No

I PP 79-84

1986

MECHANICAL PROPERTIES OF LUMBAR SPINE MOTION SEGMENTS UNDER LARGE LOADS

J. A. A. MILLER*, A. B. SCHULTZ*, D. N.
Biomtzhanics Research Laboratories and WARWICK and D. L. SPEK;CER~

Department of Orthopaedic Chicago

Surgery. University of Illinois a~

Abstract-The mechanical behavior of fourteen fresh human lumbar motion scgmenrs taken at autop\y from males with an average age of 29 yr was studied. Forces up IO 1029 N wrc applied in anterior. posterior and lateral shear;and moments up to 95 Nm were applied in flexion,extension. lateral bending and torsion. In response to these loads endplate displacements up to 9 mm and rotations up lo 111 were measured. Stitfncss values ranged from 53 to 140 N mm- in response to the shear forces and 6-l I Nm degree- in response to the moments. Lumbar motion segments can develop significant passive resistances to loads in situarions where they are allowed to undergo substantial deformations.

INTRODUCTION Lumbar large spine motion segments frequently must resist

iMATERIALS AND METHODS

loads in axial compression. Both analytical (e.g. Schultz et al.. 1983) and experimental (e.g. Nachemson and Elfstrom, 1970) studies show that compression forces of up to eleven times the superincumbent body weight can be imposed on the lumbar spine by daily activities. The mechanical response of cadaver lumbar spine motion segments to compression forces of these magnitudes has been well established (for example, Panjabi er al., 1977; Berkson ct al., 1979). The analytical studies and their accompanying validation experiments indicate that routine daily activities seldom impose large loads on the spine in shear, bending or torsion. In bending and torsion in particular it usually falls to the trunk muscles rather than the motion segments to balance moments. This occurs because few physical activities require lumbar motion segments to flex, extend, bend laterally or twist more than a few degrees. Few physical activities involve significant motions in shear. In response to only small motions, the motion segments can develop only small moment and shear resistances (Berkson ef al., 1979). Situations which call for large lumbar spine motion segment resistances to shear, bending or torsion may occur. Large resistances may develop in tasks involving large trunk flexions, lateral bends or twists; during traumatic events such as a fall; or when trunk muscle contractions are recruited inappropriately in unfamiliar tasks. Few data are presently available to describe the resistances of lumbar spine segments to large shear, bending, or torsional loads, and how those resistances relate to segment motions in shear, bending or torsion. The purpose of this study was to collect such data.
Rrceitrd 19 April 1954; in rrcisrd form 19 Junr 1985. *Current address: Department of Mechanical Engineering and Applied Mechanics, University of Michigan, Ann Arbor, MI48109-2125, U.S.A. 79

Spine sections from TIZ-LS were excised from nine male fresh cadavers of ages 1841 yr, with a mean age of 28.7 yr (Table 1). Deaths were due to trauma to body parts other than the spine. No radiographic abnormalities were found in any of the spine sections. The sections were sealed and stored at -20C until just before testing. They were divided into fourteen motion segments, leaving all intervertebral ligaments and joints intact. The anteroposterior and lateral disc diameters of each segment were measured with dial calipers. Similarly, the disc and vertebral body heights were each found by averaging anterior and posterior margin dimensions. Mean lateral and anteroposterior disc diameters were 52 and 38 mm, respectively. The segments were assigned into test groups so that each group had three segments of roughly comparable dimensions. To enable tight gripping of the segments in the test apparatus at the intended loads, much of the cancellous bone of each vertebral body centrum was removed and replaced with polymethylmethacrylate cement. Each vertebral body was then gripped radially by forcing curved serrated metal pieces against it with a hoseclamp (Fig. I). Thisconstruct in turn wasclamped to test apparatus fixture plates by passing multistrand steel wires over each pedicle and through the spinous process, and drawing those wires tight. The end twothirds of each vertebra so prepared, with the gripping pieces, hose clamp and wires was then cast into the fixtures using more acrylic cement, so as to obtain a nearly rigid fixation. Care was taken not to disturb motion segment tissues outside of the regions used for gripping. The methods and apparatus used for applying loads and measuring motions were similar to thosedescribed by Schultz et al. (1979).In these tests static shears and moments were applied in increments to the top vertebra of the motion segment by way of suitably arranged cables and pneumatic cylinders, while the

