The Spine Journal 7 (2007) 308–317
`
`Efficacy of silicated calcium phosphate graft
`in posterolateral lumbar fusion in sheep
`Donna L. Wheeler, PhDa,b,c,d,*, Louis G. Jenis, MDa, Matthew E. Kovach, MSb,
`Jason Marini, MSb, A. Simon Turner, BVSc, MS, Dipl ACVSc
`aDepartment of Mechanical Engineering, College of Engineering, Orthopaedic Bioengineering Research Laboratory, Colorado State University,
`A101 Engineering Building, Fort Collins, CO 80523-1374, USA
`bBoston Spine Group, New England Baptist Hospital, Boston, MA
`cCharles River Laboratory, Interventional Surgical Services, Worchester, MA
`dDepartment of Clinical Sciences, Colorado State University, Fort Collins, CO
`Received 25 August 2005; accepted 12 January 2006
`
`Abstract
`
`BACKGROUND CONTEXT: Conditions requiring posterior lumbar spinal fusion remain a clin-
`ical challenge. Achieving arthrodesis using autogenous bone graft is inconsistent when rigid inter-
`nal fixation such as transpedicular instrumentation is applied. Synthetic materials, particularly
`calcium phosphate–based ceramics, have shown promise for spine fusion applications, especially
`when combined with autograft. Silicate substitution has been shown to enhance the bioactivity
`of calcium phosphates and may obviate the need for autologous supplementation.
`PURPOSE: Determine efficacy of silicated calcium phosphate (Si-CaP) compared with autograft
`to generate solid lumbar fusion.
`STUDY DESIGN: Comparison of healing of instrumented posterolateral lumbar fusion in ewes at
`2 and 6 months using Si-CaP or iliac crest autograft.
`METHODS: Eighteen skeletally mature ewes underwent implantation of either autograft or Si-
`CaP in the space spanning the L4–L5 transverse process. In vivo quantitative computed tomography
`(CT) scans were made at 2-month intervals and after euthanasia. Harvested spine segments were
`radiographed and biomechanically tested in bending at 6 months. Histological assessments were
`made at 2 and 6 months.
`RESULTS: Animals receiving Si-CaP graft were biomechanically and radiographically equivalent
`to those receiving autograft. Fusion mass density and volume were higher for the Si-CaP group
`throughout the healing period. Si-CaP regenerated normal bone tissue morphology, cellularity,
`and maturation with no inflammatory responses despite the fact that no autograft, bone marrow as-
`pirate, or blood was mixed with the material. Histomorphometrically, fusion mass was higher for
`Si-CaP and bony bridging was equivalent when compared with autograft treatment.
`CONCLUSIONS: Si-CaP was biomechanically, radiographically, and histologically equivalent to
`autograft in generating a solid, bony, intertransverse process fusion in an ovine model. Both treat-
`ment groups achieved 100% bridging fusion after 6 months of healing. Ó 2007 Elsevier Inc. All
`rights reserved.
`
`Keywords:
`
`Ovine; Posterior-lateral intertransverse process fusion; Lumbar; Biomechanics; Imaging; Histology; X-ray; CT;
`Calcium phosphate; Silicate
`
`Introduction
`
`FDA device/drug status: not approved for this indication (Actifuse
`Synthetic Bone Graft).
`Support was received from ApaTech Ltd, London, UK. No other funds
`were received from a commercial entity related to this manuscript.
`* Corresponding author. BioSolutions Consulting, LLC, 345 East
`Street, East Walpole, MA 02032. Tel.: (508) 850-6992; fax: (508) 850-
`9596.
`E-mail address: BiosolutionsConsulting@comcast.net (D.L. Wheeler)
`1529-9430/07/$ – see front matter Ó 2007 Elsevier Inc. All rights reserved.
`doi:10.1016/j.spinee.2006.01.005
`(cid:20)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:19)
`
`Autograft has been considered the ‘‘gold standard’’ for
`posterolateral intertransverse process fusion yet arthrodesis
`success is inconsistent. Effective alternatives to autograft
`would eliminate the need for a second surgical site and
`the associated donor site pain and morbidity [1,2]. Syn-
`thetic materials, particularly calcium phosphate–based
`ceramics, have been investigated as both grafting materials
`
`MILLENIUM EXHIBIT 2034
`Baxter Healthcare Corp. et al. v. Millenium Biologix, LLC
`IPR2013-00582, -00590
`
`

