finite element investigation of the loadin grate effect on the spinal load-sharing changes under...
TRANSCRIPT
-
7/27/2019 Finite Element Investigation of the Loadin Grate Effect on the Spinal Load-sharing Changes Under Impact Conditions
1/11
Finite element investigation of the loading rate effect on the spinalload-sharing changes under impact conditions
Marwan El-Rich a,b, Pierre-Jean Arnoux a,, Eric Wagnac a,b, Christian Brunet a, Carl-Eric Aubin b
a Laboratoire de Biomecanique Appliquee (LBA-INRETS), Marseille, Franceb Department of Mechanical Engineering, Ecole Polytechnique, Montreal, Quebec, Canada
a r t i c l e i n f o
Article hi story:
Accepted 11 March 2009
Keywords:
High-speed impact
Lumbar spine
Stress
Bone fracture
Injury
Finite-element analysis
a b s t r a c t
Sudden deceleration and frontal/rear impact configurations involve rapid movements that can cause
spinal injuries. This study aimed to investigate the rotation rate effect on the L2L3 motion segmentload-sharing and to identify which spinal structure is at risk of failure and at what rotation velocity the
failure may initiate?
Five degrees of sagittal rotations at different rates were applied in a detailed finite-element model to
analyze the responses of the soft tissues and the bony structures until possible fractures. The structural
response was markedly different under the highest velocity that caused high peaks of stresses in the
segment compared to the intermediate and low velocities. Under flexion, the stress was concentrated at
the upper pedicle region of L2 and fractures were firstly initiated in this region and then in the lower
endplate of L2. Under extension, maximum stress was located in the lower pedicle region of L2 and
fractures started in the left facet joint, then they expanded in the lower endplate and in the pedicle
region of L2. No rupture has resulted at the lower or intermediate velocities. The intradiscal pressure
was higher under flexion and decreased when the endplate was fractured, while the contact forces were
greater under extension and decreased when the facet surface was cracked. The highest ligaments
stresses were obtained under flexion and did not reach the rupture values. The endplate, pedicle and
facet surface represented the potential sites of bone fracture. Results showed that spinal injuries can
result at sagittal rotation velocity exceeding 0.51
/ms.&2009 Elsevier Ltd. All rights reserved.
1. Introduction
Epidemiologic and biomechanical studies have shown the role
of the mechanical loads acting on the human spine during daily
activities in the onset of low back pain (LBP) disorders and
symptoms (Damkot et al., 1984; Kelsey et al., 1984). Among the
various loading conditions, impact loading at high-velocity
rate releases important energy over a short time period and can
induce spinal fractures. Although spinal fractures (burst fracture,
disc protrusion and narrowing of the spinal canal by bone
fragment) are of immense clinical significance (Tran et al., 1995;Wilcox et al., 2004), the biomechanics of spinal injury has been
insufficiently analyzed under dynamic loadings.
Under high-velocity impact, the spine can be injured with
small displacement and angulation compared with a low-velocity
injury (Neumann et al., 1995, 1996). For a short duration of impact,
the high dynamic stiffness increases the stability of the spinal
segment against the impact load (Lee et al., 2000). However, the
corresponding increase in stresses within the vertebral body and
endplate may induce fractures. Yingling et al. (1997) found that
failure at low load rates occurred exclusively in the endplate,
whereas failure of the vertebral body appeared more frequently at
higher load rates. In healthy spine, the excessive pressurization of
nucleus under high dynamic loading can cause endplate fracture
(Brown et al., 2008).
Numerous studies investigating the effect of loading rate on
the biomechanical properties of the boneligamentbone com-
plex have reported an increase in failure load, failure strain,
stiffness and energy absorbed to failure with increasing loading
rate (Panjabi et al., 1998; Bass et al., 2007). High loading rateincreased the intradiscal pressure (IDP), resistant bending mo-
ment, ligament stress and annulus fiber stress (Wang et al., 2000).
However, the pressure is independent of the impact duration and
depends only on the magnitude of the impact force (Lee et al.,
2000).
The lumbar spine component (L2L3) mobility seems to be
responsible for many severe injuries such as bone fractures
(Sances et al., 1984). Mechanisms postulated in trauma situations
(vehicle crash, aircraft ejection) are related to flexion/extension of
the spinal unit. From these previous works, we assumed that the
lumbar motion segment could demonstrate diverse mechanical
behaviors leading to trauma under different loading velocities.
ARTICLE IN PRESS
Contents lists available atScienceDirect
journal homepage: www.elsevier.com/locate/jbiomechwww.JBiomech.com
Journal of Biomechanics
0021-9290/$- see front matter& 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jbiomech.2009.03.036
Corresponding author. Tel.: +334 9165 80 00; fax: +334 9165 8019.
E-mail address: [email protected] (P.-J. Arnoux).
Journal of Biomechanics 42 (2009) 12521262
http://www.sciencedirect.com/science/journal/jbiomechhttp://www.elsevier.com/locate/jbiomechhttp://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.jbiomech.2009.03.036mailto:[email protected]:[email protected]://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.jbiomech.2009.03.036http://www.elsevier.com/locate/jbiomechhttp://www.sciencedirect.com/science/journal/jbiomech -
7/27/2019 Finite Element Investigation of the Loadin Grate Effect on the Spinal Load-sharing Changes Under Impact Conditions
2/11
We developed a refined three-dimensional finite-element model
(FEM) of the L2L3 segment enriched by advanced visco-elastic,
hyper-elastic and elasto-plastic material properties (Garo et al.,
2007) to evaluate the rotation rate effect on the load-sharing
changes among the segment components during rapid sagittal
movements. These movements could result from sudden decel-
eration, frontal or rear impact configurations and lead to potential
damage in the structure. The current study aimed particularly to
identify which spinal component is at risk of failure and at whatvelocity the failure may occur.
