Biomechanical responses due to discitis infection of a juvenile ...

Medical Engineering & Physics 36 (2014) 938–943

Contents lists available at ScienceDirect

Medical Engineering & Physics journal homepage: www.elsevier.com/locate/medengphy

Technical note

Biomechanical responses due to discitis infection of a juvenile thoracolumbar spine using finite element modeling D. Davidson Jebaseelan a,∗ , C. Jebaraj a , N. Yoganandan b , S. Rajasekaran c , J. Yerramshetty c a b c

School of Mechanical and Building Sciences, VIT Chennai, Chennai, India Department of Neurosurgery, Medical College of Wisconsin, Milwaukee, WI, USA Department of Orthopaedics and Spine Surgery, Ganga Hospitals Pvt. Ltd., Coimbatore, India

a r t i c l e

i n f o

Article history: Received 1 May 2013 Received in revised form 8 January 2014 Accepted 12 March 2014 Keywords: Discitis Pediatric spine Finite element analysis Instability Load sharing

a b s t r a c t Growth modulation changes occur in pediatric spines and lead to kyphotic deformity during discitis infection from mechanical forces. The present study was done to understand the consequences of discitis by simulating inflammatory puss at the T12/L1 disc space using a validated eight-year-old thoracolumbar spine finite element model. Changes in the biomechanical responses of the bone, disc and ligaments were determined under physiological compression and flexion loads in the intact and discitis models. During flexion, the angular-displacement increased by 3.33 times the intact spine and localized at the infected junction (IJ). The IJ became a virtual hinge. During compression loading, higher stresses occurred in the growth plate superior to the IJ. The components of the principal stresses in the growth plates at the T12/L1 junction indicated differential stresses. The strain increased by 143% during flexion loading in the posterior ligaments. The study indicates that the flexible pediatric spine increases the motion of the infected spine during physiological loadings. Understanding intrinsic responses around growth plates is important within the context of growth modulation in children. These results are clinically relevant as it might help surgeons to come up with better decisions while developing treatment protocols or performing surgeries. © 2014 IPEM. Published by Elsevier Ltd. All rights reserved.

1. Introduction Pediatric spine disorders may have debilitating consequences on the progressive growth of the spine. Among them discitis is a self-limiting inflammation of the intervertebral disc space caused by low grade viral or bacterial infection [1,2]. The infection changes the physical properties of the involved disc leading to disc space narrowing, changes in the geometry of the adjacent vertebral bodies, erosion of the endplate, and finally to collapse of spinal structures [3,4]. Loss of disc height can be accompanied by changes in spinal curvature and the ensuing spinal instability can cause neurological deficit that impedes individuals from performing daily activities. The pediatric spine differs from the adult spine in its anatomical structures and geometry. These differences alter the biomechanical

∗ Corresponding author at: School of Mechanical and Building Sciences, VIT Chennai, Vandalur-Kelambakkam Road, Chennai 600 127, India. Tel.: +91 44 3993 1555; fax: +91 44 3993 2555; mobile: +91 9445122485. E-mail addresses: [email protected], [email protected] (D. Davidson Jebaseelan). http://dx.doi.org/10.1016/j.medengphy.2014.03.003 1350-4533/© 2014 IPEM. Published by Elsevier Ltd. All rights reserved.

behavior of the pediatric spine. The flexibilities of the pediatric spine, because of its inherent material properties, play a major role during disorders such as disc infections [5–9]. Pediatric spines are immature and as they are in developmental stage, physiological motions during inflammatory conditions like discitis cause excessive movement leading to hypermobility and instability in spine [1–4]. Clinical case studies suggest that posterior structures compensate for the required stability, especially during pathological infections when other structures fail or become less effective. Also some studies indicate the change in stiffness in various bodies during infection [2,10]. The relationship between mechanical factors and growth changes are known [11–13]. Villemure and Stokes’ 2009 studies on growth modulation describe that at the simplest level, bone remodeling is governed by Wolff’s law, while mechanical influence on longitudinal bone growth is controlled by Hueter–Volkmann Law [14]. This mechanical growth modulation can be used to control or optimally reverse musculoskeletal deformities with suitable corrective surgical procedures. The sensitivity of growth plates to their mechanical environment is known [6,14,15]. Also the cited authors mentioned that dynamic compression on the growth plate is essential for bone development, and excessive loading can lead to bone

