The effects of bone density and disc degeneration on the structural property distributions in the lower lumbar vertebral endplates

Journal of Orthopaedic Research

ELSEVIER

Journal of Orthopaedic Research 20 (2002) I 1 15-1 120 www.elsevier.com/locate/orthres

The effects of bone density and disc degeneration on the structural property distributions in the lower lumbar vertebral endplates J.P. Grant

a,

Thomas R. Oxland

a,*,

Marcel F. Dvorak ’, Charles G. Fisher

Diuision of Orthopaedic Engineering Research, Departments of 0rthopuerlic.s and Mechanical Engineering, Uniiwsity of’ Briti.sIi Columhirr und Vancouwr Hospitul und Health Sciences Centre, 3114-910 West 10th Avenue, Vancourw, BC, Canudn V5Z 4E3 Division q f Spine, Depurtment of Orthopaedics, University of’ British Colunihiu und Vancouver Hospital and Healih Sciences Centre, 3114-910 West 10th Avenue. Vancouver. BC, Canudu V5Z 4E3

Abstract

In this study, we hypothesized that vertebral bone density and disc degeneration would affect the structural property distributions of the lower lumbar vertebral endplates (L3-L5). The results may have implications for improving interbody implant designs to better resist subsidence. A 3 mm diameter hemispherical indenter was used to perform indentation tests at 0.2 mm/s to a depth of 3 mm at 27 standardized locations in 55 bony endplates of intact human lumbar vertebrae (L3-L5). The resulting load-displacement curves were used to extract the failure load and stiffness of each test site. Bone density was measured using lateral DEXA scans. Disc condition was determined using a four-point grading scale. Three-way analyses of variance were used to analyze the relationships between the data. The overall failure load decreased with bone mineral density (BMD) in the superior (p < 0.0001) and inferior (p = 0.01 1) lumbar endplates. In both endplates, the posterolateral regions were significantly stronger than more central regions. With increasing BMD, this difference became more pronounced in the superior endplates only (p = 0.005). Increased disc degeneration was associated with an overall failure load decrease in the inferior lumbar endplates (p = 0.002). The strength in the central regions of the superior endplates was reduced with increasing degeneration, but this was not observed peripherally (p = 0.001). Stiffness magnitude or distribution was not significantly affected by BMD or disc degeneration. The locations of the strongest regions of the endplate did not change with either bone density o r disc degeneration. This implies that implant shapes designed using the basic structural property maps for the L3--L5 endplates are appropriate for use in patients with a wide range of pathologies, even though overall failure loads are generally lower in patients with reduced bone density and greater degrees of disc degeneration. 0 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved.

Introduction Interbody fusion is a surgical technique in which a bone o r synthetic spacer is implanted between two adjacent vertebrae to provide structural support following removal of diseased or damaged tissue from the spinal column. One mode of failure for this type of procedure is subsidence, in which the implant sinks into one or both of the vertebrae. This can lead to pain, deformity or nerve damage and can only be corrected through revision surgery. One means of reducing the incidence of subsidence is to modify implant designs to take advantage of the strongest regions of the vertebral bodies. *Corresponding author. Tel.: +1-604-875-5475: fax: +1-604-8754677. E-mail address: [email protected] (T.R. Oxland).

It was shown in a previous paper [S] that there is significant variation in the structural properties over the upper and lower surfaces of the lower lumbosacral vertebral bodies (L3-Sl). The superior and inferior lumbar endplates are strongest around the periphery, particularly in the posterolateral region, in front of the pedicles. The sacral endplate is strongest at the posterior margin, with a steady decrease in strength from the posterior to anterior margin. The majority of existing implants are placed in the center of the intervertebral space, where the vertebral body is weak, thus modifying implants to engage the stronger bone regions may significantly reduce the incidence of subsidence. Degenerative changes in the spine may affect the structural property distribution, which would have implications for implant design. The compressive strength of vertebrae is related to bone density [3,4,9]. Disc

0736-0266/02/$ - see front matter 0 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved. PII: SO7 3 6 - 0 2 6 6 ( 02)OOO 3 9 - 6

1116

J. P. G r m i

et ui.

