Biomechanical Comparison of Anterolateral Plate, Lateral Plate, and Pedicle Screws-Rods for Enhancing Anterolateral Lumbar Interbody Cage Stabilization

SPINE Volume 29, Number 6, pp 635–641 ©2004, Lippincott Williams & Wilkins, Inc.

Biomechanical Comparison of Anterolateral Plate, Lateral Plate, and Pedicle Screws-Rods for Enhancing Anterolateral Lumbar Interbody Cage Stabilization Hakan Bozkus, MD, Robert H. Chamberlain, MS, BSE, Luis E. Perez Garza, MD, Neil R. Crawford, PhD, and Curtis A. Dickman, MD

Study Design. A repeated measures in vitro flexibility experiment was performed in calf spines. Objectives. To determine the biomechanical differences among three techniques for augmenting stability of an anterolateral lumbar threaded interbody cage. Background. Stand-alone interbody cages are known to inadequately stabilize the spine. Surgeons often add supplementary instrumentation for a more stable construct. Methods. Six L2–L5 calf spines (L3–L4 level instrumented) were tested: 1) intact; 2) with a single anterolateral interbody cage; 3) with cage plus anterolateral plating; 4) with cage plus lateral plating; and 5) with cage plus pedicle screw fixation. Specimens were loaded in each anatomic plane quasistatically (maximum 5.0 Nm). Angular motion was measured stereophotogrammetrically. Results. The stand-alone interbody cage allowed significantly less range of motion than normal during all loading modes except axial rotation. Addition of pedicle screws-rods, anterolateral plate, or lateral plate significantly further reduced range of motion in all planes. Pedicle screws slightly outperformed the anterolateral plate during extension and lateral bending and slightly outperformed the lateral plate during flexion, extension, and left axial rotation (range of motion differences ⬍0.65°, P ⬍ 0.05). The anterolateral plate outperformed the lateral plate during flexion and extension, whereas the lateral plate outperformed the anterolateral plate during lateral bending (range of motion difference ⬍0.57°, P ⬍ 0.05). Conclusion. Anterolateral or lateral lumbar plating increases stability significantly compared to stand-alone interbody cage fixation. These findings support anterolateral or lateral plate fixation as a potential clinical alternative to pedicle screws-rods in this role and may obviate the need for combined anterior and posterior approaches when spinal instability exists. [Key words: biomechanics, cage, interbody fusion, lumbar spine, spinal instrumentation] Spine 2004;29:635– 641

From the Spinal Biomechanics Research Laboratory, Barrow Neurological Institute, Phoenix, Arizona. Funding and instrumentation for this project provided by Medtronic Sofamor Danek, Memphis, Tennessee. Acknowledgment date: February 17, 2003. First revision date: April 9, 2003. Acceptance date: June 13, 2003. The device(s)/drug(s) is/are FDA approved or approved by corresponding national agency for this indication. Corporate/Industry funds were received to support this work. One or more of the author(s) has/have received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this manuscript, e.g., honoraria, gifts, consultancies, royalties, stocks, stock options, or decision-making position. Address correspondence and reprint requests to Curtis A. Dickman, MD, c/o Neuroscience Publications, Barrow Neurological Institute, St. Joseph’s Hospital & Medical Center, 350 W. Thomas Road, Phoenix, AZ 85013, USA; E-mail: [email protected]

The lumbar threaded interbody cage technique is often used to treat spinal degenerative disc disease and spinal instability. In some cases, however, stand-alone interbody cages may provide inadequate stability for fusion to occur.1 Biomechanical studies of stand-alone anterior, anterolateral, lateral, posterior, and posterolateral lumbar interbody cages have shown that all these cage orientations can provide adequate stabilization during certain loading modes, but none consistently stabilizes the spine well during all loading modes.1–10 To enhance stability, additional devices can be added after cage insertion. Cagli et al showed that the addition of pedicle screws-rods significantly stabilized injured spondylolisthetic spines during all loading modes compared to dual anterior cages alone.1 Lund et al4 added pedicle screws to three different posterior cage designs. All cage types provided the greatest stabilization during flexion– extension and lateral bending when used together with pedicle screw fixation. Other posterior techniques that increase spinal stiffness when added after interbody cage insertion include translaminar facet screws and transfacet screws.7,11 If an anterior or anterolateral interbody cage is placed, however, posterior implantation of pedicle screws, transfacet screws, or translaminar facet screws would require an additional surgical approach. In contrast, an anterolateral or lateral plate can be placed immediately after insertion of an anterior or anterolateral cage. Recently, Le Huec et al12 found that an anterolateral plate significantly increased stabilization regardless of whether it was connected to a lateral cage. This study, however, did not compare the biomechanical features of anterolateral plate augmentation to the proven technique of pedicle screw-rod augmentation. To our knowledge, no studies have compared the biomechanical advantage of different locations of plate placement as an adjunct to anterolateral interbody cage fixation, nor have they compared anterolateral lumbar plate augmentation to pedicle screw augmentation. This study therefore compared the biomechanical stability achieved by adding an anterolateral plate, a lateral plate, or pedicle screws-rods to specimens with single anterolateral interbody cage fixation. Materials and Methods Specimen Preparation. Six fresh calf lumbar spine segments (L2–L5) were studied. The mean age was 5 months, and the mean weight was 60 kg. Specimens were examined radiograph635