D. Davidson Jebaseelan et al. / Medical Engineering & Physics 36 (2014) 938–943

939

Fig. 1. (a) Deformed model of the intact and discitis T2/S1 model with discitis modeled at T12/L1 during compression. (b) The superior vertebra settling down over the inferior vertebra at the site of infection has been observed clinically [4].

growth alteration [6,14]. Stokes in 2002 reported that an extreme case of mechanical growth modulation is practiced clinically, when growth plates are stapled. Biomechanical studies on clinical conditions of the adult spine have provided insights on the changes in various responses [16,17]. But there is a lack of similar studies in the pediatric spine. Hence, in this research an anatomically accurate juvenile thoracolumbar (T2/S1) finite element model validated in an earlier study was used to study the consequences of discitis at T12/L1 disc space [8,18]. The first objective of this study was to study the displacement behavior of the discitis model, and the second objective was to study changes in intrinsic responses between the various pediatric components of a thoracolumbar spine due to discitis at the T12/L1 disc space junction. The decrease in the bending stiffness at the infected junction (IJ) would alter the intrinsic responses at the infected region which could have a bearing on the growth of a pediatric spine. The analysis of changes in stresses in various spine components and strains in the ligaments under physiological flexion bending moment and compression loading modes, will add to our present understanding of the consequences of discitis. This information would be useful in evaluating surgical and instrumentation methodologies considered for the treatment of a discitic spine.

from T2 to S1 and 16 discs which comprised of 244,207 elements and 103,593 nodes (Fig. 1a).

2.1. Simulation of discitis at T12/L1 and loading details The inflammatory puss in the IJ, due to discitis in the T12/L1 disc space, and its inferior growth plate were modeled with Young’s modulus and Poisson ratio of 0.1 MPa and 0.4995 [1–4,17,20,25]. The material model was chosen based on the etiology of discitis. Rationale for the choice is provided in Section 4. The thoracolumbar junction, T12/L1 disc, choice as the IJ was based on clinical literature on discitis [20,21]. Compressive force and flexion bending moment loading modes were studied, as these loads were considered to cause increase in kyphosis [22]. A flexion moment of 500 N mm and a compressive load of 100 N were applied in this study. The loads were chosen, based on suggestions from literature [16,17,23,24]. Additional details are provided in Section 4. To simulate the clinical phenomena at the IJ [21], a bonded self-contact was established between the disc and the inferior endplate at the IJ. All degrees-of-freedom at the S1 level were constrained, and the model was solved using the finite element analysis software, Ansys version 10.0 (Ansys Inc.).

2. Methods A three dimensional eight-year-old thorocolumbar (T2/S1) finite element model was used in this study. Model development and validation are detailed in earlier studies and hence not repeated [7–9,18]. The various pediatric components modeled include: the vertebral centrum, cartilaginous growth plate, intervertebral disc, posterior elements, ligaments and the articular cartilage of the facet joints (Fig. 1a). The vertebral centrum, growth plate and the intervertebral disc were modeled using isoparametric eight noded hexahedral elements. The posterior elements along with the facet capsules were modeled with isoparametric four noded tetrahedral elements. The ligaments and annular fibers were modeled using tension-only link elements. Non-linear contact and target elements were assigned to the facet capsules. The pediatric nucleus pulposus was modeled such that it occupies 50% of the disc volume [19]. Idealized juvenile material properties used in earlier studies [5,18] were adopted in the present study (Table 1). The intact pediatric thorocolumbar spine finite element model had 17 vertebral levels

Table 1 Material properties assigned in the current model [5,18]. Cortical bone E (MPa)/Poisson ratio Cancellous bone E (MPa)/Poisson ratio Growth plate E (MPa)/Poisson ratio Posterior complex E (MPa)/Poisson ratio Disc Annulus fibrosis E (MPa)/Poisson ratio Nucleus pulposus E (MPa)/Poisson ratio Ligaments 90% of adult ligament E (MPa) – bilinear model valuesb Anterior longitudinal (AL) Posterior longitudinal (PL) Ligamentum flavum (LF) Transverse ligament (TL) Capsular ligament (CL) Interspinous ligament (ISL) Supraspinous ligament (SSL) a b

Cartilage. Bilinear stress–strain model was used for all ligaments.

