OBJECTIVE: To assess the safety and efficacy of artificial disc replacement (ADR) technology for degenerative disc disease (DDD). CLINICAL NEED: Degenerative disc disease is the term used to describe the deterioration of 1 or more intervertebral discs of the spine. The prevalence of DDD is roughly described in proportion to age such that 40% of people aged 40 years have DDD, increasing to 80% among those aged 80 years or older. Low back pain is a common symptom of lumbar DDD; neck and arm pain are common symptoms of cervical DDD. Nonsurgical treatments can be used to relieve pain and minimize disability associated with DDD. However, it is estimated that about 10% to 20% of people with lumbar DDD and up to 30% with cervical DDD will be unresponsive to nonsurgical treatments. In these cases, surgical treatment is considered. Spinal fusion (arthrodesis) is the process of fusing or joining 2 bones and is considered the surgical gold standard for DDD. Artificial disc replacement is the replacement of the degenerated intervertebral disc with an artificial disc in people with DDD of the lumbar or cervical spine that has been unresponsive to nonsurgical treatments for at least 6 months. Unlike spinal fusion, ADR preserves movement of the spine, which is thought to reduce or prevent the development of adjacent segment degeneration. Additionally, a bone graft is not required for ADR, and this alleviates complications, including bone graft donor site pain and pseudoarthrosis. It is estimated that about 5% of patients who require surgery for DDD will be candidates for ADR. REVIEW STRATEGY: The Medical Advisory Secretariat conducted a computerized search of the literature published between 2003 and September 2005 to answer the following questions: What is the effectiveness of ADR in people with DDD of the lumbar or cervical regions of the spine compared with spinal fusion surgery?Does an artificial disc reduce the incidence of adjacent segment degeneration (ASD) compared with spinal fusion?What is the rate of major complications (device failure, reoperation) with artificial discs compared with surgical spinal fusion?One reviewer evaluated the internal validity of the primary studies using the criteria outlined in the Cochrane Musculoskeletal Injuries Group Quality Assessment Tool. The quality of concealment allocation was rated as: A, clearly yes; B, unclear; or C, clearly no. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) system was used to evaluate the overall quality of the body of evidence (defined as 1 or more studies) supporting the research questions explored in this systematic review. A random effects model meta-analysis was conducted when data were available from 2 or more randomized controlled trials (RCTs) and when there was no statistical and or clinical heterogeneity among studies. Bayesian analyses were undertaken to do the following: Examine the influence of missing data on clinical success rates;Compute the probability that artificial discs were superior to spinal fusion (on the basis of clinical success rates);Examine whether the results were sensitive to the choice of noninferiority margin. SUMMARY OF FINDINGS: The literature search yielded 140 citations. Of these, 1 Cochrane systematic review, 1 RCT, and 10 case series were included in this review. Unpublished data from an RCT reported in the grey literature were obtained from the manufacturer of the device. The search also yielded 8 health technology assessments evaluating ADR that are also included in this review. Six of the 8 health technology assessments concluded that there is insufficient evidence to support the use of either lumbar or cervical ADR. The results of the remaining 2 assessments (one each for lumbar and cervical ADR) led to a National Institute for Clinical Excellence guidance document supporting the safety and effectiveness of lumbar and cervical ADR with the proviso that an ongoing audit of all clinical outcomes be undertaken owing to a lack of long-term outcome data from clinical trials. Regarding lumbar ADR, data were available from 2 noninferiority RCTs to complete a meta-analysis. The following clinical, health systems, and adverse event outcome measures were synthesized: primary outcome of clinical success, Oswestry Disability Index (ODI) scores, pain VAS scores, patient satisfaction, duration of surgery, amount of blood loss, length of hospital stay, rate of device failure, and rate of reoperation. The meta-analysis of overall clinical success supported the noninferiority of lumbar ADR compared with spinal fusion at 24-month follow-up. Of the remaining clinical outcome measures (ODI, pain VAS scores, SF-36 scores [mental and physical components], patient satisfaction, and return to work status), only patient satisfaction and scores on the physical component scale of the SF-36 questionnaire were significantly improved in favour of lumbar ADR compared with spinal fusion at 24 months follow-up. Blood loss and surgical time showed statistical heterogeneity; therefore, meta-analysis results are not interpretable. Length of hospital stay was significantly shorter in patients receiving the ADR compared with controls. Neither the number of device failures nor the number of neurological complications at 24 months was statistically significantly different between the ADR and fusion treatment groups. However, there was a trend towards fewer neurological complications at 24 months in the ADR treatment group compared with the spinal fusion treatment group. Results of the Bayesian analyses indicated that the influence of missing data on the outcome measure of clinical success was minimal. The Bayesian model indicated that the probability for ADR being better than spinal fusion was 79%. The probability of ADR being noninferior to spinal fusion using a -10% noninferiority bound was 92%, and using a -15% noninferiority bound was 94%. The probability of artificial discs being superior to spinal fusion in a future trial was 73%. Six case series were reviewed, mainly to characterize the rate of major complications for lumbar ADR. The Medical Advisory Secretariat defined a major complication as any reoperation; device failure necessitating a revision, removal or reoperation; or life-threatening event. The rates of major complications ranged from 0% to 13% per device implanted. Only 1 study reported the rate of ASD, which was detected in 2 (2%) of the 100 people 11 years after surgery. There were no RCT data available for cervical ADR; therefore, data from 4 case series were reviewed for evidence of effectiveness and safety. Because data were sparse, the effectiveness of cervical ADR compared with spinal fusion cannot be determined at this time. The rate of major complications was assessed up to 2 years after surgery. It was found to range from 0% to 8.1% per device implanted. The rate of ASD is not reported in the clinical trial literature. The total cost of a lumbar ADR procedure is $15,371 (Cdn; including costs related to the device, physician, and procedure). The total cost of a lumbar fusion surgery procedure is $11,311 (Cdn; including physicians' and procedural costs). CONCLUSIONS: Lumbar Artificial Disc Replacement Since the 2004 Medical Advisory Secretariat health technology policy assessment, data from 2 RCTs and 6 case series assessing the effectiveness and adverse events profile of lumbar ADR to treat DDD has become available. The GRADE quality of this evidence is moderate for effectiveness and for short-term (2-year follow-up) complications; it is very low for ASD.The effectiveness of lumbar ADR is not inferior to that of spinal fusion for the treatment of lumbar DDD. The rates for device failure and neurological complications 2 years after surgery did not differ between ADR and fusion patients. Based on a Bayesian meta-analysis, lumbar ADR is 79% superior to lumbar spinal fusion.The rate of major complications after lumbar ADR is between 0% and 13% per device implanted. The rate of ASD in 1 case series was 2% over an 11-year follow-up period.Outcome data for lumbar ADR beyond a 2-year follow-up are not yet available.Cervical Artificial Disc Replacement Since the 2004 Medical Advisory Secretariat health technology policy assessment, 4 case series have been added to the body of evidence assessing the effectiveness and adverse events profile of cervical ADR to treat DDD. The GRADE quality of this evidence is very low for effectiveness as well as for the adverse events profile. Sparse outcome data are available.Because data are sparse, the effectiveness of cervical ADR compared with spinal fusion cannot be determined at this time.The rate of major complications was assessed up to 2 years after surgery; it ranged from 0% to 8.1% per device implanted. The rate of ASD is not reported in the clinical trial literature.