Biomechanical investigation of a novel integrated device for intra-articular stabilization of the...

The Spine Journal 12 (2012) 136–142

Basic Science

Biomechanical investigation of a novel integrated devicefor intra-articular stabilization of the C1–C2 (atlantoaxial) joint

Peter A. Robertson, MD, FRACSa,*, Parmenion P. Tsitsopoulos, MD, PhDb,c,Leonard I. Voronov, MD, PhDb,c, Robert M. Havey, BSb,c, Avinash G. Patwardhan, PhDb,c

aThe Orthopaedic Clinic, Mercy Specialist Centre, 100 Mountain Rd, Epsom, Auckland 1023, New ZealandbMusculoskeletal Biomechanics Laboratory, Edward Hines Jr. VA Hospital, PO Box 5000 (151), Hines, IL 60141, USA

cDepartment of Orthopaedic Surgery and Rehabilitation, Loyola University Chicago, 2160 S. First Ave, Maywood, IL 60153, USA

Received 28 January 2011; revised 14 September 2011; accepted 5 January 2012

Abstract BACKGROUND CONTEXT: The anatomy

FDA device/drug

Author disclosure

Johnson DePuy (B); S

cial); Trips/Travel: M

Institute, Auckland, N

Medtronic (Financial)

employer). PPT: Not

Nothing to disclose.

shares); Consulting: A

rangements: Aesculap

1529-9430/$ - see fro

doi:10.1016/j.spinee.2

of the atlantoaxial joint makes stabilization at thislevel challenging. Current techniques that use transarticular screw fixation (Magerl) or segmentalscrew fixation (Harms) give dramatically improved stability but risk damage to the vertebral artery.A novel integrated device was designed and developed to obtain intra-articular stabilization via pri-mary interference fixation within the C1–C2 lateral mass articulation.PURPOSE: To assess the atlantoaxial stability achieved with a novel integrated device when com-pared with the intact, destabilized, and stabilized state using the Harms technique.STUDY DESIGN: A biomechanical study of implants in human cadaveric cervical spines.METHODS: Six human cadaveric specimens were used. Biomechanical testing was performedwith moment control in flexion-extension, lateral bending, and axial rotation. Range of motion(ROM) was measured in the intact state, after both destabilization by creation of a Type II odontoidpeg fracture and sequential stabilization using the integrated device and the Harms technique.RESULTS: Mean flexion-extension ROM of the intact specimens at C1–C2 was 14.1�62.9�. Desta-bilization increased the ROM to 31.6�64.6�. Instrumentation with the Harms technique reducedflexion-extension motion to 4.0�61.4� (p!.01). The integrated device reduced flexion-extension mo-tion to 3.6�61.8� (p!.01). In lateral bending, the respective mean angular motions were 1.8�61.1�,14.1�65.8�, 1.4�60.7�, and 0.4�60.3� for the intact destabilized Harms technique and integrated de-vice. For axial rotation, the respective mean values were 67.3�613.8�, 74.2�616.1�, 1.4�60.7� and0.9�60.7�. Both the Harms technique and integrated device significantly reduced motion comparedwith the destabilized spine in flexion-extension, lateral bending, and axial rotation (p!.05).Direct com-parison of the Harms technique and the integrated device revealed no significant difference (pO.10).CONCLUSIONS: The integrated device resulted in interference fixation at the C1–C2 lateral massjoints with comparable stability to the Harms technique. Perceived advantages with the integrated de-vice include avoidance of fixation below theC2 lateral masswhere the vertebral artery is susceptible toinjury, and access to the C1 screw entry point through the blade of the integrated device avoiding ex-tended dissection superior to the C2 nerve root and its surrounding venous plexus. � 2012 ElsevierInc. All rights reserved.

Keywords: Atlantoaxial; Biomechanics; Fusion device; Stabilization

status: Not applicable.

s: PAR: Consulting: Medtronic (F), Johnson and

peaking/Teaching Arrangements: Medtronic (Finan-

edtronic (Financial); Board of Directors: The Back

ew Zealand (Financial); Scientific Advisory Board:

; Grants: Medtronic (C, Paid directly to institution/

hing to disclose. LIV: Nothing to disclose. RMH:

AGP: Stock Ownership: Spinal Kinetics (10,000

lphatec (B), Aesculap (B); Speaking/Teaching Ar-

(B); Trips/Travel: Spinal Kinetics (B), Aesculap

(A); Scientific Advisory Board: Ortho Kinematics (Financial), Axiomed

(Financial), Spinal Kinetics (Financial); Research Support (Staff/Mate-

rials): Simpirica (E, Paid directly to institution/employer); Grants: Synthes

(E, Paid directly to istitution/employer).

