j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Biomechanical effects of cement distribution in the fractured area on osteoporotic vertebral compression fractures: a three-dimensional finite element analysis De Liang, MM,a,1 Lin-Qiang Ye, MB,a,b,1 Xiao-Bing Jiang, MD,a,c,* Pan Yang, MB,b,d Guang-Quan Zhou, MD,a,c Zhen-Song Yao, MD,a Shun-Cong Zhang, MM,a and Zhi-Dong Yang, MBa a
Department of Spinal Surgery, The First Affiliated Hospital of Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, People’s Republic of China b Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, People’s Republic of China c Department of Digital Orthopaedics and Biomechanics, Laboratory Affiliated to National Key Discipline of Orthopaedics and Traumatology of Chinese Medicine, Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, People’s Republic of China d Orthopaedics Hospital, Guangzhou General Hospital of Guangzhou Military Command of PLA, Guangzhou, Guangdong, People’s Republic of China
article info
abstract
Article history:
Background: According to some clinical studies, insufficient cement distribution (ID) in the
Received 15 August 2014
fractured area and asymmetrical cement distribution around the fractured area were
Received in revised form
thought to be the reasons for unrelieved pain and recollapse after percutaneous vertebral
7 December 2014
augmentation (PVA) in the treatment of symptomatic osteoporotic vertebral compression
Accepted 31 December 2014
fractures.
Available online 7 January 2015
Methods: Finite element methods were used to investigate the biomechanical variance among three patterns of cement distribution (ID and sufficient cement distribution in the
Keywords:
fractured area and asymmetrical cement distribution around the fractured area including
Osteoporotic vertebral compression
upward [BU] and downward [BD] cement distribution).
fracture Percutaneous vertebral augmentation
Results: Compared with fractured vertebra before PVA, distribution of von Mises stress in the cancellous bone was transferred to be concentrated at the cancellous bone surrounding cement after PVA, whereas it was not changed in the cortical bone. Compared with suf-
Cement distribution
ficient cement distribution group, maximum von Mises stress in the cancellous bone and
Finite element analysis
cortical bone and maximum displacement of augmented vertebra increased significantly in
Fractured area
the ID group, whereas asymmetrical cement distribution around the fractured area in BU and BD groups mainly increased maximum von Mises stress in the cancellous bone significantly. Similar results could be seen in all loading conditions.
* Corresponding author. Department of Spinal Surgery, The First Affiliated Hospital of Guangzhou University of Chinese Medicine, Airport Road 16, Guangzhou 510405, Guangdong, China. Tel.: þ86 13632494486; fax: þ86 02036591604. E-mail address: [email protected] (X.-B. Jiang). 1 Co-first authors: D.L. and L.-Q.Y. contributed equally to this work. 0022-4804/$ e see front matter ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jss.2014.12.053
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
247
Conclusions: ID in the fractured area may lead to unrelieved pain after PVA in the treatment of symptomatic osteoporotic vertebral compression fractures as maximum displacement of augmented vertebral body increased significantly. Both ID in the fractured area and asymmetrical cement distribution around the fractured area are more likely to induce recollapse of augmented vertebra because they increased maximum von Mises stress in the cancellous bone and cortical bone of augmented vertebra significantly. ª 2015 Elsevier Inc. All rights reserved.
1.
Introduction
Osteoporotic vertebral compression fractures (OVCFs) are very common in the elderly, with an estimated 1.4 million new fractures occurring every year worldwide [1]. Until recently, symptomatic OVCFs were treated commonly with conservative methods including bed rest, analgesics, braces, and physical therapy. However, percutaneous vertebral augmentation (PVA), such as percutaneous vertebroplasty (PVP) and percutaneous kyphoplasty (PKP), has been introduced as an alternative treatment option [2,3]. Biomechanical studies showed significant increases in the stiffness and strength of individual-fractured vertebra after PVA [4,5]. Apart from rapid pain relief, another immediate effect of PVA was an increase of anterior vertebral height [6e8], which reduced kyphosis in patients [8]. The realigned spinal column and regained height in the augmented vertebra may decrease pulmonary and gastrointestinal complications and early morbidity related to compression fractures [9]. However, some studies reported that pain could not be relieved after PVA [10], and recollapse of the augmented vertebral bodies had been observed in some patients during the follow-ups [11,12]. Insufficient cement distribution (ID) in the fractured area was thought to be the reason for unrelieved pain, and asymmetrical cement distribution around the fractured area was assumed to be the main risk factor of recollapse of the augmented vertebral bodies [10e12]. To date, however, few biomechanical studies have been performed to research the reasons why ID in the fractured area and asymmetrical cement distribution around the fractured area can induce previously mentioned complications. Moreover, a better understanding of this biomechanical behavior is critical in optimizing the ultimate result of our treatment. The purpose of this study was to investigate biomechanical effects of ID in the fractured area and asymmetrical cement distribution around the fractured area on symptomatic OVCFs.
2.
Materials and methods
computer and then imported to Mimics software (version 14.11; Materialise, Inc, Leuven, Belgium) for generation of the 3-dimensional FE model of T11-L1 vertebra, including cortical (1-mm-thick) and cancellous bone and posterior elements. The geometry of other structures (the annulus fibrosis, nucleus pulpous, facet cartilage, and end plate), which were difficult to separate from the CT images, were modeled using the solid modeling software, SolidWorks 2012 (SolidWorks Corp, Dassault Systemes, Concord, MA). The nucleus pulpous occupied 43% of the total disc [13]. The element types of cortical bone, cancellous bone, bony end plate, facet joint cartilage, annulus, and nucleus pulpous were defined as solid elements. Seven different ligaments including anterior longitudinal ligament, posterior longitudinal ligament, interspinous ligament, supraspinal ligament, capsular ligament, ligamentum flavum, and intertransverse ligaments in tension only were modeled with truss elements. These elements were orientated along the respective ligament directions obtained from anatomic textbooks. The assigned material properties were assumed to be linear, homogeneous, and isotropic. Tied contact interfaces were used to ensure the disc and ligament attachment to the vertebra and to prevent any relative movement during the simulations. Surface-based, finite-sliding contact with a friction coefficient 0.0026 was defined for facet joints. As a result, the model of the T11-L1 was developed, consisting of 147,355 solid elements, 388 truss elements, and a total of 235,594 nodes. The validation of the normal model was conducted according to the published FE model and human cadaveric thoracolumbar spines. The inferior end plate of L1 vertebra was fixed in all degrees of freedom. Pure moment of 7.5 Nm was applied on the superior end plate of T11 for validation. Because vertebral compression fractures are related clinically to osteoporosis, a model of an osteoporotic T11-L1 was built. According to the methods reported by Polikeit et al. [14], a model with osteoporosis was defined as follows. The elastic moduli of all bony structures were reduced by 66% for the cancellous bone and by 33% for the cortical shell, the end plates, and the posterior elements. The other structures were left unchanged. The material properties of different components are listed in Table [14e17].
2.1. The construction of normal and osteoporotic T11-L1 finite element model
2.2.
A normal three-dimensional digital anatomic finite element (FE) model of T11-L1 was built using digitized image data of a T11-L1 motion segment. The image data of T11-L1 were obtained from a computed tomography (CT) scan of the thoracolumbar spine from a healthy volunteer who had no abnormal findings on roentgenograms and were taken at 1-mm intervals. The slice images were preserved in a
Similar to simulation methods reported by Chiang et al. [18], the model was constructed with the following steps to simulate compression fracture on T12. The cleft was horizontally penetrated into the vertebral body by 20 mm through the center of the anterior cortical shell. The size of the cleft was approximately 20, 30, and 2 mm in depth, width, and height, respectively (Fig. 1A).
