SPINE Volume 33, Number 1, pp 19–26©2008, Lippincott Williams & Wilkins, Inc.

The Effect of Dynamic Posterior Stabilization on FacetJoint Contact ForcesAn In Vitro Investigation

Christina A. Niosi, MASc,*† Derek C. Wilson, BSc,† Qingan Zhu, PhD,† Ory Keynan, MD,†David R. Wilson, DPhil,*† and Thomas R. Oxland, PhD*†

Study Design. Facet contact forces in the lumbar spinewere measured during flexibility tests using thin film elec-troresistive sensors in intact cadaveric spine specimensand in injured specimens stabilized with a dynamic pos-terior system.

Objective. The purpose of this study was to investigatethe effect of the Dynesys system on the loading in thefacet joints.

Summary of Background Data. The Dynesys, a poste-rior nonfusion device, aims to preserve intersegmentalkinematics and reduce facet loads. Recent biomechanicalevidence showed that overall motion is less with theDynesys than in the intact spine, but no studies haveshown its effect on facet loads.

Methods. Ten human cadaveric lumbar spine speci-mens (L2–L5) were tested by applying a pure moment of�7.5 N m in 3 directions of loading with and without afollower preload of 600 N. Test conditions included anintact specimen and an injured specimen stabilized with 3Dynesys spacer lengths. Bilateral facet contact forceswere measured during flexibility tests using thin film elec-troresistive sensors (Tekscan 6900).

Results. Implanting the Dynesys significantly in-creased peak facet contact forces in flexion (from 3 N to22 N per side) and lateral bending (from 14 N to 24 N perside), but had no significant effect on the magnitude of thepeak forces in extension and axial rotation. Peak facetloads were significantly lower with the long spacer com-pared with the short spacer in flexion and lateral bending.

Conclusion. Implantation of the Dynesys did not affectpeak facet contact forces in extension or axial rotationcompared with an intact specimen, but did alter theseloads in flexion and lateral bending. The spacer lengthaffected the compression of the posterior elements, witha shorter spacer typically producing greater facets loadsthan a longer one.

Key words: lumbar facets, facet contact force, nonfu-sion, biomechanics, stabilization. Spine 2008;33:19–26

Dynamic posterior stabilization is an alternative to fu-sion for the treatment of degenerative problems in thelumbar spine, although both are performed in an attemptto reduce pain. Several clinical studies have reported thatfusion may accelerate degeneration at adjacent levels.1–4

The Dynesys (Zimmer GmbH, Winterthur, Switzer-land), a dynamic posterior stabilization system, is de-signed to preserve intersegmental kinematics and reduceloading at the facet joints. It is a bilateral device thatconsists of titanium alloy pedicle screws and polycarbon-ate urethane spacers that surround tensioned polyethyl-ene terephthalate cords. The spacers are designed to sup-port compressive loads whereas the tensioned cordsstabilize the segment and act against tensile loads.

Assessing load transmitted by the facet joints is a crit-ical component of the biomechanical evaluation of manyspinal implants. The facet joints contribute to the stabil-ity of the spine, transfer load through the spinal column,and restrict motion between vertebrae, specifically in ax-ial rotation, extension, and translation in the anteropos-terior direction.5–7 Under varying degrees of extension,facet contact forces have been found to range between10% and 40% of the applied load5,8–11 and in flexion,facet forces haves been shown to be minimal.10–12 Facetjoints can also be a common site of low back pain.13

Facet contact forces have been measured in intactloaded cadaver specimens using both indirect and directtechniques. Indirect measurement of facet loads involvedsequential removal of spinal structures,8 the insertion ofpressure transducers,14–16 or intervertebral load cells5

into the intervertebral disc to quantify anterior columnload. The load in the facet joints was then inferred fromthe difference between the measured intradiscal load andthe total applied axial compressive load. Facet contactforces were also estimated by placing strain gauges onthe superior articular processes.17 This technique ishighly sensitive to the placement and orientation of thegauges18 and is destructive because it requires the motionsegment to be disarticulated for calibration. Facet loadshave been measured directly by insertion of Fuji PrescaleFilm into the joint space to statically measure facetloads.9,10,19 This approach is invasive because the jointcapsule must be sectioned to accommodate the film.

