Introduction
Failed Back Surgery Syndrome (FBSS) refers to intractable pain or discomfort in the lumbar spine, buttocks, or lower limbs that occurs after laminectomy or discectomy. FBSS is a common and challenging complication following lumbar spine surgery, with reports indicating that approximately 30% of patients experience FBSS after lumbar disc surgery.1 As the number of lumbar spine surgeries has increased in recent years, the incidence of patients diagnosed with FBSS has also risen.2
Spinal cord stimulation (SCS) is a widely accepted treatment for FBSS that has been used clinically for over 30 years. SCS employs electrical impulse signals to stimulate spinal cord nerves, effectively blocking the conduction of pain signals to provide relief. SCS can be classified into traditional/tonic low-frequency SCS, subperception SCS, high-frequency SCS, and burst SCS. Traditional low-frequency SCS operates at a frequency of about 40 to 60 Hz,3 generating an anesthetic sensation that “covers” the painful area. High-frequency SCS utilizes electrical pulse frequencies ranging from 5 to 10 kHz, with 10 kHz being the most common, and does not produce abnormal sensations.4 Subperception SCS functions at a frequency between low-frequency and high-frequency SCS, typically around 1000 Hz, and does not elicit abnormal sensations. Burst SCS mainly consists of 40 Hz stimulus clusters, each containing five 500 Hz spike pulses, and is less likely to cause abnormal sensations.5
The emergence of novel electrical stimulation modalities has introduced new ideas and options for SCS in the treatment of FBSS. As these innovative SCS techniques become more widely applied in clinical practice, the challenge of selecting the appropriate method arises, making a comparative evaluation of the effectiveness and advantages of different SCS modalities increasingly urgent.
Several reviews have evaluated the effectiveness of different SCS in patients with FBSS.6 Studies have shown that SCS can significantly improve pain, daily functioning, and quality of life in these patients.7,8 Previous studies have also demonstrated that novel electrical stimulation is superior to conventional low-frequency electrical stimulation for pain relief.9–11 Some studies have grouped subperception SCS with traditional low-frequency SCS. North et al noted that traditional SCS (40–60 Hz) requires paresthesia coverage for therapeutic efficacy, whereas subperception SCS (1000–1200 Hz) functions without inducing sensory phenomena.12 Regarding pain relief, subperception SCS is noninferior to conventional low-frequency SCS and may be an effective alternative for patients who do not respond to traditional therapy.4 This mechanistic divergence is corroborated by two randomized, crossover, controlled trials by Sokal et al and Rigoard et al. These studies compared traditional low-frequency SCS (40–60 Hz) with subperception SCS (1000–1200 Hz) and found significant differences in pain relief between the two.13,14 Consequently, our review classifies conventional low-frequency SCS and subperception SCS as separate therapeutic categories due to the pooling of these distinct modalities and divergent clinical efficacy profiles.
Additionally, there have been no network meta-analyses evaluating the effectiveness of different SCS modalities for FBSS, despite their advantages over traditional meta-analyses, such as improved statistical power through the use of direct and indirect estimations, the ability to compare treatments not originally assessed, and ranking capabilities.15 The aim of our review was to conduct the first systematic network meta-analysis of RCT studies comparing different SCS frequencies for FBSS. We compared SCS with various frequencies directly and indirectly, and sequenced them to evaluate their effectiveness and superiority in terms of pain relief, daily functioning, and quality of life.
Materials and Methods
Data Sources and Search Strategies
The systematic review was meticulously conducted in full compliance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to ensure methodological rigor and transparency.16 This systematic review has been officially registered in PROSPERO (registration no. CRD42024496483). Bayesian random-effects models were employed for network meta-analysis to comprehensively evaluate treatment effects on pain relief, functional improvement, and quality of life enhancement. A thorough exploration of the database was conducted up to October 31, 2024, including PubMed, MEDLINE, Embase and Cochrane Central Register of Controlled Trials (CENTRAL), systematically searching for RCTs that described the use of SCS for FBSS. The retrieval strategy was meticulously developed and systematically executed by an experienced librarian in close collaboration with the study’s principal investigator (X.F).
