KEY WORDS: Arthroplasty, Cervical spine, Disc prothesis, Meta-analysis

Transcription

1 RAPID COMMUNICATION Ronald H.M.A. Bartels, MD, PhD Radboud University Nijmegen Medical Centre, Department of Neurosurgery, Nijmegen, Netherlands Roland Donk, MD Canisius Wilhelmina Hospital, Department of Orthopedic Surgery, Nijmegen, Netherlands André L.M. Verbeek, MD, PhD Radboud University Nijmegen Medical Centre, Department of Epidemiology, Biostatistics and Health Technology Assessment, Nijmegen, Netherlands Reprint requests: R.H.M.A. Bartels, MD, PhD, Radboud University Nijmegen Medical Centre, Department of Neurosurgery, R. Postlaan 4, 6500 HB Nijmegen, Netherlands. Received, March 11, Accepted, January 2, Copyright 2010 by the Congress of Neurological Surgeons No Justification for Cervical Disk Prostheses in Clinical Practice: A Meta-Analysis of Randomized Controlled Trials OBJECTIVE: A meta-analysis was performed to evaluate whether a beneficial clinical effect of cervical disk prostheses over conventional cervical diskectomy with fusion exists. METHODS: A literature search was completed ending February 4, 2009, that included the abstract books of recent major spine congresses. All studies reported the results of singlelevel cervical disease without myelopathy. The Visual Analog Score (VAS) of the arm, VAS of the neck, Neck Disability Index, Physical Composite Scores of the Short Form 36, and Mental Composite Score of the Short Form 36, as well as adverse events, were evaluated. RESULTS: Nine records were found, totaling 1533 patients. Of these, 1165 were evaluable at the last follow-up at 12 or 24 months. As an effect measure, a pooled odds ratio (OR) was calculated at 12 and 24 months. At 12 months, the VAS arm reached statistical significance (OR = 0.698; 95% confidence interval [CI], ), as did the VAS neck (OR = 0.690; 95% CI, ), and the Physical Composite Scores (OR = 1.362; 95% CI, ) and the Mental Composite Score (OR = 1.270; 95% CI, ) of the Short Form 36, favoring arthroplasty. The Neck Disability Index at 24 months also reached statistical difference (OR = 0.794; 95% CI, ). All other measurements did not reveal any statistical difference. The number of complications, including secondary surgeries for adjacent segment disease, did not differ. CONCLUSION: A clinical benefit for the cervical disk prosthesis is not proven. Because none of the studies were blinded, bias of the patient or researcher is a probable explanation for the differences found. Therefore, these costly devices should not be used in daily clinical practice. KEY WORDS: Arthroplasty, Cervical spine, Disc prothesis, Meta-analysis Neurosurgery 66: , 2010 DOI: /01.NEU F online.com ABBREVIATIONS: ACDF, anterior cervical diskectomy with fusion; CI, confidence interval; FDA, Food and Drug Administration; MCS, Mental Component Score; NCI, Neck Disability Index; OR, odds ratio; PCS, Physical Component Score; SF, Short Form; VAS, Visual Analog Score Cervical disk prostheses are relatively new implants in the field of spinal surgery and the subject of many clinical and biomechanical investigations. In 2005, the authors of a review of spinal prostheses concluded that artificial cervical disks in particular should be considered at an experimental stage. 1 Although the clinical relevance of disk arthroplasty is not determined yet, they are frequently implanted in patients. Several observational reports have been reported, as well as results of randomized controlled trials. This metaanalysis was performed to obtain a more conclusive answer about the clinical relevance. METHODS AND RESULTS Search Strategy and Selection of Studies This review was conducted under the suggested QUORUM guideline standards. 2 The literature was searched using the entire Medline, PubMed, Embase and Cochrane databases ending February 4, The following search string was applied: ([prosthesis or prostheses or arthroplasty] or [artificial {disc or disk}]) and (cervical anterior [discectomy or diskectomy]). There was no language restriction. The search string did not specifically select for randomized controlled trials because these studies may have been excluded from our search if they had not been listed as such. NE UROSURGERY VOLUME 66 NUMBER 6 JUNE