80

J. A. A. MILLER. A. B. SCHULTZ, D. N.

WARWICK and D. L. SPENCER

on each segment

Table

1. Segment data and tests conducted Frontal

Specimen/ spine No. I:1 2, 10 3.6 4, IO 5;7 6, I s:*; 7 9/6 IO,2 I l/5 I?,7 l3:8 I4;9 Mean (S.D.) Range Failure

Level L3-4 L4-5 L4-5 L?-3 Ll-2 Ll-2 Ll-2 L-3 L2-3 L4-5 L3-4 L4-5 Ll-2 L2-3

Age (yr) 41 28 37 28 30 41 27 I8 37 27 27 30 IS 31 29(7) IS-41

disc diameter (mm) 53 59 51 52 51 49 44 56 53 47 56 46 47 50 52(4) 44-59

Sagittal disc diameter (mm) 3: 37 40 42 40 39 36 35 37 3s 37 39 36 33 38(?) 33-42

Test directions Anterior shear Antrrior, posterior shear Anterior, posterior shear Posterior. lateral shear Lateral shear, torsion Lateral shear Flexion. extension* lateral bending* Flexion. extension Extension Lateral bending Lateral bending Torsion* Torsion*

occured at maximum

load.

Fig. I. Method of mounting specimens. Rigid fixation to the mounting plates is enhanced by the six tensioned steel cables passing over and through the posterior elements. Test loads are applied to the upper vertebra.

the superior disc end plate were calculated lo the nearest mm or half degree. Since we did not measure full three-dimensional motions, the three dial gages were rearranged for each test so as to be able to record the three main motions for that test (see Table 2 for motions measured). Stiffness in a given mode of loading was defined as the load applied divided by the translation or rotation of the center (0) of the superior vertebral body or that bodys rotation in that direction at that load. Note that stiffness by this definition is equal lo the inverse of the flexibility coe5cient in matrix notation; but it is not equal lo the matrix stiffness coefficient.

RESULTS

was held rigidly. Loads of up to 1029 N in shear and up to 95 Nm in bending were applied through a point which lay a mean distance of 54.5 mm above the disc center. Load increments were 50 N in shear and 5 Nm in torsion and bending tests. Each segment from its group of three was loaded in one of the seven test directions: anterior, posterior or lateral shear, or flexion, extension, right lateral bending or torsion counter clockwise when viewed from above. Loads were applied in this direction until either overt rupture occurred, or the maximum load capability of the test apparatus was reached. Seven of the fourteen segments showed no overt signs of tissue failure when this was done. These seven segments were then loaded in a second test direction (Table I). Displacements were measured 15 s after each load was applied using three dial gages. From these measurements and a knowledge of vertebral geometry the three main motions (for example, anterior and superior translations and 5exion rotation in response to an anteroposterior shear force) of the superior vertebral body center (0, Fig. 1)and of the center (S) of
lower vertebra

Overall specimen geometry is reported in Table 1. The mean (S.D.) disc height measured as the average of the anterior and posterior heights was 12.1 (2.2) mm. The mean distance between the center (0) of the superior vertebra and the center (S) of the disc superior endplate was 13.8 (1.3) mm. The motions measured in response to the loads applied are reported in two forms; in terms of equivalent motions occurring at the center of the superior vertebral body (Table 2), and in terms of equivalent motions occurring at the center of the superior endplate of the intervertebral disc for a common nominal load. In either set of terms, the motions varied considerably. For example, in response to anterior shear forces of 980 N, anterior motions of the vertebral body center (0) ranged from 8.3 to 13.2 mm, while the corresponding motions at the center (S) of the superior disc endplate ranged from 3.6 to 9.2 mm. In general the transverse motions at S in the anterior, posterior and lateral shear tests averaged 59.71 and 85 y0 respectively of those at 0. For the moment tests the translational