The effect of the Dynesys system on intervertebral ki-nematics has been studied, but its effect on facet loads isnot clear. One study showed that the Dynesys restrictedrotation in flexion–extension compared with an intactspine.20 Other more recent investigations found that sta-

From the Departments of *Mechanical Engineering and †Orthopae-dics, University of British Columbia and Vancouver Coastal HealthResearch Institute, Vancouver, British Columbia, Canada.Acknowledgment date: February 17, 2007. Revision date: May 22,2007. Acceptance date: June 24, 2007.The legal regulatory status of the device(s)/drug(s) that is/are the sub-ject of this manuscript is not applicable in my country.Corporate/Industry, Federal, and Foundation funds were received insupport of this work. Although one or more of the author(s) has/havereceived or will receive benefits for personal or professional use from acommercial party related directly or indirectly to the subject of thismanuscript, benefits will be directed solely to a research fund, founda-tion, educational institution, or other nonprofit organization which theauthor(s) has/have been associated.Address correspondence and reprint requests to Thomas R. Oxland, PhD,Division of Orthopaedic Engineering Research, Department of Orthopae-dics, Vancouver General Hospital, 828 W 10th Avenue, 5th Floor, Van-couver, BC, Canada V5Z 1L8; E-mail: toxland@interchange.ubc.ca

19

bilization with the Dynesys yielded a more flexible seg-ment than internal fixation, but was generally stiffer thanthe intact spine.21,22 In one of these studies, variation ofthe spacer length of the Dynesys resulted in significantdifferences in range of motion,22 with the range of mo-tion being larger with a longer spacer. In addition, therewere significant changes observed in the motion pattern,as described by the helical axis of motion (HAM), inspecimens stabilized with the Dynesys system and thesechanges in the motion pattern could lead to alterations infacet loading. Although the Dynesys was designed toreduce loads at the facet joints, there have been no stud-ies that show how the contact forces are affected by sta-bilization with the Dynesys system. Also, the spacerlength is presumably an important parameter that di-rectly influences intersegmental motion and loading, be-cause it determines the segmental position and there havebeen no studies that show how the contact forces areaffected by spacer length.

The purpose of this in vitro study was to investigatethe effect of the Dynesys dynamic stabilization system onthe transmission of load through the facet joints and toexamine the effect of the length of the Dynesys spacer onfacet contact forces.

Materials and Methods

Ten fresh-frozen cadaveric lumbar spine segments (L2–L5)were used in this study. The average age of the donors was 77years (range, 70–88) and there were 6 males, 3 females, and 1donor of unknown gender. Each specimen was dissected ofmusculature, such that the soft tissues were not damaged. TheL2 and L5 vertebrae were potted in dental stone mounts withthe L3–L4 disc space horizontal.

Flexibility tests were performed on each specimen under 4different conditions:

i. Intact (sectioned facet capsules)ii. Injury and Dynesys implant at L3–L4 (standard length

spacer)iii. Injury and Dynesys implant at L3–L4 (long spacer)iv. Injury and Dynesys implant at L3–L4 (short spacer).

The testing order of conditions (ii) through (iv) was random-ized to eliminate variability because of test sequence. A custom-designed and fabricated spine testing machine applied a puremoment of �7.5 N–m to the top vertebra while the specimenwas allowed to move without constraint in 3 dimensions23

(Figure 1). The continuous load was applied in displacementcontrol at a rate of approximately 1.3°/s in flexion–extension,lateral bending, and axial rotation to a load limit for 3 completelyreversed cycles. All tests were performed both with and without acompressive follower preload of 600 N, which was used to simu-late physiologic compressive loading.24 The path of the followerpreload followed the contour of the specimen and was optimizedsuch that each level experienced nearly pure compression.

Injury representing severe instability was performed by aspine surgeon. This consisted of sectioned facet joint capsules,a posterolateral nucleotomy, and transection of the supraspi-nous and interspinous ligaments at the L3–L4 level.