2 BARTELS ET AL Two authors (R.H.M.A.B. and R.D.) reviewed the list of articles and their abstracts. Each article of interest was marked. The full text of the marked articles was retrieved, and only those reporting the results of a randomized controlled trial were included in this analysis. The references were checked for additional publications. Finally, the abstract books of meetings (Spine Week 2008 in Geneva, 24th Annual Meeting of the AANS/CNS Section on the Disorders of the Spine and Peripheral Nerves) dealing with the cervical spine in 2008 were checked. If 2 or more articles or abstracts of oral presentations reported on the same (sub) set of patients, then either the most recently published article was included or the article reporting on the largest patient sample was used. If studies did not report the actual number or the standard deviation but rather presented the data only in graph format, the authors were contacted. Most authors responded but were not able to provide additional clarification because of personal circumstances, because the data belonged to the US Food and Drug Administration (FDA) (if the product was still under investigation), or because the data presented were preliminary and not available for scientific research. Outcome The following outcome measurements were evaluable: Visual Analog Score (VAS) or analog of the arm, VAS or analog of the neck, Neck Disability Index (NDI), Physical Composite Score (PCS) of the Short Form (SF) 36, and the Mental Composite Score (MCS) of the SF 36. Analogs to the VAS score were also considered evaluable because the score will be expressed as a ratio. Therefore, the contribution of factors such as frequency or intensity will be minimal. In the pooled evaluation, radiological measurements were not evaluated. Adverse events were noted for both groups. Secondary surgery for a neck-related problem was considered an adverse event. Heterogeneity To establish inconsistency in the study results, a test for heterogeneity (Cochrane Q) was performed. However, because the test is susceptible to the number of trials included in the metaanalysis, we also calculated I 2. I 2, directly calculated from the Q statistic, describes the percentage of variation across the studies that is due to heterogeneity rather than change. I 2 ranges from 0% to 100%, with 0% indicating the absence of any heterogeneity. 3 Although absolute numbers for I 2 are not available, values <25% are considered low heterogeneity. When I 2 is <25%, low heterogeneity is assumed, and the effect is thought to be due to change. Conversely, when I exceeds 25%, then heterogeneity is thought to exist and the effect is random. Risk for Bias To assess the validity of the study, the risk for bias was emphasized according to the guidelines recommended by the Cochrane Collaboration. 4 Two authors (R.H.M.A.B. and A.L.M.V.) independently assessed the risk for bias of the included studies. If the authors disagreed on a topic, a discussion followed. If they still disagreed, the opinion of a third external and independent epidemiologist was requested. For the 6 domains, a judgment was made for the risk for bias. An eventual publication bias was graphically explored by funnel plots of standard error by log odds ratio (OR). Statistical Analysis For statistical analysis, Comprehensive Meta-Analysis version , USA was used. If the means were not explicitly reported, they were recalculated from the graphs. When the standard deviation was not provided, it was estimated with the statistical software, taking into account the sample size, the difference in effect, and the reported P value and assuming a 2-tailed test. If a P value was not provided, the value below which statistical significance is assumed was taken (generally P =.05) when statistical significance was not reached. If the authors reported statistical significance without mentioning the P value, the P value below which statistical significance is assumed was set at.001. The effect measure is expressed as the OR comparing arthroplasty and anterior cervical diskectomy with fusion (ACDF). Results are graphically presented, as well as the actual estimated pooled OR with 95% confidence intervals (95% CIs). A result of 1 within the 95% CI indicates a lack of superiority of one procedure over the other, whereas a result of <1 favors arthroplasty and vice versa. RESULTS The search yielded 6 articles on randomized controlled trials 5-10 and 3 abstracts of oral presentations (Figure 1) In total, 1533 patients were randomized, but only 1165 randomized patients could be evaluated at the last follow-up. The maximum follow- FIGURE 1. Diagram showing results of the literature and reasons for deleting references VOLUME 66 NUMBER 6 JUNE