`

`D.L. Wheeler et al. / The Spine Journal 7 (2007) 308–317
`
`309
`
`and autograft expanders for spine fusion applications
`[3–6]. In general, grafting options that incorporated autolo-
`gous graft improved fusion over ceramic alone, because of
`improved cellularity and biological activity [3,5,7,8]. Im-
`proving healing responses to synthetic graft materials in
`the hope of obviating the need for autograft supplementa-
`tion has been a primary focus of biomaterials development.
`The influence of silicon in bone formation and repair has
`been established [9,10]. The role of silicon in promoting
`bone bonding and induction in silicon-containing bioactive
`glasses and glass ceramics has been studied both in vivo
`[11–13] and in vitro [14–18]. Low levels of ionic silicate,
`particularly in the presence of calcium ions, have been
`shown to up-regulate osteoblast proliferation and differenti-
`ation [19–21], promote osteoinductive gene expression
`[14,20], and increase Type I collagen synthesis [19].
`Recently, a phase-pure silicated calcium-phosphate ma-
`terial (Si-CaP), Actifuse Synthetic Bone Graft (ApaTech
`Limited, London), has been developed to exploit the os-
`teo-stimulatory effects of silicon. This material is similar
`to the structure of bone mineral but with site-specific sub-
`4-) ionsstitution of phosphate (PO43-) ions with silicate (SiO4
`
`
`resulting in a biomaterial containing 0.8 wt% silicon
`[22,23]. The interconnected porous structure provides
`a scaffold for osteoconduction, and the release of silicate
`ions from the matrix promotes rapid bone formation [22].
`Other in vivo studies have shown this material to be effec-
`tive in healing lapine and large ovine osseous defects [22–
`24]. The aforementioned in vitro and in vivo studies have
`established that Si-CaP enhances the bioactivity when com-
`pared with phase-pure calcium phosphates of equivalent
`structure. No study has ascertained whether this silicate en-
`hancement of calcium phosphates can achieve the healing
`standards of autograft, particularly in the challenging
`healing environment of posterior
`spinal
`fusion. We
`hypothesized that Si-CaP would generate solid lumbar
`intertransverse process fusion, similar to that seen with au-
`tograft, in an ovine model. Fusion was quantitatively and
`qualitatively assessed using plain radiographs, computed
`tomography (CT), biomechanics, and histological evalua-
`tions to prove this hypothesis.
`
`Materials and methods
`
`All animal procedures were approved by the Animal
`Care and Use Committee at Colorado State University.
`Eighteen skeletally mature ewes were used for the study.
`Each surgery was performed under general anesthesia and
`strict sterile conditions. With the animal in the sternal re-
`cumbency, a 6–8 cm longitudinal incision was made over-
`lying the palpable spinous processes from L2 to L5. The
`fascia was split bilaterally, approximately 2–3 cm from
`the midline, and an intermuscular plane developed to the
`transverse processes. The transverse processes of L4 and
`L5 and the pars interarticularis were cleared of soft tissues
`with electrocautery, and the L4–L5 facet joint was excised.
