Radiographical adjacent segment pathology (RASP) refers to the progressive degenerative changes that may occur at the adjacent segment after spinal surgery. If RASP becomes symptomatic, further treatment including surgery may become necessary. One theory to explain the phenomenon of RASP is that the elimination of motion at the index level after fusion leads to increased compensatory motion at the adjacent segments. These alterations in adjacent segment kinematics and biomechanics may lead to more rapid degeneration than would occur without fusion. Biomechanical studies and finite element models have confirmed these changes, showing that fusion results in an increase in adjacent segment range of motion (ROM), altered centers of rotation (COR), increased bending moment, increased intradiscal pressures, and increased energy absorption of impulse loads and vibrational stresses.1–3 Furthermore, in vitro studies4,5 have shown that total disc replacement (TDR) may help to avoid these abnormal effects. However, whether there are kinematic differences at the adjacent segments after TDR versus fusion surgery in patients has not been firmly established. Clinical studies comparing arthroplasty and fusion offer a unique opportunity to evaluate the changes in kinematics after TDR compared with fusion because these changes may serve as surrogates to demonstrate the possibility that clinical ASP (CASP) may be minimized by TDR.
The purpose of this systematic review was to examine the following key questions regarding the kinematics of the adjacent levels and the whole cervical spine using data from studies that compared cervical-TDR (C-TDR) with anterior cervical discectomy and fusion (ACDF):
- What are the differences in ROM, COR, and sagittal alignment at adjacent segments after C-TDR versus ACDF?
- What are the differences in ROM and sagittal alignment for the entire cervical spine after C-TDR versus ACDF?
MATERIALS AND METHODS
Electronic Literature Search
A systematic search was conducted for articles published anytime through February 2012 using terms related to C-TDR, ACDF, and specific Medical Subject Headings (MeSH) terms related to kinematics (“kinematics,” “motion,” “mobility,” “alignment,” “angular,” “translation,” or “biomechanics”). Articles were reviewed at the title/abstract level and many were excluded for the following reasons: the study did not compare C-TDR with ACDF, kinematics did not seem to be evaluated, or the study was not a comparative clinical study. We evaluated the remaining articles at the full-text level to determine whether the adjacent segment(s) were evaluated for ROM, COR, or alignment; or whether the ROM or alignment of the global cervical spine (i.e., C2–C6 or C2–C7) was evaluated. We included comparative studies that evaluated adult cervical patients with degenerative disc disease undergoing C-TDR versus ACDF. Articles were excluded if patients were younger than 18 years or treated for tumor, trauma, or infection. Other exclusions included posterior fusions, case reports, or studies with less than 10 subjects, as described in Table 1.
From the included articles, the following data were extracted: patient demographics, study interventions, inclusion and exclusion criteria, follow-up duration and rate, and details of the radiography and the method by which motion was measured. Specific outcomes of interest were abstracted, including angular ROM in flexion-extension, anteroposterior translation, global cervical motion from C2-C6 or C2-C7, changes in COR, alignment of adjacent segments, and global sagittal alignment from C2-C6 or C2-C7 (Figure 1A–C).
In cases where outcomes were measured using different methodologies between studies, we did not pool outcomes and instead performed analyses on a study level. For outcomes that were evaluated by similar methodologies, we pooled results. We employed a random-effects model to pool all relevant comparative studies that evaluated 1-level C-TDR versus ACDF at up to 24 months of follow-up. The decision to combine studies for our meta-analysis was systematic in nature. We first considered clinical heterogeneity. As all studies included patients with similar demographics and disease processes, we felt justified in combining data. Although there was some heterogeneity in the magnitude of change across the 12- and 24-month time points, the directions of effects were consistent and we decided to combine data from these different time points. Next, we considered methodological heterogeneity. Because there were no large differences in magnitude or effect between randomized controlled studies (RCTs) and cohort studies, and pooling these 2 different study types did not change the results, we combined outcomes from RCTs with those from cohort studies. Finally, we considered statistical heterogeneity. We thought it was important to combine the results despite the statistical heterogeneity across studies to determine the overall effect of arthroplasty versus arthrodesis because the effects varied across studies.