Inclusion and Exclusion Criteria
The analysis included only English-language prospective RCTs that compared two or more SCS frequency protocols in patients with failed FBSS. FBSS is defined as persistent back or leg pain following at least one lumbar surgery.17 Inclusion criteria were: 1) randomized controlled trials, 2) comparisons of different frequencies of SCS, 3) study population with FBSS, and 4) reporting outcomes of back and/or leg pain and/or overall pain at baseline and follow-up. Exclusion criteria included: 1) fewer than 10 patients, 2) research subjects who have not been clearly diagnosed with FBSS or whose primary population is not FBSS, 3) review articles, case reports, case series, letters to the editor, commentaries, conference proceedings, laboratory science studies, and other non-relevant studies, and 4) studies that did not report outcomes of interest. If a trial reported multiple time points for the same outcome, we selected the three-month or closest follow-up outcome.
Screening Process
We imported the database-retrieved entries into EndNote 20.4.1 (Clarivate Analytics, Philadelphia, PA, USA) and removed duplicates, and evaluated them in relation to findings from other sources. The screening process comprised three distinct phases. Initially, two reviewers (QYY and ZCP) independently screened the articles by title, with any uncertain entries retained for further review. Next, all articles selected in the first stage underwent an abstract review, and any disagreements among the reviewers were resolved through discussion. Finally, full-text review was conducted for articles with eligible titles and abstracts, following predefined inclusion and exclusion criteria.
Data Collection
Two independent reviewers systematically extracted data using a standardized electronic form, with discrepancies resolved through consensus or third-party adjudication. From each included study, we collected study identifiers (first author, publication year, country), methodological characteristics (study design [RCT/RCD], follow-up duration, attrition rate), participant demographics (sample size, sex distribution, age [mean ± SD or median (IQR)], proportion of FBSS patients [range: 45–100%]), intervention specifications including spinal cord stimulation modality (conventional low-frequency [LF: 40–100 Hz], subperception stimulation [500–1200 Hz], high-frequency [HF: 10 kHz], or burst stimulation). The outcome measures included pain intensity assessed using the visual analog scale (VAS) and numeric pain rating scale (NPRS) at baseline and post-intervention with site-specific changes (back/leg/global), functional disability evaluated by the Oswestry Disability Index (ODI),18 health-related quality of life measured with the EuroQol-5 Dimensions (EQ-5D),19 and responder rate defined as either ≥50% pain reduction or ≥2-point improvement in pain related scores.
Quality Assessment of Evidence
Two independent reviewers (Q.Y.Y. and Z.C.P.) assessed the risk of bias using the revised Cochrane risk-of-bias tool for randomized trials (RoB 2.0).20 The overall risk of bias was evaluated as low when all domains were assessed as low risk, high when at least one domain was high risk or multiple domains raised some concern, and some concern if the overall judgment was neither low nor high. To provide a comprehensive methodological evaluation, we employed the CINeMA (Confidence in Network Meta-Analysis) framework to assess the certainty of the evidence across six key domains (within-study bias, reporting bias, indirectness, imprecision, heterogeneity, and inconsistency).21,22 This approach addresses issues specific to network meta-analysis that extend beyond the scope of RoB 2.0. Using the CINeMA framework, we examined intransitivity (violation of the transitivity assumption) by comparing key effect modifiers, including age, sex, and baseline pain scores, across studies that provided both direct and indirect evidence for each comparison.
Statistical Analysis
A frequency network meta-analysis (NMA) was performed using Stata 18.0 (command: mvmeta23–25). The analysis incorporated both direct and indirect evidence from a variety of SCS modalities. Prior to performing the NMA, the assumption of cross-sectionality was validated by assessing demographic comparability (age, sex, baseline pain score) and risk of bias distribution across studies. The presence of heterogeneity among the studies was addressed by implementing a random-effects model with restricted maximum likelihood estimation. This heterogeneity was then quantified using the τ² metric, which is commonly interpreted as follows: low < 0.04; low-moderate 0.04–0.16; moderate-high 0.16–0.36; high > 0.36.8,26,27 It was hypothesized that a common τ² existed among the various contrasts, and that the correlation coefficient of the inter-study covariance matrix was equivalent to 0.5. The consistency between direct and indirect evidence was verified by means of a node-split analysis.28
Missing standard deviations of baseline changes were derived using correlation coefficients25 or estimated from comparable studies. For data reported as median and range, means and standard deviations were calculated using the Hozo method.29 Treatment effects were synthesized as pooled estimates with 95% confidence intervals.
