Transforaminal Lumbar Interbody Fusion

SPINE Volume 29, Number 4, pp E65–E70 ©2004, Lippincott Williams & Wilkins, Inc.

Transforaminal Lumbar Interbody Fusion The Effect of Various Instrumentation Techniques on the Flexibility of the Lumbar Spine Basil M. Harris, MD, PhD,† Alan S. Hilibrand, MD,† Paul E. Savas, MD,† Anthony Pellegrino, BS,* Alexander R. Vaccaro, MD,† Sorin Siegler, PhD,* and Todd J. Albert, MD†

Study Design. In vitro comparison of four reconstruction techniques following transforaminal lumbar interbody fusion in a human cadaveric model. Introduction. Transforaminal lumbar interbody fusion (TLIF) is a relatively new technique that avoids the morbidity of an anterior approach and the nerve root manipulation of a posterior interbody fusion. This study measured the effects of a TLIF on the overall and segmental flexibility of the lumbar spine using four different spinal implant configurations. Summary of Background Data. Anterior lumbar interbody fusion, posterior lumbar interbody fusion, and combined anterior–posterior spinal procedures are gaining wide acceptance for the treatment of selected patients with segmental spinal instability and spondylolisthesis with associated degenerative changes. Each fusion technique may have different effects on the overall flexibility of the lumbar spine. The unilateral TLIF procedure with adjunctive pedicular fixation is one variation of an interbody fusion technique that requires less bony and soft tissue dissection and minimizes nerve root manipulation compared with other interbody fusion methods. Methods. Five fresh-frozen, human lumbar spines were nondestructively subjected to flexion, extension, lateral bending, and axial rotation moments using a previously validated spine flexibility tester, and displacements were measured. Testing the intact lumbar spine was followed by testing of a unilateral L4 –L5 TLIF using a single ramp carbon fiber cage without adjunctive internal fixation. The single carbon fiber (Brantigan) cage was inserted obliquely in a posterolateral to anteromedial position in the L4 –L5 disc space. Following testing of the cage alone, three different adjunctive stabilization techniques were tested. Posterior stabilization involved one of the following: a contralateral translaminar facet screw, single side/ipsilateral nonsegmental pedicle screw fixation, and bilateral nonsegmental pedicle screw fixation. The overall flexibility of each lumbar spine was calculated from load-displacement curves for each axis of rotation. The flexibility of the L4 –L5 segment of each spine was computed from kinematic motion data acquired via attached LED sensors to the L4 and L5 vertebral bodies. Statistical testing was performed with paired t tests.

From the *Biomechanics Laboratory of Drexel University and the †Rothman Institute of Thomas Jefferson University, Philadelphia, Pennsylvania. Acknowledgment date: December 13, 2002. First revision date: April 21, 2003. Acceptance date: August 4, 2003. The device(s)/drug(s) that is/are the subject of this manuscript is/are not FDA-approved for this indication and is/are not commercially available in the United States. No funds were received in support of this work. No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript. Address correspondence to Alan S. Hilibrand, MD, 925 Chestunt Street, 5th Floor, Philadelphia, PA 19107; E-mail: [email protected]

Results. The flexibility of the entire (T12–S1) destabilized spine after TLIF with interbody cage alone and with all three reconstructive techniques was comparable with the intact spine. However, the motion at the L4 –L5 segment was significantly increased for the TLIF with interbody cage alone in axial rotation (299% of intact, P ⬍ 0.01), with no significant change in flexion– extension (79% of intact, P ⫽ 0.22) or lateral bending (87% of intact, P ⫽ 0.39). With the addition of a contralateral translaminar facet screw, the motion at the L4 –L5 segment remained significantly more flexible in axial rotation (250% of intact, P ⫽ 0.06) although less than with the cage alone. With the unilateral pedicle screw construct, the L4 –L5 segment remained more flexible in axial rotation (182% of intact, P ⫽ 0.07) although significantly less than with the facet screw construct (P ⬍ 0.05). The addition of bilateral pedicle screws most closely reapproximated the flexibility of the intact spine with no significant difference in axial rotation (91% of intact, P ⫽ 0.30), flexion– extension (93% of intact, P ⫽ 0.19), or lateral bending (99% of intact, P ⫽ 0.47). The motion at the L4 –L5 segment with bilateral pedicle screws was not significantly different than for the intact specimen in axial rotation (144% of intact, P ⫽ 0.17), flexion– extension (81% of intact, P ⫽ 0.21), or lateral bending (86% of intact, P ⫽ 0.27). Conclusions. TLIF reconstruction with a solitary cage did not increase overall spine flexibility from the intact condition but significantly increased segmental flexibility at L4 –L5 in axial rotation. A unilateral translaminar facet screw had minimal stabilizing effect at L4 –L5. Unilateral pedicle screws further increased stiffness at the L4 –L5 segment. However, TLIF with bilateral pedicle screws most closely approximated the L4 –L5 segmental flexibility of the intact spine. [Key words: lumbar spine, transforaminal lumbar interbody fusion, biomechanics] Spine 2004;29:E65–E70

