This review concluded that spinal cord stimulation was effective in the treatment of failed back surgery syndrome in terms of pain reduction and was less costly in the long term. This was a well-conducted review and although the evidence was limited the authors conclusions are likely reliable.

To assess the effectiveness and cost effectiveness of spinal cord stimulation for people with chronic pain as a result of failed back surgery syndrome.

MEDLINE, EMBASE, LILACS, CINAHL and The Cochrane Library databases were searched from inception to Sept 2006; an update search of MEDLINE and EMBASE was conducted in Jan 2008. The search terms were reported. The reference lists of included studies were checked for additional studies, as were relevant reviews and Health Technology Assessment reports. A manufacturer of spinal cord stimulation devices was contacted for unpublished studies.

Studies of clinical effectiveness in adults with chronic pain as a result of failed back surgery syndrome were eligible for inclusion. Studies needed to be randomised controlled trials (RCTs), non-randomised controlled trials or prospective cohort studies (with a parallel control group) that evaluated all types and methods of spinal cord stimulation in addition to usual care in comparison with usual care (when other treatments had failed). Chronic pain was defined as six months or more. The reviewers also included case series if the study had at least 50 participants with permanent implants, if at least 60% of participants had failed back surgery syndrome and if there was at least one year follow up. Studies of participants with angina pectoris, malignancies, complex regional pain syndrome and pain in the peripheral nervous system were excluded.

Primary outcome measures were reduction of pain assessed with validated instruments, and secondary outcomes were quality of life, activities of daily living, functional status, depression and adverse events. Observational studies that evaluated serious adverse events (those that resulted in hospital admission or prolonged hospital stay) were also eligible for inclusion. Studies of cost effectiveness had to include an explicit analysis of cost and health-related outcomes for spinal cord stimulation and have at least one comparator.

In the included studies of effectiveness, mean age ranged from 47 to 54 years (where reported). The percentage of participants with failed back surgery syndrome ranged from 61% to 100% (100% of participants in the RCTs and cohort study). Electrode types used for spinal cord stimulation were described; most studies used percutaneous electrodes. Where reported, the trial stimulation period ranged from 24 hours to at least one month. Mean duration of follow-up varied between six months and 8.8 years. Various definitions and outcome measures for pain were reported.

Two reviewers independently selected the studies. It was unclear how any discrepancies may have been resolved.

Quality assessment of RCTs was based on nine criteria: randomisation; allocation concealment; comparability of groups at baseline; eligibility criteria specified; outcome assessors blinded; care provider blinded; patient blinded; point estimates and measure off variability presented for the primary outcome measure; and intention-to-treat analysis, withdrawals and dropouts described.

Quality assessment of cohort studies was based on a 10-item checklist. Case series were assessed using a six-item checklist. Quality assessment was undertaken by one reviewer and checked by a second.

Economic evaluations were assessed using a published 10-point quality checklist. Two reviewers independently assessed the papers; any disagreements were resolved by consensus.

For cross-over studies, data were extracted for the period up to the first cross-over. For case series, data were extracted from after the intervention only, or before and after with difference data or results as described by the authors. For binary outcomes, relative risks with 95% confidence intervals (CIs) were calculated. Numbers needed to treat, with 95% CIs, were calculated. The data were extracted by one reviewer and checked by a second.

The studies were combined in a narrative synthesis. Data were analysed by length of follow-up (≤six months, six to 12 months, one to three years, ≥three years).

For clinical effectiveness, 16 studies (overall n=not clear) were included in the review: two RCTs; one retrospective cohort study; and 13 case series. The RCTs were considered to be of moderate quality; all other studies were of low quality.

Pain reduction (two RCTs, 13 case series): Spinal cord stimulation treatment resulted in significant improvements in pain reduction compared to the control treatment (two RCTs). In one RCT, number needed to treat was three (95% CI 2 to 7) for at least 50% pain relief in the leg. In another RCT, number needed to treat was four (95% CI 3 to 47). All 13 case series demonstrated success or very good to good pain relief with spinal cord stimulation treatment (statistical results reported in only one study).

Quality of life (one RCT, three case series: There were significant improvements in quality of life in seven of eight domains of SF-36 after six months follow up in participants treated with spinal cord stimulation compared to controls (one RCT), p values varied from 0.01 to <0.001. Significant improvements in quality of life with spinal cord stimulation treatment were also reported in three case series.

Activities of daily living (one RCT, one retrospective cohort, seven case series): One RCT demonstrated no significant differences between treatment groups in the percentage of patients who returned to work. All other studies reported improvements in activities of daily living with spinal cord stimulation treatment (statistical results reported in only one study).

Functional status (one RCT, one retrospective cohort, two case series): The RCT reported significant improvements in functional status with spinal cord stimulation treatment after six months follow up (p<0.001). Significant improvements were reported in the other studies.

Depression (one case series): A significant improvement in depression was reported after one year follow up.

Drug use (one RCT, one retrospective cohort, five case series): There were no significant differences between the treatment and comparison groups in opioids use and in percentages of participants who used opioids, nonsteroidal anti-inflammatory drugs, antidepressants and anticonvulsants after six months (one RCT). Results from the other studies varied.

Patient satisfaction (two RCTs, one retrospective cohort, three case series): Both RCTs reported significantly improved patients satisfaction with spinal cord stimulation treatment after six months follow up. The other studies reported that 60% or more patients were satisfied with spinal cord stimulation treatment after more than three years follow up.

Adverse events: Any infection, mean 5% (range 2% to 14%; 13 studies); infection that required removal of the device, mean 3% (range 1% to 8%; nine studies); pain along the lead or pulse generator, mean 12% (range 2% to 23%; six studies); other event, mean 11% (range 3% to 18%; seven studies); hardware failure, mean 28% (range 2% to 71%; 10 studies).

Three economic evaluations were included (cost data for two of the three came from one study). Limitations of these analyses were presented.

In one study the incremental cost per quality-adjusted life years gained for spinal cord stimulation over conventional medical management was US $64,266 (2007) over two years; spinal cord stimulation was dominant in comparison with conventional medical management in the lifetime analysis. In another study, the mean cost for spinal cord stimulation over five years was US $24,799/patient and US $33,722 for conventional pain therapy (costs were higher in the first 2.5 years for spinal cord stimulation). In the last study, the mean cost for spinal cord stimulation therapy was US $42,897, with a mean cost in the reoperation group of US $51,917 (the difference in favour of spinal cord stimulation was not significant).

The evidence demonstrated that spinal cord stimulation was effective in the treatment of failed back surgery syndrome in terms of pain reduction. It was less costly in the long term, but there was an initial high cost associated with device implantation and maintenance.

The review addressed a clear question and was supported by appropriate inclusion criteria. Steps were taken to minimise publication bias by searching for published and unpublished data. No language restrictions appeared to be applied to the search strategy. Quality of the included studies was assessed according to appropriate criteria and used to support the review findings. A number of reviewers were involved in the review process, thus minimising the potential for reviewer bias. Comprehensive details of the included studies were reported, although some data were not clearly presented (for example, sample sizes). Given clinical differences between the studies, they were appropriately summarised in a narrative synthesis. This was a well-conducted review and although the evidence was limited the authors' conclusions are likely reliable.

Implications for practice: The authors stated that the data used to inform the economic evaluations had some reliability limitations that made their use in policy change uncertain.

The authors did not state any implications for research.

Funded by the Dutch Health Care Insurance Board (CVZ).

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.