The Dynesys was implanted using manufacturer suppliedinstruments and recommended procedures. Pretension of 300

N was applied to the cord during implantation and the order oftensioning was alternated randomly between the right and leftsides. Although not part of the normal operative procedure, thepedicle screws were cemented in place using polymethyl-methacrylate to ensure that there was no loosening at the bone–screw interface, thereby reducing test variability. The spacersused in these cadaver experiments were manufactured with amodified stiffness to account for testing in a room temperatureenvironment. Thus, the stiffness of the tested spacer at roomtemperature was equivalent to the stiffness of the Dynesysspacer at body temperature. In this investigation, 3 differentspacer lengths were examined: standard length as determinedby a spine surgeon to fit between the pedicle screws and tomaintain a neutral position of the spine; 2 mm longer than thestandard length; and 2 mm shorter than the standard length.

Facet contact forces were measured dynamically duringeach of the flexibility tests using thin film electroresistive sen-sors and I-Scan software (Tekscan Inc., South Boston, MA)(Figure 2). One sensing “finger” of the 4 finger Tekscan 6900Quad sensor was inserted into each of the right and left facetjoints after sectioning of the joint capsules. New sensors wereused for each specimen to minimize the effect of sensor deteri-oration. Conditioning and calibration of the sensors followedmanufacturer recommendations and the methodology was pre-sented previously.25–27 The expected maximum load (for thepurposes of calibration and conditioning), based on previousnoninvasive measurements in our lab, was 100 N. The sensorswere conditioned before initial use by loading the sensor be-tween 2 layers of 3.2-mm thick lubricated rubber covering ma-chined aluminum plates in a materials testing machine (InstronDynaMight 8841, Canton, MA). The sensors were loaded to120% of the expected maximum load (120 N) for 5 loading

Figure 1. Flexibility test set-up for flexion– extension. The figureshows the posterior aspect of the specimen with Tekscan sensorsinserted into the facet joint and Dynesys system implanted.

20 Spine • Volume 33 • Number 1 • 2008

cycles. A linear calibration was performed after each test con-dition, using a similar loading protocol as in the conditioningphase, by loading each sensor to 80% of the expected maxi-mum load (80 N). Accuracy and repeatability of these sensorsfor measurement of facet loads in the lumbar spine was de-scribed previously.27

We tested the hypothesis that the Dynesys reduced contactforces in the facet joint when compared with the natural jointusing a two-way repeated measures analysis of variance with a95% level of significance. We also tested the hypothesis thatlonger spacers reduced facet joint loads more than short spac-ers using the same tests. This was done in 2 analyses; the firstlooking at the effect of the Dynesys when compared with anintact specimen and the second looking at the effect of spacerlength. The first factor was specimen condition [(intact, Dyne-sys), (Dynesys standard spacer, Dynesys long spacer, Dynesysshort spacer)] and the second factor was side (left, right). Theanalysis was performed separately with and without a followerpreload. Student–Newman-Keuls post hoc analysis (P � 0.05)was used to determine differences between specific cases.

Results

Application of a follower preload did not result in signif-icant differences in facet load compared with testing per-formed without a follower preload (P � 0.16). Onlythose results without a follower preload will be pre-sented here. In the intact spines, the average total peakfacet loads (average of right and left for lateral bendingand axial rotation, sum of right and left for flexion andextension) was greatest in axial rotation (56 N), fol-lowed by extension (27 N), lateral bending (13 N), andlowest in flexion (7 N) (Table 1).

In the intact spine, right axial rotation loaded the leftfacet joint whereas left axial rotation loaded the rightfacet joint (Figure 3). The contact force increased withincreasing extension and was minimal in flexion (Figure4). In lateral bending the contact force pattern was lessconsistent among specimens, but typically alternated be-tween sides for left and right applied moments.

Implantation of the Dynesys system created a smallinitial load at the facet joints before starting the flexibil-ity tests (Figure 4, Table 2). The differences were signif-icant between the short and long spacers on both the left(P � 0.004) and right sides (P � 0.014). However, therewas no significant difference between the right and leftsides.

