Bio-Medical Materials and Engineering 20 (2010) 349359 349

DOI 10.3233/BME-2010-0648
IOS Press

Influence of loading cycle profile and

frequency on a biomechanical parameter of
a model of a balloon kyphoplasty-augmented
lumbar spine segment: A finite element
analysis study
Yuan Li and Gladius Lewis
Department of Mechanical Engineering, The University of Memphis, TN, USA

Received 7 October 2009

Accepted 19 October 2010

Abstract. For patients who are suffering debilitating and persistent pain due to vertebral compression fracture(s) and for whom
conservative therapies have not provided relief, balloon kyphoplasty (BKP) is used as a surgical option. There are only a very
few literature reports on the use of the finite element analysis (FEA) method to obtain biomechanical parameters of models of
spine segments that include BKP augmentation at a given level. In each of these studies, the applied loading used was quasi-
static. During normal activities of daily living, the patients spine would be subject to dynamically-applied loading. Thus, the
question of the influence of the characteristics of a dynamically-applied loading cycle on biomechanical parameters of a spine
that includes BKP-augmented segment(s) is germane; however, a study of this issue is lacking. We investigated this issue in the
present FEA work, with the spine segment model being the L1L3 motion segment units (MSUs) (a segment that is commonly
augmented using BKP) and prophylactic BKP simulated at L2. The dynamic load was the compressive load-versus-time cycle
to which the L3L4 MSU is subjected during gait. Four cases of the cycle were considered, corresponding to slow-, normal-,
fast- and very fast-paced gait. The loading cycle was applied to the superior surface of L1 while the inferior surface of L3 was
fully constrained. It was found that (1) the global mean von Mises stress during the loading cycle (VMG ), in each tissue in the
model increased in going from a slow-paced gait cycle to a very fast-paced gait cycle; and (2) for the slow-paced gait cycle,
with increase in frequency of the cycle, f (1  f  3 Hz), VMG in each of these tissues increased. Potential uses of the
present findings are identified.
Keywords: Balloon kyphoplasty, finite element analysis, von Mises stress

1. Introduction

Pathologic vertebral compression fractures resulting from severe osteoporosis commonly occur at the
T6L5 levels, have a high incidence (for example, in the US, there are 750,000 new cases per an-
num [1]), exert many deleterious effects on patients (such as decrease in the quality of life [2], increase
in morbidity [3] and increase in mortality [4]) and impose very high financial burdens on healthcare sys-
*
Address for correspondence: Dr. Gladius Lewis, Department of Mechanical Engineering, The University of Memphis,
Memphis, TN 38152-3180, USA. E-mail: glewis@memphis.edu.

0959-2989/10/$27.50 2010 IOS Press and the authors. All rights reserved
350 Y. Li and G. Lewis / A finite element analysis study

