Surgical Radiologic Anatomy

SRA (1986) 8 : 175-182

9 Springer.Verlag 1986

Original articles

Comparison of the structure of human intervertebral discs in the cervical, thoracic and lumbar regions of the spine JS Pooni ~, DWL Hukins 2, PF Harris 1, RC Hilton 3, KE Davies 2 t Department of Anatomy, University of Manchester, Manchester M13 9PT, England 2 Department of Medical Biophysics, University of Manchester, Manchester M13 9PT, England 3 Hope Hospital, Salford M6 8HD, England

Summary.

Posterior and anterior heights, crosssectional area and shape were measured for all the intervertebral discs in four spines from elderly human cadavers. Disc height was a minimum at the T4-5 level; thoracic discs were less wedge-shaped than those in the cervical and lumbar regions. Cross-sectional area increased from the cranial to caudal extremity; at the L5-SI level the nucleus pulposus occupied a high proportion of this area. Cervical discs tended to have an elliptical cross-sectional shape, thoracic discs were more circular and lumbar discs tended to have an elliptical cross-section which was flattened or re-entrant posteriorly. This shape distribution was quantified by defining a shape index which had a maximum value of 1 for a circular cross-section. Orientations of the reinforcing fibres in the outer lamellae of the anterior annulus fibrosus were measured from 27 discs by X-ray diffraction. For these measurements, C3-4, T7-8 and L2-3 were chosen as representative of cervical, thoracic and lumbar discs. The fibre tilt, with respect to the axis of the spine, was significantly less in the cervical discs (at 65 ~) than in the thoracic and lumbar discs (about 70~ These findings are interpreted in relation to differing functional requirements and possible mechanisms of failure in the cervical, thoracic and lumbar regions of the spine in the light of current knowledge on the biomechanics of the intervertebral disc.

Offprint requests :

Dr DWL Hukins

Comparaison de la structure des disques intervert6braux humains dans les r6gions cervicale, thoracique et lombaire de la colonne vert6brale R6sum6. Les hauteurs post6rieure et ant6rieure, la superficie et la forme en section transversale ont 6t6 mesur6es pour tousles disques intervert6braux de quatre colonnes vert6brales pr61ev6es chez des sujets ~g6s. La hauteur du disque 6tait minimale au niveau T4-5; les disques thoraciques 6taient moins cun6iformes que ceux des r6gions cervicale et lombaire. La superficie en section transversale augmentait de l'extr6mit6 crS.aiale jusqu'~ l'extr6mit6 caudale; au niveau LS-S1 le noyau g61atineux occupait une grande proportion de cette surface. Les disques cervicaux avaient tendance ~t poss6der une forme elliptique en section transversale, les disques thoraciques 6taient plus circulaires et les disques lombaires avaient tendance h poss6der une section transversale elliptique qui 6tait plane ou concave en arri6re. Cette distribution de formes a 6t6 quantifi6e en d6finissant un indice de forme qui avait une valeur maximum de 1 pour une section transversale circulaire. Les orientations des fibres de renfort dans les lamelles ext6rieures de l'anneau fibreux ant6rieur de 27 disques ont 6t6 mesur6es au moyen de la diffraction des rayons X. Pour ces mesures, nous avons choisi C3-4, T7-8 et L2-3 comme repr6sentatifs des disques cervicaux, thoraciques et lombaires. L'inclinaison des fibres par rapport l'axe de la colonne vert6brale, 6tait significativement moindre dans les disques cervicaux (vers 65 ~ que dans les disques thoraciques et lombaires (vers 70~ Ces

176 rrsultats sont interprrtrs selon les exigenees fonctionnelles diffrrentes et tes mrcanismes possibles de rupture dans les rrgions cervicale, thoracique et lombaire de la colonne vertrbrale, en tenant compte des connaissances contemporaines de la biomrcanique du disque intervertrbral.

