Vertebroplasty Using Injectable Calcium Phosphate Cement Compared Topolymethylmethacrylate in a Unique Canine Vertebral Body Large Defect Model

Introduction: Vertebroplasty was developed to mechanically reinforce and restore height of weaken or fractured vertebral bodies. Clinically, polymethylmethacrylate bone cement (PMMA) has been used but carries risks of thermal injury to the spinal cord and may allow continued collapse. Use of calcium phosphate (CaP) cement offers the potential for resorption of the cement over time and replacement with new bone as a biological method to restore vertebral body mass and avoid potential thermal cord injury. A previous study from our laboratory in human cadaveric spines showed that CaP cement was effective at increasing the compressive strength of vertebral bodies [1]. In the present study, the mechanical vertebral strength and histological assessment of the bone response to these two materials are compared using a canine vertebral body large defect model. Methods and Materials: Two level vertebroplasties were performed at L1 and L3 in 20 skeletally mature large hounds weighing 30 to 41 kg and studied for 1 mo. (N=10) and 6 mos. (N=10) under an IACUC protocol. In each dog, one vertebral defect was injected with the test material, CaP cement (BoneSource), and the other defect with PMMA containing 30% BaSO4 as a control with material location randomized. Both materials were supplied sterile by the manufacturer, Stryker Howmedica Osteonics, Rutherford, NJ. The canine model vertebroplasty was performed through a dorsal spinal approach with the defect created through the lateral vertebral body wall extending into and across the width of the central body. Defects of reproducible size were created using custom instrumentation. Accessible cancellous bone was excavated without perforation of the end plates or spinal canal resulting in a central rectangular vertebral defect nominally 18x5x22 mm, representing loss of approximately 3/5 the length, 1/2 the height, and 3⁄4 of the width of the body. The materials were injected using a syringe until the defects were filled and the volumes were recorded (mean, 1.5cc of CaP cement and 1.6cc of PMMA). The animals were followed clinically for neurological changes and radiographs were obtained preoperative, postoperative and at 1, 3, 4 and 6 mos. After euthanasia with a super saturated solution of pentobarbital, 5 dogs from each group were used for mechanical tests and the remaining 5 per group processed for histological evaluation. Prior to sectioning, orthogonal radiographs and CT scans were obtained of the explanted spines. Transaxial CT sections were used to calculate the percent vertebral body fill. The dorsal lamina and the spinal cord were removed and canals graded for extrusion of cement. Mechanical tests performed on an Instron materials test machine consisted of placing the isolated specimen axially in a custom jig and compressing at 15mm/min to 50% of original height. The two treated vertebrae and adjacent non-implanted vertebrae in each dog were tested for compressive fracture strength. The pretest height was compared to post compression height with all results analyzed using an ANOVA. Histological specimens were serially sectioned axially and processed for undecalcified, plastic-embedded histology. Basic fuchsin and toluidine blue-stained sections were studied by light microscopy to evaluate the remaining cement, the cement-bone interface, new bone formation, foreign body reaction, and the surrounding vertebral bone. Results: All dogs ambulated within 5 days without neurological deficits, or fractures, and no infections occurred throughout the study. Radiographically, postoperative views revealed that the implant materials were consistently placed well within the vertebral bodies. There was no extrusion through any endplate. A density difference was apparent between PMMA and CaP in postoperative radiographs and persisted throughout the study with PMMA appearing denser than CaP. Compared to the immediate postoperative films, there was a decrease in density of the CaP cement sites of 50% in 1 of 10 dogs at 1 mo. and of 25% in 4 of 10 at 6 mos., but no change over time in the PMMA sites. Material extrusion was evident postoperatively into the spinal canal area however this was actually intrusion into the ventral vertebral sinus vessels with a greater frequency in PMMA 7/20 than CaP 1/20. Cement intrusion of the central vertebral body vessels was seen in 5/20 CaP and 8/20 PMMA