tems (for example, direct costs in the European Union countries amount to $440 million per year [5]).
In cases where the pain due to these fractures is debilitating and persistent and conservative therapies,
such as analgesics and bracing [6], have not provided sufficient relief, the surgical option of verte-
broplasty (VP) or balloon kyphoplasty (BKP) [7,8] is used. Both procedures involve the percutaneous
injection, under fluoroscopic guidance, of a bolus of a highly-radiopaque injectable bone cement either
directly into the body of the fractured vertebral body, VB (VP) or into a void created in the VB by an
inflatable bone tamp (BKP). Both VP and BKP have been shown to be safe and to provide pain relief
(as demonstrated through, for example, a significant reduction in the Visual Analog Scale pain score)
[911]. However, each procedure has its shortcomings. For example, (1) with VP, there are high inci-
dences of cement extravasation (leakage of the cement out of the augmented VB into peri-augmentation
soft tissues) and new compression fractures [911]; (2) the high pressures that are typically employed
to inject the cement bolus into the VB during VP (up to 1800 kPa [12]) could result in embolization
of the bolus through the venous channels in the VB into the lungs [13]; and (3) the direct medical cost
per VB treated is significantly higher when BKP is used compared to when VP is used [14]. Prospec-
tive, randomized, double-blind, controlled studies directly comparing the outcomes of VP and BKP are
sparse [15,16]. Thus, the choice of procedure used in a given patient depends on guidelines developed by
the physician; for example, Carrino et al. [17] have stated that BKP may be the option to use in patients
with subacute fracture(s) limited to 1 or 2 vertebral levels with loss of VB height of 4050%. On the
basis of the aforementioned observations, we chose to focus, in the present study, on BKP.
The literature on the use of the finite element analysis (FEA) method to study simulated BKP is very
limited [1821]. In all of these reports [1821], the model of the spine segment was subjected to quasi-
static loading. However, in the period of physical therapy following BKP (which, typically, lasts about
two weeks [22]) and in the performance of activities of daily living, the patients spine is subjected to
dynamically-applied loading. The question arises, then, as to the influence of characteristics of a loading
cycle on the mechanical state of the augmented spine segment. The objective of the present study was to
provide an answer to this question, with the segment considered being the L1L3 motion segment units
(MSUs), which are a common site for BKP [23,24], the dynamic loading being that estimated as being
experienced at the L3L4 MSUs during gait [25], and the variables being (1) the profile of the gait load-
ing cycle at a given gait speed and (2) the frequency of a given gait cycle. The biomechanical parameter
determined was the von Mises stress, during the loading cycle, in each of the tissues in the model.

2. Finite element model and analysis

An anatomically correct three-dimensional solid model of the L1L3 MSUs, comprising the VBs,
the endplates, the bony posterior elements (transverse process, pedicles, spinous process, laminae and
facet joints), the intervertebral discs (IVDs) and seven ligaments (see Table 1 for the specific ligaments
Table 1
a
Values of the elastic constants of tissues/materials in the FEA model
Tissue/material Element type Density (kg m3 ) Elastic constantsb
Cortical bone 3-noded triangular 1900 E11 = 1344 MPa; E22 = 1344 MPa
(osteoporotic) general-purpose shell Source: Cowin [28] E33 = 2492 MPa; G12 = 434 MPa
G13 = 491 MPa; G23 = 491 MPa
12 = 0.55; 13 = 0.30; 23 = 0.30
Source: Lewis and Xu [29]
 Y. Li and G. Lewis / A finite element analysis study 351

Table 1
(Continued)
Tissue/material Element type Density (kg m3 ) Elastic constantsb
Cancellous bone 4-noded tetrahedral 575 E11 = 20 MPa; E22 = 14 MPa
(osteoporotic) Source: Cowin [28] E33 = 48 MPa; G12 = 7 MPa
G13 = 6 MPa; G23 = 9 MPa
12 = 0.23; 13 = 0.17; 23 = 0.11
Source: Lewis and Xu [29]
Endplate 4-noded tetrahedral 1900 E = 510 MPa; = 0.40
(osteoporotic) Source: Cowin [28] Source: Lewis and Xu [29]
Annulus fibrosus of 4-noded tetrahedral 700d Ground substance: E = 4.2 MPa; = 0.45;
intervertebral discc tension-only spar Elastic fibers: E = 450 MPa; = 0.30
Source: Zhong et al. [30]
Nucleus pulposus of 4-noded tetrahedral 805e E = 1 MPa; = 0.499
intervertebral disc Source: Zhong et al. [30]
Bony posterior elements 4-noded tetrahedral 1900 E = 2345 MPa; = 0.25
(osteoporotic) Source: Cowin [28] Source: Lewis [33]
Injectable acrylic bone 8-noded brick 1165 E = 3500 MPa; = 0.30
cement Source: Kurtz et al. [34] Source: Kurtz et al. [34]
ALLf Nonlinear tension- E = 14 MPa
(anterior longitudinal only-spar Sources: Zhong et al. [30]; Shin et al. [35]
ligament)
PLLf Nonlinear tension- E = 8 MPa
(posterior longitudinal only-spar Sources: Zhong et al. [30]; Shin et al. [35]
ligament)
LFf Nonlinear tension- E = 9 MPa
(ligamentum flavum) only-spar Sources: Zhong et al. [30]; Shin et al. [35]
ITLf Nonlinear tension- E = 34 MPa
(intertransverse only-spar Sources: Zhong et al. [30]; Shin et al. [35]
ligament)
CLf Nonlinear tension- E = 20 MPa
(capsular ligament) only-spar Sources: Zhong et al. [30]; Shin et al. [35]
ISLf Nonlinear tension- E = 6 MPa
(interspinous ligament) only-spar Sources: Zhong et al. [30]; Shin et al. [35]
SSLf Nonlinear tension- E = 8 MPa;
(supraspinous ligament) only-spar Sources: Zhong et al. [30]; Shin et al. [35]
Notes: a E elastic modulus; Poissons ratio. b 11, 22 and 33 refer to the radial, tangential and longitudinal axes of the
bone, respectively. c Annulus fibrosus taken to be a filamentary composite material comprising a matrix of ground substance
reinforced with elastic fibers (volume fraction: 20%) oriented at 30 with respect to the horizontal plane and arranged in a
criss-cross manner (Zhong et al. [30]). d Estimated using the mean water content of the annulus fibrosus to be 70% [31] and
the density of water = 1000 kg m3 . e Estimated using the mean water content of the nucleus pulposus to be between 78%
[31] and 83% [32] and the density of water = 1000 kg m3 . f For each ligament, (i) no density value is shown because the
ligaments are not physical parts of the finite element mesh; and (ii) the value of E given is the mean of values given by Zhong
et al. [30] and Shin et al. [35].
352 Y. Li and G. Lewis / A finite element analysis study