2. Methods
2.1. Model description
The geometry of the vertebrae was reconstructed from 0.6-mm-thick CT-scan
slices of a 50th percentile healthy male with no recent back complication (Fig. 1).
The vertebral bodies and posterior elements were modeled by taking into account
the separation of the cortical shell (including bony endplate and facet joints) and
cancellous bone using 3-nodes shell and 4-nodes solid elements, respectively.
All shell elements had 0.7mm thickness and characteristic length close to 0.5 mm.
The intervertebral disc was created between the intervening endplates. It was
subdivided into nucleus pulposus and annulus fibrosus with a proportion
according to the histological findings (44%_nucleus, 56%_annulus). The disc
was filled with 5 layers of 8-nodes solid elements and the annulus was
reinforced in the radial direction by 8 collagenous fiber layers using
unidirectional springs organized in concentric lamellae with crosswise pattern
ARTICLE IN PRESS
Annulus Matrix
Collagenous Fibers
ALL
ITL
FL
JC
Contact Facet
Nucleus
SSL
ISL
PLL
Bony Endplate
Cancellous Bone
Cortical Shell
191000 elements40300 nodes
Fig. 1. L2L3 finite-element model.
Fig. 2. JohnsonCook elasto-plastic material law used to model the structural
behavior of the bone structure.
M. El-Rich et al. / Journal of Biomechanics 42 (2009) 12521262 1253
-
7/27/2019 Finite Element Investigation of the Loadin Grate Effect on the Spinal Load-sharing Changes Under Impact Conditions
3/11
close to 7351 ( Schmidt et al., 2007). Fiber stiffness increased by 110% from the
center to the middle layers and by 150% from the center to the outer layers of the
annulus (Shirazi-Adl et al., 1986;Cheung et al., 2003).
The surrounding ligaments, the anterior (ALL) and posterior (PLL) longitudinal
ligaments, intertransverse (ITL), flavum (FL), capsular (JC) and intertransverse (ITL)
ligaments were represented by envelops of 1 mm uniform thickness. The
geometrical properties were taken from the literature (Pintar et al., 1992). All
ligaments were modeled with 4- and 3-nodes (JC) shell elements.
Tied contact interfaces were used to ensure the disc and ligament attachment
to the vertebrae and to prevent any relative movement during the simulations.
Frictionless contact interfaces were assumed between the diverse parts of themodel to avoid any possible penetration. This interface was also used between the
facet surfaces to calculate the contact forces.
The bone structure was assumed as a homogeneous material. It was modeled
with a symmetric elasto-plastic material law (JohnsonCook) allowing computing
von Mises hardening with ductile damage until potential rupture ( Fig. 2). In this
law, the material behaves as linear elastic when equivalent stress is lower than the
yield stress and as plastic for higher values of stress. When the maximum stress is
reached during computation, the stress remains constant while the elements
deformation continues until the plastic strain reaches the maximum value.
Element rupture occurs if the plastic strain reaches the maximum value. If the
element is a shell, the ruptured element is deleted. If the element is a solid, the
ruptured element has its deviatoric stress tensor permanently set to zero, but the
element is not deleted. Therefore, the material rupture is modeled without any
damage effect. The plastic strain threshold used in the model is ranged from 1% to
3% (Schileo et al., 2008; Kimpara et al., 2006; Arnoux et al., 2005). Strain-rate
dependency of the bone structure was investigated through a sensitivity analysis.
The ligaments and the disc were governed by visco-elastic (generalized
MaxwellKelvinVoigt; Fig. 3) and hyper-elastic (MooneyRivlin) material
laws, respectively, while the fibers were modeled using nonlinear elastic
material (Table 1). Damage occurrence of soft tissues is usually described in
terms of ultimate strain levels. Therefore, ligament failure was based on straincriterions based onPintar et al. (1992) findings. Ultimate strain levels were thus
calculated for all ligaments and used to identify their potential failure. All
simulations were performed using the explicit dynamic finite-element solver
Radioss (Version 4.4, Altair HyperWorks Inc.).
2.2. Model validation under quasi-static loading conditions
The calculated quasi-static compressive stiffness of the disc was compared
with the in-vitro values obtained on cadaveric samples (three women donors:
ARTICLE IN PRESS
Table 1
Summary of the material properties used for the modeling.