I Journul OJ Orihopcwdic Research 20 (2002) I I IS I I20

degeneration also adversely affects the strength of the implant-vertebral body interface [2I]. These findings, however, are based on the strength of the entire vertebral body or the load causing implant subsidence into a vertebral body, rather than regional strength variation. Keller et al. [14] tested 1 cm cubes of trabecular bone taken from the superior and inferior surfaces of lumbar vertebrae and found that there was regional variation in the strength and stiffness, which changed with disc degeneration. The effects of the bony endplate, however, were not addressed in that study. Other authors have looked at the effects of age and osteoporosis on the trabecular bone morphometry within the vertebral body and found that the changes in bone structure are not uniform [1,5,12,18,20,25-271. The intervertebral discs have been shown to distribute load less uniformly over the vertebral endplates as they degenerate [6,11,17,19,23], which may have regional effects on bone remodelling [28]. The regional variation of endplate properties may change, reflecting these underlying structural changes. The purpose of this study was, therefore, to determine the effects of bone density and disc degeneration on the structural property distributions that were identified previously [7,8], specifically the superior and inferior lumbar distributions.

Materials and methods Spc4nen puepumtion

The specimens used in this study were the same as those used in the initial study that identified patterns in the strength and stiffness distributions in the L3-SI vertebral bodies [8]. Lateral radiographs of the full spine were used to screen the 1 1 L3-SI specimens used in the study. Donors were selected by the first author to represent a range of bone density and disc health based on the presence or absence of osteophytes, osteoporotic rractures, and decreased disc height. Based o n the radiographs, three spines were considered healthy, three showed signs of disc degeneration, two showed evidence of osteoporosis, and three showed evidence of both disc degeneration and osteoporosis. Donor age ranged from 48 to 90 years, with a mean age of 74.5 years. Three donors were female and eight were male. Of the 77 endplates in the I I L3-SI specimens, 55 L3-L5 endplates were included: nine of the remaining endplates were randomly selected to be used in a secondary study on endplate removal, six were too degenerated to test, and the seven sacral endplates used previously [8] were not included because there was an insufficient number to assess the effects of bone density and disc degeneration. Lateral DEXA scans were used to record the bone mineral density (BMD) of the superior and inferior halves of each vertebral body (LUNAR Corp., Madison, WI). Bags of long-grained white rice with a total height of 7 cm were placed beneath the spine segments to simulate soft tissue [21]. Once a specimen had been scanned, the vertebrae were separated. The discs were sectioned through the center using a 7 cm long knife to obtain a smooth cut surface. The top and bottom half of each disc were photographed using a digital camera (Nikon Coolpix 950, Nikon Corp., Tokyo, Japan] with a pixel count of 1600 x 1200. After the discs had been photographed, the endplates were cleaned using a scalpel to remove the disc and cartilage tissue, leaving the bony endplate exposed. The disc photographs were used to grade the level of disc degeneration using a four-point grading scheme [19]. Both disc halves were used together to grade the L3-L4, L 4 L 5 and L5-SI discs, and only the lower half was used to grade the L2-L3 disc. Three of the authors

graded the discs, following a training session designed to equalize the rankings. In no case did the ratings differ by more than one point. The disc grade used for the study was the majority ranking from among the three rankings. Due to the advanced age of most specimens, only one disc was given a grade of 1. For the purpose of analysis, this specimen was included in the grade 2 disc group, since a single specimen would preclude statistical analysis. The superior and inferior L3-L5 endplates of intact vertebrae were tested. The specimens were potted using plaster of Paris to hold the vertebra securely and distribute load evenly over the lower (supporting) endplate. The vertebra was wrapped in a thin sheet of plastic secured with an elastic band to prevent infiltration of the submerged endplate by the plaster. lndentution tests