636 Spine • Volume 29 • Number 6 • 2004 ically to exclude pathology or spinal deformity. Specimens were wrapped in plastic bags and stored at ⫺20° C until tested. The specimens were thawed in a bath of normal saline at 30° C and carefully cleaned of muscular tissue while all ligaments, joint capsules, and discs were kept intact. For testing, L5 was firmly embedded in a metal fixture with polymethylmethacrylate and household wood screws and then attached to the base of the testing apparatus. L2 was similarly embedded in a metal fixture for the application of loads.

Experimental Protocol. Nonconstraining, nondestructive pure moment (torque) loading was applied to each specimen through a system of cables and pulleys in conjunction with a standard servohydraulic test system (858 Mini Bionix, MTS, Minneapolis, MN), as described previously.13 Torques were applied about the appropriate anatomic axes (maximum 5 Nm) to induce three different types of motion: flexion– extension, lateral bending, and axial rotation. Although the spine is loaded in vivo through multiple linear muscle attachments, it is infeasible to attempt to replicate this complex loading in the laboratory. Torque or pure moment loading can accurately reproduce physiologic bending and twisting of the spine14 and has the advantage over other simple loading techniques, such as offset compressive or shear force, that it is uniformly applied among specimens regardless of the site of load application or the dimensions of the specimen.15 For preconditioning, each load was applied to the maximum value and released after holding for 60 seconds. Preconditioning was performed three times, resulting in three load/unload cycles. After the third preconditioning cycle, the specimen was allowed to relax for 60 seconds before the data collection cycle began. During the data collection cycle, loads were applied quasistatically by 1.0 Nm increments until the maximum load of 5.0 Nm was reached. At each increment, the load was maintained for 45 seconds to allow viscoelastic creep. During the data collection cycle, three-dimensional motion was measured with the assistance of the Optotrak 3020 system (Northern Digital, Waterloo, Ontario, Canada). This system measures stereophotogrammetrically the three-dimensional displacement of infrared-emitting markers rigidly attached in a noncollinear arrangement to each vertebra. Custom software converted the marker coordinates to angles about each anatomic axis as described elsewhere.16,17 The resolution of the angular output was 0.01°, and the experimental error associated with this method was ⫾ 0.1° based on experiments in which the angle as registered on a machinist’s high-precision angular index table was compared to the value calculated from the optical system.

Surgical Procedure. All surgical procedures were performed at L3–L4. Specimens were tested: 1) intact; 2) after a single cage was inserted in an anterolateral trajectory; 3) with a cage plus an anterolaterally placed plate; 4) with a cage plus a laterally placed plate; and 5) with a cage plus pedicle screw-rod fixation. Each specimen was tested over 2 to 3 days. Specimens were refrozen at the end of each testing day to preserve their mechanical properties. The order in which conditions 3, 4, and 5 were tested was systematically varied so that 2 of 6 specimens were tested first in each of the 3 conditions involving the cage together with additional devices. This procedure was intended to minimize the chance of bias caused by loosening of the cage with progressive testing. Before an anterolateral cage was placed, a partial discec-

Figure 1. Transected specimen demonstrating typical positions and trajectories of two pedicle screws, two lateral screws for lateral plate attachment, two anterolateral screws for anterolateral plate attachment, and anterolaterally oriented threaded interbody cage.