75/0.29 75/0.29 25/0.40a 200/0.25 4.2/0.45 1/0.49 14.04, 18 9, 45 13.5, 16.92 9, 53.01 6.75, 29.43 8.82, 10.71 3.78, 13.41

940


100 80 60 40

Annulus Ground substance

20

Nucleus Pulposus Superior growth Plate

0

Inferior growth plate

T2 /T 3 T3 /T 4 T4 /T 5 T5 /T 6 T6 /T 7 T7 /T 8 T8 /T T9 9 /T T1 10 0/ T T1 11 1/ T1 T1 2 2/ L1 L1 /L 2 L2 /L 3 L3 /L 4 L4 /L 5

% Change in Von-mises stress of various components during compression

120

-20 -40 -60 -80

-100 Fig. 2. Change in the Von mises stress of the pediatric components during compression at various levels, with discitis modeled at the T12/L1 disc space.

2.2. Responses studied during sagittal bending moment and compression loading The hypermobility was analyzed by studying changes in the angular displacement in all the functional units due to flexion moment loading. Since its known that intrinsic responses like stresses have correlation with growth modulation, Von-mises stresses were assessed for flexion bending moment and compressive loading in various pediatric components: vertebral bodies, superior and inferior growth plates, and disc components, which included the nucleus pulposus and annulus fibrosus at the IJ. The component stresses that make up the principal stress that includes the directional and shear stresses were also computed at the growth regions to give more insight into the contributing intrinsic responses toward the growth modulation. To further understand the ligament restraints, axial strains of the posterior ligament structures were computed and normalized with respect to the intact case. 3. Results

% Change in angular rotation for the discitis model

The axial displacement during the compression loading increased the curvature of the infected spine (Fig. 1a). The superior vertebra was observed to settle down on the inferior vertebra at the site of the infection (Fig. 1b). The intrinsic responses indicated higher Von-mises stress in the growth plate superior to the IJ, while other pediatric spine components showed reduced stresses (Fig. 2). The changes in the stress distribution of various components at levels adjacent to the IJ were found to be minimal. During flexion, the change in displacement in the discitis model was found to be 3.33

times the angular displacement observed in the intact model at the IJ and it was found to be a localized phenomenon (Fig. 3). The computation of intrinsic responses indicated considerable changes at the T12/L1 level, the IJ. The changes in the Von mises stress for the vertebral centrum, annular ground substance, nucleus pulposus, superior and inferior growth plates and the vertebral bodies were less than the stresses found in the intact condition. The percentage change in the Von-mises stresses for the discitis model with respect to the intact model reduced between 19% for the nucleus pulpous of the disc and 85% for the inferior growth plate (Fig. 4). The posterior ligaments demonstrated an increase in the axial strain, with the percentage change in the strain being 142.67% for the posterior longitudinal ligament, 62.4% for the ligamentum flavum, 130.82% for the interspinous ligament and 129.33% for the supraspinous ligament (Fig. 5). Also, tensile stresses around the facet capsule showed an increase from 0.54 MPa in the intact to 0.93 MPa in the discitis model. A study of the components of the principal stresses, viz., directional and shear stresses in the superior and inferior growth plates during compression loading indicated considerable increases in the normal and shear stresses for the superior growth plate. The increase in stress was found to be 3.5 times higher and was deemed considerable. But there was a reduction in stresses for the inferior growth plate (small in z but considerably higher in shear stress). 4. Discussion Conditions like discitis, osteomyelitis and tuberculosis of the spine are common infections that affect pediatric spines, and these

350

250

150

Flexion

50

-50

T2T3

T3T4

T4T5

T5T6

T6T7

T7T8

T8T9

T9- T10- T11- T12- L1T10 T11 T12 L1 L2

L2L3

L3L4

L4L5

-150

-250

Fig. 3. Change in the angular rotation during flexion bending moment at various levels with discitis modeled at the T12/L1 disc space.


