With the steady development of internal fixation techniques, lateral mass screw fixation (LSF) with plates or rods has become the standard method for posterior cervical spine fixation and stability on various surgical indications. Many studies emphasized that LSF techniques could provide the same biomechanical stability, as compared with anterior cervical fixations or posterior wiring techniques. Additionally, due to the development of the polyaxial screw-rod system, cervical fixation surgery has now become easy to perform. Therefore, most surgeons believe the LSF techniques are the optimum methods for reconstructing the stability of the cervical spine, following decompressive surgery [1, 2].

LSF techniques have been constantly studied and modified, since it was first introduced by Roy-Camille in 1972; and the techniques of Magerl, Anderson and An were extensively used as well. Even though numerous anatomical and clinical studies pointed out that the techniques of Roy-Camille and Magerl had the best practical applications, inevitable and undesirable complications still occurred, such as adjacent facet joint injury, nerve root, and vertebral artery or screw loosening and fracture [3]. In fact, all LSF techniques may result to diverse complications, depending on the surgeon’s optimization of the LSF techniques [4, 5]. For the last two decades, many studies have evaluated the advantages of different LSF techniques, based on its complications [6–8]. However, we are unaware of any report that directly compares LSF techniques by analyzing fixation fractures as a severe complication -- which has a relatively low incident rate [9, 10].

Due to limited case reports, it is rare find studies related to the fixation fractures of the LSF system; making it impossible to have a case analysis and difficult to simulate by traditional cadaveric experiments. A novel biomechanical approach by computer could provide a possible way to reveal the rupture risks of fixation in different LSF techniques. In recent years, finite element (FE) analysis is a technique that has been generally adopted by orthopedic experts to address biomechanical problems, especially for stress distribution [11–13]. The CT scan data obtained from a healthy volunteer was used in our study; and the four LSF techniques were applied to establish reconstructive FE models of the lower cervical spine, after a three-segment (C4-C6) laminectomy. The biomechanical stress distribution of the reconstructed models were analyzed by applying different LSF techniques via FE analysis -- aiming to provide experimental and theoretical reference for choosing the LSF technique for the lower cervical spine.

The 3D FE model is expected to provide theoretical reference for clinical practice. We established an intact model of the lower cervical spine (C3-C7) that contained bony structures, annulus fibrosus, nucleus pulposus and related ligaments (i.e. anterior and posterior longitudinal ligaments, ligamentum flavum, spinous ligament, and capsular ligaments). The intact model consisted of 503,911 elements and 93,390 nodes (Table 3), which were more elaborate than the previous models [20–23]. We simulated ligamentous structures based on the biomechanical data that was previously reported [24] and ligament functions were simulated as uniaxial nonlinear element connections. The physiological lordosis, caused by its unloaded position, recovered to accord with the authentic biomechanics of the human vertical spine. The intact model’s angular intersegmental range of motions was in good agreement with the data reported by the literature [17–19]. These results validated the model and that this model could be used for other analysis. Additionally, the mesh quantity determines the accuracy of the biomechanical analysis; our FE model promises better results in our future studies.

Based on the validated model, laminectomy was simulated to establish the postoperative model, and the four reconstructive models using the four different techniques (i.e. Roy-Camille, Magerl, Anderson, and An techniques). All simulations were performed in Solidworks 2012. With its powerful design and assembly functions, the reconstructive models were established more accurately. Fault recognition was effectively avoided, when files were exported to the other software programs – due to its software processing.

Magerl, Anderson, and An techniques were based on the Roy-Camille technique. The biomechanical stability of all four techniques could provide similar results, so as not to destroy the nucleus pulposus. In recent years, the comparisons of the above four LSF techniques mainly focused on the following aspects, facet joint violation, nerve injury, vertebral artery injury and screw pullout force. Firstly, for the facet joint violation, the point of entry and orientation of the screw were different from the four techniques. The facet joint may be violated during the screw’s implantation. Studies [25, 26] have shown that approximately 50% of patients had complications due to facet joint violation, after undergoing fixation surgery with the technique of Roy-Camille. However, these complications rarely happened to patients who underwent fixation surgery using the Magerl and Anderson techniques. Secondly, for nerve injury, chronic postoperative neck pain may occur in patients after undergoing surgery with the technique of Magerl. This symptom occurs due the protrusion of the screw tip, which is close to the dorsal branch of the nerve root [27, 28]. Xu et al. reported that potential risks of nerve root violation were higher in patients who underwent the Magerl and Anderson techniques, than the An technique [29]. The potential risk of nerve root injury was lower in patients who used the Roy-Camille technique, than the Magerl technique [30]. Third, for the vertebral artery injury, due to the distance of the vertebral artery from the screw’s trajectory, this artery is not likely to get injured. Heller et al. [30] found that the vertebral arteries of 26 cadaver specimens were not threatened, after undergoing LSF surgery of the lower cervical spine with the techniques of Roy-Camille and Magerl. Based on the findings of Heller, Katonis et al. [4] found no cases of vertebral artery injury among the 225 patients who underwent LSF surgery. Lastly, although comparisons had been previously conducted, the pullout force of the implanted screws, with the application of different LSF techniques, is still controversial [31]. Ulrich et al. [32] found that the imprecise fixation of screws could decrease the pullout force while undergoing surgery with the Roy-Camille technique. Magerl and An techniques paid more attention to the orientation of the screw to acquire the maximum pullout force – fixing these screws firmly to prevent extraction [25].

At present, there are still no reports on stress distribution comparisons for the four LSF techniques following laminectomy in the lower cervical spine -- for neither cadaveric study nor FE analysis. In our study, the FE analysis showed obvious differences among the four LSF techniques in response to stress distribution and maximum von Mises stress. Under various loading conditions, stress distribution in both Magerl and Roy-Camille groups were more dispersive as compared with the other two groups. The fixations in the Anderson and An groups had higher stress concentrations on the upper-side of the connecting rods and in the rod-screw’s interface. The difference of the screw’s point of entry and orientation contribute mostly to the difference in stress distribution. However, there were no significant stress distribution differences between the Magerl and Roy-Camille groups, and the Anderson and An groups. The maximum von Mises stress in the Anderson and An groups were obviously higher than the other two groups -- especially under extension and left-right axial rotating conditions. Nearly 100 MPa was concentrated on the upper side of the rod, which led to higher risks of fracture. According to the comprehensive analysis above, fixation by Magerl and Roy-Camille techniques were safer than the Anderson and An techniques. These conformed results were based on the routine choice of surgeons [3]. Although, the Magerl and Roy-Camille techniques are still better, looking from the standpoint of stress distribution.

In summary, stress distribution was more dispersive and the maximum von Mises stress levels were lower in the Magerl and Roy-Camille groups for various conditions. Magerl and Roy-Camille techniques were safer for stabilizing the lower cervical spine, due to its potentially low risk of fixation fracture. We also suggest that surgeons should use new alloy materials for fixation or should limit extension movement, to retain and yield better strength and to decrease the risk of fracture, when opting for applying the Anderson and An techniques.