3 CERVICAL DISK PROSTHESES IN CLINICAL PRACTICE TABLE 1. Main Characteristics of Included Studies a Study Author and Year Source of Data a ACDF, anterior cervical diskectomy with fusion. Treated Levels Patients per Group (Arthro plasty/acdf), n up was either 12 or 24 months (Table 1). All patients underwent a surgical procedure for a symptomatic herniated cervical disk at 1 level. Myelopathic signs or symptoms were exclusion criteria for all of the included studies. The results for the test on heterogeneity are depicted in Table 2. Seven analyses had I 2 < 25, indicating low heterogeneity. In 1 study, a high risk for bias was assumed because of inadequate follow-up data. 14 Even after the authors were contacted, the outcomes could not be recalculated because the exact numbers of patients in the arthroplasty group and the ACDF group were not known. Therefore, this study was not included. For 1 study, the risk for bias was unclear because of a language barrier (Table 3). 8 Patients and assessors of the outcome were not blinded in any of the studies, but the risk for bias from this was debatable. For many studies, the allocation concealment was unclear. However, some of them reported on results of studies that had been approved by Mean Age (Arthroplasty/ ACDF), y Prosthesis Used Maximum Follow- Up, mo Patients at Last Follow Up (Arthroplasty/ACDF), n Riina et al, Article Single 10/9 40.8/38.1 Prestige ST 24 9/7 Mummaneni et al, Article Single 276/ /43.9 Prestige ST /125 Nabhan et al, Article Single 19/21?/? Prodisc-C 12 19/21 Wang et al, Article Single 28/31?/? Bryan 24 21/23 Abitbol et al, Abstract Single 38/38?/? Cervicore 12 18/24 Murrey et al, Article Single 103/ /43.5 Prodisc-C /106 Fischgrund et al, Abstract Single 41/44?/? Cervicore 12 23/22 Guyer et al, Abstract Single 21/ /41.5 Kineflex C 12 21/20 Heller et al, Article Single 242/ /44.7 Bryan /194 Total 778/ /542 TABLE 2. Results for Testing for Heterogeneity With the Q Statistic and I2 Analysis Q df P I 2 NDI at 12 mo NDI at 24 mo VAS arm at 12 mo VAS arm at 24 mo VAS neck at 12 mo VAS neck at 24 mo SF36 MCS at 12 mo SF 36 MCS at 24 mo SF 36 PCS at 12 mo SF 36 PCS at 24 mo the US FDA. Therefore, a low risk for bias was assumed. The risk for publication bias was also considered low because grossly asymmetrical funnel plots were not found (Figure 2). Clinical outcome measures were VAS arm, VAS neck, NDI, SF 36 PCS, and/or SF 36 MCS. Not all studies provided results on these outcome measures. The results are graphically represented in Figure 3 and Table 4. The OR at 12 months for the VAS arm, VAS neck, and PCS and MCS of the SF 36 reached statistical significance in favor of arthroplasty. At 24 months, only the NDI reached statistical difference by a small margin, favoring arthroplasty. Adverse events were explicitly mentioned in 3 studies, 6,9,10 whereas 2 studies explicitly stated that no complications occurred. 5,7 The pooled OR for the occurrence of complications was (95% CI, ). A remarkable finding is that in 1 study secondary surgeries were performed for adjacent segment disease. 6 The OR for this event was (95% CI, ). DISCUSSION Symptomatic cervical radiculopathy caused by a herniated intervertebral disk is encountered frequently in daily practice. Although conservative management is the treatment of choice, many patients will visit an orthopedic surgeon or neurosurgeon for surgical treatment of persistent pain. Approximately 0.3% of all hospital discharges were associated with the surgical treatment of degenerative diseases of the cervical spine. The majority of admissions were related to symptomatic cervical herniated disk. 15 Various surgical procedures exist, among which the anterior cervical diskectomy with arthroplasty is relatively new. Cervical disk prostheses are the subject of many investigations. Numerous cervical disk prostheses are on the market. Presently, the reasons for implanting a cervical disk prosthesis are purely theoretical. The main assumption is that maintaining motion pre- NE UROSURGERY VOLUME 66 NUMBER 6 JUNE