`
`Decortication of the dorsal aspect of each transverse pro-
`cess and lateral pars was performed with a high-speed burr.
`Nine animals were randomized to receive 20 cc Si-CaP, and
`the other nine received 20 cc cortico-cancellous autograft
`(AG) harvested from the iliac crest. Using a separate fascial
`incision, the graft was placed bilaterally in the region span-
`ning the L4–L5 transverse processes and along the lateral
`pars area. The fusion site was stabilized with bilateral trans-
`pedicular
`instrumentation (Custom-fabricated, ApaTech
`Limited, London). Wounds were closed in layers. Postoper-
`ative pain management and wound care were provided until
`the animals returned to activities ad lib. CT scans were ac-
`quired during healing for three animals per treatment group
`(randomly chosen). Scans were acquired at 60, 120, and
`180 days for the Si-CaP group and at 60 and 180 days
`for the AG group. At necropsy, the spine was harvested
`en bloc from L1 to L7, X-rays were taken, spines were
`cleaned of soft tissue, nondestructive biomechanical testing
`was completed, and spines were processed for histological
`analyses to assess fusion efficacy.
`
`Computed tomography
`
`Standard transverse CT scans were acquired of the fused
`lumbar spine region and a standard density phantom at 130
`kV and 175 mA with a slice thickness of 1.5 mm and a res-
`olution of 0.35 mm/pixel (Picker International, Cleveland,
`OH). The individual slice images were stacked to form
`a three-dimensional (3D) volume using public domain soft-
`ware. Gray scale thresholding and edge detection algo-
`rithms were used to differentiate between bone, muscle,
`and the metal implants. Edge detection algorithms were
`used to differentiate among bone, muscle, and metal for
`each 1.5-mm slice [25]. Boundaries were interpolated be-
`tween adjacent scan layers using a cubic spline to create
`a 3D rendering of the spine and fusion mass. The fusion
`mass volume was calculated from the 3D reconstructions
`for the right and left sides from the edge cranial to caudal
`transverse process spanning the fusion site. Bilateral quali-
`tative fusion scores (05no fusion, 15moderate fusion and
`thin connectivity, 25extensive fusion and connectivity)
`were made based on rendered 3D image by a single quali-
`fied investigator blinded to the treatments. The density of
`the fusion mass was categorized based on the density phan-
`tom into four categories: 400–600, 600–800, 800–1000,
`1000–1250 mg/cc. The percentage of the total fusion vol-
`ume at each density range was calculated to quantify den-
`sification of the tissue. One animal received a CT scan of
`the Si-CaP graft material immediately after implantation
`to characterize initial graft volume and density. To measure
`graft resorption and new bone formation, specific image sli-
`ces within the 3D image stack were evaluated at each time
`point. Changes in fusion mass density based on shifts in the
`four density categories and fusion volume were calculated
`and provided information to estimate Si-CaP graft resorp-
`tion and bone formation over time.
`
`2 of 10
`
`