One method of surgical treatment of symptomatic lumbar disc degeneration is the “transforaminal” lumbar interbody fusion (TLIF), originally described by Blume and Rojas1 and Harms and Rolinger2 in the early 1980s. The TLIF approach was popularized in the late 1990s by Harms and Jeszensky.3,4 The procedure allows stabilization and support of the anterior column of the spine while allowing direct nerve root decompression from a posterior approach. Fundamental to the TLIF procedure is the necessity of performing a complete unilateral facetectomy. Although it is generally accepted that at least 50% of the facet joint complex must be preserved to maintain stability of the motion segment, the degree of instability caused by this particular aspect of the TLIF procedure has not been E65

E66 Spine • Volume 29 • Number 4 • 2004

studied in conjunction with the placement of an interbody spacer (i.e., titanium mesh cages or carbon fiber chamber). It is possible that facet screw fixation or unilateral pedicle screw fixation may provide adequate stability without the need to expose and instrument the contralateral side. As described by Harms et al,4 the TLIF technique includes additional posterior stabilization with bilateral pedicle screw fixation. This study was designed to identify the effect of the TLIF procedure on overall lumbar and individual segmental flexibility in vitro and to compare the stability provided by translaminar facet screws, and unilateral and bilateral pedicle screws on overall and segmental lumbar flexibility in a cadaveric model. We also discuss these findings relative to the normative values of 12.8° axial rotation, 36.8° lateral bending, and 77.0° flexion– extension as reported by White and Panjabi.5 Materials and Methods Specimen Preparation. Five intact lumbar spines were isolated from fresh-frozen human cadavers and were found to be without any history of neoplastic or metabolic abnormality other than age-appropriate osteoporosis. They averaged 81 years in age with a range of 61 to 92 years. The 12th thoracic vertebra as well as the first portion of the sacrum were included in each specimen. The specimens were cleaned of all soft tissue from T12 and S1, taking care to preserve all ligamentous and capsular attachments. The T12 and S1 vertebral bodies were each potted in 4-cm deep, 7.5-cm diameter cylindrical molds with polymethylmethacrylate cement (Dentsply International Inc., York, PA). Two wood screws were placed in the vertebral bodies (T12 and S1 only) and buried in the molds to ensure fixation during testing. Movement at the T12–L1 and L5–S1 joints was preserved during the preparation. The T12 vertebral body in its mold constituted the superior grip in the loading apparatus; likewise, the S1 vertebra in its mold was the inferior grip. Each lumbar vertebral level was instrumented with light emitting diodes (LEDs) for kinematic data acquisition.

Testing Setup. The overall testing setup consisted of a flexibility testing apparatus and a three-dimensional kinematic data acquisition system. The apparatus used to apply loads and measure flexibility was a modification of the 6 df instrumented linkage used to quantify the load-displacement characteristics of the in vivo as well as in vitro cervical spine described by McClure et al.6 This “neck flexibility tester” allows the application of pure moments across the spine while allowing it to move freely in response to these moments. That is, the device, because of its special 6 df structure, did not interfere or restrict in any way the natural motion of the spine. Voltage output from the angular and linear potentiometers and from the force and/or torque sensors are fed into an A-D converter, sampled at 20 Hz, and stored on a computer. The neck flexibility tester was modified for this study by installing a base platform aligned with the axial plane of the testing equipment and a linkage designed to attach to the upper pot of the specimen. This apparatus was used to apply loads to the lumbar spine and record the resulting displacement data of the entire specimen. To measure motion at each spinal segment, a threedimensional optoelectric, kinematic data acquisition system (WATSMART by Northern Digital, Alberta, Canada) was

used. The WATSMART system measured the threedimensional coordinates of small LEDs attached to each vertebral body. To record the motion of the L4 and L5 vertebral bodies, a triad consisting of three LEDs was fixed to each vertebra. The first sacral vertebra was potted into the inferior grip, which remained stationary relative to the testing apparatus and was not instrumented with LEDs. From the relative change in position of each of the three LED markers, the threedimensional motion of the vertebrae, including three rotations and three translations, was computed. Data were recorded synchronously with the load and position sensor data obtained from the loading apparatus at a rate of 20 Hz.

