Cervical disk replacement was designed to be an ideal substitute for fusion, preserving motion at the operative level for selected single-level cervical radiculopathy and myelopathy.1
Clinical comparisons with fusion found that motion-preserving surgery tended to yield superior outcomes,1–3
which led to cervical arthroplasty’s growing popularity. Advantages of disk replacement include improved neurologic status, better Neck Disability Index (NDI) score, preserved segmental mobility, reduced incidence of reoperation, and fewer major surgical complications.2,4–6
However, no protective effect of arthroplasty on adjacent segments has been found at 1 to 2 years.1,7
Thus, it remains unclear whether arthroplasty accomplishes its major aim—preventing symptoms of adjacent segment degeneration—in the long term.1,7,8
As a result, the debate between advocates of cervical motion-preserving and fusion surgery is likely to continue until a long-term follow-up study reveals whether cervical arthroplasty brings more advantages to patients.
The Bryan Cervical Disc can be implanted with either the Bryan Cervical Disc System or the ACCEL system (Medtronic Inc, Minneapolis, Minnesota). The ACCEL system has not obtained FDA approval but has been approved in China. No clinical comparison of the safety and effectiveness of these 2 implant systems has been previously reported. The current study compared the outcomes of the 2 systems in single-level cervical disk replacement.
The suitable size for each implant was estimated based on a preoperative CT scan. The patient was intubated for general anesthesia and positioned supine. The physiological cervical anterior arch was maintained and fixed in the neutral position by affixing the chin with tape. The shoulders were taped to the table and pulled caudally for optimal lateral fluoroscopic C-arm visualization. A standard ACDF procedure was used to expose the target disk, which was confirmed by fluoroscopy. Diskectomy was performed on the target disk. The dura was carefully exposed, and neural elements were completely explored and decompressed. A high-speed burr was used to clean the osteophytes at the target intervertebral space, ensuring that the anterior edges of the upper and lower vertebral bodies were in the same plane. An 8.5-mm-height cam distracter tested the intervertebral space height. If the distracter stretched too tight or loose, the space would be deemed unsuitable for replacement, and ACDF would be performed instead of motion-preserving surgery.
For patients with the Bryan Cervical Disc System, the retractor frame was accommodated on the table and adjusted horizontally. The bone-milling rig was secured on the retractor frame after the midpoint of the vertebra, as confirmed by a gravity-guided caliper device. The milling guide was carefully positioned precisely in the sagittal middle section of the vertebral and middle cross-section of the intervertebral space, referring to the sagittal wedge placed into the diskectomy site and the gradienter on the rig tail. Precise milling ensured a proper bone-implant interface. The prosthesis could then be implanted.
For patients with the ACCEL system, the milling guide was fixed on the adjacent vertebrae with 2 single screws based on the guidance of a sagittal wedge and a gradienter but without the accommodation of the retractor frame. The milling and implantation followed essentially the same process as that with the Bryan Cervical Disc System.
Before implanting the prepared prosthesis, a second check was made to confirm complete decompression and to ensure that the space had been washed with saline repeatedly. After implantation, final anteroposterior and lateral fluoroscopic images were taken to ensure proper sizing and positioning of the implant. The prevertebral fascia, platysma, subcutanea, and skin were sutured in layers. A drainage tube was placed.
The first 3 days postoperatively, 40 mg of methylprednisolone and 40 mg of omeprazole were given intravenously twice a day. Patients also received a 200-mg oral dose of celecoxib twice a day for 1 week. Soft collar protection lasted for 1 week. Cervical motion rehabilitation training began on postoperative day 3.
Seventy-five patients were screened for possible participation in the study. Eleven patients did not meet the inclusion criteria, and 8 others refused to participate. The first 3 motion-preserving surgeries with each implanting system were not included. Ultimately, 50 patients were included in the study, 24 in group A (the Bryan Cervical Disc System) and 26 in group B (the ACCEL system).