The disclosure key can be found on the Table of Contents and at www.

TheSpineJournalOnline.com.

* Corresponding author. The Orthopaedic Clinic, Mercy Specialist

Centre, 100 Mountain Rd, Epsom, Auckland 1023, New Zealand. Tel.:

(64) 9-630-0214; fax: (64) 9-630-3981.

E-mail address: p.a.robertson@xtra.co.nz (P.A. Robertson)

nt matter � 2012 Elsevier Inc. All rights reserved.

012.01.004

137P.A. Robertson et al. / The Spine Journal 12 (2012) 136–142

Introduction C1–C2—atlantoaxial instability. This may cause pain, cord

The C1–C2 (atlantoaxial) joint is anatomically unique inthe human spine. The near flat lateral mass articulation sur-faces on the inferior aspect of the C1 lateral mass and the su-perior aspect of the C2 lateral mass allow significanttranslation of one surface against the other. The coupled op-posing translation (anterior translation of C1 on C2 on oneside and posterior translation on the contralateral side) ofthe paired lateral mass joints about the odontoid peg allowsaxial rotation at the C1–C2 joint that is many times morethan at any other cervical or spinal articulation and accountsfor 50% of the cervical spine rotation. This unique anatomydoes not favor other motion at C1–C2. Flexion-extension issimilar in range of motion (ROM) to other cervical levels.Lateral bending is markedly reduced at C1–C2 comparedwith other spinal levels, an observation that is of no surprisewhen the joint shapes are considered [1,2].

Although the anatomy of the joints at C1–C2 creates re-markable rotation, it renders this articulation particularlyvulnerable to any pathological process that affects the in-tegrity of the odontoid peg, the transverse ligament andthe capsular structures around the C1–C2 synovial joint.

Trauma to the odontoid, congenital abnormalities of for-mation, ligamentous injuries, and inflammatory disordersthat lead to ligament laxity and insufficiency all have thepotential to ‘‘uncouple’’ the reciprocal translational motionoccurring at the C1–C2 lateral mass joints, leading toabnormal anterior or posterior translational instability of

Fig. 1. The integrated device for C1–C2 stabilization. (A) Schematic diagram of

radiograph of integrated device inserted at C1–C2 in a cadaveric spine. (C) Sche

(D) Posteroanterior radiograph of the integrated device inserted at C1–C2 in a c

compression with major neurological deficits, and death.The surgical challenge to stabilize a painful or unstable

C1–C2 joint to achieve fusion has led to a multitude of op-erations that provide varying stiffness of this joint balancedwith an increasing risk to the vascular structures that pro-vide the arterial supply to the Circle of Willis.

Initial posterior wiring procedures [3,4] and subsequentposterior clamp devices [4–9] resulted in inadequate stabil-ity and fusion. Magerl transarticular screws reduced rotation[10] and improved fusion and outcomes [8,9,11–17]. Seg-mental screw fixation [18] was popularized by Harms [19].It gives the most stiffness and is associated with good clini-cal results at this level [20–25]. Nevertheless, both theMagerl and Harms techniques are associated with significantrisk to the vertebral artery (VA) in the vertebral artery fora-men (VAF) of C1 and C2 [15,26–31].

Attempts to avoid the potential vascular risks of theMagerl/Harms techniques have led to several alternate con-structs that avoid the lateral masses of C1 and C2 [32,33].Further stabilization using the posterior elements has led toalternative implant descriptions [34], and there have beenlimited attempts to use the lateral mass articulations as a sitefor the basis of a C1–C2 construct [21,35,36].

The risks associated with current C1–C2 stabilizationtechniques lead us to design a new integrated device to im-mobilize this joint (Fig. 1). Our primary aims were to avoidboth the VAF in C1 and C2 and the need for sublaminar

the integrated device (lateral view) for stabilization of C1–C2. (B) Lateral

matic diagram of the integrated device (posterior view) inserted at C1–C2.

adaveric spine.