We calculated the standardized mean difference (SMD) for each study by dividing the difference in the change score (from baseline to last follow-up) between treatment groups (i.e., ACDF−C-TDR) by the mean SD of the treatment group follow-up scores. Because the standard deviation (SD) of the follow-up scores was different between the treatment groups, we used the mean of the SD of the follow-up scores for each treatment group. In cases where the study did not report the SD of the follow-up scores, we imputed the SD. The imputed value was equal to the mean of the SDs of the other studies (for that outcome and that study type). The data used to calculate the SMD are available in supplemental Table 7 (see Supplemental Digital Content 1, available at http://links.lww.com/BRS/A698). A positive SMD value indicates that the change in the reported kinematics measurement is greater in the ACDF group; conversely, a negative SMD value indicates that it is greater in the fusion group. After calculating the SMDs for each study, the effect sizes were pooled together and weighted by their inverse variance using RevMan (Version 5.1. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2011). Using this approach gives greater importance to studies that have less deviation. A random-effects model was used to correct for a consistently high I2 measure of heterogeneity across studies. A significance level of P < 0.05 was selected and confidence intervals (CIs) were calculated at a 95% level.
Study Quality and Overall Strength of Body of Literature
Level-of-evidence (LoE) ratings were assigned to each article independently by 2 reviewers (RH, AR) using criteria set by the Journal of Bone and Joint Surgery, American Volume6 for therapeutic studies and modified to delineate criteria associated with methodological quality and described elsewhere7 (see Web Appendix, Supplemental Digital Content 1, available at http://links.lww.com/BRS/A698 for individual study ratings). The quality of the overall body of evidence with respect to each outcome was determined on the basis of precepts outlined by the Grades of Recommendation Assessment, Development and Evaluation Working Group8 and recommendations made by the Agency for Healthcare Research and Quality.9 This system derives a strength of evidence (SoE) grade for each outcome or clinical question of “High,” “Moderate,” “Low,” or “Insufficient” and is described in further detail in the methods article for this Focus Issue.7
Clinical Recommendations and Consensus Statements
Clinical recommendations or consensus statements were made through a modified Delphi approach by applying the Grades of Recommendation Assessment, Development, and Evaluation/Agency for Healthcare Research and Quality criteria that imparts a deliberate separation between the SoE (i.e., high, moderate, low, or insufficient) from the strength of the recommendation. When appropriate, recommendations or statements “for” or “against” were given “strong” or “weak” designations based on the quality of the evidence, the balance of benefits/harms, and values and patient preferences. In some instances, costs may have been considered. A more thorough description of this process can be found in the focus issue methods article.7
We identified 196 total citations from our literature search (Figure 2). Of these, 150 were excluded by the title/abstract and 46 full-text articles were evaluated to determine if they met the inclusion criteria, details in Web Appendix (see Supplemental Digital Content 1, available at http://links.lww.com/BRS/A698). From these 46 studies, 34 were excluded after full-text review; details of the excluded articles and reasons for exclusion can also be found in the supplemental digital content. The 12 included studies consist of 7 RCTs10–16 (of 5 different trials), 1 prospective cohort study,17 3 retrospective cohort studies,18–20 and 1 case-control study.21 All but 1 of the studies evaluated patients with single-level disc disease; this prospective cohort study17 enrolled patients with 1- or 2-level disc disease and stratified outcomes based on the number of operated levels. The studies utilized a variety of disc prostheses. In general, patients had symptomatic cervical disc disease with radiculopathy and/or myelopathy, although diagnoses varied slightly by study. All RCTs were graded LoE II and the cohort and case-control studies LoE III. Details of the included studies are available in Table 2 and in Web (see Supplemental Digital Content 1, available at http://links.lww.com/BRS/A698).
