slice thickness of 0.75 mm and the slice increment of 0.69 mm thickness from C2 to T1. The Cobb angle was in the reported range of the subaxial cervical spine (Linder 2000). Lordosis of C2-C7 in current model was 23.7 . The DICOM images were imported into the software Mimics (Materialise NV, Belgium) to construct the geometric structure of C2-T1. The mesh structure was prepared using the preprocessing software Hypermesh (Altair Technologies Inc) and then was imported into Abaqus (Simulia, Providence, RI) to solve. The FE model included cortical bone, cancellous bone, bony posterior elements, annulus fibrosus (AF), nucleus pulposus (NP), posterior facets, end plates, anterior longitudinal ligament, posterior longitudinal ligament, ligamentum flavum, interspinous ligament, and capsular ligaments. The insertion points and areas of the ligaments were closely matched with published data (Yoganandan et al. 2000). The model components, the element type and the material properties are shown in Table 1 (Yoganandan et al. 2000(Yoganandan et al. , 2001Hussain et al. 2009;Tchako and Sadegh 2009;Lee et al. 2011;Wang et al. 2017). The fluid-like behavior of the disc was simulated to be nearly incompressible using eight-node hybrid hexahedral elements. The facet joint was built as a nonlinear three-dimensional contact problem with surface-to-surface contact elements. Hexahedral elements with an isotropic-elastic material model were used to model facet cartilage. Major ligaments of the cervical region were represented by truss elements.
To validate the FE model, range of motion (RoM) of each cervical segment was the major indicator. Subaxial RoM and functional spinal unit RoM (movement in sagittal plane, coronal plane and axial plane) were chosen for comparison with the published experimental results (Miura et al. 2002;Wheeldon et al. 2006;Yoganandan et al. 2007Yoganandan et al. , 2008. The subaxial RoM was defined as the measurement of the total motion between the C2 and C7 vertebrae. The functional spinal unit RoM, intersegmental motion, was the motion between two adjacent vertebrae. The same boundary and loading conditions were simulated with the controlled experiments. All degrees of freedom (DoF) were constrained on the lower surface of T1 in our FE model. Pure moments of 2-Nm were applied to the superior surface of C2 in the three main planes to produce flexion, extension, lateral bending and axial rotation respectively. Using the follower load technique, a compressive follower load of 100 N was also applied to the upper surface of C2 to simulate physiologic compressive loads. The validation and following FE analysis were performed in ABAQUS (Simulia, Providence, RI). Subaxial RoM and functional spinal unit RoM were recorded.

Surgery simulation, generation of implant models and boundary conditions
From the intact C2-T1 FE model ( Figure 1), a singleor double-level ACDF was performed. Fusion models were created by removing the anterior longitudinal ligament, the intervertebral disc and cartilaginous endplate. To study the efficacy of fusion on the adjacent level biomechanics, and to make the target level be the same segment, six fusion conditions to be tested were designed according to Prasarn et al.'s method (Prasarn et al. 2012). As shown in Figure 2, six fusion models were created. C4-C5 was chosen as the segment for comparing the fusion action on the superior adjacent level in Model A-D ( Figure 2). In order to test the influence of fusion on the inferior adjacent segment, biomechanical features of C6-C7 segment was observed in Model C-F ( Figure 2). In single-level conditions, C5-C6 intervertebral disc was removed. In double-level models, C4-C5 and C5-C6, or C5-C6 and C6-C7 intervertebral discs were removed. Thereafter, six FE models received instrumentation with PCC or Zero-p (Figure 3, Model A-F). In the standard PCC fusion model, a rigid  anterior cervical plate was used. Plate (width 16 mm, thickness 2 mm) and fixed screws (diameter 4 mm, length 14 mm) were rigidly fixed in the fused segment ( Figure 3). The 4-screw anchored Zero-P system (Synthes, Oberdorf, Switzerland) were adopted in the current study ( Figure 3). Fixed screws (diameter 3 mm, length 16 mm) were adopted in the Zero-P system. The same PEEK interbody spacer (width 15 mm, length 16 mm, and height 6 mm) was used in both PCC and Zero-P fusion model to maintain the sagittal alignment after operation. The titanium alloy plate and PEEK material properties were assigned to the respective implants (Table 1). In the standard PCC surgical model, the above and below distance of the plate to the adjacent disc were 5.5 mm and 5 mm in one level PCC construct model, 5.5 mm and 5.2 mm in C4-C5 and C5-C6 fusion model, and 5.7 mm and 5.3 mm in C5-C6 and C6-C7 fusion model. The C2-7 Cobb angle were 22 for one level (C5-C6) fusion, 24 for C4-C5 and C5-C6 fusion model and 23 for C5-C6 and C6-C7 fusion model. In all loading conditions, the inferior surface of T1 vertebra was fully constrained. A pure moment of 2 Nm combined with a follower load of 100 N were similarly imposed on C2. The external (range of motion, RoM) and internal (endplate, disc and facets load sharing) responses adjacent to fusion of single-and double level ACDF surgeries in six models were assessed. Stresses in the endplate and disc were calculated using the average von-Mises stresses. The facet loads at a motion segment was defined as the average loads on the right and left articulating facets.