tomy was performed using a scalpel, pituitary rongeurs, and a curette to remove the disc material where the cage was to be placed. In the calf spine, the endplates are naturally concave at L3 and convex at L4. We therefore shaped the exposed endplates using the Midas Rex high-speed drill (Medtronic Midas Rex, Fort Worth, TX) with a cylindrical bit to make them more uniformly parallel before reaming a hole for the threaded cage. A distractor was used to ensure proper interbody height, and a hand reamer was used to create a pilot hole for the cage. The cage was inserted by hand until seated 1 to 2 mm below the entry hole rim. Cages were InterFix™ (Medtronic Sofamor Danek, Memphis, TN) hollow, fenestrated titanium alloy cylinders with a diameter of 14 mm (5 specimens) or 16 mm (1 specimen) appropriate to the anatomy of a specimen. Cage length was 35 mm. The cage was inserted anterolaterally, oriented 45° from the midsagittal plane (Figure 1). For anterolateral or lateral fixation, a trapezoidal plate with bicortical screws was used (Xantus™, Medtronic Sofamor Danek; Figure 2). At their longest base, plates measured 54 mm by 19 mm wide. With each plate, four 6.5 mm ⫻ 35 mm screws were used. Pilot holes were drilled using a 4.4 mm diameter drill bit and a power drill. Plates were attached with the longer base of the trapezoidal plate oriented dorsally, and bicortical screw fixation was verified radiographically in each specimen. When the location of a plate was changed from lateral to anterolateral (or vice versa), new pilot holes were drilled, avoiding crossing the holes used previously (Figure 1). Posteriorly, a fixed cantilever pedicle screw-rod fixation system was used (TSRH威, Medtronic Sofamor Danek). All screws were 6.5 mm in diameter and either 40 or 45 mm long to accommodate the anatomy of a specimen. Screws were made of titanium alloy. A power drill with a 4.4 mm diameter bit was used to create the pilot holes for the pedicle screws. Rods were stainless steel and measured 5.5 mm in diameter.

Anterolateral Lumbar Interbody Cage Stabilization • Bozkus et al 637

Results

Figure 2. Titanium trapezoidal plate used for lateral or anterolateral augmentation of anterolateral interbody threaded cage. Bicortical screws are inserted with a variable trajectory. A thin locking cover plate prevents the screws from backing out. The longer base of the trapezoidal plate was oriented dorsally.

Data Analysis. From the raw data, well-established parameters were calculated to quantify spinal stability: range of motion (ROM), neutral zone (NZ), elastic zone (EZ), and stiffness coefficient (SC). The ROM was defined as the angle at maximum load (5.0 Nm). The NZ and EZ represent, respectively, the portion of the ROM in which the ligaments and instrumentation are under minimal loading and the portion of the ROM in which the ligaments are under stress. The boundary between NZ and EZ was defined as the resting angle after releasing the third preconditioning load and allowing 60 seconds for creep.18 The SC is the amount of resistance exerted by the motion segment per unit of angular deformation. The SC was calculated by the method of least squares from the linear portion of the EZ; it is the inverse slope of the load-deformation data at 2.0, 3.0, 4.0, and 5.0 Nm and their corresponding deformations. Cage alone was compared to the normal intact condition using paired two-tailed Student t tests. Combined instrumented conditions were compared to each other using analysis of variance followed by Student-Newman-Keuls procedure or compared to cage only using one-tailed paired Student t tests. P values ⬍ 0.05 were considered significant. For all statistical tests, power (probability of avoiding a type 2 or falsenegative error) was calculated using SigmaStat 2.0 (SPSS Inc., Chicago, IL). Posttesting Inspection. Secondary devices—anterolateral plate, lateral plate, or pedicle screw-rod fixation—were added in three sequential configurations (sequence order varied) to specimens already instrumented with the interbody cage. After completing all biomechanical tests, specimens were inspected to determine whether the screw trajectories for the different configurations overlapped, thereby potentially placing a device configuration tested later at a disadvantage compared to a configuration tested earlier. Screw crossing was evaluated by simultaneously inserting 1.5 mm surgical guide wires into the screw holes and evaluating whether any 2 wires contacted each other within the vertebra.