Mechanical properties of lumbar spine

81

Table 1. Msltn (SD.)

and range of the superior vertebral Displacements m

body center (0)

motions for maximum Coupled motions

nominal

test loads

Test direction

(Load)

loading direction Rotation 17.0 (1.7) Is.@-l8.8~5 7.8 (0.8) 7.M.6jl 10.6 (1.5) 9.&l l.SC Forwards 4.9 (0.3) 4.7-5.3 mm -4.1 (1.0) -3.3--5.3 5.7 (1.7) 9.6 (0.4) 3.5-6.9 9.2-lO.l$ Downward 0.S (0.1) 0.8-1.0 mm 1.1 (0.4) 0.7-1.5 2.5 (OS) 1.7-3.1 Downward I.5 (1.0) 0.8-2.7 mm 1.9 (0.9) 0.9-2.7 2.0 (0.5) 3.0 (0.3) I.>2.6 ?.S-3.4t

Antrrtor (980 Kl Posterior (9SO N) Left lateral (490 N) Flcxion (70 Nm) Extension (70 Nm) Right lateral bending (60 Nm) CCW torsion (70Nm) Right lateral translation. tAnterior translation. $Left translation. SFlcxion rotation. /IExtension rotation. (hteral rotdtion.

10.1 (2.7) 7.0 (1.4) 9.3 (1.3) 12.7 (1.0) 9. (0.4) 13.8 (1.4) 6.4 (0.4)

8.3- 13.2 mm 5.3-7.9 5.5-10.7 11.7-13.8 8.7-9.5 12.&15.4 6.G6.8

those at 0 in the flexion, extension and lateral bending tests. In torsion the translations were essentially identical at both points. Load-deformation response in shear was quite linear (Fig. 2). Large coupled motions occurred (Table 2), particularly in response to the shear forces. There, rotations of up to 18 occurred in flexion coupled with compressions of up to 3 mm. This presumably resulted because the shear forces were applied on average 5 cm superior to the disc center, SO that they were accompanied by substantial bending moments. Load-deformation response in bending was non-linear. No specimen showed any overt signs of failure at loads below 539 N in shear or below 59 Nm in bending or torsion. No specimens failed in response to loads
1000

motions of S averaged 39,49 and 49 yOof

iti 2 5 E v) 400 B ._ h s Q 200 600

applied in anterior shear, posterior shear or Bexion (Table 1). One of the three specimens tested in each direction failed in lateral shear and in extension, while two of three tested failed in lateral bending. In each case, failure was by separation of the superior endplate from the vertebral body. All three specimens tested in torsion failed due to avul$ion of the apophyseal joints. The smallest shear stiffness values were found in lateral shear and the largest in posterior shear; the smallest moment stiffness values were found in lateral bending and the largest in torsion (Table 3). The motion segment stiffnesses found here in response to large loads were compared to those found, using the same methods and test apparatus, in response to small loads. In general, the large-load shear stiffness values were similar to small-load values (Berkson er al., 1979). reflecting the Iinearity of the loaddisplacement curves observed in the present tests. In contrast, the large-load moment stiffness values were significantly larger than small-load values (Schultz et al., 1979). They were 1.5 to 6 times as large, reflecting the non-linear moment-rotation relationships observed in the present tests.
DISCUSSIOS

0 -12 -8 -4 0 4 0 12 Displacement (degrees or mm) Fig. 2. An example of motion segment response in posterior shear. The posterior and downward displacements of the vertebral body center are represented by the solid triangle and square symbols; similarly those of the superior disc endplate center are represented by the diamonds and plus signs, while rotation in extension is represented by the open circles.

