Available online at www.sciencedirect.com
Clinical Biomechanics 23 (2008) 859–869
www.elsevier.com/locate/clinbiomech
Stress distribution in the intervertebral disc correlates with
strength distribution in subdiscal trabecular bone
in the porcine lumbar spine
Garrett Ryan a,b,*, Abhay Pandit a,b, Dimitrios Apatsidis a
a
Department of Mechanical and Biomedical Engineering, National University of Ireland, Galway, Ireland
b
National Centre for Biomedical Engineering Science, National University of Ireland, Galway, Ireland
Received 18 June 2007; accepted 20 March 2008
Abstract
Background. It is understood that an interdependence of properties exists between the intervertebral disc and the subdiscal trabecular
bone. Determining the biomechanics of this relationship is important in the development of novel spinal implants and instruments. The
aim of this study was to analyze this relationship for the porcine lumbar spine and to compare it with that of the human spine.
Methods. The stress distribution within the intervertebral disc of 10 porcine lumbar (L4/L5) motion segments was recorded using a
1.5 mm needle pressure transducer. For dynamic loading a specialized testing rig was developed to apply flexion/extension and medial/
lateral bending while intervertebral disc stress was simultaneously recorded. The regional variation in mechanical properties of trabecular
bone was also examined for an additional 10 porcine (L5) vertebral bodies. For compressive testing of the subdiscal bone, columns were
prepared using a low speed cutting saw and subjected to axial compression.
Findings. Under pure compressive loading, stress levels within the intervertebral disc were relatively uniform. However, during asymmetric loading large peak stresses were evident in the periphery of the intervertebral disc in areas underlying the annulus fibrosus. The
mechanical properties of trabecular bone demonstrated regional variations within the vertebral body. The ratio of compressive yield
strength of bone underlying the outer annulus fibrosus to that of bone underlying the nucleus pulposus averaged 1.36.
Interpretation. Although the effects of stress distribution and bone mass adaptation cannot be directly related, it is probable that peak
stresses arising in the annulus fibrosus during asymmetric loading provide greater stimuli for the underlying bone to undergo adaptive
remodeling to withstand the greater forces experienced. Findings of intervertebral stress distribution and strength distribution of subdiscal trabecular bone for the porcine spine may aid in the development of strategies for preclinical animal testing of spinal implants.
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: Biomechanics; Intervertebral disc; Trabecular bone; Adaptive remodeling; Porcine
1. Introduction
Animal spines are widely used in the biomechanical
analysis of prospective spinal instruments, due to the low
availability and high cost of human tissues. For practical
reasons, bovine and porcine spines, especially the latter,
*
Corresponding author. Present address: CIMRU Building, Nuns
Island, National University of Ireland, Galway, Ireland.
E-mail address: [email protected] (G. Ryan).
0268-0033/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.clinbiomech.2008.03.066
are most popularly selected as the model for both in vitro
and in vivo studies (Kaigle et al., 1997; Keller et al., 1990;
Pfeiffer et al., 1994; Teo et al., 2006). Interpretation of data
from such studies usually considers the limitations related
to the mismatch of properties between the human and animal
tissues. Although the physical and mechanical response of
the human lumbar spine have been widely studied (White
and Panjabi, 1990), less efforts have been made in comparing its properties with those of porcine spines, despite
ongoing clinical studies using these animals. Understanding the mechanism of load transfer through the various
860
G. Ryan et al. / Clinical Biomechanics 23 (2008) 859–869
tissues of the porcine spine is important in the testing and
validation of new spinal implants and in the interpretation
of mechanical data from studies using these animals.
The porcine intervertebral disc (IVD) is constructed, as
in humans, of a central gel-like nucleus pulposus (NP) surrounded by a peripheral firm annulus fibrosus (AF). Since
it is essentially composed of two parts, it is important to
determine how each of these structures contributes to the
IVDs load bearing capacity and resulting redistribution
of load to subdiscal trabecular bone. In studying the
human spine, researchers have shown that several factors
such as posture (Adams et al., 2000a,b; Edwards et al.,
2001; Steffen et al., 1998), muscle tone (Wilke et al.,
1996a,b), spinal load (McNally and Adams, 1992; Ranu,
1990), and time (Adams et al., 1996a,b) are significant in
influencing the distribution of stress within the IVD. With
regards the distribution of strength within the vertebral
body, research has aimed towards better prediction of osteoporosis and related fractures. In these efforts, a number of
researchers have shown an uneven distribution in mechanical and morphological properties of trabecular bone inside
the vertebral body of human and quadrupedal species
(Gong et al., 2006; Keller et al., 1989; Lin et al., 1997). Keller et al. (1989) found that the mechanical properties
(strength, stiffness) and density of subdiscal trabecular
bone specimens prepared from 12 regions within the vertebral body were highest in regions underlying the NP and
lowest in regions underlying the AF. Similarly, a study
by Lin et al. (1997) analyzing the regional strength distribution of trabecular bone in the porcine lumbar vertebral
body found that the average ultimate strength of the bone
underlying the NP was consistently higher than the bone
underlying the AF. Recently, Gong et al. (2006) investigated the three-dimensional microstructural properties of
L4 vertebral bodies using micro-computed tomography
(lCT). Specimens were divided into two groups according
to the average structure model index (SMI) of the 15 specimens inside the vertebral body. Trabecular specimens with
lower mass were liable to form a high-SMI group and it
was found that the anterior column (underlying the AF)
in this group was more susceptible to vertebral body wedge
fracture. However, in the low-SMI group, off-axis bone
damage was found to be most harmful to the central column of vertebral trabeculae, which underly the NP.