Range of Motion Insertion of an anterolateral interbody cage reduced the ROM to significantly less than normal during all loading modes except axial rotation (Figure 3). Addition of pedicle screw fixation, an anterolateral plate, or a lateral plate significantly further reduced ROM compared to normal (P ⬍ 0.013) or cage alone (P ⬍ 0.007) during all loading modes (Tables 1 and 2). Pedicle screws and rods allowed a significantly smaller ROM than the anterolateral plate during extension and left and right lateral bending (P ⬍ 0.046). Although statistically significant, these differences in ROM were 0.65° or less. Pedicle screws and rods allowed a significantly smaller ROM than the lateral plate during flexion, extension, and left axial rotation (P ⬍ 0.028). Although statistically significant, these differences in ROM were 0.40° or less. The anterolateral plate allowed a significantly smaller ROM than the lateral plate during flexion and extension, but the lateral plate allowed a significantly smaller ROM than the anterolateral plate during lateral bending. Although statistically significant, these differences in ROM were 0.57° or less. Stiffness Coefficient Insertion of an anterolateral interbody cage increased the SC to significantly greater than normal only during extension and right axial rotation (Figure 3). Addition of pedicle screws and rods, an anterolateral plate, or a lateral plate made the construct significantly stiffer than normal (P ⬍ 0.014) or cage alone (P ⬍ 0.017) during all loading modes (Tables 1 and 2). Pedicle screws and rods made the construct significantly stiffer than one or both orientations of the plate during flexion, extension, and right lateral bending. The anterolateral plate made the construct significantly stiffer than the lateral plate during flexion and extension, but the lateral plate made the construct significantly stiffer than the anterolateral plate during right lateral bending. During axial rotation, there was no significant difference in SC among any of the three combined instrumentation systems. Neutral Zone Insertion of an anterolateral interbody cage reduced the NZ to significantly less than normal only during flexion– extension (Figure 3). Addition of pedicle screw fixation, an anterolateral plate, or a lateral plate significantly further reduced NZ compared to normal during all loading modes (P ⬍ 0.032). Addition of pedicle screw fixation, an anterolateral plate, or a lateral plate significantly further reduced NZ compared to cage alone except for the lateral plate during flexion– extension (P ⫽ 0.493; Table 2). Pedicle screws and rods allowed a significantly smaller NZ than one or both orientations of the plate during all modes. The anterolateral plate allowed a significantly smaller NZ than the lateral plate during flexion– extension, but the lateral plate allowed a significantly smaller NZ than the anterolateral plate during

638 Spine • Volume 29 • Number 6 • 2004

Figure 3. Comparison of range of motion, stiffness coefficient, neutral zone, and elastic zone in each condition studied. *Represents no significant difference from normal (P ⬎ 0.05). Neutral zone is quantified bilaterally; for other parameters, flexion, left lateral bending, and left axial rotation are depicted as positive values whereas extension, right lateral bending, and right axial rotation are negative. Error bars show standard deviation.

lateral bending. The two placements of the plate performed similarly during axial rotation. Elastic Zone Insertion of an anterolateral interbody cage reduced the EZ to significantly less than normal during all loading modes except axial rotation (Figure 3). Addition of pedicle screw fixation, an anterolateral plate, or a lateral plate significantly further reduced EZ compared to normal (P ⬍ 0.017) or cage alone (P ⬍ 0.013) during all loading modes (Tables 1 and 2). Pedicle screws and rods allowed a significantly smaller EZ than one or both orientations of the plate during flexion, extension, and lat-

eral bending. The anterolateral plate allowed a significantly smaller EZ than the lateral plate during flexion and extension, but the lateral plate allowed a significantly smaller EZ than the anterolateral plate during lateral bending. There was no significant difference among any of the three combined instrumentation systems during axial rotation. Validation of Screw Paths In one specimen, the hole in which a dorsal screw of the lateral plate was placed crossed the path of an inferior pedicle screw at the caudal vertebra of the fixated level.

Anterolateral Lumbar Interbody Cage Stabilization • Bozkus et al 639

Table 1. P Values and Power for Statistical Comparisons of Normal Versus Instrumented Conditions Parameter

Flexion

Extension

Left Lateral Bending

Right Lateral Bending

Left Axial Rotation

Right Axial Rotation

Normal vs. cage only

ROM SC NZ EZ

0.001 (0.99) 0.118 (0.25) 0.006 (0.93) 0.002 (0.99)

0.010 (0.87) 0.006 (0.93) 0.006 (0.93) 0.016 (0.78)

0.002 (0.99) 0.462 (0.05) 0.050 (0.47) 0.001 (0.99)

0.004 (0.96) 0.056 (0.44) 0.050 (0.47) 0.016 (0.78)

0.597 (0.05) 0.488 (0.05) 0.396 (0.05) 0.416 (0.05)

0.084 (0.33) 0.014 (0.81) 0.396 (0.05) 0.029 (0.63)

Normal vs. anterolateral plate ⫹ cage

ROM SC NZ EZ

0.000 (1.00) 0.003 (0.98) 0.003 (0.97) 0.000 (1.00)