Key w o r d s : Intervertebral disc - Spine

The paper is concerned with a comparison of the structure of the intervertebral discs along the spinal column in order to relate the results to different functional requirements in different regions of the spine. Measurements which were made on discs included : (i) height, (ii) cross-sectional area, (iii) shape, and (iv) orientation of the reinforcing fibres of the annulus fibrosus. It is well established that disc height contributes between 20% and 33% of the total length of the spine [22, 23], However, there are slight regional differences in the percentage in the cervical, thoracic and lumbar spine [12] and it also changes with age [18, 25]. Furthermore, adult cervical and lumbar (but not thoracic) discs are thicker anteriorly than posteriorly [4, 12, 20]. Farfan [4] measured the cross-sectional shapes of lumbar discs and related the results to differing functional requirements within the lumbar spine. We are aware of no previous quantitative studies of disc structure in different regions of the spine of the kind presented in this paper. Some of our measurements can be more readily related to disc structure and function than others. Disc height is related to the mobility of the intervertebral joint while cross-sectional shape is related to the ability to withstand the effects of flexion and torsion [4]. Furthermore, there are conflicting criteria for disc shapes which confer strength in flexion and in torsion - the two movements most likely to damage the intervertebral disc [1, 5, 6, 9]. The inner region of the disc, the nucleus pulposus, has a gelatinous texture which becomes fibrous during ageing [18, 25]. It is surrounded by the lamellae of the annulus fibrosus which is reinforced by fibres of collagen. In any one lamella the fibres are parallel and tilted, by about 65 ~, with respect to the longitudinal axis of the spine, the direction of tilt alternates in successive lamellae [13]. This pattern of fibre orientations is believed to be related to the mechanical stability of the disc [9]. Although many qualitative observations of disc structure have been recorded, there has previously been little attempt to relate structure to differing functions and patterns of failure in the cervical, thoracic and lumbar regions of the spine.

JS Pooni et al. : Disc structure and function

Materials and Methods

Materials The whole vertebral column, from the atlanto-occipital joint to the caudal extremity of the sacrum was removed from four cadavers which had been embalmed in a mixture of formalin (5%), phenol (6%), glycerol (12%), methylated spirits (62%) and water (15%). Details of the specimens are given in Table 1. These spines were used for investigating the gross structures of the discs, ie. measurement of height, cross-sectional area and shape. Table 1. Spines used for investigating the gross structures of the discs

1 2 3 4

Sex

Age/years

Cause of death

M F F F

73 86 85 80

Bronchial carcinoma * Arteriosclerosis Bronchopneumonia Pneumonia

9 Radiographs showed no evidence of bony metastasis Results from these four specimens were sufficiently close, as judged from the low standard errors shown in the figures of the "Results" section, that this small number of specimens was adequate for a useful comparison of structure along the spinal column. In order to compare orientations of the reinforcing fibres of the annulus fibrosus, a much larger number of specimens was required. Discs C3-4, T7-8 and L2-3 were removed by dissection from 27 cadavers (18 male and 9 female) whose age range is shown in Fig. 1. These discs were preserved in formol saline, which does not affect the orientation of the fibres in the annulus fibrosus [8]. 6

2 z

2O

40

60

80

iOO

AGE / Y E A R S

Fig. 1 Ages of cadavers used as sources of C3-4, T7-8 and L2-3 discs for measurementof the orientations of the reinforcing fibres of the annulus fibrosus Les hges des cadavresutilisrs pour les prrl~vementsdes disques C3-4, T7-8 et L2-3 afin de mesurer les orientations des fibres de renfort de l'anneau fibreux

Js Pooni et al. : Disc structure and function

177

Disc height The length of each spine was measured intact. A fine thread was placed carefully, along the mid-line to conform with the curvature between the tip of the odontoid process and the inferior border of the L5-S 1 disc. The thread was cut and measured to obtain the length of the spine. Lateral radiographs were taken of the vertebral column, both before and after removal of the posterior elements, in order to measure the height of the disc space ie. the distances between the bony surfaces of adjacent vertebral bodies. They also aided assessment of spinal pathology in the specimens. Radiographs were calibrated by means of a scale placed on the specimen. Standard reference points were used to measure anterior, central and posterior heights [16, 17]. Discs heights were also measured directly from the spines. Anterior heights were measured with engineering callipers. Neural arches were then removed by severing the pedicles along the entire length of the spine. Posterior disc heights could then be measured and anterior heights were remeasured.

