**2.2.1 Axial static and fatigue strength**

To determine the static and fatigue strength of the device, the smallest footprint (11x20 mm) and height (8 mm) was utilized. These were considered the worst-case scenario for contact stress distribution. The specimens tested were representative of the smallest true contact stress area for all current designs. Testing was performed on an MTS 858 Mini-Bionix II test frame with steel fixturing. A sample size of 3 was used for the static tests with a displacement rate of 2 mm/min. The acceptance criteria was an offset yield load of greater than 2200 N (Lowe, 2004) for the axial static testing. For the fatigue tests, a sample size of 6 was used with a runout load criteria of 3000 N to 5 million cycles (FDA guidance), with R=10 at 10 Hz.

The results of the axial static testing showed that the mean offset yield was 10868 ± 413 N. The primary failure mode was excessive plastic deformation of the top component of the device. For the axial fatigue test, two specimens were tested at a peak compressive load of 9238 N with both failing within 200 cycles due to excessive plastic deformation of the top component, similar to the axial static testing. Two other specimens ran out to 10 million cycles at a peak compressive load of 8151 N. However, upon visual inspection, fatigue cracks emanating from the radiographic marker hole were discovered. Therefore, the load was dropped down to 7065 N for an addition two samples, with a successful outcome. This represented 85%, 75% and 65% of the mean offset yield compressive load, respectively. These results are thought to be well beyond those of the expected *in vivo* physiological loads, and therefore failure was not expected to occur in either static axial compression or dynmaic axial compression.

### **2.2.2 Shear fatigue strength**

To determine the static and fatigue strength of the device, the smallest footprint (11x20 mm) and height (8 mm) was utilized. The specimens tested were also representative of the smallest true contact stress area for all current designs. Testing was performed on an MTS 858 Mini-Bionix II test frame with steel fixturing. A shear angle of 15° was selected based

To determine the static and fatigue strength of the device, the smallest footprint (11x20 mm) and height (8 mm) was utilized. These were considered the worst-case scenario for contact stress distribution. The specimens tested were representative of the smallest true contact stress area for all current designs. Testing was performed on an MTS 858 Mini-Bionix II test frame with steel fixturing. A sample size of 3 was used for the static tests with a displacement rate of 2 mm/min. The acceptance criteria was an offset yield load of greater than 2200 N (Lowe, 2004) for the axial static testing. For the fatigue tests, a sample size of 6 was used with a runout load criteria of 3000 N to 5 million cycles (FDA guidance), with

The results of the axial static testing showed that the mean offset yield was 10868 ± 413 N. The primary failure mode was excessive plastic deformation of the top component of the device. For the axial fatigue test, two specimens were tested at a peak compressive load of 9238 N with both failing within 200 cycles due to excessive plastic deformation of the top component, similar to the axial static testing. Two other specimens ran out to 10 million cycles at a peak compressive load of 8151 N. However, upon visual inspection, fatigue cracks emanating from the radiographic marker hole were discovered. Therefore, the load was dropped down to 7065 N for an addition two samples, with a successful outcome. This represented 85%, 75% and 65% of the mean offset yield compressive load, respectively. These results are thought to be well beyond those of the expected *in vivo* physiological loads, and therefore failure was not expected to occur in either static axial compression or dynmaic

To determine the static and fatigue strength of the device, the smallest footprint (11x20 mm) and height (8 mm) was utilized. The specimens tested were also representative of the smallest true contact stress area for all current designs. Testing was performed on an MTS 858 Mini-Bionix II test frame with steel fixturing. A shear angle of 15° was selected based

Fig. 6. PEEK nucleus replacement device.

**2.2.1 Axial static and fatigue strength** 

R=10 at 10 Hz.

axial compression.

**2.2.2 Shear fatigue strength** 

upon a review of the relevant literature. Extreme flexion of the lumbar spine appears to be approximately 10° on average, with slightly more at the levels L2L3 to L4L5 and slightly less at L1L2 and L5S1 when measured *in vivo* (Pearcy, 1984). Therefore, a 15° angle was considered to be at the physiological extreme. Two specimens were used to characterize the shear fatigue strength of the device.