The ensuing results are presented in the following ways: network plots were used to plot SCS pattern comparisons; forest plots were used to show pairwise treatment effects; and Surface Under Cumulative Ranking (SUCRA) values (range: 0–1) were used to objectively rank the interventions. This metric quantifies the relative superiority of each treatment option, where higher SUCRA scores indicate a greater likelihood of being among the most effective interventions.26,30
Results
Study Characteristics
A total of 1913 studies were identified after searching PubMed, MEDLINE, Embase, and Cochrane, and the search strategy is shown in Tables S1–S4. An additional 20 studies were identified after performing manual searches, such as reviewing references of included studies/reviews and additional searches for related/cited articles in PubMed. After eliminating duplicates, reviewing titles and abstracts, and reading the full text, our systematic review finalized 11 studies, as detailed in the flow chart (Figure 1). The list of data extracted from the included studies is in Table S5. Of the included articles, six studies were two-armed, three analyzed HF vs LF,31–33 two analyzed subperception vs LF SCS,4,12 one analyzed burst vs sham.34 Four studies were three-armed, three of which analyzed burst vs subperception vs sham,11,35 one study analyzed LF vs subperception vs burst,13 and one study analyzed HF vs subperception vs sham.36 There was also 1 study that was a four-arm study that analyzed burst vs subperception vs LF vs sham.14 In the study by Kapural et al, 77% of patients were diagnosed with FBSS.32 In two independent studies by North et al, one publication4 reported 46% of patients with FBSS while the other12 reported 45%. In Sokal et al’s research, this proportion was 72%.14 All other included studies enrolled only FBSS patients.
|
Figure 1 Flow diagram of preferred reporting items identified, included, and excluded for systematic reviews and meta-analyses (PRISMA).
|
Table 1 provides a summary of the characteristics of the trials included in the systematic review and network meta-analysis.
 |
Table 1 Summarizes the Characteristics of the Trials Included in the Systematic Review and Network Meta-Analysis
|
Risk of Bias, Certainty of Evidence, and Consistency
Figure 2 presents the risk of bias for each trial, see Figure S1 and Table S6 for more details. One major limitation was the lack of sufficient details in some studies regarding the implementation of blinding for participants, investigators, and outcome assessors, with certain studies also failing to comprehensively report how the blinding process was conducted. For our consistency evaluation (ie, alignment of direct and indirect evidence), side-splitting results suggested there is no strong statistical evidence of global inconsistency was reported for most outcomes (Tables S7–S16). Consistency assessments also prompted vigilance over waist circumference data (P=0.08), which warrants more rigorous randomized controlled trials in the future. The τ2 result did not identify any high heterogeneity in the network, and most comparisons were low or low-moderate levels of heterogeneity (Tables S7–S16). After assessing the level of evidence using CINeMA, most of the results of the pairwise comparisons were of low or moderate confidence (Supplementary Material 1, Figures S2–S8 and Tables S17–S19). All networks met the principle of transitivity, endowing the validity of indirect comparisons (Supplementary Material 1, Figures S2–S8 and Tables S17–S19).
 |
Figure 2 Plot of the risk of bias for each trial.
|
Synthesizing the Data
For each specific outcome, we generated network plots (Figures 3–8). Within these networks, edge thickness is proportional to the number of direct comparison trials (ranging from 3 to 12 studies), while edge length represents the inverse variance of the direct effect estimates. All metrics were standardized according to Cochrane network meta-analysis protocols to reflect evidence intensity and statistical precision.
 |
Figure 3 Network geometry of spinal cord stimulation (SCS) modalities for global pain.
|
 |
Figure 4 Network geometry of spinal cord stimulation (SCS) modalities for back pain.
|
 |
Figure 5 Network geometry of spinal cord stimulation (SCS) modalities for leg pain.
|
 |
Figure 6 Network geometry of spinal cord stimulation (SCS) modalities for responder rate (50% of pain im-provement or 2-point decrease).
|
 |
Figure 7 Network geometry of spinal cord stimulation (SCS) modalities for ODI (Oswestry Disability Index).
|
 |
Figure 8 Network geometry of spinal cord stimulation (SCS) modalities for EQ-5D (used to describe health-related quality of life).
|
We also created forest plots for pairwise comparisons (Figures 9–14). Orange dots denote the mean difference, with horizontal lines indicating 95% confidence intervals. Numerical results (mean difference [95% CI]) are listed adjacent to each comparison (eg, LF vs HF: 0.08 [−1.08, 1.24]).
