Introduction

Anterior cervical discectomy and fusion (ACDF) is up to now the most common procedure for symptomatic degenerative cervical disc disease [5, 17, 37]. Radiographic and clinical studies have shown that with time, the disc adjacent to the fused spinal segment occasionally degenerates or becomes unstable [1, 8, 17, 18, 24, 41, 47]. However, it is still debatable, whether cervical fusion creates an unfavourable biomechanical situation at adjacent levels causing accelerated disc degeneration or the actually observed degeneration of the adjacent segment is causally related to a natural development in a predisposed person.

There are many potentially important factors associated with the development of adjacent segment disease, starting with increased segmental motion, stress, load, and intradiscal pressure at levels adjacent to the fusion site [4, 6, 26, 33, 34, 43, 45, 46].

There is also a considerable number of authors who reported an increase in adjacent segment disease following anterior cervical fusion, indeed most of their studies were retrospective [10, 11, 14, 15, 28].

Hilibrand et al. [17] reported a large retrospective series in which patients underwent ACDF, and the authors concluded that symptomatic adjacent-segment disease occurred at a relatively constant incidence of 2.9% annually (in 18% of the single-level cases symptomatic cervical disc disease developed at the adjacent level). They predicted that in 25.6% of the patients who underwent anterior cervical fusion, new symptomatic disease would occur at an adjacent segment within 10 years of the operation.

Goffin et al. [11] showed that after at least a 60-month follow-up, they had a 6.11% reoperation rate due to symptomatic adjacent level degeneration.

Therefore, adjacent-segment disease has generated considerable interest, and a new implant, the disc prosthesis, is being developed in response to concerns about this disorder. Motion preservation procedures have become increasingly popular during the last years.

The primary goal for introducing cervical disc prostheses is to protect the patients from developing symptomatic adjacent disc disease by preserving motion.

Meanwhile there are many studies, which came to the results that disc replacement preserves segmental motion [2, 12, 13, 29, 30, 32]. However, the effect of motion preservation to the adjacent level is still unclear. Therefore, the aim of this current study was to determine the segmental motion following cervical disc replacement as well as the asymptomatic upper segment and compare both segmental motions with patients who received ACDF.

Materials and methods

Study design

This is a prospective randomized controlled study, approved by the local ethical committee of Saarland (Germany) Nr. 21/06.

Twenty patients [13 men and 7 women, mean age 43 years, standard deviation (SD) 9 years] suffering from symptomatic degenerative soft disc disease with single-level radiculopathy, not responding to a trial of conservative treatment were treated between January 2006 and August 2007. All patients have had confirmatory imaging studies and were informed of the purpose of the study and given written informed consent at least 24 h before surgery. Randomization evolved by drawing cards in sealed envelopes. Ten patients received ACDF in single level with ABC titanium plate fixation (control group) and 10 patients received single-level disc replacement with ProDisc C prostheses (study group). Surgery was performed by one of the authors (A. N.). A standard left-sided anterior approach was performed; and the lateral parts of the annulus were preserved. However, the symptomatic disc and the posterior longitudinal ligament were removed. Either prosthesis or a titanium plate and a cage filled with Tribone were inserted according to the manufacturers’ recommendation. Finally, 7–9 tantalum markers of 0.8-mm diameter (RSA Biomedical, Box 7972, S-907 19 Umea, Stockholm, Sweden) were placed into the vertebrae adjacent to the treated segment, as well as 4–5 into the adjacent asymptomatic upper segment (Fig. 1) [21].

Fig. 1
figure 1

Lateral X-ray of a cervical spine showing the tantalum markers of the vertebral body C4, C5 and C6. a Incorporated tantalum markers after disc replacement. b The same with titanium plate fixation

Full size image

Clinical symptoms such as cervical and brachial pain were investigated preoperatively and 1 week, as well as 6 and 12 months postoperatively. Visual analogue scale (VAS) was used for grading brachial pain. The Neck Disability Index (NDI) was used to assess cervical neck pain.

