JBJS | A Comparative Biomechanical Investigation of Anterior Lumbar Interbody Cages: Central and Bilateral Approaches*

The Journal of Bone & Joint Surgery, Volume 82, Issue 3

Articles | March 01, 2000

A Comparative Biomechanical Investigation of Anterior Lumbar Interbody Cages: Central and Bilateral Approaches*

THOMAS R. OXLAND, PH.D.†; ZOLTAN HOFFER, M.D.‡; THOMAS NYDEGGER, DIPL.ING.§; GABOR C. RATHONYI, M.D.‡; LUTZ-P. NOLTE, PH.D.§, BERN, SWITZERLAND

View Disclosures and Other Information

Investigation performed at the Maurice E. Müller Institute for Biomechanics, University of Bern, Bern

The Journal of Bone & Joint Surgery. 2000; 82:383-93

5 Recommendations (Recommend) | 3 Comments | Saved by 3 Users Save Case

Article

Figures

References

text A A A

Abstract

Background: Some biomechanical studies have been performed to evaluate the stabilization provided by interbody cages, but there are virtually no comparative data for the different designs. Furthermore, most investigators have used animal models, which may have led to different results due to morphological variation in the end plates and articular facets. The objectives of the current study were to evaluate whether two different anterior cage designs (BAK and SynCage) performed differently with respect to immediate stabilization of the spine, whether the cages stabilized the spine significantly compared with its intact condition, and whether the addition of supplementary translaminar screw fixation further stabilized the spine. Stabilization was defined as a reduction in motion after insertion of an implant.

Methods: Twelve lumbar functional spinal units from human cadavera were tested under pure moments of flexion, extension, bilateral axial rotation, and bilateral lateral bending to a maximum of ten newton-meters. The relative intervertebral motions were measured, with use of an optoelectronic camera system, under three test conditions: with the spine intact, after insertion of anterior interbody cages, and after insertion of anterior interbody cages supplemented with translaminar screw fixation. Six specimens were tested for each type of cage: a bilateral, porous, threaded cylinder (BAK) and a central, porous, contoured implant with end-plate fit (SynCage).

Results: The cages performed in a similar manner in all directions of loading, with no significant differences between the two designs. The cages significantly stabilized the spine compared with its intact condition in flexion, axial rotation, and lateral bending (the median value for motion was 40, 48, and 29 percent of the value for the intact condition, respectively; p = 0.002 for all three directions). Compared with the cages alone, translaminar screw fixation provided no additional stabilizing effect in these directions but it significantly increased the stability of the spine in extension (the median value for motion was 34 percent of the value with the cages alone; p = 0.013).

Conclusions: There were no differences in the stabilization provided by the two different cage designs. Use of the cages alone stabilized the spine in all directions except extension, and use of supplementary translaminar screw fixation provided additional stabilization only in extension.

Clinical Relevance: This study demonstrated that interbody cages do not stabilize the lumbar spine in extension, and this observation was not altered by the use of substantially different designs. If the lack of stabilization in extension is a clinical problem, possible solutions include the avoidance of extension postoperatively or the use of supplementary fixation.

Figures in this Article

Introduction

Introduction | Materials and Methods | Results | Discussion | References

In the last decade, several vertebral interbody cages of different designs were developed for use with either an anterior or a posterior approach^2,12,24. These implants are based on the distraction-compression principles of Bagby¹, who advocated stabilizing the intervertebral joints and thereby providing an appropriate environment for interbody fusion. Clinical trials in humans, conducted to investigate the safety and efficacy of these implants for the treatment of discogenic low-back pain, have shown promising results. Kuslich et al.¹² reported an average rate of fusion of 96 percent after single-level procedures (144 of 147 with use of an anterior approach and 119 of 127 with use of a posterior approach) and of approximately 80 percent after two-level procedures (eighty of 100 with use of an anterior approach and twenty-seven of thirty-eight with use of a posterior approach). Ray²⁴ reported an average rate of fusion of 96 percent in a combined series (202 of 211 procedures).

Despite these encouraging early results, numerous questions remain regarding many areas of interbody cage technology, including the proper surgical indication for use of a cage, the assessment of bone growth into and through the cage (that is, fusion), the importance of settling of the implant into adjacent vertebral bodies, and the initial stabilization provided by the cage. With respect to the latter point, a major focus is the importance of supplementary fixation and the different results obtained with different cage designs.

Some biomechanical studies have provided basic data on the immediate stabilization (the reduction in intervertebral motion) achieved with use of interbody cages. Butts et al.⁴ investigated the immediate stabilization provided by one or two cylindrical implants in calf and pig spines and found that two implants yielded more stability. In a multifaceted study of a calf-spine model, Brodke et al.³ found that the BAK system provided stabilization similar to that provided by pedicle fixation with interbody bone blocks. Tencer et al.²⁶ evaluated the effects of a cylindrical cage (Ray TFC) in different positions and orientations in the interbody space of calf and human spines. They found that the cages affected only the laxity (that is, the neutral zone) of the intervertebral joint and not its stiffness. However, none of these investigators, except Tencer et al.²⁶, who studied five specimens, used a human cadaveric model for the biomechanical evaluation. There could be important differences between these in vitro animal models and the human situation, particularly due to the immature growth plates in many animal models⁶ and the fact that interbody implants rely on this area of the vertebral body for fixation.