Of the 50 patients, 1 had the C3–C4 segment replaced, 7 had the C4–C5 segment replaced, 32 had the C5–C6 segment replaced, and 10 had the C6–C7 segment replaced. Twenty-eight were men and 22 were women. Mean patient age was 43.7±7.3 years (range, 29–58 years). All patients completed at least 2 years of follow-up. Mean follow-up was 36.3±10.9 months (range, 24–57 months). There were no statistically significant differences in the distributions of demographic characteristics between the 2 groups (Table 1
Pain intensity was similar between the 2 groups preoperatively (P
>.05) (Table 2
). Patients in both groups reported significant pain relief postoperatively (P
<.05). In the 2 years following replacement, pain intensities for both groups remained low. No pain recurrence was found for either group (P
>.05). No significant difference in pain was found between the groups at any follow-up point (P
VAS Scores for Neck and Arm Pain
Cervical function was not significantly different between the 2 groups preoperatively (P
>.05) (Table 3
). Disk replacement significantly improved cervical function postoperatively in both groups (P
<.05) and maintained it stably over the following 2 years (P
>.05). No significant difference in cervical function was found between the groups at any follow-up point (P
Short Form 36 physical scores, mental scores, and sum scores were not statistically different between the 2 groups preoperatively (P
>.05) (Table 4
). Both groups achieved significant improvement after implantation (P
<.05) and maintained the improvement during the 2-year follow-up. Changes in quality of life were not significantly different between the 2 groups at any follow-up point (P
Change in Patient Quality of Life
All patients (100%) rated their surgery outcome as either excellent (n=6; 20%) or good (n=24; 80%) with the Odom scale. No distribution difference was found between the groups (P=.507; Pearson value=1.360).
In group A, mean operative time was 164.6±42.4 minutes and average blood loss was 243.6±151.3 mL (Table 5
). In group B, mean operative time was 137.5±19.3 minutes and average blood loss was 138.1 ± 86.7 mL. Both operative time and blood loss were significantly lower in group B.
Three (12.5%) patients in group A experienced complications. The epidural vein was ruptured during decompression in 1 patient, leading to a 400-mL blood loss and a 185-minute operative time. The patient recovered well, with symptoms significantly relieved postoperatively. The other 2 patients experienced temporary throat discomfort postoperatively but had no hoarseness or difficulty swallowing or breathing. Local symptomatic treatment completely alleviated the symptom prior to discharge. In group B, prostheses were found to have shifted slightly forward in 2 (7.6%) patients 1 month after implantation (Figure 1
). Those patients felt no obvious discomfort. Soft collar protection and intensive observation were applied. No further loosening or dislocation was observed on fluoroscopic images. The overall complication incidence rates were similar between the 2 groups (P
=.571; Pearson value=0.321).
Bryan Cervical Disc (Medtronic Inc, Minneapolis, Minnesota) migration after replacement of the C5–C6 segment. Preoperative lateral cervical radiograph showing degeneration of C5–C6 (A). Preoperative cervical sagittal magnetic resonance image showing degeneration and herniation of the C5–C6 intervertebral disk compressing the dura (B). Lateral cervical radiograph 2 weeks postoperatively showing the prosthesis in a good position (C). Lateral cervical radiograph 1 month postoperatively showing the prosthesis migrating slightly forward, although the patient was asymptomatic (D). Lateral cervical radiograph 1 year postoperatively showing no significant changes in displacement (E). Lateral cervical radiograph 2 years postoperatively showing that the prosthesis was stable and maintaining its position (F).
Radiological evaluation revealed that all prostheses functioned well in flexion, extension, and translation (Figure 2
). No cervical instability, prostheses subsidence, or ectopic ossification was found during follow-up.
Bryan Cervical Disc (Medtronic Inc, Minneapolis, Minnesota) replacement for the C4–C5 segment. Preoperative lateral cervical radiograph showing degeneration at C4–C5 (A). Preoperative sagittal 3-dimensional computed tomography reconstruction section of the cervical spine showing posterior disk osteophyte formation at C4–C5 (B). Preoperative sagittal cervical magnetic resonance image showing degeneration and herniation of the C4–C5 intervertebral disk compressing the dura (C). Lateral cervical radiograph in flexion (D), neutral position (E), and extension (F) 3 years postoperatively showing the prosthesis maintained in an ideal position and motion, without loosening or subsiding.