138 P.A. Robertson et al. / The Spine Journal 12 (2012) 136–142

wires or cables, as well as optimize the use of interferencefixation to fix the two lateral mass joint articular surfaces.The design integrated an intra-articular blade with a serratedsurface for interference stabilization of the two lateral massjoint surfaces, an oblique passage within the posterior part ofthe blade that allowed insertion of an oblique bicortical fixedangle screw from the posterior inferior corner of the C1 lat-eral mass in an anterior and superior direction medial to theC1 VAF, and a posterior leg for mini screw attachment to thelamina of C2. Most important, we designed the integrateddevice to be inserted through the C1–C2 joint, with accessto the joint from a posterior approach using the superiorand lateral aspect of the C2 lamina to guide to the joint. Pre-vious clinical experience has demonstrated that this route isrelatively avascular. Initial prototypes were trialed in cadav-eric specimens, and with minor modifications, it was veryclear that the implant could be inserted without difficulty,and excellent rigidity of C1–C2 was achievable when as-sessed with manual palpation subject to imaging and visualobservations.

The purpose of this study was to compare the stiffness ofthe C1–C2 articulation in the intact state, after destabiliza-tion by the creation of a Type II odontoid peg fracture, thenafter restabilization with the new integrated device, andthen finally with a Harms technique at C1–C2 as the stan-dard technique at present for C1–C2 stabilization.

Methods

Specimens

Six fresh frozen human cadaveric cervical spine speci-mens (occiput-C4, mean age 54.666.6 years) were usedfor this study (Table 1). Radiographic screening was per-formed to exclude specimens with fractures, metastaticdisease, bridging osteophytes, or other conditions that couldsignificantly affect the biomechanics of the spine. Thespecimens were thawed and stripped of the paraspinal mus-culature while preserving the discs, facet joints, and osteoli-gamentous structures. Vertebral bodies from C2 through C4were stabilized to simulate fusion, using screws and bone ce-ment, leaving the occiput–C1 and C1–C2 segments mobile.The specimens were wrapped in saline soaked towels to

Table 1

Specimen demographics

Specimen Age (y) Gender Cause of death

1 59 M Lung cancer

2 53 M COPD

3 55 M Cardiac arrest

4 49 M Pancreatic cancer

5 47 M Myocardial infarction

6 65 M Heart failure

Mean6SD 54.666.6

M, male; COPD, chronic obstructive pulmonary disease; SD, standard

deviation.

prevent dehydration of the soft tissues. All tests were per-formed at room temperature.

Experimental setup

The specimens were fixed to the apparatus at the caudalend and free to move in any plane at the proximal end(Fig. 2). A moment was applied by controlling the flow ofwater into bags attached to loading arms fixed to the occiput.The apparatus allowed for continuous cycling of the speci-men between specified maximum moment end points inflexion-extension, lateral bending, and axial rotation. Theload displacement data were collected until two reproduc-ible load-displacement loops were obtained. This generallyrequired a maximum of three loading cycles.

The angular motion of the occiput and C1 vertebra rela-tive to C2 was measured using an optoelectronic motionmeasurement system (Model 3020, Optotrak; Northern Dig-ital, Waterloo, Ontario, Canada). In addition, biaxial anglesensors (Model 902-45; Applied Geomechanics, Santa Cruz,CA, USA) were mounted on each vertebra. A six-componentload cell (Model MC3A-6-1000, AMTI Multicomponenttransducers; AMTI Inc., Newton, MA, USA) was placed un-der the specimen to measure the applied moments. Fluoro-scopic imaging (GE OEC 9800 Plus digital fluoroscopymachine, Salt Lake City, UT, USA) was used in flexionand extension to monitor vertebra and implant motion,including implant-vertebra motion.

Testing protocol

Each specimen was tested in flexion-extension, lateralbending, and axial rotation to 61.5-Nm moments withoutpreload under the following sequential conditions:

1. Intact spine2. Destabilization of C1–C2 by creation of an odontoid

process fracture (Type II)3. Single-level stabilization at C1–C2 using the inte-

grated device (Fig. 1)

Fig. 2. Experimental testing setup of cadaveric cervical spine fixed dis-

tally and mobile cranially. Optical sensors are attached to occiput, C1,

and C2. The bending arm is applied to the occiput. Radiologic monitoring

was performed.

139P.A. Robertson et al. / The Spine Journal 12 (2012) 136–142

4. Single-level stabilization at C1–C2 using the Harmstechnique

Creation of instability

The odontoid fracture was created at the base of theodontoid to mirror a Type II fracture [37]. The base and lat-eral extent of the peg was identified, and anterior holeswere drilled to weaken the structure. The ‘‘fracture’’ wasthen established with an osteotome. Care was taken to en-sure complete division of the odontoid. No division of theligaments or joint capsule was performed, except for theopening of the posterior capsule of the C1–C2 lateral massjoint for placement of the integrated device.