More recently, other investigators studied stabilization provided by cages in a human model and found evidence that cages do not reduce motion in extension^10,14,16. In a comparative study of posterior lumbar interbody fusion implants, Lund et al.¹⁴ found that axial rotation was a problem for all types of implants, with no differences between them. Volkman et al.²⁷ reported that cages did not stabilize the spine in any direction, but they used only single implants with off-axis compression loading.

There have been no comparative studies, to our knowledge, of cages inserted from an anterior approach. Due to substantial differences between anterior cages in terms of vertebral body fixation and soft-tissue destruction during insertion, we hypothesized that there may be differences in the stabilization provided to the spine by these implants. Therefore, using two types of anterior implants—a two-cage construct (BAK; Sulzer Spine-Tech, Minneapolis, Minnesota) and a single-cage construct (SynCage; Mathys Medical, Bettlach, Switzerland)—we designed this study to determine whether (1) there are differences in the stabilizing abilities of the two types of cages, (2) the cages stabilize the spine significantly compared with its intact condition, and (3) use of supplementary translaminar screw fixation improves stabilization compared with use of cages alone. We defined stabilization as a reduction in motion (flexibility) after insertion of an implant.

*Although none of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article, benefits have been or will be received, but are directed solely to a research fund, foundation, educational institution, or other nonprofit organization with which one or more of the authors is associated. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was Mathys Medical, Bettlach, Switzerland.

†Departments of Orthopaedics and Mechanical Engineering, University of British Columbia, Vancouver Hospital, 910 West 10th Avenue, Vancouver, British Columbia V5Z 4E3, Canada. E-mail address: toxland@interchange.ubc.ca.

‡National Center for Spinal Disorders, Semmelweis University of Medicine, Kiralghago utca 1—3, 1126 Budapest, Hungary.

§Department of Orthopaedics, Maurice E. Müller Institute for Biomechanics, University of Bern, Murtenstrasse 35, P.O. Box 30, CH-3010 Bern, Switzerland.

‡National Center for Spinal Disorders, Semmelweis University of Medicine, Kiralghago utca 1—3, 1126 Budapest, Hungary.

§Department of Orthopaedics, Maurice E. Müller Institute for Biomechanics, University of Bern, Murtenstrasse 35, P.O. Box 30, CH-3010 Bern, Switzerland.

View Large | Download Slide(.ppt)
Add to My JBJS

FIG1-A:: Figs. 1-A, 1-B, and 1-C: Anteroposterior and lateral radiographs of the two types of interbody cages and the translaminar screw fixation used in the current study. Fig. 1-A: The BAK implant is a threaded, porous, titanium, cylindrical device, which was placed bilaterally into the disc space, penetrating both vertebral end plates.

View Large | Download Slide(.ppt)
Add to My JBJS

FIG1-B:: Fig. 1-B The SynCage is a porous titanium device with contoured surfaces, which was placed centrally into the disc space. It has sharp teeth to engage both end plates.

View Large | Download Slide(.ppt)
Add to My JBJS

FIG1-C:: Fig. 1-C Translaminar screw fixation consists of two bone screws crossing through the lamina and the facet joints. The screws are shown supplementing the fixation of the BAK implant.

View Large | Download Slide(.ppt)
Add to My JBJS

FIG2:: Fig. 2 Anteroposterior views of three different setups for multidirectional flexibility testing. Pneumatic actuators (not shown) applied pure moments to the cephalad vertebra by delivering equal and opposite forces (F) to cables and torque pulleys. The applied moments were measured and controlled with use of strain gauges on one of the pulleys. Infrared light-emitting diodes (LEDs) were attached to each vertebra to track the relative intervertebral motions. The position of the pulleys was changed to produce flexion-extension moments (A), axial rotation moments (B), and lateral bending moments (C).

View Large | Download Slide(.ppt)
Add to My JBJS

FIG3-A:: Fig. 3-A Average moment-rotation curves in flexion-extension for all specimen conditions are shown for the BAK device (Fig. 3-A) and the SynCage device (Fig. 3-B). Motion of the intact specimens was much greater in flexion than in extension. In flexion, both implants had an obvious, significant stabilizing effect (p = 0.002) and supplementary translaminar screw fixation (TLS) had a marginal additional effect. In extension, the cages did not reduce motion substantially, although the SynCage had a modest (but nonsignificant) effect.