0.001 (1.00) 0.013 (0.81) 0.003 (0.97) 0.000 (1.00)

0.000 (1.00) 0.008 (0.91) 0.004 (0.96) 0.000 (1.00)

0.000 (1.00) 0.001 (1.00) 0.004 (0.96) 0.000 (1.00)

0.008 (0.90) 0.012 (0.84) 0.031 (0.37) 0.011 (0.85)

0.005 (0.95) 0.005 (0.95) 0.031 (0.37) 0.004 (0.96)

Normal vs. lateral plate ⫹ cage

ROM SC NZ EZ

0.001 (0.99) 0.011 (0.85) 0.024 (0.68) 0.000 (1.00)

0.001 (1.00) 0.004 (0.97) 0.024 (0.68) 0.000 (1.00)

0.000 (1.00) 0.008 (0.91) 0.005 (0.96) 0.000 (1.00)

0.000 (1.00) 0.001 (1.00) 0.005 (0.96) 0.000 (1.00)

0.013 (0.82) 0.001 (0.99) 0.026 (0.53) 0.017 (0.76)

0.008 (0.90) 0.006 (0.94) 0.026 (0.53) 0.008 (0.90)

Normal vs. pedicle screws ⫹ cage

ROM SC NZ EZ

0.000 (0.99) 0.000 (1.00) 0.002 (0.99) 0.000 (1.00)

0.001 (1.00) 0.001 (0.99) 0.002 (0.99) 0.000 (1.00)

0.000 (1.00) 0.003 (0.98) 0.005 (0.95) 0.000 (1.00)

0.000 (1.00) 0.002 (0.99) 0.005 (0.95) 0.000 (1.00)

0.005 (0.95) 0.003 (0.98) 0.006 (0.74) 0.006 (0.93)

0.005 (0.95) 0.003 (0.97) 0.006 (0.74) 0.005 (0.96)

Comparison

Values in bold are statistically significant (P ⬍ 0.05, power ⬎ 0.80). ROM ⫽ range of motion; SC ⫽ stiffness coefficient; NZ ⫽ neutral zone; EZ ⫽ elastic zone.

No other crossing paths were identified. This specimen was tested with pedicle screws and rods before being tested with the lateral plate. The ROM for this specimen with the lateral plate attached was an average of 0.06° smaller than the mean ROM of all 6 specimens with the lateral plate attached. The ROM for this specimen with

pedicle screws and rods attached was an average of 0.05° smaller than the mean ROM of all 6 specimens with pedicle screws and rods attached. These findings support our assumption that the influence of testing order on biomechanical outcomes was minimal and was well controlled.

Table 2. P Values and Power for Statistical Comparisons Among Instrumented Conditions Parameter

Flexion

Extension

Left Lateral Bending

Right Lateral Bending

Anterolateral plate vs. cage only

ROM SC NZ EZ

0.001 (0.99) 0.001 (0.99) 0.012 (0.69) 0.000 (1.00)

0.000 (1.00) 0.010 (0.72) 0.012 (0.69) 0.000 (1.00)

0.002 (0.97) 0.002 (0.96) 0.009 (0.75) 0.002 (0.97)

0.001 (0.99) 0.001 (0.99) 0.009 (0.75) 0.001 (0.99)

0.002 (0.98) 0.004 (0.90) 0.005 (0.65) 0.002 (0.97)

0.002 (0.98) 0.007 (0.82) 0.005 (0.65) 0.002 (0.97)

Lateral plate ⫹ cage vs. cage only

ROM SC NZ EZ

0.001 (0.99) 0.002 (0.98) 0.493 (0.50) 0.001 (0.99)

0.000 (1.00) 0.005 (0.89) 0.493 (0.50) 0.000 (1.00)

0.001 (0.99) 0.004 (0.91) 0.001 (0.99) 0.001 (0.99)

0.000 (1.00) 0.000 (1.00) 0.001 (0.99) 0.000 (1.00)

0.007 (0.82) 0.017 (0.59) 0.005 (0.67) 0.013 (0.65)

0.002 (0.97) 0.003 (0.94) 0.005 (0.67) 0.003 (0/95)

Pedicle screws ⫹ cage vs. cage only

ROM SC NZ EZ

0.001 (0.99) 0.001 (0.99) 0.002 (0.97) 0.001 (0.99)

0.000 (1.00) 0.001 (0.99) 0.002 (0.97) 0.000 (1.00)

0.001 (0.99) 0.002 (0.97) 0.002 (0.98) 0.001 (0.99)