Surgical stabilization

The surgical technique for insertion of the integrated de-vice required dissection anteriorly along the dorsolateral sur-face of the C2 lamina, until the C1–C2 lateral mass capsulewas identified. The capsule was opened, allowing access tothe lateral mass joint. After defining the joint space, a smallstraight curette allowed removal of the articular cartilagedown to subchondral bone. The integrated device was thenloaded onto an inserter and placed into the joint betweenthe articular surfaces. It was impacted into place fully untilthe lamina leg of the device contacted the C2 lamina. Afterscrew hole preparation, the oblique C1 screw was insertedthrough the screw hole in the blade to achieve bicorticalfixation. Finally the C2 lamina leg was fixed to the C2 laminawith 5-mm screws. The procedure was repeated to theopposite side.

The Harms procedure was performed as described previ-ously [19]. The two procedures were performed sequen-tially on the same specimen, as the initial use of theintegrated device did not compromise bone integrity thatwould disadvantage subsequent Harms fixation.

Statistical analysis

Normality of data for the C1–C2 ROM was checked us-ing the Kolmogorov-Smirnov (K-S) test. All data sets werenormally distributed (pO.30). C1/C2 ROM was analyzedwith repeated measures analysis of variance with Bonferronicorrection for three comparisons (Table 2). The four com-parisons made were:

Table 2

Bonferroni-corrected p values for flexion-extension, lateral bending, and

axial rotation ROM

C1–C2

Integrated device

vs. destabilized

Harms technique

vs. destabilized

Integrated device

vs. Harms

technique

Flexion-

extension

!.01 !.01 1.00

Lateral bending !.01 .011 .10

Axial rotation !.01 !.01 1.00

ROM, range of motion.

1. Intact versus destabilized2. Destabilized versus stabilized using the integrated

device3. Destabilized versus stabilized using the Harms technique4. Integrated device versus Harms technique

These comparisons were done separately for flexion-extension, lateral bending, and axial rotation, as no compar-isons across load types were intended in the study design.The statistical data analyses were performed with use ofthe Systat 10.2 software package (Systat Software, Rich-mond, CA, USA).

Results

All six specimens were successfully tested as per theprotocol, and the mean and standard deviation of each ofthe angular ROMs are tabulated for flexion-extension, lat-eral bending, and axial rotation at the C1–C2 articulationfor each of the four testing protocols (Table 3). The angularROM for each of the four testing protocols for flexion-extension, lateral bending, and axial rotation are repre-sented graphically in Fig. 3.

In the intact specimen, the flexion-extension ROM was14.1�62.9�. Destabilization increased this ROM to31.6�64.6�. The integrated device reduced flexion-extension ROM to 3.6�61.8� (p!.01 for comparison be-tween the destabilized state and the surgical technique).The Harms technique reduced flexion-extension ROM to4.0�61.4� (p!.01). In lateral bending, the respective angularrotations were 1.8�61.1�, 14.1�65.8�, 0.4�60.3�, and1.4�60.7�. For axial rotation, the intact ROM was67.3�613.8�, increasing to 74.2�616.1� with destabiliza-tion. The integrated device reduced ROM to 0.9�60.7�

(p!.01), and the Harms technique reduced ROM to1.4�60.7� (p!.01) There was no statistically significant dif-ference between the integrated device and the Harms tech-nique in flexion-extension (p51.0), lateral bending(p5.10), and axial rotation (p51.0).

Discussion

Rotation of the head on the trunk is a critical musculo-skeletal function that requires a large degree of rotation

Table 3

Mean and standard deviations for the C1–C2 ROM (degrees) in flexion-

extension, lateral bending, and axial rotation for each protocol step

C1–C2 Intact Destabilized

Integrated

device

Harms

technique

Flexion-

extension

14.162.9 31.664.6 3.661.8 4.061.4

Lateral bending 1.861.1 14.165.8 0.460.3 1.460.7

Axial rotation 67.3613.8 74.2616.1 0.960.7 1.460.7

ROM, range of motion.

Fig. 3. Flexion-extension, lateral bending, and axial rotation range of mo-

tion (meanþstandard deviation) for the intact specimens, destabilized, sta-

bilized with integrated device, and stabilized with the Harms technique.