Flexibility Testing Protocol. The testing involved applying a torque about a single axis while both angular and linear displacements were recorded through the position sensors on the loading apparatus and through change of position of the threedimensional markers of the kinematic system. Each test began with a calibration and digitization procedure in which a digitizer was used to identify anatomic sites on each vertebral body and other points of interest. This allowed definition of fixed relations between an anatomic reference frame and a measuring frame defined by the LEDs located on each vertebra. The tests on each specimen were conducted in flexion, extension, lateral bending (right and left), and axial rotation (right and left). The loading apparatus was used to apply a torque of up to 5 N*m in each direction at a rate of 1 N*m per second in both loading and unloading. The loading– unloading was repeated in each direction for three cycles before continuing to the next test (e.g., right lateral bending tests from neutral to ⫹5 N*m and back to neutral three times before proceeding to left lateral bending). Kinematic data were recorded from the third cycle of each test. After motion in all six directions was achieved, the spine was subjected to the next treatment and retested. The TLIF procedure was performed following the technique described by Harms and Jeszensky,4 with the insertion of a single, obliquely oriented carbon fiber cage (DePuy Acromed, Raynham, MA) measuring 9 ⫻ 9 ⫻ 20 mm in place of two titanium mesh cages. The single, obliquely oriented carbon fiber cage technique was chosen because it relies on insertion of a single implant, which can consistently be placed across the midline in the same location within the disc space, as opposed to the placement of two implants with the technique described by Harms et al.3 Preparation of the disc space included removal of all disc material and endplate cartilage while preserving the lateral and anterior anulus and avoiding disruption of the bony endplates. Testing proceeded with the following treatments performed in sequence: 1) intact spine; 2) following the right-sided TLIF exposure with placement of the carbon fiber cage inserted obliquely in a posterolateral to anteromedial position in the L4 –L5 disc space; 3) the carbon fiber cage and a 3.5 mm ⫻ 40 mm L4 –L5 translaminar facet screw (DePuy Acromed); 4) the carbon fiber cage and unilateral (TLIF side) single-segment transpedicular fixation with 6.0 mm ⫻ 45 mm titanium screws connected to a 5.5 mm diameter rod (Moss-Miami, DePuy Acromed); and 5) the carbon fiber cage and bilateral single segment transpedicular fixation with the same system as in treatment 4. Care was taken in placing the translaminar facet screw to avoid disruption of the facet capsule, which might compromise its stability during subsequent tests (treatments 4 and 5).

Transforaminal Lumbar Interbody Fusion • Harris et al E67

Table 1. Lumbar Spine Flexibility by Reconstruction Technique Axial Rotation

Instrumentation Technique Intact spine Overall lumbar L4–L5 Cage alone Overall lumbar L4–L5 Cage and facet screw Overall lumbar L4–L5 Cage and unilateral pedicle screws Overall lumbar L4–L5 Cage and bilateral pedicle screws Overall lumbar L4–L5

Spine Flexibility (% of intact)

vs. Intact (P value)

Lateral Bending

vs. Cage (P value)

100 100

Spine Flexibility (% of intact)

vs. Intact (P value)

Flexion–Extension

vs. Cage (P value)

100 100

103 299

0.45 ⬍0.01

94 250

0.37 0.06

94 182 91 144

0.16 0.39

0.07 0.12

105 85

0.33 0.26

0.36 0.07

⬍0.01 ⬍0.05

104 117

0.30 0.17

⬍0.05 ⬍0.05

99 81

was calculated from load-displacement curves for each construct for each axis of rotation. The flexibility was defined as the range of motion with an applied moment of 5 N*m. The flexibility of the lumbar spine was defined as the movement of the T12 –S1 segment. The flexibility of the L4 –L5 segment of each spine was computed from the kinematic motion data acquired via attached LED sensors to the anterior of the L4 and L5 vertebral bodies. Statistical testing was performed with paired t tests. Statistical significance was established for a P value of less than 0.05.