included), was built using digitized quantitative axial computed tomography (CT) scans/images of a
male cadaver (Visible Human Project dataset; National Library of Medicine, Bethesda, MD, USA),
a commercially-available 3D scanning software package (Mimics , version 8.1; Materialise, Inc., Leu-
ven, Belgium), a commercially-available 3D medical image processing and editing software pack-
age (RapidForm , version 2006; INUS Technology, Inc., Seoul, Korea) and a commercially-available
computer-aided solid modeling software package (ProEngineer Wildfire 2.0; Parametric Technology
Corporation, Needham, MA, USA). Complex bony posterior elements, such as the facet articulation
surfaces, were refined manually. The endplates and the IVDs were generated by extruding the surface
patches on the anterior and posterior surfaces of the VBs. For each IVD, the nucleus pulposus and the
annulus fibrosus occupied 43 and 57% of the total IVD volume, respectively [26].
Prophylactic BKP at L2, whose volume was taken to be 40 ml [27], was simulated with an oblate
spheroid void in the L2 VB, oriented about the midline, filled with the cement. This configuration of
the cement augmentation zone is consistent with views of the cement augmentation zone, as seen in
post-operative CT axial scans and X-radiographs and reported in the literature [22,24]. The cement fill
ratio (ratio of total volume of cement that filled the void to the volume of L2) used was 18%, which is
within the range used in BKP [24].
A commercially-available FEA software package (ABAQUS , version 6.6; Abaqus, Inc., Providence,
RI, USA) was used to generate the finite element meshes from the solid models of the cases when
there was no cement augmentation at L2 (intact model) and when there was (cement-augmented model)
(Fig. 1(A) and (B)). Details of the element type and material property values used for each of the tissues
in these models are given in Table 1. The values of the elastic properties and densities of all the tissues in
the model, as used in the study, are given in Table 1. For each of the two models, the h-convergence test,
based on the mean von Mises stress in each tissue in the model, when it was subjected to a compression
load of 800 N applied quasi-statically as a uniformly distributed load to the superior surface of L1 while
the inferior surface was fully constrained, was used to obtain the final mesh density for the model. The
total numbers of elements in the converged models (Fig. 1(C)) were 432,566 and 440,000 for the intact
and cement-augmented models, respectively.
Validation was conducted using the intact model and it took the form of comparisons between results
obtained using our FEA intact model and experimental results reported by Renner et al. [36], Yamamoto
et al. [37] and Panjabi et al. [38] for each of the following four parameters: (1) the longitudinal displace-
ment of the intervertebral disc at the L2L3 level, under a follower load of 1200 N at L1 (2-3 ), (2) the
range of motion under flexionextension moments (ROMFE), (3) range of motion under axial torsional
moment (ROMTM) and (4) range of motion under lateral bending moment (ROMLB). In our FEA work,
each loading was applied to the superior surface of L1 while the inferior surface of L3 was fully con-
strained. In the case of 2-3 , the experimental result, given by Renner et al. [36], was 1.50 0.80 mm,
while our FEA result was 1.1 mm. Results of the comparisons of the range of the motion parameters
(ROMFE, ROMTM and ROMLB) are given in Fig. 2. Taken as a whole, these results indicate validation
of our FEA model.
The cement-augmented finite element model was subjected to a loading that comprised a compressive
load having a sine wave profile applied to the superior surface of L1, while the inferior surface of L3 was
fully constrained in all degrees of freedom. Four cases were studied, representing the compressive force
cycle exerted on the L3L4 MSU during slow-, normal-, fast- and very fast-paced gait [25] (Fig. 3). For
the FEA, we used a sub-package (Linear Perturbation ) available within the ABAQUS , version 6.6
package. As the output, for each tissue, the variation of the von Mises stress during the loading cycle,
at each of the integration points within the finite element mesh of the tissue was obtained. From this
 Y. Li and G. Lewis / A finite element analysis study 353