The purpose of this investigation was to obtain the load-displacement behavior of young adult lumbar motion segments under large shear and moment loads. The results show that the motion segment soft tissues alone can passively resist loads of over 500 N in shear and 55 Nm in bending and torsion. To place these numbers in perspective, if an adult with a body mass of 75 kg could lie horizontally with neither external support superior to L5 nor the internal support of his trunk muscles, his L5 motion segment would experience loads of about 400 N in shear and 100 Nm in bending. Thus the motion segments could supply all of the shear resistance, but only about half of the bending

82

J. A. A. MILLER.

A. 8. SCHULTZ, D. N. WARWICK and D. L. SPENCER

Table 3. A comparison of mean stiffnessin the direction of the load, at small and large loads. Stiffnesses were calculated by dividing the motion of the vertebral body center by the load Small load response Stiffness (N mm-) Berkson* Lint Liu$ 145 143 128 294 400 384 643 Large load response
Panjabit Load IN) StitTncss (N mm- )

Test direction Anterior shear Posterior shear Right lateral shear

I25 250

980 N 980 N 5OON Load (Nm) 68.6 68.6 58.8 68.6 Nm Nm Nm Nm

97 I40 53 Stiffness (Nm degree- ) 5.5 7.6 4.4 10.9

Stiffness (Nmdegree-) Schultz11 Markolf![ Panjabi** Flexion Extension Right lateral bending CCW torsion 0.9 2.2 1.1 6.8 2.7 3.6 2.0 7.7 I.7 2.3 18.8

l Berkson et ul. (1979) data. at 86 N

force and 400 N compression.

tLin ef al. (1978) data, at IO0 N and 440 N compression. $Liu rr al. (1975) data, initial stiffness, no compressive preload. $Panjabi et al. (1984) data, at I50 N. no compressive load. l[Schulta et al. (1979) data, at 4.7 Nm moment and 400 N compression. YMarkolf (1971) data, at 5.4 Nm and no compressive load. l *Panjabi rt al. (1984) data, at 7.5 Nm and no compressive preload.

resistance stretching.

needed

to

do

this,

through

passive

Computer models have recently been used to predict muscle contraction forces and spine loading in performance of static tasks (for example, Schultz et ol., 1983). One of the assumptions of these models is that lumbar motion segments can resist substantial compression and shear loads, but that their bending resistance, in comparison to the moments which can be devefoped by the trunk musculature, is negligibIe. The present results confirm those assumptions in general for shear, and for rotations of a few degrees in bending and torsion, For rotations of about lo or more, however, the present results show that the motion segments can resist passively moments between 30 and 70% of those which the trunk muscles can develop voluntarily (McNeil1 et al., 1980). So, in positions involving large amounts of flexion, extension, lateral bending or torsion, the tissues of the spine may passively resist considerable moments. The average maximum flexion of a lumbar motion segment in F~FOis IS (Adams and Hutton, 1982; Pearcy et al, 1984). The present results suggest that at such large values of flexion, the lumbar motion segments may provide substantial bending resistance. This is also suggested by the phenomenon of back muscle relaxation in full trunk flexion (see, for example, Schultz et al., in press). Similarly maximal axial rotation between lumbar vertebrae has been reported to be 6 (Pearcy and Tibrewal, 1984). so the large torsional resistances found in the present studyarealso probably developed in I;ico. In extension and lateral bending, however, the maximum intervertebral tilt in the lumbar spine has been reported to rarely exceed 5 in t&o (Bakke, 1931; Pearcy and Tibrewal, 1984). So the largest extension and lateral bending resistances