Specimens with the standard length Dynesys im-planted had significantly higher facet loads in flexion(P � 0.003) (Figures 4, 5) and lateral bending (P �0.044) (Figure 6), compared with the intact condition.The most substantial change was in flexion where theDynesys increased facet loads compared with the intactcondition from an average peak load of 2 and 4 N for theleft and right sides to 16 and 27 N for the left and rightsides, respectively. There were no significant differencesin facet loads for extension and axial rotation comparedwith an intact specimen (Figures 3, 7, and 8); however,the implant seemed to alter the contact load pattern withvertebral rotation in both axial rotation (Figure 3) andflexion–extension (Figure 4).

Variation of the Dynesys spacer length had an effecton facet loads. Generally, contact force was highest withthe short spacer and lowest with the long spacer in alldirections of loading (Figures 5–8). There was a signifi-cant difference in contact force between the long andshort spacers in flexion (P � 0.011) and lateral bending(P � 0.014).

In spines stabilized with the Dynesys, the asymmetryin the peak contact force during flexibility tests betweenthe 2 facet joints increased. The difference between theleft and right sides with the Dynesys was statisticallysignificant only in extension (P � 0.024). There was asignificant difference based on side, as a main factor, inlateral bending (P � 0.016).

Discussion

The effects of motion-preserving implant systems onfacet joint mechanics are likely of importance to theirclinical performance. Therefore, this study was designedto investigate the effect of the Dynesys dynamic posteriorstabilization system on contact forces in the lumbar spinefacet joints and to examine how these forces were af-fected by variation in the Dynesys spacer length. Overall,we found that the Dynesys dynamic posterior stabiliza-

Figure 2. Tekscan thin film electroresistive sensor (Tekscan Inc.,MA). The figure shows the relative size of one “finger” of theTekscan 6900 Quad sensor. The sensing element is a 14 � 14 mm2

area with 121 measurement sensels.

Table 1. Average Facet Contact Force

Loading Direction

Intact Dynesys

Left (N) Right (N) Left (N) Right (N)

Axial Rotation 56 � 17 55 � 18 50 � 24 63 � 20Lateral Bending 11 � 11 16 � 14 16 � 13 31 � 21Extension 13 � 14 14 � 10 9 � 11 21 � 18Flexion 2 � 5 4 � 4 16 � 16 27 � 22

Facet loads are average and standard deviation for intact specimens and forinjured specimens stabilized with the standard length Dynesys spacer. Re-ported contact forces are without a compressive follower preload. Forces arein Newtons (N).

21Dynamic Posterior Stabilization on Facet Joint • Niosi et al

tion system did not alter peak contact forces within thefacet joints in axial rotation and extension, when com-pared with the intact case. In flexion and lateral bending,the device created a small increase in facet loads. Thelength of the Dynesys spacer had a significant effect on facetloads in flexion and lateral bending, with the long spacerresulting in lower facet loads than the short spacer.

Strengths of this study include direct measurement ofthe contact force within the facet joints and evaluation ofthe facet loads with a nonfusion implant. In contrast toprevious studies where the load was also measured di-

rectly, the Tekscan allowed dynamic measurements andelectronic recording of results, making it a relatively easysystem to use. All testing was done with and without acompressive follower preload and no effect of the pre-load was observed.

The accuracy and repeatability of the use of Tekscansensors to directly measure facet contact forces in thelumbar spine has been evaluated27 and the accuracy ofthe Tekscan was considered as a limitation in the presentstudy. The Tekscan 6900 sensor was accurate in measur-ing facet loads to within 18% at 100 N and to within

Figure 3. Typical contact forcesfor left and right facet joints ver-sus rotation (in degrees) in axialrotation for the Intact conditionand with standard Dynesysspacer. Contralateral facet jointwas loaded in compression dur-ing left and right axial rotation.Negative facet load indicatescompression.

Figure 4. Typical contact forces forleft and right facet joints versusrotation (in degrees) in flexion-extension for the Intact conditionand with standard Dynesys spacer.Negative facet force indicates com-pression. Also shown clearly is aninitial facet load created after im-plantation of the device.