4 BARTELS ET AL TABLE 3. Risk for Bias Represented for the Included Studies a Riina et al, Mummaneni et al, Nabhan et al, Wang et al, Fischgrund et al, Murrey et al, Guyer et al, Abitbol et al, Heller et al, Adequate Sequence Generation? Allocation Concealment? Blinding of Participants, Personnel, and Outcome Assessors? Incomplete Outcome Data Addressed? Free of Selective Outcome Reporting? Free of Other Sources of Bias Comments Yes U No Yes Yes Yes Although a sample size estimation and approval of an ethics board were not clearly described, the study was approved by US FDA, so we assume a low risk for bias Yes Yes No Yes Yes Yes Because the study is FDA approved, a low risk for bias is assumed Yes Yes No Yes Yes Yes The sample size estimation is unclear, as is the testing hypothesis; however, the presentation of the data is correct and therefore has a low risk for bias Yes U No U U U Because of the language (Chinese), a thorough understanding of the text was not possible; we repeatedly tried to contact the first author without result Yes U No Yes Yes Yes The design of the study is not explicitly described because it is presented as an abstract of oral presentation of the results of a US FDA trial Yes Yes No Yes Yes Yes Nearly all outcomes are adequately represented except the PCS and MCS; they were provided by the first author after request Yes U No Yes Yes Yes The design of the study is not explicitly described because it is presented as an abstract of oral presentation of the results of a US FDA trial Yes U No Yes Yes Yes The design of the study is not explicitly described because it is presented as an abstract of oral presentation of the results of a US FDA trial Yes Yes No Yes Yes Yes Sample size and results are clearly presented. However, an interim analysis was performed, and it is not clear whether this is corrected in the presentation of the results. Furthermore, 1-sided P values are reported, whereas the authors claim equality. Small differences can become statistically significant. Finally, it explicitly stated that the company producing the investigational device sponsored the trial a U, unclear; Yes, a low risk for bias; No, a high risk for bias; FDA, Food and Drug Administration VOLUME 66 NUMBER 6 JUNE

5 CERVICAL DISK PROSTHESES IN CLINICAL PRACTICE A B C D E F G H I J FIGURE 2. Funnel plots of standard error by log odds ratio of VAS arm (A, F), VAS neck ( B, G), NDI (C, H), MCS of SF 36 ( D, I), and PCS of SF 36 (E, J) at 12 months (left column) and 24 months (right column). vents adjacent disk disease that might occur after ACDF. How - ever, the relevance of this radiological phenomenon is currently the subject of debate. 16 It is questioned whether adjacent disk disease is related to the fusion procedure or to natural history. To the best of our knowledge, a series examining clinically symptomatic adjacent disk disease has never been reported, whereas ACDF has been acknowledged since the 1950s. Furthermore, in one of the included studies, secondary surgical procedures were performed in the arth - roplasty and control groups because of adjacent segment disease. 6 The validity of the studies was assessed by emphasizing the risk for bias. None of the studies blinded the patients or assessors of the outcome. It is impossible to blind the surgeon. It is not certain whether the lack of blinding introduces bias. The difference in arm pain may be an argument for bias because theoretically the amount of decompression of the nerve root should be the same for arthroplasty and ACDF. Therefore, the VAS for arm pain should be the same in both postoperative groups. However, our review shows that at 12 months the pooled OR favors arthroplasty. Because this difference cannot theoretically be explained by the use of a disk prosthesis, something else must be causing this difference. A potential explanation is that the patients and investigators are biased toward a positive result or that the surgeons altered their technique. The latter possibility was also suggested in the literature. 5 The difference in VAS of the neck is also remarkable. However, at 12 months, the NDI, a self-assessed questionnaire that measures the NE UROSURGERY VOLUME 66 NUMBER 6 JUNE

