Anterior cervical discectomy and fusion (ACDF) has been a standard surgical procedure for cervical disc decompression. It functions to decompress affected neural components, provide mechanical stability and lordosis, and preserve intradiscal height.²⁶ However, investigators have also reported an increase in motion, shear strain, and intradiscal pressure in adjacent vertebrae after treatment.^10,21 The displacement of motion and mechanical stress to adjacent segments is a major concern because force and motion translocation are believed to lead to increased rates of adjacent-segment degeneration in patients treated with ACDF.^{6,10,16,17,21} It is also hypothesized that adjacent-segment degeneration is further heightened in multilevel ACDF treatment.¹¹

Cervical total disc replacement (TDR) is a treatment option for symptomatic radiculopathy and myelopathy. A number of clinical trials have shown that TDR is a safe and effective alternative to ACDF for 1- and 2-level cervical decompression.^{3,5,7,11,15,19,22,27–29} As with ACDF, cervical TDR acts to decompress the affected segment, provide stabilization, and preserve intradiscal height while maintaining mobility.^3,22,28 Investigators have suggested that the preservation of mobility may result in a decreased frequency of adjacent-segment degeneration compared with that observed in patients treated with ACDF.^1,5,31

The results of multiple independent studies suggest that cervical TDR at 1 level may also result in decreased rates of subsequent operations at the treatment and adjacent levels.^{5,9,12,15,22,23,28–30} Long-term studies have shown that treatment with ACDF results in significantly higher subsequent surgery rates than cervical TDR, although few studies have analyzed 2-level subsequent surgery outcomes.^2,9,24 In long-term studies, the ACDF subsequent surgery rate was observed to be as high as 5 times the rate of TDR subsequent surgical intervention.⁹ The purpose of this study was to evaluate 5-year subsequent surgery rates at index and adjacent levels in patients treated at 1 or 2 contiguous levels with TDR or with the ACDF control procedure, as part of an FDA Investigational Device Exemption (IDE) clinical trial.

Methods

Study Interventions

The investigational device is the Mobi-C Cervical Disc Prosthesis (LDR Medical). The implant is composed of an ultra-high-molecular-weight polyethylene (UHMWPE per ISO 5834–2) mobile insert between 2 endplates (Fig. 1). The control device is ACDF, using either the Slim-Loc Anterior Cervical Plate System (DePuy Spine) or the Sofamor Danek Atlantis or Atlantis Vision Anterior Cervical Plate Systems (Medtronic) with corticocancellous allograft. Postoperative care for both groups was left to the discretion of the treating surgeon.

Mobi-C Cervical Disc Prosthesis (LDR Medical). Copyright LDR Holding Corporation. Published with permission. Figure is available in color online only.

Results

For both the 1- and 2-level arms, a total of 599 patients were treated with the investigational or control device. For the 1-level arm, 179 patients received TDR and 81 received ACDF treatment. For the 2-level arm, 234 patients received TDR and 105 received ACDF treatment. No significant differences were found between the demographic profiles of the investigational and control groups. The 60-month follow-up rate was 85.5% (TDR) and 78.9% (ACDF) for the 1-level group and 90.7% (TDR) and 86.7% (ACDF) for the 2-level group.

A subsequent surgery was considered to be any operation that occurred at the initial treatment level or at adjacent levels after the primary operation. All TDR and ACDF subsequent surgery cases are listed in Tables 3 and 4.

For the 1-level ACDF group, 14 of 81 patients underwent subsequent surgeries and 1 patient required multiple subsequent surgeries. In the 1-level TDR group, 8 of 179 patients underwent subsequent surgeries and 2 patients required multiple subsequent surgeries. The number of 1-level patients receiving subsequent surgeries was significantly higher for ACDF at 60 months (TDR 4.5% vs ACDF 17.3%; p = 0.0012).

For the 2-level ACDF group, 22 of 105 patients underwent subsequent surgeries and 3 patients required multiple subsequent surgeries. In the 2-level TDR group, 17 of 234 patients underwent subsequent surgeries and 2 patients required multiple subsequent surgeries at the 60-month follow-up point. The percentage of 2-level ACDF patients receiving subsequent surgery was significantly higher than that of 2-level TDR patients at 60 months (TDR 7.3% vs ACDF 21.0%; p = 0.0007).

Subsequent surgeries were classified by operative level as an index- and/or adjacent-level surgery (Fig. 2). For the 1-level arm at 60 months, there were a total of 8 TDR (4 index, 2 adjacent, 2 index and adjacent) and 14 ACDF (5 index, 4 adjacent, 5 index and adjacent) subsequent surgeries classified by operative level. For the 2-level arm, there were a total of 17 TDR (9 index, 6 adjacent, 2 index and adjacent) and 22 ACDF (10 index, 3 adjacent, 9 index and adjacent) subsequent surgeries classified by operative level at the 60-month follow-up.