Given the compressive loads obtained during the axial testing, only shear fatigue testing was performed. The two test specimens were successfully tested to 10 million cycles at 6600 N. There is limited data on flexion fatigue testing of human cadaveric lumbar motion segments. The available data states that fatigue failure can occur on average at 263 ±646 cycles with 9 kg of cyclic loading (representing a peak load of 3150 N) at 6° of flexion (Gallagher, 2005). It was therefore thought that the device has adequate compressive-shear fatigue strength for its intended application.

### **2.2.3 Wear testing and wear particulate characterization**

All implants used for wear testing were applicable to a clinical setting and were of the smallest height (8 mm) and footprint (11x20 mm) available. Three separate groups of six implants each were utilized in this wear study. Table 2 summarizes the test methodology. The peak compressive load of 1024 N was chosen based upon the physiological load sharing expected of nucleus replacement devices. The peak compressive load of 2000 N represents a worst case scenario in the event that the device must support the entire compressive load on the anterior column. Group 2 was exposed to 200 kGy of gamma radiation followed by simulated aging. The simulated aging process was similar to ASTM F2003-02, which was developed to measure accelerated aging of UHMWPE in air at 70° C and 5 atm of oxygen, except that the aging time was extended from 14 days to 40 days. All implants were presoaked in test fluid for at least 6 weeks. All groups were tested on a six-station spine wear simulator (EndoLab GmbH). The testing fluid consisted of newborn calf serum diluted with phosphate buffered saline (PBS) to a final protein content of 20 g/L. Ethylene-diamine tetraacetic acid (EDTA) was added to the serum at a concentration of 20 mM to bind the calcium ions present in the serum. EDTA is a known preservative, and together with the low protein content and PBS, the addition of sodium azide or other anti-bacterial agent was not used. The test fluid temperature was kept at 37 ± 3° C. Although not part of the test methodology, group 2 was temporarily stopped at 5 million cycles due to simulator availability whereupon the implants were immersed in saline at 37 ± 3° C for 33 weeks. Although unintended, this represented an additional test parameter. The test was subsequently resumed until 10 million cycles was reached. Unloaded soak controls were used for all groups to account for moisture uptake. Group 3 was tested at approximately double the compressive load. The tests were stopped at half million-cycle intervals to clean, dry and gravimetrically assess the wear rates. The average wear rates were determined using linear regression analysis with one-way-analysis-of-variance (ANOVA) used to determine if significant differences (p < 0.05) in the wear rates was present. The proteinacious test serum containing PEEK wear debris underwent enzymatic digestion (5x Trypsin digestion and 1.5 mg/mL of Proteinase K per sample at 37° C for 24 hours) followed by a mild acid treatment (10% HCl at 37° C for 24 hours), meeting or exceeding similar protocols previously established to be equivalent for acid and base digestion (Niedzwiecki, 2001). The particles were then analyzed by SEM analysis and characterized

The Use of PEEK in Spine Arthroplasty 227

Fig. 8. SEM analysis of the PEEK particulate generated from the wear tests.

The first study sought to assess the restoration of disc height, the potential adverse effects of the device on the underlying vertebral endplate under physiological loading conditions and an initial assessment of the extrusion risk. A total of eight fresh-frozen human lumbar functional spinal units (L2L3 and L4L5) were harvested en-bloc and utilized in this investigation. Prior to biomechanical analysis, standard anterior/posterior and standard lateral plain films were obtained to exclude specimens demonstrating intervertebral disc or osseous pathology. Bone mineral density (BMD) scans (DEXA) were performed to exclude specimens demonstrating BMD's less than 0.9 g/cm3. Each test specimen consisted of two contiguous functional spinal units (three vertebra and two intervertebral discs). One intervertebral disc served as a control and the other intervertebral disc was implanted with the device. To determine restoration of disc height, a 1200 N compressive load was applied. An annular incision followed by a box cut (6 mm x 10 mm) was created on the anatomic right, followed by a complete nucleotomy. A fatigue test was conducted under a left lateral bending moment ranging from 2.5 to 7.5 Nm (250 to 750 N compressive load offset 10 mm). The test was run at 2 Hz for 100,000 cycles. The left lateral bending mode represented a "worst case" scenario with regards to implant extrusion. Upon completion of the biomechanical analyses, each specimen was radiographed and examined for possible fractures and/or migration. Segments were then dissected transversely through the intervertebral discs. Both the intact and implanted discs were examined macroscopically for potential changes. Photographs of the disc were taken using a macro lens to document the condition of the endplates and the annulus. The L1 and L4 vertebrae were scanned on a mCT 80 (ScancoMedical AG, Bassersdorf, Switzerland) to examine the trabecular bone for

**2.2.4 Biomechanical analyses** 

according to their equivalent circle diameter (ECD), aspect ration (AR), roundness (R) and form factor (FF) in accordance with ASTM F 1877-05.

