A history of spine biomechanics. Focus on 20th century progress.

The application of mechanical principles to problems of the spine dates to antiquity. Significant developments related to spinal anatomy and biomechanical behaviour made by Renaissance and post-Renaissance scholars through the end of the 19th century laid a strong foundation for the developments since that time. The objective of this article is to provide a historical overview of spine biomechanics with a focus on the developments in the 20th century. The topics of spine loading, spinal posture and stability, spinal kinematics, spinal injury, and surgical strategies were reviewed.

Cervical spine injuries and flexibilities following axial impact with lateral eccentricity.

PURPOSE: Determine the effects of dynamic injurious axial compression applied at various lateral eccentricities (lateral distance to the centre of the spine) on mechanical flexibilities and structural injury patterns of the cervical spine. METHODS: 13 three-vertebra human cadaver cervical spine specimens (6 C3-5, 3 C4-6, 2 C5-7, 2 C6-T1) were subjected to pure moment flexibility tests (±1.5 Nm) before and after impact trauma was applied in two groups: low and high lateral eccentricity (1 and 150 % of the lateral diameter of the vertebral body, respectively). Relative range of motion (ROM) and relative neutral zone (NZ) were calculated as the ratio of post and pre-trauma values. Injuries were diagnosed by a spine surgeon and scored. Classification functions were developed using discriminant analysis. RESULTS: Low and high eccentric loading resulted in primarily bony fractures and soft tissue injuries, respectively. Axial impacts with high lateral eccentricities resulted in greater spinal motion in lateral bending [median relative ROM 3.5 (interquartile range, IQR 2.3) vs. 1.4 (IQR 0.5) and median relative NZ 4.7 (IQR 3.7) vs. 2.3 (IQR 1.1)] and in axial rotation [median relative ROM 5.3 (IQR 13.7) vs. 1.3 (IQR 0.5), p < 0.05 for all comparisons] than those that resulted from low eccentricity impacts. The developed classification functions had 92 % classification accuracy. CONCLUSIONS: Dynamic axial compression loading of the cervical spine with high lateral eccentricities produced primarily soft tissue injuries resulting in more post-injury spinal flexibility in lateral bending and axial rotation than that associated with the bony fractures resulting from low eccentricity impacts.

The effect of lateral eccentricity on failure loads, kinematics, and canal occlusions of the cervical spine in axial loading.

Current neck injury criteria do not include limits for lateral bending combined with axial compression and this has been observed as a clinically relevant mechanism, particularly for rollover motor vehicle crashes. The primary objectives of this study were to evaluate the effects of lateral eccentricity (the perpendicular distance from the axial force to the centre of the spine) on peak loads, kinematics, and spinal canal occlusions of subaxial cervical spine specimens tested in dynamic axial compression (0.5 m/s). Twelve 3-vertebra human cadaver cervical spine specimens were tested in two groups: low and high eccentricity with initial eccentricities of 1 and 150% of the lateral diameter of the vertebral body. Six-axis loads inferior to the specimen, kinematics of the superior-most vertebra, and spinal canal occlusions were measured. High speed video was collected and acoustic emission (AE) sensors were used to define the time of injury. The effects of eccentricity on peak loads, kinematics, and canal occlusions were evaluated using unpaired Student t-tests. The high eccentricity group had lower peak axial forces (1544 ± 629 vs. 4296 ± 1693 N), inferior displacements (0.2 ± 1.0 vs. 6.6 ± 2.0 mm), and canal occlusions (27 ± 5 vs. 53 ± 15%) and higher peak ipsilateral bending moments (53 ± 17 vs. 3 ± 18 Nm), ipsilateral bending rotations (22 ± 3 vs. 1 ± 2°), and ipsilateral displacements (4.5 ± 1.4 vs. -1.0 ± 1.3 mm, p<0.05 for all comparisons). These results provide new insights to develop prevention, recognition, and treatment strategies for compressive cervical spine injuries with lateral eccentricities.

Acoustic emission signals can discriminate between compressive bone fractures and tensile ligament injuries in the spine during dynamic loading.