The simulation of compression fracture
248
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
Table e The material properties of the FE models. Component name Cortical bone Cancellous bone Posterior structure Bony endplate Cartilage Nucleus pulposus Annulus ALL PLL LF ISL SSL ITL CL Cement
Young modulus (MPa)
Poisson ratio
Cross-section area (mm2)
Status
8040 (67% normal) 34 (34% normal) 2345 (67% normal) 670 (67% normal) 10 1 4.2 20 70 50 28 28 50 26 3000
0.3 0.2 0.25 0.4 0.4 0.4999 0.45 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.4
d d d d d d d 60 21 60 40 30 10 67.5 d
Osteoporotic Osteoporotic Osteoporotic Osteoporotic Normal Normal Normal Normal Normal Normal Normal Normal Normal Normal Grafting
ALL ¼ anterior longitudinal ligament; CL ¼ capsular ligament; ISL ¼ interspinous ligament; ITL ¼ intertransverse ligament; LF ¼ ligamentum flavum; PLL ¼ posterior longitudinal ligament; SSL ¼ supraspinal ligament.
2.3.
The simulation of cement augmentation
For bone cement, the material properties of polymethylmethacrylate were applied (Young modulus, 3000 MPa; Poisson ratio, 0.4) because polymethylmethacrylate represents the most common clinically used material. Two cement cylinders with the same volume were vertically implanted into the fractured vertebra to mimic bipedicular PVA. The volume of each cement cylinder was approximately 2 mL. For ID in the fractured area, one cement cylinder was implanted around the fractured area symmetrically, whereas the other was not across the fractured area (Fig. 1B). In contrast, both cement cylinders were embedded symmetrically around the fractured area to mimic symmetric and sufficient cement distribution (SD; Fig. 1C). Being different from SD, the two cement cylinders, which were still across the fractured area, were relocated upward (BU) or downward (BD) in the vertebral body to simulate asymmetrical cement distribution (Fig. 1D and E). Eventually, we got five different models for the test, including the fractured model before PVA and four augmented models.
2.4.
Boundary and loading conditions of FE models
Considering the effects of the paraspinal muscles and intraabdominal pressure, all models were implemented with a 500-N vertical compression load for a balanced standing [19,20]; a pure moment of 7.5 Nm combined with the precompressive load of 500 N was implemented for flexion, extension, and left and/or right lateral bending. According to the spinal three-column concept, the load and moment were applied to the superior end plate and articular facets of T11, with 85% of these on the anterior-middle column and 15% on the posterior column [16,21]. All computational processes were performed with Abaqus software (version 6.12; Abaqus, Inc, Providence, RI). The magnitudes and distributions of the von Mises stress in the cortical bone and cancellous bone and maximum displacement of T12 vertebral body were recorded. Von Mises stress has been proposed as a parameter of failure criteria for the bone [14] and maximum displacement a parameter of stability [22].
3.
Results
3.1.
Validation of the normal FE model
Under flexion, extension, left and/or right lateral bending, and axial rotation, the range of motions predicted by our FE model of T11wL1 were 7.0 , 4.5 , 7.5 , and 3.1 , which were similar to those published in the literature [23e25].
3.2. The distributions and magnitudes of the von Mises stress in cortical bone of T12 in fractured models before and after PVA The results of distributions of the von Mises stress under vertical compression in the cortical bone of T12 in fractured models before and after PVA are shown in Figure 2. Compared with the distribution before PVA, it was unchanged after PVA, which was still concentrated at the posterior unfractured area. Similar results could be seen in the flexion, extension, and left and/or right lateral bending. Under vertical compression, maximum von Mises stress in the cortical bone of T12 in ID, SD, BU, and BD groups was about 53.64%, 47.34%, 50.25%, and 49.12% of that in the fractured model before PVA, respectively. Therefore, comparing BU, BS, and SD groups, asymmetrical cement distribution around the fractured area slightly increased maximum von Mises stress. However, comparing ID and SD groups, ID in the fractured area increased maximum von Mises stress significantly. Similar changes could be seen in flexion, extension, and left and/or right lateral bending (Figs. 3 and 4).
3.3. The magnitudes and distributions of the von Mises stress in cancellous bone of T12 for fractured models before and after PVA The results of distributions of the von Mises stress under vertical compression in the cancellous bone of T12 for fractured models before and after PVA are shown in Figure 5 and demonstrate that von Mises stress was concentrated at the
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
249
Fig. 1 e Fractured models before and after PVA. (A) A cleft horizontally penetrated into the vertebral body simulating compression fracture (Pre-augmented). (B) One cement cylinder implanted around the fractured area symmetrically, whereas the other not across the fractured area simulating ID in the fractured area. (C) Both cement cylinders embedded symmetrically around the fractured area mimicking symmetric cement distribution and SD. (D and E) Both cement cylinders located upward (BU) or downward (BD) in the vertebral body but still across the fractured area simulating asymmetrical cement distribution. (Color version of figure is available online.)
posterior unfractured area before PVA but transferred to be concentrated at the cancellous bone surrounding the bone cement after PVA. Furthermore, variance was notable among the augmented models, in which von Mises stress was concentrated below and above the fractured area symmetrically in ID and SD groups, whereas it was mainly concentrated at the side of the cancellous bone with less cement distribution in BU and BD groups. Similar results could be seen in flexion, extension, and left and/or right lateral bending. Under vertical compression, maximum von Mises stress in the cancellous bone of T12 in ID, SD, BU, and BD groups was about 242.18%, 119.35%, 313.41%, and 266.05% of that in the fractured model before PVA, respectively. Therefore, compared with the SD group, both ID in the fractured area and
asymmetrical cement distribution around the fractured area could significantly increase the maximum von Mises stress in the cancellous bone. Similar changes could be seen in flexion, extension, and left and/or right lateral bending (Figs. 6 and 7).
3.4. The maximum displacement of T12 for fractured models before and after PVA Under vertical compression, maximum displacement of T12 in ID, SD, BU, and BD groups was about 51.07%, 38.18%, 40.22%, and 40.18% of that in the fractured model before PVA, respectively. Therefore, comparing BU, BS, and SD groups, asymmetrical cement distribution around the fractured area slightly increased maximum displacement. However,
250
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
Fig. 2 e The nephograms of the von Mises stress under vertical compression in the cortical bone of T12 for fractured models before and after PVA showing PVA did not change the distribution of the von Mises stress in the cortical bone. (AeE) The nephograms of pre-augmented, ID, SD, BU, BD groups all showing the von Mises stress was concentrated at the posterior unfractured area. (Color version of figure is available online.)
comparing ID and SD groups, ID in the fractured area increased maximum displacement significantly. Similar changes could be seen in flexion, extension, and left and/or right lateral bending (Figs. 8 and 9).
4.