The structural properties were found using indentation tests conducted with a custom-built uniaxial testing machine. A 3 mm diameter hemispherical indenter was pressed into the bone at 0.2 mmls to a depth of 3 mm by a stepper motor-controlled electromechanical linear actuator (Dynact Model I-PP3-B5, Orchard Park, NY). The motor position and load through the indenter (Omega Model LCIOI-50 load cell, Omega Engineering, Stamford, CT) were recorded by a personal computer at 35 Hz. The locations of the test sites were defined using the dimensions of the endplate in order t o standardize the results (Fig. 1). The anteriorposterior (AP) dimension was divided into increments of 20%, while the lateral lateral (LAT) dimension was divided into increments of 15%. In each specimen there was a minimum of 6 mm center-to-center between adjacent test sites and 4.5 mm between the centers of the outermost test sites and the edge of the endplate. Preliminary tests showed no interaction between tests separated by this distance. In most cases, 27 tests were done in each endplate, though the actual number varied based on the endplate shape. An q-translating table was used to move accurately between the test sites. The specimens were levelled prior to testing using a mounting device with adjustable legs. A tripod bubble level [7] placed in the center of the endplate was used to determine when the endplate was horizontal. Dutu unulysis

The failure load and stiffness values were extracted from the loaddisplacement curves generated at each test site. The failure load was defined as the maximum load reached prior to a load decrease of greater than 5%. The stiffness was the slope of the linear region of the load-displacement curve based on a linear regression analysis of the

100% LAT

Fig. 1. Test site layout based on endplate dimensions. Indentation tests were performed at the intersection points between the orthogonal lines. The shaded region identifies the area for which a complete AP-LAT statistical analysis was performed. A P is the anterior-posterior depth of the endplate and LAT is the lateral width of the endplate.

J.P. Grant ef al. I Journul of' Ortliopueilic Rc~seuvch20 (2002) 1115 -1120

data set. These values represent the structural properties of the bone, rather than tissue properties, due to the size of the indenter relative to the components of the vertebral body (endplate and trabeculae). The data were analyzed using three-way analyses of variance (ANOVA) with post-hoc Student-Newman-Keuls (SNK) tests at a 95% level of significance. The AP and LAT test co-ordinates were two of the [actors and were included as repeated measures so that each specimen acted as its own control. The third factor was either disc grade ( 2 4 ) or bone density group (low density with DEXA < 0.65 g/ cm2, mid-density, and high density with DEXA > 0.90 g/cm2) [8]. All 27 test sites could not be considered in a single ANOVA, since not every AP co-ordinate had a full complement of LAT co-ordinates, and vice versa. For this reason, three ANOVAs were done for each analinaximum AP, LAT analysis (-40'K to 20% AP, -30% to 30% LAT) to look at the largest number of sites at once, a full A P analysis (-40'K to 40'% AP, - 1 S'%, to 15'% LAT) to look at the full A P excursion, and a full LAT analysis (-200/0 to O'% AP, -45% to 45%)LAT) to look at the full lateral excursion. The table and graphs included here are from the maximum AP, LAT analyses. The results from the other ANOVAs were consistent with these analyses and were therefore not included. The reported p-values for map shape comparisons indicate whether o r not the property distribution differed in specimens with different bone densities or disc grades (statistically, this is a threeway interaction effect). The p-values for mean comparisons indicate whether the overall mean value of the failure load o r stiffness differed between specimens with different densities or disc grades (statistically, this is a main effect). Previous authors [10,16,21] have shown that bone density and disc degeneration may be interrelated. To check the interaction of these two variables in this study, a series of linear regressions were done, with disc degeneration as the independent variable and bone density as the dependent variable. All statistical analyses were performed using the STATISTICA 5.1 H statistics software package (Statsoft Inc., Tulsa, OK). The superior and inferior lumbar endplates were analyzed separately, since the structural property distributions were found to be significantly different in these subgroups [8].