In light of Wolff’s law that stress-morphology relationships exist for most biological structures (Wolff, 1884), load
transfer through the IVD and subdiscal vertebral bone
properties should be interrelated. This is apparent in a
recent study by Adams et al. (2006), who showed that
IVD degeneration was associated with reduced loading of
the anterior vertebral body in upright postures. Removal
of stimulus to the bone underlying the anterior IVD permitted negatively balanced remodeling and corresponded
with reduced bone mineral density (BMD) and inferior trabecular architecture in this region. This predisposed the
anterior vertebral bone to compressive fracture when the
spine was flexed.
For the human and quadrupedal spine, it is evident that
an interdependence of properties exists between the IVD
and subdiscal trabecular bone. Accordingly, the objectives
of this research were twofold: firstly, to examine the stress
distribution in the IVD of porcine lumbar functional spine
units or ‘motion segments’ under physiological activities
using multi-localized pressure measurements, and secondly,
to analyze the regional distribution of compressive properties of the subdiscal trabecular bone of porcine lumbar vertebral bodies. Results are compared with those of human
spines and the effect of anatomical and age-related variation is addressed.
2. Methods
Twenty fresh porcine lumbar spines were harvested from
8-month-old landrace pigs weighing 90–100 kg immediately after sacrificing. The spines, with attached muscles
and ligaments, were vacuum sealed and kept in frozen storage at 20 °C until testing. Before testing, each specimen
was thawed to room temperature overnight and then
wrapped in normal saline soaked cloth for 4–5 h. Separate
spines were used for each of the two studies as outlined
below.
2.1. Measurement of stress distribution in the IVD
A pressure probe similar to that described by McNally
and Adams (1992) was developed for use in this study.
The probe consisted of a miniature pressure transducer
embedded in a 1.5 mm diameter stainless steel needle
(Gaeltec Ltd., Isle of Skye, Scotland). The transducer diaphragm was rectangular in shape and measured
2.5 mm 1 mm. The needle had a sharp pointed tip to
facilitate easy insertion into the IVD and had been graduated by the authors with laser markings so that the depth of
insertion could be gauged. The probe was calibrated using
a sealed pneumatic chamber (Instrument Technology Ltd.,
Dunboyne, Co., Meath, Ireland) and was found to have a
linear ramp up to 4 MPa without damage. A testing rig was
developed to introduce controlled flexion/extension and
medial/lateral bending to the porcine motion segment,
while data from the needle pressure transducer could be
simultaneously recorded (Fig. 2). The testing rig was
designed for mounting with the Instron 8874Ò servohydraulic testing machine (Instron Corp., Norwood, MA,
USA), which can apply axial and torsional loads to the
motion segment.
Ten L4–L5 motion segments were harvested from the
porcine lumbar spines. Care was taken to preserve the ligaments while other soft tissues were completely removed.
The major coronal and sagittal diameters of the L3/L4
and L5/L6 IVD’s were recorded. An average of these values was used to approximate the size of the L4/L5 IVD.
This data was later used to determine the placement of
the pressure transducer probe in predefined locations
within the IVD according to the arrangement in Fig. 1.
G. Ryan et al. / Clinical Biomechanics 23 (2008) 859–869
0.85
L 0.9
D
0.6
L 0.2
D
L 0.4
L 0.6
L
D
Fig. 1. Schematic showing pressure measurement points in the IVD.
The intersection of the major coronal and sagittal diameters was taken as the centre of rotations of each motion
segment based on these average values. The motion segments were secured with bone screws and potted using
low melting point alloy (Tm: 40 °C). A dowel pin was positioned in the centre of the cementing pot to hold the
motion segment in position while it was being secured.
The testing rig consisted of a number of tiers, separated
by large deep-groove ball bearings as illustrated in Fig. 2.
The potted motion segment was attached to the upper tier,
which was essentially a geared platform. The centre of rotation lay 15 mm from the base of the cementing pot. The NP
was placed at the centre of rotation of the geared platform
with careful reaming of the L5 vertebral body so that the
base of the vertebral body was 15 mm from the centre of
the IVD. Also, an aligning hole for the central dowel pin
was drilled into the vertebral body directly below the centre
of the NP. This was taken from previous measurements of
the L3/L4 and L5/L6 IVD’s. The geared platform revolved
to introduce varying degrees of flexion/extension and medial/lateral bending to the motion segment. The bearings
allowed the entire platform to be rotated to alternate
between flexion/extension and medial/lateral bending without disruption of the motion segment.
Using a 1mm diameter needle, incisions were made into
the IVD prior to introducing the needle pressure probe in
order to limit the risk of damaging the transducer diaphragm. The needle pressure probe was gradually
advanced into the IVD with the transducer diaphragm facing upward so that only axial stress was recorded. A clamp
held the needle in position while measurements were carried out. For positioning in the posterior AF the needle
was first pushed through the IVD, then retracted slowly
until a sharp increase in stress was noted. The needle’s position was then confirmed on the basis of the earlier standard
measurements. Initially, stress data was recorded at each
measurement location, with incremental axial compressive
loads of 200 N up to 800 N, applied to the motion segment.
The axial loads adopted in this study were chosen based on
previous studies using porcine spines (Park et al., 2005; van
Deursen et al., 2001). Subsequently, a sequence of movements was executed with the testing rig. This involved flexion/extension from 1° to 4° followed by medial and lateral
bending from 1° to 4°. The range of motion was adopted
L4/L5 Spinal
Motion Segment &
Intervertebral Disc
Needle Pressure
Transducer
Upper Tier
(Attaches To Servohydraulic
Testing Machine)
Stepper
Motor
Internal Gear
And Pinion
861
Universal Joint
Assembly
Middle Tier
(Introduces Flex./Ex.
& Lat. Bending)
Deep Groove
Ball Bearings
Stepper
Motor
Bottom Tier
(Rotates Middle Tier
Between Flex/Ex.
& Lat. Bending)
X-Y Table
Fig. 2. Schematic showing spinal testing apparatus and relevant components.