reported here may seldom be developed in ho. When large shear or moment loads on the spine do occur in niuo. they usually occur in conjunction with large compressive loads. Schultz et al. (in press) have predicted, for example, that in maximal forward flexion spine compression can range up to 4500 N while anteroposterior shear forces can range up to 350 N. How would such compressive loads affect the present results, which were obtained without a compressive preload? Preload has been reported to decrease motion segment bending stiffness depending on the magnitude and direction of the preload (Panjabi et al., 1977; Miller et al., 1982). On the other hand, in anteroposterior or lateral shear, a 200 N preload stiffened the lumbar motion segment by 25%; although adding a further 200 N to preload increased the stiffness only by an additional 5% (Lin et al., 1978). In axial torsion a lOOON preload has been shown to stiffen lumbar segments by about 50% (Panjabi et ul., 1977). So, a compressive preload can be expected to increase motion segment resistance to shear or torsional loading, but decrease resistance to bending moments when compared to the present, no-compression-preload values. In the present study the shear test forces were applied, on average, 35mm above 0. In a 980 N test this would yield a moment of 34 Nm about 0. Thus our shear test was in fact a combined shear and moment test. This type of loading can occur in ho when a transverse force is applied to the trunk superior to the spine level in question and the trunk muscles are quiescent; for instance in the forward flexed position (Schultz er ol., in press). This associated moment should be considered when comparing the present shear results with those of others. In a shear test the

Mechanical properties of lumbar spine translations at 0 depend on the amount of vertebral rotation, which depends in part on the material properties and in part on the load point. Since Berkson er al. (1979) and Schultz et al. (1979) used the same test apparatus and loading point, our results are comparable to theirs. Table 3 shows a comparison with their results, together with other published values. in comparing the shear data with those of Liu er al. (1975) and Lin er al. (1978), it should be noted that their data were obtained using pure shear applied at the level of the disc and with the resulting motions constrained to be translational only. It seems clear that while the present shear results are valid for the combined loading that was used, they yield lower-bound stiffness values due to the increased translations that resulted from the vertebral rotations. These considerations do not apply to the moment and torsion results. Since we calculated vertebral motions by measuring the motions of the base mounting plates, it was important to obtain rigid fixation of the specimens. Specimens were inspected after application of each load increment for signs of slippage or failure. Usually slippage was accompanied by an audible sound and a large increase in displacement. We excluded from our results the response of any specimen in which slippage was suspected. Due to the paucity of specimens in the age range studied here, only three specimens were tested in each direction. Large variation in the mechanical properties of motion segments had been noted earlier (Nachemson er al., 1979) and was observed here also. Our results nevertheless are probably reliable, at least as order-of-magnitude estimates for mechanical properties at large loads. The shear results that we obtained are similar (Table 3) to earlier results obtained using the same test procedures (Schultz et al., 1979). Our larger moment stiffnesses reflect the nonlinearities we observed in the load-displacement relations, so they seem consistent with earlier results. Considerable translations of the superior endplate center were observed in shear and bending. These motions are of interest in assessing available foraminal
space for the lumbosacral nerve roots (Reuber et ol., 1982; Panjabi et al., 1983). Few data are available on motion segment failure loads. To our knowledge only Farfan (1973) tested segments to failure in torsion. Our results agree with those of his much larger sample. Adams and Hutton (1982) tested motion segments under large loads in

83 CONCLUSlOSS

Nm. The results show that (I) These motion segments werecapable of resisting large loads: over 539 N in shear and over 59 Nm in bending without failure. (2) End-plate translations of up to 9 mm occurred in shear and moment tests. Rotations of up to 18 without failures were recorded in response to 980 N of anterior shear with its accompanying bending moment. (3) Large load stiffnesses were similar to low load stiffnesses in shear tests, but were generally two to six times larger in bending and torsion tests. Ac~nowledgemettr-We gratefully acknowledge the technical assistance of John Dulf and Karol Haderspeck-Grib, and the support of PHS Grants NS 20536 and OH 01962.