Discussion
Controversy still exists with respect to the overall benefits of PVP. Clinical practice guidelines published by the Society of Interventional Radiology in 2003 [26], Cardiovascular and
Interventional Radiological Society of Europe in 2006 [27], and the American College of Radiology in 2014 [28], respectively, all recommended PVP for symptomatic OVCFs refractory to conservative or traditional management. However, two highprofile, multicenter, prospective, randomized controlled trials compared PVP with a sham control group, both of which found that patients randomized to PVP did not experience decreases in pain or disability relative to patients in the placebo arm [29,30]. Although these two studies have been criticized for a variety of reasons, the American Academy of Orthopaedic Surgeons believed that these two studies did
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
251
Fig. 3 e The maximum von Mises stress in the cortical bone of T12 for pre-augmented, ID, and SD FE models. (Color version of figure is available online.)
have sufficient power to detect the minimal clinically important difference in pain, and clinical practice guideline issued by the American Academy of Orthopaedic Surgeons strongly recommended against PVP for symptomatic VOCFs [31]. To the best of our knowledge, there have not been studies or clinical practice guidelines recommending against PKP for symptomatic OVCFs refractory to conservative or traditional management so far. In the present study, we studied biomechanical effects of cement distribution in the fractured area on OVCFs, not only for PVP but also for PKP. Thus, findings from this study might provide references, at least, for PKP in the treatment of symptomatic OVCFs in the clinical practice when controversy still exists with respect to the overall benefits of PVP around the world. This FE model was developed on the data collected from a spinal CT scan of a healthy volunteer’s thoracolumbar region
and biomechanical material properties reflecting the pathologic characteristics of vertebral osteoporosis. A validated three-vertebra segment (T11-T12-L1) was constructed, but not individual vertebra (T12), because a three-vertebra segment model with intervertebral discs and facet joints might not only highly simulate the motion and load transfer of the thoracolumbar junction of the spine when compression fracture was simulated in the middle vertebra but also avoid loading positions and boundary conditions being directly connected to the target vertebra of T12, which may have an influence on biomechanical behavior of T12 [32]. The validation test proved that the constructed three-dimensional FE model could accurately simulate physiological activity at the thoracolumbar region and, therefore, could be a valuable tool for later research. Although stress measurements were not validated experimentally, this could not be an influential factor in our
Fig. 4 e The maximum von Mises stress in the cortical bone of T12 for SD, BD, and BU FE models. (Color version of figure is available online.)
252
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
Fig. 5 e The nephograms of the von Mises stress under vertical compression in the cancellous bone of T12 for fractured models before and after PVA showing von Mises stress, which was concentrated at posterior unfractured area before PVA, was transferred to be concentrated at the cancellous bone surrounding bone cement after PVA. (A) The nephogram of preaugmented group showing the von Mises stress was concentrated at the posterior unfractured area. (B and C) The nephograms of ID and SD groups showing von Mises stress was concentrated at the cancellous bone surrounding cement below and above the fractured area symmetrically. (D and E) The nephograms of BU and BD groups showing von Mises stress was concentrated at the side of the cancellous bone with less cement distribution. (Color version of figure is available online.)
study because the parameters in this study were compared in terms of relative differences among fractured models before and after PVA with different distributing patterns of cement. The anterior cortex of compressed vertebra is usually injured and not continuous before union. In the present study, we attempted to re-create a compression fracture model where a cleft was horizontally penetrated into a T12 vertebral
body through the center of the anterior cortical shell. This failure model may be attributed to the type A1.2 fracture, that is, the wedge impaction fracture [33]. For bipedicular augmentation, two vertically orientated cement cylinders were implanted across the fractured area in the center of the cancellous bone, but not only in the fractured area. That is because cement not only distributed in the fractured area but
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
253
Fig. 6 e The maximum von Mises stress in the cancellous bone of T12 for pre-augmented, ID, and SD FE models. (Color version of figure is available online.)
also interdigitated into the surrounding cancellous bone as we observed on radiographs of treated patients. Furthermore, similar methods to simulate PVA had been reported by Polikeit et al. [14] and Liebschner et al. [34], and they also agreed that this shape was comparable with the cement distribution seen on radiographs of treated patients. We acknowledge that using cylinders to simulate PVA is done for the ease of computer simulation, and not a direct reflection of the reality of PVA, because the simulation of PVA in the present study can be easily created through the FE analysis method and ensures the repeatability of the study. We are aware that different shapes of cement might occur in the vertebral bodies. Thus, preliminary experiments, in which another shape of cement (cement cake) with equivalent volume was used to simulate PVA, had been done to analyze how cement shape affects findings in the present study. We found that different stress
and displacement can be produced using cement cakes to simulate PVA, but the same conclusion can be drawn using cement cakes to simulate PVA-like cement cylinders. It is expected that the current simulation of injury creation and cement augmentation could be the reasonable scenario of PVA procedure, after which the injured vertebra restored to its original height [18]. Our present study, focusing on the magnitudes and distributions of the von Mises stress in cortical bone and the cancellous bone and maximum displacement of T12 for fractured models before and after PVA, showed that von Mises stress in the cancellous bone was transferred to be concentrated at the cancellous bone surrounding cement and maximum von Mises stress increased in the cancellous bone, decreased in the cortical bone after PVA, which is consistent with an FE analysis conducted by Polikeit et al. [14]. Thus, this
Fig. 7 e The maximum von Mises stress in the cancellous bone of T12 for SD, BD, and BU FE models. (Color version of figure is available online.)
254
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
Fig. 8 e The maximum displacement for the vertebral body of T12 for pre-augmented, ID, and SD FE models. (Color version of figure is available online.)
may, to some extent, validate that findings in our study using these models to simulate injury creation and cement augmentation were reasonable. Compared with the SD group, maximum von Mises stress in the cortical bone and cancellous bone and maximum displacement of T12 vertebral body were increased significantly in the ID group. Increase in maximum displacement indicates that stability cannot be well restored due to ID in the fractured area, which is in agreement with the finding from a clinical study conducted by He et al. [10] who claimed absent or inadequate cement filling in the fractured area of the vertebral body may be responsible for the unrelieved pain after the initial PVA because stabilization of micro-motions in the fractured vertebra was thought to be the main mechanism for pain relief after PVA. In addition, although pain can be relieved after PVA with ID in the fractured area; from the perspective of biomechanics,
recollapse is more likely to happen in these augmented vertebral bodies because of the fact that maximum von Mises stress and displacement of the augmented vertebra increased in the case of ID in the fractured area. Unfortunately, as far as we know, no clinical study regarding ID in the fractured area has been performed to support our finding except one previous study conducted by us [35] and published in a Chinese peer-reviewed journal, in which we found that augmented vertebral bodies with ID in the fractured area were inclined to recollapse during 1-y follow-ups in comparison with those with cement filled sufficiently in the fractured area. Theoretically, if the adequate amount of cement is injected into the vertebral bodies and sufficiently distributed in the fractured area, the height of augmented vertebral bodies could be retained after PVA. However, recollapse has been observed even in these augmented vertebral bodies. A previous clinical
Fig. 9 e The maximum displacement for the vertebral body of T12 for SD, BD, and BU FE models. (Color version of figure is available online.)
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
study reported 13.3% of vertebral height loss at augmented vertebral bodies at mean follow-ups of 27 mo, in which asymmetrical cement distribution around the fractured area was defined as a main risk factor for the reason that asymmetrical cement distribution around the fractured area more probably damaged the cancellous bone in the unfilled area [12]. Our study showed that, compared with SD group, asymmetrical cement distribution around the fractured area did not increase maximum von Mises stress in the cortical bone and maximum displacement of the augmented vertebral body significantly; however, it increased maximum von Mises stress in the cancellous bone surrounding cement significantly, which indicates symmetrical cement distribution around the fractured area may provide better structural support and decrease incidence of recollapse. Although previously mentioned findings in this study might be meaningful for the clinical practice, some limitations of this study need to be mentioned. First, the assumptions of linear, isotropic, and homogeneous material properties for FE models and cement were a simplification of the real situation. Different stress and displacement may be produced with a more physiologic material distribution. Even so, we believe that an alteration in the stress and displacement because of different cement distributions still would occur. Second, wedge-like vertebra-simulating compression fracture was not created, which may have an impact on stress and displacement. Therefore, only the idealized status of the anterior vertebral height correction was simulated in this study; however, PVA only partially restores the vertebral shape in most cases. Future work evaluating the resulting differences would be desirable. Furthermore, clinical study evaluating the findings from this study also would be expected in the future.
5.
Conclusions
ID in the fractured area may lead to unrelieved pain after PVA because of the fact that it increases displacement of augmented vertebral body significantly. Both ID in the fractured area and asymmetrical cement distribution around the fractured area are more likely to induce recollapse of augmented vertebral bodies because maximum von Mises stress in cancellous bone and cortical bone of augmented vertebral body increased significantly. Hence, to guarantee pain relief after PVA and decrease incidence of recollapse in the coming years, SD in the fractured area should be involved in the surgical plan of PVA, and symmetrical cement distribution around the fractured area might be the optimal pattern we should pursue in the clinical practice.