862
G. Ryan et al. / Clinical Biomechanics 23 (2008) 859–869
from previous research involving ovine and bovine spines
(Wilke et al., 1996a,b, 1997), as no data was available for
the range of motion of porcine spines. The rig’s stepper
motors were employed to introduce the movements at a
rate of 1°/s. We developed a software programme using
LabViewÒ (National Instruments Corp., Austin, TX,
USA) to synchronize control of the rig’s two stepper
motors and the data recording process. Data was collected
using a USB 6008Ò data acquisition device (National
Instruments Corp.). Throughout this movement sequence
the axial load was maintained at 300 N. Care was taken
to keep the motion segment hydrated at all times by continuously spraying with saline solution. Following testing, the
L4/L5 motion segment was isolated and cut along the
transverse plane so that it’s cross sectional area (CSA)
could be measured. This was performed by placing a transparent plastic grid over the IVD and sketching the outline
onto the plastic. By counting the squares of the grid that
fell within the outline, the CSA could be determined. The
outer perimeter of the NP was defined by the presence of
laminae of fibrous AF tissue. By outlining this boundary
on the transparent plastic grid the CSA of the NP was also
determined.
2.2. Measurements of strength distribution in subdiscal
trabecular bone
Ten L5 vertebral bodies were harvested from the porcine
spines. The vertebrae had all soft tissues and bony processes removed using a bone saw. A low speed cutting
saw (Buehler Ltd., Lake Bluff, IL, USA) with specially
designed clamping fixtures was used to section each vertebral body into 5 mm 5 mm 12 mm cubes from predetermined anatomical regions as portrayed in Fig. 3. In
total, 16 rectangular columns of trabecular bone were prepared from each vertebral body. The dimensions of each
specimen were measured using a precision micrometer.
The long axis of each specimen was kept parallel to the lonMid-Coronal Plane
D E F
A B C
G
H
Mid-Sagittal
Plane
Fig. 3. Schematic showing locations of trabecular bone columns sectioned
from the L5 porcine vertebral body.
gitudinal axis of the spinal canal. Six of these pieces (A–F)
were cut from trabecular bone along the mid-coronal
plane, 4 mm from the posterior wall of the vertebral body.
The remaining two (G and H) were cut along the mid-sagittal plane. The middle two specimens of the six in the midcoronal plane (C and D) were chosen as being strictly
beneath the NP. This particular configuration and sample
size was chosen to maximize the number of samples that
could be cut from the vertebral body while care was taken
to avoid any cortical bone in the outermost samples. This
was governed by the size and shape of the smallest vertebral body. Uniaxial compression was carried out using a
Zwick universal testing machine (Zwick GmbH, Ulm, Germany) at a deformation rate of 5 mm/min along the long
axis of the specimen. A video extensometer (Messphysik
Materials Testing Ltd., Furstenfeld, Austria) was used to
gauge specimen displacement. Load–displacement curves
were recorded for later data analysis. Each specimen was
loaded until a decrease in the load was observed, at which
point the load was removed. The yield stress and strain
were calculated using a 0.2% offset. The material parameters, including elastic modulus (E), yield stress (ryield) and
ultimate compressive stress (rucs), were obtained from calculations using the load–displacement curves.
2.3. Statistical analysis
For both studies, statistical analyses were carried out
using statistical software (MinitabTM, v.13.32). Statistical
variances between regions were determined by one-way
analysis of variance (ANOVA). Tukey’s honesty significant
difference test was used for post hoc evaluation of differences between groups. A P-value of <0.05 was considered
to be statistically significant.
3. Results
4.1. Stress distribution in the IVD
The mean CSA of the 10 porcine IVDs was 8.68 cm2
(SD 0.24) and the average IVD height was 5.7 mm (SD
0.2). The porcine endplate revealed a slightly concave curvature shifted posteriorly. The stationary placement of the
pressure probe provided identical measurement sites so
that stress readings obtained under different loading regimens could be compared. The axial stress recorded in both
the AF and NP displayed a closely linear relationship to
the applied load. The average intrinsic stress on transducer
insertion for all measurement sites was 288.7 kPa (SD
70.5). Thereafter, the mean ratio of stress measured to
the theoretical IVD stress was 2.13 (SD 0.43) where theoretical IVD stress is simply the axial force applied divided
by the CSA of the IVD. Stresses measured in the NP were
on average, slightly lower than those measured in the AF.
The mean ratio of NP stress to the theoretical stress was
1.95 (SD 0.19) over the complete loading range. Fig. 4 illustrates the relationship between stress and applied axial load
G. Ryan et al. / Clinical Biomechanics 23 (2008) 859–869
863
Fig. 4. Stress across the mid-coronal and mid-saggital planes due to increasing axial load (values reported as mean (+SD) n = 10).
at the various positions across the IVD under static compression. In the mid-coronal plane and under no axial load,
the stress measured across the IVD was lowest in the outer
AF (Positions 1 and 5). Values of stress in the NP and inner
AF (Positions 2 and 4) were slightly greater and almost
identical. With rising axial load the stress became distributed evenly over the area of the IVD and there is no statistically important difference between the stress at any
location across the mid-coronal or mid-sagittal planes
above 400 N compression (analysis of variance [ANOVA],
P > 0.05). At 800 N the average IVD stress measured
1611.2 kPa (SD 198.4). Increasing axial load coincided with
a slight bulging in the annular part of the IVD.