400

1117

-

350

-t 300 V

g

250

-1

-f 200 ._ 2

5

150

? l 100

50

AP

-40% -20% 0% 20% High pO.90)

-40% -20% 0% Mid

20%

-40% -20% 0% 20% Low (<0.65)

Fig. 2. Effect of bone density on the superior lumbar endplate failure load map. There was a significant change in the map shape as the bone density decreased (I,= 0.005). The failure load decreased more rapidly posteriorly than anteriorly as the bone density dropped, resulting in a flatter A P failure load prolile. The overall Failure load also decreased as bone density dropped (r, < 0.0001).

(A)

Superior Endplate

Low (
Results The overall failure load magnitude decreased as the bone density dropped in both the superior (p < 0.0001) and inferior (p = 0.01 1) lumbar endplates. For all bone density levels, the posterolateral corners of the endplate were significantly stronger than the central region (Fig. 2). This difference became less pronounced in the superior lumbar endplates as the bone density decreased (Fig. 3A), while the strength decrease was relatively uniform across the inferior endplates (Fig. 3B). In general, the strength loss with decreasing bone density was uniform over the superior and inferior lumbar endplates. However, there was a greater loss posteriorly in the superior endplate that was not observed in the inferior endplate. Despite these changes with decreasing bone density, the strongest part of the endplate remained the posterolateral region in front of the pedicles in both the superior and inferior lumbar endplates. Statistically, increasing disc degeneration was associated with an overall loss of strength in the inferior lumbar endplates (p = 0.002), but not the superior lumbar endplates (p = 0.735). However, the shape of the strength distribution did change in the superior lumbar endplates (Fig. 4, p = 0.001); the central region showed decreased strength with disc degeneration, while the

0'

-40%

-20%

20%

0%

I

AP Position Inferior Endplate

(B)

I

350 I

-3 2

300 o....,

250

4Low (<0.65)

[email protected]

......_... -.e..

0

""

-40%

-20%

0%

20%

AP Position

Fig. 3 . Effect of bone density on the A P failure loads for both the superior (A, top) and the inferior (B, bottom) endplates. The overall strength magnitude decreased with lower bone density for both endplates. The change in the superior endplate map shape can be observed for the most posterior point (i.e., -40'%), while the inferior failure load map shape did not change significantly as the bone density decreased 67 = 0.901).

peripheral strength increased. A more general depiction of this trend for the superior endplate can be seen in

J . P . Grunt et ul. I Journul of Orthopuedic Reseurch 20 (2002) 1115-1120

1118 250

--z

+-30%

200

-.t3-. -15%

Table I Effects of bone density and disc degeneration on the failure load (FL) and stiffness (ST) map shapes and overall mean values in the superior and inferior lumbar endplates

-4 0%

~~

V

150 6 l

-m L

B

Superior lumbar Inferior lumbar Disc degeneration Superior lumbar Inferior lumbar

D.-.EI

50

0

AP

-40% -20%

0% Grade 2

20%

0% Grade 3

-40% -20%

20%

-40% -20%

0%

20%

Grade 4

Fig. 4. Effecl of disc degeneration on the superior endplate maximum AP. LAT failure load map. The map shape was significantly affected by disc degeneration (p = 0.001). The center of the endplate weakened, while the periphery remained strong, resulting in a less uniform failure load profile. The overall failure load did not change significantly with disc degeneration (p = 0.735).

(4

Superior Endplate

140 130.

0"

MeanFL

ST map shape

Mean ST

0.005 0.901

0.0001 0.011

0.764 0.967

0.084 0.632

0.001 0.453

0.735 0.002

0.230 0.868

0.391 0.183

Eot7e density

CI

; 100 m

-I

FLmap shape

..

100

2

-z U

c

.

90.

; :

a-......

.. I

---u.:

80-

7060-

a....... ......4.............d.' .-*-Grade 3 *-Grade 4

50 40

.