Acoustic emission (AE) sensors are a reliable tool in detecting fracture; however they have not been used to differentiate between compressive osseous and tensile ligamentous failures in the spine. This study evaluated the effectiveness of AE data in detecting the time of injury of ligamentum flavum (LF) and vertebral body (VB) specimens tested in tension and compression, respectively, and in differentiating between these failures. AE signals were collected while LF (n=7) and VB (n=7) specimens from human cadavers were tested in tension and compression (0.4m/s), respectively. Times of injury (time of peak AE amplitude) were compared to those using traditional methods (VB: time of peak force, LF: visual evidence in high speed video). Peak AE signal amplitudes and frequencies (using Fourier and wavelet transformations) for the LF and VB specimens were compared. In each group, six specimens failed (VB, fracture; LF, periosteal stripping or attenuation) and one did not. Time of injury using AE signals for VB and LF specimens produced average absolute differences to traditional methods of 0.7 (SD=0.2) ms and 2.4 (SD=1.5) ms (representing 14% and 20% of the average loading time), respectively. AE signals from VB fractures had higher amplitudes and frequencies than those from LF failures (average peak amplitude 87.7 (SD=6.9) dB vs. 71.8 (SD=9.8)dB for the inferior sensor, p<0.05; median characteristic frequency from the inferior sensor 97 (interquartile range, IQR, 41) kHz vs. 31 (IQR 2) kHz, p<0.05). These findings demonstrate that AE signals could be used to delineate complex failures of the spine.

Design and biomechanical evaluation of a rodent spinal fixation device.

STUDY DESIGN: An in vitro and in vivo study in rats. OBJECTIVES: To design a novel rat spinal fixation device and investigate its biomechanical effectiveness in stabilizing the spine up to 8 weeks post injury. METHODS: A fixation device made of polyetheretherketone was designed to stabilize the spine via bilateral clamping pieces. The device effectiveness was assessed in a Sprague-Dawley rat model after it was applied to a spine with a fracture-dislocation injury produced at C5-C6. Animals were euthanized either immediately (n=6) or 8 weeks (n=9) post-injury and the C3-T1 segment of the cervical spine was removed for biomechanical evaluation. Segments of intact spinal columns (C3-T1) (n=6) served as uninjured controls. In these tests, anterior-posterior shear forces were applied to the C3 vertebra to produce flexion and extension bending moments at the injury site (peak 12.8 Nmm). The resultant two-dimensional motions at the injury site (that is, C5-C6) were measured using digital imaging and reported as ranges of motion (ROM) or neutral zones (NZ). RESULTS: Flexion/extension ROMs (average±s.d.) were 18.1±3.3°, 19.9±7.5° and 1.5±0.7°, respectively for the intact, injured/fixed, and injured/8-week groups, with the differences being highly significant for the injured/8-week group (P=0.0002). Flexion/extension NZs were 3.4±2.8°, 5.0±2.4°, and 0.7±0.5°, respectively for the intact, injured/fixed, and injured/8-week groups, with the differences being significant for the injured/8-week group (P=0.04). CONCLUSION: The device acutely stabilizes the spine and promotes fusion at the site of injury.

Can extra-articular strains be used to measure facet contact forces in the lumbar spine? An in-vitro biomechanical study.

Experimental measurement of the load-bearing patterns of the facet joints in the lumbar spine remains a challenge, thereby limiting the assessment of facet joint function under various surgical conditions and the validation of computational models. The extra-articular strain (EAS) technique, a non-invasive measurement of the contact load, has been used for unilateral facet joints but does not incorporate strain coupling, i.e. ipsilateral EASs due to forces on the contralateral facet joint. The objectives of the present study were to establish a bilateral model for facet contact force measurement using the EAS technique and to determine its effectiveness in measuring these facet joint contact forces during three-dimensional flexibility tests in the lumbar spine. Specific goals were to assess the accuracy and repeatability of the technique and to assess the effect of soft-tissue artefacts. In the accuracy and repeatability tests, ten uniaxial strain gauges were bonded to the external surface of the inferior facets of L3 of ten fresh lumbar spine specimens. Two pressure-sensitive sensors (Tekscan) were inserted into the joints after the capsules were cut. Facet contact forces were measured with the EAS and Tekscan techniques for each specimen in flexion, extension, axial rotation, and lateral bending under a +/- 7.5 N m pure moment. Four of the ten specimens were tested five times in axial rotation and extension for repeatability. These same specimens were disarticulated and known forces were applied across the facet joint using a manual probe (direct accuracy) and a materials-testing system (disarticulated accuracy). In soft-tissue artefact tests, a separate set of six lumbar spine specimens was used to document the virtual facet joint contact forces during a flexibility test following removal of the superior facet processes. Linear strain coupling was observed in all specimens. The average peak facet joint contact forces during flexibility testing was greatest in axial rotation (71 +/- 25 N), followed by extension (27 +/- 35 N) and lateral bending (25 +/- 28 N), and they were most repeatable in axial rotation (coefficient of variation, 5 per cent). The EAS accuracy was about 20 per cent in the direct accuracy assessment and about 30 per cent in the disarticulated accuracy test. The latter was very similar to the Tekscan accuracy in the same test. Virtual facet loads (r.m.s.) were small in axial rotation (12 N) and lateral bending (20 N), but relatively large in flexion (34 N) and extension (35 N). The results suggested that the bilateral EAS model could be used to determine the facet joint contact forces in axial rotation but may result in considerable error in flexion, extension, and lateral bending.