Fig. 1. The solid model of the L1L3 motion segment units (A), a cutaway view of the cement augmentation zone (oblate
spheroid) at the L2 level (B) and the converged finite element mesh of the model (C). (Colors are visible in the online version
of the article; http://dx.doi.org/10.3233/BME-2010-0648.)

body of results, we obtained the von Mises stress at the time point at which the maximum value of the
von Mises occurred in the von Mises stress-versus-time plot for each integration point (in fact, this time
point, designated tmax , was the same for all the integration points); we designated this von Mises stress as
the reduced von Mises stress. Finally, we computed the mean of all the reduced von Mises stress values
at all of the integration points and designated this parameter the global mean von Mises stress (VMG ).
In the second part of the study, we determined the influence of the frequency (f ) of the slow-paced gait
cycle (1  f  3 Hz) on VMG in each of these tissues.

3. Results and discussion

Consistent with a priori expectation, the variation of the von Mises stress, in a given tissue in the
model, for a given gait cycle, is sinusoidal (for example, see Fig. 4). In each of the tissues in the model,
VMG increased (1) in going from a slow-paced gait cycle to a very fast-paced gait cycle (Fig. 5); and
(2) with increase in f in the case of the slow-paced gait cycle (Fig. 6).
354 Y. Li and G. Lewis / A finite element analysis study

Fig. 2. (A) Comparison of the main motion results obtained in experimental tests reported by Renner et al. [36] and those
obtained using the present intact finite element model (8 N m flexion moment, 6 N m extension moment, 6 N m clockwise
axial torsional moment, 6 N m counter-clockwise axial torsional moment, 4 N m right lateral bending moment and 4 N m
left lateral bending moment); (B) comparison of the main motion results obtained in experimental tests reported by Yamamoto
et al. [37] and by Panjabi et al. [38] and those obtained using the present intact finite element model (10 N m moment applied
about sagittal plane (flexion/extension), coronal plane (clockwise and counter-clockwise) and transverse plane (left and right)).

Three caveats of the study are recognized. First, only direct compressive load-versus-time data [25]
were used because, to the best of the present workers knowledge, details of the other types of dynamic
loading on the spine, such as lateral bending moment-versus-time and axial torsional moment-versus-
time, are not available in the literature. Second, the experimental data used in the validation work were
obtained using quasi-static loading [3638]. In light of the absence of results from experimental studies
in which lumbar spine segment(s) were subjected to dynamically-applied loading(s), we deemed our
approach acceptable. Furthermore, in the report by Cheung et al. [39], which, to the best of the present
workers knowledge, is the only literature report on the stress analysis of a model of a lumbar spine
segment subjected to a dynamically-applied loading, in which the FEA method was used, there is no
information given on validation of the model. (There is another relevant literature report [40] but that
study did not involve a lumbar spine segment; rather, it involved the use of a finite element model of
 Y. Li and G. Lewis / A finite element analysis study 355

Fig. 3. The compressive load-versus-time curves used (adapted from Nordin and Frankel [25], with subject mass and time of a
normal-paced gait cycle taken to be 61 kg and 1 s, respectively).