Adjacent Segment ROM
Five RCTs,11–15 4 cohort studies,17–20 and 1 case-control study21 reported angular ROM at the caudal and/or cranial adjacent segments after C-TDR versus ACDF, Web Appendix Table 2 (see Supplemental Digital Content 1, available at http://links.lww.com/BRS/A698). Unless otherwise noted, ROM was calculated using flexion-extension radiographs. In all but 1 retrospective cohort study,19 measurements were taken with software (Quantitative Motion Analysis [Medical Metrics, Inc, Houston, TX]; Infinitt PiviewSTAR 5051; or picture archiving and communication system). The study by Nabhan et al12 used radiostereometric analysis. The length of follow-up ranged from 12 to a mean of 25 months.
Cranial Adjacent Segment Angular Motion
Meta-analysis of 3 RCTs,11,13,15 3 cohort studies,17–19 and 1 case-control study21 showed no statistically significant differences in the change in cranial adjacent segment ROM between the arthroplasty versus fusion treatment groups compared with baseline values, with an SMD of 0.23° (95% CI, −0.10° to 0.56°) (Figure 3A). Data from the majority of the individual studies similarly showed no statistically significant difference between the treatment groups; however, data from 1 RCT11 and 1 cohort study19 suggested a higher cranial adjacent segment ROM after ACDF, whereas data from 1 case-control study21 showed the opposite. There was a high level of statistical heterogeneity between the RCTs (I2 = 76%). An additional RCT reported by Nabhan et al12 reported that the cranial adjacent segment motion from neutral (rather than flexion, as in the other studies, thus this study could not be pooled with the rest) to extension after surgery was similar in both the arthroplasty and ACDF groups 1 week (15.5° vs. 15.95°, respectively; P > 0.05) and 12 months (14.5° vs. 18.1°, respectively; P > 0.05) after surgery (baseline data were not reported).12
Caudal Adjacent Segment Angular Motion
At the caudal adjacent segment, ROM was reported by 3 RCTs11,15,19 and 3 cohort studies17–19 (Figure 3B). Data from the 3 RCTs provide inconsistent results, with 1 RCT reporting a modest increase in ROM at the caudal adjacent segment that was similar in both treatment groups;13 another reporting significantly higher ROM in the arthroplasty group;11 and yet another reporting higher ROM in the ACDF group.15 Similarly, inconsistent directions of effects were reported by the cohort studies. The SMD was 0.04° (95% CI, −0.32° to 0.39°) (I2 = 80%), which indicates that there is no statistically significant difference in change in ROM at the caudal adjacent segment after arthroplasty versus fusion as measured compared with baseline values after surgery (Figure 3B).
Kim et al17 reported that the change in ROM at the caudal adjacent segment was statistically higher in the arthroplasty groups after both single- (0.9° vs. −0.6°, respectively; P < 0.005) and 2-level (1.1° vs. 0.7°, respectively; P < 0.0001) procedures; however, the differences between treatment groups are in reality relatively small.
Two additional small studies reported adjacent segment ROM but did not specify whether it was the cranial or caudal adjacent. Peng-Fei14 found that at 12 months after surgery, the ROM of the adjacent segment remained relatively constant in both the arthroplasty and fusion groups, with a change in ROM of −1.6° in the arthroplasty group and −0.5 in the ACDF group. Yi et al20 reported a small and likely clinically insignificant difference in the change of ROM at the adjacent segment between the arthroplasty and ACDF groups (0.3° vs. 1.7°, respectively).
Adjacent Segment Translation
Two RCTs15,19 reported on adjacent segment translation; because of differences in how the outcomes were reported, we did not pool the results.
