**ISO 10993, Part 10: Test for Irritation and Delayed-Type Hypersensititivity-**

A guinea pig (Hartley albino) maximization test was conducted to evaluate the potential for delayed dermal contact sensitization. HA coated PEEK and coated PEEK that had the coating removed, were extracted in 0.9% sodium chloride USP solution (SC) and sesame oil, NF (SO). Each extract was intradermally injected and occlusively patched to ten test guinea pigs per extract in an attempt to incude sensitization. Five additional guinea pigs served as a control. All sites were scored at 24 and 48 hours after patch removal. HA coated PEEK and coated PEEK that had the coating removed, were extracted in 0.9% sodium chloride USP solution (SC) and sesame oil, NF (SO). A 0.2 mL dose was injected by the intracutaneous route into five separate sites on the right side of the the back of two New Zealand white rabbits. Similarly, the corresponding control was injected on the left side of the back of each rabbit. Observations for erythema and edema were conducted at 24, 48 and 72 hours after injection. Under the conditions of this study, there was no evidence of the extracts causing delayed dermal contact sensitization in the guinea pigs or no erythema or edema from the extracts injected intracutaneously into the rabbits. The requirements of the study were met.

### **ISO 10993, Part 11: Tests for Systemic Toxicity-**

HA coated PEEK and coated PEEK that had the coating removed, were extracted in 0.9% sodium chloride USP solution and sesame oil, NF. A single dose of the extract was injected into five mice per extract by either the intravenous or intraperitoneal route. Five mice served as a control. The animals were observed immediately and at 4, 24, 48 and 72 hours after injection. Under the conditions of this study, there was no mortality or evidence of systemic toxicity from the extracts, and therefore the requirements of the study were met.

### **2.1.6 Biomechanical analysis**

220 Recent Advances in Arthroplasty

readily removed with a mild solvent, it is highly unlikely that any covalent bonding occurred between the HA and the PEEK material. Furthermore, there was no measurable change in the molecular structure of the PEEK induced by adding the HA layer. Therefore, no chemical bonding was noted to occur, the surface properties and subsequently the biocompatibility profile of the PEEK substrate have not changed, and it can be concluded

Biocompatibility tests in accordance with ISO 10993 parts 5, 10 and 11 were conducted at NAMSA. These tests were carried out on coated and an equal number of substrates having the coating removed with a mild solvent (phosphoric acid) in order to expose the

HA coated PEEK and coated PEEK that had the coating removed, were prepared using single strength Miminum Essential Medium supplemented with 5% serum and 2% antibiotics (1X MEM). The test extracts were placed onto three separate monolayers of L-929 mouse fibroblast cells propagated in 5% CO2. Three separate monolayers were prepared for the reagent control, negative control and for the posititve control. All monolayers were incubated at 37° C in the presence of 5% CO2 for 48 hours at which point they were examined microscopically to determine any change in cell morphology. Under the conditions of this study, the 1X MEM test extract showed no evidence of causing cell lysis or toxicity. The reagent control, negative control and the positive control performed as

A guinea pig (Hartley albino) maximization test was conducted to evaluate the potential for delayed dermal contact sensitization. HA coated PEEK and coated PEEK that had the coating removed, were extracted in 0.9% sodium chloride USP solution (SC) and sesame oil, NF (SO). Each extract was intradermally injected and occlusively patched to ten test guinea pigs per extract in an attempt to incude sensitization. Five additional guinea pigs served as a control. All sites were scored at 24 and 48 hours after patch removal. HA coated PEEK and coated PEEK that had the coating removed, were extracted in 0.9% sodium chloride USP solution (SC) and sesame oil, NF (SO). A 0.2 mL dose was injected by the intracutaneous route into five separate sites on the right side of the the back of two New Zealand white rabbits. Similarly, the corresponding control was injected on the left side of the back of each rabbit. Observations for erythema and edema were conducted at 24, 48 and 72 hours after injection. Under the conditions of this study, there was no evidence of the extracts causing delayed dermal contact sensitization in the guinea pigs or no erythema or edema from the extracts injected intracutaneously into the rabbits. The

HA coated PEEK and coated PEEK that had the coating removed, were extracted in 0.9% sodium chloride USP solution and sesame oil, NF. A single dose of the extract was injected into five mice per extract by either the intravenous or intraperitoneal route. Five mice served as a control. The animals were observed immediately and at 4, 24, 48 and 72 hours after injection. Under the conditions of this study, there was no mortality or evidence of systemic

toxicity from the extracts, and therefore the requirements of the study were met.

that only a mechanical bond exists between the PEEK and HA coating.

underlying substrate. The control consisted of virgin PEEK:

anticitpated. Therefore the requirements of the study were met.

