Available online at www.sciencedirect.

com

Medical Engineering & Physics 30 (2008) 11271133

Cervical spine biomechanics following implantation of a disc prosthesis

Fabio Galbusera a, , Chiara M. Bellini a , Manuela T. Raimondi a,b , Maurizio Fornari a , Roberto Assietti c
b a IRCCS Istituto Ortopedico Galeazzi, via R. Galeazzi 4, 20161 Milan, Italy LaBS, Department of Structural Engineering, Politecnico di Milano, Milan, Italy c Ospedale Fatebenefratelli e Oftalmico, Milan, Italy

Received 10 July 2007; received in revised form 11 January 2008; accepted 12 February 2008

Abstract This study presents a nite element model of the C4C7 segment in healthy conditions and after implantation of a disc prosthesis at a single level, in order to investigate of the inuence of disc arthroplasty on the biomechanics of the cervical spine. A nonlinear nite element model of the C4C7 segment in intact conditions was developed and run in exion and extension. A detailed model of the Bryan disc prosthesis, including contacts between the different components of the device, was built and positioned at C5C6. The calculated segmental motion resulted preserved after disc arthroplasty, with respect to the model of the intact spine, in both exion and extension. A general preservation of the forces transmitted through the facet joints was obtained; a minor force increase at the implanted level was detected. The analysis of the instantaneous centers of rotation (ICR) in exionextension showed the preservation of a physiological kinematics. The mechanical behaviour showed an asymmetry between exion and extension, probably due to the removal of the anterior longitudinal ligament and the anterior part of the annulus brosus, and the preservation of the posterior structures. In general, the disc prosthesis showed to be able to reproduce a nearly physiological motion. However, other important mechanical aspects, such as the possible micromotion at the boneimplant interface and the possible degenerative conditions of the spine, need to be evaluated before drawing a conclusion about total disc arthroplasty from an engineering point of view. 2008 IPEM. Published by Elsevier Ltd. All rights reserved.
Keywords: Finite element; Cervical arthroplasty; Articial disc; Cervical disc prosthesis

1. Introduction Cervical fusion is currently the gold-standard treatment for herniated cervical discs [1]. However, many clinical studies show that fusion can be associated with symptomatic adjacent-segment disease [2,3]. Early adjacent level degeneration appears to be due to the abnormal kinematics and load transfer patterns at adjacent segments induced by fusion. Cervical disc arthroplasty, which allows the preservation of the mobility of the implanted segment, could reduce the stress sustained by the adjacent levels and avoid, or at least slow down, their early degeneration. Preliminary clinical studies on cervical arthroplasty, based on short or intermediate

Corresponding author. Tel.: +39 02 66214939; fax: +39 02 23994365. E-mail address: fabio.galbusera@polimi.it (F. Galbusera).

follow-up times, report no evidence of adjacent symptomatic cervical disc degeneration, and no increased motion at adjacent levels after implantation of a disc prosthesis [3,4]. Thus, based on the current clinical experience, cervical arthroplasty seems to be a promising option. One of the most commonly implanted articial cervical discs is the Bryan disc prosthesis (Medtronic Ltd., Memphis, TN, USA), which contains a low-friction, wear-resistant polyurethane core, which is located between, and articulates with, shaped titanium shells that include convex porous in growth surfaces, to allow bony xation to the adjacent vertebral endplates. A exible polyurethane membrane surrounds the interior articulating shell surfaces, to separate the internal structures of the device from the external in vivo environment and to contain a lubricant. Short and intermediate-term randomized controlled clinical trials have been published [3,5],

1350-4533/$ see front matter 2008 IPEM. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.medengphy.2008.02.002

1128

F. Galbusera et al. / Medical Engineering & Physics 30 (2008) 11271133

Fig. 1. FE meshes of (a) the model of the intact spine, (b) the model of the spine after implantation of the Bryan disc prosthesis at the C5C6 level and (c) the disc prosthesis and the peri-implantar region.