6 BARTELS ET AL A B C E FIGURE 3. Forest plots depicting the odds ratio with 95% confidence intervals (CIs) of Visual Analog Score (VAS) of the arm (A, D), VAS of the neck ( B, E), and Neck Disability Index (NDI) (C, F) at 12 months (left column) and 24 limitations caused by neck pain, did not reach statistically significant difference. Whether bias contributed to this phenomenon remains unclear. Another (more hypothetical) explanation is that a single-level fusion does not alter global cervical mobility. The adjustment period immediately after surgery, during which time each of the other segments accommodates a little extra motion to make up D F months (right column). The last line of each plot represents the pooled OR (diamond). Each study symbol is proportional to the size of the study. CADF indicates anterior cervical diskectomy with fusion. for that which has been lost at the fused segment, may be uncomfortable. Muscles acting on the fused segment will do so in vain, which also could result in some pain. This does not happen in arthroplasty. At 12 months, the PCS and MCS of the SF 36, a scale not specific for cervical disease, differed in favor of arthroplasty. In combination with results of the NDI at 12 months, bias could also 1158 VOLUME 66 NUMBER 6 JUNE

7 CERVICAL DISK PROSTHESES IN CLINICAL PRACTICE TABLE 4. Estimated Pooled Odds Ratio of Clinical Outcome Measures at 12 and 24 Months a Patients (Arthroplasty/ACDF), n Estimated OR 95% CI At 12 mo VAS arm 667/ VAS neck 645/ NDI 649/ SF 36 PCS 612/ SF 36 MCS 612/ At 24 mo VAS arm 586/ VAS neck 586/ NDI 586/ SF 36 PCS 565/ SF 36 MCS 565/ a ACDF, anterior cervical diskectomy with fusion; OR, odds ratio; CI, confidence interval; VAS, Visual Analog Score; NDI, Neck Disability Index; SF, Short Form; PCS, Physical Component Score; MCS, Mental Component Score. Bold numbers are 95% CIs indicating statistical significance favoring arthroplasty. contribute to the difference. Only a fully blinded study would answer this question. However, any clinical relevant difference at 24 months was not found. None of the outcome measures, especially the VAS neck or VAS arm, favored arthroplasty. An exception was the NDI, which differed only very slightly in favor of arthroplasty. The relatively large 95% CI that nearly included 1.0, combined with the absent differences in the other scores, raises doubt about the clinical relevance of this finding, Although all included studies treated patients with a singlelevel symptomatic herniated cervical disk, several disk prostheses were used. Biomechanical properties are not investigated in a standard way; therefore, the data from the literature cannot be compared. 17 On the other hand, the disk prostheses are designed to maintain cervical motion within a physiological range. Certainly within the period studied, differences in the design of the prostheses will not alter the clinical outcome. Therefore, we do not consider this point a major flaw. Publication bias is frequently cited as a reason for a lack of validity in meta-analyses. It may occur if studies that do not find any difference between arthroplasty and ACDF remain unpublished. However, many of the studies reviewed concluded that the clinical results were the same for arthroplasty and ACDF. Although only 3 studies explicitly described the allocation concealment, allocation bias is not considered a problem because all studies without this information except one 8 were related to or approved by the US FDA. Therefore, a correct procedure of allocation is assumed. To justify the use of a costly device (in the Netherlands, a prosthesis costs about 2500 [US $3200]), a clear clinical benefit should be proven. However, in all studies reviewed, a plate was used for ACDF. Although the implant is less costly, it still costs about 700 in the Netherlands. These remarks do not resemble a cost-efficiency or cost-minimization analysis. To calculate costs, many facts need to be known that are not provided in all studies. Therefore, we focused on the clinical effects. On the basis of the results of this meta-analysis, the use of the cervical disk prosthesis for a single-level herniated disk cannot be recommended. Especially in this era of evidence-based medicine, these and other costly devices should be appraised very critically because they greatly affect the costs of health care. Therefore, a proper cost analysis should be performed. A cervical disk