Subsequent surgery classification by operative levels.

Surgeries Involving an Index Level

At 60 months, the rate of subsequent surgeries that involved the index level for the 1-level arm was significantly different at 3.4% (6/179) for TDR and 12.3% (10/81) for ACDF (p = 0.0097). Of the 10 1-level ACDF surgeries involving an index level, 7 ACDF surgeries were a result of index-level indications and 3 surgeries resulted from removal of the anterior plate when treating adjacent-level disease. When censoring patients undergoing plate removal due to adjacent-level indications only, the ACDF group retained a substantially higher subsequent surgery rate, although this difference lost significance (3.4% vs 8.6%; p = 0.1194). The difference in index-level subsequent surgeries was also significant in the 2-level arm, at 4.7% (11/234) for TDR and 18.1% (19/105) for ACDF. Of the 19 2-level ACDF surgeries involving an index level, 13 surgeries were due to index-level indications and 6 surgeries were due to hardware removal for adjacent-level disease. When censoring patients undergoing plate removal due to adjacent-level indications only, the ACDF group maintained a higher rate of subsequent surgeries (4.7% vs 12.4%; p = 0.0197).

Of these subsequent surgeries involving the index level at 1 level, 2.8% (5/179) TDR and 11.1% (9/81) ACDF patients (p = 0.014) failed to meet the primary end point criteria due to subsequent surgical intervention. Similarly, at 2 levels, 3.8% (9/234) TDR and 16.2% (17/105) ACDF patients (p = 0.0002) did not meet the study's primary end point criteria (Fig. 3). There was no statistically significant difference within treatment types between 1- and 2-level rates. The most prevalent reasons for subsequent surgeries at the index level for 1- and 2-level ACDF were radiculopathy, neck pain, and pseudarthrosis. Radiculopathy was the most common indication for subsequent surgery among TDR patients.

Subsequent surgery at index level leading to study failure. *p= 0.014; **p = 0.0002.

Surgeries Involving an Adjacent Level

The rate of subsequent surgeries involving an adjacent level was calculated at 60 months for both 1- and 2-level arms (Fig. 4). The involvement of adjacent level could overlap with a subsequent surgery at the index level. For the 1-level arm at 60 months, the rate of adjacent-level subsequent surgery was significantly higher for ACDF patients, at 11.1% (9/81), than for TDR patients, at 2.2% (4/179) (p = 0.0043). For the 2-level arm at 60 months, the ACDF group also demonstrated a significantly higher rate of adjacent-level subsequent surgical intervention (TDR 3.4% [8/234] vs ACDF 11.4% [12/105]; p = 0.0059). There was no statistically significant difference observed within treatment groups between 1- and 2-level treatments.

Subsequent surgery at adjacent level. *p = 0.0043; **p = 0.0059.

The most common reasons for adjacent-level ACDF surgeries were adjacent-level disease and neck pain. Radiculopathy and adjacent-level disease were the most frequent indications for subsequent surgery among TDR patients. Adjacent-level disease was an indication for surgical intervention for 8 of 9 ACDF and 4 of 4 TDR 1-level patients. For 2-level patients, adjacent-level surgery was initiated by adjacent-level disease for 11 of 12 ACDF and 5 of 8 TDR surgeries. The average time from diagnosis of adjacent-level disease to surgery was 35.8 months for ACDF patients and 32.1 months for TDR patients.

Multiple Surgeries

Several patients underwent more than 1 subsequent surgery. In the 1-level TDR group, 1 patient initially had the TDR device removed (C4–5) and replaced with an ACDF at 5 months postsurgery due to worsening radiculopathy and spondylosis, possibly due to an oversized implant. Eight months later, the patient underwent surgery again, with an ACDF at 3 adjacent levels (C3–4, C5–7) due to symptomatic adjacent-level disease. Another patient in the 1-level TDR group had the device removed 38 months postsurgery and underwent fusion at the index level (C5–6) and the inferior adjacent level (C6–7) after experiencing neck pain, radiculopathy, and adjacent-level disease following an injury. Seventeen months later, the patient underwent a revision fusion with supplemental fixation due to pseudarthrosis at C5–6 and foraminal stenosis at C6–7.