and report clinical outcome at least on par with cervical fusion. However, long-term follow-up studies demonstrating the advantages of disc arthroplasty versus fusion are not available at present. Some biomechanical studies on cervical arthroplasty are currently available in the literature [6], mostly referring to ball-and-socket disc prostheses. In particular, some experimental ex vivo studies on cadaver implanted spine specimens provide momentrotation curves in exionextension at the adjacent levels [79]. Ex vivo comparisons of the intradiscal pressures between fusion and arthroplasty are also available [10,11], and generally show increased intradiscal pressures and range of motion (ROM) in the segments adjacent to fusion, with respect to arthroplasty. Pickett et al. [12] conducted a radiographic kinematic analysis of cervical motion after implantation of the Bryan disc prosthesis in 20 patients. The authors reported a general preservation of the exibility and motion patterns of the implanted and the adjacent segments. Some computational studies of disc arthroplasty are also currently available in the literature. Most of them are targeted to the investigation of ball-and-socket disc prostheses [1315]. Ha [16] considered a generic elastomer cervical disc prosthesis, and calculated the elastic modulus that induced an optimal biomechanics at the implanted and the adjacent segments. A paper concerning the Bryan cervical disc prosthesis is currently available [17], but it includes some simplications in the modeling of the disc prosthesis and is limited to the C5C6 unit, thus neglecting the inuence of the disc prosthesis on the adjacent levels. The present study is aimed to overcome the main limitations of the former model [17], by including the whole C4C7 spinal segment and a detailed nite element (FE) modeling of the Bryan disc prosthesis, taking into account the interactions between the different components constituting the devices. The main target of the study is the analysis of the motion patterns of the spine in exionextension, by determining the momentrotation curves for all levels and the location of the instantaneous center of rotation (ICR) of the implanted level. The load-bearing behaviour of the implanted spine and of the core of the disc prosthesis is also investigated.

2. Materials and methods Two different FE models of the C4C7 spinal segment were built. The rst model was targeted to the simulation of an intact segment in healthy conditions, while in the second one the Bryan cervical disc prosthesis was included at the C5C6 level. The anatomy and loading conditions imposed were symmetrical with respect to the sagittal plane, then only one half of the complete structure was analyzed in both models. The vertebral geometry was built based on CT images. Commercial software (Amira 4.1, TGS, San Diego, CA, USA) was employed to convert the CT images into a point cloud describing the bony surfaces. A 8-node hexahedral mesh was then built for each vertebra. Cortical bone was modeled with solid hexahedral elements, having thickness of 1 mm both on the lateral walls and on the endplates [18]. The intervertebral discs were then added, by subdividing the intradiscal space in an inner zone representing the nucleus pulposus (NP), and an external layer constituting the annulus brosus (AF); both were discretized in a hexahedral mesh (Fig. 1a). Annular bers were included as 5160 tension-only trusses in 8 layers, having cross-sectional area of 0.1 mm2 . All the solid materials were modeled as linear elastic isotropic (Table 1). Ligaments were modeled as nonlinear spring elements (Table 2) [19,20]. Ligaments included in the models were the anterior longitudinal (ALL), posterior longitudinal (PLL), capsular (CL), aval (FL) and interspinous (ISL). The mechanics of the facet joints was modeled by including a frictionless surface-based contact between the relevant bony surfaces.

Table 1 Mechanical properties of the components of the FE models E (MPa) Cancellous bone Cortical bone Nucleus pulposus Annulus brosusmatrix Annulus brosusbres Titanium Polyurethane 100 12,000 1 2.5 500 110,000 70 0.3 0.3 0.499 0.45 0.35 0.3 0.3

F. Galbusera et al. / Medical Engineering & Physics 30 (2008) 11271133 Table 2 Initial average length and force-deection data of cervical ligaments ALL (l0 = 6 mm) F (N) 0 32 60 81 100 115 l (mm) 0 1.2 2.5 3.7 4.8 6 PLL (l0 = 3.6 mm) F (N) 0 28 50 66 79 88 l (mm) 0 1.2 2.2 3.2 3.4 5 FL (l0 = 9.5 mm) F (N) 0 30 55 71 95 105 l (mm) 0 1.8 3.5 5.1 6.9 8 ISL (l0 = 11.8 mm) F (N) 0 8.5 10 23 28 32 l (mm) 0 1.3 2.8 4.1 5.5 7 CL (l0 = 2.5 mm) F (N) 0 1.5 29 52 86 104