941

Fig. 4. Change in the Von mises stress of pediatric disc components and vertebral bodies during flexion at various levels, with discitis modeled at the T12/L1 disc space.

disorders might develop into kyphotic deformities and interfere with the quality of life of younger populations [4]. Studies using human cadaver experiments and numerical investigations of the pediatric spine for these clinical conditions have the potential to shed light on their biomechanical consequences. The biomechanical evaluation of such clinical conditions on the pediatric spine using numerical methods has not been reported to the best knowledge of the authors of the present study. This study was an attempt to partly fill the gap in the pediatric spine literature. The current model used to study discitis had an anatomically accurate geometry, as it was directly developed from an eight-year-old CT scan data, rather than using scale-down models of adults, which does not have exact geometry and structures. The study also adopted an age-specific appropriate material property [5]. A compression load study of the discitis model revealed that the growth plate superior to IJ sustained greater Von-mises stresses than the intact model and when probed further, the axial compressive stresses were greater at the disc space and superior growth plate. The compressive loading part of the study was associated with the clinical phenomena, showed increase in kyphosis. Also the superior vertebral settling down with wide contact area over the inferior vertebral bodies (fusion) was observed in this modeling study, a phenomenon observed in patients with discitis [4,22] (Fig. 1a and b). The sagittal moment load, with the discitis modeled at the T12/L1, showed greater angulations at the IJ, probably due to the reduction in the bending stiffness at the infected disc space leading to hypermobility. The other important observation, during the

Fig. 5. Change in the axial strain of the posterior longitudinal ligament (PLL), ligamentus flavum (LF), interspinous ligament (ISL), supraspinous ligament (SSL)–during flexion at various levels with discitis modeled at the T12/L1 disc space.

application of flexion moment was that, rotations in the L1/L2 level, an immediate adjacent level below the IJ, had change in rotations in the order of 50% but inferior to these levels the change was found to be minimal. Above the IJ, almost all the functional units experienced reduced rotations of more than 10% and the maximum was up to 70%. The maximum reduced change in rotation of 70% occurred at the level of the thoracic curvature change (T6–T7). Thus, the caudal motion segments, below T12/L1, had normal rotations, while the cephalic motion segments displayed reduced motion range. The displacement results showed that the IJ was the center of the curvature change, may have an accentuated role as described below. The higher displacement reflects the increased flexibility of the pediatric spine. This is very relevant, especially during trauma and physiological motions, as infections like discitis could cause momentary neurological conditions, which could have serious consequences [1–3,25]. From this perspective, the present numerical model has provided clinically pertinent data. The stress distribution on the various pediatric components and strains in the ligaments during the flexion would help to answer the question of instability and the biomechanical understanding of the discitis infection. Figs. 3 and 4 show that the loads during flexion shift toward the posterior soft tissue structures at the T12/L1 motion segment. The changes in the peak stresses with respect to the intact model in various pediatric components were found to be minimal at levels other than the IJ. At the IJ, the disc and the bone components sustained reduced stresses, but the ligament strains substantially increased: the highest increase of 142% was found in the posterior longitudinal ligament, followed by 139% in the supraspinous ligaments. During flexion, four ligamentous structures, viz., the posterior longitudinal ligament, the ligamentum flavum, the interspinous ligament and the supraspinous ligament provides spine stability. So, under traumatic conditions, the flexible pediatric spine experiences increased bending, but the reduced range of motion observed in levels other than IJ is probably due to the inherent ligament behavior. The posterior ligaments were modeled in this study using non-linear material property and tension-only elemental modeling. Thus, in flexion, the increased strain in the posterior ligament structures restricted the motion of the spine during instability, weakening of which could lead to increased kyphosis. These results are also relevant in the context of surgical procedures, where extra care should be taken while operating such that posterior ligaments are preserved. There was an increase in the tensile stress around the L1 superior facet zone and in the posterior region of the vertebral centrum, while anterior regions experienced higher compressive stresses. Similar clinical behavior was reported in the literature [22]. The intrinsic responses of growth plate were studied, in terms of normal and shear stresses, as they are important in the context