Results The flexibility data of the lumbar spines following an applied 5 N*m moment in axial rotation, lateral bending, and flexion– extension is shown in Table 1. For the intact lumbar spines, the average spine yielded a total of 24.5° in axial rotation, 29.1° in lateral bending, and 31.7° in flexion– extension. Following performance of a TLIF with resection of the right L4 –L5 facet joint complex and carbon fiber cage placement, overall lumbar flexibility was similar, with 25.1° in axial rotation (P ⫽ 0.45), 32.5° in lateral bending (P ⫽ 0.16), and 32.1° in flexion– extension (P ⫽ 0.43). With the addition of a translaminar facet screw, similar overall lumbar flexibility was measured; 23.1° in axial rotation (P ⫽ 0.37), 30.5° in lateral bending (P ⫽ 0.33), and 28.8° in flexion– extension (P ⫽ 0.13). Stabilization of the TLIF construct using a unilateral pedicle screw construct provided 23.1° in axial rotation (P ⫽ 0.36), 30.1° in lateral bending (P ⫽ 0.38), and 28.2° in flexion– extension (P ⬍ 0.05). Bilateral pedicle screw stabilization allowed the least lumbar motion, with 22.4° of axial rotation (P ⫽ 0.30), 28.9° of lateral bending (P ⫽ 0.47), and 29.4° of flexion– extension (P ⫽ 0.19). The overall lumbar spine flexibility data are shown graphically in Figure 1a. The average L4 –L5 segmental flexibility for the intact spines was 3.0° in axial rotation, 4.1° in lateral bending, and 8.2° in flexion– extension. Following destabilization

vs. Intact (P value)

vs. Cage (P value)

100 100

112 87

Data Analysis. The overall flexibility of each lumbar spine

Spine Flexibility (% of intact)

102 79

0.43 0.22

⬍0.01 0.47

90 73

0.13 0.23

0.07 0.33

0.38 0.25

⬍0.01 0.20

89 51

⬍0.05 0.20

⬍0.05 0.23

0.47 0.21

⬍0.05 0.45

93 86

0.19 0.27

0.07 0.37

of the right L4 –L5 facet joint due to the TLIF exposure and placement of the carbon fiber cage, this increased to 8.9° in axial rotation (299% of intact, P ⬍ 0.01), but did not change significantly for lateral bending (3.5°, 87% of intact, P ⫽ 0.39) or flexion– extension (6.5°, 79% of intact, P ⫽ 0.22). With the addition of a translaminar facet screw, the segmental flexibility of the L4 –L5 (TLIF) level remained significantly increased above the intact condition in axial rotation (7.4°, 250% of intact, P ⫽ 0.06), but was not significantly changed in lateral bending (3.4°, 85% of intact, P ⫽ 0.26), nor in flexion– extension (6.0°, 73% of intact, P ⫽ 0.23). The L4 –L5 segmental flexibility with a unilateral pedicle screw construct measured 5.4° in axial rotation (182% of intact, P ⫽ 0.07), although this was significantly more stable than with the translaminar facet screw (P ⬍ 0.05). L4 –L5 segmental motion with the unilateral pedicle screw construct was 4.8° in lateral bending (117% of intact, P ⫽ 0.25), and 4.2° in flexion– extension (51% of intact, P ⫽ 0.20). The L4 –L5 segmental flexibility with a bilateral pedicle screw construct was 4.3° in axial rotation (144% of intact, P ⫽ 0.17), 3.3° in lateral bending (86% of intact, P ⫽ 0.21), and 7.0° in flexion– extension (81% of intact, P ⫽ 0.27). The L4 –L5 segment flexibility data are shown in Figure 1b. Discussion The goals of a segmental spinal fusion are to provide stability to a spinal segment in order to allow interbody space graft consolidation and prevent cage displacement. Ideally, the operated segment should have a flexibility that approaches the intact specimen. If the reconstruction is too flexible, loosening of the implants and cage displacement may occur. If the construct is too stiff (i.e., not flexible enough), this might cause transfer of loads to adjacent levels and accelerate degeneration at these levels requiring further surgery.