Strength indices from pQCT imaging predict up to 85% of variance in bone failure properties at tibial epiphysis and diaphysis.

Our primary objective was to validate the Bone Strength Index for compression (BSIC) by determining the amount of variance in failure load and stiffness that was explained by BSIC and bone properties at two distal sites in human cadaveric tibiae when tested in axial compression. Our secondary objective was to assess the variance in failure moment and flexural rigidity that was explained by bone properties, geometry and strength indices in the tibial diaphysis when tested in 4-point bending. Twenty cadaver tibiae pairs from 5 female and 5 male donors (mean age 74 yrs, SD 6 yrs) were measured at the distal epiphysis (4 and 10% sites of the tibial length from the distal end) and diaphysis (50 and 66% sites) by peripheral Quantitative Computed Tomography (pQCT; XCT 2000, Stratec). After imaging, we conducted axial compression tests on the distal tibia and 4-point bending tests on the diaphysis. Total bone mineral content and BSIC (product of total area and squared density of the cross-section) at the 4% site predicted 75% and 85% of the variance in the failure load and 52% and 57% in stiffness, respectively. At the diaphyseal sites 80% or more of the variance in failure moment and/or flexural rigidity was predicted by total and cortical area and content, geometry and strength indices corresponding to the axes of bending.

Tibial geometry is associated with failure load ex vivo: a MRI, pQCT and DXA study.

UNLABELLED: We studied the relations between bone geometry and density and the mechanical properties of human cadaveric tibiae. Bone geometry, assessed by MRI and pQCT, and bone density, assessed by DXA, were significantly associated with bone's mechanical properties. However, cortical density assessed by pQCT was not associated with mechanical properties. INTRODUCTION: The primary objective of this study was to determine the contribution of cross-sectional geometry (by MRI and pQCT) and density (by pQCT and DXA) to mechanical properties of the human cadaveric tibia. METHODS: We assessed 20 human cadaveric tibiae. Bone cross-sectional geometry variables (total area, cortical area, and section modulus) were measured with MRI and pQCT. Cortical density and areal BMD were measured with pQCT and DXA, respectively. The specimens were tested to failure in a four-point bending apparatus. Coefficients of determination between imaging variables of interest and mechanical properties were determined. RESULTS: Cross-sectional geometry measurements from MRI and pQCT were strongly correlated with bone mechanical properties (r(2) range from 0.55 to 0.85). Bone cross-sectional geometry measured by MRI explained a proportion of variance in mechanical properties similar to that explained by pQCT bone cross-sectional geometry measurements and DXA measurements. CONCLUSIONS: We found that there was a close association between geometry and mechanical properties regardless of the imaging modality (MRI or pQCT) used.

Femoral neck cortical geometry measured with magnetic resonance imaging is associated with proximal femur strength.

INTRODUCTION: Magnetic resonance imaging (MRI) is a promising medical imaging technique that we used to assess femoral neck cortical geometry. OBJECTIVES: Our primary objective was to assess whether cortical bone in the femoral neck assessed by MRI was associated with failure load in a simulated sideways fall, with and without adjustment for total bone size. Our secondary objective was to assess the reliability of the MRI measurements. MATERIALS AND METHODS: We imaged 34 human cadaveric proximal femora using MRI and dual-energy X-ray absorptiometry (DXA). MRI measurements of cross-sectional geometry at the femoral neck were the cortical cross-sectional area (CoCSA(MRI)), second area moment of inertia (x axis; Ix(MRI)), and section modulus (x axis; Zx(MRI)). DXA images were analyzed with the standard Hologic protocol. From DXA, we report the areal bone mineral density (aBMD(DXA)) in the femoral neck and trochanteric subregions of interest. The femora were loaded to failure at 100 mm/s in a sideways fall configuration (15 degrees internal rotation, 10 degrees adduction). RESULTS AND OBSERVATIONS: Failure load (N) was the primary outcome. We observed that the femoral neck CoCSA(MRI) and Ix(MRI) were strongly associated with failure load (r (2)=0.46 and 0.48, respectively). These associations were similar to those between femoral neck aBMD and failure load (r (2)=0.40), but lower than the associations between trochanteric aBMD and failure load (r (2)=0.70). CONCLUSION: We report that MRI holds considerable promise for measuring cortical bone geometry in the femoral neck and for predicting strength at the proximal femur.