`

`310
`
`Radiography
`
`D.L. Wheeler et al. / The Spine Journal 7 (2007) 308–317
`
`Routine ventrodorsal and lateral radiographs of the ex-
`planted spinal segments were performed. Radiographs were
`qualitatively scored, blinded to treatment group, based on
`the extent of right and left fusion mass and connectivity
`using the same scale applied for the CT scans (0,1,2).
`
`Biomechanics
`
`Soft tissues were carefully dissected from the resected
`spine spanning from L3 to L6. Spinal
`instrumentation
`was removed and a qualitative fusion score was assigned
`(0,1,2) based on gentle manual manipulation at the fusion
`site (blinded to treatment). A score of ‘‘2’’ indicated no mo-
`tion and obvious fusion while a ‘‘0’’ was associated with
`visible motion between the spinal segments. The cranial-
`and caudal-most vertebral bodies were potted in high-
`strength dental plaster (Snapstone, WhipMix) in a base
`mold using a custom alignment jig. A triad of retroreflec-
`tive markers was attached to each vertebral body, and the
`markers were tracked during loading by an optical kine-
`matic system (Peak Performance Technologies, Denver,
`CO). All spines underwent five 60-Hz cycles of nondestruc-
`tive loading in flexion/extension and lateral bending with
`loads ranging from 5.0 N-m to 5.0 N-m using a biaxial
`servohydraulic materials testing machine (MTS, Eden Prai-
`rie, MN). Pure bending was applied using a custom 4-point
`bending fixture to ensure a constant bending moment across
`the fusion site. The last data cycle was processed to deter-
`mine stiffness (N-m/deg) and range of motion (deg) at the
`fused motion segment [26]. Stiffness was calculated as
`the linear trend-line of the load-angular deformation curve.
`Range of motion was obtained from the span between max-
`imum and minimum angles during the loading cycle. Fu-
`sion of experimental spines (Si-CaP, AG) was confirmed
`using biomechanical data obtained from historical control
`spines that received L4–L5 facetectomies (n58).
`
`Histology
`
`Immediately after nondestructive mechanical testing, the
`fused spinal segments were cut in the sagittal plane and
`fixed in 70% ethyl alcohol fixative for histology. Tissues
`were dehydrated, infiltrated, and embedded using standard
`methods for undecalcified methyl methacrylate processing
`[27]. One 20–30 mm section was made using an Exakt
`saw and grinder (Exakt Technologies, Oklahoma, OK) from
`the right and left side. Sections were made at the center of
`the fusion mass perpendicular to the transverse processes.
`Each slide was stained with a Van Gieson bone stain for
`quantitative and histopathological assessments. Healing,
`bone quality, and graft incorporation were scored using
`a semiquantitative scale. The scale graded graft–tissue in-
`terface (05gap, 15fibrous, 25fibrous and bone, 35bone),
`remodeling (05woven, 15wovenOlamellar, 25woven
`!lamellar, 35lamellar), osteoblasts (05minimal, 15some,
`
`25many), osteoclasts (05minimal, 15some, 25many),
`and inflammation (05none, 15some, 25many). Scores
`were assigned by a single pathologist in 0.5 increments.
`
`Histomorphometry
`
`High-resolution digital images were acquired by field for
`each histological slide using an Image Pro Imaging system
`(Media Cybernetics, Silver Spring, MD) and a Nikon E800
`microscope (AG Heinze, Lake Forest, CA), Spot digital
`camera (Diagnostic Instruments, Sterling, Heights, MI),
`and a personal computer. The histomorphometric parameters
`measured or calculated included: total reactive (area of the
`fusion) area (mm2), bone within reactive area (mm2), percent
`bone within reactive area (%), graft within reactive area
`(mm2), percent graft within reactive area (%), percent
`graftþbone within the reactive area (%), distance between
`transverse processes (mm), connecting bone within the trans-
`verse process span (mm), and percent distance between trans-
`verse processes consisting of bone (bone union) (%).
`
`Statistical analysis
`
`Continuous responses were evaluated statistically using
`a one-way analysis of variance to determine differences be-
`tween treatments (Si-CaP, AG). Categorical data were ana-
`lyzed using a Wilcoxon signed-rank test. All statistical tests
`were run using SAS statistical software (Cary, NC) at a sig-
`nificance level of a50.05. No statistics were run for the CT
`data, which were completed for screening purposes with
`insufficient sample size for data analysis.
`
`Results
`
`All animals tolerated the procedure and were ambulatory
`within 12 hours. One animal in the Si-CaP group experi-
`enced hardware failure during the healing period and was
`eliminated from the data analysis. No wound complications
`or infections were noted during the study period.
`All animals (Si-CaP and AG) received a qualitative ra-
`diographic fusion score of 0 at 60 days (0% fusion rate).
`By 120 days the Si-CaP group fusion score improved to
`an average score of 1.2560.5 (mean6SD). The average
`radiographic fusion scores for the Si-CaP and AG groups
`after 180 days of healing were equivalent, 1.7560.43 and
`respectively (mean6SD, pO.05). Fusion
`1.5860.51,
`connectivity scores using CT images were similar to plain
`radiographic scores showing no differences between treat-
`ment groups (pO.05). CT fusion scores for the AG treat-
`ment were 0 after 60 days (0% fusion) and 2 after 180
`days of healing (100% fusion). CT fusion scores for the
`Si-CaP treated animals were 0, 1.560.55, and 2.0 after
`60, 120, and 180 days, respectively.
`Initial AG and Si-CaP graft placement is illustrated in
`Figure 1. The transverse CT slices shown in Figure 2 illus-
`trate the development of bony fusion from 60 to 180 days in
`
`3 of 10
`
`