Plots of typical stress profiles during dynamic loading
are shown in Fig. 5. Also shown is the sequence of movements produced by the testing rig. It is evident that as
the testing rig introduces flexion/extension and medial/lateral bending, stresses in the AF rise rapidly when the movement is directed towards the sensor location. Peak stress
values always occurred at highest angles of flexion/exten-
sion and medial/lateral bending. Stress values below those
experienced during pure compressive loading were often
noted in the AF when flexion/extension or medial/lateral
bending was directed away from the sensor location. For
these instances, it may be that the annular tissue lost slight
contact with the sensor. A drop in stress was very uncommon in the NP. There was no statistical difference between
peak stresses measured for left and right AF positions during ipsilateral and contralateral bending. These findings
underline the reproducibility of the technique. Therefore,
peak stresses obtained for Positions 1 and 2 were pooled
with those of Positions 5 and 4, respectively for data analysis. For the posterior AF, stress peaks were greatest during extension. Similarly, for the anterior AF, stress peaks
were greatest during flexion. However, no statistical difference was evident between stress peaks in the posterior and
anterior AF. In all specimens, greatest peak stresses were
observed in the AF as opposed to the NP (P < 0.05). The
peak stresses at each measurement location for the movement sequence are presented in Figs. 6 and 7.
Fig. 5. An example of three stress history profiles at various locations within the IVD during dynamic loading of the porcine motion segment. Also shown
is the change in angle produced by the testing apparatus, beginning with flexion/extension and followed by lateral bending.
864
G. Ryan et al. / Clinical Biomechanics 23 (2008) 859–869
Fig. 6. Regional peak stress values across the mid-coronal plane recorded during dynamic loading (values reported as mean (+SD) n = 10).
Fig. 7. Regional peak stress values across the mid-sagittal plane recorded during dynamic loading (values reported as mean (+SD) n = 10).
4.2. Strength distribution in subdiscal trabecular bone
The constant deformation conditions of the uniaxial
compressive tests resulted in a mean strain rate of
0.008 s 1 (SD 0.001). The typical load-deformation
response involved an initial nonlinear region followed by
a linear, elastic region until yield. Mean compressive yield
strain was 2.7% (SD 0.7). Loading continued until failure,
defined as the first detectable decrease in load. At failure,
the mean compressive ultimate strain was 4.2% (SD 1.3)
or 1.5% strain post-yield. The mean values of material
parameters were: compressive elastic modulus (E):
394.4 MPa (SD 183.2); compressive yield stress (ryield):
11.9 MPa (SD 4.2), and ultimate compressive stress (rucs):
12.8 MPa (SD 4.0). Results obtained for all positions are
displayed in Table 1. No macroscopic changes in the specimen appearance were evident, and the microscopic
appearance of the specimens was not examined. No significant differences were observed for material parameters (E,
rult, ryield) between the cranial and caudal regions in the
vertebral body so these values were pooled for statistical
significance. Variations in the strength of trabecular bone
in relation to anatomical position were, however, observed
within each region for all eight vertebral segments that
were examined. Statistical analysis revealed that the variations in the material parameters were significant within the
eight positions studied (P < 0.05).
In order to assess the differences in specimen strength in
these regions, the data was separated into the six segments
lying along the mid-coronal plane and four segments lying
along the mid-sagittal plane. An average of the two central
specimens was taken, as these underlie the NP and showed
almost identical results. For further statistical analysis the
segments were grouped according to their symmetry and
their anatomical position into the following sub-groups;
specimens A and F, specimens B and E, specimens C and
D, and specimens G and H. All material properties tended
to be greater on the periphery compared to the central
regions. Differences between sub-groups were determined
using ANOVA and the corresponding P-value for each
material property is reported in Table 2. The most striking
differences with respect to anatomical position were found
G. Ryan et al. / Clinical Biomechanics 23 (2008) 859–869
865
Table 1
Compressive properties for trabecular bone columns taken from distinct locations within the porcine vertebral body (values reported as mean (SD) n = 10)
Property
Cranial position
A
B
C
D
E
F
G
H
Young’s modulus (E) (MPa)
Yield strength (ryield) (MPa)
Ultimate compressive
strength (rucs) (MPa)
437.8 (199.3)
13.8 (4.6)
14.9 (3.3)
267.4 (129.5)
10.4 (3.7)
10.5 (3.7)
368.1 (144.7)
10.0 (2.4)
10.7 (2.2)
333.2 (103.2)
10.7 (2.6)
11.3 (2.6)
268.8 (170.8)
9.1 (3.4)
9.8 (3.3)
457.1 (148.5)
12.9 (3.7)
14.7 (3.6)
467.1 (239.8)
14.1 (5.8)
16.8 (6.6)
408.8 (159.4)
13.9 (3.8)
15.7 (3.5)
Caudal position
Young’s modulus (MPa)
Yield strength (ryield) (MPa)
Ultimate compressive
strength (rucs) (MPa)
A
B
C
D
E
F
G
H
488.9 (173.7)
15.5 (4.8)
15.2 (2.7)
299.8 (191.9)
9.1 (2.6)
9.6 (2.3)
337.1 (138.5)
10.4 (2.3)
10.6 (1.4)
295.3 (166.2)
9.6 (2.7)
10.8 (1.9)
342.6 (218.8)
9.1 (4.4)
9.7 (2.7)
399.2 (186.4)
13.2 (4.7)
14.3 (3.5)
544.6 (236.0)
13.1 (4.5)
13.9 (2.9)
400.7 (115.4)
14.6 (5.2)
16.1 (3.7)
Table 2
Statistical variances between trabecular bone column sub-groups as determined by one-way analysis of variance (ANOVA)
Young’s modulus (E)
ANOVA: P < 0.001
Tukey’s HSD test
B and E
C and D
G and H
*
Yield stress (ry)
ANOVA: P < 0.0001
Tukey’s HSD test
A and F
B and E
C and D
*
n/a
n/a
n/a
*
*
B and E
C and D
G and H
Ultimate compressive stress (ru)
ANOVA: P < 0.0001
Tukey’s HSD test
A and F
B and E
C and D
*
n/a
n/a
n/a
*
*
*
B and E
C and D
G and H
A and F
B and E
C and D
*
n/a
n/a
n/a
*
*
*
P < 0.05.
for compressive strength with the outer regions, underlying
the AF, tending to be stronger than central regions underlying the AF (P < 0.05).