Material properties Vertebra components
Density (kg/mm3) 1.83E-06 (Lee et al., 2000) 0.17E-06 (Kopperdahl and
Keaveny, 1998)
1.06E-06 (Kasra et al., 1992)
Young modulus,
E(MPa)
14000 (Kopperdahl and Keaveny,
1998;Wirtz et al., 2000)
291 (Kopperdahl and
Keaveny, 1998)
10000 (Lee et al., 2000)
Poisson ratio, n 0.3 (Qiu et al., 2006) 0.25 (Qiu et al., 2006) 0.3 (Qiu et al., 2006)Yield stress a(MPa) 110 (Kopperdahl and Keaveny,
1998,Wirtz et al., 2000)
1.92 (Kopperdahl and
Keaveny, 1998)
6 (Ochia et al., 2003)
Hardening modulus
b(MPa)
100 20 100
Hardening exponent,
n
0.1 1 1
Failure plastic strain,
ep
9.68E-03 14.5E-03 (Kopperdahl and
Keaveny, 1998)
0.02
Maximum stress,
(MPa)
155 (Kopperdahl and Keaveny,
1998,Wirtz et al., 2000)
2.23 (Kopperdahl and
Keaveny, 1998)
7.5 (Ochia et al., 2003)
Strain rate coefficient,
c
1 1 3
Disc components
Nucleus pulposus Annulus matrix References Collagenous fibers Reference
Density (kg/mm3) 1.00E-06 1.20E-06 (Lee et al., 2000) Nonlinear elastic curve (ShiraziAdl et al., 1986)
Poisson ratio, n 0.495 0.45 (Schmidt et al., 2007)C10 0.12 0.18 (Schmidt et al., 2007)
C01 0.03 0.045 (Schmidt et al., 2007)
Ligaments
ALL PLL ITL ISL LF SSL JC References
Density (kg/mm3) 1.0E-06 1.0E-06 1.0E-06 1.0E-06 1.0E-06 1.0E-06 1.0E-06
Young modulus, E(MPa) 11.4 9.12 11.4 4.56 5.7 8.55 22.8 (Yang et al. 1998)
Poisson ratio, n 0.4 0.4 0.4 0.4 0.4 0.4 0.4 (Yang et al. 1998)Tangent modulus, Et(MPa) 10.0 9.0 11.0 4.0 5.0 8.0 22.0
Tangent poisson ratio, nt 0.42 0.42 0.42 0.42 0.42 0.42 0.42 Viscosity coefficient,Z0 28 28 28 28 28 28 28 Naviers constant, l 1.0E06 1.0E06 1.0E06 1.0E06 1.0E06 1.0E06 1.0E06
ke += , ke +=
E
e = ,
t
vk
E
= ,
v
=k
tE
E
k
ke
+=+=
, kt
E +=
k
tt
EEEEE ++=
)(
Fig. 3. Generalized MaxwellKelvinVoigt visco-elastic material law used to model the structural behavior of the spinal ligaments.
M. El-Rich et al. / Journal of Biomechanics 42 (2009) 125212621254
-
7/27/2019 Finite Element Investigation of the Loadin Grate Effect on the Spinal Load-sharing Changes Under Impact Conditions
4/11
subject1: 85years, 177 cm, 86 kg; subject2: 55years, 156cm, 38 kg; subject3: 80years,
159cm, 69 kg). These samples containing the disc and adjacent endplates (Fig. 4a)
were taken from T9 to L4. The disc was loaded until damage (over 60% of axial strain)
by applying an axial compressive displacement on the lower endplate of the proximal
vertebra at a constant velocity of 1.267mm/s, while the upper endplate of the distal
vertebra was fixed (Fig. 4b). These endplates were embedded with a rigid resin.
The IDP changes under a follower preload simulating gravity (Rohlmann et al.,
2006) and a combination of a preload and moments in the principal planes were
also evaluated and compared with the published values. These loads were 500N of
preload combined with 7.5 N m in the principal planes (Schmidt et al., 2007) and
1000 N of preload combined with 20Nm in the sagittal plane (Shirazi-Adl andDrouin, 1988).
The ligaments strains were compared within-vitrovalues measured on L3L4
and L4L5 segments under 15 N m of physiological moments (Panjabi et al., 1982).
These deformations were calculated as the percentage change in length with
respect to the original length (shortest distance between the insertion points of the
ligaments), and an average value was considered.
2.3. Loading rate investigation
Five degrees of flexion and extension were applied on the L2 upper endplate
at three rates (0.05, 0.5 and 51/ms) while the model was fixed in the L3 lower
endplate. These endplates were considered as rigid and the rest of the structures as
flexible bodies. The IDP changes in the nucleus and von Mises stress in the annulus
were calculated. The ligaments stresses were evaluated in their fibers directions
and the contact forces were assessed in the facet joints. Equivalent stress and
plastic strain in the bone until possible fracture were also evaluated.
3. Model validation results
The calculated forcedeformation curve of the disc showed a
nonlinear compressive behavior and a stiffness increase with load.
The curve falls within the experimental corridor (Fig. 4c), however
the modeled disc appeared stiffer than the corresponding
experimental one, and the decrease in stiffness seen in the
experimental curves was not obtained by the model.
The IDP was in agreement with the previous results ofShirazi-Adl and Drouin (1988)(Fig. 5a) andSchmidt et al. (2007)(Fig. 5b).
Ligament strains that were inside the experimental range was
defined byPanjabi et al. (1982)(Table 2). In flexion all ligaments
were recruited except ALL while only ALL and JC ligaments were
recruited under extension. The right bending generated greater
deformation than the left one. In the axial plane, the deformation
was similar under the right and left rotations and the greatest
values were obtained for the JCligaments (Table 2).
4. Loading rate investigation results
The structural response was markedly different under the
highest velocity (Figs. 6 and 7). It exhibited vibrations and caused
ARTICLE IN PRESS
0
500
1000
1500
2000
2500
3000
3500
4000
4500
Axial Displacement (mm)
CompressiveForce(N)
L2_L3_FEM
T11-T12 (Sub_1)
T9-T10 (Sub_1)
L1-L2 (Sub_1)
L3-L4 (Sub_1)
L2-L3 (Sub_2)
T10-T11 (Sub_2)
T12-L1 (Sub_2)
L1-L2 (Sub_3)
T12-L1 (Sub_3)
0.0 0.5 1.0 1.5 2.0 2.5
Fig. 4. (a) Experimental set up for quasi-static compressive testing of the disc, (b) finite-element simulation of the compressive testing and (c) the compressive
forcedisplacement curve of the disc: FE model of the lumbar L2L3 level versus experimental curves of several thoracic, thoraco-lumbar and lumbar levels of the three
postmortem human subjects (Sub_1, Sub_2, Sub_3).