Acknowledgment The authors gratefully acknowledge the support from Guangdong Province Administration of Traditional Chinese Medicine Program (NO. 20111185), Project of The Health Ministry of China (NO. W2012ZT07), and Guangdong Province Medical Science and Technology Research Program (NO. B2014175).
255
The authors also thank Prof Zhihui Pang for the advice of biomechanics. Authors’ contributions: D.L. and X.-B.J. contributed to conceiving and designing the study. L.-Q.Y., P.Y., and G.-Q.Z. performed the experiment. X.-B.J., Z.-S.Y., S.-C.Z., and Z.-D.Y. analyzed the data. X.-B.J., D.L., and L.-Q.Y. play the main role in writing the article.
Disclosure The authors reported no proprietary or commercial interest in any product mentioned or concept discussed in the article.
references
[1] Johnell O, Kanis JA. An estimate of the worldwide prevalence and disability associated with osteoporotic fractures. Osteoporos Int 2006;17:1726. [2] Gray LA, Jarvik JG, Heagerty PJ, et al. INvestigational Vertebroplasty Efficacy and Safety Trial (INVEST): a randomized controlled trial of percutaneous vertebroplasty. BMC Musculoskelet Disord 2007;8:126. [3] Steinmann J, Tingey CT, Cruz G, et al. Biomechanical comparison of unipedicular versus bipedicular kyphoplasty. Spine (Phila Pa 1976) 2005;30:201. [4] Belkoff SM, Mathis JM, Jasper LE, et al. The biomechanics of vertebroplasty. The effect of cement volume on mechanical behavior. Spine (Phila Pa 1976) 2001;26:1537. [5] Chen B, Li Y, Xie D, et al. Comparison of unipedicular and bipedicular kyphoplasty on the stiffness and biomechanical balance of compression fractured vertebrae. Eur Spine J 2011; 20:1272. [6] Hiwatashi A, Moritani T, Numaguchi Y, et al. Increase in vertebral body height after vertebroplasty. AJNR Am J Neuroradiol 2003;24:185. [7] McKiernan F, Faciszewski T, Jensen R. Reporting height restoration in vertebral compression fractures. Spine (Phila Pa 1976) 2003;28:2517. [8] Teng MM, Wei CJ, Wei LC, et al. Kyphosis correction and height restoration effects of percutaneous vertebroplasty. AJNR Am J Neuroradiol 2003;24:1893. [9] Garfin SR, Yuan HA, Reiley MA. New technologies in spine: kyphoplasty and vertebroplasty for the treatment of painful osteoporotic compression fractures. Spine (Phila Pa 1976) 2001;26:1511. [10] He SC, Teng GJ, Deng G, et al. Repeat vertebroplasty for unrelieved pain at previously treated vertebral levels with osteoporotic vertebral compression fractures. Spine (Phila Pa 1976) 2008;33:640. [11] Kim YY, Rhyu KW. Recompression of vertebral body after balloon kyphoplasty for osteoporotic vertebral compression fracture. Eur Spine J 2010;19:1907. [12] Lin WC, Lee YC, Lee CH, et al. Refractures in cemented vertebrae after percutaneous vertebroplasty: a retrospective analysis. Eur Spine J 2008;17:592. [13] Polikeit A, Ferguson SJ, Nolte LP, et al. Factors influencing stresses in the lumbar spine after the insertion of intervertebral cages: finite element analysis. Eur Spine J 2003; 12:413. [14] Polikeit A, Nolte LP, Ferguson SJ. The effect of cement augmentation on the load transfer in an osteoporotic functional spinal unit: finite-element analysis. Spine (Phila Pa 1976) 2003;28:991.
256
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
[15] Baroud G, Nemes J, Heini P, et al. Load shift of the intervertebral disc after a vertebroplasty: a finite-element study. Eur Spine J 2003;12:421. [16] Goel VK, Kong W, Han JS, et al. A combined finite element and optimization investigation of lumbar spine mechanics with and without muscles. Spine (Phila Pa 1976) 1993;18:1531. [17] Noailly J, Lacroix D, Planell JA. Finite element study of a novel intervertebral disc substitute. Spine (Phila Pa 1976) 2005; 30:2257. [18] Chiang CK, Wang YH, Yang CY, et al. Prophylactic vertebroplasty may reduce the risk of adjacent intact vertebra from fatigue injury: an ex vivo biomechanical study. Spine (Phila Pa 1976) 2009;34:356. [19] Rohlmann A, Zander T, Rao M, et al. Applying a follower load delivers realistic results for simulating standing. J Biomech 2009;42:1520. [20] Rohlmann A, Zander T, Rao M, et al. Realistic loading conditions for upper body bending. J Biomech 2009;42:884. [21] Denis F. The three column spine and its significance in the classification of acute thoracolumbar spinal injuries. Spine (Phila Pa 1976) 1983;8:817. [22] Li QL, Li XZ, Liu Y, et al. Treatment of thoracolumbar fracture with pedicle screws at injury level: a biomechanical study based on three-dimensional finite element analysis. Eur J Orthop Surg Traumatol 2013;23:775. [23] Panjabi MM, Oxland TR, Lin RM, et al. Thoracolumbar burst fracture. A biomechanical investigation of its multidirectional flexibility. Spine (Phila Pa 1976) 1994;19:578. [24] Qiu TX, Teo EC, Zhang QH. Effect of bilateral facetectomy of thoracolumbar spine T11-L1 on spinal stability. Med Biol Eng Comput 2006;44:363. [25] Xu G, Fu X, Du C, et al. Biomechanical effects of vertebroplasty on thoracolumbar burst fracture with
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
transpedicular fixation: a finite element model analysis. Orthop Traumatol Surg Res 2014;100:379. McGraw JK, Cardella J, Barr JD, et al. Society of Interventional Radiology quality improvement guidelines for percutaneous vertebroplasty. J Vasc Interv Radiol 2003;14:S311. Gangi A, Sabharwal T, Irani FG, et al. Quality assurance guidelines for percutaneous vertebroplasty. Cardiovasc Intervent Radiol 2006;29:173. McConnell CT Jr, Wippold FJ 2nd, Ray CE Jr, et al. ACR appropriateness criteria management of vertebral compression fractures. J Am Coll Radiol 2014;11:757. Buchbinder R, Osborne RH, Ebeling PR, et al. A randomized trial of vertebroplasty for painful osteoporotic vertebral fractures. N Engl J Med 2009;361:557. Kallmes DF, Comstock BA, Heagerty PJ, et al. A randomized trial of vertebroplasty for osteoporotic spinal fractures. N Engl J Med 2009;361:569. McGuire R. AAOS clinical practice guideline: the treatment of symptomatic osteoporotic spinal compression fractures. J Am Acad Orthop Surg 2011;19:183. Wijayathunga VN, Oakland RJ, Jones AC, et al. Vertebroplasty: patient and treatment variations studied through parametric computational models. Clin Biomech (bristol, Avon) 2013;28:860. Magerl F, Aebi M, Gertzbein SD, et al. A comprehensive classification of thoracic and lumbar injuries. Eur Spine J 1994;3:184. Liebschner MA, Rosenberg WS, Keaveny TM. Effects of bone cement volume and distribution on vertebral stiffness after vertebroplasty. Spine (Phila Pa 1976) 2001;26:1547. Jiang XB, Mo L, Liang D, et al. The distribution of bone cement in the fractured lines on the outcome of percutaneous vertebroplasty. Chin J Spine Spinal Cord 2014;24:144.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Biomechanical effects of cement distribution in the fractured area on osteoporotic vertebral compression fractures: a three-dimensional finite element analysis De Liang, MM,a,1 Lin-Qiang Ye, MB,a,b,1 Xiao-Bing Jiang, MD,a,c,* Pan Yang, MB,b,d Guang-Quan Zhou, MD,a,c Zhen-Song Yao, MD,a Shun-Cong Zhang, MM,a and Zhi-Dong Yang, MBa a
Department of Spinal Surgery, The First Affiliated Hospital of Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, People’s Republic of China b Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, People’s Republic of China c Department of Digital Orthopaedics and Biomechanics, Laboratory Affiliated to National Key Discipline of Orthopaedics and Traumatology of Chinese Medicine, Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, People’s Republic of China d Orthopaedics Hospital, Guangzhou General Hospital of Guangzhou Military Command of PLA, Guangzhou, Guangdong, People’s Republic of China
article info
abstract
Article history:
Background: According to some clinical studies, insufficient cement distribution (ID) in the
Received 15 August 2014
fractured area and asymmetrical cement distribution around the fractured area were
Received in revised form
thought to be the reasons for unrelieved pain and recollapse after percutaneous vertebral
7 December 2014
augmentation (PVA) in the treatment of symptomatic osteoporotic vertebral compression
Accepted 31 December 2014
fractures.