0.000 (1.00) 0.001 (1.00) 0.002 (0.98) 0.000 (1.00)

0.004 (0.90) 0.004 (0.92) 0.003 (0.77) 0.007 (0.83)

0.002 (0.98) 0.006 (0.85) 0.003 (0.77) 0.003 (0.95)

Lateral plate ⫹ cage vs. anterolateral plate ⫹ cage

ROM SC NZ EZ

0.017 (0.94) 0.043 (0.93) 0.027 (0.72) 0.025 (0.95)

0.003 (0.99) 0.016 (0.96) 0.027 (0.72) 0.035 (0.95)

0.019 (0.68) ⬎0.095 (0.31) 0.012 (0.71) 0.032 (0.61)

0.012 (0.84) 0.033 (0.55) 0.012 (0.71) 0.009 (0.86)

0.075 (0.56) ⬎0.071 (0.38) 0.125 (0.70) ⬎0.061 (0.41)

⬎0.129 (0.25) ⬎0.071 (0.38) 0.125 (0.70) ⬎0.265 (0.11)

Anterolateral plate ⫹ cage vs. pedicle screws ⫹ cage

ROM SC NZ EZ

0.101 (0.94) 0.043 (0.93) 0.411 (0.72) 0.059 (0.95)

0.046 (0.99) 0.077 (0.96) 0.411 (0.72) 0.104 (0.87)

0.024 (0.68) ⬎0.095 (0.31) 0.027 (0.71) 0.029 (0.61)

0.006 (0.84) 0.048 (0.55) 0.027 (0.71) 0.007 (0.86)

0.293 (0.56) ⬎0.071 (0.38) 0.093 (0.70) ⬎0.061 (0.41)

⬎0.129 (0.25) ⬎0.071 (0.38) 0.093 (0.70) ⬎0.265 (0.11)

Lateral plate ⫹ cage vs. pedicle screws ⫹ cage

ROM SC NZ EZ

0.002 (0.94) 0.003 (0.93) 0.016 (0.72) 0.002 (0.95)

0.000 (0.99) 0.002 (0.96) 0.016 (0.72) 0.005 (0.87)

0.698 (0.68) ⬎0.095 (0.31) 0.937 (0.71) 0.576 (0.61)

0.886 (0.84) 0.475 (0.55) 0.937 (0.71) 0.706 (0.86)

0.028 (0.56) ⬎0.071 (0.38) 0.014 (0.70) ⬎0.061 (0.41)

⬎0.129 (0.25) ⬎0.071 (0.38) 0.014 (0.70) ⬎0.265 (0.11)

Comparison

Values in bold are statistically significant (P ⬍ 0.05, power ⬎ 0.80). ROM ⫽ range of motion; SC ⫽ stiffness coefficient; NZ ⫽ neutral zone; EZ ⫽ elastic zone.

Left Axial Rotation

Right Axial Rotation

640 Spine • Volume 29 • Number 6 • 2004

Discussion Comparison of Anterolateral or Lateral Plate to Anterolateral Cage Alone The amount of immobilization required for solid fusion to occur is unclear, but it is generally accepted that motion must be reduced to within the normal range, preferably well within the normal range. However, a single anterolateral cage alone failed to reduce motion consistently to within normal ranges or to make the construct stiffer than normal (Figure 3). These findings support the conclusion that a stand-alone anterolateral interbody cage may be biomechanically inadequate and may potentially lead to poor clinical outcomes. The addition of an anterolateral or lateral plate to specimens with a cage alone significantly reduced the NZ, EZ, and ROM and significantly increased the SC. These findings were true in all directions of loading, even those not expected to perform well (e.g., bending in the plane in which the plate was oriented). These findings indicate that significantly better immobilization followed the addition of a plate in either position. Reduction in NZ means that there was less “play” between the instrumentation and bone during small loads. Reduction in EZ and increased SC means that the instrumentation provided more rigid resistance once under moderate loading. Comparison of Anterolateral or Lateral Plate to Pedicle Screw Fixation Pedicle screw fixation added to an interbody construct is already known to enhance stability and the likelihood for fusion.1 Our results show that pedicle screw-rod fixation outperformed both plate orientations in terms of relative percentages. Compared to pedicle screws and rods, the anterolaterally placed plate allowed as much as a 169% greater NZ, 88% greater EZ, and 107% greater ROM. Compared to pedicle screws and rods, the laterally placed plate allowed as much as 158% greater NZ, 97% greater EZ, and 109% greater ROM. However, in terms of angular differences, these percentages corresponded to very small values. Compared to anterolaterally placed plates, pedicle screw fixation provided as much as 0.65° smaller ROM, 0.52° smaller NZ, and 0.39° smaller EZ. Compared to laterally placed plates, pedicle screw fixation provided as much as 0.40° smaller ROM, 0.27° smaller NZ, and 0.27° smaller EZ. Because these absolute differences are very small, whether the superior stability from pedicle screw fixation would translate to a significant difference in clinical fusion rates is debatable. Comparison of Anterolateral and Lateral Plates Stability provided by a plate during different modes of loading depended on its placement. The anterolaterally attached plate resisted flexion and extension better than the laterally attached plate. However, it performed worse in resisting lateral bending. These findings indicate that the plate is an excellent buttress and tension band in resisting motions, but it is less effective in preventing toggling around the axis of the screws.