22 Spine • Volume 33 • Number 1 • 2008

49% at 25 N. One explanation for the lower accuracyfound at 25 N was that the measured loads were at thelow end of the range of the sensor, which was typicallywhere the loads in this study fell. Repeatability of forcemeasurement in the facet joint was 4% and 7% in axialrotation and extension, respectively.27 Despite low accu-racy for measurement of small forces, the repeatability offorce measurements suggests that relative differences inloading can still be assessed.

The results of this study must be interpreted in light ofsome other limitations. Direct facet load measurementrequired sectioning of the joint capsule, which was as-sumed equivalent to the intact case. Studies using doglumbar spines have reported that the effect of capsuletransection on facet loads was inconsistent but minimaloverall.28 In addition, the effect of inserting a film intothe articular joint on contact mechanics is important. Afinite element study using Fuij Pressensor film found thatthe film changed the maximum true contact pressure by10% to 26% depending on the loading, geometry of thejoints, and properties of the cartilage.29 The effectivethickness of the Fuji Pressensor film and the Tekscansensors was 0.3 and 0.1 mm, respectively. The thinner

Tekscan sensor would be expected to have a smaller ef-fect than that which was determined with the Fuji film.

Another limitation of this study was not foreseenwhen using spacers that were less stiff than in vivo spac-ers, to account for testing at room temperature. Becausethe spacer stiffness was lower, the pretension appliedto the cord may have produced a smaller distance be-tween the screw heads, and thus, greater compressionof the system than what would be expected in vivo. Inaddition, during clinical use, a small amount is normallyadded to the required spacer length to compensate forthe loss in spacer height that results from cord preten-sioning. In this study, the spacers were sized to put thespine in a neutral position without this compensatinglength. These 2 factors may have resulted in a greatersegmental compression than what would be anticipatedintraoperatively. As a result, the long spacer may bemore representative of the in vivo situation than the stan-dard spacer.

Comparison of our results with previous work waslimited strictly to the intact condition because there havebeen no published biomechanical studies that examinethe facet loads with the Dynesys. In flexion, the facetjoints have been found to support very minimal or noload.10–12 This is consistent with the findings of ourstudy. Under varying degrees of extension, previouswork has found that the facet joints support about 10%to 40%5,9–11 of the applied load or between 52 and 130N12,30 for an applied moment comparable with the onein our study (Table 3). Conversion of our results, basedon average 3-dimensional quantitative lumbar anato-my,31 showed an average facet load of approximately13% of the applied load or 27 N. This is within the rangeof most of the literature, although at the lower end of the

Table 2. Average Initial Facet Contact Force

Short (N) Standard (N) Long (N)

Left 16 � 12* 13 � 13 4 � 8*Right 27 � 27† 18 � 18 11 � 10†

Values shown are mean and standard deviation for the initial force created atthe left and right facet joints by implantation of the Dynesys system (short,standard, and long spacers). Differences were statistically significant betweenthe short and long spacers, on the left side (*P � 0.004) and right side (†P �0.014). Forces are in Newtons (N).

Figure 5. Average peak facetload in flexion for 10 specimenswithout a follower preload. Con-tact forces for the left and rightfacets for intact and 3 Dynesysspacer lengths are shown. Sta-tistical analysis was between in-tact and standard Dynesysspacer and between the 3spacer lengths. There was nosignificant difference based onside (right or left).

23Dynamic Posterior Stabilization on Facet Joint • Niosi et al

spectrum. In lateral bending the facet loads measured inour study were smaller than the values in the literature(Table 3). Small discrepancies between our results andthe work of others are likely due to differences in facetload measurement technique. The studies by Goel et al30

and Shirazi-Adl32 were finite element analyses andSchendel et al12 indirectly measured facet loads usingstrain gauges, in contrast to our direct measurementtechnique. There are also differences qualitatively infacet loads reported in the literature, therefore, highlight-ing the ambiguity surrounding the precise function of thejoint.