In the 1-level ACDF group, 1 patient had a subsequent surgery at 15 months postoperatively for symptomatic pseudarthrosis with radiculopathy, undergoing a posterior foraminotomy and medial facetectomy plus posterior fusion with instrumentation (C5–6). Due to worsening dysesthesia and spinal cord changes, the subject underwent posterior decompression and fusion with allograft at C3–6, as well as removal of the posterior hardware at C5–6 at 26 months postoperatively.

In the 2-level arm, 1 patient with a TDR device continued to experience neck and arm pain following the primary surgery, which was attributed to poor device stability (Fig. 5). The patient underwent removal of both prostheses and received a fusion at both index levels (C4–6) approximately 11 months postsurgery. Nine months later, the patient presented with symptomatic pseudarthrosis at both levels and underwent another anterior and posterior fusion procedure. A second patient in the TDR group had the inferior prosthesis (C5–6) removed after experiencing multiple motor vehicle accidents with concurrent neck pain 23 months postsurgery. The patient had a surgery to remove the superior prosthesis (C4–5) 7 months later at a noninvestigational site and chose not to release their medical records to the investigator.

TDR failure requiring removal and multiple subsequent fusions.

In the 2-level ACDF group, 3 patients had multiple subsequent surgeries. One patient had a subsequent surgery at 10 months postoperatively due to continuing neck and arm pain (Fig. 6). The patient underwent removal of the index-level (C4–6) hardware and underwent ACDF at an adjacent level (C3–4). A nonunion (C5–6) was detected intraoperatively, and a redo fusion takedown with decompression was performed. One year later, the subject presented with disabling neck pain and was admitted. This patient underwent removal of hardware C3–6 anteriorly; exploration of fusion with a finding of nonunion at C5–6; reinstrumentation of C3–6 with a plate and screw system; and C3–4, C4–5, and C5–6 bilateral posterior-lateral fusion with instrumentation using a posterior cervical fixation system. Seven months later, the patient returned with complaints of recurrent neck pain and underwent removal of hardware from C3–6 and exploration of the fusion, which showed solid union.

ACDF nonunion and development of adjacent-level disease leading to subsequent fusion surgeries.

A second 2-level ACDF subject, who was initially treated at C5–7, developed pseudarthrosis with neck pain 10 months postoperatively. The subject underwent a C5–6 posterior arthrodesis with lateral mass screws and local autologous graft. After the patient reported swelling in her neck 22 months later, it was revealed by MRI that the patient had a large posterior disc protrusion at C3–4 with cord indentation and bilateral foraminal narrowing. The subject underwent anterior cervical discectomy and C3–4 arthroplasty. After presenting with severe cervical and thoracic instability 11 months later, the patient underwent C5–T2 posterior fusion.

A third patient in the 2-level ACDF group treated at C3–5 underwent an additional surgery for adjacent-level disease at C5–6, 22 months after the primary surgery. The patient was treated with discectomy and TDR at C5–6 and removal of the initial hardware at C3–5. Thirty months later, the patient presented with persistent and worsening neck pain, and radiographs showed loosening of the TDR. The subject then underwent removal of the TDR, corpectomy, and anterior cervical fusion.

Discussion

The safety and effectiveness of TDR has been validated across a number of studies at many different follow-up periods.^{3,5,15,22,23,27–29,33} Overall, TDR has demonstrated an advantage over ACDF with regard to motion preservation. The nature of ACDF eliminates motion at treated levels, whereas TDR has been shown to preserve segment mobility with high success.^3,22,28 Both short- and long-term results have trended toward similar or greater improvements in NDI, neck pain, and arm pain visual analog scale scores in TDR populations when compared with ACDF, although the significance of these results remains controversial.^{5,11,12,14,22,24,27–29}

The results of many single-level TDR clinical trials suggest that TDR may also result in a lower incidence of secondary operations.^{5,7,9,12,15,22,23,28,29} Sasso et al. reported a 2-fold increase in secondary surgery rates in patients with 1-level ACDF compared with Bryan cervical disc replacement counterparts after 2 years (TDR 2/56 vs ACDF 4/59).²⁸ Garrido et al. reported a 4-fold increase in subsequent surgical intervention in patients with 1-level ACDF compared with Bryan cervical disc replacement at 4 years (TDR 1/23 vs ACDF 5/26).¹² In a 5-year study on secondary surgery rates, Delamarter and Zigler reported a significant decrease in secondary surgery rates in patients who received a ProDisc-C artificial cervical disc (2.9%) versus ACDF (14.5%).⁹

The intent of this study was to further demonstrate the benefits of TDR in terms of subsequent surgical intervention rates. In agreement with previous studies, we found that patients with 1-level ACDF who received a TDR device had a significantly lower occurrence of subsequent surgical intervention at the treated level compared with ACDF-treated patients (2.8% TDR vs 11.1% ACDF; p < 0.05). Patients with 2-level ACDF who received a TDR device also demonstrated significantly fewer index-level surgeries at 60 months (3.8% TDR vs 16.2% ACDF; p < 0.001).

