Mechanical function of the human lumbar interspinous and supraspinous ligaments

R.J. Hindle, M.J. Pearcy and A. Cross*

Bioengineering Research Group, University of Durham, Durham City, UK and *Department of Orthopaedic Surgery, Sunderland General Hospital, Sunderland, UK
Received May 1989, accepted November 1989

ABSTRACT iF3.e mechanical finction of the interspinous and supraspinous ligaments has been examined by simulating, on excised specimens, the o!eformation caused duringforwardflexion of the spine in real life. The load extension curves showed that during the$rst half offlexion the ligaments camed very little load but towards the end of the range ofjexion they resisted up to 734 N. When the supraspinous ligament was sectioned the interspinous ligament alone resisted 75% of this load. Testing at high strain rates showed an increase in load-carrying capacity of up to 30%. The maximum extension moment that can be produced by these ligaments was calculated to be approximately 7Nm, or 5% of the moment the back muscles can produce across any intervertebraljoint. Hence, during active l@ag, these ligaments in isolation provide little mechanical assistance. Keywords: Interspinous
ligament, supraspinous ligament, back extension moments

INTRODUCTION
One of the most contentious issues in biomechanics at resent is the mechanism of the human back during P ifting. In particular, the source of the moment required to extend the back as a weight is lifted is a subject of considerable debate>2. McGill and Norman3,4, as a result of their modelling, believe the back muscles to be capable of providing all the extensor moment required when raising the weight of the trunk and an external load, claiming that posterior ligaments and fascia have no role to play. Other workers, however, take the view that the back muscles alone are unable to generate sufficient force to overcome the significant moments exerted by body weight and an external load. Bogduk et aL5 conclude that, in heavy lifting, the back muscles must be assisted by some other mechanism. Those suggested include the intra-abdominal balloon mechanism6, the thoraco-lumbar fascia mechanism7 and the posterior ligamentous system**. However, the magnitude of the extensor moments produced by these mechanisms has been questioned. This paper seeks to clarify the mechanical role of the interspinous and su raspinous ligaments of the lumbar spine. It is hope cf that this information will be of use to modellers interested in the role of the posterior ligaments in the production of an extensor moment, and also to clinicians, since these ligaments are innervated and damage to them could be expected to lead directly to pain. Heylings, following earlier work by Rissanen12 studied the structure
Correspondence and reprint requests to: Dr M.J. Pearcy, Department of Orthopaedic Surgery and Trauma, Royal Adelaide Hospital, North Terrace, Adelaide, South Australia 5000, Australia 0 1990 Butterworth-Heinemann for BES 0141-5425&O/004340-05 340 J. Biomed. Eng. 1990, Vol. 12, July

and attachments of the supraspinous and intraspinous ligaments in 28 human lumbar spines by dissection and histological examination. He found that the interspinous ligament crossed the inters inous space in a posterocranial direction. Heylings P1 also found that ventrally the interspinous ligament joined with fibres of the ligamenturn flavum and that dorsally fibres assed into the supraspinous ligament or the media P tendons of the erector spinae. On the other hand, Prestar13 found no connections between the interspinous and supraspinous ligaments. The view of Heylings would seem to be that most widely accepted. Heylings hypothesized that the inters inous and supraspinous ligaments are clearly B esigned to limit flexion. Chazal et al. I4 suggested that in maximum physiological flexion the interspinous and su raspinous ligaments had reached their maximum g iomechanical possibilities. Silver15 noted that the interspinous ligaments seemed to be stretched at the limit of flexion and Panjabi et al. I6 concluded that it was the posterior elements which provided the stability in flexion. Adams et al. l7 found the supras inous/interspinous ligaments to be slack at small ang P es of flexion and to come into tension only for the final few degrees of flexion, but to be the first to fail immediately after the limit of flexion was reached. They found that, on average, the supraspinous and interspinous ligaments accounted for only 19% of the overall bending moment of the whole intervertebral joint. As a result of their X-ray analysis of in uiuo ligament deformations during flexion, Pearcy and Tibrewal* concluded that the interspinous ligament could be active only in the extremes of flexion. Various workers have noted that the centre of rotation of a motion segment varies instantaneously

Spinal ligaments: RJ Hide

et al.