Validation of the FE model
The intact cervical spine FEM consisted of 152608 elements and 41797 nodes (Figure 1). The present intact cervical model of C2-T1 vertebrae was compared with previously published experimental results to assess the validity. The predicted segmental RoM of the flexion-extension, lateral bending, and axial rotation of the intact cervical model were well in agreement with previous experiments studies ( Figure  4) (Miura et al. 2002;Wheeldon et al. 2006;Yoganandan et al. 2007Yoganandan et al. , 2008.

Kinematics of adjacent segments after oneand two-level fusion
When comparing single-level PCC fusion versus Zero-P fusion, RoM in proximal adjacent segment (C4-C5) showed an increase of 6.5% in the sagittal plane, 4.9% in the transverse plane, and 5.6% in the frontal plane ( Figure 5, Model C, D). In distal adjacent segment (C6-C7), RoM of 3.8%, 2.6%, 4.8% increases were observed in the sagittal, transverse and frontal planes, respectively ( Figure 5, Model C, D). Whereas, in double-level fusion models, the above (C4-C5) ( Figure 5, Model A,B) and below (C6-C7) ( Figure 5, Model E,F) adjacent segments had RoM increases of 16.7% and 5% in the sagittal motion, 7.7% and 6.1% in the transverse motion, and 13.7% and 5.6% in the frontal motion, respectively.

Biomechanical changes in disc, endplate and facet in adjacent levels
As shown in Figure 6, at the superior adjacent segment (C4-C5), single level fixation produced mildly higher flexion, extension and lateral bending intradiscal and endplate stress, and facet load in the PCC fusion model than that in the Zero-P fusion model. When compared with the single level Zero-P fusion, during flexion, extension and lateral bending, the biomechanical value in the PCC fixation surgery increased by 1.5%, 2.2% and 1.0% of the intervertebral disc stresses ( Figure 6, Model C, D); by 3.8%, 4.2% and 1.2% of the endplate stresses ( Figure 6, Model C, D); and by 7.6%, 14.9% and 12.2% of the facet loads ( Figure 6     In the double level procedure, the fold changes of PCC versus Zero-P value were obviously enlarged. Compared with the Zero-P fusion condition, there were significant increases of flexion, extension, axial rotation and lateral bending: 21.3%, 43.6%, 97.8%, 2.9% of intradiscal stresses (Figure 7, Model E, F); 35.2%, 56.9%, 15.6%, 12.1% of intraendplate stresses (Figure 7, Model E, F); and 61.3%, 67.6%, 12.3%, 34.8% of intrafacet loads ( Figure  7, Model E, F). The increased values of the fold changes of PCC versus Zero-P rate were larger at the superior than that at the inferior segment.