Cranial Adjacent Segment Translation
Park et al13 reported little difference in anterior translation at the cranial adjacent segment between arthroplasty and fusion groups at baseline (1.4 ± 0.8 mm vs. 1.3 ± 0.8 mm, respectively) and 12 months (1.5 ± 0.8 mm vs. 1.5 ± 0.9 mm, respectively). Furthermore, there were no statistically significant within-group changes (P > 0.05). Powell et al15 similarly reported that similar translation values (which they reported as percentage of the vertebral endplate width during flexion/extension) were not statistically different between treatment groups preoperatively (∼8.2% vs. ∼8.3%, respectively) and at 24 months (∼7.8% vs. ∼10.5%, respectively) (P > 0.2 for both). In a retrospective study, Rabin et al21 also found no statistically significant difference in translation at the cranial adjacent level between treatment groups at 24 months of follow-up (P > 0.05), but did not report the corresponding data.
Caudal Adjacent Segment Translation
As for the cranial adjacent segment, Park et al13 reported similar anterior translation values between arthroplasty and fusion groups at baseline (0.7 ± 0.5 mm vs. 0.7 ± 0.5 mm, respectively) and 12 months (0.8 ± 0.6 mm vs. 0.9 ± 0.6 mm, respectively). Furthermore, there were no significant within-group changes (P > 0.05). Powell et al15 found no statistically significant difference in translation between treatment groups at baseline (∼3.8% vs. ∼6.2% of endplate width, respectively) or 24 months (6.2% vs. ∼1.8%, respectively) (P > 0.2 for both).
Adjacent Segment COR
Two RCTs13,15 reported on changes in COR at the cranial and caudal adjacent segments after C-TDR compared with ACDF (Web Appendix Table 4; see Supplemental Digital Content 1, available at http://links.lww.com/BRS/A698). The sagittal plane coordinates of COR in the horizontal (COR-X) and vertical (COR-Y) directions were obtained by independent radiographers using software analysis. The COR was calculated as the offset (mm) from the center of the superior13 or inferior15 endplate of the vertebrae caudal to the one of interest. Negative values correlate to posterior and cephalad CORs, respectively. Because of differences in the way in which the COR was calculated, the data were not pooled.
Cranial Adjacent Segment COR
Park et al13 reported that COR-X and COR-Y at the cranial adjacent segment were similar between the arthroplasty and ACDF groups at baseline and at 12 months. The change in COR-X at the cranial adjacent segment lordosis was −0.1 mm in the arthroplasty group and −1.4 mm in the fusion group (P value not reported). Although the cranial COR-X fell on average 1.4 mm more posterior in the ACDF group at follow-up than at baseline, this difference was not statistically meaningful (P > 0.4). The cranial segment COR-Y changed minimally in both treatment groups (0.2 mm vs. 0.1 mm, respectively). The authors commented that the CORs after C-TDR and ACDF remained within normal physiologic parameters, being altered up to 1.5 mm of the preoperative location in the anterior and/or inferior position.13 Powell et al15 reported that at the cranial level, COR-X was significantly more posterior 24 months after C-TDR versus ACDF (−1.0 mm vs. −2.6 mm, respectively; P < 0.01); however, there were similar differences at baseline that may not have been controlled for (−1.4 mm vs. −2.4 mm, respectively). The change in COR-X was similar in both treatment groups (0.4 mm vs. −0.1 mm, respectively; P value not reported). COR-Y was similar in both treatment groups at 24 months of follow-up (−7.8 mm vs. −8.6 mm, respectively; P > 0.1). There was little difference between treatment groups in the change in COR-Y from baseline (0.3 mm vs. 0.8 mm, respectively). There were no significant within-group differences (P > 0.1 in all cases).15
Caudal Adjacent Segment COR