Static load-deformation tests of lumbar motion segments from male mean age of 29 yr were performed, bending loads of up to 1029 N and 95

fourteen intact cadavers with a with shear and

REFERENCES
Adams, M. A. and Hutton W. C. (1982) Prolapsed intervertebral disc: a hyperflexion injury. Spine 7, 184-191. Bakke. S. (1931) Roentgenologische Beobachtungen uber die Bewegungen der Wirbelsaule. Acta Radial. Suppl. 13. Berkson, M. H., Nachemson, A. and Schultz, A. B. (1979) Mechanical properties of human lumbar spine motion segments-Part II: Responses in compression and shear; influence of gross morphology. J. biomech. Engng 101, 53-57. Farfan, H. (1973) Mechanical Disorders ofthe Low Back. Lea and Febiger. Philadelphia. Lin, H. S.. Liu, Y. K. and Adams, K. H. (1978) Mechanical response of the lumbar intervertebral joint under physiological complex loading. J. Bone Jr Surg. 6OA, 41-55. Liu, Y. K., Ray, G. and Hirsch, C. (1975) The resistance of the lumbar spine to direct shear. C/in. Orthop. N. Am. 6,33-49. Markolf, K. L. (1971) Deformation of the thoracolumbar intervertebral joints in response to external loads. A biomechanical study using autopsy material. J Bone Jr
Surg.

54-A. 51I-533.

combined compression and flexion, but did not specify the magnitude of the applied flexion moments. It was noteworthy that no specimens failed at loads below 750 N or 75 Nm in anterior or posterior shear or in flexion tests; the motion segments failed most easily in torsion (two out of three) and lateral bending (one out of three). Specimen mounting is critical under high loads, and it is possible that our fixation method influenced both the failure mode and load level achieved. Our data provide at least lower bounds to specimen failure loads.

McNeil1 T., Warwick, D., Andersson, G. and Schultz, A. (1980)Trunk strengths in attempted flexion, extension and lateral bending in healthy subjects and patients with low back disorders. Spine 5, 529-538. Miller, J. A. A., Skogland, L. B. and Ludvigsen. P. L. (1982) On the load-displacement behavior of adolescent spinal motion segments. On the importance of growth in idiopathic scoliosis. Ph.D. Dissertation, University of Oslo. Nachemson, A. and Elfstrom, G. (1970) Intravital dynamic pressure measurements in lumbar discs. Scand. J. Rebab.
Med. Suppl. 1. Nachemson, A., Schullz, A. and Berkson, M. (1979)

Mechanical properties of human lumbar spine motion segments, Part III, Influences of age, sex, disc level and degeneration. Spine 4. I-8. Panjabi, M. M., Krag M. H.. White, A. A. and Southwick, W. D. (1977) Effects of preload on load-displacement curves of the lumbar spine. Orthop. Clin. N. Am. 8, 181-192. Panjabi, M. M.. Takata, K. and Gael, V. K. (1983) Kinematics of lumbar intervertebral foramen. Spine 8, 348-357. Panjabi, M. M., Krag, M. H. and Chung, T. Q. (1984) EIfec~s

84

J. A. A. MILLER,A. B. SCHULTZ, D. N. WARWICK and D. L.

SPENCER

of disc injury on mechanical behavior of the human spine. 9. 707-713. Pearcy, M., Portek, I. and Shepherd, J. (1984) Threedimensional X-ray analysis of normal movement in the lumbar spine. Spine 9, 294-300. Pearcy, M. and Tibrewal, S. (1984) Axial rotation and lateral bending in the normal lumbar spine measured by threedimensional radiography. Spine 9, 582-587. Reuber, M., Schultz, A., Denis, F. and Spencer, D. (1982) Bulging of lumbar intervertebral disks. J. biomech. Engng
Spine 104, 187-192.

Schultz A., Warwick, D., Berkson. M. and Nachemson, A. (1979) Mechanical properties of human lumbar spine motion segments, Part 1, Responses in flexion, extension, lateral bending, and torsion. 1. biomech. Engng 101.4652. Schultz, A., Haderspeck, K.. Warwick, D. and Portillo, D. (1983) Use of lumbar trunk muscles in isometric performance of mechanically-complex standing tasks. J. Or&p. RKS. 1, 77-91. Schultz A., Haderspeck-Grib, K., Sinkora, G. and Warick, D. (in press) Quantitative studies of the flexion-relaxation phenomenon in the back muscles. J. Ouhop. RKS.