prosthesis acts clinically as an expensive spacer. Considering the doubtful clinical relevance of a radiological entity and the unknown preventive effect of a cervical disk prosthesis on this effect, cervical arthroplasty should be discouraged in daily practice. Disclosure The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article. REFERENCES 1. WCB Evidence Based Practice Group. Artificial cervical and lumbar disc implants: a review of the literature. Worksafe BC care_providers/assets/pdf/artificial_cervical_lumbar_disc.pdf. 2. Shea B, Dubé C, Moher D. Assessing the quality of reports os systematic reviews: the Quorum statement compared to other tools. In: Egger M, Smith GD, Altman DG, eds. Systematic Reviews in Health Care: Meta-Analysis in Context. London: BMJ Publishing Group; 2006: Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414): Higgins JPT, Altman DG. Assessing the risk of bias in included studies. In: Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions. The Cochrane Collaboration; Available at: 5. Riina J, Patel A, Dietz JW, Hoskins JS, Trammell TR, Schwartz DD. Comparison of single-level cervical fusion and a metal-on-metal cervical disc replacement. Am J Orthop (Belle Mead NJ). 2008;37(4):E71-E Mummaneni P, Burkus K, Haid R, Traynelis VC, Zdeblick TA. Clinical and radiographic analysis of cervical disc arthroplasty compared with allograft fusion: a randomized controlled clinical trial. J Neurosurg Spine. 2007;6(3): Nabhan A, Ahlhelm F, Shariat K, et al. The Prodisc-C prosthesis: clinical and radiological experience 1 year after surgery. Spine. 2007;32(18): Wang Y, Cai B, Zhang XS, et al. Clinical outcomes of single level Bryan cervical disc arthroplasty: a prospective controlled study [in Chinese]. Zhonghua Wai Ke Az Zhi. 2008;46(5): Murrey D, Janssen M, Delamarter R, et al. Results of the prospective, randomized, controlled multicenter Food and Drug Administration investigational device exemption study of the Prodisc-C total disc replacement vs anterior discectomy and fusion for the treatment of 1-level symptomatic cervical disc disease. Spine J. 2009;9(4): Heller J, Sasso RC, Papadopoulos SM, et al. Comparison of BRYAN cervical disc arthroplasty with anterior cervical decompression and fusion: clinical and radiographic results of a randomized, controlled clinical trial. Spine (Phila Pa 1976). 2009; 34(2): Guyer R, Lauryssen C, Blementhal S. A prospective randomized comparison of cervical total disc replacement to anterior cervical fusion. Abstractbook Spineweek Geneva. 2008;C73: Abitbol JJ, Baldwin NG, Youssef JA, Wright NM. Cervicore disc replacement vs fusion for cervical nerve root compression: functional and occupational outcomes from a prospective, randomized multicenter trial. Abstractbook Spineweek Geneva. 2008;C74:46. NE UROSURGERY VOLUME 66 NUMBER 6 JUNE

8 BARTELS ET AL 13. Fischgrund J, Garcia R, Gratch M, Wright N. Cervicore disc replacement vs. fusion, results from four study sites in a prospective, randomized, multicenter trial. Abstractbook Spineweek Geneva. 2008;AB5: Darden BV, Murrey DB, Zigler JE, et al. Results of the prospective, randomized, multi-center food and drug administration investigational device exemption study of the Prodisc C total disc replacement versus anterior discectomy and fusion for the treatment of 1-level symptomatic cervical disc disease. Abstractbook Spineweek Geneva. 2008;C4: Murrey D, Janssen M, Delamarter R, et al. Results of the prospective, randomized, multi-center Food and Drug Administration investigational device exemption of the Prodisc-C total disc replacement (TDR) vs. anterior discectomy and fusion. Abstractbook Spineweek Geneva. 2008;AB7: Heller J. Bryan cervical disc: results of the U.S. randomized clinical trial v ACDF. Abstractbook Spineweek Geneva. 2008;AB6: Fessler RG, Papadopoulos S, Anderson P, Heller J, Sasso R. Comparison of Bryan cervical disc arthroplasty with anterior cervical decompression and fusion: clinical and radiographic results of a randomized controlled clinical trial. Presented at: 24th Annual Meeting of the AANS/CNS Section on the Disorders of Spine and Peripheral Nerves. Abstractbook. 