View Large | Download Slide(.ppt)
Add to My JBJS

FIG3-B:: Fig. 3-B Average moment-rotation curves in flexion-extension for all specimen conditions are shown for the BAK device (Fig. 3-A) and the SynCage device (Fig. 3-B). Motion of the intact specimens was much greater in flexion than in extension. In flexion, both implants had an obvious, significant stabilizing effect (p = 0.002) and supplementary translaminar screw fixation (TLS) had a marginal additional effect. In extension, the cages did not reduce motion substantially, although the SynCage had a modest (but nonsignificant) effect.

View Large | Download Slide(.ppt)
Add to My JBJS

FIG4-A:: Fig. 4-A Average moment-rotation curves in axial rotation for all specimen conditions are shown for the BAK device (Fig. 4-A) and the SynCage device (Fig. 4-B). The motion of the intact specimens in left and right rotation was virtually symmetrical. The cages had an obvious, significant stabilizing effect (p = 0.002). Translaminar screw fixation (TLS) did not significantly reduce motion.

View Large | Download Slide(.ppt)
Add to My JBJS

FIG4-B:: Fig. 4-B Average moment-rotation curves in axial rotation for all specimen conditions are shown for the BAK device (Fig. 4-A) and the SynCage device (Fig. 4-B). The motion of the intact specimens in left and right rotation was virtually symmetrical. The cages had an obvious, significant stabilizing effect (p = 0.002). Translaminar screw fixation (TLS) did not significantly reduce motion.

View Large | Download Slide(.ppt)
Add to My JBJS

FIG5-A:: Fig. 5-A Average moment-rotation curves in lateral bending for all specimen conditions are shown for the BAK device (Fig. 5-A) and the SynCage device (Fig. 5-B). The motion of the intact specimens in left and right bending was virtually symmetrical. The cages had an obvious, significant stabilizing effect (p = 0.002). Translaminar screw fixation (TLS) did not significantly reduce motion.

View Large | Download Slide(.ppt)
Add to My JBJS

FIG5-B:: Fig. 5-B Average moment-rotation curves in lateral bending for all specimen conditions are shown for the BAK device (Fig. 5-A) and the SynCage device (Fig. 5-B). The motion of the intact specimens in left and right bending was virtually symmetrical. The cages had an obvious, significant stabilizing effect (p = 0.002). Translaminar screw fixation (TLS) did not significantly reduce motion.

View Large | Download Slide(.ppt)
Add to My JBJS

FIG6:: Fig. 6 Box-stem plot comparing the two cage designs. For each loading direction and cage type, the ratio of motion of the intact specimens to that of motion of the specimens with cages is plotted as the median, the twenty-fifth to seventy-fifth percentiles, and the non-outlier range. The nonparametric data analysis revealed no significant differences between the BAK and SynCage implants in flexion (p = 0.08), extension (p = 0.47), axial rotation (p = 0.15), or lateral bending (p = 0.75).

View Large | Download Slide(.ppt)
Add to My JBJS

FIG7:: Fig. 7 Box-stem plot summarizing the effect, on spinal stabilization, of insertion of a cage alone and a cage with supplementary translaminar screw fixation (TLS). For each loading direction and specimen condition, the range of motion in degrees is plotted as the median, the twenty-fifth to seventy-fifth percentiles, and the non-outlier range. In the comparison of the intact and cage groups, the nonparametric data analysis revealed a significant reduction of motion in flexion, axial rotation, and lateral bending (p = 0.002 for all three directions) but not in extension (p = 0.48). In the comparison of the effect of the cage and that of the cage with translaminar screw fixation, the analysis demonstrated that the screw fixation significantly reduced motion only in extension (p = 0.013).

Materials and Methods

Introduction | Materials and Methods | Results | Discussion | References

Preparation of Specimens

Twelve lumbar functional spinal units were harvested from human cadavera. Nonligamentous soft tissue was carefully dissected from all specimens, and the specimens were kept frozen at -20 degrees Celsius before and between test procedures. There was no evidence of any bone-related disease on dissection of the specimens or on pretest radiographs.

The bone-mineral density of the cephalad and caudad vertebrae was measured with use of dual-energy x-ray absorptiometry (model QDR 1000; Hologic, Waltham, Massachusetts) from a lateral and an anteroposterior direction. The vertebrae were mounted in polymethylmethacrylate blocks, with the intervertebral disc kept in the horizontal plane. Lateral and anteroposterior radiographs were made of all specimens.

Experimental Protocol

The specimens were divided into two groups of six specimens each, and each group received either the BAK device or the SynCage construct.

The BAK implant is a threaded, porous, titanium cylinder, which penetrates the vertebral end plates to engage the bone of the vertebral bodies superior and inferior to the disc space (Fig. 1-A). The vertebral bodies are prepared with a reamer and tap before insertion of the implant. We defined the end plate as the central shelf of perforated subchondral bone on the ends of the vertebral body⁸. The size of the two, bilaterally placed implants was selected to obtain a tight annulus after insertion in the disc space. The procedure maintained approximately a five-millimeter width of the anterior annulus and the anterior longitudinal ligament.