-30%

-15%

0%

f 5%

30%

LAT Position

(B)

Inferior Endplate

The p-values shown are for the maximum AP, LAT analyses. Significant p-values are given in bold, indicating that the map shape o r mean value of FL o r ST was significantly affected by bone density or disc degeneration.

strength was maintained much better than in the central regions. Again, for both the superior and inferior endplates, the locations of the strongest sites did not change with disc degeneration, remaining in the posterolateral region in front of the pedicles. The stiffness distributions as a function of both bone density and disc degeneration were similar in appearance to the corresponding failure load maps, but they did exhibit more variability. There were no significant effects of either bone density or disc degeneration on either the stiffness magnitude or the shape of the stiffness distributions (Table I). For the superior endplate, the stiffness magnitudes were nominally lower for lower bone density, but this trend was not significant (p = 0.084). The lack of a difference may have been due to insufficient power. The correlation coefficient between bone density and disc degeneration for all 55 specimens was T- = 0.088 (p = 0.53), indicating that there was no interaction between these variables. The 27 superior lumbar specimens gave r = 0.061 (p = 0.76), and the 28 inferior lumbar specimens gave Y = 0.13 0, = 0.51). In other words, the bone density and degree of disc degeneration were not related in these specimens.

Discussion 0 -30%

-15%

0%

15%

30%

LAT Position

Fig. 5. Effect of disc degeneration on the LAT failure loads for both the superior (A, top) and the inferior (B, bottom) endplates. For bath endplates, the central strength decreased with increasing disc degeneration. The peripheral strength did not change for the superior endplate (p = 0.735), but it did decrease slightly for the inferior ( p = 0.002).

It was hypothesized, based on the literature, that reduced bone density and disc degeneration would result in changes to the structural property distributions in the lumbar endplates. As expected, changes were observed in both the superior and inferior lumbar endplate properties, and each variable seemed to have a different effect.

Effects qf bone density Fig. 5A and the corresponding plot for the inferior endplate in Fig. 5B. Note the more uniform strength decrease in the inferior endplates, though the peripheral

Both the superior and inferior lumbar endplates showed a significant overall decrease in strength as the

J.P. Grunt et ul. i Jvurnul of’ Oi-thopcidic Rrseurrh 20 (2002) III.F-II20

bone density dropped. Generally, the decreases were relatively uniform across the endplate surfaces (Fig. 3A and B). This finding suggests that older patients, osteoporotic patients, or those with unusually low bone density are at higher risk of subsidence following interbody fusion, as supported by in vitro implant performance studies [13,22,24] and clinical observations. The sole exception to a uniform strength decrease with decreasing bone density was observed in the superior lumbar endplate map shapes, which demonstrated a more substantial decrease in the posterior failure loads than elsewhere on the endplate surface (Fig. 3A). The reasons for the differences between the superior and inferior endplates are unclear. We observed in our earlier study that the inferior endplate was stronger than the superior [8]. Recent histomorphometric data supports this observation, since the cancellous bone near the inferior endplate was found to be approximately 15% denser than that near the superior endplate [2]. Earlier work in a single specimen also suggests that the cancellous bone near the inferior endplate may be somewhat more dense than that near the superior endplate [15]. However, given that we do not understand these differences well, the nonuniform changes seen in the endplate properties with overall bone density are difficult to address. We suspect that the intersection with the pedicle in the superior vertebra may have an influence, but this is conjecture. If the structural support of the pedicles is substantial in the posterolateral region of the superior vertebral body, the posterior region of the superior endplate could sustain more profound changes in bone structure, such as those associated with aging and osteoporosis, than could the inferior region. Effects

1119

center of the vertebral body would weaken and the periphery would become stronger as the disc degenerates. The superior endplate strength maps shown here clearly support this hypothesis, and the general trends observed for the inferior endplate are also consistent with this hypothesis. Stifness