Cervical motion-preserving surgery with artificial disk replacement has emerged as an alternative to spinal fusion for the treatment of selected cases of cervical radiculopathy and myelopathy. A recent Cochrane systematic review by Boselie et al1
compared the 2 surgeries and found significant differences in alleviation of arm pain, neck-related function, neurological outcome, and segmental mobility at 1 to 2 years, with arthroplasty yielding better outcomes. However, differences in effect size were invariably small and not clinically relevant due to the small number of studies, publishing bias, and non-blinded designs. In addition, the primary purported advantage of arthroplasty over fusion, protection of adjacent levels, was not found at 1- to 2-year follow-up. Other systematic reviews and meta-analyses had similar findings.2–4,7,15
In a meta-analysis by Yin et al,2
cervical disk arthroplasty appeared to provide better function, a lower incidence of reoperation related to index surgery at 1 to 5 years, and lower major complication rates compared with fusion. However, cervical disk arthroplasty did not reduce the reoperation rate attributable to adjacent segment degeneration. In a study by Zechmeister et al,15
statistically significant noninferiority of cervical arthroplasty vs fusion was demonstrated at 2-year follow-up. Both technologies showed similar pain alleviation, improvement of disability or quality of life, and complication rates. Fallah et al7
found no strong evidence to support the routine use of cervical arthroplasty over fusion in single-level cervical spondylosis. Due to the lack of evidence supporting the theoretical advantages of cervical arthroplasty in decreasing the incidence of adjacent segment disease, the debate between cervical motion-preserving and fusion surgery will likely continue until long-term (5 years or more) follow-up studies reveal any longer-term differences in outcomes between the 2 surgeries.1,7,16–19
Comparison of cervical motion-preserving and fusion surgery is made more complicated by the variety of disk prostheses and instruments used for cervical motion-preserving surgery, which may impact its clinical outcomes. Some studies have explored the outcomes associated with different varieties of artificial disk prostheses. Sekhon et al20
reported that devices containing nontitanium metals (eg, PCM [NuVasive, Inc, San Diego, California] and ProDisc-C [Synthes Spine, Inc, West Chester, Pennsylvania]) prevent accurate postoperative assessment with MRI at the surgical and adjacent levels, whereas titanium devices (eg, Bryan Cervical Disc and Prestige LP [Medtronic Inc, Minneapolis, Minnesota]) allow for satisfactory monitoring of the adjacent and surgical levels. In finite element modeling studies by Lin et al,10
Bryan Cervical Disc, Prestige LP, and ProDisc-C prostheses had differences in interface stress and load transfer pattern, which may explain the mechanism of subsidence and affect segmental motion. Another finite element modeling study by Kang et al11
showed that in unconstrained and semiconstrained prostheses with different core rigidities, the shared loads at the joints differ, and greater flexibility may result in greater joint loads. Kowalczyk et al’s21
kinematic analysis of 3 different cervical devices revealed adequately maintained range of motion (ROM) at the surgical level but different motion models. Park et al’s22
kinematic analysis of the cervical spine according to prosthesis design revealed that devices with an unconstrained design may not be as beneficial to adjacent segment kinematics as are semi-constrained prostheses.
Theoretically, the different mechanical and kinematic characteristics of different disk designs may lead to different effects on clinical outcome. More detailed information about these variations is needed to fully validate cervical motion-preserving surgery.
In the current study, both the group treated with the Bryan Cervical Disc System and the group treated with the ACCEL system achieved satisfactory clinical outcomes, including significant pain relief, cervical function improvement, and quality of life amelioration, consistent with the results of previous studies.12,14,23,24
The 2 implantation systems had the same clinical effectiveness over the 2-year follow-up period. The critical factor for relieving symptoms in the short term after replacement was whether complete decompression had been achieved. However, precise milling, proper prosthesis size, and correct implantation position may impact patients’ outcomes and prosthesis functioning in the long term. In the current study, 1 experienced surgeon performed all of the spinal decompressions and prosthesis implantations, which ensured the 2 groups experienced comparable symptom relief postoperatively.