 |
Figure 9 Forest plot of network effect sizes between different frequencies of SCS for global pain.
|
 |
Figure 10 Forest plot of network effect sizes between different frequencies of SCS for back pain.
|
 |
Figure 11 Forest plot of network effect sizes between different frequencies of SCS for leg pain.
|
 |
Figure 12 Forest plot of network effect sizes between different frequencies of SCS for responder rate.
|
 |
Figure 13 Forest plot of network effect sizes between different frequencies of SCS for ODI.
|
 |
Figure 14 Forest plot of network effect sizes between different frequencies of SCS for EQ-5D.
|
Additionally, we present league tables (Tables S20–S25) and plots of average expected rankings alongside SUCRA cumulative probabilities for each treatment regimen (Figures 15–20). SUCRA values (range: 0–1) provide an objective intervention ranking by quantifying relative treatment superiority, with higher scores indicating greater likelihood of being among the most effective options. Specifically, SUCRA = 1.0 (100%) denotes the theoretically optimal treatment (consistent first rank), while 0 (0%) represents the least favorable option (consistent last rank); a value of 0.80 indicates an 80% probability of superiority over alternative regimens. Extended results are available in Figures S9–S14 and Tables S26–S31.
 |
Figure 15 Plots of SUCRA values for each SCS to global pain.
|
 |
Figure 16 Plots of SUCRA values for each SCS to back pain.
|
 |
Figure 17 Plots of SUCRA values for each SCS to leg pain.
|
 |
Figure 18 Plots of SUCRA values for each SCS to responder rate.
|
 |
Figure 19 Plots of SUCRA values for each SCS to ODI.
|
 |
Figure 20 Plots of SUCRA values for each SCS to EQ-5D.
|
Global Pain Relief
Seven studies (583 patients) in total measured overall pain before and after different SCS treatments. All of which underwent NMA, and Figure 3 shows a network diagram of all eligible comparative endpoints. All 5 treatment frequencies(nodes) were connected to the network and no network inconsistencies were found [χ2 (3)=2.12, P=0.5469]. Figure 15 shows the expected mean rankings and SUCRA cumulative probability plots for each intervention. According to the SUCRA cumulative ranking plot, for global pain, the efficacy of SCS in patients with FBSS was ranked as follows: subperception SCS (64.0%), LF SCS (59.7%), burst SCS (59.6%), HF SCS (50.0%), and sham (16.7%). The reticular forest plot for two-by-two comparisons (Figure 9) showed no significant differences between the various SCSs.
Back Pain Relief
Five studies (730 patients) measured back pain before and after treatment with different SCSs. Figure 4 shows the network plot of all eligible comparisons for this endpoint. All 5 administrative modalities (nodes) were connected to the network, and no network inconsistencies were observed [χ2 (3)=1.41, P=0.7032]. Figure 16 shows the anticipated average rankings and SUCRA cumulative probability plots for each intervention, according to which, for back pain, the efficacy rankings were in descending order, HF SCS (99.7%), subperception SCS (64.6%), burst SCS (47.2%), LF SCS (26.3%), and sham (12.2%). In accordance with the reticulated forest plot, HF SCS was superior to other SCSs in relieving back pain, and the difference was statistically significant (Figure 10).
Leg Pain Relief
A total of 4 studies (658 patients) measured leg pain scores before and after different SCS treatments. The network plot of all eligible comparisons for this endpoint is shown in Figure 5. All 5 administrative modalities (nodes) were connected to the network, and no network inconsistency was observed [χ2 (1)=0.03, P=0.8633]. The expected mean rankings and SUCRA cumulative probability plots for each intervention are shown in Figure 17. According to the probability ranking graph SUCRA cumulative probability plot, it is known that for leg pain, the efficacy rankings were HF SCS (93.2%), subperception SCS (68.4%), LF SCS (40.6%), burst SCS (33.3%), and sham (14.6%) in that order. According to the reticulated forest plot, HF relieved leg pain better than traditional LF SCS, and the difference was statistically significant (Figure 11).