M. El-Rich et al. / Journal of Biomechanics 42 (2009) 12521262 1255
-
7/27/2019 Finite Element Investigation of the Loadin Grate Effect on the Spinal Load-sharing Changes Under Impact Conditions
5/11
high peaks of stresses in the ligaments, the disc and the bony
structures. The stress and the IDP increased with the rotation rate.
At the faster rotation, the stress and the plastic strain exceeded
the yield and ultimate values, respectively, in L2 vertebra. The
stress was concentrated in the upper and lower pedicle regions
under flexion and extension, respectively (Table 3). Under flexion,
ARTICLE IN PRESS
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
Moment (N.m)
IntradiscalPressure(MPa)
Flexion_Shirazi & Drouin, 1988
Flexion_FE Model
Extension_Shirazi & Drouin, 1988
Extension_FE Model
PreloadP = 1000N
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
IDP(MPa)
FEM_L2-L3
Schmidt et al. 2007_L4-L5
500N of preload &7.5Nm of moment
0 2 4 6 8 10 12 14 16 18 20
Flexion Lat. Bending (R)Extension Lat. Bending (L) Ax. Rotation (R)
Fig. 5. IDP changes under combination of preload and moments in the principal planes: (a) 1000 N of preload combined with large sagittal moments and (b) 500 N of
preload combined with 7.5N m of moment in the principal planes.
Table 2
Ligaments deformation under 15 N m of moment in the principal planes: negative values mean unloaded ligaments (FE Model versusin-vitro data).
Ligament Flexion Extension Right bending Left bending Right rotation Left rotation
FE model In-vitro
studyaFE model In-vitro
studyaFE model In-vitro
studyaFE model In-vitro
studyaFE model In-vitro
studyaFE model In-vitro
studya
ALL 7.9 871.4 7.1 6.771.4 2.6 3.773 1.25 2.672.8 1.9 1.671.4 2.2 2.773.3
PLL 7.2 7.373.3 4.6 4.673.4 7.8 8.873.5 2.9 4.273 0.9 1.170.4 3.7 3.271.6
ITL_R 4.3 7.472.8 4 4.572.6 8.4 6.871.4 10.9 9.971.8 3.3 3.271.7 0 1.171.5
ITL_L 3.7 7.672.7 3.2 3.971.8 12.7 15.973.8 11.9 1273.9 0.9 0.771.3 2.7 4.972.6
FL_R 9.5 9.173.1 6 6.573.1 2.4 2.772.1 1.9 0.971.9 2.1 1.870.8 2.4 2.271.1
FL_L 10.7 9.173.1 7.6 6.372.9 6.8 7.772.9 3.7 3.972.7 1.0 0.970.6 2.8 371.7
JC_R 10.3 10.475 5 672.1 1.3 1.874.7 1.1 1.274.3 7.7 771.8 2.1 3.672.7
JC_L 12.7 1375.5 3.8 3.771.7 9.9 8.875.4 2.9 3.274 3.7 3.372.7 8.2 8.774.3
ISL 17.3 1775.3 4.9 5.172.2 4.6 4.274.2 2.3 0.873.1 4.2 471.7 4.1 5.973.2
SSL 19.2 1874.9 11 13.174.1 1.4 4.473.4 1.1 0.173.5 2.5 3.371.8 2.6 5.272.9
a Mean7SD, Panjabi et al., (1982).
M. El-Rich et al. / Journal of Biomechanics 42 (2009) 125212621256
-
7/27/2019 Finite Element Investigation of the Loadin Grate Effect on the Spinal Load-sharing Changes Under Impact Conditions
6/11
fractures were initiated in the lower endplate at 2.41 and in the
upper pedicle region at 3.51 (Fig. 6a). Under extension, fractures
were initiated in the left facet joint at 1.51, in the lower endplate at
2.31and in the upper pedicle region at 2.81(Figs. 6b and8). Under
the intermediate and lower flexion, stresses were concentrated in
the same regions as for the fastest flexion except for the stress
in the endplate that was higher in L3 than in L2 (Table 3).
Under the intermediate and lower extension, higher stresses were
concentrated in the L2 cortical shell, L3 cancellous bone and
endplate (Table 3). However, no rupture has resulted at the
intermediate or lower rate.
The fastest rotation of L2 has increased significantly the IDP
(3.1 MPa_flexion) and the stress in the outer annulus (4.6 MPa_ex-
tension, 12.7 MPa_flexion). The fastest extension caused also
important contact forces (920 N of total force) but increased
slightly the IDP (1 MPa). The fracture occurrence in the endplate
and facet surfaces caused decreases in the IDP (Fig. 6a) and the
contact forces (Fig. 6b), respectively.
The fastest flexion caused high stresses in the capsular and
posterior ligaments. Maximal stress was concentrated in the
lower attachment of left JC ligament to the articular facet, in the
middle region of SSL ligament and in the lower posterior
attachment of ISL ligament to the L3 spinous process. Highest
stress was also concentrated in the middle region of the right ITL
ligament, in the lower region of LF ligament and in the upper
attachment ofPLL ligament to the L2 lower endplate. Under the
ARTICLE IN PRESS
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
Rotation ()
IDP
(MPa)
5/ms_F
5/ms_E
0.5/ms _F
0,5/ms_E
0.05/ms_F
0.05/ms_E
Fracture of the L2
Lower Endplate
Fracture in the L2 Pedicle Region
0
100
200
300
400
500
600
700
800
900
1000
Rotation ()
TotalContactForce(N)
5/ms_F
5/ms_E
0.5/ms_F
0.5/ms_E
0.05/ms_F
0.05/ms_E
Fracture of the L2
Left Facet Surface
0 1 2 3 4 5
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
Fig. 6. (a) IDP changes under flexion and extension at different rates and fracture occurrence at different rotation rate levels from 0.05 to 5 1/ms under flexion (_F) and
extension (_E) and (b) total contact force at facet surfaces under flexion and extension at different rotation rate levels from 0.05 to 5 1/ms under flexion (_F) and extension
(_E).
