SPINE Volurne 28, Nurnber 16, pp 1821-1828@2003, Lippinco!! Williarns & Wilkins, 1nc.

11 Pedicle Screws Enhance Primary Stability in MultilevelCervical Corpectomies: Biomechanical In VitroComparison of Different Implants IncludingConstrained and NonconstrainedPosterior Instumentations

ReneSchmidt.MD,*Hans-JoachimWilke,PhD,tLutzClaes,PhD,tWolfhartPuhl,MD,*andMarcusRichter,MD*

Study Design.6 human eerviealspines wefe tested invitro in a biomeehanieal nondestruetive set-up to eom-pare different anterior, posterior and eombined instru-mentations after a eorpeetomy C4-C6.

Objectives. To evaluate the primary three-dimensionalstability of the different instrumentations.

Summary of Background Data. The elinieal results af-ter stabilization of multilevel eorpeetomies are often dis-appointing. Higher biomeehanical stability could enhaneethe rate of sueeessful outeomes. The best instrumentationfor these high-grade instabilities has yet to be found.

Methods. Six human eervieal speeimens were loadednondestruetively with pure moments and uneonstrainedmotion at C3n was measured. The six speeimens wereinstrumented with eaeh of the following fixation tech-niques: 1. Cage 2. Noneonstrained posterior serew androd system with lateral mass (NC-LM)3. and pedicleserews (NC-P) 4. Constrained posterior serew and rodsystem with lateral mass (C-LM)and 5. pedicle serews(C-P)6. Circumferential (C-Pand anterior plate) 7. Anteriorplate (OAP)

Results. For flexion/extension and axial rotation theeireumferential instrumentation showed lowest ROMval-ues, followed by C-P. The use of pedicle serews showedonlya lower ROMwhen using the eonstrained system. Nodifferenee was found between the two serew types in thenoneonstrained system. The anterior plating had the low-est stabilizing effeet of all instrumentations, exeept for theeage alone.

Conclusions. Usage of pedicle screws enhanees pri-mary stability only when using an eonstrained serew androd system. In axial rotation the nonconstrained systemshowed no distinet differenee eompared to the intaetstate, independent of the serew type. [Keywords: multi-level eorpeetomy, high-grade instability, biomechanies,primary stability, eonstrained and noneonstrained serewand rod systems, pedicle screws] Spine 2003;28:1821-1828

Anterior cervical corpectomy is an. increasingly usedtechnique for a wide diversity of disorders, including

From the *Departrnent of Orthopedics and SCI, University of Ulrn,Ulm, and the tDepartrnent of Orthopedic Research and Biornechanics,University of Ulrn, Ulrn, Germany.SuPpOrt was granted by Ulrich medizintechnik, Ulm, Gerrnany.Acknowledgrnent date: September 24, 2002. .first revision date: De-cember 18, 2002. Acceptance date: Decernber 19,2002.Address correspondence to Marcus Richter, MD, RKU, OrthopädischeKlinik mit Querschnittgelähmtenzentrum der Universität Ulrn, ObererEselsberg 45, Ulrn 89081, Gerrnany; E-mail: marcus.richter@RKU.de

degenerative changes,t-3 trauma,4,S infections, kyphoticdeformities after laminectomy,6,7 and neoplasms.8 Ac-cording to some authors, corpectomy is even recom-mended instead of multilevel interbody fusions due tobetter results.3,9,10The clinical results for one- or two-level corpectomies are good, with high fusion rates.lO,l1In contrast, the results and complications in multilevelcorpectomies (more than two levels) with early failurerates from 30% to 100%, dependent on the fusionlength, are often disappointing.3,6,10-13 As many of thecausing conditions underlie a progressive devolution, thesuccess rates of these technical challenging proceduresshould be enhanced to better serve the patients. Saunderset a/14 found that there is no unique morbidity in four-level corpectomies compared to a shorter decompressionrange, which would exclude this procedure. In vitro ex-periments with posterior fixations, including posteriorplating, have shown superior primary stability whencompared with anterior procedures1S-20in multilevel cor-pectomies.21 Recent screw and rod systems showed betterbiomechanical stability than posterior plate fixations, par-ticular for flexion-extension.22,23 Also, new constrainedscrew and rod systems were developed. Here, the linkagebetween screw, connector, and rod is rigid, thereby mini-mizing the motion between screw and rod under load. Pedi-cle screw fixations showed increased stability compared toconventional anterior or posterior constructs in multilevelcervical instabilityP When compared with lateral massscrews, they also showed in vitro higher pullout forces thatcould be beneficial in these cases.24Especially in the multi-level instability after corpectomy, a higher biomechanicalstability could enhance success rates and is therefore desir-able. Therefore, the purpose of this study was to evaluatethe primary stability of screw and rod systems in a multi-levelcorpectomy model under inclusion of lateral mass andpedicle screws and compare them to anterior plating andcircumferential fixation. This information could be a fur-ther step to improve methods for reconstruction in thesechallenging clinical cases.