Implantation of the Dynesys created a small compres-sion of the posterior elements, because of pretensioningof the cord, as was evident by the presence of a static loadat the facet joints immediately after the device was im-planted. The initial load was lowest with the long spacer,which as mentioned previously, probably better repre-sents the in vivo situation. The significantly larger initialload observed with the short spacer can be attributed toan increase in segmental compression that resulted withthe short spacer. Nonsignificant asymmetry in the initialload occurred between the right and left sides. Duringflexibility tests, this asymmetry was also apparent, with

Figure 6. Average peak facetload in lateral bending for 10specimens without a followerpreload. Contact forces for theleft and right facets for intact and3 Dynesys spacer lengths areshown. Statistical analysis wasbetween intact and standardDynesys spacer and between the3 spacer lengths. In addition,there was a significant differ-ence between right and left sidesin intact and with the standardDynesys spacer (P � 0.016).

Figure 7. Average peak facetload in extension for 10 speci-mens without a follower preload.Contact forces for the left andright facets for intact and 3Dynesys spacer lengths areshown. Statistical analysis wasbetween intact and standardDynesys spacer and between the3 spacer lengths. There were nostatistically significant differ-ences (P � 0.05).

24 Spine • Volume 33 • Number 1 • 2008

significantly different facet load magnitudes between theright and left sides in lateral bending and extension. TheDynesys had a tendency to increase the peak load atthe right facet and decrease the peak load at the left facetjoint in axial rotation and extension compared with theintact condition. The asymmetry did not seem to dependon which side of the implant was tightened first or whichspacer was longer. Reasons for this observation couldinclude small mismatches in cord pretension between theleft and right side or in sizing the spacer length.

It has been commonly accepted that in flexion, thefacet joints are distracted and therefore the contact loadis very minimal8,11,17; however, in this study, the Dyne-sys system seemed to reverse the loading pattern com-pared with that seen in the intact specimen, such that thecontact load increased with greater degree of flexion.When the device was implanted, the load at the facetjoints increased during flexion and became larger thanthat in extension, which is in contrast to the findings in theintact specimen. This observation can be attributed tothe significant posterior shift in the location of the HAM

in flexion–extension that occurs with implantation ofthe Dynesys from its central position in the intact speci-men.22 The HAM is the unique axis about which a bodyrotates and parallel to which it translates.33 The Dynesysaffected the position and orientation of the HAM andmay have also led to alteration in the natural contactmechanism between the articulating surfaces, resultingin increasing facet loads during flexion.

The kinematic behavior of the Dynesys is consistentwith changes found in loading at the facet joints. Inter-segmental ROM at the implanted level decreased signif-icantly after stabilization with the Dynesys such that inflexion, extension, lateral bending, and axial rotation,the ROM was 27%, 33%, 26%, and 76% that of theintact specimen, respectively.22 The long spacer pro-duced a significantly greater ROM than the short spacerin all directions of loading. The smaller ROM with theshort spacer is consistent with findings that increased seg-mental compression generates a decrease in ROM.22,34

There was evidence in both the initial static facet loads anddynamic facet loads (in flexion and lateral bending) that theposterior elements were placed under a larger degree ofcompression with the short spacer than with the longspacer.

Comparison of facet loads at identical rotations, as ina displacement controlled investigation, would describethe effect of the implant on posterior element loadingbecause facet load is dependent on the degree of rotation.However, we felt that a load-controlled study was moreapplicable and demonstrated greater clinical relevancebecause the motion with these devices is not the same asin an intact condition.21,22 Even so, the forces with theDynesys implanted were higher than those for the intactcondition at the same rotation (Figures 3, 4). In addition,the long spacer typically caused an increase in motion,

Figure 8. Average peak facetload in axial rotation for 10 spec-imens without a follower preload.Contact loads for the left andright facets for intact and 3Dynesys spacer lengths areshown. Statistical analysis wasbetween intact and standardDynesys spacer and between the3 spacer lengths. There were nostatistically significant differ-ences (P � 0.05).