Responder Rate
A total of 4 studies (400 patients) measured responder rates before and after treatment with different SCSs. Figure 6 displays a network plot that covers all eligible comparisons made for this endpoint. All 4 administrative modalities (nodes) were connected to the network, and no network inconsistency was observed [χ2 (1)=0.00, P=0.9648]. The expected mean rankings and SUCRA cumulative probability plots for each intervention are shown in Figure 18. According to the probability ranking plot SUCRA cumulative probability plot, the qualitative analysis of SCS for pain relief in FBSS patients in terms of response rate showed that the efficacy rankings were subperception SCS (75.3%), HF SCS (69.6%), burst SCS (44.2%), and LF SCS (10.8%) in that order. According to the reticulated forest plot (Figure 12), Patients with subperception SCS had a higher level of pain relief response than those with LF SCS.
ODI Improvement
A total of 4 studies (360 patients) measured ODI scores before and after treatment with different SCSs. The network plot of all eligible comparisons for this endpoint is shown in Figure 7. All 5 management modalities (nodes) were connected to the network, and no network inconsistency was observed [χ2(2)=3.16, P=0.1646]. The expected mean rankings and SUCRA cumulative probability plots for each intervention are shown in Figure 19. According to the probability ranking plot SUCRA cumulative probability plot shows that the order ranking of SCSs for improvement of lumbar function in FBSS patients was HF SCS (85.0%), burst SCS (64.4%), LF SCS (64.4%), subperception SCS (30.8%), and sham (5.1%) in that order. None of the differences were statistically different according to the reticular forest plot (Figure 13).
EQ-5D Improvement
A total of 4 studies measured A total of 4 studies (618 patients) measured EQ-5D scores before and after different SCS treatments. The network plot of all eligible comparisons for this endpoint is shown in Figure 8. All 5 administrative modalities (nodes) were connected to the network, and no network inconsistency was observed [χ2 (1)=1.90, P=0.1681]. The expected mean rankings and SUCRA cumulative probability plots for each intervention are shown in Figure 20. According to the probability ranking plot SUCRA cumulative probability plot shows that the effect of SCS on the improvement of healthy quality of life of FBSS patients in the order of subperception SCS (80.3%), LF SCS (64%), burst SCS (48.7%), sham (29%), and HF SCS (28%) in descending order. None of the differences were statistically different according to the forest plot of the two-by-two comparison (Figure 14).
Discussion
Our network meta-analysis comprehensively evaluated the efficacy of different SCS frequencies in FBSS. This study addresses a notable literature gap by distinctly comparing traditional low-frequency SCS with subperception SCS—modalities frequently conflated in prior analyses despite evidence suggesting divergent mechanisms of action and clinical outcomes. Results suggest subperception SCS may be associated with superior global pain improvement, with higher rates of ≥50% pain responders versus traditional low-frequency SCS (P < 0.05).
Current research indicates that SCS can significantly alleviate pain in patients with FBSS and improve their quality of life. Kurt et al confirmed the benefits of SCS from multiple perspectives,7 while Nissen et al conducted a 20-year follow-up study of 224 patients, demonstrating the sustained therapeutic effects of SCS.37 However, with the increasing variety of SCS methods available, selecting the appropriate option has become a challenge. This study compared the efficacy of different frequencies of SCS commonly used in patients with FBSS to provide evidence for clinical selection.
In Morales and Karri’s systematic reviews, it was shown that HF SCS provided better pain relief compared to low-frequency SCS.10,38 Our analysis reveals that HF SCS is potentially the most effective option for relieving back and leg pain. SUCRA values and statistical significance indicate its superiority in these areas: 99.7% for back pain (P < 0.05 vs comparators; consistency χ² = 1.41, P = 0.703) and 93.2% for leg pain (P < 0.05 vs low-frequency SCS). Conversely, subperception SCS exhibits potential in overall pain management and quality of life enhancement. SUCRA analysis showed that subperception SCS (500–1200 Hz) had the highest probability of providing global pain relief (64.0%), a responder rate of ≥50% (75.3%), and an improvement in health-related quality of life (EQ-5D, 80.3%). Therefore, subperception SCS may be clinically applicable in managing overall pain and improving quality of life.
As SCS technology evolves, the emergence of novel modalities provides a broader spectrum of options for clinicians and patients alike. To our knowledge, this study is the most comprehensive and up-to-date systematic review and network meta-analysis, covering almost the entire range of available data on the frequency of SCS usage for patients with failed back surgery syndrome. By rigorously applying the CINeMA quality assessment approach, we enhanced the credibility of the findings. This network meta-analysis lays the groundwork for such inquiries and emphasizes the need for further exploration into the optimal use of SCS in FBSS management.
