RUNNING HEAD: Effect of a Cervical Intervertebral Spacer on Construct Stiffness Effect of an Intervertebral Disk Spacer on Stiffness after Monocortical Screw/Polymethylmethacrylate Fixation in Simulated and Cadaveric Canine Cervical Vertebral Columns

Objective: To determine the biomechanical effect of an intervertebral spacer on construct stiffness in a PVC model and cadaveric canine cervical vertebral columns stabilized with monocortical screws/polymethylmethacrylate (PMMA). Study Design: Biomechanical study. Sample population: PVC pipe; cadaveric canine vertebral columns Methods: PVC model – PVC pipe was used to create a gap model mimicking vertebral endplate orientation and disk space width of large-breed canine cervical vertebrae; 6 models had a 4-mm gap with no spacer (PVC group 1); 6 had a PVC pipe ring spacer filling the gap (PCV group 2). Animals – large breed cadaveric canine cervical vertebral columns (C2-C7) from skeletally mature dogs without (cadaveric group 1, n=6, historical data) and with an intervertebral disk spacer (cadaveric group 2, n=6) were used. All PVC models and cadaver specimens were instrumented with monocortical titanium screws/PMMA. Stiffness of the 2 PVC groups was compared in extension, flexion, and lateral bending using non-destructive 4point bend testing. Stiffness testing in all 3 directions was performed of the unaltered C4-C5 vertebral motion unit in cadaveric spines and repeated after placement of an intervertebral cortical allograft ring and instrumentation. Data were compared using a linear mixed model approach that also incorporated data from previously tested spines with the same screw/PMMA construct but without disk spacer (cadaveric group 1). Results: Addition of a spacer increased construct stiffness in both the PVC model (P< .001) and cadaveric vertebral columns (P<.001) compared to fixation without a spacer. Conclusions: Addition of an intervertebral spacer significantly increased construct stiffness of monocortical screw/PMMA fixation. INTRODUCTION Canine cervical vertebral column stabilization is an emerging treatment option for dogs with traumatic injuries and cervical spondylomyelopathy (CSM). Few surgical techniques for stabilization of the CSM vertebral column few have been assessed biomechanically. In addition to stabilization, the goal of several canine surgical procedures is to create distraction between 2 vertebrae and ultimately achieve bony fusion across the intervertebral space. Bony fusion of the instrumented vertebrae should improve long-term stability and allow load-sharing with spinal implants, without which fatigue failure of the hardware is likely. Bicortical pins and polymethylmethacrylate constructs (PMMA) have been a landmark technique for cervical vertebral column stabilization in dogs; however, there is a high risk of implant violation into the vertebral canal and some recommend avoiding bicortical implants in the cervical vertebral column. A recent study documented the biomechanical equivalence of monoand bicortical screw and PMMA constructs in the cadaveric canine cervical vertebral column. In that study 3.5mm cortical stainless steel and titanium screws inserted monocortically performed similarly to bicortical 1/8 inch positive profile Steinman pins. Locking plates further enable the use of monocortical screw fixation. Whereas few comparative biomechanical studies have been performed, clinical reports support use of this approach for the canine cervical vertebral column. Treatment options for the intervertebral disk include leaving it intact, performing a ventral slot or partial diskectomy with bone grafting, or adding something structurally stronger such as washers, cement plugs or cortical grafts. Intervertebral spacers have been promoted as a method to improve construct biomechanics and maintain distraction after spinal fusion in people. Intervertebral spacers of various designs have been used in conjunction with vertebral column stabilization in animals but have not been biomechanically evaluated in dogs. Our purpose was to determine the biomechanical effects of an intervertebral spacer in cadaveric canine cervical vertebrae stabilized with monocortical screws and PMMA. A synthetic bone substitute model (polyvinyl chloride pipe) was developed and tested before evaluation in cadavers. We hypothesized that the addition of an intervertebral disk spacer would significantly increase construct stiffness compared with specimens instrumented without a spacer in a synthetic and cadaveric model. MATERIALS AND METHODS PVC models A cervical intervertebral space was simulated using two 20 cm long segments of 12.5 mm diameter PVC pipe (12.7 mm outer diameter, 2 mm wall thickness) separated by a gap. The PVC pipe was cut at a 30° angle to create a gap model mimicking vertebral endplate orientation and cervical disk space width similar to that of a large-breed dog. Twelve models were created using screw and PMMA fixation: 6 with a 4-mm gap with no spacer (PVC group 1) and 6 with a spacer filling the gap (PVC group 2). Spacers were created from 4-mm thick rings cut perpendicularly from PVC pipes to simulate a cortical ring allograft used in clinical patients with cervical distraction/fusion. All models were instrumented with 6 self-tapping titanium alloy cortical screws (Synthes Vet, West Chester, PA) inserted monocortically and PMMA (Simplex P Bone Cement, Stryker, Mahwah, NJ) for fixation of the simulated disk space. Three screws were inserted in a triangular fashion adjacent and parallel to the angled cut surface simulating the disk space. Holes were drilled with a 2.5mm drill bit through 1 wall of the PVC pipe and screws were advanced until they contacted the opposite inner surface of the ring. Screws protruded 12-15 mm from the PCV pipe and 20g PMMA were applied to form a uniform cement mantle (Fig. 1). Mechanical Data Collection – PVC models Small K-wires were placed in the dorsal and lateral plane to connect an extensometer which was used to measure motion at the interface. Construct stiffness in extension, flexion and lateral bending for both PVC groups was determined using a custom 4-point bend fixture. (Fig. 2 A) The