Learn · Module
The evidence, including the parts the brochures omit
Every therapy on this site rests on published trials. Those trials are real, and most are positive — but each has a design, a population and a set of limitations that determine what it can honestly be said to prove. Both halves are given equal weight here.
What You'll Learn
Approx. 14 min read- Which trials established each therapy
- What 'responder rate' actually measures
- Why blinding is unusually difficult in this field
- How to read an industry-sponsored study fairly
- Why cross-manufacturer comparisons barely exist
- What the evidence does not yet answer
How to Read This Field
Four things that apply to nearly every study below
Almost everything here is industry-sponsored
Neuromodulation trials are expensive and the manufacturers fund them. That does not make the results false, but it does mean the comparator, the endpoint and the timing of publication were all chosen by a party with an interest in the outcome. Read every headline number with that in mind.
Blinding is genuinely hard, not merely neglected
You cannot easily blind a patient to a therapy that makes their leg tingle. EVOKE's closed-loop versus open-loop comparison used the same device in both arms to support blinding. Earlier double-blind SCS trials also exist, including sham-controlled studies; neither approach answers every head-to-head question.
The 50% responder threshold hides a lot
A patient who goes from 8/10 to 5/10, sleeps through the night and returns to work is not a responder by this definition. A patient who goes from 4/10 to 2/10 with no functional change is. Responder rates make trials comparable; they do not describe lives.
Head-to-head comparisons between manufacturers barely exist
Each company compares its new therapy to conventional tonic stimulation, not to its competitors' current systems. So the claim 'superior to traditional SCS' is true for several platforms simultaneously and tells you almost nothing about how they compare to each other.
The Studies
Landmark trials, in chronological order
Each entry gives the design, the population, the question, the finding — and then the limitation, in the same amount of detail.
Melzack & Wall — Gate Control Theory
Foundational1965- Design
- Theoretical paper, Science
- Population
- n/a
The question
Why does rubbing an injury reduce the pain, and can pain transmission be modulated within the spinal cord itself?
What it found
Proposed that activity in large-diameter sensory fibres can inhibit transmission of pain signals carried by smaller fibres at the level of the dorsal horn, with descending control also influencing the 'gate'. Shealy and colleagues reported an early clinical dorsal-column stimulation case in 1967.
Limitations
A theoretical framework, not a clinical test of spinal cord stimulation. The original model is incomplete and does not establish how later paresthesia-free waveforms work. It remains historically foundational rather than mechanistically sufficient.
PROCESS
SCS2007- Design
- Randomised controlled trial
- Population
- 100 patients with neuropathic radicular leg pain after spine surgery
The question
Does spinal cord stimulation plus conventional medical management outperform conventional medical management alone?
What it found
At six months, 24 of 50 assessed stimulation patients (48%) achieved at least 50% leg pain relief versus four of 44 assessed medical-management patients (9%). Function and quality of life also favoured stimulation. Substantial crossover occurred after this six-month comparison.
Limitations
Unblinded — patients knew which arm they were in, and stimulation at the time produced paresthesia, making blinding effectively impossible. Crossover after six months complicates the longer-term comparison.
SENZA-RCT
SCS2015–2016- Design
- Prospective randomised controlled trial
- Population
- 198 patients with chronic back and leg pain randomised; 171 implanted
The question
Is 10 kHz stimulation non-inferior — or superior — to traditional tonic stimulation?
What it found
At three months, 84.5% of assessed 10 kHz recipients had at least 50% back-pain relief versus 43.8% with tonic stimulation. An advantage in back and leg pain was reported through 12 months; a separate 24-month report also found an advantage among participants assessed at that time. These results do not describe all 198 randomised patients.
Limitations
Not blinded, since the two therapies feel different — one produces tingling and one does not. Only 171 of 198 randomised patients were implanted; longer follow-up depends on available assessments. Industry-sponsored. The tonic comparator used a fixed programming protocol that some argue did not represent best available tonic practice.
ACCURATE
DRG2017- Design
- Prospective randomised comparative trial
- Population
- 152 patients with CRPS or causalgia of the lower extremity
The question
Does DRG stimulation outperform conventional spinal cord stimulation in focal limb pain?
What it found
At three months, 56 of 69 assessed DRG patients (81%) met the trial's composite treatment-success endpoint versus 39 of 70 assessed SCS patients (56%). An advantage was also reported at 12 months. DRG produced more targeted coverage with less postural variation in the study assessment.
Limitations
Treatment success required a successful trial, permanent implant, at least 50% pain relief and no stimulation-related neurological deficit; the reported percentages are not of all 152 randomised patients. Unblinded and limited to lower-extremity CRPS or causalgia, not axial back pain. SCS position adjustment was disabled for the postural comparison. Industry-sponsored.
SUNBURST
SCS2017–2018- Design
- Randomised crossover trial
- Population
- 100 randomised patients with chronic trunk and/or limb pain after a successful tonic-stimulation trial
The question
Does burst stimulation outperform tonic stimulation when the same patients try both?