**ISO 10993, Part 10: Test for Irritation and Delayed-Type Hypersensititivity-**

**ISO 10993, Part 5: Tests for Cytotoxicity-**

requirements of the study were met.

**ISO 10993, Part 11: Tests for Systemic Toxicity-**

Using an *in-vitro* human cadaveric model, the range of motion (ROM) of the device was evaluated at a single level, with data analysis based on the resulting operative and adjacent level multidirectional flexibility properties. A total of twelve fresh-frozen human cadaveric cervical spines (C1T3) were harvested en-bloc and utilized. Prior to biomechanical analysis, standard anteroposterior and lateral plain films were obtained to exclude specimens demonstrating intervertebral disc or osseous pathology. In preparation for biomechanical testing, the specimens were thawed to room temperature and cleaned of all residual musculature, with care taken to preserve all ligamentous attachments and facet joint capsules. The proximal (C1C3) and distal (T1T3) ends of the specimen were stabilized in metal tubing containers using four, four point compression screws. An optoelectronic motion measurement system (3020 Optotrak System) was used for kinematic analysis under a pure moment loading to evaluate the flexion/extension, lateral bending and axial rotational response after reconstruction of the C4C5 cervical level. A comparision was made between the intact, destabilized, reconstructed and fusion procedures.

Flexion/extension testing demonstrated the diskectomy condition as producing a five degree increase (10.5 to 15.6°) in operative level range of motion compared to the intact spine and was statistically greater than intact and anterior plate-cage groups (p<0.05). Implantation of the PEEK cervical artificial disc restored motion to near intact level (11.9°) which was not statistically different from the intact spine (10.5°) (p > 0.05). The final reconstruction using the anterior plate-cage decreased the mean range of motion to 8.71° at the operative level, which resulted in significantly less range of motion than the PEEK artificial disc (11.9°) (p<0.05) but not intact condition (10.5 dg) (p>0.05). The elevated mean range of motion value for this reconstruction is attributable to the segmental instability created by the diskectomy condition (ANOVA F=3.72, p=0.018). Biomechanical testing in lateral bending loading did not reveal any significant differences in range of motion when comparing the intact spine and three treatment groups (p>0.05). Diskectomy (7.63°), PEEK artificial disc (6.90°) or anterior plate-cage construct (6.77°) afforded a statistical change in flexibility compared to the native intact condition (7.10°). The high standard deviation in the range of motion account for the lack of statistical significance under this loading modality (ROM: ANOVA F=0.58, p=0.633). Biomechanical testing in axial rotation indicated no statistical differences in ROM between the intact spine, diskectomy, PEEK total disc replacement or anterior plate-cage group (p>0.05). The cervical disc replacement procedure (7.21°) resulted in a slight increase in operative level flexibility relative to the intact (6.05°) and diskectomy conditions (6.44°) (p>0.05). The trends under this biomechanical loading mode failed to reach statistical significance secondary to specimen variability and high standard deviations. (ROM: ANOVA F=0.22, P=0.882). Despite the use of the Panjabi Hybrid testing protocol, there were very minimal changes in adjacent level ROM when comparing the intact spine, diskectomy condition and two reconstruction groups (p>0.05). The only significant change was observed at the adjacent level in flexion extension loading condition. The normalized-to-intact ROM at C6C7 level in these two reconstruction conditions was significantly higher than that was detected in diskectomy condition (p<0.05). Similar to the operative level ROM, the trends under this biomechanical loading mode failed to reach statistical significance secondary to specimen variability and high standard deviations.