140 P.A. Robertson et al. / The Spine Journal 12 (2012) 136–142

not provided by normal intervertebral articulations but ca-tered for by the unique anatomy of the C1–C2 joint. Asnoted, this specific anatomy renders the C1–C2 level liableto major instability when pathological changes occur, andthere has been considerable difficulty achieving surgicalstability at this level.

Initial surgical procedures focused on the posterior wireconstructs attached to the laminae of C1 and spinous pro-cess or laminae of C2 along with various options to braceacross or against structural corticocancellous bone graft[3,4]. Modification of fixation with clamps to the posteriorelements avoided passage of wire or cable within the canalat C1, yet did not improve the biomechanical stability[5–9]. These constructs are useful to restrict flexion andwill control anterior translation to some degree if the odon-toid peg is intact but have suboptimal stability in axialrotation or when atlantoaxial instability has developed.

Magerl and Seemann [10] introduced transarticularscrew fixation, with screws inserted obliquely from the in-ferior articular process of C2 across the pars of C2, acrossthe joint, and into the lateral mass of C1. Not surprisingly,this dramatically reduced axial rotation by fixing the lateralmass joints [8,9,11]. It was associated with much improvedstiffness of C1–C2, improved fusion rates and clinical out-comes [12–17], but required a screw pathway close to theVA in its foramen in C2, with potential for risk to the vessel[15,26–29]. Biomechanically, the Magerl technique did notcontrol flexion-extension as well as it limited axial rotation,and the addition of a posterior Gallie type construct wasrecommended to optimize multiplanar stability [8,11].

In 1994, Goel and Laheri [18] introduced the concept ofsegmental lateral mass fixation at C1–C2 with bilateralscrew/plate fixation. Harms modified this technique to usebilateral screw/rod instrumentation, and this technique hasthe most stiffness of the instrumentation constructs atC1–C2 [19–21]. The technique is associated with good

clinical outcomes and fusion rates [22–25]. It introducesa further point of risk to the vascular structures, requiringnot only a screw into the lateral mass of C2 adjacent tothe VAF [38] but also a screw into C1 medial to the VAF[39]. The C1 insertion point is relatively deep and can beassociated with significant venous bleeding, which mayimpede accurate starting and direction of the screw holepreparation.

In both the Magerl and Harms techniques, there is risk ofinjuring the VA in the VAF during screw hole preparationand screw placement. Several authors have identified theserisks and noted that the risks are increased if there is atlan-toaxial subluxation that does not reduce before screw place-ment, or if there is destructive disease to the joint surfacesof the lateral mass joint reducing the bone mass above theVAF [15,26–29]. A percentage of older patients seem tohave ‘‘high-riding’’ VAFs that will again reduce the spacefor a Magerl screw. There is a recommendation that allthese cases should undergo routine preoperative computedtomography assessment of the bone corridor [26], althoughstandard three-plane scans may still underestimate the riskto the VA even in ideal situations [31] and safe placementmay not be possible. Several authors have attempted to useother regions of the C1 and C2 vertebrae to obtain mechan-ical purchase [32–34].

Our goal was to design and test an implant that woulduse the C1 and C2 bony features that avoids the VAF inboth vertebrae and also avoids the sublaminar approachto the C1 lateral mass required in the Harms technique,which can be associated with difficult access because ofvenous bleeding.

This study demonstrates that the integrated deviceachieves similar stiffness at the C1–C2 level when com-pared with the Harms technique in a destabilized cadavericcervical spine. Previous studies have shown the Harmstechnique to be superior to the Magerl transarticular screwsand equivalent to a combined Magerl/Gallie technique[8,20]. In this study, the novel integrated device achievedincreased stiffness that was biomechanically equivalent tothat achieved by the Harms technique. Both the integrateddevice and the Harms technique have the advantage of min-imizing the flexion-extension motion without the need forsublaminar wires at C1.

The mechanisms of action of the integrated device arelikely to be multiple. The device blade separates the lateralmass joint surfaces of the C1–C2 joint, with visibly in-creased tension to the facet joint capsule. Subsequent serialligament division of a sample-stabilized specimen demon-strated the importance of the medial facet joint capsule inthis. It is likely that the integrated device has a further ef-fect on increasing the tension in the alar and apical liga-ments as the C1 lateral masses and occiput are elevatedin relation to C2 and its odontoid peg.