The SynCage implant is a porous titanium block with contoured surfaces and sharp teeth to engage both vertebral end plates (Fig. 1-B). Since the contours of the implant were fixed and the shape of the vertebral end plate was not, there was some variability in the bone-implant interface. The size of the single central implant was selected to obtain a tight annulus after insertion in the disc space.

The BAK group comprised three second and third lumbar and three fourth and fifth lumbar functional spinal units, and the SynCage group contained four second and third lumbar and two fourth and fifth lumbar functional spinal units. All specimens were tested in multidirectional flexibility under three conditions: (1) intact, (2) after insertion of the corresponding interbody-cage construct, and (3) with the addition of translaminar screw fixation to the cage construct. All implants were inserted according to the manufacturers' guidelines on surgical technique, from an anterior direction. The translaminar screws were inserted with use of the technique described by Magerl¹⁵ (Fig. 1-C).

Multidirectional Flexibility Protocol

Each specimen was mounted in a specially designed apparatus in order to apply known pure moments to the spine and to measure intervertebral motion in an unconstrained manner (Fig. 2)^21,22. In each test condition, pure moments of flexion-extension, bilateral axial rotation, and bilateral lateral bending were applied individually to the cephalad vertebra in an incremental stepwise fashion to a maximum of ten newton-meters (that is, four steps of 2.5 newton-meters). At each load step, the specimen was allowed to creep for thirty seconds. This loading regimen was repeated for two cycles. Due to the design of the apparatus, the moments remained pure along the length of the specimen.

The rigid body motion of the cephalad vertebra with respect to the caudad vertebra was measured with use of an optoelectronic camera system (Optotrak 3020; Northern Digital, Waterloo, Ontario, Canada). This system monitors the spatial position of infrared light-emitting diodes. Marker carriers, each having four noncollinear light-emitting diodes, were attached on both polymethylmethacrylate blocks of all specimens.

For simplicity, only rotation in the direction of the applied moment (for example, flexion rotation for the applied flexion moment) as the main motion was investigated. Two motion parameters, the neutral zone and the range of motion, were determined for each specimen²¹. The range of motion represents the magnitude of motion under the maximum moment from the initial, unloaded position of the specimen. This parameter was determined in flexion, extension, axial rotation (left to right), and lateral bending (right to left). The neutral zone, defined as the magnitude of motion under zero moment at the start of the second load cycle from the initial, unloaded position, represents a laxity in the functional spinal unit. This parameter was determined in flexion, extension, axial rotation (left to right), and lateral bending (right to left).

Data Analysis

Nonparametric statistical methods were used since we could not be sure that the assumptions inherent in parametric analysis (normally distributed groups and homogeneity of variance) were valid. For all tests, we used a significance level of 0.05.

The first question addressed the differences between the two cage designs. The ratio of the neutral zone or the range of motion in the cage condition to that in the intact condition was calculated and compared between the two cage designs in each direction of loading with use of a Mann-Whitney U test (a nonparametric analog of the unpaired Student t test). For this comparison, we determined that, for 80 percent power (a type-II error of 20 percent), six specimens per group were required. The sample sizes in this study were based upon these comparisons between cage designs. This calculation was based upon detecting a difference of 30 percent in the ratio to the intact condition, with a standard deviation of 15 percent. This standard deviation was similar to observed variability in other recent studies^10,14. A 30 percent difference between the two types of cages was assumed to be clinically relevant.

The second and third questions dealt with whether the cages stabilized the spine compared with its intact condition and whether use of supplementary translaminar screw fixation provided more stabilization than did cages alone. In each loading direction, a Friedman analysis of variance was conducted across the three test conditions (that is, intact, cages alone, and cages with translaminar screw fixation). For the analysis of variance, the two types of cages were grouped together in any loading directions where they were not significantly different. If significant differences were found, then pairwise comparisons were conducted with use of the Wilcoxon matched-pairs test with Bonferroni corrections.

The intact and cage conditions were compared to address the second question, and the cage condition and the condition consisting of the cage with translaminar screw fixation were compared to address the third question. Given that there were six specimens per group as calculated previously, the power to detect 40 percent differences in motion between the intact and cage conditions was expected to exceed 80 percent and the power to detect 50 percent differences in motion between the cage and translaminar screw-fixation conditions was expected to be about 80 percent. The critical percent difference in motion between the cage and translaminar screw-fixation groups (that is, 50 percent) was greater than that between the cage and intact conditions (that is, 40 percent) since the magnitudes of motion were expected to be smaller.

The bone-density values in the two cage groups were compared since it had been reported previously that bone density has an effect on intervertebral stabilization¹⁸. Given that there were six specimens per group, the power to detect a difference of 0.15 gram per square centimeter between the two groups was 90 percent.

The same software package (Statistica for Windows 5.1; Statsoft, Tulsa, Oklahoma) was used for all analyses.