`

`D.L. Wheeler et al. / The Spine Journal 7 (2007) 308–317
`
`311
`
`All spines received a manual manipulation score of 2,
`with the exception of one specimen in the AG treatment
`group that received a score of 1. Biomechanical testing re-
`vealed equivalent stiffness for both flexion/extension and
`lateral bending in Si-CaP and AG treated spines (Fig. 6,
`pO.05). Range of motion was also equivalent between
`treatments (pO.05). The comparison to historical controls
`of unfused, facetectomized spinal segments confirmed that
`all spines in this study were solidly fused.
`Histological assessment revealed a lack of continuous
`bony fusion after 60 days with development of bony conti-
`nuity by 180 days for both treatment groups. The Si-CaP
`fused animals showed good bone apposition and incorpora-
`tion within the newly formed fusion mass (Fig. 7). No dif-
`ferences were detected between histopathology scores for
`bone remodeling, osteoblast activity, and osteoclast activ-
`ity, yet a trend towards increased cellular activity was noted
`for Si-CaP grafted tissues after 180 days compared with 60
`days, despite the fact that no autograft, bone marrow aspi-
`rate, or blood was mixed with the material at the time of
`surgery (pO.05). There was no evidence of a significant in-
`flammatory response on histological analysis. Morphomet-
`rically, the total fusion mass (reactive area) was greater for
`the Si-CaP treatment (Fig. 8) because of the added volume
`of the unresorbed graft material (p!.05). Similar contigu-
`ous bone was measured between the transverse processes
`for both Si-CaP and AG treatments (Fig. 9, pO.05).
`
`Discussion
`
`This study provided a comprehensive analysis by assess-
`ing fusion from radiographic, functional, and biological ap-
`proaches. Grafting with Si-CaP was equivalent to using
`autograft in an instrumented sheep lumbar intertransverse
`process fusion model. Although minor differences were de-
`tected in morphometric indices, the overall biomechanical
`function and radiographic fusion scores between treatment
`groups were equivalent.
`Plain radiographs and CT are clinical tools often used to
`appraise fusion in animal models, often through a grading
`system [28,29], but rarely are clinical CT scans quantified.
`Micro-computed tomography is used commonly to assess
`regenerated bone structure with the analysis limited to
`small samples of tissues, which is impractical when evalu-
`ating an entire fusion mass [30–32]. The quantitative CT
`data obtained in this study involved a simple process of
`calculation of fusion volumes and density categorization.
`Although the original objective of the scans was to obtain
`screening data from a few animals in each treatment group
`for fusion progression over the 6 months of healing, the
`quantitative data obtained were more informative than antic-
`ipated. The CT volume and density quantification showed
`two distinct phases in the healing process, a graft resorptive
`phase followed by callus growth. The Si-CaP grafted fusions
`had more fusion volume than the autograft fusion after 6
`
`Fig. 1. Computed tomography slices illustrating autograft (AG) and sili-
`cated calcium phosphate (Si-CaP) graft placement after initial implanta-
`tion. (A) Si-CaP coronal; (B) Si-CaP sagittal; (C) AG coronal; and (D)
`AG sagittal.
`
`a Si-CaP treated animal. Both Si-CaP and AG treated ani-
`mals experienced a loss of approximately 50% in fusion
`volume after 60 days compared with the initial graft vol-
`ume. By 180 days the AG treated animals regenerated bony
`fusion approximately equivalent to the volume of the ini-
`tially implanted graft whereas Si-CaP treated animals dou-
`bled fusion volume (Fig. 3). Differences in fusion volume
`between treatments quantified by 3D CT reconstructions
`were not significant at any time point (pO.05). Although
`statistical analyses were not run on density data, high-
`density tissue within the fusion mass increased during
`maturation of the fusion for both treatments. However,
`the Si-CaP treatment had a greater percentage of high-
`density tissue and less low-density tissue when compared
`with the AG treatment (Fig. 4). The initial Si-CaP graft
`density was quantified to be between 550 and 710 mg/cc.
`Fusion volume with the density of Si-CaP graft decreased
`whereas densities associated with bone increased with time
`in vivo (Fig. 5).
`
`4 of 10
`
`