injected vertebrae. Grossly, at examination of the explanted vertebrae from the 10 dogs in the histological group, extrusion of cement into the vertebral sinus on the spinal canal floor was confirmed in 3/10 CaP vertebrae and 6/10 PMMA vertebrae. Examination of the cut surface of the discs of the motion segment revealed a low incidence of degenerated discs. CT scan analysis indicated that the average % volume fill with cement was not statistically different between CaP at 1 mo. (31.27%) versus 6 mos. (27.13%) and between PMMA at 1 mo. (42.74%) versus 6 months (39.43%). However, at both 1 and 6 mos. PMMA had a significantly larger percentage fill than CaP (p =0.005 and p =0.002) Mechanical tests of compression strength revealed no significant difference in vertebral body height and compressive strength among the groups. There was a trend for CaP treated vertebrae to increase in compressive strength from 1mo. to 6 mos. (6490+ 372 N v. 7211+ 963 N) and for PMMA to decrease (8305+1674 N v. 6502+1916 N) compared to adjacent non-treated vertebrae (5771+927) v. (5866+971). Motion segment testing found no statistical differences among groups for elastic modulus or hysteresis, dynamic modulii from cyclic loading, and for decrease in disc stresses (stress relaxation). These suggest that injection of CaP or PMMA did not cause significant changes in the biomechanical behavior of the adjacent discs up to 6 months. Histology showed both materials were well integrated into the vertebral body bone. Some differences were evident in bone response. Vertebral sites containing PMMA were characterized by new bone present on cement surfaces and new bone forming on the surrounding trabecular surfaces. Often a thin fibrous membrane incompletely surrounded the residual PMMA bolus centrally and cement intruded into trabecular spaces, vascular channels and the canal. Comparison of 1month and 6 months sections was characterized by an increase in the amount and extent of fibrous tissue surrounding the PMMA cement. PMMA was frequently intruded into the vessels on the canal floor while this occurred infrequently with CaP. CaP sites, in contrast, had a majority of the material residual in the central vertebral body and surrounding trabecular spaces but infrequently in the canal or vessels. Unlike PMMA, the CaP was undergoing resorption and remodeling, which was apparent as fragmentation with vascular invasion and bone ingrowth into the material bolus and encompassing smaller fragments dispersed into the trabecular spaces. Abundant woven and lamellar bone was found on the CaP cement surfaces, wi thin the remodeled material, and on the surrounding trabeculae surfaces. This gave an overall appearance of condensation of bone around the implant site. From 1 mo. to 6 mos., CaP containing sites showed prominent resorption of the material, incorporation of residual CaP into newly formed trabeculae, and a greater amount of lamellar bone on the implant surface. No histological changes in spinal cord segments overlying treated vertebrae were noted. Discussion: CaP (Bone Source) was as effective as PMMA in restoring and maintaining vertebral body strength in this canine vertebral body large defect model for up to 6 months. Neither material caused significant visible degeneration in the adjacent discs or differences in biomechanical behavior of the motion segments. Histologically, CaP demonstrated excellent biocompatibility without adverse inflammatory reaction or foreign body response. The injectable CaP set in situ and provided a highly osteoconductive surface that allowed remodeling, resorption and replacement by bone. Although some injected cement, which intruded into the venous sinus, could have been diluted away by blood flow, CaP appears to pose less of a risk for canal extrusion. The relationship of a compressive strength equal to or greater than adjacent non-treated vertebrae mechanically and considerable bone formation histologically demonstrates successful restoration of bone mass in a spinal vertebrae with a large central defect. Overall, in short and intermediate time periods, CaP cement can be an effective method to treat large vertebral defects. There is the implication that continued remodeling might result in resorption of a majority of the cement with replacement by bone. This will require additional study in this model over a longer time period. Reference: Lim TH, et. al. Spine 12:1297-1302, 2002 Acknowledgement: Funded by Stryker Howmedica Osteonics E-mail: [email protected]