Available online 7 January 2015
Methods: Finite element methods were used to investigate the biomechanical variance among three patterns of cement distribution (ID and sufficient cement distribution in the
Keywords:
fractured area and asymmetrical cement distribution around the fractured area including
Osteoporotic vertebral compression
upward [BU] and downward [BD] cement distribution).
fracture Percutaneous vertebral augmentation
Results: Compared with fractured vertebra before PVA, distribution of von Mises stress in the cancellous bone was transferred to be concentrated at the cancellous bone surrounding cement after PVA, whereas it was not changed in the cortical bone. Compared with suf-
Cement distribution
ficient cement distribution group, maximum von Mises stress in the cancellous bone and
Finite element analysis
cortical bone and maximum displacement of augmented vertebra increased significantly in
Fractured area
the ID group, whereas asymmetrical cement distribution around the fractured area in BU and BD groups mainly increased maximum von Mises stress in the cancellous bone significantly. Similar results could be seen in all loading conditions.
* Corresponding author. Department of Spinal Surgery, The First Affiliated Hospital of Guangzhou University of Chinese Medicine, Airport Road 16, Guangzhou 510405, Guangdong, China. Tel.: þ86 13632494486; fax: þ86 02036591604. E-mail address: [email protected] (X.-B. Jiang). 1 Co-first authors: D.L. and L.-Q.Y. contributed equally to this work. 0022-4804/$ e see front matter ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jss.2014.12.053
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
247
Conclusions: ID in the fractured area may lead to unrelieved pain after PVA in the treatment of symptomatic osteoporotic vertebral compression fractures as maximum displacement of augmented vertebral body increased significantly. Both ID in the fractured area and asymmetrical cement distribution around the fractured area are more likely to induce recollapse of augmented vertebra because they increased maximum von Mises stress in the cancellous bone and cortical bone of augmented vertebra significantly. ª 2015 Elsevier Inc. All rights reserved.
1.
Introduction
Osteoporotic vertebral compression fractures (OVCFs) are very common in the elderly, with an estimated 1.4 million new fractures occurring every year worldwide [1]. Until recently, symptomatic OVCFs were treated commonly with conservative methods including bed rest, analgesics, braces, and physical therapy. However, percutaneous vertebral augmentation (PVA), such as percutaneous vertebroplasty (PVP) and percutaneous kyphoplasty (PKP), has been introduced as an alternative treatment option [2,3]. Biomechanical studies showed significant increases in the stiffness and strength of individual-fractured vertebra after PVA [4,5]. Apart from rapid pain relief, another immediate effect of PVA was an increase of anterior vertebral height [6e8], which reduced kyphosis in patients [8]. The realigned spinal column and regained height in the augmented vertebra may decrease pulmonary and gastrointestinal complications and early morbidity related to compression fractures [9]. However, some studies reported that pain could not be relieved after PVA [10], and recollapse of the augmented vertebral bodies had been observed in some patients during the follow-ups [11,12]. Insufficient cement distribution (ID) in the fractured area was thought to be the reason for unrelieved pain, and asymmetrical cement distribution around the fractured area was assumed to be the main risk factor of recollapse of the augmented vertebral bodies [10e12]. To date, however, few biomechanical studies have been performed to research the reasons why ID in the fractured area and asymmetrical cement distribution around the fractured area can induce previously mentioned complications. Moreover, a better understanding of this biomechanical behavior is critical in optimizing the ultimate result of our treatment. The purpose of this study was to investigate biomechanical effects of ID in the fractured area and asymmetrical cement distribution around the fractured area on symptomatic OVCFs.
2.
Materials and methods
computer and then imported to Mimics software (version 14.11; Materialise, Inc, Leuven, Belgium) for generation of the 3-dimensional FE model of T11-L1 vertebra, including cortical (1-mm-thick) and cancellous bone and posterior elements. The geometry of other structures (the annulus fibrosis, nucleus pulpous, facet cartilage, and end plate), which were difficult to separate from the CT images, were modeled using the solid modeling software, SolidWorks 2012 (SolidWorks Corp, Dassault Systemes, Concord, MA). The nucleus pulpous occupied 43% of the total disc [13]. The element types of cortical bone, cancellous bone, bony end plate, facet joint cartilage, annulus, and nucleus pulpous were defined as solid elements. Seven different ligaments including anterior longitudinal ligament, posterior longitudinal ligament, interspinous ligament, supraspinal ligament, capsular ligament, ligamentum flavum, and intertransverse ligaments in tension only were modeled with truss elements. These elements were orientated along the respective ligament directions obtained from anatomic textbooks. The assigned material properties were assumed to be linear, homogeneous, and isotropic. Tied contact interfaces were used to ensure the disc and ligament attachment to the vertebra and to prevent any relative movement during the simulations. Surface-based, finite-sliding contact with a friction coefficient 0.0026 was defined for facet joints. As a result, the model of the T11-L1 was developed, consisting of 147,355 solid elements, 388 truss elements, and a total of 235,594 nodes. The validation of the normal model was conducted according to the published FE model and human cadaveric thoracolumbar spines. The inferior end plate of L1 vertebra was fixed in all degrees of freedom. Pure moment of 7.5 Nm was applied on the superior end plate of T11 for validation. Because vertebral compression fractures are related clinically to osteoporosis, a model of an osteoporotic T11-L1 was built. According to the methods reported by Polikeit et al. [14], a model with osteoporosis was defined as follows. The elastic moduli of all bony structures were reduced by 66% for the cancellous bone and by 33% for the cortical shell, the end plates, and the posterior elements. The other structures were left unchanged. The material properties of different components are listed in Table [14e17].
2.1. The construction of normal and osteoporotic T11-L1 finite element model
2.2.
A normal three-dimensional digital anatomic finite element (FE) model of T11-L1 was built using digitized image data of a T11-L1 motion segment. The image data of T11-L1 were obtained from a computed tomography (CT) scan of the thoracolumbar spine from a healthy volunteer who had no abnormal findings on roentgenograms and were taken at 1-mm intervals. The slice images were preserved in a
Similar to simulation methods reported by Chiang et al. [18], the model was constructed with the following steps to simulate compression fracture on T12. The cleft was horizontally penetrated into the vertebral body by 20 mm through the center of the anterior cortical shell. The size of the cleft was approximately 20, 30, and 2 mm in depth, width, and height, respectively (Fig. 1A).