. Material and Methods

Six human cadaveric specimens, 2 male and 4 female, mean age80 years (range 66-92 years), consisting of C2 to at least Th1or Th2, as obtained, were used. The specimens were examined,and plain radiographs were taken to exclude soft tissue or bone

1821

1822 Spine. Volume28. Number 16 . 2003

Figure 1. 1., Sawbonemodelwith anteriorplating and cage.B, Sawbonemodelwith nonconstrainedsystem:pediclescrews (left) andlateral mass screws (right). C, Sawbone model with constrained system: pedicle screws (left) and lateral mass screws (rightl.

damage and then stored frozen at - 20 C in triple-sealed plasticbags. After thawing, the muscle tissue was carefuIly removed,and aIl the ligaments and bony structures were preserved. Toprevent dehydration, specimens were kept moist with salinesolution. Handling specimens in the above-described mannerdoes not affect their biomechanical properties.25 Bone mineraldensity was measured in the vertebral bodies of C4 and C6 byquantitative computed tomography (QCT) after calibration ofthe CT (XCT 960, Stratec Medizintechnik GmbH, Pforzheim,Germany) with a standardized phantom. The cranial and cau-dal vertebrae were embedded in polymethylmetacrylate(PMMA; Technovit 3040, Heraeus Kulzer GmbH, Wehrheim,Germany). To obtain a better anchorage of the vertebrae in thePMMA, short screws were partiaIly driven into the embeddedparts of the vertebrae. Screws were inserted lateraIly in thevertebral bodies of C3 and C7 to fix the motion analysis systemto the specimen.

The corpectomy C4, C5, and C6 was created using rongeursand a high-speed air drill. The posterior longitudinal ligament(PLL)was preserved. The corpectomy was at least 16 mm wide.After decompression, a smaIl heilewas drilled in the cranial andcaudal endplate to accommodate the cage with a force sensor.The force sensor consisted of a miniature load ceIl (Miniatur-Druckkraftsensor Typ 8413, Burster Präzisionsmeßtechnik,Gernsbach, Germany) capable of.measuring axial compressiveforces only in a range between 0 and 500 N. The load ceIl wasmounted in a speciaIly modified cage on the basis of a routinelyused cage (ADD, Ulrich Medizintechnik, DIrn, Germany).Spikes at the superior and inferior end prevented the cage fromslipping. The cage was continuously adjustable to the desiredgraft height by either a screw thread at the lower end of the cageor by the use of two stainless steel tubes of different lengths. Theforce sensor was adjusted to 40 Nm preload while the specimenwas mounted in the spine tester with free movement of thespecimen in all directions. Thereby, every specimen could

adopt its own neutral position, and an identical axial preloadcould be adjusted.

The spinal instrumentations included the Cervifix system26(Synthes, Umkirch, Germany) with 3.5 mm lateral mass screws(NC-LM) and pedicle screws (NC-P) (Fig. 1B). The other pos-terior spinal implant was the neon occipitocervical system27.28(Ulrich Medizintechnik, DIrn, Germany) with 4.0 mm cannu-lated lateral mass (C-LM) and pedicle screws (C-P) (Fig. 1C).The anterior plate (OAP) was from the Osmium system (UlrichMedizintechnik, DIrn, Germany; Fig. 1A). The lateral massscrews were inserted according to the technique described byJeanneret et al.29The pedicle screws were inserted after visual-ization of the pedicle by dissecting the intervertebral foramen.The screws for the anterior plate were inserted with monocor-tical bone purchase. The circumferential instrumentation was a

Figure 2. Simulated screw back-out with the nonconstrained sys-tem with pedicle screws, left extension and right flexion. See themovement of the connector between screw head and bone sur-face by the initiated motion.