testing protocol was similar to an earlier study using canine cervical vertebral specimens wherein testing was load-controlled at 50 N/min to 150 N in flexion and extension and to 100 N in right lateral bending. Each PVC model underwent only 2 full cycles of extension, flexion, and lateral bending because data were very similar during pilot PVC model testing. Actuator displacement (mm), applied load (N), and extensometer displacement (mm) data were collected continuously and recorded at 0.5 N intervals using the data acquisition system integral with the servohydraulic test frame (MTS Bionix 858 Test System, MTS, Eden Prairie, Minnesota). Load and extensometer displacement data from the 2 cycle of each loading direction were used to calculate load-displacement curves for bending moment of the PVC model with gap and with the PVC spacer. Stiffness (Nm/m) was determined by calculating the slope of the linear portion of each load-displacement curve. One PVC model from either group (without and with spacer) was subjected to fatigue testing to 200,000 cycles without evidence of failure. Vertebral Specimens The study was approved by the Institutional Animal Care and Use Committee. Cervical vertebral columns (C2-C7) were collected from 12 skeletally mature dogs (21-30 kg body weight) euthanatized for reasons unrelated to this study. Lateral and dorsoventral radiographic projections were made to exclude dogs with radiographic evidence of open physes or evidence of vertebral column deformities and other conditions affecting vertebrae or disk spaces. Six specimens were tested as part of another study under the same inclusion criteria and procedural protocols; these 6 spines had monocortical titanium screw and PMMA fixation without a spacer (cadaveric group 1). The other 6 spines represented specimens with an intervertebral spacer and fixation (group 2). Cervical vertebral columns were sorted into balanced groups based on dual-energy x-ray absorptiometry (DEXA) measures of bone mineral density at C4 and C5 (Lunar Prodigy; GE Healthcare, Milwaukee, WI). Surrounding soft tissues were removed except for paravertebral musculature, joint capsules and ligaments associated with the C3-C6 vertebrae. Specimens were wrapped in moist towels soaked in sterile saline (0.9% NaCl) solution and frozen at -20oC until testing. Specimens were kept moist during processing and testing with sterile saline solution. Cortical ring allografts from cadaveric canine tibiae were collected before vertebral column instrumentation and frozen at 20oC until testing. Ring dimensions (height and width) were adjusted to disk space dimensions of each specimen measured on radiographs. Ring depths varied between 4.0 and 5.2mm as determined with digital calipers. Mechanical Data Collection – Cadaveric Specimens Potting of the cervical vertebral columns was performed and allowed for isolated testing of the C4-C5 vertebral motion unit (VMU). Briefly, after removal of soft tissues (sparing C4-C5), the VMUs of C2-C4 and C5-C7 were stabilized and potted. K-wires were placed on midline in the dorsal and lateral plane in C4 and C5 to connect an extensometer to measure localized deformation at the C4-C5 VMU. Each specimen was allowed to thaw to room temperature and tested in extension, flexion and right lateral bending using a custom made four-point bending fixture (Fig. 2 B). A preload of 5 Newtons (N) was applied to stabilize the specimen and to assure that all specimen tests were initiated under the same conditions. Testing was load-controlled at 50 N/min to 150 N in flexion and extension and to 100 N in right lateral bending. Each specimen underwent 4 full cycles of extension, flexion and lateral bending and was allowed to rest in neutral for 30 seconds between each cycle to allow for tissue recovery. Actuator displacement (mm), applied load (N), and extensometer displacement (mm) data were collected as with the PCV model. Load and extensometer displacement data from the 4 cycle of each loading direction were used to calculate loaddisplacement curves for each bending moment of the unaltered and instrumented C4-C5 motion unit. Stiffness (Nm/m) was calculated by selecting the linear portion of each loaddisplacement curve. The same loading and data collection protocol was used in both groups (spacer, no spacer) after instrumentation with monocortical screws/PMMA. Surgical Fixation – Cadaveric Specimens The longus colli musculature was resected in an ~5cm long by 3.5cm wide area centered over the C4-C5 intervertebral disk. In group 1, the intervertebral disk remained unaltered. In group 2, a partial diskectomy was performed by removing the ventral annulus fibrosus, the nucleus pulposus and part of the remaining annulus, leaving only a thin rim of annulus intact along the lateral and dorsal borders. Manual distraction was applied to facilitate placement of a cortical ring allograft previously harvested from the tibial diaphysis of 2 of the study dogs. The C4-C5 VMUs of both groups were stabilized with 6 monocortical selftapping titanium alloy cortical screws (Synthes Vet, West Chester, PA) and PMMA as previously described (Fig. 3). Briefly, after predrilling the ciscortex, 3 screws each were inserted into C4 and C5 and advanced until they contacted the inner cortex of the vertebral canal. Screw orientation was parallel to the vertebral endplate for all but the most caudal screw, which was oriented in a cranioventral to caudodorsal direction toward the caudal endplate of C5 because of physical obstruction by the potting construct. Screws protruded 1215 mm from the ventral vertebral body surface to allow incorporation into 20 g PMMA to create a uniform cement mantle. Cement was allowed to harden for a minimum of 20 minutes before testing. Postoperative Implant Assessment – Cadaveric Specimens Post-testing, orthogonal radiographs were obtained to assess monocortical screw and cortical ring allograft position as well as potential bony damage from mechanical testing (Fig. 4) Specimens were then cleared of remaining soft tissues using a dermestid beetle colony and evaluated for potential screw violation of the vertebral canal and position of the cortical ring