M. El-Rich et al. / Journal of Biomechanics 42 (2009) 12521262 1257
-
7/27/2019 Finite Element Investigation of the Loadin Grate Effect on the Spinal Load-sharing Changes Under Impact Conditions
7/11
fastest extension, the highest stress was concentrated in lower
attachment of the left JCligament to the articular facet and in the
left upper attachment of theALL ligament to the lower endplate of
L2 (Fig. 9). No failure was observed in ligaments as the strain level
remained below the ultimate thresholds.
5. Discussion
The current study used a FEM with bio-realistic geometry
and refined mesh. It allowed investigating the different injury
phenomena that could arise under impact loads. This model
ARTICLE IN PRESS
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
Stress(MPa)
0.05/ms
0.5/ms
5/ms
Extension () Flexion ()
JC
SSL
PLL
JC
ALL
Fracture of the L2Left Facet Surface
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
Stress(MPa)
0.05/ms
0.5/ms
5/ms
Flexion ()
ISL
ITL
FL
0.0 1.0 2.0 3.0 4.0 5.0 6.0
0.5 1.5 2.5 3.5 4.5 5.5 6.5-6.5 -5.5 -4.5 -3.5 -2.5 -1.5 -0.5
Fig. 7. Ligaments stresses evaluated in their fiber directions under flexion (left) and extension (right) at the different rates: (a)JC,SSL,PLLligaments under flexion andJCand
ALL ligaments under extension and (b) ISL, FL and ITL ligaments under flexion.
Table 3
Maximum stress values and locations in the vertebrae at different rotation rates of flexion and extension movements.
Bony location Flexion Extension
Low rate Intermediate rate High rate Low rate Intermediate rate High rate
Cancellous bone 2.2 (L2) 4.5 (L2) 20 (L2) 2.1 (L3) 4.1 (L3) 22 (L2)
Cortical shell 80.5 (L2) 86.5 (L2) 154 (L2) 89 (L2) 91 (L2) 154 (L2)
Endplates 6.5 (L2) 6.6 (L2) 96 (L2) 6.2 (L3) 6.3 (L3) 65 (L2)
M. El-Rich et al. / Journal of Biomechanics 42 (2009) 125212621258
-
7/27/2019 Finite Element Investigation of the Loadin Grate Effect on the Spinal Load-sharing Changes Under Impact Conditions
8/11
allowed a detailed/realistic evaluation of failure occurrence and
propagation over the bone and not only to define failure risk
regions based on high-stress concentrations (Lee et al., 2000;Qiu
et al., 2006;Wilcox et al., 2004).Kopperdahl and Keaveny (1998)
demonstrated that the cancellous bone in human vertebrae has
similar compressive and tensile mechanical properties expect
the yield strain, which was significantly higher in compression.
Thus, similar compressive and tensile mechanical properties were
considered in modeling the bone. The microstructure and
anisotropy of the cancellous bone were not modeled as well asthe bone fragments at the fracture site. Also, the fracture
depended on the ultimate plastic strain value and location. Such
simplifications were already used in many other studies (Schileo
et al., 2008;Kimpara et al., 2006) and offered acceptable results.
Future work could benefit from the implementation of a user
pseudo-elasto-plastic material law based on energy formulation
that includes unsymmetrical behavior, damage and failure (Jundt
et al., 2007).
Ligament failure was not supported by this model. However,
the ligament stress/strain values were compared with the
published failure data (Pintar et al., 1992). Stress was not
uniformly distributed over the ligaments and was concentrated
in specific region of each ligament. This confirmed the benefit
of the 2D geometrical representation in ligament modeling. Theuse of viscous material properties allowed studying the ligaments
under various loading rates. However, the results depend on the
viscous properties given in the model. Since very limited
information was available on the structural behavior of spinal
ligaments in dynamic loading conditions, the used viscous
parameters were based on the values measured in quasi-static
conditions (Pintar et al., 1992). Consequently, in the present study,
it was assumed that any slight variation in the viscosity
parameters would not modify significantly the results trend and
the observed phenomenon, which is not related to ligaments
injury but to bone fracture. Recent findings (Arnoux et al., 2005)
showed that loading of the ligaments at high strain rate could lead
to a saturation phenomenon (no more viscous effect even when
the strain rate increases). Thus, the objective of this work was
focused on the investigation of injury mechanisms up to bone
failure using a single set of viscous parameters for ligaments
structure behavior. Obviously, further improvements of the
ligaments model (integration of the toe-in region, implementation
of threshold data for damage and failure process in dynamic
loading conditions, extended validation of the strain-rate effects
according to the range of tested velocity, etc.) could be performed
once experimental data will be available.
The fluid-like behavior of the disc was simulated with nearly
incompressible hyper-elastic material (Schmidt et al., 2007;Noailly et al., 2007). The fluid-flow in the disc and porous bone
was not simulated since the current model studied only the
immediate effect of fast movements on the load-sharing changes
over the spinal structures.
The disc stiffness curve obtained by the model under quasi-
static compressive test corroborated the experimental ones.
However, the modeled disc was stiffer than the corresponding
experiments that may be explained by the decrease of the discs
strength related to the donors age (Skrzypiec et al., 2007). The
decrease in stiffness showed by the experimental curves repre-
senting the failure behavior was not demonstrated by the model.
The IDP changes under preload only and a combination of pre-
load and moments followed the same trends as the published
values (Shirazi-Adl and Drouin, 1988; Schmidt et al., 2007). Theligaments strains under physiological moments corroborated the
experimental values (Panjabi et al., 1982) and the differences may
be related to the dissimilarities in geometry and stiffness between
the modeled segment (L2L3) and the experimental ones (L3L4
and L4L5).