Results

Introduction | Materials and Methods | Results | Discussion | References

The bone-mineral-density values were not significantly different between the BAK and SynCage groups (0.79 ± 0.09 compared with 0.81 ± 0.05 gram per square centimeter; p = 0.54) and therefore were not included in additional analyses.

Flexion-Extension

Before insertion of both the BAK and the SynCage implants, the curves for the intact specimens in flexion exhibited a characteristic nonlinear moment-rotation behavior (Figs. 3-A and 3-B). Also, the range of motion for the intact specimens in extension was less than that in flexion before insertion of both cage types. Qualitatively, these differences were not apparent after insertion of the interbody cages.

In flexion, the median motion with use of the BAK device was 33 percent of that of the intact specimens and the median motion with use of the SynCage device was 41 percent of that of the intact specimens (Fig. 6). This difference between the two types of cages was not significant (p = 0.08). Insertion of the interbody cages decreased the range of motion in flexion significantly compared with that in the intact condition; the median range of motion with the cages was 40 percent of that in the intact spine (p = 0.002; Fig. 7). The addition of translaminar screw fixation did not significantly increase the stability of the spine compared with use of cages alone (p = 0.13; Fig. 7).

In extension, use of the BAK device alone resulted in a median range of motion of 109 percent of that of the intact specimens and use of the SynCage device alone decreased the range of motion to 68 percent of that of the intact specimens (Fig. 6); this difference between the cages was not significant (p = 0.47). Overall, the cages did not significantly stabilize the spine in extension (p = 0.48; Fig. 7), but the addition of translaminar screw fixation did significantly stabilize it; the median range of motion with the cages and screw fixation was 34 percent of that with the cages alone (p = 0.013; Fig. 7).

The changes in the neutral-zone parameter mirrored those observed for range of motion. There were no significant differences between the two types of cages in either flexion (p = 0.067) or extension (p = 0.95). Insertion of the interbody cages significantly reduced the neutral zone in flexion compared with that of the specimens in the intact condition; the decrease was from 1.3 degrees with the specimens intact to 0.7 degree after insertion of the cages (p = 0.021). In extension, the cages significantly increased the neutral zone, from 0.5 degree in the intact specimens to 0.8 degree after insertion of the cages (p = 0.041).

Axial Rotation

In both the BAK and the SynCage group, the behavior of the specimens in left and right rotation was virtually symmetrical in all test conditions (Figs. 4-A and 4-B). Therefore, only the total rotational motion (left plus right) was analyzed.

The median range of motion after insertion of the interbody cages was 56 percent of that of the intact specimens in the BAK group and only 34 percent of that of the intact specimens in the SynCage group (Fig. 6); this difference between the cages was not significant (p = 0.15). The cages significantly reduced the range of motion compared with that in the intact condition; the median range of motion with the cages was 48 percent of that in the intact spine (p = 0.002; Fig. 7). The addition of translaminar screw fixation did not significantly increase the stability of the spine compared with the cage condition (p = 0.53; Fig. 7).

For the neutral-zone parameter in axial rotation, there were no significant differences between the two types of cages (median range of motion, 53 percent of that of the intact specimens in the BAK group and 39 percent of that of the intact specimens in the SynCage group; p = 0.20). The neutral zone was reduced from 0.7 degree in the intact specimens to 0.5 degree after insertion of the cages, and this change was marginally significant (p = 0.055).

Lateral Bending

The nonlinear behavior in the intact condition was apparent before insertion of both the BAK and the SynCage device (Figs. 5-A and 5-B). The behavior in left and right lateral bending was virtually symmetrical for both cage groups and all test conditions; therefore, only the total rotational motion was analyzed.

The median range of motion after insertion of the interbody cages was 29 percent of that of the intact specimens in the BAK group and 35 percent of that of the intact specimens in the SynCage group; this difference between the cages was not significant (p = 0.75; Fig. 6). The cages significantly reduced the range of motion compared with that of the intact specimens; the median range of motion with the cages was 29 percent of that in the intact spine (p = 0.002; Fig. 7). The addition of translaminar screw fixation did not significantly increase the stability of the spine compared with use of the cages alone (p = 0.40; Fig. 7).

For the neutral-zone parameter in lateral bending, there were no significant differences between the two types of cages (median range of motion, 49 percent of that of the intact specimens in the BAK group and 41 percent of that of the intact specimens in the SynCage group; p = 0.63). The neutral zone was significantly reduced from 3.2 degrees in the intact specimens to 1.0 degree after insertion of the cages (p = 0.041).