Fig. 4. Typical variation of the maximum von Mises stress with time, for a given gait cycle: slow-paced gait cycle; L2 cortical
bone. (Note that maximum von Mises stress at a given time point is the value of the von Mises stress when the von Mises
stress values at all the integration points within the finite element mesh of the tissue, at that time point, are considered.)

a sitting man and the prediction of static compression, dynamic compression, static shear and dynamic
shear forces acting on the lumbar discs during whole-body vibration.) Third, the present work was not an
explicit dynamic FEA study per se; rather, it involved the imposition of a dynamically-applied loading
cycle. In most commercially-available FEA software packages, such as the one used in the present study
(ABAQUS , version 6.6), the mean time increment when the explicit dynamic sub-package is used is
automatically assigned (and, thus, cannot be adjusted by the user) to satisfy stability limit requirements.
For the present L1L2L3 MSU model (comprising on the order of 440,000 elements), the sub-package
(ABAQUS/Explicit ) automatically assigned a mean time increment of 3.2 108 s, which means that
it would have taken 1 year to run one case (on a typical personal computer or workstation, such as
the one we used that has a quad processor with speed and memory of 2.4 GHz and 8 MB SDRAM,
respectively). Clearly, this would make the study prohibitively long to complete. That consideration,
together with the fact that the accuracy of the solution needs to be checked (the explicit dynamic analysis
356 Y. Li and G. Lewis / A finite element analysis study

Fig. 5. A summary of the values of the global mean von Mises stress, during the loading cycle, in each of the tissues in the
model, for the four gait cycles. (Note that tmax for the slow-, normal-, fast- and very fast-paced loading cycles were 0.29, 0.66,
0.51 and 0.41 s, respectively; see text for definition of tmax .)

Fig. 6. A summary of the values of the global von Mises stress, during the loading cycle, in each of the tissues in the model, for
three different frequencies for the slow-paced gait cycle. (Note that tmax when frequencies were 1, 2 and 3 Hz were 0.66, 0.33
and 0.22 s, respectively; see text for definition of tmax .)
 Y. Li and G. Lewis / A finite element analysis study 357