The Use of PEEK in Spine Arthroplasty 223

Fig. 5. Left - Good bone apposition to the HA coated devices was seen. Fibrous membranes are around the edges of both components where there is no coating applied, but there is no significant fibrosis at the bone/implant interface. Right – mostly fibrous tissue surrounded

uncoated group, only focal bone ingrowth to the Ti cam's was noted, with no bone apposition to the PEEK, consistent with the relatively inertness of the material not allowing for bone apposition. For all groups, plain and polarized light microscopy of local tissues overlying the operative sites from experimental levels of both groups indicated a mild inflammatory reaction. The histiocytic infiltration responses were considered secondary to the surgical procedure and healing response. The spinal cord sections exhibited no significant pathological changes based on plain and polarized review. There was no evidence of giant cell reaction or other significant pathological changes, no debris associated with the device, aggresive inflammatory reactions, degenerative changes or

The PEEK nucleus replacement device is a two-piece implant comprised of a top and a bottom plate, with an inner ball and socket articulation (Figure 6). Since PEEK is radiolucent, tantalum (Ta) markers are used to facilitate the visualization of the device location when using radiographic methods. The articulation allows for rotational motions along all three major axes. The motion of the device in axial rotation is unconstrained, while the motions in flexion/extension and lateral bending are semi-constrained slightly beyond the physiological ROM. Similar to all nucleus replacement devices and unlike a total disc replacement, the device ranges of motion are largely dependent upon the constraints from the adjacent tissues, and is designed to restore or maintain the load sharing capabilities of

the uncoated devices. For all devices, no significant particles of debris were evident.

autolysis noted.

the index disc.

**2.2 PEEK nucleus replacement device** 

The greatest differences between the intact spine and three treatment groups were detected during flexion/extension loading. The fundamental trend in each loading condition demonstrated the diskectomy condition as producing, as expected, an increase in segmental motion. Reconstruction using the PEEK artificial disc restored motion to near intact levels. The final reconstruction using the anterior plate-cage construct reduced motion at the operative level. These reductions in segmental motion, however, were not as low as expected. These findings are attributable to the instability created by the diskectomy condition and follow patterns similar to previous biomechanical studies using anterior plate and interbody cages.

### **2.1.7 Pilot functional animal study**

Using an *in-vivo* goat model, evaluation of implant osseointegration and biocompatibility properties of HA coated and uncoated devices was performed. These properties were radiographically and histologically evaluated at three months post-operatively. Mature female crossbred goats served as the experimental model in this pilot study as these animals are gentle, easy to care for and represent excellent models for cervical spine procedures. There were n = 4 animals for the HA coated and uncoated groups. Following the three month sacrifice, the operative cervical motion segments were sagittally sectioned along the geometric centerline of the implanted device using a Beuhler Isomet saw. Histologic preparation of all samples included dehydration in 100% ethanol, staining using the Villanueva Osteochrome Bone Stain, undecalcified solution processing and embedding in polymethyl-methacrylate (PMMA). Using the EXAKT Microgrinding Device, the embedded specimens were cut at 250 to 300 um thick, ground and polished to 75 um. Microradiographs were obtained of the slide-mounted specimens using Faxitron radiography. The slides were placed twelve inches from the beam and exposed for two minutes, using a technique of 45 kVp and 3mA while in direct contact with the single emulsion high-resolution graphics arts film. For all animals, the soft tissue organs and structures were resected, sectioned and prepared by a veterinarian pathologist. The specimens were fixed in a 10% formalin solution, underwent routine paraffin processing and slide preparation. 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 and local tissue overlying the anterior aspect of the C3-C4 operative disc level were evaluated using routine H&E and Macrophage Staining Method (HAM-56). The use of this staining method highlights the presence of activated macrophages within the local tissues. Using plain and polarized light microscopy, histopathological readings of the slide-mounted tissue sections, 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.

All operative cervical motion segments were examined for gross evidence of infection and local pathologic response at the time of necropsy. There was no apparent infection in any of the operative C3-C4 levels or adjacent motion segments. Based on gross examination at the time of necropsy, the local tissues from all animals were considered unremarkable and without evidence of ectopic bone formation or other significant histopathologic changes.