The serrated surface of the blade provides interference tothe normal translation of the joint surfaces that occurs inaxial rotation and the abnormal translation of the joint

141P.A. Robertson et al. / The Spine Journal 12 (2012) 136–142

surfaces that occurs in atlantoaxial instability. Interferencefixation has been used in other orthopedic procedures wheretwo parallel surfaces may move against each other [39–41].The technique relies on serrations, normally provided bya screw, to limit translation of two surfaces. Further stabili-zation of the blade interface in the integrated device isachieved with the screw fixation to C1 and C2. The C1 screwis placed through the blade of the integrated device, avoidingthe need for any other approach to C1. An oblique bicorticalscrew pathway further stabilizes the blade to C1 and uses themaximum bone mass available in the C1 vertebra. The C1oblique screw is angled parallel both to the pars of C2 andleg of the implant, giving an anterior cortical perforationpoint that is similar to the ideal points of anterior purchasewith the Magerl technique [15]. It is directed medially by20� to steer away from the VAF of C1 [30] and avoid theinternal carotid artery [42].

Bicortical fixation of this screw is necessary to obtainflexion-extension stability. It must be noted that in the firstspecimen tested, a unicortical screw was placed in C1 andthe flexion-extension ROM was 6.9� versus a mean of 2.9�

for the remaining five specimens where bicortical screwswere inserted.

The lower surface of the integrated device blade is fur-ther stabilized by its posterior or lamina leg, with screw fix-ation to the cortical bone in the superolateral surface of theC2 lamina (Fig. 1). This stabilization is as oblique as pos-sible to the C1–C2 lateral mass joint to optimize stability ofthe device on C2.

Serial deconstruction of the implanted C1–C2 articula-tion demonstrated that all these features combined to max-imize the rigidity of the fixed C1–C2 segment whensubjected to manipulation and inspection.

The comparison with the Harms technique was per-formed, as previous studies have demonstrated this to bethe most consistently effective stabilization technique forC1–C2 [20,21]. The integrated device required an intactC2 lamina, so it was tested in each specimen before theHarms technique was performed. A valid criticism of thismethodology is that the Harms technique was performed af-ter use and subsequent removal of the integrated device ineach specimen, and it is possible that the results from theHarms techniquemay have been ‘‘disadvantaged’’ by this or-der of testing. We do not think that this materially affectedthe outcomes for the following reasons. The destabilizationcreated by the performance of an osteotomy of the base ofthe odontoid peg to recreate a Type II peg fracture is fargreater than any destruction to the articular surface of theC1–C2 joints created by insertion of the integrated device.Further, we were careful to insert the C1 screws for theHarms technique with the C1 lateral mass screw of the inte-grated device still in situ ensuring that the C1 Harms screwhad a new bicortical pathway through the C1 lateral masswith a direct anterior approach and a path that avoided anyprevious screw tract. The C2 lateral mass screw for theHarms technique also used a pathway through intact bone

anterior and lateral to the C2 lamina leg screws of the inte-grated device. Finally, the validity of this order of testingwas proven by finding that the raw values for reduction inROM provided by the Harms technique were comparableto previously published values [20,21].

The clinical effectiveness of this integrated device is yetto be determined. Surgeons operating in this region shouldhave little difficulty accessing the posterior aspect of theC1–C2 lateral mass joint via the lamina of C2. This is a com-mon dissection plane free of the venous plexus that is oftenencountered under the lamina of C1 in the Harms technique.

Use of the integrated device would require reduction ofany atlantoaxial subluxation with anatomical alignment ofthe C1 and C2 vertebrae, as is required for the Magerl tech-nique. Inability to achieve reduction of atlantoaxial sublux-ation would require alternative approaches to stabilizationof C1 and C2.

Vertical migration of the odontoid (basilar invagination)secondary to C1–C2 joint destruction may be reducible us-ing this technique. Elevation of the C1 lateral mass wouldhave the potential to correct the vertical relation of C1and the occiput to the odontoid peg, reducing basilar invag-ination. Similar mechanisms were described by Goel [35]to achieve reduction of basilar invagination.

The integrated device has the potential to facilitate pos-terior bone grafting for fusion by virtue of the absence ofany fixation to the lamina of C1 and minimal coverage ofthe lamina of C2 by its lamina leg. The relative anterior po-sitioning of the device also gives the advantage of better vi-sualization of a posterior fusion bed on lateral radiographswithout instrumentation obscuring this view as occurs withthe Harms technique.