results are not automatically checked), led us to use another sub-package for our FEA study; namely,
ABAQUS Linear Perturbation . This choice of sub-package is acceptable because the present study is
parametric in nature; that is, the interest is the change in the value of VMG in a given tissue when one
input condition, such as loading cycle profile, is changed, all others remaining the same.
It is useful to comment on the similarities and differences in features of the Cheung et al. study [39]
and the present work. In the study by Cheung et al. [39], (1) a model of the intact L4L5 MSU was
constructed; (2) the model consisted of the VBs, the endplates, the IVD, the facet joints and the seven
associated ligaments; (3) the cortical and cancellous bones were each taken to have linear isotropic
elastic properties; (4) there was no information regarding validation of the model; and (5) the model was
subjected to a vibratory force (a sinusoidal pattern with 10% variation of each of three concentrated
forces of 420, 400 and 180 N applied at the anterior boundary, at the center of the surface of L4 and at
the tip of the L4 spinous process, respectively). Cheung et al. [39] reported that, when the frequency of
the vibratory force (f  ) increased from 1 to 4 Hz, (1) the variations of the total facet load (Wf ) and the
mean pore pressure in the nucleus (Pn ) with time were sinusoidal; (2) within the first cycle of loading,
the maximum values of both Wf and Pn occurred at the same time (for example, with f  = 4 Hz, the
aforementioned maxima occurred at 0.06 s); (3) within the first cycle of loading, f  had no influence
on the maximum value of either Wf or Pn ; and (4) within the first cycle of loading, the time at which
the maximum of Wf or Pn occurs, relative to the time for the cycle to be complete, is the same at
f  = 1 Hz as it was at f  = 4 Hz. In the present study, the model was of two MSUs (L1L3); the
model consisted of the VBs, the endplates, the IVDs and the seven associated ligaments; the articulating
tissues of the facet joints were not explicitly included in the model (the hard tissues of the facet joints
were included as part of the bony posterior elements, which also included the spinous processes, the
transverse processes, the pedicles and the laminae); prophylactic BKP was simulated at one level in the
model (L2); the cortical and cancellous bones were taken to have orthotropic elastic properties; the values
of the elastic moduli for the cortical bone, cancellous bone, endplates and bony posterior elements used
were for osteoporotic tissues; the model was validated; and the dynamically-applied compressive force
was uniformly distributed on one location (the superior surface of L1). In both the present study and in
the Cheung et al. study [39], the influence of the frequency of a loading cycle on various biomechanical
parameters were obtained but the trends in the results were different. While, in the present study, VMG
in each tissue of the model increased with increase in frequency (in the case of the slow-paced gait
cycle), Cheung et al. [39] found that, within the first cycle of loading, frequency had no influence on the
maximum value of either Wf or Pn . This difference in trends may be attributed to differences in various
features of the model used in these two studies, as summarized above.
From a computational mechanics perspective, the significance of the present work lies in highlighting
the challenges involved in performing FEA of models of musculoskeletal structures when the loading is
dynamically applied and in demonstrating practical ways of surmounting these challenges, culminating
in a successful analysis. Also, the present results could be used in, for example, developing a physical
therapy regimen for patients in whom BKP is performed, particularly in the first few weeks following
the procedure.

4. Conclusions

For a model of the L1L3 MSUs, with simulated prophylactic BKP at the L2 level, subjected to a
compressive force-versus-time cycle (which the L3L4 MSU experiences during gait) at the superior
358 Y. Li and G. Lewis / A finite element analysis study

surface of L1, while the inferior surface of L3 was fully constrained, (1) the global mean von Mises
stress during the loading cycle (VMG ), in each of the tissues in the model, increased as the profile of
the cycle changed with increase in gait pace; and (2) for the slow-paced gait cycle, with increase in
frequency of the cycle, f (1  f  3 Hz), VMG in each of these tissues increased. These findings have
importance from both computational mechanics and clinical perspectives.