Stiffness was not significantly affected by either bone density or disc degeneration. This may be, in part, due to the fact that a hemispherical indenter was used, which would tend to introduce error into the stiffness measurement, since the indenter area in contact with the bone changes as the indentation depth increases. This would, however, be a consistent error in every test, as would any errors associated with equipment stiffness. Implications for implmt design

Despite the changes in the shape of the lumbar endplate failure load maps with bone density and disc degeneration, the locations of the strongest bone regions did not change. This implies that implant shapes developed using the basic structural property maps [8] will theoretically remain the best shapes for patients with a wide range of bone density and disc conditions. Nevertheless, since the strength of the central endplate drops with both decrease in bone density and increased disc degeneration, patients with these conditions will still be at higher risk of implant subsidence. Clearly, the ability to generalize these results may be limited by the average age of the specimens in this study compared to the age of patients undergoing interbody fusions.

of disc degeneration Limitations and .future research

The analysis of the effects of increasing disc degeneration may have appeared to produce conflicting results in the superior and inferior endplates, but a rather consistent pattern can be seen. In the superior lumbar endplates, the map shape changed significantly, with a decrease in the central endplate strength and a slight increase in the peripheral strength (Figs. 4 and 5A). In the inferior lumbar endplates, a more uniform loss of endplate strength was observed with increasing disc degeneration (Fig. 5B). However, the strength decreases were much greater centrally than on the periphery, which is the same trend seen in the superior endplate. Therefore, disc degeneration seems to generally lead to decreased central endplate strength with little strength change on the endplate periphery. Several researchers have demonstrated that as the disc degenerates, the center of the disc transmits progressively less of the load, and the periphery of the disc supports the additional load [6,19,23]. Current understanding of adaptive bone remodelling predicts that the

It should be noted that the specimens used in this study were taken from the lower lumbar spine (L3-L5), thus the results cannot necessarily be extrapolated to other vertebrae. Also, the results given in this paper are mean values for a large number of specimens with considerable variation between specimens. The results are structural properties of the vertebral body, not tissue properties of the bone. Ongoing work will look at the effects of removing the endplate on the structural properties of the vertebral surfaces.

Acknowledgements The authors would like to acknowledge the contributions of Alston Bonamis and Jesse Chen for their assistance with dissection and photography and Darrell Goertzen for his technical advice. Financial support from the George W. Bagby Research Fund and the

1120

J. P. Griinr er ul. I Jourticrl of’ Ortliopacdic Ri>.wurcIi20 (2002) 1115-1120

Natural Sciences and Engineering Research Council of Canada (NSERC) is gratefully acknowledged. References [I] Atkinson PJ. Variation in trabecular structure of vertebrae with age. Calcif Tissue Res 1967;1(1):24-32. [2J Banse X. Devogelaer JP, Delloye C, Cornu 0, Munting E. Grynpas M . Vertical inhoinogenity of the density within the human vertebral body. In: Trans 47th Annual Meeting ORS, San Francisco, CA. 2001. p. 930. [3] Bell G H , Dunbar 0, Beck JS, Gibb A. Variations in strength of vertebrae with age and their relation to osteoporosis. Calcif Tissue Res 1967;1(1):75-86. [4] Biggemann M, Hilweg D, Brinckmann P. Prediction of the compressive strength of vertebral bodies of the lumbar spine by quantitative computed tomography. Skeletal Radio1 1988;17(4): 264--9. [5] Cassucio C. An introduction to the study of osteoporosis. Proc Roy Soc Med 1962:55:663-8. [6] Frei HP. Oxland TR, Rathonyi GC. Nolte LP. The effect of nucleotomy on lumbar spine mechanics in compression and shear loading. Spine 2001;26( 19):208O-9. [7] Grant JP. Mapping the structural properties of the lumbosacral vertebral endplates. Master of Applied Science, University of British Columbia, Vancouver, BC, Canada, November 2000. [8] Grant JP, Oxland TR, Dvorak M F . Mapping the structural properties of the lumbosacral vertebral endplates. Spine 2001 ;26(8): 889-96. [9] Hansson T, Roos B, Nachemson A. The bone mineral content and ultimate compressive strength of lumbar vertebrae. Spine 1980; 5( I ):46-55. [lo] Harada A, Okuizumi H, Miyagi N. Genda E. Correlation between bone mineral density and intervertebral disc degeneration. Spine 1998;23(8):857-62. [ I l l Horst M, Brinckmann P. 1980 Volvo Award in biomechanics. Measurement of the distribution of axial stress on the endplate of the vertebral body. Spine 1981;6(3):217 32. [ I 21 Jayasinghe JAP. Jones SJ, Boyde A. Three-dimensional photographic study of cancellous bone in human fourth lumbar vertebral bodies. Anat Embryo1 (Berl) 1994;I89(3):259--74. [I31 Jost B. Cripton PA, Lund T, Oxland TR, Lippuner K, Jaeger P. ct al. Compressive strength of interbody cages in the lumbar spine:

the effect of cage shape, posterior instrumentation and bone density. Eur Spine J 1998;7(2):13241. [I41 Keller TS, Hansson TH, Abram AC, Spengler DM. Panjabi MM. Regional variations in the compressive properties of lumbar vertebral trabeculae. Effects of disc degeneration. Spine 1989: 14(9): 1012-9. [I51 Keller TS, Moeljanto E. Main JA, Spengler D M . Distribution and orientation of bone in the human lumbar vertebral centrum. J Spin Dis 1 9 9 2 3l):60- 74. [I61 Margulies JY, Payzer A, Nyska M , Neuwirth MG. Floman Y, Robin GC. The relationship between degenerative changes and osteoporosis in the lumbar spine. Clin Orthop 1996;(324):145-52. [I71 McNally DS, Adams MA. Internal intervertebral disc mechanics as revealed by stress profilometry. Spine 1992:17(1):66 73. [I81 Mosekilde L. Age-related changes in vertebral trabecular bone architecture assessed by a new method. Bone 1988:9(4):247-SO. [ 191 Nachemson A. Lumbar intradiscal pressure: Experimental studies on post-mortem material. Acta Orthop Scand Suppl 1960;43: 21- 10s. [20] Oda K, Shibayama Y , Abe M, Onomura T. Morphogenesis of vertebral deformities in involutional osteoporosis: age-related, three-dimensional trabecular structure. Spine 1998;23(9):IOS&6. [21] Oxland TR. Lund T, Jost B, Cripton P, Lippuner K. Jaeger P, et al. The relative importance of vertebral bone density and disc degeneration in spinal flexibility and interbody implant performance. An in vitro study. Spine 1996;21(22):2558-69. [22] Oxland TR, Lund T. Biomechanics of stand-alone cages and cages in combination with posterior fixation: a literature review. Eur Spine J 2000:9(Suppl I):S95 -101. [23] Shirazi-Ad1 SA, Shrivastava SC, Ahmed AM. Stress analysis of the lumbar disc-body unit in compression. A three-dimensional nonlinear finite element study. Spine 1984:9(2):120-34. [24] Steffen T, Tsantrizos A, Aebi M . Effect of implant design and endplate preparation on the compressive strength of interbody fusion constructs. Spine 2000:25(9): 1077-84. [25] Twomey L, Taylor J. Furniss B. Age changes in the bone density and structure of the lumbar vertebral column. J Anat 1983;136(1): 15-25. 1261 Twomey LT, Taylor JR. Age changes in lumbar vertebrae and intervertebral discs. Clin Orthop Relat Res 1987;224:97-104. [27] Vesterby A, Mosekilde L. Gunderson HJ, Melsen F, Mosekilde L, Holme K, et al. Biologically meaningful determinants of the in vitro strength of lumbar vertebrae. Bone 1991;12(3):219-24. [28] Wolff J. Das Gesetz der Transformation der knochen. Berlin: Hirschwald: 1892.