Histopathological interpretation of the slide-mounted undecalcified specimens indicated no evidence of significant pathological changes in tissues within or surrounding the specimens. For the HA coated group, there was good intercalation of cancellous bone with the Ti fixation cams and bone apposition to both endplates of the devce (Figure 5). For the

The greatest differences between the intact spine and three treatment groups were detected during flexion/extension loading. The fundamental trend in each loading condition demonstrated the diskectomy condition as producing, as expected, an increase in segmental motion. Reconstruction using the PEEK artificial disc restored motion to near intact levels. The final reconstruction using the anterior plate-cage construct reduced motion at the operative level. These reductions in segmental motion, however, were not as low as expected. These findings are attributable to the instability created by the diskectomy condition and follow patterns similar to previous biomechanical studies using anterior plate

Using an *in-vivo* goat model, evaluation of implant osseointegration and biocompatibility properties of HA coated and uncoated devices was performed. These properties were radiographically and histologically evaluated at three months post-operatively. Mature female crossbred goats served as the experimental model in this pilot study as these animals are gentle, easy to care for and represent excellent models for cervical spine procedures. There were n = 4 animals for the HA coated and uncoated groups. Following the three month sacrifice, the operative cervical motion segments were sagittally sectioned along the geometric centerline of the implanted device using a Beuhler Isomet saw. Histologic preparation of all samples included dehydration in 100% ethanol, staining using the Villanueva Osteochrome Bone Stain, undecalcified solution processing and embedding in polymethyl-methacrylate (PMMA). Using the EXAKT Microgrinding Device, the embedded specimens were cut at 250 to 300 um thick, ground and polished to 75 um. Microradiographs were obtained of the slide-mounted specimens using Faxitron radiography. The slides were placed twelve inches from the beam and exposed for two minutes, using a technique of 45 kVp and 3mA while in direct contact with the single emulsion high-resolution graphics arts film. For all animals, the soft tissue organs and structures were resected, sectioned and prepared by a veterinarian pathologist. The specimens were fixed in a 10% formalin solution, underwent routine paraffin processing and slide preparation. 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 and local tissue overlying the anterior aspect of the C3-C4 operative disc level were evaluated using routine H&E and Macrophage Staining Method (HAM-56). The use of this staining method highlights the presence of activated macrophages within the local tissues. Using plain and polarized light microscopy, histopathological readings of the slide-mounted tissue sections, 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.

All operative cervical motion segments were examined for gross evidence of infection and local pathologic response at the time of necropsy. There was no apparent infection in any of the operative C3-C4 levels or adjacent motion segments. Based on gross examination at the time of necropsy, the local tissues from all animals were considered unremarkable and without evidence of ectopic bone formation or other significant histopathologic changes. Histopathological interpretation of the slide-mounted undecalcified specimens indicated no evidence of significant pathological changes in tissues within or surrounding the specimens. For the HA coated group, there was good intercalation of cancellous bone with the Ti fixation cams and bone apposition to both endplates of the devce (Figure 5). For the

and interbody cages.

**2.1.7 Pilot functional animal study** 

Fig. 5. Left - Good bone apposition to the HA coated devices was seen. Fibrous membranes are around the edges of both components where there is no coating applied, but there is no significant fibrosis at the bone/implant interface. Right – mostly fibrous tissue surrounded the uncoated devices. For all devices, no significant particles of debris were evident.

uncoated group, only focal bone ingrowth to the Ti cam's was noted, with no bone apposition to the PEEK, consistent with the relatively inertness of the material not allowing for bone apposition. For all groups, plain and polarized light microscopy of local tissues overlying the operative sites from experimental levels of both groups indicated a mild inflammatory reaction. The histiocytic infiltration responses were considered secondary to the surgical procedure and healing response. The spinal cord sections exhibited no significant pathological changes based on plain and polarized review. There was no evidence of giant cell reaction or other significant pathological changes, no debris associated with the device, aggresive inflammatory reactions, degenerative changes or autolysis noted.

### **2.2 PEEK nucleus replacement device**

The PEEK nucleus replacement device is a two-piece implant comprised of a top and a bottom plate, with an inner ball and socket articulation (Figure 6). Since PEEK is radiolucent, tantalum (Ta) markers are used to facilitate the visualization of the device location when using radiographic methods. The articulation allows for rotational motions along all three major axes. The motion of the device in axial rotation is unconstrained, while the motions in flexion/extension and lateral bending are semi-constrained slightly beyond the physiological ROM. Similar to all nucleus replacement devices and unlike a total disc replacement, the device ranges of motion are largely dependent upon the constraints from the adjacent tissues, and is designed to restore or maintain the load sharing capabilities of the index disc.

The Use of PEEK in Spine Arthroplasty 225

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

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

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

shear fatigue strength of the device.

fatigue strength for its intended application.

**2.2.3 Wear testing and wear particulate characterization** 

Fig. 6. PEEK nucleus replacement device.