The vertebral body failure load under dynamic compressive
tests was previously investigated using the current model (Garo
et al., 2007). These authors have found that under axial displace-
ment test (2.5 m/s), the vertebral body failed when the load
reached 10.4kN, which agrees with the published values of
9699.9372110.63 kN (Ochia et al., 2003). Moreover, a comple-
mentary sensitivity study was performed to investigate the strain
rate effects on the bone structure. The analysis was performed on
the Youngs modulus, the yield stress, the maximum stress and
ARTICLE IN PRESS
At 2.8
At 5.0
At 1.5
Fractures
Propagation
Fractures
Initiation
Top View
Bottom View
Sagittal View
Top View
Fig. 8. Initiation and propagation of fracture in the L2 vertebra under the fastest extension movement.
M. El-Rich et al. / Journal of Biomechanics 42 (2009) 12521262 1259
-
7/27/2019 Finite Element Investigation of the Loadin Grate Effect on the Spinal Load-sharing Changes Under Impact Conditions
9/11
the maximum strain of the bony components (endplate, cortical
and cancellous bones) and considered two sets of material
properties. The first one was based on material propertiesmeasured in quasi-static loading conditions (eo1 s1) (Wirtz
et al., 2000;Lee et al., 2000;Kopperdahl and Keaveny, 1998) while
the other set was based on material properties measured in
dynamic loading conditions (e428 s1) (Hansen et al., 2008;Shim
et al., 2005). Accordingly, the values of the parameters provided in
the sensitivity analysis covered a wide range of material proper-
ties, thus providing an alternate method to evaluate the potential
effect of bone viscosity (represented by strain-rate dependency of
the material properties) on spinal injuries. Results showed that for
both flexion and extension, changes in the material properties
of the bony components had no effects on the location of the site
where fractures are occurring, and slight effect on the angle at
which they initiate (o31). Thus, despite a limited validation of the
FE model against experimental data measured in dynamic loading
conditions, the complementary analysis confirmed the conclusion
drawn from the current model.
Faster movement of L2 has increased considerably the IDP inthe nucleus and the stresses over the rest of the structures,
whereas a slight change was found between the intermediate
and lower rates except in the ligaments stress under flexion. This
increase was more significant under faster flexion than extension.
However, the faster extension movement generated the highest
contact forces causing facets fractures. This confirmed the distinct
role of the posterior ligaments and facet joints in supporting the
extension moment (Shirazi-Adl and Drouin, 1988). Under flexion,
the substantial increase of the IDP caused the initiation of
fractures in the L2 lower endplate (Brown et al., 2008). The L2
pedicle fractures may result from the posterior ligaments (viscous
structure) resistance to the rapid movement of L2. These results
supported the hypothesis of the inertial and visco-elastic
resistance of the spine when exposed to high-speed traumatic
ARTICLE IN PRESS
Max
Extension
ALL
Max
Extension
JC_RJC_L
Max
Flexion
ISL
Max
Flexion
SSL
Max
Flexion
ITL_RITL_L
Max
Flexion
FL
Max
Flexion
PLL
Max
Flexion
JC_RJC_L
Fig. 9. Ligaments stress (von Mises) distribution under the fastest extension and flexion.
M. El-Rich et al. / Journal of Biomechanics 42 (2009) 125212621260
-
7/27/2019 Finite Element Investigation of the Loadin Grate Effect on the Spinal Load-sharing Changes Under Impact Conditions
10/11
loading (Viano and Lau,1988). Fractures were initiated in different
regions of L2 at smaller angles, which confirmed the vulnerability
of the segment under high-rate loading (Neumann et al., 1995,
1996). This study confirmed also that rapid movements may
reduce the margin of safety for the spine and increase the risk
of injury (Adams and Dolan, 1996). Results demonstrated also that
under the same rate, fractures may occur at different times and
regions depending on the movement direction and the strength
of these regions (Lee et al., 2000). The facet surfaces, endplatesand pedicle were the weakest regions when the motion segment
moved rapidly in the sagittal plane. Experimental studies have
shown that failure caused by impact loading occurs in the
endplate or posterior region of the cortical shell, (Willen et al.,
1984;Yingling et al., 1997) although there is a lack of consensus as
to which region fails first.
The stress values and distribution in the different ligaments
depended on their stiffness and orientation with respect to the
center of rotation (Panjabi et al., 1982). Stress was mostly
concentrated in the attachment regions of ligaments to the bone
that may lead to ligaments tear. The capsular and posterior
ligaments were highly loaded; however, it is quite speculative at
this time to connect these results to the invocation of pain. We
assumed also that the disc will not fail under the highest stress
obtained in this study based on the failure load values obtained
experimentally under quasi-static compressive tests and the
increase of the disc strength with loading rate (Adams and Dolan,
1996;Kemper et al., 2007).
This study investigated sagittal symmetric movements and did
not consider neuromuscular responses, which may underestimate
the effect on the load-sharing changes over the spinal compo-
nents. The risk of injury increases as a result of the higher stress/
strain caused by additional lateral or axial rotation (Shirazi-Adl,
1989) and/or higher internal loads (Lavander et al., 1999; Marras
and Mirka, 1990; Fathallah et al., 1998). In real life conditions,
more complex loads are present. To our knowledge, no study
quantified precisely those complex loads. This study is a first step
to investigate such mechanical loads and identify the potential
risk of injuries with an increasing loading rate. Future work willaddress more complex types of loading such as combined
rotations and translations that could potentially lead to more
severe injuries due to the early contact between bone components
and to an amplified strain level on ligaments.