Implants used in the study

The cage used here is named Solis (Stryker Howmedia GmbH, Gewerbeallee 18, 45478 Mülheim, Germany). It is made from polyetheretherketone (PEEK), comes in two different diameters (12 and 14 mm), and in different heights (4, 5, 6, 7, 8, and 9 mm). The cage has a large central perforation to allow bony ingrowths. Titanium anchoring spikes and bone-hugging serrations should enhance a secure fit.

The titanium alloy plate is the ABC (advanced biomechanical concept) plate (Aesculap AG, AM Aesculap-Platz, Tuttlingen, Germany) for anterior osteosynthesis of the cervical spine. It is a dynamic, titanium plate that can be bent to match the patient’s individual lordosis. The screw for monocortical fixation is a titanium, self-tapping, conical screw of 10, 12, 14, 16, and 18 mm length with an outer diameter of 4.0 mm. The single level plate is offered with a length of 20, 22, 24, 26, 28, 30, and 32 mm.

The prosthesis used here was the ProDisc C, which consists of a modular design, two metal plates, and a polyethylene inlay that is safely secured into the lower end plate. The metal end plates have a keel design for enhanced primary stability and fixation, and the end plate coverage with a coating that consists of a titanium plasma spray allows bony ingrowth and long-term fixation. The polyethylene inlay determines the height of the prosthesis (Synthes Spine, Paoli, PA).

Roentgen stereometric analysis (RSA)

Patients underwent RSA postoperatively, after 6 and 12 months. For more details about the Roentgen stereometric analysis and images, please rivet on our article published in [29].

Statistic

t-Test for paired values was used to determine a statistical difference of residual intervertebral motion within the same group (P < 0.05).

Mann–Whitney U test for unpaired values was used to determine a statistical difference of residual intervertebral motion between both groups in the three axes of motion (P < 0.05).

Results

No perioperative or postoperative complications were observed and no patients of both groups required revision of the device.

Radiological analysis

Study group (prosthesis)

ROM of treated segment

Between 1 week and 6 months postoperatively, segmental motion decreased significantly in extension, axial rotation and bending (P = 0.02, P = 0.03 and P = 0.01). However, 1 year postoperative, no further decrease in segmental motion in extension, axial rotation and bending (P > 0.05 for all three axes) was seen. Data for each axis are contained in Table 1 and Figs. 2, 3, 4.

Table 1 Rotations in degree (°) 1 week, 6 and 12 months after surgery for the treated segment
Full size table
Fig. 2
figure 2

Graph illustrating segmental motion in extension (degree) (mean value and standard deviation) for each follow-up examination

Full size image
Fig. 3
figure 3

Graph illustrating segmental motion in right-sided axial rotation (degree) (mean value and standard deviation) for each follow-up examination

Full size image
Fig. 4
figure 4

Graph illustrating segmental motion in right-sided lateral bending (degree) (mean value and standard deviation) for each follow-up examination

Full size image
ROM of adjacent segment

Segmental motion at adjacent cranial levels (Table 2; Figs. 2, 3, 4) did not exhibit a significant change between 1 week and 6 months after surgery in extension, axial rotation and bending (P = 0.3, P = 0.07 and P = 0.5). One year after surgery, there was also no significant change in segmental motion in all three directions, when compared to the 6-month results (P = 0.2, P = 0.6 and P = 0.4).

Table 2 Rotations in degree (°) 1 week, 6 and 12 months after surgery for the adjacent segment
Full size table

The segmental motion of the adjacent segment was significantly higher in extension (P = 0.001), axial rotation (P = 0.002) and bending (P < 0.0001) after 1 week in comparison to the index level. The segmental motion 6 months and 1 year after surgery was also significantly higher when compared to the values measured for the segment treated with prosthesis (P < 0.05) (Figs. 2, 3, 4).

Control group (ACDF)

ROM of treated segment

Six months postoperative, there was a statically significant decrease in segmental motion in extension, axial rotation and bending (P = 0.04, P = 0.03 and P = 0.03), when compared to the values obtained after 1 week. However, 1 year postoperative, no significant difference could be noticed when compared with the 6 months value (P > 0.05). Data for each axis are contained in Table 1 and Figs. 2, 3, 4.