as the joint flexes or extends, this point being known as the instantaneous axis of rotation OAR). Ogston et al. , using data obtained from an X-ray stud , demonstrated that vertebrae rotate about a variab re axis of rotation. Seligman et a1. showed, b computer analysis of specimens in a test bed, that x is axis moved on a locus, which changed as a disc degenerated. They indicated that the position of this centre of rotation would seem to lie slightly posterior to the centre of the intervertebral disc in the upright position, movin anteriorly during flexion. A recent study of the su%ject by Pearcy and Bogduk21 has shown that the instantaneous axis of rotation the (IAR) in normal subects is located on the superior vertebral endplate o f the lower vertebra of a lumbar motion segment, just posterior to the centre of the disc. Their workgoin agreement with the findings of Seligman et al. , suggests that for flexion from upright, the IAR lies anterior to that for extension from upright; in other words the locus of the IAR would seem to move anteriorly as a joint moves from extension to flexion. When Adams et al. l7 tested their whole motion segments they imposed a fixed axis of rotation anterior to the centre of the disc. Other in vitro tests of the isolated ligaments also loaded the ligaments in a manner different from the loading that occurs in life, without a variable axis of rotation. It can be deduced that the IAR moves through a locus of up to 20mm during flexion from extension, moving forwards through the disc. Assuming that, in the normal motion segment, the locus of the IAR moves from the posterior of the intervertebral disc to the centre as the movement passes from extension to flexion, it can be calculated from various anthropometric studies22T23 of lumbar vertebrae that the distance from the centre of the ligament to the IAR increases from 30 mm at extension to between 45 and 50mm at full flexion. To clarify the mechanical role of the interspinous and su raspinous ligaments an apparatus was designed to B eform them in a manner representing their true deformations in life.

Figure

Specimen

in place ready for testing

PROCEDURFA
The normal range of lumbar sa$ittal flexion varies from 13 for Li-2 to 16 for L&s2 . It was decided to test all the specimens in this study over a range of lo, this value being deemed sufficient since it was not possible to assess accurately the zero, or standing position. Also, since repeated tests were being carried out, it would have been unwise to risk straining the ligaments beyond the normal physiological range of motion. As mentioned previously, it can be deduced that the distance between the centre of the ligament and the IAR moves between 30mm and 50mm, moving forwards through the disc, from extension to flexion. To achieve this, the ligaments were mounted between two arms which were fixed to an axle such that the arms rotated about a set axis. The linear translation of a Hounsfield testing machine, used for this study, was converted to rotation using a rack and pinion arrangement. Specimens were tested with the axis of rotation fixed in different positions to build up a picture of their function with a variable axis of rotation. Experiments by Hasberry and Pearcf5, following earlier work by Viidik26, showed the

importance of testing specimens under conditions as close as possible to those that exist inside the body. Therefore, the whole apparatus was enclosed in a humidity chamber which maintained conditions of 100% relative humidity and 37C. The s ecimens used in this study were removed from ca s avers and consisted of two or more adacent lumbar spinous processes with their associate dl inter and supraspinous ligaments. Specimens were frozen at -20C until required. Before testing, specimens were divided into units of two adjacent spinous recesses (with the inter- and supra-spinous ligaments g etween them), by cutting through the centre of the spinous process. The specimens were then ready for mounting in the two ligament housing blocks. During the whole of this procedure, Ringers saline solution was used to keep the material moist. The specimen was attached to the housing blocks usin four metal pins inserted through holes drilled in tB e bone. In Figure 7, a specimen is shown in place inside the humidity chamber before testing. Tests were carried out initially with the inters inous/supras inous ligament complex left intact, tI! e centre of t! e ligament being 30, 35, 40, 45 and 50mm from the centre of rotation. Half the specimens were tested with the position of the axis in this order, and half in reverse. The test cycle consisted of five individual cycles, the fifth being recorded for analysis. This was done because with successively changes its applied strain cycles the ligament slightly, a characteristic of visco-elastic response materials. Thus by pre-conditioning the specimens, a consistent res onse was achieved by the fifth cycle. Throughout tx e tests, the specimen was observed carefully for any signs of failure of the bone around the pins, slip ing at the pins or failure of the actual ligament. Fol P owing these tests, the mechanical action of the supraspinous ligament was removed by sectioning it at the top and bottom of the space between the two adjacent spinous processes. The procedure was then repeated to obtain the characteristics of the isolated interspinous ligament. At the end of each test cycle each interspinous ligament was extended by 30 in an attempt to cause it to fail. As ligaments display properties characteristic of a visco-elastic material, it could be expected that

J. Biomed.

Eng. 1990, Vol. 12, July

341

Spinal ligamnts: RJ. Hid% et al.