References
[1] N.B. Watts, S.T. Harris and H.K. Genant, Treatment of painful osteoporotic vertebral fractures with percutaneous verte-
broplasty or kyphoplasty, Osteopor. Int. 12 (2001), 429437.
[2] O. Strom, F. Borgstrom, N. Zethraeus, O. Johnell, L. Lidgren, S. Ponzer, O. Svensson, P. Abdon, E. Ornstein, L. Ceder,
K.G. Thorngren, I. Sernbo and B. Jonsson, Long-term cost and effect on quality of life of osteoporosis-related fractures
in Sweden, Acta Orthop. 79 (2008), 269280.
[3] J. Adachi, G. Ioannidis, W. Olszynski, J. Brown, D. Hanley, R.J. Sebaldt, A. Petrie, A. Tenenhouse, G. Stephenson,
A. Papaioannou, G. Guyatt and C. Goldsmith, The impact of incident vertebral and non-vertebral fractures of health
related quality of life in postmenopausal women, BMC Muscul. Disord. (2002), 311.
[4] E. Lau., K. Ong, S. Kurtz, J. Schmier and A. Edidin, Mortality following the diagnosis of vertebral compression fractures
in the Medicare population, J. Bone Joint Surg. 90-B (2008), 14791486.
[5] O. Johnell, Economic implication of osteoporotic spine disease: cost to society, Eur. Spine J. 12(Suppl. 2) (2003), S168
S169.
[6] J.T. Lin and L.M. Lane, Nonmedical management of osteoporosis, Curr. Opin. Rheumatol. 14 (2002), 441446.
[7] D.B. Pateder, A.J. Khanna and I.H. Lieberman, Vertebroplasty and kyphoplasty for the management of osteoporotic
vertebral compression fractures, Orthop. Clin. North Amer. 38 (2007), 409418.
[8] W.C.G. Peh, P.L. Munk and L.A. Gilula, Percutaneous vertebral augmentation: vertebroplasty, kyphoplasty, and skypho-
plasty, Radiol. Clin. North Amer. 46 (2008), 611635.
[9] P.A. Hulme, J. Krebs, S.J. Ferguson and U. Berelemann, Vertebroplasty and kyphoplasty: a systematic review of 69
clinical studies, Spine 31 (2006), 19832001.
[10] R.S. Taylor, R.J. Taylor and P. Fritzell, Balloon kyphoplasty and vertebroplasty for vertebral compression fractures: a com-
parative systematic review of efficacy and safety, Spine 31 (2006), 27472755.
[11] J.C. Eck, D. Nachtigall, S.C. Humphreys and S.D. Hodges, Comparison of vertebroplasty and balloon kyphoplasty for
treatment of vertebral compression fractures: a meta-analysis of the literature, Spine J. 8 (2008), 488497.
[12] G. Baroud, M. Bohner, P. Heini and T. Steffen, Injection biomechanics of bone cements used in vertebroplasty, Bio-Med.
Mater. Eng. 14 (2008), 487504.
[13] W. Lavelle, A. Carl, E.D. Lavelle and M.A. Khaleel, Vertebroplasty and kyphoplasty, Med. Clin. North Amer. 91 (2007),
299314.
[14] H.J. Cloft and M.E. Jensen, Kyphoplasty: an assessment of a new technology, AJNR Am. J. Neuroradiol. 28 (2007),
200203.
[15] C.A.H. Klazen, P.N.M. Lohle, J. de Vries, F.H. Jansen, A.V. Tielbeek et al., Vertebroplasty versus conservative treatment
in acute osteoporotic vertebral compression fractures (VERTOS II): an open-label randomized trial, Lancet 376 (2010),
10851092.
[16] D. Wardlaw, S.R. Cummings, J. Van Meirhaeghe, L. Bastain, J.B. Tilman, J. Ranstam, R. Eastell, P. Shabe, K. Talmadge
and S. Boonen, Efficacy and safety of balloon kyphoplasty compared with non-surgical care for vertebral compression
fracture (FREE): a randomized controlled trial, Lancet 373 (2009), 10161024.
[17] J.A. Carrino, R. Chan and A.R. Vaccaro, Vertebral augmentation: vertebroplasty and kyphoplasty, Semin. Roentgenol. 39
(2004), 6884.
[18] M.L. Villarraga, A.J. Bellezza, T.P. Harrigan, P.A. Cripton, S.M. Kurtz and A.A. Edidin, The biomechanical effects of
kyphoplasty on treated and adjacent nontreated vertebral bodies, J. Spinal Disord. Tech. 18 (2005), 8491.
[19] T.S. Keller, V. Kosmopoulos and I.H. Lieberman, Vertebroplasty and kyphoplasty affect vertebral motion segment stiffness
and stress distributions, Spine 30 (2005), 12581265.
[20] A. Rohlmann, T. Zander and G. Bergmann, Spinal loads after osteoporotic vertebral fractures treated by vertebroplasty or
kyphoplasty, Eur. Spine J. 15 (2006), 12551264.
[21] L. Zhang, G. Yang, L. Wu and B. Yu, The biomechanical effects of osteoporosis vertebral augmentation with cancellous
bone granules or bone cement on treated and adjacent non-treated vertebral bodies: a finite element evaluation, Clin.
Biomech. 25 (2010), 166172.
[22] G. Maestretti, C. Creme, P. Otten and R. Peter, Prospective study of standalone balloon kyphoplasty with calcium phos-
phate cement augmentation in traumatic fractures, Eur. Spine J. 16 (2007), 601610.
 Y. Li and G. Lewis / A finite element analysis study 359