Cross-sectional area and shape To determine cross-sectional area and shape, each disc was viewed in transverse section. Spines were first deep frozen at - 5 0 ~ for 48 hours and then sectioned serially in a transverse plane through the superior surface of each disc using a band saw. Thus after sectioning the spine there were 23 discs, each attached to its caudal vertebra. Specimens were thawed at room temperature, the cartilage end-plates removed to reveal the structure of the disc and the superior surface photographed after placing a steel rule adjacent to the disc. A Magiscan Image Analysis System (Joyce-Loebl Ltd., Gateshead, England) was used to measure the photographs. All measurements were made using the Simple Picture Evaluation Language SPEL [21]. The cross-sectional area of the whole intact disc and of the nucleus pulposus as well as the perimeter of the disc were measured.

Fig. 2 Typical X-ray diffraction pattern recorded from the outer lamellae of the annulus fibrosus of an L2-3 disc

Un diagramme typique de diffraction par les rayons X enregistr6 partir des lamelles extrrieures de l'annean fibreux d'un disque L2-3

Optical Comparator (Mitutoyo Manufacturing Co., Tokyo, Japan) and halved to obtain the tilt of the fibres with respect to the axis of the spine [8, 10, 11].

Results

Spinal pathology Spines 1 and 2 (see Table 1) appeared, from radiographs and on inspection, to be typical of healthy spines for cadavers of that age. There was some evidence of osteophyte formation along most of their length, especially in the cervical region where the C5-6 disc spaces were almost fused. Spines 3 and 4 showed slight lateral curvature in the upper lumbar and lower thoracic region (spine 3) and the lumbar (spine 4). There were fewer cervical osteophytes in these spines.

Disc height Fibre orientations The outer lamellae were dissected from the anterior face of each disc to a depth of a few millimetres. In this part of the investigation three discs were used from 27 spines, as described previously. X-ray diffraction patterns were recorded from the resulting specimens using a previously established technique [8]. Figure 2 shows a typical diffraction pattern which consists of a clear cross. The obtuse angle between the arms of the cross was measured using a Mitutoyo Model PJ-250C

Disc heights were expressed as a percentage of the total length of the spine from which they were removed. The mean values for all spines and the standard errors of the means could then be calculated and the results plotted as shown in Figure 3. Figures 3 et 4 show how the anterior heights of the discs vary along the length of the spine. Figure 3 is based on measurements taken directly from the spine while the data for Figure 4 were measured from radiographs. Cranial and caudal extremities were not,

178

JS Pooni et al. : Disc structure and function 7.

6 E 5 84

4.

3

+

<

"I2

0 CERVICAL

THORACIC

LUMBAR

I

CERVI CAL

Fig. 3 Mean anterior disc height, expressed as a percentage of the total length of the spine, measured directly from C2-3 to L5-S1. Error bars represent the standard error of the mean La hauteur antrrieure moyenne du disque, exprim6e comme pourcentage de la longueur totale de la colonne vertrbrale, mesurre directement de C2-3 jusqu'~t L5-S1. Les barres des erreurs reprrsentent l'erreur-type de la moyenne

I

THORACI C

LUMBAR

Fig. 5 Mean posterior disc height, expressed as a percentage of the total length of the spine, measured from lateral radiographs. Values are recorded for the length of the spine from C2-3 to L5-$1. Error bars represent the standard error of the mean La hauteur postrrieure du disque, exprimre comme pourcentage de la longueur totale de la colonne vertrbrale, mesurre h partir des radiographies de profil. Les valeurs sont notres tout au long de la colonne vertrbrale de C2-3 jusqu'h L5-$1. Les barres des erreurs reprrsentent l'erreur-type de la moyenne