1129

l (mm) 0 1.7 3.6 5 7.5 9.5

The model of the implanted segment were derived from the model of the intact segment, by including the Bryan cervical disc prosthesis at C5C6 (Fig. 1b and c), being one the levels where arthroplasty is most frequently performed [3]. A proper prosthesis size (diameter 18 mm) was chosen, by analyzing the CT scans of the spinal segment. The intervertebral disc was removed, except for the lateral portion of the annulus brosus, in order to simulate the surgical procedure. The relevant endplates were removed, and a spherical socket tting the Bryan disc prosthesis was obtained in the trabecular vertebral bone. The uncinate processes were preserved, as indicated by the standard surgical technique. The ALL was removed at the implanted level; the PLL was preserved [17]. The C5 and C6 vertebrae were distracted by 1 mm, and a FE hexahedral mesh of the Bryan prosthesis was then inserted into the socket. The prestresses in the PLL, FL and ISL due to a homogenous distraction of 1 mm were taken into account. A homogenous tensile prestress of 0.5 MPa was imposed to the remaining annulus matrix at C5C6, to model the distraction due to the insertion of the disc prosthesis. A bonded contact was imposed between the outer shells of the disc prosthesis and the vertebral socket, in order to simulate full osteointegration of the device. A detailed FE model of the prosthesis was implemented (Fig. 2). The model included four components: the upper and the lower shells, the exible external membrane and the internal core. Five contact pairs were dened (Fig. 2): two for the core-upper shell contact, two for the core-lower shell contact and one for the core-membrane contact. For all contacts, a friction coefcient of 0.1 and a contact stiffness factor of 0.1 were assumed. Simulation were conducted with commercial software (ANSYS, ANSYS Inc., Canonsburg, PA, USA). Nonlinear geometry effects were accounted for in all simulations. The lower endplate and the lower surface of the spinous process of C7 were xed in all simulations. A hybrid loading protocol was employed [21]. The model of the intact spine was loaded in exion and extension, by imposing a pure moment of 2.5 Nm [22] combined with a compressive preload of 100 N to the C4 superior endplate. Then, the global C4C7 rotation was registered and imposed to the model of the implanted spine, combined to the 100 N compressive preload. This approach allowed the comparison of the segmental motion between the intact and the implanted spine,

to study the motion compensation due to the insertion of the disc prosthesis at the implanted and the adjacent segments. Momentrotation curves in exion and extension were obtained for both models at all levels. The results obtained with the intact model were compared with data obtained from the literature, to assess the validity of the model. Then, the momentrotation curves from the intact and the implanted models were analyzed and compared, in order to study the inuence of the disc prosthesis on cervical spine motion. To better understand the motion patterns in exion and extension [23], the locations of the instantaneous center of rotation of the C5C6 level were determined, by the use of a least squares procedure [24] implemented in the GNU Octave language [25]. In order to obtain a nite number of ICRs, their locations were calculated for moment increments of 0.5 Nm, thus obtaining ve points for both exion and extension. Vertebrae were assumed to be rigid for the calculation of the ICR locations. The load-bearing behaviour of the cervical spine in exion and extension was analyzed, by comparing the forces transmitted through the facet joints at the various level, in intact and healthy conditions and after disc arthroplasty, and by examining the stress distribution in the prosthesis core.

Fig. 2. Contact pairs considered in the FE modeling of the Bryan disc prosthesis. Each color is associated to a pair; solid lines and dots represents the contours of the axisymmetric contact surfaces in a sagittal section view.

1130

F. Galbusera et al. / Medical Engineering & Physics 30 (2008) 11271133

3. Results The momentrotation curves in exion and extension obtained with the two models are reported in Fig. 3. All the curves show a marked nonlinearity, particularly in exion. The agreement between the curves calculated with the model of the intact, healthy spine and the average curves published by Wheeldon et al. [26] is good, except for the C5C6 level in extension, where an error of about 40% in correspondence of the peak extension moment was obtained. The curves obtained with the model including the disc prosthesis appear rather similar to those obtained with the model of the intact segment. However, some differences are observable, in particular with reference to the implanted segment. In exion, the ROM of the C5C6 level appears reduced if compared to the intact segment, but still significantly preserved. In extension, a minor reduction of the ROM is observable; however, the mobility of the implanted segment appears to be globally preserved. The asymmetry between exion and extension in the ROM alteration induced by the disc prosthesis is likely to be due to the resection of the ALL and the anterior part of the AF, which are able to bear relatively great forces mainly in extension, and the total preservation of the posterior structures which provide a signicant load-bearing capacity mainly in exion. The quantication of the segmental motion compensation induced by the disc prosthesis, by using the hybrid load protocol, is reported in Fig. 4. The total C4C7 ranges of motion were 29.2 degrees in exion and 20.4 degrees in extension. The ROM reductions at the implanted segment were 29% in exion and 19% in extension. The mobility of the segments adjacent to the implanted one results not to be signicantly altered by the placement of the disc prosthesis, with the largest difference being an increase of 22% at C6C7 in extension. Fig. 5 shows the ratios between the forces transmitted through the facet joints in extension calculated with the model of the implanted segment and those calculated with the model of the intact segment. The highest value is scored by the implanted segment, which shows an increase of the facet forces of 8% that is likely to be related to the different motion pattern imposed by the disc prosthesis. Overall, the values of the forces transmitted through the facet joints are not signicantly altered by the implantation of the disc prosthesis. The von Mises stress distribution in the polyurethane core of the prosthesis is reported in Fig. 6, for the maximal values of the exion and extension moments. Peak stress values in exion and in extension are in the same order of magnitude. Some areas subjected to high stresses, where wear or failure may be more likely, can be observed in both load cases: the sharp edges on the articular surfaces, the outer rim (due to the contact with the external containing membrane) and the inner surfaces (due to the contact with the metallic posts).