Discussion

Introduction | Materials and Methods | Results | Discussion | References

The current study adds to the increasing body of literature regarding the immediate stabilization provided by interbody cages. In this study, stabilization was defined as a decrease in motion after insertion of an implant. Previous comparative evaluation of posterior interbody cages revealed no significant differences between different designs¹⁴, but there have been no previous studies, to our knowledge, on different types of cages inserted with use of an anterior approach. The BAK system requires that two implants be placed, bilaterally, in the intervertebral space, with each implant penetrating both vertebrae for fixation. The SynCage implant is a single device, which is placed centrally into the disc space and achieves fixation through sharp teeth that engage both vertebrae. It is reasonable to expect that these two anterior implants would not stabilize the spine in similar ways, given the vastly different interfaces between the implants and the vertebral bone.

The findings of the current study are in remarkable agreement with data in the recent literature. The absence of any difference in the immediate stabilization provided by the two anterior cage designs was similar to the finding in a recent comparative study that there was no difference among three different types of posterior interbody cages¹⁴. Therefore, apparently large differences in design between interbody cages do not seem to have a significant effect on the immediate stabilization provided to the spine.

The ability of the anterior cages to stabilize the spine substantially in all loading directions except extension was observed previously^10,16. The lack of stabilization with the spine in extension is probably due to distraction of the articular facets. It is known that the facets provide resistance to extension by contact of the articular surfaces^13,29. During distraction of the intervertebral space, the facets also are distracted⁵, thereby reducing the posterior element resistance to extension. The damage to the anterior longitudinal ligament during insertion of the anterior cages may also be an explanation for the lack of stabilization in extension. However, in the comparative study of posterior lumbar interbody cages, a lack of stabilization in extension also was observed when the anterior longitudinal ligament was intact¹⁴. Therefore, we believe that distraction of the facets is the most reasonable explanation for the findings in extension.

Axial rotation loading is often a problem for spinal fixation devices²⁰. The comparative study of different types of posterior interbody cages revealed that they did not stabilize the spine in axial rotation¹⁴. Therefore, it is remarkable that the two types of anterior devices used in the current study stabilized the spine so effectively in axial rotation. The difference between the performances of the anterior and posterior implants may be due to destruction of the facets in the posterior procedure or possibly to the manner in which the anterior implants engage the vertebrae. Additional work in this area may be warranted.

The effect of supplementary posterior fixation with interbody cages as observed in the present study also has been noted previously. We found that translaminar screw fixation was effective in extension, the only direction in which the cages did not provide stabilization. This trend also was noted previously with respect to transfacet²⁷ and translaminar²³ screws. Pedicle screw fixation seems to provide stabilization in all loading directions, as noted by several investigators^3,10,14. In a recent clinical study, Fidler⁷ suggested that a simple interspinous H-graft with cerclage wire may vastly improve the results of interbody arthrodesis.

The use of human cadaveric specimens to evaluate these implant devices is important due to the different methods of vertebral fixation. Often, morphological characteristics of animals differ from those of humans⁶. Many animal models that have been used for the evaluation of spinal implants, most notably calf models, have unfused growth plates. Since interbody implants provide fixation near the growth plate, this may explain why the findings of some early studies of stabilization provided by interbody implants in calf or pig spines^3,4 differ from those of more recent studies of human cadaveric specimens^10,14,16. The findings of the current study are similar to those of earlier studies of human models but differ substantially from those of animal studies^3,4, particularly with respect to extension and axial rotation. These differences may be further explained by morphological differences of the articular facets between the human and animal models since the facets are known to be important in these loading directions.

The current study had several limitations. First, the loads that were applied were pure moments without a compressive preload. Actual in vivo loading of the spine is more complex. However, the use of pure moments is a well defined protocol in spinal studies for the three-dimensional evaluation of injury and stabilization^17,18,21. Such a protocol has been found to result in intervertebral rotations similar to those observed in in vivo motion analyses²². Second, although the number of specimens per group (six) in our study is comparable with the number used in many investigations of spinal biomechanics, the statistical power can be relatively low, depending upon the critical differences between the cage designs and the variability in the data. Our data were analyzed with nonparametric methods since their normality could not be assumed. These methods are approximately 95 percent as powerful as parametric methods for normal distributions and are more powerful for non-normal distributions⁹. Our experimental design of six specimens per group was based upon a power of 80 percent and assumed a critical difference of 30 percent and a standard deviation of 15 percent in the ratio of motion in the cage condition to motion in the intact condition between the two designs. The observed variances were at or below this level in flexion and lateral bending but not in extension or axial rotation. In axial rotation, the average standard deviation was 20 percent of motion in the intact condition, which corresponded to a power of about 65 percent. In extension, the average standard deviation was 40 percent of motion in the intact specimens, which corresponded to a power of only 20 percent in detecting a difference between the cages. Thus, it is possible that there was a difference between the cages in extension that was not detected in this experiment.

Given the six specimens per group, we anticipated that our power to compare the intact, cage, and cage-with-fixation groups was 80 percent. On the basis of the measured standard deviations, the actual power values for detection of 40 percent differences in motion between the intact and cage conditions ranged from 70 to 88 percent in the different loading directions. To detect 50 percent differences in motion between the cage and cage-with-fixation conditions, the power ranged from 50 to 70 percent in the different loading directions. In general, substantial nonsignificant numerical differences between groups were not observed; therefore, low power was not of great concern in the comparisons of the intact and cage conditions or those of the cage and cage-with-fixation conditions.