Pedicle Screws and Stability in Cervical Corpectomy .Schmidt et al 1823

Figure3. Simulatedscrew back-outwith the constrainedsystemwith pediclescrews,left extensionandrightflexion.Nomovementbetweenconnectorand screw headvisible.

combination of the OAP and the constrained screw and rod

system with pedicle screws (C-P).The specimen were mounted in a previous\y described spine

tester,30 where the caudal vertebrae were rigidly fixed in thetesting apparatus, and the cranial vertebra (C2) was fixed in aCardan joint containing integrated stepper motors that couldintroduce pure moments separately around the t'hree axes. Theother five out of six degrees of freedom were free, enabling thespecimen to move unconstrained. The segmental motion be-tween C3 and C7 was measured by a high-resolution, noncon-tacting ultrasound motion analysis system (Zebris, Isny, Ger-many; resolution 0.06°). Nondestructive loads were applied aspure moments in alternating sequences for right/left lateralbending (:!:Mx), flexion-extension (:!: My) and right/left axialrotation (:!: Mz). The different instrumentations were testedwith :!:1.5 Nm for all directions. To precondition the speci-mens and to minimize viscoelastic effects, they were tested withthree cyeles, but only the data of the third cyele was evaluat-ed.31 The total range of motion (ROM) of the segments wasdetermined for each direction of loading and for the differentloads. Range of motion was defined as the angular deformationat maximum load. The neutral zone as a measure of laxiry wasdefined at zero load as the difference in angulation between thetwo phases of motion. It is the area where the specimen movesbasically free of applied load. The spinal implants were testedaccording to the recommendations for the standardization of invitra stability testing of spinal implants.32

For the statistical analysis, the results were analyzed by theWilcoxon signed rank test as a nonparametric test by commer-cial statistic software (StatView 4.0, Abacus Concepts). AI-though we tested many conditions and several parameters, wedid not adjust the calculated P values for multiple testing. This

would have resulted in a great loss of information. If a P valuewas smaller than 0.05, we rated this as a tendency towards arelevant difference of the results compared.

. Results

The BMD was 179.9 mg/cm3 (median) in the vertebralbodies of C4 (range 99.3-303.6 mg/cm3) and 194.4mg/cm3 in C6 (range 159.4-274.1 mg/cm3).

The median of the ROM for the intact specimen was17.8° for lateral bending (Table 1). All the other instru-mentations had a lower median ROM value with P <0.05, except for the cage and the anterior plate (Table 1;Figure 4). The two pedicle screw fixations had ROMvalues of 0.4° (C-P) and 0.7° (NC-P) with P < 0.05 whencomparing them to the nonconstrained system with lat-eral mass screws (NC-LM). For the comparison with theconstrained system (C-LM), only the C-P had a P < 0.05(NC-P vs. C-LM, P > 0.05). Comparing the two lateralmass screws containing instrumentations showed no statis-tical difference;this was the case also in comparing the twopedicle screw instrumentations. For the circumferential in-strumentation only for the comparison with the noncon-strained screw and rod system with lateral mass screws(NC-LM) a statistical differenceoccurred (Table 1).

For flexion-extension, again the highest medianROM value occurred for the intact specimen with 26.6°,followed by the cage with 19.3° (P > 0.05; Table 2). Theanterior fixation (OAP) had the highest median ROMwhen compared to the posterior and circumferential in-strumentations (Figure 5; Table 2). For the two fixationswith lateral mass screws and the nonconstrained systemwith pedicle screws, no statistical difference occurred(P> 0.05; Table 2). The constrained system with pediclescrews (C-P) had the lowest value of the posterior instru-mentations with 1.4° (P < 0.05) when compared to theother posterior fixations. The lowest median ROM valueoverall was measured for the circumferential with 0.4°and a P < 0.05 when compared to all the other testedfixations (Table 2; Figure 5).

For axial rotation, the highest median ROM wasfound for the cage with 25.8°, followed by the anteriorplate (20.1°) and the intact state with 18.5° (Table 3;Figure 6). The anterior plate, the nonconstrained systemwith lateral mass screws and pedicle screws, had a P >

ROM = range of motion; NC-LM = nonconstrained screw and rod system with lateral mass screws; C-LM = constrained system with lateral mass screws;NC-p = nonconstrained screw and rod system with pedicle screws; Cop = constrained system with pedicle screws; 3600 = circumferential fixation; OAP =anterior plate.