Park et al13 reported that the COR-X and COR-Y at the caudal adjacent segment were each similar between the arthroplasty and ACDF groups at baseline and at 12 months. Differences between treatment groups in the changes from baseline were small for both COR-X (0.0 mm vs. 0.1 mm, respectively) and COR-Y (0.4 mm vs. 0.3 mm, respectively) and unlikely to be clinically meaningful. There were no statistically significant within-group differences between baseline and 12 months (P > 0.1 in all cases).13 Powell et al15 reported similar COR-X values 24 months after arthroplasty and ACDF (−2.1 mm vs. −3.2 mm; P > 0.3, respectively). There were no clinically meaningful differences in COR-Y between the 2 treatment groups at 24 months (−13.2 mm vs. −13.4 mm, respectively; P > 0.1). Similarly, there was no clinically meaningful difference between the treatment groups in the change in COR-X and COR-Y from baseline (COR-X: 0.5 mm vs. −0.4 mm, respectively) (COR-Y: −0.2 mm vs. −0.8 mm, respectively).15
Adjacent Segment Sagittal Alignment/Lordosis
Two RCTs10,13 were identified that reported adjacent segment sagittal alignment at 12 to 24 months after surgery (Web Appendix Table 5; see Supplemental Digital Content 1, available at http://links.lww.com/BRS/A698). Sagittal alignment was determined using neutral lateral cervical spine radiographs. A positive disc angle value indicates lordosis. The SMD was −0.14° (95% CI, −0.30 to 0.02) at cranial and −0.23° (95% CI, −0.40 to −0.06) at caudal adjacent segments and was thus more lordotic after arthroplasty compared with ACDF (Figure 4A, B). Minimal statistical heterogeneity was present between studies (cranial, I2 = 0%; caudal, I2 = 10%).
Global Cervical Spine (C2–C7) Angular ROM
One RCT11 and 2 retrospective cohort studies reported global ROM for the cervical spine at C2–C7 at a mean range of 14 to 25 months (Web Appendix Table 3; see Supplemental Digital Content 1, available at http://links.lww.com/BRS/A698). The SMD was −0.39° (95% CI, −0.63° to −0.14°) favoring arthroplasty, and this effect was statistically significant (Figure 5A). There was little statistical heterogeneity between the studies (I2 = 6%).
Cervical Sagittal Alignment/Lordosis
Two RCTs10,16 and 2 retrospective cohort studies18,19 reported the sagittal alignment of the entire cervical spine, which was measured between C2 and C7 in all but 1 study10 (Web Appendix Table 6; see Supplemental Digital Content 1, available at http://links.lww.com/BRS/A698). The pooled data suggest that the change in global cervical sagittal alignment from baseline to follow-up (range, mean of 6–25 mo) is similar between treatment groups (SMD, −0.15°; 95% CI, −0.42° to 0.12°) (I2 = 22%) (Figure 5B).
The overall SoE evaluating the changes in cranial and caudal adjacent segment ROM, translation, COR, cranial adjacent segment sagittal alignment; and global cervical sagittal alignment is “moderate,” that is, we have moderate confidence that the evidence reflects the true effect (Table 3). Further research may change our confidence in the estimate of effect and may change the estimate. The overall SoE evaluating the changes in caudal adjacent segment sagittal alignment between baseline and follow-up after C-TDR compared with ACDF is “high,” that is, we have high confidence that the evidence reflects the true effect. Further research is unlikely to change our confidence in the estimate of effect. The overall SoE evaluating the changes in global cervical ROM after C-TDR compared with ACDF is “low,” that is, we have low confidence that the evidence reflects the true effect. Further research is likely to change the confidence in the estimate of effect and likely to change the estimate.
Kinematic changes at the adjacent segment are presumed to be potentially harmful and may lead to RASP or CASP. The purpose of this systematic review was to compare changes that occur at the adjacent segments after C-TDR with those that occur after fusion. Although the studies varied in length of follow-up from 6 to a mean of 25 months, they were in many regards homogeneous and employed narrow inclusion and exclusion criteria. All but 1 study included only patients with single-level disease, and the demographics and baseline disability were similar between groups across studies. The method of measurement was, with few exceptions, similar and utilized accurate and validated techniques. However, there was statistical heterogeneity across studies for some outcomes as demonstrated by large I2; therefore, a random-effects model was used for the meta-analyses.