Ipsilateral shoulder and elbow replacements: on the risk of periprosthetic fracture.

BACKGROUND: Ipsilateral shoulder and elbow replacements may leave only a short segment of bone bridging the two implants in the humerus. The potential for high stress concentrations as a result of this geometry has been a concern with regard to periprosthetic fracture, especially with osteoporotic bone. The study aims to determine the optimum length of the bone-bridge between shoulder and elbow humeral implants, and to assess the effect of filling the canal with cement. METHODS: A three-dimensional finite element model was used to compare the stresses between a humerus with a solitary prosthesis and a humerus with both proximal and distal cemented prostheses. The length of the bone-bridge and the effect of filling the canal with cement were studied under bending and torsion. FINDINGS: Gradual load transfer from prosthesis to bone was observed for all cases, and no stress concentration was evident. The length of the bone-bridge had no deleterious effect on stresses in the humerus, and filling the canal with cement did not appreciably decrease the loads carried by the humerus. INTERPRETATION: The length of the bone-bridge between stem tips has little effect on the resultant stresses in the humerus. Filling the canal with cement adds little benefit to the structural integrity of the humerus. Ipsilateral shoulder and elbow prostheses may be considered independent of one another in terms of risk of periprosthetic fracture.

Allograft impaction and cement penetration after revision hip replacement. A histomorphometric analysis in the cadaver femur.

We studied various aspects of graft impaction and penetration of cement in an experimental model. Cancellous bone was removed proximally and local diaphyseal lytic defects were simulated in six human cadaver femora. After impaction grafting the specimens were sectioned and prepared for histomorphometric analysis. The porosity of the graft was lowest in Gruen zone 4 (52%) and highest in Gruen zone 1 (76%). At the levels of Gruen zones 6 and 2 the entire cross-section was almost filled with cement. Cement sometimes reached the endosteal surface in other Gruen zones. The mean peak impaction forces exerted with the impactors were negatively correlated with the porosity of the graft.

Fixation of trochanteric slide osteotomies: a biomechanical study.

OBJECTIVE: (1) Determine the effect of a compressive force on the stability of trochanteric slide osteotomies repaired with a cable repair system or a suturing technique. (2) Develop an approach to surgical decision making for trochanteric repair. DESIGN: Muscle forces acting on the greater trochanter were experimentally modeled by the application of shear and compressive loads to osteotomized greater trochanters. A repeated measures design was used to compare suture and cable fixation. BACKGROUND: The use of cables and wires for trochanteric repair has been associated with a high incidence of acetabular loosening and trochanteric bursitis. With trochanteric slide osteotomies, the vastus lateralis remains attached to the trochanter, which results in a compressive force being generated across the osteotomy and relatively small shear forces. The use of less rigid fixation techniques for trochanteric repair, such as sutures, may reduce the complications of cables and wires. METHODS: Seven cadaveric femora with trochanteric osteotomies were tested sequentially after repair with a cable system and with a suturing technique. A cyclic shear load of constant amplitude was applied while a compressive load was decreased in a stepwise fashion. Migration and cyclic motion of the trochanter were measured, and the coefficient of friction was also determined. RESULTS: Cyclic motions of the trochanter in both superior and anterior directions were generally less than 0.5 mm and were not significantly different between the cables and sutures at high compressive loads. At low compressive loads, cyclic motion was significantly lower with the cable system. CONCLUSIONS: Compression across the trochanteric slide osteotomy has a significant effect on stability. Cyclic motion of the trochanter is similar for both suture or cable repair of a trochanteric slide with good preservation of soft tissue attachments. RELEVANCE: Based on theoretical and experimental evidence, repair of trochanteric slide osteotomies with a suture technique may be a viable alternative to the use of cables and wires in selected cases.

The effect of nucleotomy on lumbar spine mechanics in compression and shear loading.