Conclusions

This study has demonstrated that the integrated deviceachieves equivalent ROM reduction at the C1–C2 levelwhen compared with the Harms technique in a destabilizedupper cervical spine model. It has the potential to provide analternative option for surgical stabilization of the C1–C2 ar-ticulation, and further development is warranted to optimizedesign before pilot clinical studies. Its clinical use andapplicability remain to be validated.

References

[1] White AA, Paniabi MM. Clinical biomechanics of the spine. 2nd ed.

Philadelphia, PA: Lipincott, 1990.

[2] Zhang H, Bai J. Development and validation of a finite element

model of the occipito-atlantoaxial complex under physiologic loads.

Spine 2007;32:968–74.

[3] Gallie W. Fractures and dislocations of the cervical spine. Am J Surg

1939;46:495–9.

[4] Brooks AL, Jenkins EB. Atlanto-axial arthrodesis by the wedge

compression method. J Bone Joint Surg Am 1978;60:279–84.

[5] Holness RO, Huestis WS, Howes WJ, Langille RA. Posterior stabili-

zation with an interlaminar clamp in cervical injuries: technical note

142 P.A. Robertson et al. / The Spine Journal 12 (2012) 136–142

and review of the long term experience with the method. Neurosur-

gery 1984;14:318–22.

[6] Cybulski GR, Stone JL, Crowell RM, et al. Use of Halifax interlam-

inar clamps for posterior C1–C2 arthrodesis. Neurosurgery 1988;22:

429–31.

[7] Moskovich R, Crockard HA. Atlantoaxial arthrodesis using interlam-

inar clamps: an improved technique. Spine 1992;17:261–7.

[8] Henriques T, Cunningham BW, Olerud C, et al. Biomechanical com-

parison of five different atlantoaxial posterior fixation techniques.

Spine 2000;25:2877–83.

[9] Grob D, Crisco JJ, Panjabi MM, et al. Biomechanical evaluation of

four different posterior atlantoaxial fixation techniques. Spine 1992;

17:480–90.

[10] Magerl F, Seemann P-S. Stable posterior fusion of the atlas and axis

by transarticular screw fixation. In: Kehr P, Weidner A, eds. Cervical

Spine I. Wien, Austria: Springer, 1987:322–7.

[11] Richter M, Schmidt R, Claes L, et al. Posterior atlantoaxial fixation:

biomechanical in vitro comparison of six different techniques. Spine

2002;27:1724–32.

[12] Grob D, Jeanneret B, Aebi M, Markwalder TM. Atlanto-axial fusion

with transarticular screw fixation. J Bone Joint SurgBr 1991;73:972–6.

[13] Dickman CA, Sonntag VK. Posterior C1-C2 transarticular screw fix-

ation for atlantoaxial arthrodesis. Neurosurgery 1998;43:275–80.

[14] Farey ID, Nadkarni S, Smith N. Modified Gallie technique versus

transarticular screw fixation in C1-C2 fusion. Clin Orthop Relat

Res 1999;359:126–35.

[15] Gluf WM, Schmidt MH, Apfelbaum RI. Atlantoaxial transarticular

screw fixation: a review of surgical indications, fusion rate, compli-

cations, and lessons learned in 191 adult patients. J Neurosurg Spine

2005;2:155–63.

[16] Wang C, Yan M, Zhou H, et al. Atlantoaxial transarticular screw fixa-

tion with morselized autograft and without additional internal fixation:

technical description and report of 57 cases. Spine 2007;32:643–6.

[17] Finn MA, Apfelbaum RI. Atlantoaxial transarticular screw fixation:

update on technique and outcomes in 269 patients. Neurosurgery

2010;66(3 Suppl):184–92.

[18] Goel A, Laheri V. Plate and screw fixation for atlanto-axial subluxa-

tion. Acta Neurochir (Wien) 1994;129:47–53.

[19] Harms J, Melcher RP. Posterior C1–C2 fusion with polyaxial screw

and rod fixation. Spine 2001;26:2467–71.

[20] Melcher RP, Puttlitz CM, Kleinstueck FS, et al. Biomechanical testing

of posterior atlantoaxial fixation techniques. Spine 2002;27:2435–40.

[21] Li S, Ni B, Xie N, et al. Biomechanical evaluation of an atlantoaxial

lateral mass fusion cage with C1-C2 pedicle fixation. Spine 2010;35:

E624–32.