Adjacent segment ROM in flexion-extension increased up to 4° in both the arthroplasty and fusion groups. This relatively small change may not actually represent structural kinematic changes but instead may result from the pain elimination and functional improvement after surgery. Our meta-analysis showed that changes in angular ROM at the adjacent segments between groups did not reach statistical significance. Similarly, pooled changes in anteroposterior and mediolateral translational motion along each axis were not statistically different between treatment groups. Thus, at relatively short-term follow-up, the ROM at the adjacent segments did not vary between fusion and arthroplasty groups. Longer follow-up is needed to determine if these results will persist.
Regarding global spinal ROM (C2–C7), pooled data from 1 RCT and 2 small cohort studies demonstrated a statistically greater change in the ROM from baseline to 2 years after C-TDR versus ACDF. These results are not unexpected due to maintained motion at the index level and the improvement of overall function as a result of relief of pain and disability from either surgery.
Although ROM is an important measurement of fusion and disc arthroplasty, the quality of the motion may be even more critical when trying to predict early degeneration of a motion segment. The COR is one of the essential components of the quality of motion. Alterations in CORs at adjacent segments after surgery will change facet loading and tensile forces in the disc annulus and may lead to degeneration. In general, there tends to be variation in published COR results. This variation likely stems from multiple factors including differences in position of COR with respect to cervical level, and a lack of standardized methodology to reference the COR. The 2 studies that reported COR at the adjacent segment in this systematic review used different reference points, making it difficult to interpret the data. However, there was no statistically significant difference between treatment groups in the change in adjacent segment horizontal or vertical CORs from baseline to up to 2 years of follow-up in either study.
Alignment in the sagittal plane at the adjacent segments and in the overall cervical spine (C2–C7) was examined. We found that the change in alignment toward lordosis at both the cranial and caudal adjacent segments from baseline was statistically greater after arthroplasty compared with fusion. In contrast, there was no statistically significant difference between treatment groups in the change in overall alignment from baseline between groups, although there was a trend toward greater lordosis after arthroplasty.
The spine kinematic measurements were obtained using the most accurate methods currently available. Quantitative Motion Analysis software (Medical Metrics, Inc), which was used in the majority of the RCTs, has a mean accuracy of 0.5° and 0.3 mm in translation.22 The accuracy of the radiostereometric analysis method, which was used by Nabhan et al,12 has also been validated, with mean errors less than 0.55° and 0.15 mm.23 These methods are far more reliable than using manual placement of landmarks and lines where the operator must subjectively identify similar landmarks in 2 different images, even if these are placed using software.24 Assuming that measurement errors are randomly distributed, the effect of measurement errors on the ability to detect a difference between treatments becomes dependent on the sample size. Variability between patients and time points also affects the ability to detect differences between treatments. In addition, it is known that there are significant differences in kinematics between the different intervertebral levels. Most studies pool all levels together. Considering the scenario where all patients were treated at the C6-C7 level, and using the SDs reported for asymptomatic volunteers (4.24° at C5–C6), 70 patients per group would be required to detect a 2° difference in the average intervertebral rotation at C5–C6.25 Most of the individual studies reviewed do not have sufficient statistical power to state that there is no statistically significant difference in the outcomes between C-TDR and ACDF patients. Calculating the change on a per-patient basis is an effective way to improve the ability to detect differences, but this is done only in some studies.
The results of the present study were contrary to those reported in biomechanical studies, which have shown that the significant kinematic changes at segments adjacent to fusion are normalized by disc arthroplasty.2–5,26 However, biomechanical studies force movement into arcs that may not be physiological and often do not include the upper cervical spine, which may significantly alter motion in the treated areas. Furthermore, active processes such as muscular contraction and inhibitions from pain are not modeled in biomechanical studies. Finally, the effect of cyclical loading is largely ignored in biomechanical studies. As with all in vitro studies, caution should be used when trying to predict clinical behavior. The lack of correlation between in vivo and in vitro performance as we have shown has implications toward adoption of new technology based on presumed biomechanical advantages.