The simulation of compression fracture
248
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
Table e The material properties of the FE models. Component name Cortical bone Cancellous bone Posterior structure Bony endplate Cartilage Nucleus pulposus Annulus ALL PLL LF ISL SSL ITL CL Cement
Young modulus (MPa)
Poisson ratio
Cross-section area (mm2)
Status
8040 (67% normal) 34 (34% normal) 2345 (67% normal) 670 (67% normal) 10 1 4.2 20 70 50 28 28 50 26 3000
0.3 0.2 0.25 0.4 0.4 0.4999 0.45 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.4
d d d d d d d 60 21 60 40 30 10 67.5 d
Osteoporotic Osteoporotic Osteoporotic Osteoporotic Normal Normal Normal Normal Normal Normal Normal Normal Normal Normal Grafting
ALL ¼ anterior longitudinal ligament; CL ¼ capsular ligament; ISL ¼ interspinous ligament; ITL ¼ intertransverse ligament; LF ¼ ligamentum flavum; PLL ¼ posterior longitudinal ligament; SSL ¼ supraspinal ligament.
2.3.
The simulation of cement augmentation
For bone cement, the material properties of polymethylmethacrylate were applied (Young modulus, 3000 MPa; Poisson ratio, 0.4) because polymethylmethacrylate represents the most common clinically used material. Two cement cylinders with the same volume were vertically implanted into the fractured vertebra to mimic bipedicular PVA. The volume of each cement cylinder was approximately 2 mL. For ID in the fractured area, one cement cylinder was implanted around the fractured area symmetrically, whereas the other was not across the fractured area (Fig. 1B). In contrast, both cement cylinders were embedded symmetrically around the fractured area to mimic symmetric and sufficient cement distribution (SD; Fig. 1C). Being different from SD, the two cement cylinders, which were still across the fractured area, were relocated upward (BU) or downward (BD) in the vertebral body to simulate asymmetrical cement distribution (Fig. 1D and E). Eventually, we got five different models for the test, including the fractured model before PVA and four augmented models.
2.4.
Boundary and loading conditions of FE models
Considering the effects of the paraspinal muscles and intraabdominal pressure, all models were implemented with a 500-N vertical compression load for a balanced standing [19,20]; a pure moment of 7.5 Nm combined with the precompressive load of 500 N was implemented for flexion, extension, and left and/or right lateral bending. According to the spinal three-column concept, the load and moment were applied to the superior end plate and articular facets of T11, with 85% of these on the anterior-middle column and 15% on the posterior column [16,21]. All computational processes were performed with Abaqus software (version 6.12; Abaqus, Inc, Providence, RI). The magnitudes and distributions of the von Mises stress in the cortical bone and cancellous bone and maximum displacement of T12 vertebral body were recorded. Von Mises stress has been proposed as a parameter of failure criteria for the bone [14] and maximum displacement a parameter of stability [22].
3.
Results
3.1.
Validation of the normal FE model
Under flexion, extension, left and/or right lateral bending, and axial rotation, the range of motions predicted by our FE model of T11wL1 were 7.0 , 4.5 , 7.5 , and 3.1 , which were similar to those published in the literature [23e25].
3.2. The distributions and magnitudes of the von Mises stress in cortical bone of T12 in fractured models before and after PVA The results of distributions of the von Mises stress under vertical compression in the cortical bone of T12 in fractured models before and after PVA are shown in Figure 2. Compared with the distribution before PVA, it was unchanged after PVA, which was still concentrated at the posterior unfractured area. Similar results could be seen in the flexion, extension, and left and/or right lateral bending. Under vertical compression, maximum von Mises stress in the cortical bone of T12 in ID, SD, BU, and BD groups was about 53.64%, 47.34%, 50.25%, and 49.12% of that in the fractured model before PVA, respectively. Therefore, comparing BU, BS, and SD groups, asymmetrical cement distribution around the fractured area slightly increased maximum von Mises stress. However, comparing ID and SD groups, ID in the fractured area increased maximum von Mises stress significantly. Similar changes could be seen in flexion, extension, and left and/or right lateral bending (Figs. 3 and 4).
3.3. The magnitudes and distributions of the von Mises stress in cancellous bone of T12 for fractured models before and after PVA The results of distributions of the von Mises stress under vertical compression in the cancellous bone of T12 for fractured models before and after PVA are shown in Figure 5 and demonstrate that von Mises stress was concentrated at the
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
249
Fig. 1 e Fractured models before and after PVA. (A) A cleft horizontally penetrated into the vertebral body simulating compression fracture (Pre-augmented). (B) One cement cylinder implanted around the fractured area symmetrically, whereas the other not across the fractured area simulating ID in the fractured area. (C) Both cement cylinders embedded symmetrically around the fractured area mimicking symmetric cement distribution and SD. (D and E) Both cement cylinders located upward (BU) or downward (BD) in the vertebral body but still across the fractured area simulating asymmetrical cement distribution. (Color version of figure is available online.)
posterior unfractured area before PVA but transferred to be concentrated at the cancellous bone surrounding the bone cement after PVA. Furthermore, variance was notable among the augmented models, in which von Mises stress was concentrated below and above the fractured area symmetrically in ID and SD groups, whereas it was mainly concentrated at the side of the cancellous bone with less cement distribution in BU and BD groups. Similar results could be seen in flexion, extension, and left and/or right lateral bending. Under vertical compression, maximum von Mises stress in the cancellous bone of T12 in ID, SD, BU, and BD groups was about 242.18%, 119.35%, 313.41%, and 266.05% of that in the fractured model before PVA, respectively. Therefore, compared with the SD group, both ID in the fractured area and
asymmetrical cement distribution around the fractured area could significantly increase the maximum von Mises stress in the cancellous bone. Similar changes could be seen in flexion, extension, and left and/or right lateral bending (Figs. 6 and 7).
3.4. The maximum displacement of T12 for fractured models before and after PVA Under vertical compression, maximum displacement of T12 in ID, SD, BU, and BD groups was about 51.07%, 38.18%, 40.22%, and 40.18% of that in the fractured model before PVA, respectively. Therefore, comparing BU, BS, and SD groups, asymmetrical cement distribution around the fractured area slightly increased maximum displacement. However,
250
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
Fig. 2 e The nephograms of the von Mises stress under vertical compression in the cortical bone of T12 for fractured models before and after PVA showing PVA did not change the distribution of the von Mises stress in the cortical bone. (AeE) The nephograms of pre-augmented, ID, SD, BU, BD groups all showing the von Mises stress was concentrated at the posterior unfractured area. (Color version of figure is available online.)
comparing ID and SD groups, ID in the fractured area increased maximum displacement significantly. Similar changes could be seen in flexion, extension, and left and/or right lateral bending (Figs. 8 and 9).
4.
Discussion
Controversy still exists with respect to the overall benefits of PVP. Clinical practice guidelines published by the Society of Interventional Radiology in 2003 [26], Cardiovascular and
Interventional Radiological Society of Europe in 2006 [27], and the American College of Radiology in 2014 [28], respectively, all recommended PVP for symptomatic OVCFs refractory to conservative or traditional management. However, two highprofile, multicenter, prospective, randomized controlled trials compared PVP with a sham control group, both of which found that patients randomized to PVP did not experience decreases in pain or disability relative to patients in the placebo arm [29,30]. Although these two studies have been criticized for a variety of reasons, the American Academy of Orthopaedic Surgeons believed that these two studies did
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
251
Fig. 3 e The maximum von Mises stress in the cortical bone of T12 for pre-augmented, ID, and SD FE models. (Color version of figure is available online.)