We did not study different orientations of cages, and it is difficult to speculate how important this factor was relative to the plate. Hypothetically, relative orientations of the cage and plate would not be important in resisting axial rotation. During axial rotation, however, the anterolateral plate provided a slightly smaller ROM, NZ, and EZ and a slightly greater SC than the lateral plate. This finding may indicate that anterolateral placement of the plate works slightly better than the lateral placement of the plate when the cage is inserted anterolaterally, possibly because the normal axis of rotation is farther from the anterolateral plate, giving it a longer lever arm for resisting the twisting moment. Further study of cage orientations relative to plate orientations and identification of the true location of the axis of rotation are needed to verify this hypothesis. Anterolateral Cage In a human cadaveric study,6 an anterolateral cage provided equivalent stabilization to dual anterior cages during flexion and extension but did not stabilize the spine as well as dual anterior cages during extension or axial rotation. It was recommended that additional translaminar screw fixation be added for stabilization in all directions. Our results also indicate that a single anterolateral stand-alone interbody cage inadequately stabilizes the spine during axial rotation. Previous research has shown that supplementary fixation is beneficial when used with dual cages1; supplementary fixation is probably even more important with a single interbody cage. Study Limitations This study compared the biomechanical effects of different supplementary fixation techniques when used with anterolateral threaded interbody cage fixation. This option for interbody cage fixation is only one of many possible options (e.g., single or dual anterior cages, single or dual posterior cages, true lateral cage, tapered or crescent cages, wedges). Our findings should be interpreted with the understanding that they apply only to this particular cage shape and trajectory and that supplementary fixation could affect the biomechanics of different types of interbody cage constructs differently. Although calf spines are a fairly good model for studies of lumbar spine biomechanics, the anatomy at the level studied (L3–L4) differs in some ways from the human lumbar spine.19 As mentioned, the endplates are concave at L3 and convex at L4. Therefore, reaming the disc space without further preparation would result in a poor trajectory that would tend to be aimed toward L3 and away from L4. The best solution to this problem was to shape the endplates using the high-speed drill. However, whether this procedure weakened the bone– cage interface is unknown. Calf bone also is much harder than human bone. A high-speed drill was needed to prepare pedicle screw holes rather than the awl typically used clinically. Again, whether a better trajectory could have been achieved with an awl is unknown. The interface between the immature calf vertebra and endplate was

Anterolateral Lumbar Interbody Cage Stabilization • Bozkus et al 641

weak and separated easily. Whether some separation that would have weakened the cage– bone interface may have occurred is another unknown factor. Another limitation is that the few number of specimens (n ⫽ 6) led in many cases to low power, which means low probability of avoiding false-negative findings (Tables 1 and 2). Therefore, some of the nonsignificant findings shown in Figure 3 and Tables 1 and 2 could reflect inadequate sample size. The final limitation of this research is that the interbody cage was inserted into specimens with no overt spinal instability or deformity. The surgical procedure may therefore be less representative of an actual clinical procedure than if some type of instability existed or had been modeled.20 Conclusion A stand-alone anterolateral threaded lumbar interbody cage inadequately stabilizes the spine during axial rotation, and supplemental spinal instrumentation is needed. Anterolateral or lateral lumbar plating significantly increases stability compared to a stand-alone cage construct, reducing the NZ, EZ, and ROM significantly below their normal ranges and significantly increasing stiffness during all loading modes. Pedicle screws and rods tended to be more rigid than lateral or anterolateral screw plates for enhancing stability. However, during several loading modes, there was no significant difference between plate and pedicle screws and rods. Anterolateral or lateral plate fixation may provide a viable clinical alternative to enhance the stability of an anterior lumbar interbody cage construct. Such plate fixation may obviate the need for combined anterior and posterior approaches to treat spinal instability. Key Points ● Stand-alone anterolateral interbody cage reduced motion to statistically significantly less than normal during all loading modes except axial rotation. ● After adding pedicle screws-rods, anterolateral plate, or lateral plate, significantly less motion was allowed during all loading modes than with a cage alone. ● Pedicle screw-rod fixation performed slightly better than anterolateral plate or lateral plate fixation in enhancing stability provided by a single cage.