Table 1. Median ROM Values for lateral Bendingin De'greesand P Values

ROM Intact Cage NC-LM C-LM NC-P Cop 360° OAP

P Value 17.8 12.0 1.9 0.9 0.7 0.4 0.5 13.4Cage 0.1159NC-LM 0.0277 0.0277C-LM 0.0277 0.0277 0.0747NC-P 0.0277 0.0277 0.0277 0.1441Cop 0.0277 0.0277 0.0277 0.0431 0.0679360° 0.0277 0.0277 0.0431 0.0679 0.5002 0.0679OAP 0.1159 0.3454 0.0277 0.0277 0.0277 0.0277 0.0277

1824 Spine' Volume28. Number 16 . 2003

Intact

Anterior plate

G-P

G-LM

NC-P

Figure 4. Bilaterallateral bend-ing under a load ot :!:1.5Nm.Me-dian. minimal.and maximal val-ues in degrees tor ROMand NZ.NC = nonconstrained system;C = constrained system; LM =lateralmassscrews;P= pediclescrews.

NG-LM

Cage

Circumferential

o

0.05 when compared to the intact state. The constrainedsystem had median va lues of 6.20 for lateral mass and3.00 for pedicle screws with a consecutive P < 0.05 whencompared to the anterior and nonconstrained posteriorfixations, as well as the intact state. The circumferentialfixation had again the lowest overall ROM value of 1.60and P < 0.05 in comparison with all other tested instru-mentations. The neutral zone as a measure of laxity forthe different instrumentations showed similar trends asthe ROM with smaller median values (Figures 4-6).. Discussion

This biomechanical in vitra study was performed to de-termine the primary stabilizing effect of different cervicalfixation devices in a three-Ievel corpectomy model.

The anterior plating achieved a higher stability thanthe intact state for lateral bending, not for flexion-extension or axial rotation. Furthermore, for all motiondirections, a lower primary stability occurred when com-pared to the posterior and circumferential instrumenta-tions. Clinical studies found higher failure rates for theanterior plating when more levels were involved10,1l,33and a potential benefit for po.sterior or circumferential

10 20 30 40

instrumentations in multilevel instabilities.34-36 Thisstands in contrast to the good results for one- or two-level instabilities. With a longer decompression, a high erinstability is produced. The resulting longer lever arm, aswell as the loss of distraction by graft or cage settling,seem to overcome the stability potential of anteriorplates. The higher failure rate leads to reoperation or, incase of pseudarthrosis, to a poor clinical outcome.37,38The benefit of corpectomy and anterior stabilization isthe prevention of an additional posterior approach to thespme.

For flexion-extension, the two systemswith lateral massscrews and the nonconstrained system with pedicle screwsshowed similar ROM values. Incontrast to lateral bending,the higher screw pullout forces of pedicle screws comparedwith lateral mass screws24did not automatically enhancethe stability of the instrumentation for flexion-extension.As for the constrained system with pedicle screws, a lowerROM was found when compared to the other posteriorinstrumentations; the effectof the screw type for stability isdependent upon the screw and rod system used. The fixa-tion between the screw and the connector to the rod is likea shim in the nonconstrained system. If the screw com-

ROM = range of motion; NC-LM = noneonstrained serew and rod system with lateral mass serews; C-LM = eonstrained system with lateral mass serews;NC-P = noneonstrained serew and rod system with pedicle serews; Cop = eonstrained system with pe diele serews; 3600 = eireumferential fixation; OAP =anterior plate.

Table 2. Median ROM Values for Flexion-Extension in Degrees and P Values

ROM Intact Cage NC-LM C-LM NC-P Cop 360' OAP

PValue 26.6 19.3 5.8 4.4 3.5 1.4 0.4 6.9Cage 0.5002NC-LM 0.0277 0.0277C-LM 0.0277 0.0277 0.1730NC-P 0.0277 0.0277 0.7532 0.7532Cop 0.0277 0.0277 0.0277 0.0464 0.0277360' 0.0277 0.0277 0.0277 0.0277 0.0277 0.0277OAP 0.0277 0.0277 0.0464 0.0277 0.0277 0.0277 0.0277

Pedicle Screws and Stability in Cervical Corpectomy .Schmidt et al 1825

Intaet

Anterior plate .

G-LM

NC-P

NC-LM

Figure5. Flexion-extensionun-der a load of :t1.5Nm. Median,minimal,and maximalvalues indegreesfor ROMand NZ.NC =noneonstrained system; C =constrainedsystem;LM = lat-eral massscrews; P = pediclescrews.