Table 3. Comparison of Intact Facet Loads

Extension Lateral Bending Axial Rotation

Goel et al30 52 N (7 N–m) 90 N (7 N–m)Schendel et al12 130 N (8 N–m) 104 N (3 N–m) 30 N (7.5 N–m)Sharma et al 11 26%Dunlop et al 9 10–40%Yang and King5 12–19%Lorenz et al10 13–30%Shirazi-Adl32 8.3 N (10 N–m) 67 N (10 N–m)Our study 13%

27 N (7.5 N–m)

13 N (7.5 N–m) 56 N (7.5 N–m)

Facet loads in extension, lateral bending, and axial rotation. Values are pre-sented as either a force in Newtons (with applied moment in brackets) or asa percentage of the applied load.

25Dynamic Posterior Stabilization on Facet Joint • Niosi et al

yet a decrease in facet loads compared with the shortspacer. Because ROM increased with the long spacer, thereduction in facet loads, relative to the short spacer, canbe attributed entirely to the implant.

Clinically, increases in facet load magnitudes couldpotentially lead to low back pain, facet joint osteoarthri-tis, or other pathologies. It is important to emphasizethat in this study, although the largest change in facetloads occurred in flexion once a specimen was stabilizedwith the Dynesys system, the relative magnitude of theseloads was still fairly low in comparison with facet loadsmeasured in other loading directions. However, it maybe possible that a change in loading pattern could resultin adverse clinical effects that cannot be evaluated bio-mechanically.

Key Points

● Nonfusion stabilization systems alter the loadingthrough the spinal column.● We directly measured dynamic facet contactforces during flexibility testing using thin-film elec-troresistive sensors.● Implantation of the Dynesys created a small ini-tial contact force at the facet joints.● Compared with an intact specimen, the Dynesyssignificantly increased peak facet contact forces inflexion and lateral bending, but had no effect inextension or axial rotation.● Generally, the longer the Dynesys spacer, thelower is the facet contact force. Differences in facetloads were significant between the long and shortspacers only in flexion and lateral bending.

AcknowledgmentsThe authors gratefully acknowledge funding from theSynos Foundation, Switzerland and Natural Sciencesand Engineering Research Council of Canada (NSERC).

References

1. Eck JC, Humphreys SC, Hodges SD. Adjacent-segment degeneration afterlumbar fusion: a review of clinical, biomechanical, and radiologic studies.Am J Orthop 1999;28:336–40.

2. Dekutoski MB, Schendel MJ, Ogilvie JW, et al. Comparison of in vivo and invitro adjacent segment motion after lumbar fusion. Spine 1994;19:1745–51.

3. Lee CK. Accelerated degeneration of the segment adjacent to a lumbar fu-sion. Spine 1988;13:375–7.

4. Schlegel JD, Smith JA, Schleusener RL. Lumbar motion segment pathologyadjacent to thoracolumbar, lumbar, and lumbosacral fusions. Spine 1996;21:970–81.

5. Yang KH, King AI. Mechanism of facet load transmission as a hypothesis forlow-back pain. Spine 1984;9:557–65.

6. Haberl H, Cripton PA, Orr TE, et al. Kinematic response of lumbar func-tional spinal units to axial torsion with and without superimposed compres-sion and flexion/extension. Eur Spine J 2004;13:560–66.

7. Adams MA, Hutton WC. The mechanical function of the lumbar apophysealjoints. Spine 1983;8:327–30.

8. Adams MA, Hutton WC. The effect of posture on the role of the apophyseal

joints in resisting intervertebral compressive forces. J Bone Joint Surg Br1980;62:358–62.

9. Dunlop RB, Adams MA, Hutton WC. Disc space narrowing and the lumbarfacet joints. J Bone Joint Surg Br 1984;66:706–10.

10. Lorenz M, Patwardhan A, Vanderby R Jr. Load-bearing characteristics oflumbar facets in normal and surgically altered spinal segments. Spine 1983;8:122–30.

11. Sharma M, Langrana NA, Rodriguez J. Role of ligaments and facets inlumbar spinal stability. Spine 1995;20:887–900.

12. Schendel MJ, Wood KB, Buttermann GR, et al. Experimental measurementof ligament force, facet force, and segment motion in the human lumbarspine. J Biomech 1993;26:427–38.