Discussion
The role of ACDF in patients with cervical spine disc disease has long been established. To reduce various intraoperative and postoperative complications associated with PCC, while maintaining the benefits of interbody cages with anterior plating, low-profile angle-stable spacer Zero-P has been developed. The current investigation was performed predominantly to evaluate the effect of fusion level on adjacent-segment biomechanics of Zero-P versus PCC in ACDF. In the present computational simulation, compared with single-level fusion, double-level fusion will lead to obviously much bigger biomechanical changes of the ratio of PCC to Zero-P at adjacent segments was shown.
Zero-p spacers with integrated fixation can simplify the surgical technique while maintaining the stability of the construct. The promising biomechanical data for two-level use has been reported (Scholz et al. 2015;Kang et al. 2017). Zero-P device showed comparable stability to traditional PCC for two-level fusion . Clinical studies showed that Zero-p device may be associated with a reduced rate of ASD compared to using the PCC construct in single and double level ACDF (Ji et al. 2015;Sun et al. 2018;Cheung et al. 2019;Lu et al. 2019). The present study provided biomechanical evidences for these clinical results. The fold changes of PCC versus Zero-P of intradiscal and endplate stresses, and facet loads at adjacent levels in the double-level fusion models were significantly larger than that in the single-level fusion models during the sagittal, the transverse, and the frontal plane motion (Figures 6 and 7). Our results showed that the fold changes of PCC versus Zero-P were most notable in flexion and extension movement (Figures 6 and 7). The current computational simulation indicated that, compared to using the PCC construct, the low rate of biomechanical effects might be occurred in Zero-p device fusion condition. And the rate of ASD might be higher in double level ACDF using Zero-P versus PCC construct than that in single level.
The number of levels included in the fusion construct has been considered as an important factor for contribution to the development of ASD following ACDF. Since fusion of cervical spinal segments may lead to excessive stress on the unfused adjacent segments, therefore, longer fusions may cause greater stresses at adjacent levels after spine fusion and likely lead to ASD. Clinical investigation has demonstrated an increased risk of degeneration with increasing length of fusion. Greater rate of ASD in double-level ACDF group than in the single-level group has been reported through 5-year follow-up (Zigler et al. 2016). ASD is a long-term complication of cervical spine fusion procedures. It has been reported that over the course of 10 years after ACDF, symptomatic ASD developed at an incidence of 2.9% per year (Hilibrand et al. 1999). Although Basques et al. (2018) considered that multi-level procedures may not be a significantly greater risk of developing ASD compared to singlelevel procedure. This might be due to short follow-up time, 2 years, in their investigation. Prasarn et al. (2012) had proven that the biomechanics affecting adjacent levels in the cervical spine after ACDF do change from a single-to a multilevel fusion. Similar to their results, the present study also showed that there was increased adjacent-segment motion at the levels above and below, after a double-level compared with a single-level ACDF with both PCC and Zero-P constructs ( Figure 5). The increased value was larger at the superior than that at the inferior segment ( Figure 5). The increased mechanical response (Figures 6 and 7), in combination with the aging process, may synergistically hasten the process of degeneration adjacent to the cervical fusion level. With time, ASD may require additional surgical intervention.
Mechanical irritation against the adjacent segment caused by a traditional PCC is regarded as a predisposing factor of ASD (Chen et al. 2015). Goffin et al. (1995) suggested that the shortest plate possible be used to avoid producing an effect on adjacent segments. Zero-P spacer implant was developed to avoid such possible irritation. Compared to traditional plates, such all-in-one fusion device is contained within the disc space allowing for maximum distance from the adjacent level disc. Our computational simulation indicated that, in both single-and double-level fusion models, higher biomechanical features were occurred in PCC fusion models (Figures 5-7). So we supposed that, Zero-P spacer may lower the biomechanical effects on adjacent level for a no-additionalplate restriction in anterior vertebrae, especially, in the double-level ACDF surgery. However, future more clinical studies with high methodological quality and long-term follow-up periods are needed to evaluate the PCC versus Zero-P ACDF procedures for multilevel cervical spondylotic myelopathy treatment.
As for the test protocols for investigation of instrumented adjacent segment biomechanics, the application of pure moments during load-controlled testing (that is, a flexibility test) versus displacement-controlled testing (that is, a hybrid test) for the evaluation of spine instrumentation in vitro has been the subject of much debate over many years (Costi et al. 2021). For investigation of adjacent segment degeneration (ASD), no conclusive consensus exists regarding test protocol and load application according to the literature reports (Arnold and Friis 2015). From a parallel review on in vivo kinematics, because none of the current test protocols can replicate the in vivo kinematics, the in vivo kinematic data might put the validity of current in vitro protocols, hybrid testing and flexibility testing protocols, into question Volkheimer et al. 2015). Hybrid approach proposed by Panjabi (2007) has recently become of increased interest as non-fusion technologies entered the clinical arena and the accompanying emphasis about limiting potential adjacent level disease (Arnold and Friis 2015). Many scholars believe that the data obtained using the hybrid approach depict the reality in a patient following surgery (Goel et al. 2005;Arnold and Friis 2015). We will make a comparison of the flexibility protocol versus hybrid protocol in future biomechanical studies, and this may yield interesting comparisons.
FE analysis is an effective simulation method for predicting the trend in biomechanics after different surgical procedures and thereby providing certain guidance for clinical management. However, there are some limitations in the current study which should be taken into account. First, caution should be taken when interpreting the results of the current study, because the intact FEM is based on a single scan of a normal man. The FE simulation aimed to provide the trend rather than the actual data. FE analysis has limitations, similar to the cadaver studies and other published FE studies. Second, Truss elements were used for ligaments modeling. The contact interaction between ligaments and vertebrae does not take into account such simplification, but this has the advantage of avoiding unrealistic shearing forces in the ligaments and thus has a reduced computation time. Third, the absence of neck muscles may affect the finite element biomechanical features, for instance, motion and stress. Because of the role of these muscles is to control the cervical range of motion. The biomechanical behavior of different fusion devices should be evaluated with future clinical studies and in vivo biomechanical works are warranted.

Conclusions
Zero-p device was associated with different biomechanical effects compared to the PCC construct in ACDF. The present computational simulation shows that, compared with Zero-P device, PCC could bring about higher ROM changes at adjacent segment in both single-and double-level fusion models, especially in double level fusion condition. Compared with single-level fusion, double-level fusion could bring about increased changes of stress in the disc and endplate, and changes of facet load ratio of PCC to Zero-P at adjacent segments. It may be possible that the rate of ASD is lower in double level ACDF with Zero-P versus PCC construct than that in single level. The results of this study may be important for surgical decision-making and may provide potential biomechanical rationale for using Zero-P implants in doublelevel ACDF.

Disclosure statement
No potential conflict of interest was reported by the authors.