From September 2014 to December 2016, Consecutive ACDF surgeries (N = 44) using a BGS-7 spacer (NOVOMAX®-C, CGBio Co., Ltd., Seoul, Republic of Korea) or allograft spacer (Cervical spacer C+, CGBio Co., Ltd., Seoul, Republic of Korea) were performed in a single medical center by a single expert surgeon. We compared fusion rate and clinical outcomes between the two groups of patients at more than 60 months after surgery. The disease entity was a degenerative cervical disease from the C3/4 to C6/7 levels, including cervical disc herniation and foraminal stenosis. The use of patient data for research purposes was approved by the institutional review boards (IRB No. 3-2021-0437), and the patients provided written informed consent.

The inclusion criteria were 1) adult men and women over the age of 19, 2) patients with bioactive ceramic spacer or allobone for cervical anterior intervertebral disc removal in single or multiple segments fusion from September 2014 to December 2016, 3) a patient who voluntarily signed a written consent form. The exclusion criteria included 1) patients who voluntarily did not agree to participate in the study, 2) Patients with other factors influencing clinical trial product effectiveness assessment.

The surgical techniques followed a standard ACDF surgery procedure, which includes a plate augmented interbody fusion, using a BGS-7 spacer or allograft. Linear anterior incisions were made on the opposite side of the main neural compression. For example, if the patient had a left-side protrusion at C5/6, the incision was made on the right side. The Smith-Robinson anterior cervical approach was employed to expose the anterior surface of the cervical disc. The disc material was removed with a No 15. blade and forceps. Meticulous curettage was performed to remove the cartilage portion of the upper and lower endplate, leaving the bony structure intact. The uncinate process was removed with a high-speed drill to obtain a wide decompression of the foraminal space, if needed. The posterior longitudinal ligament was resected only when it was needed for decompression, such as for the removal of a ruptured disc material that exited through the ligament. After confirmation of a complete neural decompression, a proper sized BGS-7 spacer or allograft was impacted into the empty disc space. Plate augmentation was applied for all patient’s entire surgical levels.

Radiologic assessment was performed using a cervical spine X-ray series at the preoperative, immediate postoperative, and long term follow-up time points. A breakage of the cage was defined as the presence of a crack within the inserted cage or a destructive change to the cage’s shape on the postoperative cervical spine X-ray. Subsidence was defined as a decrease in interbody height of more than 3 mm on the follow-up plain film. The difference of subsidence was based on the length between the center of the upper and lower endplate in a lateral cervical spine X-ray. Fusion status was evaluated by the Brantigan and Steffee fusion grade using an X-ray at the final follow-up visit (Table 1) (Brantigan and Steffee, 1993), and also by stability based inter-spinous motion measurement methods. (Song et al., 2014). The clinical outcome was evaluated using neck visual analogue scale (VAS), neck disability index (NDI) and Japanese Orthopaedic Association (JOA) neck questionnaire on the patient’s follow-up time points.

The mechanical performance of the BGS-7 spacer was confirmed using finite element analysis. The comparative model is the allograft spacer. Implant models and a three-dimensional cervical spine vertebral body model were constructed. The region of interest between implants and bone-implant interface was set up using different element sizes that could be distinguished from other parts. The Young’s modulus and Poisson’s ratio were applied to each moiety using previously published data in the literature. (Table 2) (Kumaresan et al., 1999; Kim et al., 2020; Kwak et al., 2021)

Two types of surgical models were constructed by placing each implant between the vertebral bodies using Abaqus software. (Hibbett, Karlsson & Sorenson Inc, JOHNSTON, United States) Sliding contact was applied as the boundary condition between the vertebral body and the implant. The inferior endplate of the most caudal vertebra (C6) was fixed in all degrees of freedom, while 1,200N loads were applied to the superior endplate of the most cephalic vertebral level (C6). (Patwardhan et al., 2000). The peak von mises stress (PVMS) and contact pressure that occurs when the same compressive load of 1,200N is applied to the two models are predicted and compared. Information on the number and type of elements used in the FEA study is as follows; element type; tetrahedral mesh (both), number of nodes; 248,215 (BGS-7 spacer) vs. 239,802 (allograft spacer), number of elements; 1,341,817 (BGS-7 spacer) vs. 1,283,606 (allograft spacer). The stress and contact pressure of the vertebral body were used to determine the risk of subsidence, and the stress and contact pressure of the implant were used to evaluate the risk of implant yield.

Stress and contact pressure of the vertebral body were used to determine the risk of subsidence, and the risk of implant yield was confirmed using the stress of the implant.

Descriptive data are represented as mean ± standard error of the means, and statistically compared using two-sample t-test or paired t-test. For the comparison of categorical variables, chi-square tests and Fisher’s exact tests were performed between two independent groups. All p-values less than 0.05 were considered statistically significant. All statistical analyses were performed using SPSS version 26.0 (SPSS Inc., Chicago, IL, United States) by a statistical professional.