Cage

Cireumferential

o

pressestheconnectorto the boneanapproximativeangle,astable (constrained) situation is achieved. But as the surfaceof the bone is uneven, in contrast to the plain surface of theconnector, the material properties of the smooth metal sur-face and the moist surrounding tissue are forwarding glid-ing processes, and the tightening of the screw to the bone isdependent on the predetermined bone stock, this approxi-mative stable angle situation is sometimes difficult toachieve. Especiallyin osteoporotic bone stock, the tighten-ing of the screw always bears the risk of overwinding thescrewand thereby damaging the bone stock. Another pointof more clinical interest is the backing up of screws, some-times seen in radiograph routine checkups. This implicatesa loosening of screw stability by shortening of the screw-bone contact surface area, but for the nonconstrained sys-tem, it implicates a further lass of stability by loosening theconjunction between screw and connector and thereby con-secutiveobligatory between screw and rod. In the simulatedscrew back-out in Figure 2, the connector is gliding to thebone surface in extension and to the screw head in flexion.In the constrained system, the fixation of the screw to theconnector and thus to the rod is independent of the tight-ening between screw and bone stock. For the constrained

10 20 30 40

system, the linkage between screw and connector is depen-dent upon tightening a screw nut to a screw thread, whichalso clamps the rod to the connector. This is only a mechan-icallinkage, which is unaffected by the surrounding ana-tomic conditions and consequently highly reproducible(Figure 3). As long as the appealing forces are not largerthan the stability of the screw thread-screw nut conjunc-tion, the stability of the system is mainly determined by thetightening of the screw to the bone. In a clinical situation,the loosening of the screw thread-screw nut linkage by ap-pealing forces is unlikely, as such forces would probablyfirst damage anatomic structures and lead, for example, tofracture of the lateral mass.

For axial rotation, the differences were even more dis-tinct than for flexion-extension (Tables 2 and 3; Figures5 and 6). Neither the anterior plate nor the noncon-strained system with both screw types had a distinctlylower ROM than the intact state. And again, within thenonconstrained system, no difference in stability for thedifferent screws appeared, in contrast to the constrainedsystem. Subsequently, the stability of the constrained sys-tem is mainly determined by the screw length and thetightening to the bone stock, resulting in differences of

ROM = range of motion; NC-LM = nonconstrained screw and rod system with lateral mass screws; C-LM = constrained system with lateral mass screws;NC-p = nonconstrained screw and rod system with pedicle screws; Cop = constrained system with pedicle screws; 3600 = circumferential fixation; OAP =anterior plate.

Table 3. Median ROM Values for Axial Rotation in Degrees and P Va lues

ROM Intaet Cage NC-LM C-LM NC-P Cop 3600 OAP

PValue 18.5 25.8 12.2 6.2 14.0 3.0 1.6 20.1Cage 0.0277NC-LM 0.2489 0.0277C-LM 0.0277 0.0277 0.0464NC-p 0.3454 0.0277 0.2489 0.0277Cop 0.0277 0.0277 0.0277 0.0431 0.02773600 0.0277 0.0277 0.0277 0.0277 0.0277 0.0277GAP 0.7532 0.0464 0.0277 0.0277 0.0277 0.0277 0.0277

1826 Spine' Volume28' Number 16 . 2003

Intact

Anterior plate

C-P

G-LM

NC-P

NG-LM

Figure6. Bilateralaxial rotationundera loadof :t 1.5Nm.Median,minimal,and maximalvalues indegreesfor ROMand NZ.NC =nonconstrained system; C =constrainedsystem;LM = lat-eral massscrews; P = pediclescrews.

Cage

Circumferential

o

stability for lateral mass and pedicle screws, whichstands in contrast to the nonconstrained system. The in-fluence of the screw-connector and connector-rod link-

age, under the used force moments in this in vitro setting,seem to be negligible for the constrained system.