5. Discussion
According to Wolff’s law the mechanical properties of
the vertebral trabecular bone should reflect the distribution
of stresses that exist in the IVD, which correlates with findings of this research. During everyday activities, the porcine spine inevitably undergoes certain flexion/extension
and medial/lateral bending. As shown in this study, peak
stresses that occur during these physical activities would
provide greater stimuli for the underlying bone to undergo
adaptive remodeling as to withstand the greater forces
present. The trabecular network of the load bearing vertebral bodies, constructed with thick plates and columns in
vertical direction and thin horizontal struts, is formed on
the basis of these forces (Frost, 1983). A drawback of the
study is that motion segments from different spines were
used for measuring the IVD stress and the strength of subdiscal trabecular bone. This was necessary as the vertebral
bodies of the motion segments used in the stress distribution study needed to be cut in order to properly locate
the NP at the centre of rotation of the testing rig. Therefore, bone from these vertebral bodies could not be sectioned and mechanically tested. However, because
interspecimen variation in age, weight, and size of porcine
motion segments should be low, we suggest that of strength
distribution of vertebral trabecular bone between samples
would be consistent.
From the mechanical point of view, a quadruped is a
complex system. There are different types of joints, and
the distribution of the muscles and their activity during
movements are largely unknown. Due to dominant muscle
and ligament forces, the quadrupedal spine is mainly
loaded under axial compression similar to the human spine,
although it is almost impossible to determine the actual
loads in the living systems (Smit, 2002). In examining the
human motion segments in vitro using similar techniques
to the present study, axial loads ranging from 600 N to
2 kN have been employed (Pollintine et al., 2004; Steffen
et al., 1998). However, because the width of the porcine
L5 endplate is approximately 0.8 times smaller than the
human equivalent (Dath et al., 2007; Panjabi et al.,
1992), and the average animal weight is approximately
95 kg, we estimated that the loading range (0–800 N) tested
in this study provided an appropriate estimate for the axial
loading experienced in daily life. During the different forms
of quadrupedal locomotion, axial torsion, flexion-extension, and lateral bending moments are important loads that
work on the spine. It has been demonstrated that the range
of motion of lumbar segments of ovine and bovine spines is
similar to that of human spines (Wilke et al., 1997,
1996a,b). For the ovine spine, the range of motion for
the L4–L5 motion segment was approximately 5° in extension, 4° in flexion and 4.5° in lateral bending (Wilke et al.,
1997). For the bovine spine, the range of motion in flexion
866
G. Ryan et al. / Clinical Biomechanics 23 (2008) 859–869
in the lumbar levels ranged from an average of 2.8° for L1–
L2 to 5.0° for L5–L6, with a similar range of motion for
extension (Wilke et al., 1996a,b). In this study, we
employed a maximum angle of 4o for flexion/extension
and lateral bending movements and on inspection found
no visible damage to the tissues, which suggests that the
motion segments remained within their physiological range
of motion. During these movements we employed an axial
compressive load of 300 N. This load was found to be an
appropriate level in previous biomechanical studies using
porcine spines (Keller et al., 1990). Although it is smaller
than the maximum force that was employed during pure
axial compression, far greater localised stresses were generated in the IVD during asymmetric loading.
The initial application of the guiding needle was necessary to facilitate an easier entry of the needle transducer
and to prevent artifacts caused by contact stresses. The
height of the porcine IVD was found to be approximately
5.7 mm, which is significantly smaller than that of the
human IVD which was found to be approximately
10.8 mm for healthy young working males (Nissan and
Gilad, 1986). We found that the introduction of the needle
transducer into the IVD created an almost equal rise in
pressure in both the NP and AF. A comparison of the estimated mean stress across the IVD to the stresses in the NP
and AF provides insight for interpreting the stress distribution across the IVD. The net stress passed from the annulus
to the endplate is the combined effect of the tension of the
annulus fibers and the stress within the matrix material.
Brinckmann and Grootenboer (1991) obtained measurements from isolated human IVDs (with posterior elements
removed) and found that the pressure in the NP was on
average 1.68 times the estimated mean stress over a range
of compression from 300 to 2000 N. For the porcine spine
we found a slightly higher ratio of approximately 2.13. This
is in fact an underestimation of the ratio as we did not take
into account the load resisted by the neural arch. In the
human spine the NP comprises about 30% the IVD area
(Brinckmann and Grootenboer, 1991). We found that the
NP in the porcine spine consisted of approximately 23%
of the CSA, and therefore would experience increased tension from the annular fibers compared to the human spine.
We observed an almost even distribution of stress across
the porcine IVD during pure compressive loading while the
spinal motion segment maintained an upright posture. In
comparison with the human spine, this finding is in contrast with early research by Nachemson and Morris
(1964), who suggested that under pure compressive loading
conditions, vertical stresses in the AF are half those in the
NP. More recent studies however, are in good agreement
with our findings and show that there is a one-to-one relationship between the stress in the NP and the AF. Horst
and Brinckmann (1981) analyzed the distribution of axial
stress under the IVD of human lumbar and thoracic
motion segments with the aid of five miniature piezoelectric
pressure transducers placed just below the vertebral endplate. Results indicated that the stress distribution depends
essentially on the state of degeneration of the IVD and on
the relative position of the adjacent endplates. Ranu (1990)
also found a one-to-one relationship between the two areas
of the IVD, by placing two different pressure transducers in
the AF and NP of cadaveric lumbar motion segments
which were then loaded to failure. Similar results were
reported by McNally and Adams (1992), using a technique
they pioneered and is known as ‘stress profilometry’, in
which a needle pressure transducer is pulled incrementally
across the mid-sagittal plane of the IVD.