STUDY DESIGN: An in vitro biomechanical investigation on human cadaveric specimens was conducted before and after nucleotomy. Endplate and vertebral body deformation patterns were measured under compression and shear loading, in addition to kinematics and disc pressure. OBJECTIVE: The working hypotheses of this study were that in compression, nucleotomy results in an altered deformation pattern of the endplate and that in shear, nucleotomy does not result in an altered endplate deformation pattern or disc pressure. SUMMARY OF BACKGROUND DATA: The pressure distributions within the intervertebral disc have been studied in compression loading but not in shear loading. Severe degeneration and surgical nucleotomy result in small nuclear pressure and altered loading distribution in compression. The effect of these changes on the vertebral endplate and the response under shear loads are not well understood. METHODS: Five L3-L4 and two L4-L5 functional spinal units were tested under compression and shear loading, intact and after nucleotomy. Vertebral body deformations, intradiscal pressure, and intervertebral kinematics were measured. A series of compression-type (maximum 1000 N) and shear-type (maximum 500 N) loads were applied. RESULTS: With nucleotomy, the disc pressure and the endplate strains decreased under compression, but the vertebral rim strains did not change. In shear, the vertebral rim and endplate strains did not change with nucleotomy. Disc pressure was lower in shear than in compression. CONCLUSION: Nucleotomy resulted in decreased disc pressure, decreased endplate deformation, and modified loading patterns onto the inferior vertebra in compression loading. However, nucleotomy did not appreciably affect the behavior of the disc in shear loading.

Does anterolateral cage insertion enhance immediate stabilization of the functional spinal unit? A biomechanical investigation.

STUDY DESIGN: The three-dimensional flexibility of six human lumbar functional spinal units was measured after the anterolateral insertion of an interbody cage. OBJECTIVES: To determine whether an interbody cage inserted from an anterolateral direction stabilizes the spine with respect to the intact state and to compare the finding with that from the same cage inserted from an anterior direction. SUMMARY OF BACKGROUND DATA: Several biomechanical studies have shown that interbody cages do not stabilize the spine in extension. It is suspected that this may be caused by the destruction of the anterior longitudinal ligament and anterior anulus fibrosus. METHODS: Six human cadaveric lumbar functional spinal units were tested under pure moments of flexion, extension, bilateral axial rotation, and bilateral lateral bending to a maximum of 10 Nm. The relative intervertebral motions were measured by an optoelectronic camera system with the spinal units in the intact condition, after discectomy, after anterolateral interbody cage stabilization, and with additional translaminar screw fixation. The implant used was a central, porous, contoured implant with endplate fit. The results were compared with those of a previous study, which used the same implant inserted from an anterior direction. RESULTS: The anterolateral cage insertion significantly decreased the motion in comparison with the intact situation in flexion and lateral bending, but not in extension or axial rotation. No differences were found between the anterior and anterolateral insertion approaches in flexion or extension, but differences were observed in axial rotation and lateral bending, in which the anterolateral approach resulted in more motion. Additional translaminar screw fixation reduced motion to below intact levels in all loading directions. None of the surgical procedures introduced asymmetrical behavior. CONCLUSIONS: Anterolateral cage insertion did not stabilize the spine in extension or axial rotation and was not different from the anterior approach in flexion and extension. Additional translaminar screw fixation stabilized in all directions.

Mapping the structural properties of the lumbosacral vertebral endplates.

STUDY DESIGN: A biomechanical investigation using indentation tests in a human cadaveric model to seek variation in the structural properties across the lower lumbar and sacral endplates. OBJECTIVES: To determine 1) if there are regional differences in endplate strength and 2) whether any differences identified are affected by spinal level (lumbar spine vs. sacrum) or endplate (superior vs. inferior). SUMMARY OF BACKGROUND DATA: It has been postulated that some regions of the vertebral body may be stronger than others. Conclusive data, either supporting or disproving this theory, would be valuable for both spine surgeons and implant designers because one mode of failure of interbody implants is subsidence into one or both adjacent vertebrae. METHODS: Indentation tests were performed at 27 standardized test sites in 62 bony endplates of intact human vertebrae (L3-S1) using a 3-mm-diameter, hemispherical indenter with a test rate of 0.2 mm/sec to a depth of 3 mm. The failure load and stiffness at each test site were determined using the load-displacement curves. Three-way analyses of variance were used to analyze the resulting data. RESULTS: Both the failure load and stiffness varied significantly across the endplate surfaces (P < 0.0001), with posterolateral regions being stronger and stiffer than the central regions. Characteristic distributions were identified in the lumbar superior, lumbar inferior, and sacral endplates. The failure load distributions were found to differ in 1) the superior lumbar and sacral endplates (P = 0.0077), 2) the inferior lumbar and sacral endplates (P = 0.0014), and 3) the superior and inferior lumbar endplates (P < 0.0001). The sacral and inferior lumbar endplates were both found to be stronger than the superior lumbar endplates (sacrum, P = 0.054; inferior, P = 0.008) but were not themselves significantly different (P = 0.89). CONCLUSIONS: Highly significant regional strength and stiffness variations were identified in the lumbar and sacral endplates. The center of the bone, where implants are currently placed, is the weakest part of the lumbar endplates and is not the strongest region of the sacral endplate.