The differences between the two systems, besides theaspect of angle stability, were the diameter of screws androds and the material used. The screw trajectories werethe same, so that this could not be a valid difference. Thescrew diameter was 4.0 mm for the constrained and 3.5mm for the nonconstrained system. This was due to thefact that the constrained system was used with cannu-lated screws, which is technically not feasible for a diam-eter less than 4.0 mm. If this had been the reason for thedifferent results for flexion-extension and axial rotation,one could expect that within the system, a difference ofthe stability achieved by the different screws would stilloccur. This was not the case, as for the nonconstrainedsystem (P > 0.05) when comparing lateral mass andpedicle screws, in contrast to the constrained system. Thediameter of the rods, as well as the material used, shouldalso not affect the results within the different systems, aswe did not see any deformation of the rods by the appliedforces.

The circumferential instrumentation had for flexion-extension and axial rotation the lowest median ROM

10 20 30 40

values, whereas for lateral bending, as among the poste-rior instrumentations, the results were not so distinct. Asthe circumferential consisted of the constrained systemwith pedicle screws, which was the most stable posteriorfixation and an additional anterior fixation, it is not sur-prising that this instrumentation was the most stable.The questions are more how much stability is needed fora solid fusion, and does this imply the necessity of acombined anterior and posterior fixation. These ques-tions can only be answered by long-term clinical studiescomparing constrained posterior instrumentations withcircumferential fixations.

This study has some limitations. We did not random-ize the testing sequence of implants to minimize the effectof additional instrumentation. Therefore, the cage wasalways tested first, followed by the posterior fixationswith lateral mass and pedicle screws. The noncon-strained system was always tested first due to the lowerscrew diameter, thereby allowing the use of the samescrew trajectories. The last tested posterior fixation (C-P)was then extended 10 a circumferential fixation by add-ing the anterior plate, and after removal of the rod sys-tem, the anterior plate was tested alone. The pediclescrews remained so that possible damage of the vertebralbody bone stock, which could weaken the fixation of theanterior plate, was prevented. As the last tested con-

lateral Bending

Table 4. InstrumentationsListedAfter Median ROM, Lowest ROM at the Head of List

Axial RotationFlexion-Extension

ConstrainedpedicleCireumferentialNoneonstrainedpedieleConstrainedlateral massNoneonstrainedlateralmassCageAnteriorplateIntaetstate

CireumferentialConstrainedpedicleNoneonstrainedpedieleConstrainedlateral massNoneonstrainedlateral massAnterior plateCageIntaet state

CireumferentialConstrainedpedieleConstrainedlateralmassNoneonstrainedlateral massNoneonstrainedpedicleIntaet stateAnterior plateCage

Pedicle Screws and Stability in Cervical Corpectomy .Schmidt et al 1827

strained posterior fixation showed a lower ROM, it seemsunlikely that the testing sequence had a significant effectforthe posterior instrumentations, but the anterior plate withhigher median ROM in all directions compared to the pos-terior fixations could have been affected. Nevertheless, thesuperior stability of posterior instrumentations was alsofound by other authors.21 The used anterior plate did notprovide a locked screw-plate conjunction, which in vitrashowed a significantly increased rigidity that could enhancestability.39,40Also, by using a bicortical fixation, we couldhave enhanced stability40 at the cost of a higher risk inclinical use. Another point that a biomechanical study canhardly accommodate for is the stabilizing potential of thespine surrounding tissues, especiallythe muscles, which cancounterbalance occurring forces in the clinical case andtherefore even out a lower biomechanical stability of im-plants. In addition, the small number of tested spedmens,the high age of the specimens, and the interindividual dif-ferencesin biomechanical properties reduce the significanceof the results obtained and limit the conclusions that can bedrawn from this study for clinical use.

Summarizing the results of this biomechanical in vitrastudy, the sole anterior plating did not provide a satisfyingprimary stability in this multilevelcorpectomy model. Thestability of the used posterior instrumentations for axialrotation and flexion-extension depends crucially onwhether constrained or nonconstrained systems were used.The stability of the constrained system can be enhanced bythe use of pedicle screws instead of lateral mass screws,whereas the higher risk of the screw trajectory for the non-constrained system only offers a slight benefit for lateralbending, not for flexion-extension and axial rotation. Fur-thermore, this benefit is not evident when comparing to theconstrained system with lateral mass screws. When a highbiomechanical stability in cervical multilevel corpecromiesis desired, a constrained posterior instrumentation shouldbe preferred. Whether this has to be combined with addi-tional anterior plating for clinical use has to be determined,as with all results of biomechanical in vitra studies, by long-term clinical srudies.