However, our study has some limitations. First, the included trials had different follow-up periods. We conducted a subgroup analysis and a sensitivity analysis. For the former, we divided the stimulation follow-up into long-term (one month) and short-term (less than one month) subgroups (see Table S32 for details). For the latter, we compared all included studies with those that had a follow-up period of up to three months (see Figure S15 for details). Although these variations in follow-up duration did not alter the overall conclusions, they may have introduced heterogeneity and influenced the interpretation of the results. Second, some of the included studies did not exclusively involve FBSS patients. These studies were included for the following reasons: First, due to limitations in sample size, studies that included only FBSS patients lacked key data, such as VAS/ODI scores. Second, in studies that did not include only FBSS patients, the primary study population was still FBSS patients, accounting for 45–77% of the total. The remainder were patients with CRPS and other SCS indications. Since the mechanisms and SCS response profiles were consistent with FBSS, we included these studies to enhance statistical power. However, the effect estimates may be imprecise, which limits the overall accuracy and generalizability. Additionally, despite our efforts to contact the authors, some trials lacked sufficient information to accurately assess randomization, task concealment, and investigator blinding. Additionally, SCS trials may be biased due to non-standardized blinding, open-label designs, and manufacturer funding. Results from different studies may also be inconsistent.39,40 In some cases, this results in a lower level of evidence. Additionally, limited literature exists on certain SCS frequencies, resulting in less precise effect estimates compared to other frequencies. Therefore, these relative efficacy estimates must be confirmed by directly comparing SCS modes in randomized controlled trials with adequate sample sizes.
Conclusion
This NMA suggests potential differences in therapeutic profiles among SCS modalities for FBSS. HF-SCS (10 kHz) showed relatively higher SUCRA values for back pain (99.7%), leg pain (93.2%), and disability improvement (ODI 85.0%). Subperception SCS (500–1200 Hz) was associated with better probability for global pain relief (64.0%), ≥50% pain reduction (75.3%), and HRQoL outcomes (EQ-5D 80.3%). These relative efficacy estimates require confirmation in adequately powered RCTs directly comparing SCS modalities.
Acknowledgments
We thank all the authors of the clinical trials and the support of Kunming Medical University joint project.
Funding
This study received partial support from the Kunming Medical University Joint Project (No. 202301AY070001-078). The research team members maintained an independent relationship with the funding agencies. The funders did not participate in or influence any part of the study, including its design and implementation, data collection, management, analysis, and interpretation, as well as the writing, review, and approval of the manuscript.
Disclosure
All authors report no conflicts of interest in this work; no financial relationships with any organizations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.
References
1. Van Buyten JP, Linderoth B. “The failed back surgery syndrome”: definition and therapeutic algorithms – an update. Eur J Pain Suppl. 2010;4:273–286. doi:10.1016/j.eujps.2010.09.006
2. Burton CV. Failed back surgery patients: the alarm bells are ringing. Surg Neurol. 2006;65(1):5–6. doi:10.1016/j.surneu.2005.08.018
3. Hoelzer BC, Edgar D, Lu SP, et al. Indirect comparison of 10 kHz Spinal Cord Stimulation (SCS) versus traditional low-frequency SCS for the treatment of painful diabetic neuropathy: a systematic review of randomized controlled trials. Biomedicines. 2022;10(10):2630. doi:10.3390/biomedicines10102630
4. North JM, Hong KJ, Cho PY. Clinical outcomes of 1 kHz subperception spinal cord stimulation in implanted patients with failed paresthesia-based stimulation: results of a prospective randomized controlled trial [J]. Neuromodulation. 2016;19(7):731–737. doi:10.1111/ner.12441