have sufficient power to detect the minimal clinically important difference in pain, and clinical practice guideline issued by the American Academy of Orthopaedic Surgeons strongly recommended against PVP for symptomatic VOCFs [31]. To the best of our knowledge, there have not been studies or clinical practice guidelines recommending against PKP for symptomatic OVCFs refractory to conservative or traditional management so far. In the present study, we studied biomechanical effects of cement distribution in the fractured area on OVCFs, not only for PVP but also for PKP. Thus, findings from this study might provide references, at least, for PKP in the treatment of symptomatic OVCFs in the clinical practice when controversy still exists with respect to the overall benefits of PVP around the world. This FE model was developed on the data collected from a spinal CT scan of a healthy volunteer’s thoracolumbar region
and biomechanical material properties reflecting the pathologic characteristics of vertebral osteoporosis. A validated three-vertebra segment (T11-T12-L1) was constructed, but not individual vertebra (T12), because a three-vertebra segment model with intervertebral discs and facet joints might not only highly simulate the motion and load transfer of the thoracolumbar junction of the spine when compression fracture was simulated in the middle vertebra but also avoid loading positions and boundary conditions being directly connected to the target vertebra of T12, which may have an influence on biomechanical behavior of T12 [32]. The validation test proved that the constructed three-dimensional FE model could accurately simulate physiological activity at the thoracolumbar region and, therefore, could be a valuable tool for later research. Although stress measurements were not validated experimentally, this could not be an influential factor in our
Fig. 4 e The maximum von Mises stress in the cortical bone of T12 for SD, BD, and BU FE models. (Color version of figure is available online.)
252
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
Fig. 5 e The nephograms of the von Mises stress under vertical compression in the cancellous bone of T12 for fractured models before and after PVA showing von Mises stress, which was concentrated at posterior unfractured area before PVA, was transferred to be concentrated at the cancellous bone surrounding bone cement after PVA. (A) The nephogram of preaugmented group showing the von Mises stress was concentrated at the posterior unfractured area. (B and C) The nephograms of ID and SD groups showing von Mises stress was concentrated at the cancellous bone surrounding cement below and above the fractured area symmetrically. (D and E) The nephograms of BU and BD groups showing von Mises stress was concentrated at the side of the cancellous bone with less cement distribution. (Color version of figure is available online.)
study because the parameters in this study were compared in terms of relative differences among fractured models before and after PVA with different distributing patterns of cement. The anterior cortex of compressed vertebra is usually injured and not continuous before union. In the present study, we attempted to re-create a compression fracture model where a cleft was horizontally penetrated into a T12 vertebral
body through the center of the anterior cortical shell. This failure model may be attributed to the type A1.2 fracture, that is, the wedge impaction fracture [33]. For bipedicular augmentation, two vertically orientated cement cylinders were implanted across the fractured area in the center of the cancellous bone, but not only in the fractured area. That is because cement not only distributed in the fractured area but
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
253
Fig. 6 e The maximum von Mises stress in the cancellous bone of T12 for pre-augmented, ID, and SD FE models. (Color version of figure is available online.)
also interdigitated into the surrounding cancellous bone as we observed on radiographs of treated patients. Furthermore, similar methods to simulate PVA had been reported by Polikeit et al. [14] and Liebschner et al. [34], and they also agreed that this shape was comparable with the cement distribution seen on radiographs of treated patients. We acknowledge that using cylinders to simulate PVA is done for the ease of computer simulation, and not a direct reflection of the reality of PVA, because the simulation of PVA in the present study can be easily created through the FE analysis method and ensures the repeatability of the study. We are aware that different shapes of cement might occur in the vertebral bodies. Thus, preliminary experiments, in which another shape of cement (cement cake) with equivalent volume was used to simulate PVA, had been done to analyze how cement shape affects findings in the present study. We found that different stress
and displacement can be produced using cement cakes to simulate PVA, but the same conclusion can be drawn using cement cakes to simulate PVA-like cement cylinders. It is expected that the current simulation of injury creation and cement augmentation could be the reasonable scenario of PVA procedure, after which the injured vertebra restored to its original height [18]. Our present study, focusing on the magnitudes and distributions of the von Mises stress in cortical bone and the cancellous bone and maximum displacement of T12 for fractured models before and after PVA, showed that von Mises stress in the cancellous bone was transferred to be concentrated at the cancellous bone surrounding cement and maximum von Mises stress increased in the cancellous bone, decreased in the cortical bone after PVA, which is consistent with an FE analysis conducted by Polikeit et al. [14]. Thus, this
Fig. 7 e The maximum von Mises stress in the cancellous bone of T12 for SD, BD, and BU FE models. (Color version of figure is available online.)
254
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
Fig. 8 e The maximum displacement for the vertebral body of T12 for pre-augmented, ID, and SD FE models. (Color version of figure is available online.)
may, to some extent, validate that findings in our study using these models to simulate injury creation and cement augmentation were reasonable. Compared with the SD group, maximum von Mises stress in the cortical bone and cancellous bone and maximum displacement of T12 vertebral body were increased significantly in the ID group. Increase in maximum displacement indicates that stability cannot be well restored due to ID in the fractured area, which is in agreement with the finding from a clinical study conducted by He et al. [10] who claimed absent or inadequate cement filling in the fractured area of the vertebral body may be responsible for the unrelieved pain after the initial PVA because stabilization of micro-motions in the fractured vertebra was thought to be the main mechanism for pain relief after PVA. In addition, although pain can be relieved after PVA with ID in the fractured area; from the perspective of biomechanics,
recollapse is more likely to happen in these augmented vertebral bodies because of the fact that maximum von Mises stress and displacement of the augmented vertebra increased in the case of ID in the fractured area. Unfortunately, as far as we know, no clinical study regarding ID in the fractured area has been performed to support our finding except one previous study conducted by us [35] and published in a Chinese peer-reviewed journal, in which we found that augmented vertebral bodies with ID in the fractured area were inclined to recollapse during 1-y follow-ups in comparison with those with cement filled sufficiently in the fractured area. Theoretically, if the adequate amount of cement is injected into the vertebral bodies and sufficiently distributed in the fractured area, the height of augmented vertebral bodies could be retained after PVA. However, recollapse has been observed even in these augmented vertebral bodies. A previous clinical
Fig. 9 e The maximum displacement for the vertebral body of T12 for SD, BD, and BU FE models. (Color version of figure is available online.)
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
study reported 13.3% of vertebral height loss at augmented vertebral bodies at mean follow-ups of 27 mo, in which asymmetrical cement distribution around the fractured area was defined as a main risk factor for the reason that asymmetrical cement distribution around the fractured area more probably damaged the cancellous bone in the unfilled area [12]. Our study showed that, compared with SD group, asymmetrical cement distribution around the fractured area did not increase maximum von Mises stress in the cortical bone and maximum displacement of the augmented vertebral body significantly; however, it increased maximum von Mises stress in the cancellous bone surrounding cement significantly, which indicates symmetrical cement distribution around the fractured area may provide better structural support and decrease incidence of recollapse. Although previously mentioned findings in this study might be meaningful for the clinical practice, some limitations of this study need to be mentioned. First, the assumptions of linear, isotropic, and homogeneous material properties for FE models and cement were a simplification of the real situation. Different stress and displacement may be produced with a more physiologic material distribution. Even so, we believe that an alteration in the stress and displacement because of different cement distributions still would occur. Second, wedge-like vertebra-simulating compression fracture was not created, which may have an impact on stress and displacement. Therefore, only the idealized status of the anterior vertebral height correction was simulated in this study; however, PVA only partially restores the vertebral shape in most cases. Future work evaluating the resulting differences would be desirable. Furthermore, clinical study evaluating the findings from this study also would be expected in the future.
5.
Conclusions
ID in the fractured area may lead to unrelieved pain after PVA because of the fact that it increases displacement of augmented vertebral body significantly. Both ID in the fractured area and asymmetrical cement distribution around the fractured area are more likely to induce recollapse of augmented vertebral bodies because maximum von Mises stress in cancellous bone and cortical bone of augmented vertebral body increased significantly. Hence, to guarantee pain relief after PVA and decrease incidence of recollapse in the coming years, SD in the fractured area should be involved in the surgical plan of PVA, and symmetrical cement distribution around the fractured area might be the optimal pattern we should pursue in the clinical practice.