Among the patients who underwent an ACDF surgery using the BGS-7 spacer and Allograft spacer, there were 30 men and 14 women. (Table 3). The average age at the operation was 47.69 ± 10.49 in allograft spacer and 51.67 ± 11.03 in BGS-7 spacer. The mean follow-up period was 89.18 ± 5.44 months. The number of fused segments ranged from a single-level to 3-level. The thirty-four patients had surgery with a 1-level fusion, the nine patients with a 2-level fusion, and one patient with a 3-level fusion.

BGS-7 spacer showed no adverse findings at 60 months after surgery when compared with allograft spacer (Table 4). Average subsidence length was 1.33 ± 0.74 mm per level in BGS-7 spacer group and 2.27 ± 1.39 mm per level in allograft spacer group. Subsidence defined as greater than 3 mm was found in 4 of 18 patients (22.22%) only in the allograft spacer group but not in the BGS-7 spacer group. However, all of them did not complain of significant axial pain and the subsidence rate didn’t increase over follow up period. No breakage was observed like fracture line or deformation of the BGS-7 spacer on X-ray during the whole follow-up period. However, six patients in the allobone spacer group showed breakage of the spacer that resulted in a decreased height rather than the intended disc height.

The result of fusion according to Brantigan and Steffee grading is compared allograft group and BGS-7 group at Table 5 and Figure 1. Patients who were followed for less than 6 months showed stable settlement without early instrumental failure such as cage extrusion. Patients who were followed for 12 months demonstrated a settled fusion progression. Allograft group showed twenty-three of 26 patients who were followed for more than 60 months had satisfactory fusion results with grades 4,5; however, they included 2,1 patients with fusion grade 2, 3, respectively. Twelve patients who were followed for 12 months demonstrated satisfactory fusion result with grades 4, 5 in BGS-7 group. Eighteen of 18 patients who were followed for more than 60 months had enough fusion results with grades 4,5; and there were no patients with fusion grade 2, 3, respectively.

The sagittal and coronal reconstruction of CT 1 year postoperatively revealed a detailed image of a bony bridge formation, which surrounded the spacer (Figure 2). An example of bioactive glass induced segmental fusion progression after ACDF surgery during 5 years is shown in Figure 3. After 5 years of surgery, it was confirmed that the endplate was in close contact with the spacer boundary, and new bone formation was observed surrounding the spacer (Figure 3).

Table 6 showed pseudoarthrosis evaluation that dynamic stability based fusion status at long-term postoperative period, and there were no statistical differences.

The result of clinical outcomes for the BGS-7 group and the allograft group are shown in Table 7 and Figure 4. Visual analogue scale (VAS) for cervical pain showed 5.6 ± 3.0 and 6.0 ± 3.8 at pre-operative status for BGS-7 spacer and allograft spacer, respectively (p = 0.706). And the VAS at the immediate postoperative status and long-term postoperative status for BGS-7 spacer and allograft spacer were 1.8 ± 2.0 and 2.1 ± 2.1 (p = 0.662), and 2.6 ± 2.5 and 3.4 ± 2.9 (p = 0.365), respectively.

Neck disability index (NDI) showed 12.3 ± 6.8 and 16.7 ± 7.50 at pre-operative status for BGS-7 spacer and allograft spacer, respectively (p = 0.064). And the NDI at the immediate postoperative status and long-term postoperative status for BGS-7 spacer and allograft spacer were 5.6 ± 4.4 and 7.3 ± 5.9 (p = 0.309), and 6.9 ± 7.1 and 9.3 ± 7.1 (p = 0.293), respectively.

Japanese orthopedic association (JOA) score showed 15.4 ± 1.5 and 15.5 ± 2.2 at pre-operative status for BGS-7 spacer and allograft spacer, respectively (p = 0.853). And the NDI at the immediate postoperative status and long-term postoperative status for BGS-7 spacer and allograft spacer were 15.8 ± 1.4 and 16.5 ± 0.6 (p = 0.037*), and 15.4 ± 2.1 and 16.2 ± 1.6 (p = 0.157), respectively.

In order to compare the risk of subsidence and implant yield of the allograft spacer and the BGS-7 spacer, the peak von mises stress (PVMS) and contact pressure of the vertebral body/implant were analyzed by finite analysis (Figures 5A, B). The peak von mises stress (PVMS) of the vertebral body due to the compressive load was measured to be about 57% higher the allograft spacer (58.57 MPa) than the BGS-7 spacer (37.26 MPa). (Figures 5C, D) The peak von mises stress generated in the implant is higher in the allograft spacer despite the lower elastic modulus of the material. (Figures 5E, F) The contact pressure of the allograft spacer is higher than that of the BGS-7 spacer, indicating that the risk of subsidence of the BGS-7 spacer is lower than that of allograft spacer. (Figures 5G, H). It can be expected that the allograft spacer has a higher risk of implant yield due to compressive load.

At first, a limited number of enrolled patients are weak point of this study. More accumulated data are required although this study revealed the long-term clinical result of ACDF surgery with BGS-7 interbody fusion spacer. And stand-alone cage insertion without a plate augmentation might be better surgical technique to reveal the direct effect of spacer and maximize the difference from previously commercialized cervical spacer. Because this study was the first long-term clinical trial using BGS-7 glass ceramic spacer, a plate augmentation had to be applied for the safety of enrolled patients.