`

`312
`
`D.L. Wheeler et al. / The Spine Journal 7 (2007) 308–317
`
`Fig. 2. Computed tomography transverse slices within the fusion region for a representative silicated calcium phosphate animal taken at: (A) 60 days and (B)
`180 days.
`
`months. This increased fusion volume is consistent with the
`increased reactive area quantified in these same specimens
`histomorphometrically. Fusion callus size is indicative of
`graft resorption and active bone incorporation. Graft area
`and graft resorption were quantified using histomorphometry
`for the Si-Cap material and not autograft since differentiation
`between newly regenerated bone and autograft is subjective
`and inaccurate. Osteoconductive matrices such as autograft,
`calcium sulfate, and some b-tricalcium phosphate materials
`resorb quickly with loss of a three-dimensional scaffold be-
`fore fusion is achieved [22,33–36]. The relative stability of
`
`the Si-CaP preserves the osteoconductive matrix and may in-
`crease callus volume when combined with the volume of
`newly regenerated bone. Others have shown similar increases
`in callus formation compared with autograft when using
`slowly resorbing synthetic graft materials in combination
`with bone marrow [37].
`By categorizing fusion volume density using CT phan-
`tom calibration, the fusion mass for both autograft and
`Si-CaP treated animals increased in density during the heal-
`ing period. This increase in mineralized tissue reflects nor-
`mal healing and maturation of
`the fusion mass. The
`
`Fig. 3. Total fusion volume over time for the silicated calcium phosphate (Si-CaP) and autograft (AG) treatment groups. No statistical differences were
`detected between treatment groups (p!.05). Mean6SD.
`
`5 of 10
`
`

`

`D.L. Wheeler et al. / The Spine Journal 7 (2007) 308–317
`
`313
`
`Fig. 4. Percent of total fusion volume at each density category for the autograft (AG) and silicated calcium phosphate (Si-CaP).
`
`percentage of high-density material in the fusion mass was
`greater for the Si-CaP than the autograft animals because of
`the presence of
`the graft material.
`Initial density of
`Si-CaP graft material obtained by CT provided a means
`to estimate bone formation and graft resorption. Errors are
`certainly inherent in this estimate owing to the uncertainty
`of the contribution of bone versus Si-CaP in the density range
`associated with the graft material. However, a clear rela-
`tionship was noted between Si-CaP resorption and bone
`formation, confirming the potential suggested by previous
`investigators [22–24] that this material is an effective bone
`regeneration matrix in clinically relevant models.
`Mechanical integrity of the regenerated fusion mass of-
`ten is assessed qualitatively with gentle manual manipula-
`tion before histological processing [38]. Yet
`subtle
`differences in fusion mass rigidity are not discernable with
`
`Fig. 5. Estimate of silicated calcium phosphate (Si-CaP) resorption and
`bone formation for a representative animals.
`
`this qualitative method. No standards have been established
`for biomechanical testing for spine fusion [39]. However,
`most studies evaluate lateral bending, flexion/extension,
`or torsion biomechanics using various loading fixtures,
`loading rates, modes of loading, and displacement measure-
`ment methods [38,40–43] and results are difficult to com-
`pare [44]. Adopting a dynamic loading method, we found
`no differences in stiffness and range of motion in bending
`between AG and Si-CaP treated spines after 6 months of
`healing, indicating functional fusion. Biomechanical as-
`sessments at earlier healing points may have provided addi-
`tional
`information on the progression of healing and
`development of functional fusion. Nondestructive loads
`used for biomechanical testing in this study are consistent
`with loads used by other researchers working with large an-
`imal models [40,43]. However, higher loads may better as-
`sess mature fusion. Comparisons of stiffness and range of
`motion to unfused control spines clearly show that func-
`tional fusion was achieved for both treatment groups by 6
`months.
`Both grafting options produced an early, normal, woven-
`bone response followed by remodeling to mature woven
`bone by 6 months. As expected, the incorporation of the
`morselized autograft was excellent, with early bony inte-
`gration. The osteoconductive ceramic synthetic bone graft
`material, Si-CaP, also produced excellent integration of
`bone within the graft matrix without fibrous encapsulation.
`The integration response was equivalent to other ceramic
`graft materials [36,45–47]. This strong integration of the
`graft material within the fusion mass would account for
`the biomechanical properties documented in this study.
`Others have shown the relationship between graft integra-
`tion and functional biomechanics [46,48].
`The Si-CaP graft material resorbed via a combination of
`cell-mediated and dissolution mechanisms documented by
`
`6 of 10
`
`