[23] P.F. Heini, B. Walchli and U. Berlemann, Percutaneous transpedicular vertebroplasty with PMMA: operative technique
and early results, Eur. Spine J. 9 (2000), 445450.
[24] I.N. Gaitanis, A.G. Hadjipavlou, P.G. Katonis, M.N. Tzermiadianos, D.S. Pasku and A.G. Patwardhan, Balloon kypho-
plasty for the treatment of pathological vertebral compressive fractures, Eur. Spine J. 14 (2005), 250260.
[25] M. Nordin and V.H. Frankel, Basic Biomechanics of the Musculoskeletal System, 3rd edn, Lippincott Williams & Wilkins,
Philadelphia, PA, USA, 2001, p. 276.
[26] T. Smit, A. Odgaard and E. Schneider, Structure and function of vertebral trabecular bone, Spine 22 (1997), 28232833.
[27] S. Molloy, J.M. Mathis and S.M. Belkoff, The effect of vertebral body percentage fill on mechanical behavior during
percutaneous vertebroplasty, Spine 14 (2003), 15491554.
[28] S.C. Cowin, Bone Mechanics Handbook, 2nd edn, CRC Press, Boca Raton, FL, USA, 2001.
[29] G. Lewis and J. Xu, Rapid and reliable ex vivo biomechanical screening of injectable bone cements for autonomous
augmentation of osteoporotic vertebral bodies: appropriate values of elastic constants for finite element models, J. Biomed.
Mater. Res. Part B: Appl. Biomater. 82B (2007), 408412.
[30] Z. Zhong, S. Wei, J. Wang, C. Feng, C. Chen and C. Yu, Finite element analysis of the lumbar spine with a new cage using
a topology optimization method, Med. Eng. Phys. 28 (2006), 9098.
[31] W.E. Gower and V. Pedrini, Age-related variations in protein-polysaccharide from human nucleus pulposus, annulus
fibrosus, and costal cartilage, J. Bone Joint Surg. 51A (1969), 11541162.
[32] G. Lyons, S.M. Eisenstein and M.B.E. Sweet, Biochemical changes in intervertebral disc degeneration, Biochim. Biophys.
Acta 673 (1981), 443453.
[33] G. Lewis, Estimation of material properties of osteoporotic tissues, Unpublished results, 2007.
[34] S.M. Kurtz, M.L. Villarraga, K. Zhao and A.A. Edidin, Static and fatigue mechanical behavior of bone cement with
elevated barium sulfate content for treatment of vertebral compression fractures, Biomaterials 26 (2005), 36993712.
[Erratum published in Biomater. 29 (2005), 5926.]
[35] D.S. Shin, K. Lee and D. Kim, Biomechanical study of lumbar spine with dynamic stabilization device using finite element
method, Computer-Aided Design 39 (2007), 559567.
[36] S.M. Renner, R.N. Natarajan, A.G. Patwardhan, R.M. Havey, L.I. Vorono, B.Y. Guo, G.B.J. Andersson and H.S. An, Novel
model to analyze the effect of a large compressive follower pre-load on range of motions in a lumbar spine, J. Biomech.
40 (2007), 13261332.
[37] I. Yamamoto, M.M. Panjabi, T. Crisco and T. Oxland, Three-dimensional movements of the whole lumbar spine and
lumbosacral joint, Spine 14 (1989), 12561260.
[38] M.M. Panjabi, T.R. Oxland, I. Yamamoto and J.J. Crisco, Mechanical behavior of the human lumbar and lumbosacral
spine as shown by three-dimensional displacement curves, J. Bone Joint Surg. 76-A (1994), 413424.
[39] J.T.-M. Cheung, M. Zhang and D.H.-K. Chow, Biomechanical responses of the intervertebral joints to static and vibra-
tional loading; a finite element study, Clin. Biomech. 18 (2003), 790799.
[40] H. Seidel, R. Bluthner and B. Hinz, Application of finite-element models to predict forces acting on the lumbar spine
during whole-body vibration, Clin. Biomech. 16(Suppl. 1) (2001), S57S63.
Copyright of Bio-Medical Materials & Engineering is the property of IOS Press and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.