All implants for each group maintained full functionality throughout each test duration with visual and light microscopy revealing no evidence of gross deformation, delamination or fatigue cracks. Since group 1 desplayed a wear-in period from 0-1.0 million cycle, the wear rates were calculated from 1 million cycles on for all groups. For group 1, the wear rate was calculated to be 0.41 ± 0.01 mm3/million cycles. For group 2, the wear rate was calculated to be 0.52 ± 0.02mm3/million cycles. For group 3, the wear rate was calculated to be 0.92 ± 0.01 mm3/million cycles, which was significantly higher than groups 1 and 2. This increase in the wear rate was to be expected, given that the peak compressive load was increased from 1024 N to 2000 N. Overall, the wear rates are less than contempary lumbar total disc arthroplasty devices, which can range from 2.8 to 20.4 mm3/million cycles (Bushelow, 2007; Nechtow, 2006; Grupp, 2009) (Figure 7). The particle analysis conducted via SEM revealed that group 1 had the smallest particulate of 1.09 µm, followed by groups 2 and 3 at 1.71 µm and 1.86 µm respectively. Other than the aspect ratio for group 1, the morphology of the wear particulate appears to be similar for all the groups tested (Figure 8). Using standard nomenclature for describing particle morphology in accordance with ASTM F1877-05, the larger particulate tended to be mostly somewhat roughened or smooth flakes, with some shards and globular particles, along with a few fibrils. Any submicron particulate was smooth, with spheroidal granules.

Fig. 7. Average wear rates for the PEEK nucleus replacement device.

according to their equivalent circle diameter (ECD), aspect ration (AR), roundness (R) and

All implants for each group maintained full functionality throughout each test duration with visual and light microscopy revealing no evidence of gross deformation, delamination or fatigue cracks. Since group 1 desplayed a wear-in period from 0-1.0 million cycle, the wear rates were calculated from 1 million cycles on for all groups. For group 1, the wear rate was calculated to be 0.41 ± 0.01 mm3/million cycles. For group 2, the wear rate was calculated to be 0.52 ± 0.02mm3/million cycles. For group 3, the wear rate was calculated to be 0.92 ± 0.01 mm3/million cycles, which was significantly higher than groups 1 and 2. This increase in the wear rate was to be expected, given that the peak compressive load was increased from 1024 N to 2000 N. Overall, the wear rates are less than contempary lumbar total disc arthroplasty devices, which can range from 2.8 to 20.4 mm3/million cycles (Bushelow, 2007; Nechtow, 2006; Grupp, 2009) (Figure 7). The particle analysis conducted via SEM revealed that group 1 had the smallest particulate of 1.09 µm, followed by groups 2 and 3 at 1.71 µm and 1.86 µm respectively. Other than the aspect ratio for group 1, the morphology of the wear particulate appears to be similar for all the groups tested (Figure 8). Using standard nomenclature for describing particle morphology in accordance with ASTM F1877-05, the larger particulate tended to be mostly somewhat roughened or smooth flakes, with some shards and globular particles, along with a few fibrils. Any submicron particulate

form factor (FF) in accordance with ASTM F 1877-05.

was smooth, with spheroidal granules.

Fig. 7. Average wear rates for the PEEK nucleus replacement device.

Fig. 8. SEM analysis of the PEEK particulate generated from the wear tests.