A third limitation of our study is that it addressed only the immediate stabilization of the spine. It is possible that immediate stabilization would decrease postoperatively due to relaxation of the load in the annulus fibrosus or settling of the implant into the adjacent vertebral bone, or both, thereby loosening the annulus. Relaxation of the annulus may be an important issue, since collagenous soft tissues relax to approximately 20 to 30 percent of the maximum load^19,28. Settling of these implants is inevitable, and the degree to which settling decreases stabilization was described recently¹¹. Incorporation of bone into and around the implant appears to enhance stabilization²⁵. In summary, the time-course of the stabilization and the development of fusion is complicated and not completely understood with respect to interbody cages. The process is obviously an interplay among immediate stabilization, the loss of stabilization postoperatively due to a variety of factors, and improved stabilization due to bone ingrowth.

The obvious question, in any biomechanical study, concerns the amount of immediate stabilization that is required for bone ingrowth and eventual intervertebral fusion. Clearly, relevant animal studies will be required to ultimately answer this question. However, when animal studies are used to compare fundamental differences in implant design, they introduce substantial variability. Therefore, in vitro experimental approaches such as those described herein are advantageous.

This study was designed to detect a 30 percent difference between two types of implants with regard to the ratio of motion associated with the implant to that in intact specimens. It was thought that such a difference in immediate stabilization might be clinically relevant. Furthermore, to detect differences between the intact and cage conditions and between the cage-only condition and that consisting of a cage with translaminar screw fixation, differences in motion of 40 and 50 percent were deemed clinically relevant. It is not possible to directly translate a significant change in motion to eventual success of an arthrodesis. The growth of bone into an interbody cage depends upon many factors, including the relative motions, or micromotions, at the interface between the implant and the host bone. Clearly, relative vertebral motions of 1 to 2 degrees, as measured herein in the specimens with cages with translaminar screw fixation, represent very small motions that likely translate to small micromotions at the bone-implant interface. Again, whether these motions are sufficiently small to permit bone ingrowth requires investigation in a living model.

Note: The authors thank Dr. Ruth Milner of the Clinical Epidemiology and Evaluation Centre at Vancouver Hospital for her assistance with the statistical analysis.

References

Introduction | Materials and Methods | Results | Discussion | References

Bagby, G. W.: Arthrodesis by the distraction-compression method using a stainless steel implant. Orthopedics,11: 931-934, 1988.11931 1988 [PubMed]

Brantigan, J. W., and Steffee, A. D.: A carbon fiber implant to aid interbody fusion. Two-year clinical results in the first 26 patients. Spine,48: 2106-2117, 1993.482106 1993

Brodke, D. S.; Dick, J. C.; Kunz, D. N.; McCabe, R.; and Zdeblick, T. A.: Posterior lumbar interbody fusion. A biomechanical comparison, including a new threaded cage. Spine,22: 26-31, 1997.2226 1997 [PubMed]

Butts, M.; Kuslich, S.; and Bechtold, J.: Biomechanical analysis of a new method for spinal interbody fixation. In Advances in Bioengineering, pp. 95-96. Edited by A. Erdman. New York, American Society of Mechanical Engineers, 1987.

Chen, D.; Fay, L. A.; Lok, J.; Yuan, P.; Edwards, W. T.; and Yuan, H. A.: Increasing neuroforaminal volume by anterior interbody distraction in degenerative lumbar spine. Spine,20: 74-79, 1995.2074 1995 [PubMed]

Cotterill, P. C.; Kostuik, J. P.; D'Angelo, G. L.; Fernie, G. R.; and Maki, B. E.: An anatomical comparison of the human and bovine thoracolumbar spine. J. Orthop. Res.,4: 298-303, 1986.4298 1986 [PubMed]

Fidler, M. W.: Spinal fusion: a combined anterior and supplementary interspinous technique. European Spine J.,6: 214-218, 1997.6214 1997

François, R. J.; Bywaters, E. G.; and Aufdermaur, M.: Illustrated glossary for spinal anatomy. With explanations and a French and German translation. Rheumatol. Internat.,5: 241-245, 1985.5241 1985

Glantz, S. A.: Primer of Biostatistics. Ed. 4, pp. 326-327. New York, McGraw-Hill, 1997.