942


of growth changes that takes place in a pediatric spine. It was observed from Fig. 2 that superior growth plate had positive % change of Von-mises stresses while the inferior growth plate was having negative % change of Von-mises stresses. The changes in the superior and inferior growth plates in the compression and flexion loading modes are relevant in the growth processes, especially during pathological manifestations. When the inherent stress profile is disrupted, as in pathological situations, overgrowth of the bone in relation to the contralateral bone is a recognized clinical phenomenon [22]. The intrinsic responses from spinal disorders in a growing and remodeling segment are some of the factors in the varied growth patterns which could lead to scoliosis and kyphosis. According to Hueter–Volkmann’s Law, increase in stress is supposed to retard growth and decrease in stress to accelerate growth [12,15,22,26] and the more realistic mechanostat theory [11] point to the role of the intrinsic responses. Incorporating growth modulation theories into the finite element modeling is important and warrants further research. The current study focused on the most prevalent single level discitis infection. The current model can be used to study multi-location infection in the future. The role of facets during physiological loading needs to be studied due to its unique orientation in a pediatric spine. With the understanding that the biomechanical behavior of a pediatric spine with discitis is localized, a few motion segments could be studied with specific importance given for facet modeling. The loads used in the study parallel literature. Suggestions for scaling the loading from the adult to the juvenile model are recognized in literature and it is also known that the adult spine is more stable than a pediatric spine [27–32]. The buckling load for the pediatric spine is lower than the adult spine [7]. Further, the human spinal column is always under the influence of compression and flexion during normal day-to-day activities. Axial compressive loads are present on the thoracolumbar spine due to the presence of the body weight/torso mass. In addition, as this mass located anteriorly with respect to the center of gravity of the vertebral column, flexion is created in the thoracolumbar spine, and this is true for all ages in the human. As the exact loading profile is unknown for the spinal column considered in the present study, methods followed in literature were used, and from this perspective, results from the present study should be considered as a first step in the accurate determination of juvenile spine responses [8,18]. The rationale for the material model is as follows. The etiology presented for the discitis and the metastatic spine and their similarities are available in Rajasekaran [4], Fernandez et al. [3], Hensey et al. [1], Whyne et al. [17] and Tschirhart et al. [32]. This material model was based on the etiology presented clinically and their failure mechanisms. In summary, the modeling of the entire thorocolumbar (T2/S1) model helped to understand the biomechanical behavior of the inferior and superior motion segments. The intrinsic response of a thoracolumbar juvenile FEM with discitis infection at the T12/L1 disc space shows that the stresses in the pediatric components and the ligament strains are localized effects. It appears that the posterior ligamental structures have a role during flexion loading, whereas growth plates during compression loading as these were the most affected areas. The intrinsic responses observed in the growth plates are of important in the context of growth modulation theories, because acceleration or retardation of growth leads to changes in the curvature of spine.

Acknowledgements The authors acknowledge the support of the Professors and engineers of the AU-FRG Institute for CAD/CAM for the work. We thank