E68 Spine • Volume 29 • Number 4 • 2004

Figure 1. a: Overall lumbar spine flexibility with applied moment of 5 N*m as percent of intact spine. b: L4 –L5 segment flexibility with applied moment of 5 N*m as percent of intact spine.

In the present study, we assessed the effect of TLIF reconstructive techniques on overall and segmental flexibility. There was little change in the overall flexibility of the entire specimen, T12–S1, for all motions (flexion– extension, axial rotation, or lateral bending). However, segmental L4 –L5 flexibility was much greater than the intact specimen following unilateral facet resection for the TLIF approach for axial rotation but was not significantly changed for flexion– extension or lateral bending. These data suggest that sacrifice of a single facet joint in the TLIF approach has a greater impact on axial rotation at the operated level than on flexion– extension or lateral bending. This may be the result of the geometry of the interbody cage distracting the disc space and limiting “bending” in the sagittal and coronal planes. Dual titanium mesh cages might have provided more stability in

torsion given their sharper spikes, although they might be expected to have a smaller area of contact with the endplates than the carbon fiber cage. The lack of a significant change in overall flexibility may be explained by accommodation of increased motion at L4 –L5 by decreased motion at adjacent levels. A drawback of this study is that the segmental data are not available for the adjacent levels. The role of different stabilization constructs was assessed. Translaminar facet screw and unilateral pedicle screws decreased but did not reapproximate native L4 –L5 segmental axial rotation. In addition, there were no significant changes in lateral bending or flexion– extension, which were already approximately the same as that of the intact specimen. Bilateral pedicle screws reapproximated axial rotation most closely to the intact

Transforaminal Lumbar Interbody Fusion • Harris et al E69

specimen (although still greater than the intact specimen). The overall flexibility was within 10% of intact in all motions, and the L4 –L5 segmental flexibility in lateral bending and flexion– extension was reduced by 10% to 20% versus intact. In a recent study by Osman et al7 on cadaveric, twovertebrae, functional spinal units, no flexibility change was found after transforaminal decompression, whereas significant increase in extension and axial rotation flexibility was noted after posterior decompression. The results, however, are not directly comparable to the present study, in which we tested after a transforaminal decompression and with the carbon fiber cage in place. In our tests, there was a significant increase in axial rotation flexibility about the L4 –L5 segment, while there was no significant flexibility change in lateral bending or flexion– extension. White and Panjabi5 compiled in vitro and in vivo data from their own and other investigations5,8 –10 to arrive at representative ranges of rotation for each interspace of the spine. The data from the current study compares favorably with the in vitro data compiled by White and Panjabi.5 They reported in vitro ranges of rotation of 12.8° (with a limit of 6.4° to 22.9°) in axial rotation, 36.8° (with a limit of 25.0° to 50.1°) in lateral bending, and 77.0° (with a limit of 44.0° to 105.0°) in flexion– extension.5,8 The in vivo data5,9,10 suggest ranges of rotation of 7.5° (with a limit of ⫺2.0° to 16.0°) in axial rotation, 27.5° (with a limit of 18.0° to 50.0°) in lateral bending, and 73.5° (with a limit of 15.0° to 130.0°) in flexion– extension. In the present study, the intact lumbar spines on average yielded 24.5° in axial rotation, 29.1° in lateral bending, and 31.7° in flexion– extension. When compared with the data reported by White and Panjabi,5 the spines included in this study on average yielded more in axial rotation, were comparable in lateral bending, and yielded less in flexion– extension. In this study, the average L4 –L5 segmental flexibility for the intact spines was 3.0° in axial rotation, 4.1° in lateral bending, and 8.2° in flexion– extension. These compare favorably with reported representative values (and ranges) for the L4 –L5 segment of 2° (with a limit of 1° to 3°) in axial rotation, 6° (with a limit of 3° to 9°) in lateral bending, and 16° (with a limit of 9° to 21°) in flexion– extension.5 However, it must be noted that comparisons between in vitro studies is complicated by variables such as differences in specimen preparation (e.g., specimen handling, conservative vs. radical soft tissue dissection), testing (e.g., testing apparatus, applied moments, load rate, and preload conditions), and data collection. Limitations of our study design include the inherent drawbacks of in vitro cadaver testing, which cannot truly reproduce in vivo behavior without the stabilizing effects of active muscles. For example, this study found that, even with bilateral transpedicular instrumentation, the flexibility of the L4 –L5 segment in axial rotation only approximated the intact condition. Although the unilat-