Glazer, P. A.; Colliou, O.; Klisch, S. M.; Bradfore, D. S.; Bueff, H. U.; and Lotz, J. C.: Biomechanical analysis of multilevel fixation methods in the lumbar spine. Spine,22: 171-182, 1997.22171 1997 [PubMed]

Hoffer, Z.; Oxland, T. R.; Cripton, P. A.; Sherman, J.; and Nolte, L. P.: Compressive loading affects the stabilization provided by lumbar interbody cages. Trans. Orthop. Res. Soc.,23: 1065, 1998.231065 1998

Kuslich, S. D.; Ulstrom, C. L.; Griffith, S. L.; Ahern, J. W.; and Dowdle, J. D.: The Bagby and Kuslich method of lumbar interbody fusion. History, techniques, and 2-year follow-up results of a United States prospective, multicenter trial. Spine,23: 1267-1279, 1998.231267 1998 [PubMed]

Lorenz, M.; Patwardhan, A.; and Vanderby, R., Jr.: Load-bearing characteristics of lumbar facets in normal and surgically altered spinal segments. Spine,8: 122-130, 1983.8122 1983 [PubMed]

Lund, T.; Oxland, T. R.; Jost, B.; Cripton, P.; Grassmann, S.; Etter, C.; and Nolte, L. P.: Interbody cage stabilisation in the lumbar spine: a biomechanical evaluation of cage design, posterior instrumentation and bone density. J. Bone and Joint Surg.,80-B(2): 351-359, 1998.80-B(2)351 1998

Magerl, F. P.: Stabilization of the lower thoracic and lumbar spine with external skeletal fixation. Clin. Orthop.,189: 125-141, 1984.189125 1984 [PubMed]

Nibu, K.; Panjabi, M. M.; Oxland, T.; and Cholewicki, J.: Multidirectional stabilizing potential of BAK interbody spinal fusion system for anterior surgery. J. Spinal Dis.,10: 357-362, 1997.10357 1997

Nolte, L.-P.; Steffen, R.; Krämer, J.; and Jergas, M.: Der Fixateur interne: Eine vergleichende biomechanische Studie verschiedener Systeme. Akt. Traumatol.,23: 20-26, 1993.2320 1993

Oxland, T. R.; Lund, T.; Jost, B.; Cripton, P.; Lippuner, K.; Jaeger, P.; and Nolte, L. P.: The relative importance of vertebral bone density and disc degeneration in spinal flexibility and interbody implant performance. An in vitro study. Spine,21: 2558-2569, 1996.212558 1996 [PubMed]

Panagiotacopulos, N. D.; Pope, M. H.; Bloch, R.; and Krag, M. H.: Water content in human intervertebral discs. Part II. Viscoelastic behavior. Spine,12: 918-924, 1987.12918 1987 [PubMed]

Panjabi, M. M.; Abumi, K.; Duranceau, J.; and Crisco, J. J.: Biomechanical evaluation of spinal fixation devices: II. Stability provided by eight internal fixation devices. Spine,13: 1135-1140, 1988.131135 1988 [PubMed]

Panjabi, M.; Abumi, K.; Duranceau, J.; and Oxland, T.: Spinal stability and intersegmental muscle forces. A biomechanical model. Spine,14: 194-200, 1989.14194 1989 [PubMed]

Panjabi, M. M.; Oxland, T. R.; Yamamoto, I.; and Crisco, J. J.: Mechanical behavior of the human lumbar and lumbosacral spine as shown by three-dimensional load-displacement curves. J. Bone and Joint Surg.,76-A: 413-424, March 1994.76-A413 1994

Rathonyi, G. C.; Oxland, T. R.; Gerich, U.; Grassmann, S.; and Nolte, L. P.: The role of supplemental translaminar screws in anterior lumbar interbody fixation: a biomechanical study. European Spine J.,7: 400-407, 1998.7400 1998

Ray, C. D.: Threaded titanium cages for lumbar interbody fusions. Spine,22: 667-679, 1997.22667 1997 [PubMed]

Sandhu, H. S.; Turner, S.; Kabo, J. M.; Kanim, L. E.; Liu, D.; Nourparvar, A.; Delamarter, R. B.; and Dawson, E. G.: Distractive properties of a threaded interbody fusion device. An in vivo model. Spine,21: 1201-1210, 1996.211201 1996 [PubMed]

Tencer, A. F.; Hampton, D.; and Eddy, S.: Biomechanical properties of threaded inserts for lumbar interbody spinal fusion. Spine,20: 2408-2414, 1995.202408 1995 [PubMed]

Volkman, T.; Horton, W. C.; and Hutton, W. C.: Transfacet screws with lumbar interbody reconstruction: biomechanical study of motion segment stiffness. J. Spinal Dis.,9: 425-432, 1996.9425 1996

Woo, S. L-Y.; An, K.-N.; Arnoczky, S. P.; Wayne, J. S.; Fithian, D. C.; and Myers, B. S.: Anatomy, biology, and biomechanics of tendon, ligament, and meniscus. In Orthopaedic Basic Science, pp. 45-87. Edited by S. R. Simon. American Academy of Orthopaedic Surgeons, 1994.

Yang, K. H., and King, A. I.: Mechanism of facet load transmission as a hypothesis for low-back pain. Spine,9: 557-565, 1984.9557 1984 [PubMed]

Topics