Design and evaluation of a cryogenic soft tissue fixation device -- load tolerances and thermal aspects.

Mechanical studies of soft connective tissues often encounter methodological difficulties, particularly in the secure fixation of the tissues. A simple, inexpensive technique which allowed stable cryofixation of soft tissues in uniaxial loading machines was developed. The cryogenic fixation device was evaluated in terms of its fixation strength and the temperature gradients within the tested tissues. Human patellar ligaments and quadriceps tendons were tested successfully to an average failure load of 2219N (S.D. 448N) with mid-substance failures occurring in 90% of the specimens. The temperature gradients within porcine flexor and extensor tendons were determined and found to exhibit a typical diffusion profile. The fixation quality was dependent upon the initial block temperature and the desired testing time. In summary, the cryofixation device presented here is an effective tool for soft tissue fixation but the effect of this type of fixation on internal tissue temperatures and possible testing times must be acknowledged.

In vitro axial preload application during spine flexibility testing: towards reduced apparatus-related artefacts.

Presently, there is little consensus about how, or even if, axial preload should be incorporated in spine flexibility tests in order to simulate the compressive loads naturally present in vivo. Some preload application methods are suspected of producing unwanted "artefact" forces as the specimen rotates and, in doing so, influencing the resulting kinematics. The objective of this study was to quantitatively compare four distinct types of preload which have roots in contemporary experimental practice. The specific quantities compared were the reaction moments and forces resulting at the intervertebral disc and specimen kinematics. The preload types incorporated increasing amounts of caudal constraint on the preload application vector ranging from an unconstrained dead-load arrangement to an apparatus that allowed the vector to follow rotations of the specimen. Six human cadaveric spine segments were tested (1-L1/L2, 3-L2/L3, 1-L3/L4 and 1-L4/L5). Pure moments were applied to the specimens with each of the four different types of compressive preload. Kinematic response was measured using an opto-electronic motion analysis system. A six-axis load cell was used to measure reaction forces and moments. Artefact reaction moments and shear forces were significantly affected by preload application method and magnitude. Unconstrained preload methods produced high artefact moments and low artefact shear forces while more constrained methods did the opposite. A mechanical trade-off is suggested by our results, whereby unwanted moment can only be prevented at the cost of shear force production. When comparing spine flexibility studies, caution should be exercised to ensure preload was applied in a similar manner for all studies. Unwanted moments or forces induced as a result of preload application method may render the comparison of two seemingly similar studies inappropriate.

Biomechanics of stand-alone cages and cages in combination with posterior fixation: a literature review.

Interbody cages in the lumbar spine have met with mixed success in clinical studies. This has led many investigators to supplement cages with posterior instrumentation. The objective of this literature review is to address the mechanics of interbody cage fixation in the lumbar spine with respect to three-dimensional stabilization and the strength of the cage-vertebra interface. The effect of supplementary posterior fixation is reviewed. Only three-dimensional stabilization evaluations in human cadaveric models are included. These studies involve the application of different loads to the spine and the measurement of vertebral motion in flexion-extension, axial rotation, and lateral bending. There are no published studies which detected any differences between different cage designs. However, it does seem that cages inserted from an anterior direction provide better stabilization to the spine than those inserted from a posterior direction. In general, anterior cages stabilize better than posterior cages in axial rotation and lateral bending. Cages from both directions stabilized well in flexion, but not in extension. Supplementary posterior fixation with pedicle or translaminar screws substantially improves the stabilization in all directions. The strength of the cage-vertebra interface from studies using human cadaveric specimens is also reviewed. The axial compressive strength of this interface is highly dependent upon vertebral body bone density. Other factors such as preservation of the subchondral bony end-plate and cage design are clearly less important in the compressive strength. Supplementary posterior instrumentation does not enhance substantially the interface strength in axial compression.