This study has limitations as the length of follow-up was relatively short (generally 12–24 mo), which minimizes the effects of structural changes that are likely to occur over time. Furthermore, any observed change may be caused or mitigated by improvements as a result of pain relief or improved functional state. We did not consider the effect of implant design, which could affect some of the variables due to the small number of relevant studies.
Future directions of research on kinematic changes after surgery should focus on reporting outcomes after a longer follow-up to see if the minimal changes between arthroplasty and fusion that were observed in this study are maintained. Furthermore, the studies comparing the effect of design differences of varying arthroplasty prostheses are needed to determine whether the mechanical differences have a long-term influence on spinal kinematics.
This study demonstrates that at short-term follow-up, there are minimal differences in the change from baseline to follow-up of in vivo kinematics between ACDF and C-TDR. The hypothesis that changes in kinematics after fusion but not after C-TDR will lead to RASP is not supported by this study.
Patients can be advised that single-level arthroplasty and ACDF result in clinically similar kinematic changes at short-term follow-up.
Strength of Statement: Strong
- Cervical fusion has been associated with kinematic changes at adjacent segments, which may lead to ASP.
- In biomechanical studies, cervical arthroplasty mitigates adverse kinematic changes; the hope is that this change will prevent degeneration and development of clinical symptoms at the adjacent segments.
- This systematic review identified 7 RCTs, 4 cohort studies, and 1 case-control study that reported adjacent segment or global kinematic changes after cervical arthroplasty compared with ACDF.
- No statistically significant differences were found between treatment groups in the changes in adjacent-segment angular ROM in flexion-extension, translation in coronal and sagittal planes, and COR as measured at up to 2 years after surgery.
- There was moderate to high evidence that the adjacent segments became significantly more lordotic after arthroplasty compared with ACDF as measured at 1 to 2 years after surgery.
- Further studies are needed to determine how the kinematics of the adjacent segments is affected in the long term after cervical arthroplasty versus fusion.
The author P.A. contributed to study design, data analysis, writing of Introduction, Discussion, and Conclusions, and editing; R.H. literature search, evaluation of studies for inclusion, data analysis, writing of Methods and Results, evaluation of SoE, and editing; A.R. evaluation of studies for inclusion, data abstraction, evaluation of LoE; and D.N. data analysis.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.spinejournal.org).
1. Dahl MC, Rouleau JP, Papadopoulos S, et al. Dynamic characteristics of the intact, fused, and prosthetic-replaced cervical disk. JBiomech Eng 2006;128:809–14.
2. Eck JC, Humphreys SC, Lim TH, et al. Biomechanical study on the effect of cervical spine fusion on adjacent-level intradiscal pressure and segmental motion. Spine 2002;27:2431–4.
3. Fuller DA, Kirkpatrick JS, Emery SE, et al. A kinematic study of the cervical spine before and after segmental arthrodesis. Spine 1998;23:1649–56.
4. Chang UK, Kim DH, Lee MC, et al. Changes in adjacent-level disc pressure and facet joint force after cervical arthroplasty compared with cervical discectomy and fusion. J Neurosurg Spine 2007;7:33–9.
5. Dmitriev AE, Cunningham BW, Hu N, et al. Adjacent level intradiscal pressure and segmental kinematics following a cervical total disc arthroplasty: an in vitro
human cadaveric model. Spine 2005;30:1165–72.
6. Wright JG, Swiontkowski MF, Heckman JD. Introducing levels of evidence to the journal. J Bone Joint Surg Am 2003;85-A:1–3.
7. Norvell DC, Dettori JR, Skelly AC, et al. Methodology for the systematic reviews on adjacent segment pathology. Spine 2012;37:S10–7.