different strain rates would produce different mechfive specimens were anical responses. Therefore, tested at two strain rates; a low rate at 0.5s~, the speed used for the other specimens, and a high rate at 12.5 s-l. The higher rate, equivalent to flexing fully from standing in 1 s, was the fastest rate that the Hounsfield machine used in this study could achieve. Tests were carried out at 30,40 and 50mm from the IAR to the centre of the ligament at low then at high rates for all five specimens. Thirteen specimens, removed from 11 cadavers (seven male and four female) were tested (Tables 7 and 2). Ages ranged from 50 to 87 years at the time of death (mean 65.7 years). All ligaments appeared healthy upon dissection, with no apparent ruptures. RESULTS When the intact supra/interspinous ligament complex was tested, load-extension curves typical of a visco-elastic material were produced (Figure 2). For hase of low an increase in extension, an initial a gradua P increase in stiffness was followed b stiffness, the ligaments on ry taking on load towards the end of flexion. When the ligament was rotated back to its original position, the load sustained by the than it had specimens dropped more rapidly increased during flexion. A comparison of different specimens from the same intervertebral level tested with the same rate of rotation showed strong similarities; even if the ligaments took a greater loading, the graph kept its characteristic shape. No consistent differences in the responses of ligaments from different levels were observed. The effect of varying the axis of rotation is more interesting. For the smallest distance to the axis of rotation, 30mm, all the specimens showed a ne ligible load-carrying characteristic even when flexe B to 10. As the distance to the axis of rotation was increased in steps to 35, 40, 45 and 5Omm, the s ecimens gradually carried a greater load (Figure 3), tR e load increasing more rapidly the greater the distance to the axis of rotation. The moment resisted by the supraspinous and interspinous ligaments, with
5 r 4-

5r
mm

mm

mm

mm

Rotation (degrees)

Figure 3 Typical response of an Ls_+ inter/supraspinous ligament

when tested at different distances from the centre of rotation

Table 1 Details of the specimens and the maximum extension moments and forces they produced when tested with the centre of rotation 50 mm from the centre of the interspinous ligament at the low strain rate of 0.5 SK Specimen number 1 2 3 4 5 6 7 8 Moment Ace 50 64 58 58 61 61 73 62 Sex M F M M M M F M Level L&s La-4
4-4

Force (N 62 44 54 76 48 46 90 72

(Nm) 3.1 2.2 2.7 3.8 2.4 2.3 4.5 3.6

L6s 2:; L&s Lk-5

the specimen 50mm from the IAR, varied between 2.2 and 6.7 Nm at the full 10 of flexion (Tables 7 and When the supraspinous ligament was sectioned and the intersoinous liaament tested in isolation, the load carried by the ligment decreased, typically by around 25%. Figure 4 shows the response of a
45 mm

4.

45,mm

-2
5
E 3-

I /I /I

E P
2/

/GOmm

I
I 2 3 4 5 Rototion 6 (degrees) 7 8 9 IO

5 6 7 Rototion (degrees)

IO

Figure 2
complex

Typical response of an L.-4 inter/supraspinous ligament

Figure 4 Typical graph showing the mechanical response of the Ls_., interspinous and supraspinoua ligaments together (-) and also the intempinous ligament in isolation (----) at three distances from the IAR

342

J. Biomed. Eng. 1990, Vol. 12, July

Spinal ligaments: RJ Hindtk et al. Table 2 Details of the specimens and the maximum extension moment and force they produced when tested with the centre of rotation 50 mm from the centre of the interspinous ligament at both low and high strain rates Specimen number Strain rate Age Sex Level 0.5 s-l Moment (Nm) 9 10 11 12 13 65 67 67 67 67 M F M F M L-4 5-4 L-4 LL -3 k-4 4.5 3.8 2.0 5.8 1.7 Force (N 90 76 40 116 34 12.5~~ Moment Wm) 5.9 4.4 2.5 6.7 2.2 Force WI 118 88 50 134 44

*r
-T z E E $ 4

High

Low

./.:

IO

Rotatlon(degrees) Figure 5 Response of a typical Lsm4 inter/supraspinous specimen when tested at low and high strain rates with an IAR 50 mm from the centre of the ligament

s inous ligament alone. Although the carried load dp ecreased, it can be seen that the characteristic sha e e of the graph is maintained. When the results for tYl ligaments tested at both the high and low strain rates are considered, it becomes apparent that when loaded at the higher speed the inter/supraspinous ligament complex takes on between 15-30% more load (Table 2). Figure 5 illustrates the typical difference in response shown by low and high strain rate applications. Attempts to produce failure in the interspinous ligaments reduced one of two results; either the ligament pu P led away from the bone of the spinous process, or the extension of 30 (the maximum possible) proved insufficient to induce failure in the ligaments. DISCUSSION The purpose of this study was to determine the mechanical characteristics of the inter/supraspinous ligaments as the locus of the IAR moves from posterior to anterior during flexion. The graph for an IAR 30mm from the ligament could be considered relevant for the first 20% of the extension range, since the specimens were tested at five different axes of rotation. Therefore, the graph for an TAR 35 mm from the ligament would only be relevant for the second 200/o, and s3 on, until the gra h for the IAR 50mm from the ligament represente B the final 20%. There-

fore, the curve for the IAR 50 mm from the ligament represents the instance when most moment is resisted by the ligament complex, although the initial few degrees of rotation will show the ligaments to be carrying more load than would actually be the case in real life. It is apparent from the results that the supraspinous and interspinous ligaments are active only in the later stages of flexion, the amount of load carried increasing rapidly towards full flexion. This observation is in close agreement with the results of Adams et al. 17, as are the actual values of moment resisted by the inter/ supras inous ligaments. The values we obtained range B between 2.2 Nm and 6.7 Nm, corn ared with a mean value of 9.33 Nm obtained by A cfams et al., whose higher values are a reflection of three experimental differences. First, their specimens were drawn ounger population; second, from a considerably their tests were Carrie dyout with a centre of rotation anterior to any used in this study, and finally the tested their specimens to the limit of flexion, whit L was more than the 10 used in this study, in a number of their tests. From the values of maximum moment resisted by the ligaments, the amount of load carried can be calculated as shown in Tables 7 and 2. At the high strain rate, a maximum moment of 6.7 Nm was resisted, approximately 10% of the value for the whole joint, working from the figures of Adams et al., or 5% of the muscle moment available to extend the spine5. Adams et al. also remarked that interjsupraspinous ligaments were the first to fail after the limit of normal flexion had been exceeded. This would tend to agree with Rissanens observation12 that More than 20% of the adult lumbar spines had visibly ligaments and that tom ruptured interspinous attachments to spinous processes were very common after 30 years of age. Our results also indicate that if forced beyond physiological ranges of flexion, the failure of the interspinous ligament at the interface with the spinous process is a likely consequence. Rupture of the ligament may result if vertebrae are forced to move with an IAR over 50mm from the ligament, as may be the case during trauma for example. It is conceivable, under these conditions of high strain rate and large moment arm, that a small rotation of the vertebrae into flexion would induce a high strain, damaging the ligament. Additionally, Seligrnan** showed that in patients with degenerative disc disease the locus of the IAR changes, greatly increasing the distance from the IAR to the ligament. For severely damaged discs it is possible that small movements overstrain the interspinous ligaments.