6. Conclusion
A detailed FE model of the spinal complex structure was build
to investigate spinal injury mechanisms (location, chronology and
macroscopic failure process) and structure effects in dynamic
loading conditions. The obtained results demonstrated that
sagittal movement of the lumbar spine during sudden decelera-
tion and rear/front impact conditions increased significantly theIDP and the contact forces and generated high stresses in the disc,
ligaments and vertebrae. Flexion generated the highest stresses
while bone fractures were firstly initiated under extension. The
endplate, pedicle and facet surface represented the potential sites
of bone fracture. The ligaments attachment and outer annulus
regions highly loaded were susceptible to failure. These spinal
injuries can result at sagittal rotation velocity exceeding 0.51/ms.
Conflict of interest
There is no conflict of interest. Authors have not received any
payment for conducting this work and are in no conflict of
interest.
Acknowledgements
The authors are particularly grateful to M. Py, C. Regnier and M.
Paglia for the experimental set up and in-vitro samples prepara-
tions. This work was funded by the Fonds de Recherche sur la
Nature et les Technologies of the Government of Quebec.
Appendix A. Supporting Information
Supplementary data associated with this article can be found
in the online version at doi:10.1016/j.jbiomech.2009.03.036.
References
Adams, M.A., Dolan, P., 1996. Time-dependent changes in the lumbar spinesresistance to bending. Clinical Biomechanics 11, 194200.
Arnoux, P.J., Subit, D., Masson, C., Chabrand, P., Brunet, C., 2005. Knee ligamentsmechanicsFrom experiments to FE simulation. Revue Europeenne deselements finis 14, 577600.
Bass, C.R., Lucas, S.R., Salzar, R.S., Oyen, M.L., Planchak, C., Shender, B.S., Paskoff, G.,2007. Failure properties of cervical spinal ligaments under fast strain rate
deformations. Spine 32, E7E13.Brown, S.H., Gregory, D.E., McGill, S.M., 2008. Vertebral end-plate fractures as a
result of high rate pressure loading in the nucleus of the young adult porcinespine. Journal of Biomechanics 41, 122127.
Cheung, J.T., Zhang, M., Chow, D.H., 2003. Biomechanical responses of theintervertebral joints to static and vibrational loading: a finite element study.Clinical Biomechanics 18, 790799.
Damkot, D.K., Pope, M.H., Lord, J., Frymoyer, J.W., 1984. The relationship betweenwork history, work environment and low-back pain in men. Spine 9, 395399.
Fathallah, F.A., Marras, W.S., Parnianpour, M., 1998. An assessment of complexspinal loads during dynamic lifting tasks. Spine 23, 706716.
Garo, A., Jundt, G., Arnoux, P.J., Aubin, C.E., El-Rich, M., 2007. Modelisationnumerique de la variabilite des caracteristiques mecaniques des tissus osseuxen situation de choc. In: Proceedings of the XVII Canadian MultidisciplinaryRoad Safety Conference. Montreal, Canada.
Hansen, U., Zioupos, P., Simpson, R., Currey, J.D., Hynd, D., 2008. The effect of strainrate on the mechanical properties of human cortical bone. Journal ofBiomechanical Engineering 130, 17.
Jundt, G., 2007. Modeles dendommagement et de rupture des materiaux
biologiques. Ph.D. Thesis, Universite de la Mediterranee, Marseille.Kasra, M., Shirazi-Adl, A., Drouin, G., 1992. Dynamics of human lumbar
intervertebral joints: experimental and finite-element investigations. Spine17, 93102.
Kelsey, J.L., Githens, P.B., White 3rd, A.A., Holford, T.R., Walter, S.D., OConnor, T.,Ostfeld, A.M., Weil, U., Southwick, W.O., Calogero, J.A., 1984. An epidemiologicstudy of lifting and twisting on the job and risk for acute prolapsed lumbarintervertebral disc. Journal of Orthopedic Research 2, 6166.
Kemper, A.R., McNally, C., Duma, S.M., 2007. The influence of strain rate on thecompressive stiffness properties of human lumbar intervertebral discs.Biomedical Sciences Instrumentation 43, 176181.
Kimpara, H., Iwamoto, M., Watanabe, I., Miki, K., Lee, J.B., Yang, K.H., King, A., 2006.Effect of assumed stiffness and mass density on the impact response of thehuman chest using a three-dimensional FE model of the human body. Journalof Biomechanical Engineering 128, 772776.
Kopperdahl, D.L., Keaveny, T.M., 1998. Yield strain behavior of trabecular bone.Journal of Biomechanics 31, 601608.
Lavander, S.A., Li, Y.C., Andersson, G.B., Natarajan, R.N., 1999. The effects of lifting
speed on the peak external forward bending, lateral bending, and twistingspine moments. Ergonomics 42, 111125.Lee, C.K., Kim, Y.E., Lee, C.S., Hong, Y.M., Jung, J.M., Goel, V.K., 2000. Impact
response of the intervertebral disc in a finite-element model. Spine 25,24312439.
Marras, W.S., Mirka, G.A., 1990. Muscle activities during asymmetric trunk angularaccelerations. Journal of Orthopedic Research 8, 824832.
Neumann, P., Nordwall, A., Osvalder, A.L., 1995. Traumatic instability of the lumbarspine. a dynamic in vitro study of flexion-distraction injury. Spine 20,11111121.
Neumann, P., Osvalder, A.L., Hansson, T.H., Nordwall, A., 1996. Flexion-distractioninjury of the lumbar spine: influence of load, loading rate, and vertebralmineral content. Journal of Spinal Disorders 9, 89102.