7 1,4-

6 1,2-

1,0<

<

0,8-

3 ,~ o

0,60,4

>T

CERVICAL

[

THORACIC

LUMBAR

La hauteur antrrieure moyenne du disque exprimre comme pourcentage de la longueur totale de la colonne vert6brale, mesurde sur les radiographies de profil. Les valeurs sont notfes tout au long de la colonne vert6brale de C2-3 jusqu'~t L5-S1. Les barres des erreurs reprrsentent l'erreur-type de la moyenne

l

Jl

0,0 CERVICAL

Fig. 4 Mean anterior disc height, expressed as a percentage of the total length of the spine, measured from lateral radiographs. Values are recorded for the length of the spine from C2-3 to L5-S1. Error bars represent the standard error of the mean

I

I

il I 0,2

I

THORACIC

1 LUMBAR

Fig. 6 Posterior height divided by anterior height of discs C2-3 to L5-S1. Results are mean values from 4 specimens with the standard error in the mean. The nearer the value of this ratio to 1, the less wedge-shaped is the disc La hauteur du disque postrrieur, divisre par la hauteur antrrieure des disques C2-3, jusqu'~t L5-$1. Les rrsultats sont les valeurs moyennes de 4 prrlrvements avec l'erreur-type de la moyenne. Plus la valeur de ce rapport approche de 1, moins le disque est cunriforme

JS Pooni et al. : Disc structure and function

179

therefore, defined in the same way for the two sets of results. Nevertheless both figures show the same trends - there is a minimum in disc height at around T4-5 while the increase in thickness is greater caudally than cranially. Figure 5 shows the posterior heights of the discs measured from lateral radiographs o f the spines. The lumbar and lower thoracic discs are thicker posteriorly than discs in the cervical and upper thoracic spine. However, the regional variation is not so marked as it was for anterior disc height. Figure 6 shows that this is because the thoracic discs are less wedge-shaped in profile than cervical and lumbar discs.

12 ~ %~"

i0 8 I

~= ~ "<'

ah

6 o

\

,\

-

\9

4

\

-

\

2

\\ \N Disc area CERVICAL

Figure 7 shows that the cross-sectional area o f the discs increases almost linearly from the cervical to the lumbar region of the spine. However, in the lumbar region itself the areas of L2-3, L3-4, L4-5 and L5-S 1 discs are all virtually equal. Like that o f the whole disc, the area o f the nucleus 18

16

\ \ \ \ \ \ \

14

N,

12

,,<,

,

\ \

10

5t t t M~

4

I .I.

\

II'

I "~''

\

\

..'f-J 0

,T CERV I CAL

\

J

. THORAC I C

.

\ LUMBAR

Fig. 7 Cross-sectional areas of discs C2-3 to L5-SI. Results are mean values from 4 specimens with the standard error of the mean. Shaded columns represent results from 3 spines only because of the difficulty of making measurements on degenerate discs Les superficies en section transversale des disques de C2-3 jusqu'a L5-S 1. Les r6sultats sont les valeurs moyennes de 4 pr61bvements avec l'erreur-type de la moyenne. Les colonnes hachur6es ne repr6sentent que les r6sultats de 3 colonnes vert6brales, du fait de la difficult6

effectuer les mesures sur les disques ddg6n6r6s

I

THORACIC

LUMBAR

8

Cross-sectional area of the nucleus pulposus in discs C2-3 to L5-S1. Results are mean values from 4 specimens with the standard error in the mean. Shaded columns represent results from only 3 spines while the crossed column is the mean from 2 spines. The low number of measurements in these columns arises from the difficulty in making measurement on denegerate discs La superficie en section transversale du noyau g61atineux dans les disques de C2-3 jusqu'~i L5-S1. Les r6sultats sont les valeurs moyennes de 4 pr61~vements avec l'erreur-type dans la moyenne. Les colonnes hachur6es ne repr6sentent que les r6sultats de 3 colonnes vert6brales, tandis que la colonne barr6e repr6sente la moyenne de 2 colonnes vert6brales. Le petit nombre de mesure dans ces colonnes r6sulte de la difficult6 ~t effectuer des mesures sur les disques d6g6n6r6s

pulposus also increases linearly, as shown in Figure 8. However, the nucleus o f the L5-S1 disc occupies a much larger proportion of the cross-sectional area.