Fig. 3. Momentrotation curves obtained with the models of the intact and the implanted segment at C4C5 (a), C5C6 (b) and C6C7 level (c) in exion and extension, compared with experimental corridors [26].

Fig. 7 reports the locations of the ICR in exion and extension, determined by a least squares procedure, obtained with the models of the intact spine and the implanted spine. The ICR locations of the intact spine are in good agreement with the positions described by Penning and Wilmink [23], and

F. Galbusera et al. / Medical Engineering & Physics 30 (2008) 11271133

1131

Fig. 4. Segmental motion compensation induced by the disc prosthesis, by using the hybrid load protocol, for the three considered levels (C4C5, C5C6 and C6C7).

generally belong to the posterior zone of the C6 vertebra, being located near the endplate in exion and moving in the posterior-caudal direction in extension. Although the ICR locations are not exactly corresponding between the intact and the implanted spine, the alteration of the motion pattern can be considered not critical. 4. Discussion In this paper, we present FE models of the C4C7 spinal segment in intact, healthy conditions and after implantation of a disc prosthesis at the C5C6 level. The models were analyzed in exion and extension. As for the model of the intact spine, the calculated results are directly comparable

Fig. 5. Bar chart of ratios between the forces transmitted through the facet joints calculated with the model of the implanted segment (Fimplanted ) and those calculated with the model of the intact segment (Fintact ), in extension.

with those published by Wheeldon et al. [26], referring to ex vivo tests targeted to help the validation of FE models. As in the former study, the momentrotation curves are asymmetric and show greater exibility in exion than in extension. As stated above, the agreement is good, except for the C5C6 level in extension, probably due to the shape of the facet joints of the specic cervical spine. In a previous study, Galbusera et al. [17] presented a FE model of the C5C6 spinal unit implanted with the Bryan disc prosthesis, and compared the results to those obtained with a model of the intact segment. The present paper overcomes

Fig. 6. Contour plot of the Von Mises stresses in the core of the Bryan disc prosthesis, in exion and extension. The areas subjected to the highest stresses are pointed out.

1132

F. Galbusera et al. / Medical Engineering & Physics 30 (2008) 11271133

Fig. 7. Average locations of the C5C6 ICR calculated with the model of the intact and of the implanted spine, during exion and extension. Each motion is subdivided into ve steps.

many of the limitations of the previous work, by implementing a more detailed modeling of the Bryan disc prosthesis and including the adjacent segments. The results of the two studies are in general agreement, except for the ICR location, which was found to belong to a more limited area in the former paper. This difference is probably due to the different detail level in the modeling of the disc prosthesis: while in the former paper the interaction between the core and the shells was modeled as a bonded contact, in the present study the contacts between the core, the shells and the containing membrane are fully taken into account. The calculated ICR locations are in agreement with those obtained in radiological studies conducted on both healthy [23] and operated patients [12]. Overall, the inuence of disc arthroplasty on the ICR location in exionextension is found to be minor. The surgical resection of the ALL and the anterior AF is found to play a signicant role in the biomechanics of the implanted segment, in both the current and the cited paper, determining the asymmetry in the results between exion and extension [17]. In particular, the contribution of the posterior structures in exion is able to preserve the stability of the disc prosthesis, thus preventing the detaching of the core from the shells. In extension, after removal of the ALL and the anterior AF, the only mechanical structures involved in transmitting loads are the facet joints and the disc prosthesis itself. Thus, the relative motion between the different prosthesis components are not as constrained as in exion. In fact, a slightly larger ROM in extension has been found at the implanted level if compared to exion. Although this ROM difference can be safely considered negligible, this result is contrasting with experimental observations on the intact spine [26], in which the extension ROM is smaller than exion one. The discrepancy between the validation corridor at C5C6 in extension and the results obtained with the model of the intact spine should also be taken into consideration, thus caution should be adopted when interpreting this result.