### **2.2.4 Biomechanical analyses**

The first study sought to assess the restoration of disc height, the potential adverse effects of the device on the underlying vertebral endplate under physiological loading conditions and an initial assessment of the extrusion risk. A total of eight fresh-frozen human lumbar functional spinal units (L2L3 and L4L5) were harvested en-bloc and utilized in this investigation. Prior to biomechanical analysis, standard anterior/posterior and standard lateral plain films were obtained to exclude specimens demonstrating intervertebral disc or osseous pathology. Bone mineral density (BMD) scans (DEXA) were performed to exclude specimens demonstrating BMD's less than 0.9 g/cm3. Each test specimen consisted of two contiguous functional spinal units (three vertebra and two intervertebral discs). One intervertebral disc served as a control and the other intervertebral disc was implanted with the device. To determine restoration of disc height, a 1200 N compressive load was applied. An annular incision followed by a box cut (6 mm x 10 mm) was created on the anatomic right, followed by a complete nucleotomy. A fatigue test was conducted under a left lateral bending moment ranging from 2.5 to 7.5 Nm (250 to 750 N compressive load offset 10 mm). The test was run at 2 Hz for 100,000 cycles. The left lateral bending mode represented a "worst case" scenario with regards to implant extrusion. Upon completion of the biomechanical analyses, each specimen was radiographed and examined for possible fractures and/or migration. Segments were then dissected transversely through the intervertebral discs. Both the intact and implanted discs were examined macroscopically for potential changes. Photographs of the disc were taken using a macro lens to document the condition of the endplates and the annulus. The L1 and L4 vertebrae were scanned on a mCT 80 (ScancoMedical AG, Bassersdorf, Switzerland) to examine the trabecular bone for

The Use of PEEK in Spine Arthroplasty 229

Evaluation of the mechanical behavior and elicited histopathological response following long-term implantation in a functional animal (baboon) model of the device was performed. Fourteen mature male baboons (Papio cynocephalus) were randomized into two postoperative time periods of six-months (n=7) and twelve-months (n=7). Each animal underwent a lateral transperitoneal surgical approach followed by complete nucleotomy at L3-L4 and L5-L6 levels. The inferior L5-L6 level was reconstructed using the nucleus

Analyses were based on gross necropsy, MRI radiography, plain X-ray, microradiography, and biocompatibility assays (local and systemic histology) performed at the six and twelve month post-operative time-points. resected, sectioned and prepared by a veterinarian pathologist. The specimens were fixed in a 10% formalin solution, underwent routine paraffin processing and slide mounting. Using thin-sectioning microtomy, the paraffin embedded sections were cut (3-5μm in thickness), slide mounted and stained using standard Hematoxylin and Eosin (H&E). The spinal cord sections, obtained from each operative disc level, were evaluated using routine H&E. The local tissues overlying the operative levels underwent the following immunocytochemistry analyses using standard immunocytochemistry techniques, including primary and secondary antibodies combined with the avidin-biotin complex (ABC)-horseradish peroxidase. Immunohistochemical localization of intracellular and membrane bound macrophage expressing cytokines included IL-1, IL-2, IL-6 and Tumor Necrosis Factor-Alpha and beta (TNF-α,β). The Macrophage Staining Method (HAM-56) was used to highlight the presence of activated macrophages within the local tissues. Using plain and polarized light microscopy, histopathological readings of the slide-mounted tissues, activated macrophages, presence of wear debris as well as any signs of foreign body giant cells, granulomas inflammatory reactions, degenerative changes or autolysis could be discerned. Pathological assessment for all systemic and local tissues included, but was not limited to, the architecture of the tissues, presence of wear debris as well as any signs of foreign body giant cell / granulomas

Multidirectional flexibility testing indicated no difference in axial rotation (p>0.05) for the six-month groups, comparing the intact spine to surgical motion segments. Flexion/extension and lateral bending exhibited reduced segmental motion for the nucleus replacement device and operative control levels at six months versus the non-operative intact spine (p<0.05). However, these findings were also noted in the operative control levels demonstrating the effect of surgical intervention itself on segmental stability. At twelve months, segmental motion was restored to the intact levels for the reconstructed levels (p>0.05). Histopathologic analysis of the nucleus replacement treatment exhibited increased densification of trabeculae along the endplate periphery, which corroborated with Modic Type I changes observed with MRI and radiographically. Evidence of implant subsidence was noted in 5/7 cases at six months and 7/7 cases at twelve months, which correlated with the absence of articular cartilage in some regions. A limitation of this study was the use of a single height nucleus replacement device for all treated discs. This resulted in difficulties in insertion and placement of the device and significant over distraction of the disc space. The difficulties in insertion and over distraction may have resulted in the use of more force for proper placement. This potentially induced endplate damage at the treated disc levels and led to abnormal high contact stress on the endplates, which may have been the precursor to the Modic changes and subsidence. Plain and polarized light microscopy of local tissues overlying the operative sites from both the experimental and control levels indicated a

replacement device while the adjacent level served as a control.

inflammatory reactions, degenerative changes or autolysis.