Disc shape

\

\ 2

Fig.

Figure 9 shows some typical cross-sectional shapes of cervical, thoracic and lumbar discs. They range from the obviously elliptical shape typical o f the cervical spine [Fig. 9(a)] to the much more circular shape o f the thoracic spine [Fig. 9((b)]. Lumbar discs tend to have an elliptical cross-section which is flattened or re-entrant posteriorly [4]. A numerical representation o f cross-sectional shape was used in order to compare disc structure in different regions of the spine objectively. A shape index, defined by S = 4:t x Area/Circumference 2 has a maximum value o f 1 for a circular cross-section; the lower the value of this index, the further the shape deviates from circular. It applies to any cross-sectional shape and is, therefore, preferable to the "ovality ratio" defined by Farfan [4] which strictly only applies to ellipses and circles.

180

JS Pooni et al. : Disc structure and function

1,0. 0,9'

4

0,8. \

-I"

-I-\

\,.,,

63

w

0,7.

• \ \9 \\~ 0,6- \ \ \\ \\ \\ \\

0,5- \ \ ~ \\ \ 0,4. \\

IN k"

\\ \\ \ \

N

\\ X] \~

0,30,2. I 0

Fig, 9 Cross-sectional shapes of (a) C7-T1, (b) T7-8 and (c) L4-5 disc

Fibre orientations Most specimens of a n n u l u s yielded a cross-shaped X-ray diffraction pattern (Fig. 2). However, some discs were so degenerate that a suitable s p e c i m e n could not be dissected from them and so the fibre orientation could not be measured. Cervical specimens, e v e n from the y o u n g e r cadavers, were especially prone to distorti6n b y osteophyte formation. Thoracic discs often posed a problem because of their small thickness (Fig. 4). However, most of the l u m b a r discs yielded satisfactory specimens. Fibre orientations were not significantly different in the annulus of thoracic and l u m b a r discs but the tilt angle was about 4 ~ less in the cervical annulus. Table 2 summarises the results from the C3-4, T7-8 and L2-3 discs which were chosen to represent the cervical, thoracic and l u m b a r regions. A n F-test showed that the three sets of data could not be regarded as m e m b e r s of the same overall population (p < 0 . 0 0 1 ) ; t-tests showed

\\

xx\

0,0

I\

N

0,I. \ \

5cm

Figure 10 shows the variation in S values along the spinal c o l u m n . It is clear that the discs o f the mid-thoracic spine are the most nearly circular while those o f the mid-cervical are the least. The presence of osteophytes made it difficult to trace the outline of some of the discs.

\\

\\

!

Les formes en section transversale des disques (a) C7-T1, (b) T7-8 et (c) L4-5

\ \\

~N

~N

\

\\ i CERV I CAL

\\ THORACIC

i LUMBAR

Fig, 10

Disc shape index, S, defined in the text, from C2-3 to L5-S 1. Results are the mean, and standard error of the mean, from 4 specimens except that shaded columns represent results from 3 specimens because of the difficulty in making measurements on degenerate discs. This index has a maximum value of 1 from a circular cross-section; the further the shape deviates from circularity the lower the value of the index L'indice de formes du disque, S, d6fini dans le texte, de C2-3 jusqu'~t L5-S1. Les r6sultats sont la moyenne, et l'erreur-type de la moyenne, de 4 prfl~vements, saul que les colonnes hachur6es ne repr6sentent que les r6sultats de 3 pr61~vements, ~ cause de la difficult6 ~ effectuer des mesures sur les disques d6g6n6rds. Cet indice poss~de une valeur maximum de 1 pour une section transversale circulaire; plus la forme s'61oigne d'un cercle, plus la valeur de l'indice est basse

Table 2. Fibre angles measured from X-ray diffraction patterns recorded from C3-4, T7-8 and L2-3 discs taken as representative of the cervical, thoracic and lumbar regions

C3-4 T7-8 L2-3

Number of discs

Mean

Standard deviation

13 17 24

65" 69~ 70~

2.5 ~ 2.8~ 1.8~

that the results from thoracic and l u m b a r discs were not significantly different (p = 0.6) but that there was a significant difference b e t w e e n cervical and thoracic (p = 0.001) as well as b e t w e e n cervical and l u m b a r (p = 0.0001).