The inuence of the disc prosthesis on the mechanics of the adjacent segments resulted to be not signicant, in terms of both facet forces and ROMs. Comparisons between the intact and the implanted spine were made possible by the use of a particular hybrid load protocol. However, in vivo load conditions are more complex than those adopted in the present paper, presumably resulting in a different motion and load sharing behaviour. More advanced load protocols have been described in the literature [8]. However, the choice of a more consolidated and extensively employed load protocol as the one here employed was preferred in the present study. The current computational results suggest that a cervical disc prosthesis may be able to preserve the motion and to induce a nearly physiological ICR at the implanted segment and limit the alteration of the biomechanics of the adjacent levels. However, the current results give just some partial light on the engineering aspect, since some topics that may have a signicance were not evaluated in the present study. The boneimplant interface was modeled as bonded, thus neglecting any possible micromotion. This may not be a realistic assumption, in particular in the immediate postoperative time, where osteointegration has not been achieved yet. Furthermore, spine degeneration was not considered, thus the anatomy and the material properties were modeled with reference to the intact condition. Other limitations derive from the modeling approaches of the NP and the AF, which have been extensively employed in past studies [27], though more sophisticated and accurate models have been presented [28,29]. Furthermore, the model validation was performed only for the model of the intact spine, whereas the momentrotation curves obtained with the model of the implanted spine were not validated against experimental data. Another missing validation of the present model is related to the facet contact forces, due to the unavailability of these data in the current literature. Due to this limitation, only relative values were provided for comparison [30,31]. In addition to the engineering aspects, the clinical experience is fundamental for the study of total disc arthroplasty. Preliminary clinical results for the Bryan disc prosthesis are encouraging [5]; however, only a long-term randomized controlled clinical trial, not currently available, may be able to assess the better benet-risk ratio of disc arthroplasty versus fusion, thus denitely indicating that disc replacement is the treatment of choice for cervical disc disease. Conict of interest All the Authors (Fabio Galbusera, Chiara M. Bellini, Manuela T. Raimondi, Maurizio Fornari, Roberto Assietti) have no proprietary, nancial, professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as inuencing the position presented in, or the review of, the present manuscript entitled Cervical spine biomechanics following implantation of a disc prosthesis.

F. Galbusera et al. / Medical Engineering & Physics 30 (2008) 11271133

1133

References
[1] Mummaneni PV, Haid RW. The future in the care of the cervical spine: interbody fusion and arthroplasty. J Neurosurg Spine 2004;1(2):1559. [2] Hilibrand AS, Robbins M. Adjacent segment degeneration and adjacent segment disease: the consequences of spinal fusion? Spine J 2004;4(6 Suppl):190S4S. Review. [3] Gofn J, Van Calenbergh F, van Loon J, Casey A, Kehr P, Liebig K, et al. Intermediate follow-up after treatment of degenerative disc disease with the Bryan cervical disc prosthesis: single-level and bi-level. Spine 2003;28(24):26738. [4] Duggal N, Pickett GE, Mitsis DK, Keller JL. Early clinical and biomechanical results following cervical arthroplasty. Neurosurg Focus 2004;17(3):E9. Review. [5] Coric D, Finger F, Boltes P. Prospective randomized controlled study of the Bryan cervical disc: early clinical results from a single investigational site. J Neurosurg Spine 2006;4(1):315. [6] Puttlitz CM, DiAngelo DJ. Cervical spine arthroplasty biomechanics. Neurosurg Clin N Am 2005;16(4):58994. Review. [7] DiAngelo DJ, Foley KT. An improved biomechanical testing protocol for evaluating spinal arthroplasty and motion preservation devices in a multilevel human cadaveric cervical model. Neurosurg Focus 2004;17(3):E4. [8] DiAngelo DJ, Foley KT, Morrow BR, Schwab JS, Song J, German JW, et al. In vitro biomechanics of cervical disc arthroplasty with the ProDisc-C total disc implant. Neurosurg Focus 2004;17(3):E7. [9] Chang UK, Kim DH, Lee MC, Willenberg R, Kim SH, Lim J. Range of motion change after cervical arthroplasty with ProDisc-C and prestige articial discs compared with anterior cervical discectomy and fusion. J Neurosurg Spine 2007;7(1):406. [10] Dmitriev AE, Cunningham BW, Hu N, Sell G, Vigna F, McAfee PC. Adjacent level intradiscal pressure and segmental kinematics following a cervical total disc arthroplasty: an in vitro human cadaveric model. Spine 2005;30(10):116572. [11] Wigeld CC, Skrzypiec D, Jackowski A, Adams MA. Internal stress distribution in cervical intervertebral discs: the inuence of an articial cervical joint and simulated anterior interbody fusion. J Spinal Disord Tech 2003;16(5):4419. [12] Pickett GE, Rouleau JP, Duggal N. Kinematic analysis of the cervical spine following implantation of an articial cervical disc. Spine 2005;30(17):194954. [13] Denoziere G, Ku DN. Biomechanical comparison between fusion of two vertebrae and implantation of an articial intervertebral disc. J Biomech 2006;39(4):76675. [14] Goel VK, Grauer JN, Patel TCh, Biyani A, Sairyo K, Vishnubhotla S, et al. Effects of Charit e articial disc on the implanted and adjacent spinal segments mechanics using a hybrid testing protocol. Spine 2005;30(24):275564.