During asymmetrical loading, we found that peak stresses predominantly occurred in the AF. We found that pressure within the nucleus was relatively insensitive to flexion/
extension and lateral bending. On occasion, a rise in stress
within the NP was noted. However, these stresses were not
to the same level as those experienced in the AF. Some needle bending due to asymmetrical loading cannot be
excluded. However, visual inspection of the pressure signal
during manual needle bending showed only marginal
effects of bending. Our findings are in good agreement with
several studies analyzing IVD stress distribution due to
asymmetrical loading in the human spine. Ongoing
research by Adams and co-workers has revealed the presence of peak stresses in the posterior AF due to IVD degeneration (Adams et al., 1996a,b, 2000a,b), after creep
loading (Adams et al., 1996a,b), and due to minor damage
of the IVD (Adams et al., 2000a,b). Steffen et al. (1998) performed simultaneous multilocalised IVD pressure measurements in human cadaveric lumbar spines to gather
information on stress distribution during asymmetrical
loading. For all loading conditions, stress gradients were
directed towards the NP. Peak stresses were recorded in
the posterolateral inner AF, during flexion and when combined with axial rotation.
Several of the aforementioned studies predominantly
show peak stresses arising in the posterior AF. In this
study, we found no statistical difference between pressures
recorded in the posterior and anterior AF. This apparent
contradiction with previous research may be explained by
the inherent differences between the anatomical features
of the porcine vertebral spine and the human spine. Compared to the human lumbar spine, there are a number of
anatomical differences in the porcine lumbar spine, including wider pedicles, longer body height and especially the
interlocking facet joints (Cotterill et al., 1986). It seems
plausible that the strong interlocking relationship of the
facet joints performs as a mechanical hinge and in doing
so acts to stress shield the posterior AF during extension.
The effect to which the vertebral endplate alters the load
transfer process is important in analyzing the relationship
between IVD stress distribution and subdiscal trabecular
bone strength. The vertebral endplate is a thin layer of
dense, subchondral bone adjacent to the IVD. Its density
and thickness have been shown to increase toward the vertebral periphery (Grant et al., 2001; Roberts et al., 1997).
In studying the vertebral endplate curvature of different
species, Langrana et al. (2006) showed that the canine
G. Ryan et al. / Clinical Biomechanics 23 (2008) 859–869
and chimpanzees, i.e. the quadrupeds, have a different curvature of their upper endplates compared with those in
humans (bipeds). Finite element analysis predicted that
the shape of the endplate alters the distribution of loads
transferred along the spine. The stress distribution varied
as the location of the maximum curvature shifted from
the centre to a more posterior position. The stresses in
the vertebral core were found to decrease, with the shell
taking more of the load, which is due to a redistribution
of load to the periphery. In the porcine spine we found that
the endplate possessed a slightly concave curvature shifted
posteriorly, which may explain the distribution of bone
strength found in this study.
The strength of the subdiscal trabecular bone was not
homogenous inside the vertebral body. We found that the
two central specimens (regions C and D), which correspond to sections overlying the NP, had lower strength
and modulus values compared to specimens approximately
underlying the AF. A number of studies have reported
regional variation in trabecular bone strength within the
vertebral body. Grant et al. (2001) and later Oxland et al.
(2003) performed a series of indentation tests at predetermined locations across the lumbar vertebral body with
the endplates both intact and removed. They found that
for both cranial and caudal endplates, the posterior region
was stronger than the anterior, and the periphery stronger
than the centre, with the strongest points in the posterolateral test sites just in front of the pedicles. With endplate
removal, although there was a 33% decrease in failure load
for indentation testing, the basic failure pattern remained
the same, with the periphery stronger than the centre. In
contrast to these findings, Keller et al. (1989) found that
bone underlying the NP is stronger than that underlying
the AF when studying the compressive properties of
human lumbar vertebral trabecular bone. For non-degenerated IVDs, they found that the ratio of strength of bone
overlying the NP to bone overlying the AF was 1.25. Similarly, Lin et al. (1997), in a porcine model, found that the
ratio of strength of bone overlying the NP to that overlying
the AF was approximately 1.16. Both authors equate their
findings to greater stress in the NP as opposed to the AF,
which we found not to be the case for the porcine spine,
especially when asymmetrical loading was applied.
It has been reported that the structure of the vertebral
body is closely related to the degenerated or aged status
of the IVD (Adams and Roughley, 2006; Adams et al.,
1996a,b, 2006; Keller et al., 1989, 1993; Simpson et al.,
2001). In the human spine, healthy IVDs contain a large
and highly hydrated NP whereas degenerative IVDs are
associated with a loss of fluid and associated height (Horst
and Brinckmann, 1981). The young healthy IVD acts to
spread the force evenly over its area (Adams et al., 2006).
With age, the reduction in size of the hydrostatic compressive region coincides with the appearance of peak stresses
in the AF (Adams et al., 1986, 1996a,b). Furthermore,
studies performed by Adams et al. have shown that a
reduction in IVD height caused by either excessive com-
867
pressive loading or prolonged creep loading can produce
peak stresses in the AF, usually posteriorly to the NP
(Adams et al., 1996a,b, 2000a,b). In this study, all porcine
spines tested were of good quality with no apparent IVD
degeneration. However, heights recorded for porcine IVDs
were considerably smaller than what is normal in human
samples, and the AF comprised a much greater fraction
of the total IVD area compared to the human spine. These
two factors appear to be responsible for the peak stresses
recorded in the AF during asymmetrical loading. It
appears that the trabecular bone of the young porcine spine
can adapt to withstand the greater forces generated in the
AF during asymmetrical loading better than how the vertebral bone of the aged human spine might respond to its
changing loading environment. Bone loss due to inactivity,
especially in the elderly is a common finding (Lunt et al.,
2001; Nguyen et al., 2007). However, it has also been demonstrated that bone growth and adaptation has a reduced
responsiveness to mechanical stimulus in the elderly (Bassey et al., 1998). It is possible that a combination of these
traits can predispose the thoracolumbar spine of the elderly
to anterior vertebral fractures when the spine is flexed
(Adams et al., 2006), which seems not to be the case for
the young porcine spine.