`

`314
`
`D.L. Wheeler et al. / The Spine Journal 7 (2007) 308–317
`
`Fig. 6. The fusions achieved with autograft (AG) and silicated calcium phosphate (Si-CaP) had equivalent stiffness and were significantly stiffer than un-
`fused facetectomized control (CTL) spines (pO.05). Mean6SD.
`
`our histopathological evaluation. The slow yet progressive
`matrix resorption combined with the observed cellular infil-
`tration, graft integration, and active bone remodeling pro-
`vide evidence of the biocompatibility and potential for
`maintaining long-term mechanical competency in bone
`grafting applications. Others have assessed silica-contain-
`ing grafts materials and have noted similar cellular
`response and bone regenerative capabilities [13,49]. The
`Si-CaP graft provided an osteoconductive matrix for bone
`infiltration, whereas autograft provided both an osteocon-
`ductive scaffold as well as osteogenic stimuli. Yet the Si-
`CaP graft was able to overcome the biological disadvantage
`and produce a robust fusion callus with osteoblast and os-
`teoclast activity on par with autograft-supplemented fusion
`tissue.
`The amount of graft within the Si-CaP fusion mass de-
`creased from approximately 17% to 11% between 2 and
`6 months of healing (rate of 1.5%/mo). Quantitative CT
`
`was used for density analyses by which the amount of graft
`within the fusion mass was estimated to decrease from
`approximately 50% to 35% during that same 6 months
`(rate of 0.4%/mo). The CT graft density calculations may
`underestimate graft
`resorption, because the density of
`bone overlaps with the density of the Si-CaP material.
`Both assays indicate resorption of graft with concurrent
`bone regeneration.
`The sheep is a popular animal model for spine fusion.
`Although more realistic than the rabbit model, the fusion
`rate in this model is high, and definitive applicability to
`clinical efficacy is uncertain [50]. Studies adopting shorter
`healing durations have found differences between fusion
`treatments using this animal model [29]. The long healing
`duration in this study may have precluded detection of dif-
`ferences between Si-CaP and AG treatments. However,
`equivalency between treatment groups was also seen at ear-
`lier time point with our histological and CT evaluations.
`
`Fig. 7. Silicated calcium phosphate (gray) integration with the surrounding bone matrix (red). (A) Evidence of graft incorporation at 6 months. Magnifica-
`tion520. (B) Evidence of early bone formation at 2 months. Magnification5100.
`
`7 of 10
`
`

`

`D.L. Wheeler et al. / The Spine Journal 7 (2007) 308–317
`
`315
`
`Fig. 8. Total fusion mass (reactive area). Significantly more reactive area was measured for the silicated calcium phosphate (Si-CaP) treatment compared
`with autograft (AG) (p!.05). Mean6SD.
`
`Another limitation of the posterolateral ovine fusion model
`used in this study is the placement of the graft material at
`the medial aspect of the transverse processes (Fig. 1).
`The graft placement might likely stimulate both facet fu-
`sion and intertransverse process fusion. Based on our
`
`intertransverse process fusion was
`results,
`histological
`observed in all animals yet facet fusion was not evaluated.
`Although limitations exist in this ovine model, the study
`objective to compare the healing response of two graft ma-
`terials was successfully achieved.
`
`Fig. 9. Continuous bone spanning transverse processes. Equivalent bone was measured spanning the transverse processes for the (A) silicated calcium phos-
`phate (Si-CaP) and (B) autograft (AG) treatment groups (p!.05). Mean6SD.
`
`8 of 10
`
`