5. Hou S, Kemp K, Grabois M. A systematic evaluation of burst spinal cord stimulation for chronic back and limb pain [J]. Neuromodulation. 2016;19(4):398–405. doi:10.1111/ner.12440
6. Taylor RS. Spinal cord stimulation in complex regional pain syndrome and refractory neuropathic back and leg pain/failed back surgery syndrome: results of a systematic review and meta-analysis. J Pain Symptom Manag. 2006;31(4 Suppl):S13–9. doi:10.1016/j.jpainsymman.2005.12.010
7. Kurt E, Noordhof RK, van Dongen R, et al. Spinal cord stimulation in failed back surgery syndrome: an integrative review of quantitative and qualitative studies. Neuromodulation. 2022;25(5):657–670. doi:10.1016/j.neurom.2021.11.013
8. Turner RM, Davey J, Clarke MJ, et al. Predicting the extent of heterogeneity in meta-analysis, using empirical data from the cochrane database of systematic reviews. Int J Epidemiol. 2012;41(3):818–827. doi:10.1093/ije/dys041
9. Sammak SE, Mualem W, Michalopoulos GD, et al. Rescue therapy with novel waveform spinal cord stimulation for patients with failed back surgery syndrome refractory to conventional stimulation: a systematic review and meta-analysis. J Neurosurg-Spine. 2022;37(5):670–679. doi:10.3171/2022.4.SPINE22331
10. Morales A, Yong RJ, Kaye AD, et al. Spinal cord stimulation: comparing traditional low-frequency tonic waveforms to novel high frequency and burst stimulation for the treatment of chronic low back pain. Curr Pain Headache R. 2019;23(4):25. doi:10.1007/s11916-019-0763-3
11. Schu S, Slotty PJ, Bara G, von Knop M, Edgar D, Vesper J. A prospective, randomised, double-blind, placebo-controlled study to examine the effectiveness of burst spinal cord stimulation patterns for the treatment of failed back surgery syndrome. Neuromodulation. 2014;17(5):443–450. doi:10.1111/ner.12197
12. North J, Loudermilk E, Lee A. Outcomes of a multicenter, prospective, crossover, randomized controlled trial evaluating subperception spinal cord stimulation at ≤1.2 kHz in previously implanted subjects. Neuromodulation. 2020;23(1):102–108. doi:10.1111/ner.13015
13. Rigoard P, Ounajim A, Moens M, et al. Should we oppose or combine waveforms for spinal cord stimulation in PSPS-T2 patients? A prospective randomized crossover trial (MULTIWAVE Study). J Pain. 2023;24(12):2319–2339. doi:10.1016/j.jpain.2023.07.015
14. Sokal P, Malukiewicz A, Kierońska S, et al. Sub-perception and supra-perception spinal cord stimulation in chronic pain syndrome: a randomized, semi-double-blind, crossover, placebo-controlled trial. J Clin Med. 2020;10(1):9. doi:10.3390/jcm10010009
15. Winter C, Kosch R, Ludlow M, et al. Network meta-analysis correlates with analysis of merged independent transcriptome expression data. BMC Bioinf. 2019;20(1):144. doi:10.1186/s12859-019-2705-9
16. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Int J Surg. 2021;88:105906. doi:10.1016/j.ijsu.2021.105906
17. Kumar K, North R, Taylor R, et al. Spinal cord stimulation vs. conventional medical management: a prospective, randomized, controlled, multicenter study of patients with failed back surgery syndrome (PROCESS study). Neuromodulation. 2005;8(4):213–218. doi:10.1111/j.1525-1403.2005.00027.x
18. Lee CP, Fu TS, Liu CY, et al. Psychometric evaluation of the Oswestry Disability Index in patients with chronic low back pain: factor and mokken analyses. Health Qual Life Outcomes. 2017;15(1):192. doi:10.1186/s12955-017-0768-8
19. Payakachat N, Ali MM, Tilford JM. Can the EQ-5D detect meaningful change? A systematic review. Pharmacoeconomics. 2015;33(11):1137–1154. doi:10.1007/s40273-015-0295-6
20. Higgins JPT, Sterne JAC, Savović J, et al. A revised tool for assessing risk of bias in randomized trials. Cochrane Database Syst Rev. 2016;10 Suppl 1:29–31.
21. Follmann D, Elliott P, Suh I, et al. Variance imputation for overviews of clinical trials with continuous response. J Clin Epidemiol. 1992;45(7):769–773. doi:10.1016/0895-4356(92)90054-Q
22. Papakonstantinou T, Nikolakopoulou A, Higgins JPT, et al. CINeMA: software for semiautomated assessment of theconfidence in the results of network meta-analysis. Campbell SystRev. 2020;16:e1080.