13. Cavanaugh JM, Ozaktay AC, Yamashita HT, et al. Lumbar facet pain:biomechanics, neuroanatomy and neurophysiology. J Biomech 1996;29:1117–29.

14. Nachemson A. Measurement of intradiscal pressure. Acta Orthop Scand1959;28:269–89.

15. Nachemson A. The load on lumbar disks in different positions of the body.Clin Orthop 1966;45:107–22.

16. Prasad P, King AI, Ewing CL. The role of articular facets during �Gz accel-eration. J Appl Mech 1974;41:321–6.

17. Buttermann GR, Kahmann RD, Lewis JL, et al. An experimental method formeasuring force on the spinal facet joint: description and application of themethod. J Biomech Eng 1991;113:375–86.

18. Luo ZP, Buttermann GR, Lewis JL. Determination of spinal facet joint loadsfrom extra articular strains—a theoretical validation. J Biomech 1996;29:785–90.

19. Hedman TP. A new transducer for facet force measurement in the lumbarspine: benchmark and in vitro test results. J Biomech 1992;25:69–80.

20. Freudiger S, Dubois G, Lorrain M. Dynamic neutralisation of the lumbarspine confirmed on a new lumbar spine simulator in vitro. Arch OrthopTrauma Surg 1999;119:127–32.

21. Schmoelz W, Huber JF, Nydegger T, et al. Dynamic stabilization of thelumbar spine and its effects on adjacent segments: an in vitro experiment.J Spinal Disord Tech 2003;16:418–23.

22. Niosi CA, Zhu QA, Wilson DC, et al. Biomechanical characterization of thethree-dimensional kinematic behaviour of the Dynesys dynamic stabilizationsystem—an in vitro study. Eur Spine J 2006;15:913–22.

23. Goertzen DJ, Lane C, Oxland TR. Neutral zone and range of motion in thespine are greater with stepwise loading than with a continuous loading pro-tocol. An in vitro porcine investigation. J Biomech 2004;37:257–61.

24. Patwardhan AG, Havey RM, Meade KP, et al. A follower load increases theload-carrying capacity of the lumbar spine in compression. Spine 1999;24:1003–9.

25. Harris ML, Morberg P, Bruce WJ, et al. An improved method for measuringtibiofemoral contact areas in total knee arthroplasty: a comparison of K-scansensor and Fuji film. J Biomech 1999;32:951–8.

26. Wilson DR, Apreleva MV, Eichler MJ, et al. Accuracy and repeatability of apressure measurement system in the patellofemoral joint. J Biomech 2003;36:1909–15.

27. Wilson DC, Niosi CA, Zhu QA, et al. Accuracy and repeatability of a newmethod for measuring facet loads in the lumbar spine. J Biomech 2006;39:348–53.

28. Kahmann RD, Buttermann GR, Lewis JL, et al. Facet loads in the caninelumbar spine before and after disc alteration. Spine 1990;15:971–8.

29. Wu JZ, Herzog W, Epstein M. Effects of inserting a pressensor film intoarticular joints on the actual contact mechanics. J Biomech Eng 1998;120:655–9.

30. Goel VK, Winterbottom JM, Weinstein JN, et al. Load sharing among spinalelements of a motion segment in extension and lateral bending. J BiomechEng 1987;109:291–7.

31. Panjabi MM, Goel V, Oxland T, et al. Human lumbar vertebrae. Quantita-tive three-dimensional anatomy. Spine 1992;17:299–306.

32. Shirazi-Adl A. Finite-element evaluation of contact loads on facets of anL2–L3 lumbar segment in complex loads. Spine 1991;16:533–41.

33. Panjabi MM, Krag MH, Goel VK. A technique for measurement and de-scription of three-dimensional six degree-of-freedom motion of a body jointwith an application to the human spine. J Biomech 1981;14:447–60.

34. Lund T, Rathonyi G, Schlenzka D, et al. The external spinal fixator does notreduce anterior column motion under axial compressive loads. A mechanicalin vitro study. Acta Orthop Scand 1999;70:37–41.

26 Spine • Volume 33 • Number 1 • 2008