.IIKey Points

. Pedicle screws enhance .primary. stability whenusing a con,strained screw and rod system in a mul-tilevel corpectomy.· Constrained screw and rod systems provided dis-tinctly higher stability fot.'axial rotation in compar-ison with thenonconstrained system, independentof screw type used.· Anterior plating had only a distinct difference instability for flexion~extension, not for lateralbending or axial rotation, when compared to themtact state.· Circumferential instrumentation showed thelowest median ROM values for flexion-extensionand axial rotation.

AcknowledgmentThe authors thank Ulrich Medizintechnik Co. for its do-nation of the funds to realize this study.

References

1. Bernard TN, jr, Whilecloud TS 3rd. Cervieal spondylotie myelopalhy andmyeloradieulopathy. Amerior deeompression and slabilizalion wilh auioge-nous /ibula Slrul grafr. Clin Orthop 1987;221:149-60.

2. Emery SE, Bohlman HH, Bolesta Mj, et al. Amerior eervieal deeompressionand arrhrodesis for the Ireatmem of eervieal spondylolie myelopalhy. Two-to sevemeen-year follow-up. J Bane Joint Surg Am 1998;80:941-51.

3. Yonenobu K, Fuji T, Ono K, el al. Choiee of surgieal treatmem for multiseg-memal eervieal spondylotie myelopathy. Spine 1985;10:710-6.

4. Anderson PA, Bohlman HH. An~erior deeompression and arrhrodesis of Iheeervieal spine: long-Ierm motor improvemem. Parr 11.improvemenr in eom-plete traumatie quadriplegia. J Bane Joint Surg AI" 1992;74:683-92.

5. Bohlman HH, Anderson PA. Amerior deeompression and arrhrodesis of theeervieal spine: long-term motor improvement. Parr I-Improvement in ineom-plete traumatie quadriparesis. J BoneJoint SurgAm 1992;74:671-82.

6. Riew KD, Hilibrand AS, Palumbo MA, el al. Amerior eervieal eorpeetomy inpatients previously managed with a lamineetomy: short-term eomplieations.J Bane Joint Surg Am 1999;81:950-7.

7. Zdebliek TA, Bohlman HH. Cervieal kyphosis and myelopathy. Treatmentby amerior eorpeetomy and srrut-grafring. J Bane Joint Surg Am 1989;71:170-82.

8. Siegal T, Tiqva P. Verrebral body reseelion for epidural eompression bymalignant tumors. Results of forry-seven eonseeutive operalive proeedures.JBane Joint Surg Am 1985;67:375-82.

9. Hilibrand AS, Fye MA, Emery SE, et al. Inereased rate of arrhrodesis withStrut grafting after multi level anterior eervieal deeompression. Spine2002;27:146-51.

10. Swank ML, Lowery GL, Bhat AL, et al. Anterior eervieal allograft arthro.desis and instrumentation: mullilevel inrerbody grafting or strut graft reeon-Struetion. Eur Spine J 1997;6:138-43.

11. Vaeearo AR, Falaryn SP, Seuderi Gj, et al. Early failure of long segmentanterior eervieal plate /ixation. J Spinal Disord 1998;11:410-5.

12. Fernyhough jC, White jl, LaRoeea H. Fusion rates in multilevel eerviealspondylosis eomparing allograft /ibula with autograft /ibula in 126 patients.Spine 1991;16(suppl):S561-4.

13. Maedonald RL, Fehlings MG, Tator CH, et al. Multilevel anterior "rviealeorpeetomy and /ibular allograft fusion for eervieal myelopathy. JNeurosurg1997;86:990-7.

14. Saunders RL, Pikus Hj, Ball P. Four-Ievel eervieal eorpeetomy. Spine 1998;23:2455-61.

15. BueffHU, Lotz jC, Colliou OK, et al. Instrumentation of the eervieothoraeiejunetion after destabilization. Spine 1995;20:1789-92.

16. Do Koh Y, Lim TH, Won You j, et al. A biomeehanieal eomparison ofmodern anterior and posterior plate /ixation of the eervieal spine. Spine2001;26:15-21.

17. Kotani Y, Cunningham BW, Abumi K, et al. Biomeehanieal analysis of cer-vieal stabilizalion systems. An assessmenr of rranspedieular serew /ixation inthe eervieal spine. Spine 1994;19:2529-39.

18. Riehman jD, Daniel TE, Anderson DD, et al. Biomeehanieal evaluation ofeervieal spine stabilization methods using aporeine model. Spine 1995;20:2192-7.