Findings of increased stresses in subdiscal trabecular
bone underlying the AF suggest that adaptive bone remodeling has occurred to withstand the greater forces present in
this region. It is widely understood that there are various
anatomical and biomechanical differences between humans
and even the closely related primates. Therefore, animal
cadaveric analyses may not directly explain phenomena
associated with human physiology. From a clinical perspective, results of regional distributions in trabecular
strength of porcine vertebral bodies provide little useful
information for the treatment of age-related degeneration
of the human spine or the design of implants meant for
human use. However, once an implant has been designed,
animal models are widely used to investigate the biologic
as well as mechanical behavior of the implants. Interpretation of data from such studies usually considers the limitations related to the mismatch of properties between the
human and animal tissues. This study raises awareness
and identifies such limitations.
6. Conclusions
In the porcine lumbar spine we found that asymmetrical
loading generated peak stresses in the periphery of the
IVD, namely the AF. Assuming asymmetrical loading is
part of everyday activities, this has implications on bone
mass adaptations in accordance with the magnitude and
direction of the forces that are applied to them over time.
We demonstrated that bone underlying the AF is significantly stronger than bone underlying the NP. Although
these two features cannot be directly related, this suggests
that there is a correlation between the stress in the IVD
and strength of underlying trabecular bone. The data
868
G. Ryan et al. / Clinical Biomechanics 23 (2008) 859–869
presented here may be helpful in determining the role of the
porcine spine in the evaluation of new implant systems and
surgical procedures. However, researchers must take into
account the anatomical and physiological differences
between human and porcine spines and the need for loads
to be scaled to match human magnitudes.
Acknowledgement
This work was supported by Enterprise Ireland (PC/
2005/012).
References
Adams, M.A., Dolan, P., Hutton, W.C., 1986. The stages of disc
degeneration as revealed by discograms. J. Bone Joint Surg. Br. 68, 36–
41.
Adams, M.A., Freeman, B.J., Morrison, H.P., Nelson, I.W., Dolan, P.,
2000a. Mechanical initiation of intervertebral disc degeneration. Spine
25, 1625–1636.
Adams, M.A., May, S., Freeman, B.J., Morrison, H.P., Dolan, P., 2000b.
Effects of backward bending on lumbar intervertebral discs. Relevance
to physical therapy treatments for low back pain. Spine 25, 431–437,
discussion 438.
Adams, M.A., McMillan, D.W., Green, T.P., Dolan, P., 1996a. Sustained
loading generates stress concentrations in lumbar intervertebral discs.
Spine 21, 434–438.
Adams, M.A., McNally, D.S., Dolan, P., 1996b. ‘Stress’ distributions
inside intervertebral discs. The effects of age and degeneration. J. Bone
Joint Surg. Br. 78, 965–972.
Adams, M.A., Pollintine, P., Tobias, J.H., Wakley, G.K., Dolan, P., 2006.
Intervertebral disc degeneration can predispose to anterior vertebral
fractures in the thoracolumbar spine. J. Bone Miner. Res. 21, 1409–
1416.
Adams, M.A., Roughley, P.J., 2006. What is intervertebral disc degeneration, and what causes it?. Spine 31 2151–2161.
Bassey, E.J., Rothwell, M.C., Littlewood, J.J., Pye, D.W., 1998. Preand postmenopausal women have different bone mineral density
responses to the same high-impact exercise. J. Bone Miner. Res. 13,
1805–1813.
Brinckmann, P., Grootenboer, H., 1991. Change of disc height, radial disc
bulge, and intradiscal pressure from discectomy. An in vitro investigation on human lumbar discs. Spine 16, 641–646.
Cotterill, P.C., Kostuik, J.P., D’Angelo, G., Fernie, G.R., Maki, B.E.,
1986. An anatomical comparison of the human and bovine thoracolumbar spine. J. Orthop. Res. 4, 298–303.
Dath, R., Ebinesan, A.D., Porter, K.M., Miles, A.W., 2007. Anatomical
measurements of porcine lumbar vertebrae. Clin. Biomech. 22, 607–
613.
Edwards, W.T., Ordway, N.R., Zheng, Y., McCullen, G., Han, Z., Yuan,
H.A., 2001. Peak stresses observed in the posterior lateral anulus.
Spine 26, 1753–1759.
Frost, H.M., 1983. A determinant of bone architecture. The minimum
effective strain. Clin. Orthop. Relat. Res., 286–292.
Gong, H., Zhang, M., Qin, L., Lee, K.K., Guo, X., Shi, S.Q., 2006.
Regional variations in microstructural properties of vertebral trabeculae with structural groups. Spine 31, 24–32.
Grant, J.P., Oxland, T.R., Dvorak, M.F., 2001. Mapping the structural
properties of the lumbosacral vertebral endplates. Spine 26, 889–896.
Horst, M., Brinckmann, P., 1981. 1980 Volvo award in biomechanics.
Measurement of the distribution of axial stress on the end-plate of the
vertebral body. Spine 6, 217–232.
Kaigle, A.M., Holm, S.H., Hansson, T.H., 1997. 1997 Volvo Award
winner in biomechanical studies Kinematic behavior of the porcine
lumbar spine: a chronic lesion model. Spine 22, 2796–2806.
Keller, T.S., Hansson, T.H., Abram, A.C., Spengler, D.M., Panjabi,
M.M., 1989. Regional variations in the compressive properties of
lumbar vertebral trabeculae Effects of disc degeneration. Spine 14,
1012–1019.