JS Pooni et al. : Disc structure and function

Discussion It has been suggested that the range of movement of the intervertebral joint is related to disc thickness [4], which would explain why thoracic discs need not be so thick as those of the cervical and lumbar spine. Their geometry dictates that the thicker the discs, the longer the reinforcing fibres of the annulus fibrosus. Thus the fractional increase in fibre length (ie. fibre strain) caused by a given movement will be less for a thick disc than for a thin one [9]. As a result the strain energy stored by the stretched fibres during flexion-extension and torsion will be less for a thick disc. Since these movements are suggested to be responsible for tearing the annulus [1, 4, 5, 6, 9], reduction in strain energy is important for guarding against prolapse. Thoracic discs need not be so thick because the range of segmented motion, for both flexion-extension and torsion, is much less in the thoracic than in the cervical or lumbar regions of the spine [7]. The cross-sectional area of the discs increases from the cranial extremity along the length of the spine (Fig. 7). The body weight acting on the discs increases from the cranial extremity along the length of the spine. However, the pressure acting on the discs will not increase to the same extent because the cross-sectional area also increases in the caudal direction. In any case it appears that the disc is unlikely to be damaged by axial pressure in vivo because, during in vitro mechanical testing, compression of the intervertebral joint to failure invariably leads to fracture of the cancellous bone of the vertebral body and not tearing of the annulus fibrosus [2, 14, 19, 24]. There are conflicting criteria for disc strength in flexion and torsion. These two motions are considered to be potentially damaging to the intervertebral disc [1, 5, 6, 9]. A disc with a circular cross-section would have an even distribution of strain around the circumference of the annulus fibrosus in torsion; however, in flexion few of the fibres in the annulus would be able to provide optimum reinforcement and so the disc could be readily damaged [9]. In complete contrast, elliptical discs (especially if flattened or re-entrant posteriorly) are expected to be strong in flexion but weak in torsion [4, 9]. Therefore it is not surprising that thoracic discs are the more nearly circular (Fig. 9) because the range of flexion-extension is much more restricted than in the cervical and lumbar regions [7]. Lumbar discs are less circular and have restricted torsional motion [7]. In the cervical region discs have to withstand a considerable range of both flexion-extension and torsion [7]; as a result they are likely to be especially susceptible to damage. Their elliptical shape would be expected to provide some protection in flexion but to be relatively weak in torsion [4, 6, 9].

181

Finally, the patterns of orientations for the reinforcing fibres of the annulus fibrosus are very similar in the cervical, thoracic and lumbar discs. The fibres are tilted by 65 ~ with respect to the axis of the spine, in the cervical region but by about 70 ~ in the thoracic and lumbar regions (Table 2). In the lumbar spine this pattern of orientation is established when the fibres are first laid down prenatally and is retained during ageing [10, 11]. A simple analysis suggests that the tilt angle must exceed 54.7 ~ if the fibres are to reinforce the annulus in withstanding the internal pressure exerted by the nucleus pulposus [9]; a more thorough analysis, which takes into account bulging of the annulus and end-plates, suggests that in reality the tilt is likely to be nearer to 65 ~ in the living disc [15]. The regional difference we have observed is difficult to explain. Nevertheless, the tilt angles we have observed are such that nuclear pressure is unlikely to damage the annulus [15] which is more likely to be torn by flexion or extension. Acknowledgments. We thank Professor J Ball for advice on pathology, Dr J Adams for advice on radiology, and Dr DS Hickey for discussion. This research was supported by the Medical Research Council.