[22] Aryan HE, Newman CB, Nottmeier EW, et al. Stabilization of the at-

lantoaxial complex via C-1 lateral mass and C-2 pedicle screw fixa-

tion in a multicenter clinical experience in 102 patients: modification

of the Harms and Goel techniques. J Neurosurg Spine 2008;8:222–9.

[23] Stulik J, Vyskocil T, Sebesta P, Kryl J. Atlantoaxial fixation using the

polyaxial screw-rod system. Eur Spine J 2007;16:479–84.

[24] Goel A, Desai KI, Muzumdar DP. Atlantoaxial fixation using plate

and screw method: a report of 160 treated patients. Neurosurgery

2002;51:1351–7.

[25] Payer M, Luzi M, Tessitore E. Posterior atlanto-axial fixation with

polyaxial C1 lateral mass screws and C2 pars screws. Acta Neurochir

(Wien) 2009;151:223–9.

[26] Madawi AA, Casey AT, Solanki GA, et al. Radiological and anatom-

ical evaluation of the atlantoaxial transarticular screw fixation tech-

nique. J Neurosurg 1997;86:961–8.

[27] Casey AT, Madawi AA, Veres R, Crockard HA. Is the technique of

posterior transarticular screw fixation suitable for rheumatoid

atlanto-axial subluxation? Br J Neurosurg 1997;11:508–19.

[28] Chen TY, Lin KL, Ho HH. Morphological characteristics of atlan-

toaxial complex in rheumatoid arthritis and surgical consideration

among Chinese. Spine 2004;29:1000–5.

[29] Lau SW, Sun LK, Lai R, et al. Study of the anatomical variations of

vertebral artery in C2 vertebra with magnetic resonance imaging and

its application in the C1-C2 transarticular screw fixation. Spine

2010;35:1136–43.

[30] Blagg SE, Don AS, Robertson PA. Anatomic determination of opti-

mal entry point and direction for C1 lateral mass screw placement.

J Spinal Disord Tech 2009;22:233–9.

[31] Spangenberg P, Coenen V, Gilsbach JM, Rohde V. Virtual placement

of posterior C1-C2 transarticular screw fixation. Neurosurg Rev

2006;29:114–7.

[32] Wright NM. Posterior C2 fixation using bilateral, crossing C2 lami-

nar screws: case series and technical note. J Spinal Disord Tech

2004;17:158–62.

[33] Mummaneni PV, Lu DC, Dhall SS, et al. C1 lateral mass fixation:

a comparison of constructs. Neurosurgery 2010;66(3 Suppl):153–60.

[34] Kelly BP, Glaser JA, DiAngelo DJ. Biomechanical comparison of

a novel C1 posterior locking plate with the harms technique in

a C1-C2 fixation model. Spine 2008;33:E920–5.

[35] Goel A. Treatment of basilar invagination by atlantoaxial joint dis-

traction and direct lateral mass fixation. J Neurosurg Spine 2004;1:

281–6.

[36] Tokuhashi Y, Ajiro Y, Oshima M, Umezawa N. C1-C2 intra-articular

screw fixation for atlantoaxial subluxation due to rheumatoid arthri-

tis. Orthopedics 2009;32:172.

[37] Anderson LD, D’Alonzo RT. Fractures of the odontoid process of the

axis. J Bone Joint Surg Am 1974;56:1663–74.

[38] Yoshida M, Neo M, Fujibayashi S, Nakamura T. Comparison of

the anatomical risk for vertebral artery injury associated with the

C2-pedicle screw and atlantoaxial transarticular screw. Spine

2006;31:E513–7.

[39] Zou D, Yoo J, Ordway N, et al. Interference screw fixation of cervical

grafts: a biomechanical study of a new method of cervical fixation.

J Spinal Disord 1991;4:168–76.

[40] Feldman KA. The principles of interference screw fixation: applica-

tion to foot and ankle surgery. J Foot Ankle Surg 2005;44:455–61.

[41] Giurea M, Zorilla P, Amis AA, Aichroth P. Comparative pull-out and

cyclic-loading strength tests of anchorage of hamstring tendon grafts

in anterior cruciate ligament reconstruction. Am J Sports Med 1999;

27:621–5.

[42] Murakami S, Mizutani J, Fukuoka M, et al. Relationship between

screw trajectory of C1 lateral mass screw and internal carotid artery.

Spine 2008;33:2581–5.