Keller, T.S., Holm, S.H., Hansson, T.H., Spengler, D.M., 1990. 1990
Volvo Award in experimental studies. The dependence of intervertebral disc mechanical properties on physiologic conditions. Spine 15,
751–761.
Keller, T.S., Ziv, I., Moeljanto, E., Spengler, D.M., 1993. Interdependence
of lumbar disc and subdiscal bone properties: a report of the normal
and degenerated spine. J. Spinal Disord. 6, 106–113.
Langrana, N.A., Kale, S.P., Edwards, W.T., Lee, C.K., Kopacz, K.J.,
2006. Measurement and analyses of the effects of adjacent endplate
curvatures on vertebral stresses. Spine J. 6, 267–278.
Lin, R.M., Tsai, K.H., Chang, G.L., 1997. Distribution and regional
strength of trabecular bone in the porcine lumbar spine. Clin.
Biomech. 12, 331–336.
Lunt, M., Masaryk, P., Scheidt-Nave, C., Nijs, J., Poor, G., Pols, H.,
Falch, J.A., Hammermeister, G., Reid, D.M., Benevolenskaya, L.,
Weber, K., Cannata, J., O’Neill, T.W., Felsenberg, D., Silman, A.J.,
Reeve, J., 2001. The effects of lifestyle, dietary dairy intake and
diabetes on bone density and vertebral deformity prevalence: the
EVOS study. Osteoporos. Int. 12, 688–698.
McNally, D.S., Adams, M.A., 1992. Internal intervertebral disc mechanics
as revealed by stress profilometry. Spine 17, 66–73.
Nachemson, A., Morris, J.M., 1964. In vivo measurements of intradiscal
pressure discometry, a method for the determination of pressure in the
lower lumbar discs. J. Bone Joint Surg. Am. 46, 1077–1092.
Nguyen, N.D., Center, J.R., Eisman, J.A., Nguyen, T.V., 2007. Bone loss,
weight loss, and weight fluctuation predict mortality risk in elderly
men and women. J. Bone Miner. Res. 22, 1147–1154.
Nissan, M., Gilad, I., 1986. Dimensions of human lumbar vertebrae in the
sagittal plane. J. Biomech. 19, 753–758.
Oxland, T.R., Grant, J.P., Dvorak, M.F., Fisher, C.G., 2003. Effects of
endplate removal on the structural properties of the lower lumbar
vertebral bodies. Spine 28, 771–777.
Panjabi, M.M., Goel, V., Oxland, T., Takata, K., Duranceau, J., Krag,
M., Price, M., 1992. Human lumbar vertebrae quantitative threedimensional anatomy. Spine 17, 299–306.
Park, C., Kim, Y.J., Lee, C.S., An, K., Shin, H.J., Lee, C.H., Kim, C.H.,
Shin, J.W., 2005. An in vitro animal study of the biomechanical
responses of anulus fibrosus with aging. Spine 30, E259–E265.
Pfeiffer, M., Griss, P., Franke, P., Bornscheuer, C., Orth, J., Wilke, A.,
Clausen, J.D., 1994. Degeneration model of the porcine lumbar
motion segment: effects of various intradiscal procedures. Eur. Spine J.
3, 8–16.
Pollintine, P., Dolan, P., Tobias, J.H., Adams, M.A., 2004. Intervertebral
disc degeneration can lead to ‘‘stress-shielding” of the anterior vertebral
body: a cause of osteoporotic vertebral fracture? Spine 29, 774–782.
Ranu, H.S., 1990. Measurement of pressures in the nucleus and within the
annulus of the human spinal disc: due to extreme loading. Proc. Inst.
Mech. Eng. [H] 204, 141–146.
Roberts, S., McCall, I.W., Menage, J., Haddaway, M.J., Eisenstein, S.M.,
1997. Does the thickness of the vertebral subchondral bone reflect the
composition of the intervertebral disc? Eur. Spine J. 6, 385–389.
Simpson, E.K., Parkinson, I.H., Manthey, B., Fazzalari, N.L., 2001.
Intervertebral disc disorganization is related to trabecular bone
architecture in the lumbar spine. J. Bone Miner. Res. 16, 681–687.
Smit, T.H., 2002. The use of a quadruped as an in vivo model for the study
of the spine – biomechanical considerations. Eur. Spine J. 11, 137–144.
Steffen, T., Baramki, H.G., Rubin, R., Antoniou, J., Aebi, M., 1998.
Lumbar intradiscal pressure measured in the anterior and posterolateral annular regions during asymmetrical loading. Clin. Biomech. 13,
495–505.
Teo, J.C., Si-Hoe, K.M., Keh, J.E., Teoh, S.H., 2006. Relationship between CT intensity, micro-architecture and mechanical
properties of porcine vertebral cancellous bone. Clin. Biomech. 21,
235–244.
G. Ryan et al. / Clinical Biomechanics 23 (2008) 859–869
van Deursen, D.L., Snijders, C.J., Kingma, I., van Dieen, J.H., 2001. In
vitro torsion-induced stress distribution changes in porcine intervertebral discs. Spine 26, 2582–2586.
White, A.A., Panjabi, M.M., 1990. Clinical Biomechanics of the Spine,
second ed. Lippincott Williams & Wilkins, Philadelphia.
Wilke, H.J., Kettler, A., Claes, L.E., 1997. Are sheep spines a valid
biomechanical model for human spines? Spine 22, 2365–2374.
869
Wilke, H.J., Krischak, S., Claes, L., 1996a. Biomechanical comparison of
calf and human spines. J. Orthop. Res. 14, 500–503.
Wilke, H.J., Wolf, S., Claes, L.E., Arand, M., Wiesend, A., 1996b.
Influence of varying muscle forces on lumbar intradiscal pressure: an
in vitro study. J. Biomech. 29, 549–555.
Wolff, J., 1884. The Law of Bone Remodelling. Des Gesetz der
Transformation der Knochen, Berlin.