Dr. Rishi M. Kanna, Ganga Hospitals, Coimbatore and Dr. S. Radhakrishnan, Professor (Retd.), IIT Madras, India for their suggestions in the text. Funding: Nil. Conflict of interest: None declared. Ethical approval: Not required. References [1] Hensey OJ, Coad N, Carty HM, Sills JM. Juvenile discitis. Arch Dis Child 1983;58:983–7. [2] Gonzalvo A, Abdulla I, Riazi A, Harpe DDL. Single-level/single-stage debridement and posterior instrumented fusion in the treatment of spontaneous pyogenic osteomyelitis/discitis long-term functional outcome and healthrelated quality of life. J Spinal Disord Tech 2011;24(2):110–5. [3] Fernandez M, Carrol CL, Baker CJ. Discitis and vertebral osteomyelitis in children: an 18-year review. Pediatrics 2000;105:1299–304. [4] Rajasekaran S. Buckling collapse of the spine in childhood spinal tuberculosis. Clin Orthop 2007;460:86–92. [5] Kumaresan S, Yoganandan N, Pintar FA, Maiman DJ, Kuppa S. Biomechanical study of pediatric human cervical spine: a finite element approach. J Biomech Eng 2000;122(1):060–71. [6] Sylvestre PL, Villemure I, Aubin CE. Finite element modeling of the growth plate in a detailed spine model. Med Biol Eng Comput 2007;45:977–88. [7] Davidson JD, Jebaraj C, Yoganandan N, Rajasekaran S. Quantification of segmental flexibilities of juvenile lumbar spinal column using a three-dimensional finite element model (FEM). In: World congress on biomechanics, Springer IFMBE proceedings, vol. 31. 2010. p. 525–8. [8] Davidson JD, Jebaraj C, Yoganandan N, Rajasekaran S. Validation efforts and flexibilities of an eight-year-old human juvenile lumbar spine using a threedimensional finite element model. Med Biol Eng Comput 2010;48:1223–31. [9] Jebaraj C, Davidson JD, Yoganandan N, Rajasekaran S. Finite element study of collapse of pediatric spine due to tuberculosis of spine. In: Proceedings of the international mechanical engineering congress & exposition. 2010. p. 1–7. [10] Spiegel PG, Kengla KW, Isaacson AS, Wilson JC. Intervertebral disc-space inflammation in children. J Bone Joint Surg Am 1972;54:284–96. [11] Frost HM. From Wolff’s law to the mechanostat: a new face of physiology. J Orthop Sci 1998;3:282–6. [12] Stokes IAF. Mechanical effects on skeletal growth. J Musculoskel Neuron Interact 2002;2(3):277–80. [13] Jee WSS. Principles of bone physiology. J Musculoskel Neuron Interact 2000;1:11–3. [14] Villemure I, Stokes IAF. Growth plate mechanics and mechanobiology, a survey of present understanding. J Biomech 2009;42:1793–803. [15] Gooding CA, Neuhauser EBD. Growth and development of the vertebral body in the presence and absence of normal stress. Am J Roentgenol Radium Ther Nucl Med 1965;93:388–94. [16] Kumaresan S, Yoganandan N, Pintar FA, Maiman DJ, Goel VK. Contribution of disc degeneration to osteophyte formation in the cervical spine: a biomechanical investigation. J Orthop Res 2001;19:977–84. [17] Whyne CM, Hu SS, Lotz JC. Parametric finite element analysis of vertebral bodies affected by tumors. J Biomech 2001;34:1317–24. [18] Davidson Jebaseelan D, Jebaraj C, Yoganandan N, Rajasekaran S, Kanna RM. Sensitivity studies of pediatric material properties on juvenile lumbar spine responses using finite element analysis. Med Biol Eng Comput 2012;50: 515–22. [19] Weinstein S. Pediatric spine: principles and practice. New York: Raven Press; 1994. [20] Kersley JB. Non tuberculosis infection of the spine. Proc R Soc Med 1976;70:177–81. [21] Halpern EM, Bacon SA, Kitagawa T, Stephen JL. Posterior transdiscal threecolumn shortening in the surgical treatment of vertebral discitis/osteomyelitis with collapse. Spine 2010;35(13):1316–22. [22] Rajasekaran S, Natarajan RN, Babu NJ, Kanna RMP, Shetty AP, Andersson GBJ. Lumbar vertebral growth is governed by “chondral growth force response curve” rather than “Hueter-Volkmann Law”: a clinico-biomechanical study of growth modulation changes in childhood spinal tuberculosis. Spine 2011;36(22):1435–45. [23] Skogland LB, Miller JA. On the importance of growth in idiopathic scoliosis, Technical report. Sweden: University of Oslo; 1980. [24] Clarke EC, appleyard RC, Bilston LE. Immature sheep spines are more flexible than mature spine: an in vitro biomechanical study. Spine 2007;32(26): 2970–9. [25] Early SD, Kayand RM, Tolo VT. Childhood diskitis. J Am Acad Orthop Surg 2003;11:413–20. [26] Lin H, Aubin CE, Parent S, Villemure I. Mechanobiological bone growth: comparative analysis of two biomechanical modeling approaches. Med Biol Eng Comput 2009;47:357–66. [27] Sances Jr A, Myklebust JB, Larson SJ, Cusick JF, Weber RC, Walsh PR. Bioengineering analysis of head and spine injuries. CRC Crit Rev Bioeng 1981;2(15): 1–79. [28] Sances Jr A, Thomas DJ, Ewing CL, Larson SJ, Unterharnscheidt F, editors. Mechanisms of head and spine trauma. Goshen, NY: Aloray; 1984. p. 189–218.

D. Davidson Jebaseelan et al. / Medical Engineering & Physics 36 (2014) 938–943 [29] Yoganandan N, Pintar FA, Sances Jr A, Maiman DJ, Myklebust J, Harris G, et al. Biomechanical investigations of the human thoracolumbar spine. SAE Trans 1989;97:676–81. [30] Yoganandan N, Kumaresan S, Pintar F, Gennarelli T. Pediatric biomechanics. In: Nahum A, Melvin J, editors. Accidental injury: biomechanics and prevention. New York: Springler-Verlag; 2000. p. 550–87.

943

[31] Yoganandan N, Myklebust, Ray G, Pintar F, Sances Jr A. A non-linear finite element model of a spinal segment. Math Model 1987;8:617–22. [32] Tschirhart CE, Finkelstein JA, Whyne CM. Metastatic burst fracture risk assessment based on complex loading of the thoracic spine. Ann Biomed Eng 2006;34(3):494–505.