● The anterolateral plate slightly outperformed the lateral plate in resisting flexion and extension. ● The lateral plate slightly outperformed the anterolateral plate in resisting lateral bending.

References 1. Cagli S, Crawford NR, Sonntag VK, et al. Biomechanics of grade I degenerative lumbar spondylolisthesis. Part 2: Treatment with threaded interbody cages/dowels and pedicle screws. J Neurosurg 2001;94:51– 60. 2. Heth JA, Hitchon PW, Goel VK, et al. A biomechanical comparison between anterior and transverse interbody fusion cages. Spine 2001;26: E261–E267. 3. Hitchon PW, Goel VK, Rogge T, et al. Spinal stability with anterior or posterior ray threaded fusion cages. J Neurosurg 2000;93:102– 8. 4. Lund T, Oxland TR, Jost B, et al. Interbody cage stabilization in the lumbar spine: Biomechanical evaluation of cage design, posterior instrumentation, and bone density. J Bone Joint Surg Br 1998;80:351–9. 5. Nibu K, Panjabi MM, Oxland T, et al. Multidirectional stabilizing potential of BAK interbody spinal fusion system for anterior surgery. J Spinal Disord 1997;10:357– 62. 6. Nydegger T, Oxland TR, Hoffer Z, et al. Does anterolateral cage insertion enhance immediate stabilization of the functional spinal unit? A biomechanical investigation. Spine 2001;26:2491–7. 7. Rathonyi GC, Oxland TR, Gerich U, et al. The role of supplemental translaminar screws in anterior lumbar interbody fixation: a biomechanical study. Eur Spine J 1998;7:400 –7. 8. Tencer AF, Hampton D, Eddy S. Biomechanical properties of threaded inserts for lumbar interbody spinal fusion. Spine 1995;20:2408 –14. 9. Tsantrizos A, Andreou A, Aebi M, et al. Biomechanical stability of five stand-alone anterior lumbar interbody fusion constructs. Eur Spine J 2000; 9:14 –22. 10. Zhao J, Hai Y, Ordway NR, et al. Posterior lumbar interbody fusion using posterolateral placement of a single cylindrical threaded cage. Spine 2000; 25:425–30. 11. Volkman T, Horton WC, Hutton WC. Transfacet screws with lumbar interbody reconstruction: biomechanical study of motion segment stiffness. J Spinal Disord 1996;9:425–32. 12. Le Huec JC, Liu M, Skalli W, et al. Lumbar lateral interbody cage with plate augmentation: In vitro biomechanical analysis. Eur Spine J 2002;11:130 – 6. 13. Crawford NR, Brantley AG, Dickman CA, et al. An apparatus for applying pure nonconstraining moments to spine segments in vitro. Spine 1995;20: 2097–100. 14. Wilke HJ, Rohlmann A, Neller S, et al. Is it possible to simulate physiologic loading conditions by applying pure moments? A comparison of in vivo and in vitro load components in an internal fixator. Spine 2001;26:636 – 42. 15. Panjabi MM. Biomechanical evaluation of spinal fixation devices: I. A conceptual framework. Spine 1988;13:1129 –34. 16. Crawford NR, Dickman CA. Construction of local vertebral coordinate systems using a digitizing probe. Technical note. Spine 1997;22:559 – 63. 17. Crawford NR, Yamaguchi GT, Dickman CA. A new technique for determining 3-D joint angles: the tilt/twist method. Clin Biomech 1999;14:153– 65. 18. Crawford NR, Peles JD, Dickman CA. The spinal lax zone and neutral zone: measurement techniques and parameter comparisons. J Spinal Disord 1998; 11:416 –29. 19. Wilke HJ, Krischak S, Claes L. Biomechanical comparison of calf and human spines. J Orthop Res 1996;14:500 –3. 20. Crawford NR, Cagli S, Sonntag VK, et al. Biomechanics of grade I degenerative lumbar spondylolisthesis. Part 1: In vitro model. J Neurosurg 2001;94: 45–50.