**2.2.5 Functional animal study** 

potential fractures. The specimens were scanned with an x-ray energy of 70Kvp and current of 114 mA and at an isotropic voxel resolution of 74 microns (image matrix of 1024 x 1024). A total of 500 projections over 180 degrees were collected for each vertebra. An integration time of 700 ms/projection was used, which resulted in a scan-time of about 2.5 - 3 hours/specimen. A repeated measures ANOVA was used to examine differences in ROM between intact, discectomy, and implant conditions. Implant extrusion was assessed by direct observation made from the pre and post fatigue radiographs. Gross macroscopic fractures were determined from the dissection and radiographs for both the control and surgical levels. The microCT images were analyzed by a board certified radiologist (Medical Metrics, Inc. Houston, Texas, USA). The disc height of 2.2 ± 0.6 mm after discectomy significantly decreased (p<0.05) to 3.7 ± 1.1 mm for all specimens in comparison to the intact condition under 1.2 kN of compressive loading. Implantation of the nucleus device significantly increased (p<0.05) the disc height to 2.6 ± 0.7 mm compared to the discectomy condition. There was no significant difference between the intact and implanted conditions. There were no gross endplate fractures observed at the completion of 100,000 cycles of lateral bending. The evaluation of the microCT images noted no fractures in the endplates or vertebral bodies. There were slight migrations of the device towards the annulotomy, but no extrusions noted.

The second study performed was to quantify the ROM and destructive load to failure properties of the device. A total of eight fresh-frozen human lumbar functional spinal units (L2L3 and L4L5) were harvested en-bloc and utilized in this investigation. Prior to biomechanical analysis, standard anterior/posterior and lateral plain films were obtained to exclude specimens demonstrating intervertebral disc or osseous pathology. BMD scans using DEXA were performed to exclude specimens demonstrating BMD's less than 0.9 g/cm3. To determine the multidirectional flexibility, six applied pure moments of 7 Nm (flexion and extension, left and right lateral bending, and left and right torsion) were applied to the superior end of the vertically oriented specimen while the caudal portion of the specimen remained fixed to a testing platform. Following the intact analysis, an annular incision followed by a box cut (10 mm x 10 mm) was created on the anatomic left, followed by a complete nucleotomy. After biomechanical testing of the destabilized condition, the nucleus device was implanted and the operative motion segment retested. As a final test, the reconstructed specimen was destructively evaluated under axial compression. Biomechanical data was normalized to the intact condition and expressed as mean ± one standard deviation, with a Repeated Measures Analysis of Variance (ANOVA) and Student-Newman-Keuls test to determine differences among individual groups (p<0.05).

Multidirectional flexibility testing indicated significant increases in the segmental range of motion and neutral zone secondary to the annulotomy/nucleotomy procedures after reconstruction of the operative segment. For both calculated parameters, the segmental rotation increased for the destabilized condition versus the intact and reconstructed specimens (p<0.05). Importantly, the neutral zone, an indicator of spinal stability of the reconstructed segment, returned to levels not statistically different from the intact condition. As a final test, the reconstructed specimens were destructively evaluated under axial compression. In all specimens except one, the observed failure mechanism was fracture of the vertebral body, without significant disruption of the vertebral endplate. The primary mode of fracture was through the vertebral bodies above and below the operative disc level. The mean failure load was 3340 ± 2029 N. This average fracture load is comparable to the compressive failure load for intact lumbar segment (Lowe, 2004).