2008;100: Sasso RC, Smucker JD, Hacker RJ, Heller JG. Clinical outcomes of BRYAN cervical disc arthroplasty: a prospective, randomized, controlled, multicenter trial with 24-month follow-up. J Spinal Disord Tech. 2007;20(7): Coric D, Finger F, Boltes P. Prospective randomized controlled study of the Bryan Cervical Disc: early clinical results from a single investigational site. J Neurosurg Spine. 2006;4(1): Porchet F, Metcalf NH. Clinical outcomes with the Prestige II cervical disc: preliminary results from a prospective randomized clinical trial. Neurosurg Focus. 2004; 17(3):E Nabhan A, Ahlhelm F, Pitzen T, et al. Disc replacement using Pro-Disc C versus fusion: a prospective randomised and controlled radiographic and clinical study. Eur Spine J. 2007;16(3): Sasso RC, Smucker JD, Hacker RJ, Heller JG. Artificial disc versus fusion: a prospective, randomized study with 2-year follow-up on 99 patients. Spine (Phila Pa 1976). 2007;32(26): Wang MC, Chan L, Maiman DJ, Kreuter W, Deyo RA. Complications and mortality associated with cervical spine surgery for degenerative disease in the United States. Spine (Phila Pa 1976). 2007;32(3): Bartels RH. Evidence based medicine: a marketing tool in spinal surgery. Neurosurgery. 2009;64(6):E Bartels RH, Donk R, Pavlov P, van Limbeek J. Comparison of biomechanical properties of cervical artificial disc prosthesis: a review. Clin Neurol Neurosurg. 2008; 110(10): Acknowledgment We want to thank H.P. Hu, MD, an orthopedic surgeon, for his assistance in translating the Chinese text of the tables in one of the articles. We also want to express our gratitude to Ben Guiot, MD, a neurosurgeon, for correction of the text. COMMENTS CONTACT THE EDITORIAL OFFICE To reach the Editorial Office, please use the following information. NEUROSURGERY Nelson M. Oyesiku, MD, PhD, FACS Editor-in-Chief, Neurosurgery Emory University School of Medicine 1510 Clifton Road, Suite G65 Atlanta, GA USA Tel: (404) Fax: (404) EICNS@emory.edu or managingeditor@1cns.org Web site: neurosurgery-online.com Bartels et al have provided a very important contribution to the spine literature. Their analysis of the literature on cervical total disc arthroplasty is extraordinarily revealing. They emphasize the importance of bias in clinical trials. We must all take pause and attempt to fully understand and appreciate both the meaning and impact of their message. They have eloquently portrayed and exposed the underlying, and often subliminal, deficiencies of the total disc arthroplasty literature. In my opinion, this literature is simply and unequivocally flawed. Their observations and conclusions support such. This is not to lay blame on only the arthroplasty literature, however. Surgical prospective randomized control trials are, in general, plagued by such inherent biases, that are human frailty based. Fundamentally, it is very difficulty for both the patient and the surgeon, both as participants in such trials, to maintain objectivity. Without such objectivity, rationale assessments of outcome are essentially impossible. Regardless, the open discussion of such flaws in the literature is extraordinarily important as we enter the new era of health care reform and the obligatory emphasis on cost containment. Bartels et al are to be congratulated for their courage and insight, as they cause us to ponder the very essence and foundation of our future. We need more many, many more such dialogues. Edward C. Benzel Cleveland, Ohio Bartels et al performed a sophisticated meta-analysis of randomized trials examining symptomatic single level cervical disc disease treated with either arthrodesis and or arthroplasty. At 12 months Visual Analog Score of arm and neck pain as well as Short Form-36 Physical Component Score and Mental Component Score favored arthroplasty at a statistically significant level. The Neck Disability Index score at 24 months showed a statistical difference in favor of arthroplasty. These results make it difficult to understand the conclusion that the use of cervical disc prostheses is not justified. Although the differences in outcome were small, they did favor arthroplasty. The authors have concerns about the overall cost of arthroplasty but no comprehensive economic analysis of these patients was performed. Although in the Netherlands there is a significant difference in the cost of the devices required for fusion as compared to arthroplasty all of the other economic parameters are not known. Vincent C. Traynelis Chicago, Illinois 1160 VOLUME 66 NUMBER 6 JUNE