References 1. Adams MA, Hutton WC (1982) Prolapsed intervertebral disc : a hyperflexion injury. Spine 7 : 184-191 2. Brown T, Hansen RJ, Yorra AJ (1957) Some mechanical tests on lumbo-sacral spine with particular reference to intervertebral discs. A preliminary report. J Bone Joint Surg 39A : 1135-1164 3. Coventry MB, Chromley RK, Kernahan JW (1945) The intervertebral disc, its microscopic anatomy and pathology. Part II: changes in the intervertebral disc concomitant with age. J Bone Joint Surg 27 : 233-247 4. Farfan HF (1973) Mechanical Disorders of the Low Back. Lea and Febiger, Philadelphia 5. Farfan HF, Cossette JW, Robertson GH, Wells RV, Kraus H (1979) The effects of torsion on the lumbar intervertebral joints : the role of torsion in the production of discs degeneration. J Bone Joint Surg 52A: 468-497 6. Farfan HF, Huberdeau RM, Dubow MI (1972) The influence of geometric factors on the pattern of disc degeneration - a post mortem study. J Bone Joint Surg 54A : 492-510 7. Grieve GP (1981) Common Vertebral Joint Problems, Churchill Livingstone, London, pp 41-43 8. Hickey DS, Hukins DWL (1979) Effect of methods of preservations on the arrangement of collagen fibrils in connective tissue matrices; an X-ray diffraction study of annulus fibrosus. Connect Tissue Res 6 : 2 2 3 - 2 2 8 9. Hickey DS, Hukins DWL (1980) Relation between the structure of the annulus fibrosus and tile function and failure of the intervertebral disc. Spine 5 : 106-116 10. Hickey DS, Hukins DWL (1980) X-ray diffraction studies of the arrangement of collagenous fibres in human fetal intervertebral disc. J Anat 131 : 81-90 11. Hickey DS, Hukins DWL (1982) Aging changes in the macromolecular organization of the intervertebral disc : an X-ray diffraction and electron microscopic study. Spine 7 : 2 3 4 - 2 4 2

182 12. Hollingshead WH (1969) Anatomy for Surgeons. Harper and Row, New York, pp 79-199 13. Horton WG (1958) Further observations on the elastic mechanism of the intervertebral disc. J Bone Joint Surg 40B : 552-557 14. Jayson MIV, Herbert CM, Barks JS (1973) Intervertebral discs : nuclear morphology and bursting pressure. Ann Rheum Dis 32 : 308-315 15. Klein JA, Hickey DS, Hukins DWL (1983) Radial bulging of the annulus fibrosus during compression of the intervertebral disc. J Biomech 16:211-217 16. Lusted LB, Keats TE (1973) An Atlas of Roentgenographic Measurement. Year Book Medical Publishers, Chicago, pp 116117 17. Nehme A-ME, Riseborough EJ, Reed RB (1980) In : Scoliosis (1979). Zorab PA and Siegler D (eds.). Academic Press, London, pp 103-109 18. Peacock A (1952) Observations on the postnatal structure of the intervertebral disc in man. J Anat 86 : 162-179

JS Pooni et al. : Disc structure and function 19. Roaf R (1960) A study of the biomechanics of spinal injuries. J Bone Joint Surg 42B : 810-823 20. Rothman RH, Simeone FA (1975) The Spine, Vol. 1. WB Saunders, Philadelphia, pp 19-68 21. Taylor CJ, Brunt JN, Dixon RN, Gregory PJ (1977) The Magiscan : a new generation, software based, automatic image analysis. In : Quantitative Analysis of Microstructures in Materials Science, Biology and Medicine. Dr Riederer-Verlag GmbH, Stuttgart, pp 433-442 22. Taylor JR (1975) Growth of the human intervertebral disc and vertebral bodies. J Anat 120:49-68 23. Todd WT, Pyle IS (1928) A quantitative study of the vertebral column by direct and roentgenographie methods. Am J Phys Anthropol 21 : 321-337 24. Virgin WJ (1951) Experimental investigations into the physical properties of intervertebral discs. J Bone Joint Surg 33B : 607-611 25. Walmsley R (1953) The development and growth of the intervertebral disc. Edinburgh Med J 60 : 341-365