### **2.2.5 Functional animal study**

228 Recent Advances in Arthroplasty

potential fractures. The specimens were scanned with an x-ray energy of 70Kvp and current of 114 mA and at an isotropic voxel resolution of 74 microns (image matrix of 1024 x 1024). A total of 500 projections over 180 degrees were collected for each vertebra. An integration time of 700 ms/projection was used, which resulted in a scan-time of about 2.5 - 3 hours/specimen. A repeated measures ANOVA was used to examine differences in ROM between intact, discectomy, and implant conditions. Implant extrusion was assessed by direct observation made from the pre and post fatigue radiographs. Gross macroscopic fractures were determined from the dissection and radiographs for both the control and surgical levels. The microCT images were analyzed by a board certified radiologist (Medical Metrics, Inc. Houston, Texas, USA). The disc height of 2.2 ± 0.6 mm after discectomy significantly decreased (p<0.05) to 3.7 ± 1.1 mm for all specimens in comparison to the intact condition under 1.2 kN of compressive loading. Implantation of the nucleus device significantly increased (p<0.05) the disc height to 2.6 ± 0.7 mm compared to the discectomy condition. There was no significant difference between the intact and implanted conditions. There were no gross endplate fractures observed at the completion of 100,000 cycles of lateral bending. The evaluation of the microCT images noted no fractures in the endplates or vertebral bodies. There were slight migrations of the device towards the annulotomy, but no

The second study performed was to quantify the ROM and destructive load to failure properties of the device. A total of eight fresh-frozen human lumbar functional spinal units (L2L3 and L4L5) were harvested en-bloc and utilized in this investigation. Prior to biomechanical analysis, standard anterior/posterior and lateral plain films were obtained to exclude specimens demonstrating intervertebral disc or osseous pathology. BMD scans using DEXA were performed to exclude specimens demonstrating BMD's less than 0.9 g/cm3. To determine the multidirectional flexibility, six applied pure moments of 7 Nm (flexion and extension, left and right lateral bending, and left and right torsion) were applied to the superior end of the vertically oriented specimen while the caudal portion of the specimen remained fixed to a testing platform. Following the intact analysis, an annular incision followed by a box cut (10 mm x 10 mm) was created on the anatomic left, followed by a complete nucleotomy. After biomechanical testing of the destabilized condition, the nucleus device was implanted and the operative motion segment retested. As a final test, the reconstructed specimen was destructively evaluated under axial compression. Biomechanical data was normalized to the intact condition and expressed as mean ± one standard deviation, with a Repeated Measures Analysis of Variance (ANOVA) and Student-

Newman-Keuls test to determine differences among individual groups (p<0.05).

compressive failure load for intact lumbar segment (Lowe, 2004).

Multidirectional flexibility testing indicated significant increases in the segmental range of motion and neutral zone secondary to the annulotomy/nucleotomy procedures after reconstruction of the operative segment. For both calculated parameters, the segmental rotation increased for the destabilized condition versus the intact and reconstructed specimens (p<0.05). Importantly, the neutral zone, an indicator of spinal stability of the reconstructed segment, returned to levels not statistically different from the intact condition. As a final test, the reconstructed specimens were destructively evaluated under axial compression. In all specimens except one, the observed failure mechanism was fracture of the vertebral body, without significant disruption of the vertebral endplate. The primary mode of fracture was through the vertebral bodies above and below the operative disc level. The mean failure load was 3340 ± 2029 N. This average fracture load is comparable to the

extrusions noted.

Evaluation of the mechanical behavior and elicited histopathological response following long-term implantation in a functional animal (baboon) model of the device was performed. Fourteen mature male baboons (Papio cynocephalus) were randomized into two postoperative time periods of six-months (n=7) and twelve-months (n=7). Each animal underwent a lateral transperitoneal surgical approach followed by complete nucleotomy at L3-L4 and L5-L6 levels. The inferior L5-L6 level was reconstructed using the nucleus replacement device while the adjacent level served as a control.