Several authors have hypothesized that TDR may reduce the incidence of adjacent-segment degeneration compared with ACDF as a consequence of maintaining segmental motion and stress profiles.^16,20,25,31 For our 1-level arm at 60 months, we found that 4 times fewer TDR patients required a subsequent operation at adjacent levels (2.2% TDR vs 11.1% ACDF; p < 0.05). Similar results were shown in the 2-level arm for adjacent-level surgeries (3.4% TDR vs 11.4% ACDF; p < 0.05). These results are in agreement with the findings of other investigators and suggest an elevated rate of adjacent-segment degeneration in the ACDF population.

In a retrospective review of anterior cervical decompression and stabilization, patients with a maximum follow-up of 21 years (range 2–21 years), including patients without cervical instrumentation, Hilibrand and Robbins¹⁷ cite studies by Bohlman et al.,⁴ Gore and Sepic,¹³ and Williams et al.³² when analyzing rates of adjacent-segment disease among ACDF patients, with an average follow-up of 4.5 years.¹⁷ From these studies, the annual incidence of adjacent-segment disease requiring additional surgery was 1.5%–4%,¹⁷ equating to 7.5%–20% at 5 years. In a long-term TDR study with the Prestige artificial cervical disc, Burkus et al. showed a lower rate of secondary surgeries involving adjacent segments in their TDR population compared with ACDF controls (TDR 2.9% vs ACDF 4.9%).⁵ Mummaneni et al. also reported a statistically significant decrease in secondary operations involving adjacent segments in their TDR population (TDR 2/276 vs ACDF 9/265).²² Davis et al. reported a significantly greater rate of adjacent-segment degeneration at both the inferior and superior index levels for 2-level ACDF compared with TDR at 4 years.⁸ Interestingly, the rate of adjacent-level operations was similar between the 1- and 2-level ACDF or TDR groups, and does not reflect the expectation that multilevel ACDF causes a greater amount of adjacent-level disease than single-level ACDF. However, this study was not powered or designed for intratreatment comparisons, and these results are suggestive, not conclusive.

Limitations of this study include the inability to blind surgeons and patients to treatment, which opens the results to the potential of confirmation bias. Although the control group in this study was limited to anterior plating with allograft, other fusion procedures and devices (e.g., standalone devices and the use of autograft) are viable treatment options. The comparative results between the control and investigational groups are limited to anterior plate and allograft and may not be consistent with those of other surgical alternatives for cervical fusion. Additionally, the control group consisted of patients receiving 3 different cervical plate systems, based on surgeon preference. This heterogeneity represents a study limitation because ACDF failures may not have been equally distributed across the 3 fusion systems implanted.

All authors were investigators for the Mobi-C IDE clinical trial, which was sponsored by LDR Spine USA, Inc. Some surgeons received compensation for their participation in the trial or have equity in LDR Spine. To ensure that these potential conflicts of interest have not affected study outcomes, an analysis was performed to compare the subsequent surgery rates between sites with and without financial interests. A site was considered financially interested if an investigator received any payment from the manufacturer or if the investigator held company equity during the study period. At 60 months, the financially interested and nonfinancially interested sites had statistically similar subsequent surgery rates within treatment groups for both treatment arms, with no trend observed. Additionally, 45.9% of patients had subsequent surgeries performed by surgeons not participating in the IDE trial.

Conclusions

The results from this clinical trial suggest that TDR may provide a substantial benefit over ACDF in providing a lower risk for subsequent surgical intervention. Furthermore, a lower rate of subsequent adjacent-level surgical procedures in patients who received TDR devices provides indirect evidence that motion preservation may lead to a lower rate of adjacent-level disease than an anterior fusion approach.

Acknowledgments

LDR Spine USA, Inc. sponsored the FDA IDE clinical trial. We would like to thank the other principal investigators for their contributions to the study, who are as follows: Guy Danielson III, MD; Charlie Gordon, MD; Daniel Peterson, MD; John Stokes, MD; Arnold Schwartz, MD; Ali Araghi, MD; David Tahernia, MD; Hazem Eltahawy, MD; Reginald Tall, MD; Douglas Wong, MD; Gerald Schell, MD; Michael Ramsey, MD; B. Christoph Meyer, MD; Robert McLain, MD; Jon Park, MD; Ed Simmons, MD; Mark Stern, MD; and Phillip S. Yuan, MD.