ROM of adjacent segment

Segmental motion at adjacent levels (Table 2) did not show a significant difference between 1 week and 6 months after surgery in extension, axial rotation and bending (P = 0.1, P = 0.5 and P = 0.7). However, 1 year after surgery, there was a slight increasing segmental motion for all three directions, but without a significant difference (P = 0.3, P = 0.2 and P = 0.3) (Table 2; Figs. 2, 3, 4).

The segmental motion of the adjacent level was significantly higher in extension (P < 0.002), axial rotation (P < 0.001) and bending (P < 0.001) after 1 week in comparison to fused cervical segment. The segmental motion 6 months and 1 year after surgery were also significantly higher compared to the values measured for the fused segment (P < 0.05) (Figs. 2, 3, 4).

Study group (prosthesis) versus control group (ACDF)

ROM of treated segment

Segmental motion was significantly higher in the prosthesis group in comparison to fusion for extension after 1 week (P = 0.001), 6 months (P = 0.01) and 1 year after surgery (P = 0.02).

There was also a significant difference in axial rotation (P = 0,002 after 1 week, P = 0.021 after 6 months and P = 0.013 after 1 year).

Segmental motion was not significantly different between both groups for bending (P = 0.3 after 1 week, P = 0.1 after 6 months and P = 0.06 1 year after surgery) (Table 1; Figs. 2, 3, 4).

ROM of adjacent segment

The segmental motion of the adjacent level shows a slightly higher segmental motion for all three axes in the fusion group, but without significant difference for any direction (extension, axial rotation and bending) at any examination time (P > 0.05) (Table 2; Figs. 2, 3, 4).

Clinical outcomes

Study group

The NDI for neck pain decreased significantly from a mean value and SD of 0.49 (0.13) preoperatively to 0.1 (0.08) 1 year postoperative (P = 0.03). The VAS for arm pain improved significantly from preoperative mean value of 8.4 (1.1) to 1.2 (1.1) 1 year postoperatively (P = 0.01) (Fig. 5).

Fig. 5
figure 5

Visual Analogue Scale (VAS) for arm pain. Mean value and SD are given for each time. Both groups could show the same pattern of pain relief in arm to all examination times without statistically significant difference (P > 0.05)

Full size image

Control group

Mean values and SD for cervical pain measured using NDI decreased significantly from 0.47 (0.15) preoperatively to 0.16 (0.1) after 1 year (P = 0.01). Mean values and SD for arm pain using VAS decreased significantly from 8.2 (1.4) preoperatively to 1.5 (1.25) 1 year postoperative (P = 0.014) (Figs. 5, 6).

Fig. 6
figure 6

Neck Disability Index (NDI) for neck pain. Mean value and SD are given for each time. Both groups could show the same pattern of pain relief in neck to all examination times without statistically significant difference (P > 0.05)

Full size image

There was no significant difference between both groups in pain relief for neck and arm pain for each examination period (P > 0.05).

Discussion

The development of adjacent level disease engaged all spine surgeons. It is supposed that the presence of a fusion could increase load and segmental range of motion at adjacent levels and cause localized trauma with subsequent accelerated disc degeneration [25, 26, 42]. Whether a cervical fusion results in accelerated adjacent disc degeneration with clinical manifestation of radiculopathy and/or myelopathy or these are related to a natural development in a predisposed person is currently under debate [7, 9, 22, 23, 25, 26, 42, 43].

In this study, at a mean follow-up of 12 months, there was no change in the average segmental motion immediately cranial to the disc prosthesis, whereas there was an increase in average segmental motion immediately cranial to the fusion but without a significant difference (P > 0.05).

By assessing segmental motion in cervical spine, the chosen method becomes of critical importance. The evidence of the study increases clearly with the accuracy of the chosen measuring method. There are several methods to assess segmental motion. A few of these high accuracy methods are technically complex and time-consuming like RSA. Furthermore, it should be considered that we have a small patient number, so the results may change if more data from more patients are available. However, the precision of RSA is high, which makes it suitable for comparisons of small study samples and small changes. Nevertheless, our data give insight into the motion of the index and adjacent spine segment under physiological loads.

Functional X-ray with flexion and extension films is used quite frequently in clinical practice, because it is readily available.