Analyses were based on gross necropsy, MRI radiography, plain X-ray, microradiography, and biocompatibility assays (local and systemic histology) performed at the six and twelve month post-operative time-points. resected, sectioned and prepared by a veterinarian pathologist. The specimens were fixed in a 10% formalin solution, underwent routine paraffin processing and slide mounting. Using thin-sectioning microtomy, the paraffin embedded sections were cut (3-5μm in thickness), slide mounted and stained using standard Hematoxylin and Eosin (H&E). The spinal cord sections, obtained from each operative disc level, were evaluated using routine H&E. The local tissues overlying the operative levels underwent the following immunocytochemistry analyses using standard immunocytochemistry techniques, including primary and secondary antibodies combined with the avidin-biotin complex (ABC)-horseradish peroxidase. Immunohistochemical localization of intracellular and membrane bound macrophage expressing cytokines included IL-1, IL-2, IL-6 and Tumor Necrosis Factor-Alpha and beta (TNF-α,β). The Macrophage Staining Method (HAM-56) was used to highlight the presence of activated macrophages within the local tissues. Using plain and polarized light microscopy, histopathological readings of the slide-mounted tissues, activated macrophages, presence of wear debris as well as any signs of foreign body giant cells, granulomas inflammatory reactions, degenerative changes or autolysis could be discerned. Pathological assessment for all systemic and local tissues included, but was not limited to, the architecture of the tissues, presence of wear debris as well as any signs of foreign body giant cell / granulomas inflammatory reactions, degenerative changes or autolysis.

Multidirectional flexibility testing indicated no difference in axial rotation (p>0.05) for the six-month groups, comparing the intact spine to surgical motion segments. Flexion/extension and lateral bending exhibited reduced segmental motion for the nucleus replacement device and operative control levels at six months versus the non-operative intact spine (p<0.05). However, these findings were also noted in the operative control levels demonstrating the effect of surgical intervention itself on segmental stability. At twelve months, segmental motion was restored to the intact levels for the reconstructed levels (p>0.05). Histopathologic analysis of the nucleus replacement treatment exhibited increased densification of trabeculae along the endplate periphery, which corroborated with Modic Type I changes observed with MRI and radiographically. Evidence of implant subsidence was noted in 5/7 cases at six months and 7/7 cases at twelve months, which correlated with the absence of articular cartilage in some regions. A limitation of this study was the use of a single height nucleus replacement device for all treated discs. This resulted in difficulties in insertion and placement of the device and significant over distraction of the disc space. The difficulties in insertion and over distraction may have resulted in the use of more force for proper placement. This potentially induced endplate damage at the treated disc levels and led to abnormal high contact stress on the endplates, which may have been the precursor to the Modic changes and subsidence. Plain and polarized light microscopy of local tissues overlying the operative sites from both the experimental and control levels indicated a

The Use of PEEK in Spine Arthroplasty 231

Fig. 10. ODI and VAS scores from the US IDE pilot study for the nucleus replacement

The concept of reconstruction of the lumbar and cervical disc is not new, as clinical experimentation goes back to the early 1950's. Since then, the scientific and clinical community has sought to improve upon the selection of available biomaterials for use in the disc arthroplasty arenas. The mainstay of these materials has been ceramics, metal alloys and polymers, in large part due to the history gained form their use in total joint reconstruction. Currently there is no universal material that is considered the ultimate biomaterial, and an in-depth knowledge of a given material is important in understanding and predicting its response when used in an application as demanding as an implantable biomaterial. Knowledge can be gained by not only performing the appropriate testing on the bulk properties of the material, but also proper preclinical testing of the constructed device. More importantly, adroit interpretation of these results is fundamental to the clinical success of the device (Fraser, 2004; Kurtz, 2009). The use of PEEK in spinal arthroplasty represents a new application of this material. Although the history of PEEK suggests that it has the necessary material properties to serve as a long-term implantable arthroplasty material, it's use in the form of any arthroplasty device has not been diligently explored. Therefore, a battery of preclinical testing was performed, and the interpretation of the results for both a cervical disc arthroplasty device and lumbar nucleus replacement device have allowed for successful advancement to the clinical usage. The results of these

device.

**4. Conclusion** 

chronic inflammatory reaction, with evidence of fibrous connective replacement and infiltration of mononuclear cells. These observations were considered secondary to the surgical procedure and, importantly, occurred at both the control and experimental operative levels. There was no evidence of cellular apoptosis, giant cell reaction or other significant pathological changes. Analysis of the immunohistochemical antibody stains for the local tissues overlying the experimental and control levels were negative in each case. In the histological analyses, there was no detectable wear debris from the device and no evidence of an osteolytic response. There was no evidence of a pro-inflammatory cytokine reaction at any of the experimental or control levels.