`

`316
`
`D.L. Wheeler et al. / The Spine Journal 7 (2007) 308–317
`
`Ceramics alone have been found to be inferior to autol-
`ogous grafts in spine fusion procedures [3,5,7] except when
`combined with osteoinductive factors [51–55]. This ovine,
`lumbar, posterolateral fusion study indicated that Si-CaP
`was able to overcome previously noted biological disadvan-
`tages of ceramic grafts, confirming the positive effects of
`trace levels of silicon on bone regeneration previously
`documented through in vitro studies [14,19–21]. The sili-
`cated ceramic graft was equivalent
`to autologous graft
`radiographically, biomechanically, and histologically. The
`results from this ovine model support the use of this unique
`ceramic bone graft substitute in spinal fusion.
`
`Acknowledgments
`
`We thank Howard Seim III, DVM and Lisa S. Klopp,
`DVM, MS, for their surgical expertise; Diane Beranek
`and Mike Karr for preparation of histological specimens;
`Amy Lyons and Tatiana Motta for their assistance in histo-
`logical evaluations; and Jon Kushner, staff and students of
`the Orthopaedic Bioengineering Research Lab for technical
`support.
`
`References
`
`[1] Goulet JA, Senunas LE, DeSilva GL, Greenfield ML. Autogenous
`iliac crest bone graft: complications and functional assessment.
`Clin Orthop Relat Res 1997;339:76–81.
`[2] Seiler JG 3rd, Johnson J. Iliac crest autogenous bone grafting: donor
`site complications. J South Orthop Assoc 2000;9:91–7.
`[3] Cavagna R, Daculsi G, Bouler J. Macroporous calcium phosphate
`ceramic: a prospective study of 106 cases in lumbar spine fusion.
`J Long-term Effects of Medical Implants 1999;9:403–12.
`[4] Delecrin J, Aguado E, Nguyen JM, Pyre D, Royer J, Passuti N. Influ-
`ence of local environment on incorporation of ceramic for lumbar fu-
`sion: comparison of laminar and intertransverse sites in a canine
`model. Spine 1997;22:1683–9.
`[5] Emery SE, Fuller DA, Stevenson S. Ceramic anterior spinal fusion.
`Biologic and biomechanical comparison in a canine model. Spine
`1996;21:2713–9.
`[6] Passuti N, Daculsi G, Rogez JM, Martin S, Bainvel JV. Macroporous
`calcium phosphate ceramic performance in human spine fusion. Clin
`Orthop Relat Res 1989;248:169–76.
`[7] Khan SN, Fraser JF, Sandhu HS, Cammisa FP Jr, Girardi FP,
`Lane JM. Use of osteopromotive growth factors, demineralized bone
`matrix, and ceramics to enhance spinal fusion. J Am Acad Orthop
`Surg 2005;13:129–37.
`[8] Kai T, Shao-qing G, Geng-ting D. In vivo evaluation of bone marrow
`stromal-derived osteoblasts-porous calcium phosphate ceramic com-
`posites as bone graft substitute for lumbar intervertebral spinal
`fusion. Spine 2003;28:1653–8.
`[9] Carlisle EM. Silicon: a possible factor in bone calcification. Science
`1970;167(916):279–80.
`[10] Hench L, Paschall H. Direct chemical bond of bioactive glass–
`ceramic materials to bone and muscle. J Biomed Mater Res
`1973;7:25–42.
`[11] De Aza PN, Luklinska ZB, Santos C, Guitian F, De Aza S. Mecha-
`nism of bone-like formation on a bioactive implant in vivo. Biomate-
`rials 2003;24:1437–45.
`[12] Oonishi H, Hench LL, Wilson J, et al. Quantitative comparison of
`bone growth behavior in granules of Bioglass, A-W glass-ceramic,
`and hydroxyapatite. J Biomed Mater Res 2000;51:37–46.
`
`[13] Wheeler D, Eschbach E, Hoellrich R, Montfort M, Chamberland D.
`Assessment of resorbable bioactive materials for grafting of critical-
`size cancellous defects. J Orthop Res 2000;18:140–8.
`[14] Gao T, Aro HT, Ylanen H, Vuorio E. Silica-based bioactive glasses
`modulate expression of bone morphogenetic protein-2 mRNA in
`Saos-2 osteoblasts in vitro. Biomaterials 2001;22:1475–83.
`[15] Gough JE, Clupper DC, Hench LL. Osteoblast responses to tape-cast
`and sintered bioactive glass ceramics.
`J Biomed Mater Res
`2004;69A:621–8.
`[16] Gough JE, Jones JR, Hench LL. Nodule formation and mineralisation
`of human primary osteoblasts cultured on a porous bioactive glass
`scaffold. Biomaterials 2004;25:2039–46.
`[17] Xynos ID, Hukkanen MV, Batten JJ, Buttery LD, Hench LL,
`Polak JM. Bioglass 45S5 stimulates osteoblast turnover and enhances
`bone formation in vitro: implications and applications for bone tissue
`engineering. Calcif Tissue Int 2000;67:321–9.
`[18] Gibson IR, Best SM, Bonfield W. Chemical c