J. Biomed. Eng. 1990, Vol. 12, July

343

Spinal ligaments: RJ. Hindle et al. This would tend to agree with clinical experience which suggests that the interspinous ligament is damaged only as a result of anterior shear fractures. The attempt during this study to assess the contribution of the su raspinous ligament should be viewed with caution. T! is ligament has fibres running across several levels of the spine, and so testing only a section of the entire ligament cannot be expected to represent the mechanical characteristics of the whole ligament. Similarly, it is quite possible that the mechanical response of the interspinous ligament, demonstrated here in isolation, is unrepresentative of the in vivo situation. It is possible that connections with the thoracolumbar fascia may tense the ligament and cause it to take on more load at smaller angles of flexion, providing a mechanism to stabilize the vertebral column when the fascia tightens during flexion27. Certainly the direction of the fibres of the interspinous ligament are ideal for retaining the fascia. CONCLUSIONS This paper has demonstrated that the human lumbar interspinous li ament, when tested in isolation and when attache 9 to the supraspinous ligament only takes on load towards the end of flexion. The interspinous and supraspinous ligaments can provide useful assistance in restraining passive flexion but, unless they act in combination with other posterior line tissues such as the fascia, can only provide an additional 5% of the anti-flexion moment which it has been calculated, should be produced by the back muscles across each intervertebral joint. Thus during active lifting these ligaments, in isolation, have little mechanical function. ACKNOWLEDGEMENTS The authors are indebted to Angela Laidler for her assistance with the specimens.
5. Bogduk N, Macintosh JE, Pearcy MJ. Universal model of the lumbar back muscles. Spine (in press). 6. Bartelink DL. The role of abdominal pressure in relieving the pressure on the lumbar intervertebral discs. JBoru Joint Surg 1957; 39B: 718-25. 7. Gracovetsky S, Farfan HF, Helleur C. The abdominal mechanism. Spine 1985; 6: 317-24. 8. Gracovetsky S. Function of the spine. J Biomed Eng 1986; 8: 217-23. 9. Gracovetsky S, Farfan HF, Lamy C. The mechanism of the lumbar spine. Spine 1986; 6: 249-62. 10. Macintosh JE, Bogduk N, Gracovetsky S. The biomechanics of the thoracolumbar fascia. Clin Biomech 1987; 2: 78-83. 11. Heylings DJA. Supraspinous and interspinous ligaments of the human lumbar spine. JAnut 1978; 125: 127-41. 12. Rissanen PM. The surgical anatomy and pathology of the supraspinous and interspinous ligament ruptures. Acta Orthop Scund 1960; 46 (Suppl): 35. 13. Prestar FJ. Morphology and function of the interspinous and supraspinous ligaments of the lumbar spine. Morph01 Med 1982; 2: 53-8. 14. Chazal J, Tanguy A, Bourges G, Gaurel G, Escande G, Guillot M, Vanneuville G. Biomechanical properties of spinal ligaments and a histological study of the supraspinal ligament in traction. JBiomech 1985; 18: 167-76. 15. Silver PHS. Direct observations of changes in tension in the supraspinous and interspinous ligaments during flexion and extension of the vertebral column of man. J Anat 1954; 88: 550. 16. Panjabi MM, White AA, Johnson RM. Cervical spine mechanics as a function of transection of components. J Biomech 1975; 8: 327-36. 17. Adams MA, Hutton WC, Stott JR. The resistance to flexion of the lumbar intervertebral joint. Spine 1980; g(2): 245-53. 18. Pearcy MJ, Tibrewal SB. Lumbar intervertebral disc and ligament deformations measured in vivo. Clin Orthop 1984; 119: 281-6. 19. Ogston NG, King GJ, Gertzbein SD, Kapasouri A, Rubenstein JD. Centrode patterns in the lumbar spine. Spine 1984; 9: 591-5. 20. Seligman JV, Gertzbein SD, Tile M, Kapasouri A. Computer analysis of the spinal segment motion in degenerative disc disease with and without axial loading. Spine 1984; 9: 566-73. 21. Pearcy MJ, Bogduk N. Instantaneous axes of rotation of the lumbar intervertebral joints. Spine 1988; 13: 1033-41. 22. Nissan M, Gilad I. The cervical and lumbar vertebrae - an anthropometric model. Eng Med 1984; 13: 111-14. 23. Vanharanta H, Korpi J, Heliovaara M, Troup JDG. Radiographic measurements of lumbar spinal canal size and their relation to back mobility. Spine 1985; 10: 461-6. 24. Pearcy MJ. Stereo radiogra hy of lumbar spine motion. Acta Orthop &and 1985; 56 P Suppl212). 25. Hasberry S, Pearcy MJ. Temperature dependence of the tensile properties of interspinous ligaments of sheep. J Biomed Eng 1986; 8: 62-6. 26. Viidik A. Functional properties of collagenous tissues. Znt Rev Connective Tissue Res 1973; 6: 127-215. 27. Bogduk N, Twomey LT. Clinical anatomy of the lumbar spine. Edinburgh: Churchill Livingstone, 1987: 89-91.

REFERENCES
Gracovetsky S. Potential of lumbodorsal fascia forces to generate back extension moments during squat lifts (letter to the editor). JBiomed Eng 1989; 11: 172-3. McGill SM, Norman RW. Potential of lumbodorsal fascia forces to generate back extension moments during squat lifts (letter to the editor). J Biomed Eng 1989; 11: 173-5. McGill SM, Norman RW. Partioning of the L,-Ls dynamic moment into disc, ligamentous and muscular components during liftmg. Spine 1986; ll(7): 666-78. McGill SM, Norman RW. Potential of lumbodorsal fascia forces to generate back extension moments during squat lifts. JBiomed Eng 1988; 10: 312-8.

344 J. Biomed. Eng. 1990, Vol. 12, July