8. Atkins D, Best D, Briss PA, et al. Grading quality of evidence and strength of recommendations. BMJ 2004;328:1490.
9. West S, King V, Carey TS, et al. Systems to Rate the Strength of Scientific Evidence. Evidence Report/Technology Assessment No. 47. Prepared by the Research Triangle Institute-University of North Carolina Evidence-Based Practice Center, Contract No. 290-97-0011. Rockville, MD: Agency for Healthcare Research and Quality; 2002.
10. Anakwenze OA, Auerbach JD, Milby AH, et al. Sagittal cervical alignment after cervical disc arthroplasty and anterior cervical discectomy and fusion: results of a prospective, randomized, controlled trial. Spine 2009;34:2001–7.
11. Auerbach JD, Anakwenze OA, Milby AH, et al. Segmental contribution toward total cervical range of motion: a comparison of cervical disc arthroplasty and fusion. Spine 2011;36:E1593–9.
12. Nabhan A, Ishak B, Steudel WI, et al. Assessment of adjacent-segment mobility after cervical disc replacement versus
fusion: RCT with 1 year's results. Eur Spine J 2011;20:934–41.
13. Park DK, Lin EL, Phillips FM. Index and adjacent level kinematics after cervical disc replacement and anterior fusion: in vivo
quantitative radiographic analysis. Spine 2011;36:721–30.
14. Peng-Fei S, Yu-Hua J. Cervical disc prosthesis replacement and interbody fusion: a comparative study. Int Orthop 2008;32:103–6.
15. Powell JW, Sasso RC, Metcalf NH, et al. Quality of spinal motion with cervical disk arthroplasty: computer-aided radiographic analysis. J Spinal Disord Tech 2010;23:89–95.
16. Sasso RC, Metcalf NH, Hipp JA, et al. Sagittal alignment after Bryan cervical arthroplasty. Spine 2011;36:991–6.
17. Kim SW, Limson MA, Kim SB, et al. Comparison of radiographic changes after ACDF versus
Bryan disc arthroplasty in single and bi-level cases. Eur Spine J 2009;18:218–31.
18. Ahn PG, Kim KN, Moon SW, et al. Changes in cervical range of motion and sagittal alignment in early and late phases after total disc replacement: radiographic follow-up exceeding 2 years. J Neurosurg Spine 2009;11:688–95.
19. Park JH, Roh KH, Cho JY, et al. Comparative analysis of cervical arthroplasty using mobi-c(r) and anterior cervical discectomy and fusion using the solis(r) -cage. J Korean Neurosurg Soc 2008;44:217–21.
20. Yi S, Lim JH, Choi KS, et al. Comparison of anterior cervical foraminotomy vs arthroplasty for unilateral cervical radiculopathy. Surg Neurol 2009;71:677–80, discussion 80.
21. Rabin D, Pickett GE, Bisnaire L, et al. The kinematics of anterior cervical discectomy and fusion versus
artificial cervical disc: a pilot study. Neurosurgery 2007;61:100–4; discussion 4–5.
22. Reitman CA, Hipp JA, Nguyen L, et al. Changes in segmental intervertebral motion adjacent to cervical arthrodesis: a prospective study. Spine 2004;29:E221–6.
23. Zoega B, Karrholm J, Lind B. Plate fixation adds stability to two-level anterior fusion in the cervical spine: a randomized study using radiostereometry. Eur Spine J 1998;7:302–7.
24. Zhao K, Yang C, Zhao C, et al. Assessment of non-invasive intervertebral motion measurements in the lumbar spine. J Biomech 2005;38:1943–6.
25. Reitman CA, Mauro KM, Nguyen L, et al. Intervertebral motion between flexion and extension in asymptomatic individuals. Spine 2004;29:2832–43.
26. Cunningham BW, Hu N, Zorn CM, et al. Biomechanical comparison of single- and two-level cervical arthroplasty versus
arthrodesis: effect on adjacent-level spinal kinematics
. Spine J 2010;10:341–9.