[15] Dooris AP, Goel VK, Grosland NM, Gilbertson LG, Wilder DG. Loadsharing between anterior and posterior elements in a lumbar motion segment implanted with an articial disc. Spine 2001;26(6):E1229. [16] Ha SK. Finite element modeling of multi-level cervical spinal segments (C3C6) and biomechanical analysis of an elastomer-type prosthetic disc. Med Eng Phys 2006;28(6):53441. [17] Galbusera F, Fantigrossi A, Raimondi MT, Sassi M, Fornari M, Assietti R. Biomechanics of the C5C6 spinal unit before and after placement of a disc prosthesis. Biomech Model Mechanobiol 2006;5(4):25361. [18] Yoganandan N, Kumaresan SC, Voo L, Pintar FA, Larson SJ. Finite element modeling of the C4C6 cervical spine unit. Med Eng Phys 1996;18(7):56974. [19] Yoganandan N, Kumaresan S, Pintar FA. Geometric and mechanical properties of human cervical spine ligaments. J Biomech Eng 2000;122(6):6239. [20] Haghpanahi M, Mapar R. Development of a parametric nite element model of lower cervical spine in sagittal plane. Conf Proc IEEE Med Eng Biol Soc 2006;1:173941. [21] Panjabi MM. Hybrid multidirectional test method to evaluate spinal adjacent-level effects. Clin Biomech 2007;22(3):25765. [22] Wilke HJ, Wenger K, Claes L. Testing criteria for spinal implants: recommendations for the standardization of in vitro stability testing of spinal implants. Eur Spine J 1998;7(2):14854. [23] Penning L, Wilmink JT. Rotation of the cervical spine. A CT study in normal subjects. Spine 1987;12(8):7328. [24] McCane B, Abbott JH, King T. On calculating the nite centre of rotation for rigid planar motion. Med Eng Phys 2004;27(1):759. [25] Eaton JW. GNU octave manual. Publisher: Network Theory Limited; 2002. [26] Wheeldon JA, Pintar FA, Knowles S, Yoganandan N. Experimental exion/extension data corridors for validation of nite element models of the young, normal cervical spine. J Biomech 2006;39(2):37580. [27] Yoganandan N, Kumaresan S, Pintar FA. Biomechanics of the cervical spine. Part 2. Cervical spine soft tissue responses and biomechanical modelling. Clin Biomech 2001;16(1):127. [28] Yin L, Elliott DM. A homogenization model of the annulus brosus. J Biomech 2005;38(8):167484. [29] Del Palomar AP, Calvo B, Doblar e M. An accurate nite element model of the cervical spine under quasi-static loading. J Biomech 2008;41(3):52331. [30] Rousseau MA, Bonnet X, Skalli W. Inuence of the geometry of a ball-and-socket intervertebral prosthesis at the cervical spine: a nite element study. Spine 2008;33(1):E104. [31] Chang UK, Kim DH, Lee MC, Willenberg R, Kim SH, Lim J. Changes in adjacent-level disc pressure and facet joint force after cervical arthroplasty compared with cervical discectomy and fusion. J Neurosurg Spine 2007;7(1):339.