When implanting the Bryan Cervical Disc, the vertebrae endplate must be burred to a concavity with the guidance of the bone-milling rig, which directly determines the location of the prosthesis. Therefore, it is critical to secure the milling rig in the proper location and direction. An ideal milling rig is located on the middle sagittal plane of the vertebrae, pointing toward the middle cross-section of the target space.
The Bryan Cervical Disc System is a sophisticated implantation instrument. Wang et al25
summarized its procedure as 6 major steps: (1) precise implant selection, patient positioning, and orientation; (2) exposure and initial recipient site preparation; (3) neural decompression; (4) assembly of the gravity-guided bone-milling rig; (5) precise machining of the adjoining endplates to ensure a proper bone-implant interface; and (6) prosthesis implantation. The instrument consists of a 3-dimensional system based on gravity, which includes the leveled retractor frame accommodated on the table, the gravity-guided caliper device, the milling rig, and the electrically driven burr. This meticulous mechanical nature ensures quality and repeatability. However, the training procedure is time-consuming due to the need to learn to control errors in complicated procedures.
The ACCEL system omits the steps that establish the leveled reference plane, measure the angle of the preoperative intervertebral space, and accommodate a retractor frame. Rather, the setup consists of a middle-line caliper device, a sagittal wedge, and a gradienter that effectively positions the milling rig, which is directly stabilized on the adjacent vertebrae bodies with 2 single screws. Milling and implantation are similar to that with the Bryan Cervical Disc System. The ACCEL system simplifies the complex positioning process. However, it brings greater requirements and challenges for the surgeon in terms of local exposure and accurate positioning.
For single-level replacement, the ACCEL system took an average of 27 minutes less operative time than did the Bryan Cervical Disc System and had half the blood loss. Reduced operative time is generally associated with less blood loss. However, the effect of study power should be taken into consideration, especially in this nonrandomized, single-surgeon intervention study.
Three patients in the Bryan Cervical Disc System group experienced temporary throat discomfort postoperatively, whereas none did in the ACCEL group. The Bryan Cervical Disc System required a wider local field exposure and rigid retraction fixed on the frame. As a result, the soft tissue was subjected to a larger and longer stretch, which may have been the cause of the throat discomfort, whereas the ACCEL system required less operative time and had less impact on the soft tissue.
The concavities created by the milling endplates in the adjacent vertebral bodies must precisely match the titanium alloy shells with a polyurethane nucleus, which is critical for the immediate stability of the bone-implant interface. Ingrowth of bone on the interface leads to long-term implantation stability. Aggressively milling the endplates could potentially result in implant subsidence. However, inadequate milling may lead to early migration. In the ACCEL system group, 2 prostheses migrated forward slightly soon after implantation, but no migration occurred in the Bryan Cervical Disc System group. A possible explanation is insufficient stability of the ACCEL system milling rig, which is fixed on the vertebral bodies by 2 single screws. In contrast, the milling rig of the Bryan Cervical Disc System is firmly secured by 4 rods to the retractor frame. However, there is no clinical or experimental evidence to support this hypothesis. If it is correct, bicortical screw fixation could increase the stability of the ACCEL system milling rig, but iatrogenic spinal cord injury should be carefully avoided. In the current study, the 2 patients who experienced migration did not have any physical discomfort. Once migration was confirmed, soft collar protective immobility for 4 weeks was prescribed, and the prostheses showed no signs of further migration.
In both groups, all implants moved well from cervical fixation to extension. In long-term follow-up, there was no evidence of cervical instability, prosthesis loosening, subsidence, or loss of function. Because the same prosthesis was implanted in both groups, comparison of the replaced disk ROM revealed no differences between the 2 groups. However, it is interesting that patients in the Bryan Cervical Disc System group recovered cervical ROM earlier than did those in the ACCEL group. It is unclear whether different instruments truly influence the duration of cervical ROM recovery or whether this was a misleading statistical finding resulting from the limited number of cases.
In terms of complications, the Bryan Cervical Disc System mainly caused throat discomfort, whereas the ACCEL system had a higher risk of early loosening of the controllable prosthesis. The 2 implantation instruments had equivalent overall complication incidence rates. The authors recommend that the choice between the 2 be made based on the surgeon’s technique and experience.