19. Sutterlin CEd, MeAfee PC, Warden KE, et al. A biomeehanieal evalualion ofeervieal spinal stabilization methods in a bovine model. Statie and eydiealloading. Spine 1988;13:795-802.

20. Ulrieh C, Wörsdorfer 0, Claes L, et al. Comparative stUdy of the stabiliry ofanterior and posterior eervieal spine /ixation proeedures. Arch OrthopTrauma Surg 1987;106:226-31.

21. Kirkparriek jS, Levy JA, Carillo j, et al. Reeonsrruetion after multileveleorpeetomy in the eervieal spine. A sagittal plane biomeehanieal sludy. Spine1999;24:1186-90; diseussion 1191.

22. Cunningham BW, Seiter jC, Shono Y, et al. Static and eydieal biomeehaniealanalysis of pedicle serew spinal eonStruets. Spine 1993;18:1677-88.

23. Grubb MR, Currier BL,Stone j, et al. Biomeehanieal evaluation of posterioreervieal stabilization after a wide lamineetomy. Spine 1997;22:1948-54.

24. jones EL, Heller jG, Silcox DH, et al. Cervieal pedide serews versus lateralmass serews. Anatomie feasibiliry and biomeehanieal eomparison. Spine1997;22:977-82.

25. Panjabi MM, Krag M, Summers D, et al. Biomeehanieal time-toleranee offresh eadaverie human spine speeimens. J Orthop Res 1985;3:292-300.

26. Horgan MA, Kellogg jX, Chesnut RM. Posterior eervieal arrhrodesis and

1828 Spine. Volume28 . Number 16.2003

stabilization: an early report using a novellateral mass screw and rod tech-nique. Neuros"rgery 1999;44:1267-71; discussion 1271-2.

27. Richter M, Schmidt R, Claes L, et al. Posterior atlantoaxial fixation: biome-chanieal in vitro comparison of six different techniques. Spine 2002;27:1724-32. .

28. Richter M, Wilke HJ, Kluger P, et al. Biomechanical evaluation of a newmodulat rod-screw implant system for posterior instrumentation of the oc-cipiro-cervical spine: in-vitro comparison with rwo established implant sys-tems. Eur Spine] 2000;9:417-25.

29. Jeanneret B, Magerl F, Ward EH, et al. Posterior stabilization of the cervicalspine with hook plates. Spine 1991;16(suppl):S56-63.

30. Wilke HJ, Claes L, Schmitt H, et al. A universal spine tester for in vitroexperiments with muscle force simulation. Eur Spine] 1994;3:91-7.

31. Panjabi MM. Dreidimensionale tesrung der stabilität von wirbelsäulen im-plantaten. Orthopäde 1991;20:106-111.

32. WiIke HJ, Wenger K, Claes L. Tesring erireria for spinal implants-recommendarions for rhe srandardizarion of in virro stabiliry resting of spinalimplanrs. Eur Spine] 1998;7:148-54.

33. Para more CG, Dickman CA, Sonnrag VK. Radiogtaphie and clinieal fol-low-up review of Caspar plares in 49 parienrs.] Neurosurg 1996;84:957-61.

34. EISaghir H, Bohm H. Anterior versus posterior plaring in eervieal eorpee-romy. Arch Orthop Trauma Surg 2000;120:549-54.

35. Foley KT, Smith MM, Wiles DA. Anrerior cervical plaring does nor prevenrstrur grafr displacement in multilevel cervieal eorpeeromy. Orthop Trans1998;22:512-513.

36. Swank ML, Sutterlin CE 3rd, Bossons CR, er al. Rigid inrernal fixarion wirh

lareral mass plares in mulrilevel anrerior and posrerior reeonsrruerion of rheeervical spine. Spine 1997;22:274-82.

37. Newman M. The ourcome of pseudarthrosis after eervical anrerior fusion.Spine 1993;18:2380-2.

38. Phillips FM, Carlson G, Emery SE, er al. Anterior eervical pseudarthrosis.Natural hisrory and rreatmenr. Spine 1997;22:1585-9.

39. Grubb MR, Currier BL,Shihj-S, er al. Biomeehanieal evaluarion of anrerior

eervieal spine srabilizarion. Spine 1998;8:886-892.40. SpivakJM, Chen D, Kummer FJ. The effeer of loeking fixation screws on rhe

srabiliry of anterior eervieal plating. Spine 1999;24:334-8.