Noailly, J., Wilke, H.J., Planell, J.A., Lacroix, D., 2007. How does the geometry affectthe internal biomechanics of a lumbar spine bi-segment finite element model?Consequences on the validation process. Journal of Biomechanics 40,24142425.
Ochia, R.S., Tencer, A.F., Ching, R.P., 2003. Effect of loading rate on endplate andvertebral body strength in human lumbar vertebrae. Journal of Biomechanics
36, 18751881.
ARTICLE IN PRESS
M. El-Rich et al. / Journal of Biomechanics 42 (2009) 12521262 1261
http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.jbiomech.2009.03.036http://localhost/var/www/apps/conversion/tmp/scratch_2/dx.doi.org/10.1016/j.jbiomech.2009.03.036 -
7/27/2019 Finite Element Investigation of the Loadin Grate Effect on the Spinal Load-sharing Changes Under Impact Conditions
11/11
Panjabi, M.M., Crisco 3rd, J.J., Lydon, C., Dvorak, J.,1998. The mechanical propertiesof human alar and transverse ligaments at slow and fast extension rates.Clinical Biomechanics 13, 112120.
Panjabi, M.M., Goel, V.K., Takata, K., 1982. Physiologic strains in the lumbar spinalligaments. An in vitro biomechanical study. Spine 7, 192203.
Pintar, F.A., Yoganandan, N., Myers, T., Elhagediab, A., Sances Jr., A., 1992.Biomechanical properties of human lumbar spine ligaments. Journal ofBiomechanics 25, 13511356.
Qiu, T.X., Tan, K.W., Lee, V.S., Teo, E.C., 2006. Investigation of thoracolumbar T12L1burst fracture mechanism using finite element method. Medical Engineering &Physics 28, 656664.
Rohlmann, A., Bauer, L., Zander, T., Bergmann, G., Wilke, H.J., 2006. Determinationof trunk muscle forces for flexion and extension by using a validated finiteelement model of the lumbar spine and measured in vivo data. Journal ofBiomechanics 39, 981989.
Sances Jr., A., Myklebust, J.B., Maiman, D.J., Larson, S.J., Cusick, J.F., Jodat, R.W., 1984.The biomechanics of spinal injuries. Critical Reviews in Biomedical Engineer-ing 11, 176.
Schmidt, H., Kettler, A., Heuer, F., Simon, U., Claes, L., Wilke, H.J., 2007. Intradiscalpressure, shear strain, and fiber strain in the intervertebral disc undercombined loading. Spine 32, 748755.
Schileo, E., Taddei, F., Cristofolini, L., Viceconti, M., 2008. Subject-specific finiteelement models implementing a maximum principal strain criterion are ableto estimate failure risk and fracture location on human femurs tested in vitro.
Journal of Biomechanics 41 (2), 356367.Shim, V.P.W., Yang, L.M., Liu, J.F., Lee, V.S., 2005. Characterisation of the dynamic
compressive mechanical properties of cancellous bone from the humancervical spine. International Journal of Impact Engineering 32, 525540.
Shirazi-Adl, A., 1989. Strain in fibers of a lumbar disc: analysis of the role of lifting
in producing disc prolapse. Spine 14, 96103.
Shirazi-Adl, A., Ahmed, A.M., Shrivastava, S.C., 1986. Mechanical response of alumbar motion segment in axial torque alone and combined with compression.Spine 11, 914927.
Shirazi-Adl, A., Drouin, G., 1988. Nonlinear gross response analysis of a lumbarmotion segment in combined sagittal loadings. Journal of BiomechanicalEngineering 110, 216222.
Skrzypiec, D., Tarala, M., Pollintine, P., Dolan, P., Adams, M.A., 2007. When areintervertebral discs stronger than their adjacent vertebrae? Spine 32, 24552461.
Tran, N.T., Watson, N.A., Tencer, A.F., Ching, R.P., Anderson, P.A.,1995. Mechanism o fthe burst fracture in the thoracolumbar spine. The effect of loading rate. Spine20, 19841988.
Viano, D.C., Lau, I.V., 1988. A viscous tolerance criterion for soft tissue injuryassessment. Journal of Biomechanics 21, 387399.
Wang, J.L., Parnianpour, M., Shirazi-Adl, A., Engin, A.E., 2000. Viscoelastic finite-element analysis of a lumbar motion segment in combined compression andsagittal flexion. Effect of loading rate. Spine 25, 310318.
Wilcox, R.K., Allen, D.J., Hall, R.M., Limb, D., Barton, D.C., Dickson, R.A., 2004. Adynamic investigation of the burst fracture process using a combinedexperimental and finite element approach. European Spine Journal 13, 481488.
Willen, J., Lindahl, S., Irstam, L., Aldman, B., Nordwall, A., 1984. The thoracolumbarcrush fracture. An experimental study on instant axial dynamic loading: theresulting fracture type and its stability. Spine 9, 624631.
Wirtz, D.C., Schiffers, N., Pandorf, T., Radermacher, K., Weichert, D., Forst, R., 2000.Critical evaluation of known bone material properties to realize anisotropic FE-simulation of the proximal femur. Journal of Biomechanics 33, 13251330.
Yang, K., Zhu, F., Luan, F., Zhao, L., Begeman, P., 1998. Development of a finiteelement model of the human neck. In: Proceeding of the 38th Stapp Car CrashConference. Ft. Lauderdale, FL, pp. 7791.
Yingling, V.R., Callaghan, J.P., McGill, S.M., 1997. Dynamic loading affects the mechanical
properties and failure site of porcine spines. Clinical Biomechanics 12, 301305.
ARTICLE IN PRESS
M. El-Rich et al. / Journal of Biomechanics 42 (2009) 125212621262