Unfortunately, this measuring method has been shown to have poor reliability and their quantitative accuracy is limited [16, 31, 40, 44]. The video fluoroscopy is an improved technique to this relative inaccurate method. Several authors have advocated the advantages of the use of video fluoroscopy in the evaluation of intervertebral motion [3, 19, 27, 35].

The video fluoroscopy-based measurement allows continuities in intervertebral motion assessment. However, one of the major disadvantages of this method is the failure to account for the possibility that patients may have a different total range of motion at different time points.

RSA has already proven to be a highly accurate method to detect segmental motion between involved vertebrae [20, 38, 39] and allow for a three-dimensional motion analysis. This insertion of 4–6 radio-opaque titanium markers at the time of surgery in each vertebra is required to determine the geometric characteristics of the vertebral anatomy. Although the RSA method is ideal for a small patient group, it is possible that some relevant statistical differences may remain hidden. As an approximate power calculation, we analysed the corresponding power of Student’s test for two independent normally distributed samples which typically also reflects the power of the Mann–Whitney U test for important distributions. Using a two-sided test with alpha = 5%, differences of mean values of segmental motion in a specific direction between the groups larger than 0.95 times the standard deviation can be detected with a power above 90%.

As in our previously published studies, cervical spine segmental motion of the segment either treated with ProDisc C or ACDF decreases over time, but with significantly higher loss of segmental motion in the fusion group in comparison to the preoperative values [29].

Our current results demonstrated that patients with prostheses had a significantly greater segmental motion at the operative level in comparison to the ACDF group at each follow-up evaluation. When we looked to the adjacent level, we found a slightly increased segmental motion after ACDF 1 year after surgery, without a significant difference, when compared with prostheses.

In a recent study, Robertson et al. [36] prospectively compared the incidence of adjacent level radiographic disease and symptomatic disease in patients 24 months after undergoing either cervical arthroplasty with the Bryan Cervical Disc or cervical arthrodesis with a cervical cage. They reported a significantly higher incidence of X-ray changes indicative of degenerative disc disease at adjacent levels in patients with arthrodesis. The incidence of clinical symptomatic of the adjacent level disease was reported with a higher rate in patients with arthrodesis [36].

Reitman et al. [35] did not find a statistically significant change in ROM at the cephalad level adjacent to ACDF. They noticed that there were no X-ray changes indicative of degenerative disc disease at adjacent levels. Reitman et al. [35] suggested that degenerative disc disease may develop with or without alteration in cervical motion. If it is true, adjacent level disease may be the result of degeneration progression as results of altered cervical kinematics post fusion. Yet, our finding 1 year after surgery did not support the contention that altered kinematics and spondylosis are dependent entities. Therefore, kinematic changes, X-ray evidence of spondylosis, and clinical presentation may be different manifestations of the same underlying disease process with variable “expression” in a given individual.

Wigfield et al. [45, 46] prospectively compared the ROM between nonrandomized patients with ACDF and patients with prostheses. The fusion group demonstrated a higher segmental motion at the adjacent level in comparison to prostheses [45]. However, the increased ROM in patients with ACDF was distributed at spinal levels with no X-ray evidence of degenerative disc disease and not at discs with radiographic degenerative changes [45]. Although the published evidence supports a 34.6% rate of radiographic adjacent segment disease after arthrodesis versus 17.5% after arthroplasty at the 24-month follow-up evaluation [36], we did not find any significant differences in kinematics between the prostheses and ACDF groups in our relatively short follow-up period and relatively small patient number. In addition, our study did not address morphological changes to the adjacent level, which is very interesting. Therefore, we are currently investigating our earlier study, which has been started in 2004 to measure kinematics of the treated level, to evaluate morphological changes.

Both operative procedures lead to significant reduction of arm and neck pain without statistical difference between both groups. Several authors have published comparable clinical results [2, 12, 13, 36]. However, it is too early to judge whether preserving segmental motion could reduce the incidence of adjacent level degeneration and improve clinical results.

Conclusion

The results of this study demonstrate that there is no significant difference of the segmental motion of the adjacent level, either treated with prostheses or fusion, 1 year after surgery. Long-term results will be needed to determine further segmental changes in motion. Clinical results did also show no significant difference in pain relief between both groups.