23. White IR. Multivariate random-effects meta-regression: updates to mvmeta. Stata J. 2011;11:255–270. doi:10.1177/1536867X1101100206
24. Rücker G. Network meta-analysis, electrical networks and graph theory. Res Synth Methods. 2012;3:312–324. doi:10.1002/jrsm.1058
25. Wiebe N, Vandermeer B, Platt RW, et al. A systematic review identifies a lack of standardization in methods for handling missing variance data. J Clin Epidemiol. 2006;59(4):342–353. doi:10.1016/j.jclinepi.2005.08.017
26. Chawla N, Anothaisintawee T, Charoenrungrueangchai K, et al. Drug treatment for panic disorder with or without agoraphobia: systematic review and network meta-analysis of randomised controlled trials. BMJ. 2022;376:e066084. doi:10.1136/bmj-2021-066084
27. Langan D, Higgins JPT, Jackson D. A comparison of heterogeneity variance estimators in simulated random-effects meta-analyses. Res Synth Method. 2019;10(1):83–98. doi:10.1002/jrsm.1316
28. van Valkenhoef G, Dias S, Ades AE, et al. Automated generation of node-splitting models for assessment of inconsistency in network meta-analysis. Res Synth Methods. 2016;7(1):80–93. doi:10.1002/jrsm.1167
29. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol. 2005;5:13. doi:10.1186/1471-2288-5-13
30. Perez C, Rojo E, Margarit C, et al. 24-month real-world study of spinal cord stimulation in failed back surgery patients with refractory pain. Pain Physician. 2021;24(6):479–488.
31. De Andres J, Monsalve-Dolz V, Fabregat-Cid G, et al. Prospective, randomized blind effect-on-outcome study of conventional vs high-frequency spinal cord stimulation in patients with pain and disability due to failed back surgery syndrome. Pain Med. 2017;18(12):2401–2421. doi:10.1093/pm/pnx241
32. Kapural L, Yu C, Doust MW, et al. Comparison of 10-kHz high-frequency and traditional low-frequency spinal cord stimulation for the treatment of chronic back and leg pain: 24-month results from a multicenter, randomized, controlled pivotal trial. Neurosurgery. 2016;79(5):667–677. doi:10.1227/NEU.0000000000001418
33. Breel J, Wille F, Wensing AGCL, et al. A comparison of 1000 Hz to 30 Hz spinal cord stimulation strategies in patients with unilateral neuropathic leg pain due to failed back surgery syndrome: a multicenter, randomized, double-blinded, crossover clinical study (HALO). Pain Ther. 2021;10(2):1189–1202. doi:10.1007/s40122-021-00268-7
34. Hara S, Andresen H, Solheim O, et al. Effect of spinal cord burst stimulation vs placebo stimulation on disability in patients with chronic radicular pain after lumbar spine surgery: a randomized clinical Trial. JAMA. 2022;328(15):1506–1514. doi:10.1001/jama.2022.18231
35. Eldabe S, Duarte R, Gulve A, et al. Analgesic efficacy of “burst” and tonic (500 Hz) spinal cord stimulation patterns: a randomized placebo-controlled crossover study. Neuromodulation. 2021;24(3):471–478. doi:10.1111/ner.13321
36. Al-Kaisy A, Palmisani S, Pang D, et al. Prospective, randomized, sham-control, double blind, crossover trial of subthreshold spinal cord stimulation at various kilohertz frequencies in subjects suffering from failed back surgery syndrome (SCS frequency study). Neuromodulation. 2018;21(5):457–465. doi:10.1111/ner.12771
37. Nissen M, Ikäheimo T-M, Huttunen J, Leinonen V, von Und Zu Fraunberg M. Long-term outcome of spinal cord stimulation in failed back surgery syndrome: 20 years of experience with 224 consecutive patients. Neurosurgery. 2019;84(5):1011–1018. doi:10.1093/neuros/nyy194
38. Karri J, Orhurhu V, Wahezi S, et al. Comparison of spinal cord stimulation waveforms for treating chronic low back pain: systematic review and meta-analysis. Pain Phys. 2020;23:451–460.
39. O’Connell NE, Ferraro MC, Gibson W, et al. Implanted spinal neuromodulation interventions for chronic pain in adults. Cochrane Database Syst Rev. 2021;12(12):CD013756. doi:10.1002/14651858.CD013756.pub2
40. Knotkova H, Hamani C, Sivanesan E, et al. Neuromodulation for chronic pain. Lancet. 2021;397(10289):2111–2124. doi:10.1016/S0140-6736(21)00794-7