eral TLIF approach would be expected to increase segmental flexibility in axial rotation, segmental fixation with pedicle screws or even a translaminar facet screw might have decreased flexibility further, if the specimens had excellent bone quality. In addition, the elderly specimens would be expected to have more degenerated discs with more osteophyte formation, making the specimens much less flexible than the spines of younger patients in their 30s and 40s who typically undergo this procedure. Since the overall flexibility (range of motion) should be greater among these younger spines, the ability of spinal instrumentation to limit this flexibility in a younger spine is probably greater than actually measured in this study. Also, the reported testing had only limited preload, equivalent to the weight of the pot and overhead apparatus, weighing approximately 5 kg. In addition, we were unable to analyze for changes at adjacent segments secondary to instrumentation at the TLIF level because of the lack of segmental data for other motion segments. Conclusion The TLIF approach had no significant effect on the overall flexibility of the lumbar spine, although axial rotation was significantly affected at the L4 –L5 (TLIF) segment. Axial rotation flexibility at L4 –L5 was reduced closest to the intact specimen with bilateral pedicle screws, although the bilateral pedicle screw construct did not cause a significant decrease in flexibility at the L4 –L5 segment for flexion– extension and lateral bending. These findings suggest that bilateral pedicle screw fixation is warranted when a Brantigan cage is inserted obliquely into the L4 –L5 interspace through the unilateral TLIF approach. Key Points ● The TLIF approach had no significant effect on the overall flexibility of the lumbar spine, although axial rotation was significantly increased at the L4 –L5 (TLIF) segment. ● A unilateral translaminar facet screw had minimal stabilizing effect at L4 –L5. Unilateral pedicle screws further increased stiffness at the L4 –L5 segment. However, TLIF with bilateral pedicle screws most closely approximated the L4 –L5 segmental flexibility of the intact spine. ● Axial rotation flexibility at L4 –L5 was reduced closest to the intact specimen with bilateral pedicle screws, although the bilateral pedicle screw construct did not cause a significant decrease in flexibility at the L4 –L5 segment for flexion– extension and lateral bending. ● These findings support the recommendation of Harms et al for bilateral pedicle screw fixation as a standard technique for reconstruction following unilateral TLIF decompression and interbody grafting.

E70 Spine • Volume 29 • Number 4 • 2004

References 1. Blume HG, Rojas CH. Unilateral lumbar interbody fusion (posterior approach) utilizing dowel graft. J Neurol Orthop Surg. 1981;2:171–175. 2. Harms J, Rolinger H. Die operative Behandlung der Spondylolisthese durch dorsale Aufrichtung und ventrale Verblockung. Z Orthop Ihre Grenzgeb. 1982;120:342–347. 3. Harms J, Jeszensky D, Stolze D, et. al. True spondylolisthesis reduction and more segmental fusion. In: Bridwell KH, DeWald RL, eds. The Textbook of Spinal Surgery, 2nd ed. Philadelphia: Lippincott-Raven, 1997:1337–1347. 4. Harms J, Jeszensky D. The unilateral transforaminal approach for posterior lumbar interbody fusion. Orthop Traumatol. 1998;6:88 –99. 5. White AA III, Panjabi MM. Clinical Biomechanics of the Spine, 2nd ed. Philadelphia: Lippincott, 1990:102–111.

6. McClure P, Siegler S, Nobilini R. Three-dimensional flexibility characteristics of the human cervical spine in vivo. Spine. 1998;23:216-223. 7. Osman SG, Nibu K, Panjabi MM, et al. Transforaminal and posterior decompressions of the lumbar spine: a comparative study of stability and intervertebral foramen area. Spine. 1997;22:1690 –1695. 8. Yamamoto I, Panjabi MM, Crisco J, et al. Three-dimensional movements of the whole lumbar spine and lumbosacral joint. Trans Int Soc Study of Lumbar Spine, Kyoto, Japan, 1989. 9. Hayes MA, Howard TC, Gruel CR, et al. Roentgenographic evaluation of lumbar spine flexion-extension in asymptomatic individuals. Spine. 1989;14: 327–331. 10. Pearcy M, Portek I, Shepherd J. Three-dimensional x-ray analysis of normal movement in the lumbar spine. Spine. 1984;9:294 –297.