Acknowledgment The authors gratefully acknowledge the support from Guangdong Province Administration of Traditional Chinese Medicine Program (NO. 20111185), Project of The Health Ministry of China (NO. W2012ZT07), and Guangdong Province Medical Science and Technology Research Program (NO. B2014175).
255
The authors also thank Prof Zhihui Pang for the advice of biomechanics. Authors’ contributions: D.L. and X.-B.J. contributed to conceiving and designing the study. L.-Q.Y., P.Y., and G.-Q.Z. performed the experiment. X.-B.J., Z.-S.Y., S.-C.Z., and Z.-D.Y. analyzed the data. X.-B.J., D.L., and L.-Q.Y. play the main role in writing the article.
Disclosure The authors reported no proprietary or commercial interest in any product mentioned or concept discussed in the article.
references
[1] Johnell O, Kanis JA. An estimate of the worldwide prevalence and disability associated with osteoporotic fractures. Osteoporos Int 2006;17:1726. [2] Gray LA, Jarvik JG, Heagerty PJ, et al. INvestigational Vertebroplasty Efficacy and Safety Trial (INVEST): a randomized controlled trial of percutaneous vertebroplasty. BMC Musculoskelet Disord 2007;8:126. [3] Steinmann J, Tingey CT, Cruz G, et al. Biomechanical comparison of unipedicular versus bipedicular kyphoplasty. Spine (Phila Pa 1976) 2005;30:201. [4] Belkoff SM, Mathis JM, Jasper LE, et al. The biomechanics of vertebroplasty. The effect of cement volume on mechanical behavior. Spine (Phila Pa 1976) 2001;26:1537. [5] Chen B, Li Y, Xie D, et al. Comparison of unipedicular and bipedicular kyphoplasty on the stiffness and biomechanical balance of compression fractured vertebrae. Eur Spine J 2011; 20:1272. [6] Hiwatashi A, Moritani T, Numaguchi Y, et al. Increase in vertebral body height after vertebroplasty. AJNR Am J Neuroradiol 2003;24:185. [7] McKiernan F, Faciszewski T, Jensen R. Reporting height restoration in vertebral compression fractures. Spine (Phila Pa 1976) 2003;28:2517. [8] Teng MM, Wei CJ, Wei LC, et al. Kyphosis correction and height restoration effects of percutaneous vertebroplasty. AJNR Am J Neuroradiol 2003;24:1893. [9] Garfin SR, Yuan HA, Reiley MA. New technologies in spine: kyphoplasty and vertebroplasty for the treatment of painful osteoporotic compression fractures. Spine (Phila Pa 1976) 2001;26:1511. [10] He SC, Teng GJ, Deng G, et al. Repeat vertebroplasty for unrelieved pain at previously treated vertebral levels with osteoporotic vertebral compression fractures. Spine (Phila Pa 1976) 2008;33:640. [11] Kim YY, Rhyu KW. Recompression of vertebral body after balloon kyphoplasty for osteoporotic vertebral compression fracture. Eur Spine J 2010;19:1907. [12] Lin WC, Lee YC, Lee CH, et al. Refractures in cemented vertebrae after percutaneous vertebroplasty: a retrospective analysis. Eur Spine J 2008;17:592. [13] Polikeit A, Ferguson SJ, Nolte LP, et al. Factors influencing stresses in the lumbar spine after the insertion of intervertebral cages: finite element analysis. Eur Spine J 2003; 12:413. [14] Polikeit A, Nolte LP, Ferguson SJ. The effect of cement augmentation on the load transfer in an osteoporotic functional spinal unit: finite-element analysis. Spine (Phila Pa 1976) 2003;28:991.
256
j o u r n a l o f s u r g i c a l r e s e a r c h 1 9 5 ( 2 0 1 5 ) 2 4 6 e2 5 6
[15] Baroud G, Nemes J, Heini P, et al. Load shift of the intervertebral disc after a vertebroplasty: a finite-element study. Eur Spine J 2003;12:421. [16] Goel VK, Kong W, Han JS, et al. A combined finite element and optimization investigation of lumbar spine mechanics with and without muscles. Spine (Phila Pa 1976) 1993;18:1531. [17] Noailly J, Lacroix D, Planell JA. Finite element study of a novel intervertebral disc substitute. Spine (Phila Pa 1976) 2005; 30:2257. [18] Chiang CK, Wang YH, Yang CY, et al. Prophylactic vertebroplasty may reduce the risk of adjacent intact vertebra from fatigue injury: an ex vivo biomechanical study. Spine (Phila Pa 1976) 2009;34:356. [19] Rohlmann A, Zander T, Rao M, et al. Applying a follower load delivers realistic results for simulating standing. J Biomech 2009;42:1520. [20] Rohlmann A, Zander T, Rao M, et al. Realistic loading conditions for upper body bending. J Biomech 2009;42:884. [21] Denis F. The three column spine and its significance in the classification of acute thoracolumbar spinal injuries. Spine (Phila Pa 1976) 1983;8:817. [22] Li QL, Li XZ, Liu Y, et al. Treatment of thoracolumbar fracture with pedicle screws at injury level: a biomechanical study based on three-dimensional finite element analysis. Eur J Orthop Surg Traumatol 2013;23:775. [23] Panjabi MM, Oxland TR, Lin RM, et al. Thoracolumbar burst fracture. A biomechanical investigation of its multidirectional flexibility. Spine (Phila Pa 1976) 1994;19:578. [24] Qiu TX, Teo EC, Zhang QH. Effect of bilateral facetectomy of thoracolumbar spine T11-L1 on spinal stability. Med Biol Eng Comput 2006;44:363. [25] Xu G, Fu X, Du C, et al. Biomechanical effects of vertebroplasty on thoracolumbar burst fracture with
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
transpedicular fixation: a finite element model analysis. Orthop Traumatol Surg Res 2014;100:379. McGraw JK, Cardella J, Barr JD, et al. Society of Interventional Radiology quality improvement guidelines for percutaneous vertebroplasty. J Vasc Interv Radiol 2003;14:S311. Gangi A, Sabharwal T, Irani FG, et al. Quality assurance guidelines for percutaneous vertebroplasty. Cardiovasc Intervent Radiol 2006;29:173. McConnell CT Jr, Wippold FJ 2nd, Ray CE Jr, et al. ACR appropriateness criteria management of vertebral compression fractures. J Am Coll Radiol 2014;11:757. Buchbinder R, Osborne RH, Ebeling PR, et al. A randomized trial of vertebroplasty for painful osteoporotic vertebral fractures. N Engl J Med 2009;361:557. Kallmes DF, Comstock BA, Heagerty PJ, et al. A randomized trial of vertebroplasty for osteoporotic spinal fractures. N Engl J Med 2009;361:569. McGuire R. AAOS clinical practice guideline: the treatment of symptomatic osteoporotic spinal compression fractures. J Am Acad Orthop Surg 2011;19:183. Wijayathunga VN, Oakland RJ, Jones AC, et al. Vertebroplasty: patient and treatment variations studied through parametric computational models. Clin Biomech (bristol, Avon) 2013;28:860. Magerl F, Aebi M, Gertzbein SD, et al. A comprehensive classification of thoracic and lumbar injuries. Eur Spine J 1994;3:184. Liebschner MA, Rosenberg WS, Keaveny TM. Effects of bone cement volume and distribution on vertebral stiffness after vertebroplasty. Spine (Phila Pa 1976) 2001;26:1547. Jiang XB, Mo L, Liang D, et al. The distribution of bone cement in the fractured lines on the outcome of percutaneous vertebroplasty. Chin J Spine Spinal Cord 2014;24:144.