What it found
Burst was non-inferior and then superior to tonic stimulation on mean daily overall pain scores. At the crossover preference assessment, 70.8% of participants with recorded preferences chose burst; absence of tingling and better pain relief were nearly tied as their reasons.
Limitations
The preference percentage is not of all 100 randomised patients. Crossover designs can carry order effects. Patients could tell the two waveforms apart, so blinding was not achievable. Industry-sponsored.
ReActiv8-B
Restorative2021–2024- Design
- Prospective randomised, double-blind, sham-controlled trial
- Population
- 204 implanted patients with chronic mechanical low back pain and impaired multifidus control
The question
Does restorative neurostimulation of the lumbar multifidus improve chronic low back pain versus sham stimulation?
What it found
At 120 days, the prespecified endpoint of at least 30% back-pain relief without increased analgesic use was reached by 57.1% with therapeutic stimulation and 46.6% with sham; the difference was not statistically significant (p=0.138), so the primary superiority endpoint was not met. Supportive secondary outcomes favoured treatment, and uncontrolled follow-up has reported sustained improvements among participants retained through five years.
Limitations
The sham group crossed to active therapy after 120 days, leaving no long-term randomised control. Five-year data were available for 126 of 204 implanted participants, creating attrition and survivor-bias concerns. Trial eligibility was narrower than the FDA indication, and the study was manufacturer-funded.
SENZA-PDN
SCS2021- Design
- Prospective randomised controlled trial
- Population
- 216 patients with refractory painful diabetic neuropathy
The question
Does 10 kHz stimulation added to conventional medical management improve pain in diabetic neuropathy?
What it found
At three months, the primary endpoint of at least 50% pain relief without worsening neurological deficit was met by 75 of 95 assessed 10 kHz stimulation patients (79%) versus five of 94 assessed medical-management patients (5%). Improved neurological examinations at six months were reported separately. A painful-diabetic-neuropathy indication was subsequently approved for the Senza system.
Limitations
Unblinded; the neurological-examination finding was separate from the primary pain endpoint and subject to examiner variability. Most eligible medical-management patients crossed over after six months, weakening later controlled comparisons. Industry-sponsored.
EVOKE
SCS2020–2024- Design
- Double-blind randomised controlled trial
- Population
- 134 patients with chronic back and leg pain
The question
Does closed-loop stimulation, adjusting output based on measured spinal cord response, outperform open-loop stimulation?
What it found
In a double-blind comparison of closed-loop versus open-loop stimulation, 49 of 59 assessed closed-loop recipients achieved at least 50% pain reduction without increased pain medication at 12 months, versus 36 of 59 with open-loop stimulation. Separate 24- and 36-month reports described a sustained advantage. The original trial appeared in Lancet Neurology.
Limitations
The 12-month responder comparison included 59 participants per arm, rather than all 134 randomised patients. Single-manufacturer technology, so results cannot be assumed to transfer to other closed-loop implementations. Industry-sponsored. Both arms used the same device, which enabled blinding but limits comparison against other platforms.
SPRINT PNS — 60-day temporary stimulation
PNS2019–2023- Design
- Post-amputation randomised study with uncontrolled follow-up; separate retrospective safety analysis
- Population
- Adults with post-amputation pain in the randomised study; other peripheral targets are separate research populations
The question
Can a 60-day course of peripheral nerve stimulation produce relief that outlasts the therapy?
What it found
The post-amputation study reported benefit during its short randomised comparison. At 12 months, six of nine assessed participants from the original active-treatment group had at least 50% pain reduction; there was no concurrent sham comparison at that point. A separate real-world analysis described records from 6,160 treated patients.
Limitations
Only nine original active-treatment participants were assessed at 12 months, so durability across patients and other pain conditions remains uncertain. The large real-world analysis used retrospective manufacturer records, not prospective safety surveillance. Central nervous system 'reconditioning' is a hypothesis; a 60-day implant limit is not proof that later treatment courses are prohibited.
SENZA-NSRBP
SCS2022- Design
- Prospective randomised controlled trial
- Population
- 159 randomised patients with refractory back pain, no prior spine surgery and surgeon-confirmed nonsurgical candidacy
The question
Does 10 kHz stimulation help people with refractory back pain who have not had spine surgery and are not surgical candidates?
What it found
At three months, 55 of 68 assessed implanted stimulation patients (80.9%) achieved at least 50% pain relief versus one of 75 assessed medical-management patients (1.3%). These are assessed-patient results, not an all-randomised comparison. The Senza system subsequently received an expanded FDA indication for nonsurgical refractory back pain.
Limitations
Unblinded; the comparison was among assessed patients rather than all 159 randomised participants. Optional crossover after six months limits longer-term comparisons. Surgeon determinations of nonsurgical candidacy can vary across practices.
COMFORT
PNS2026- Design
- Multicentre randomised trial with a three-month control phase; uncontrolled follow-up after crossover
- Population
- Patients with chronic shoulder, low-back, knee or foot/ankle pain treated with a battery-free PNS system
The question
Do outcomes with a miniaturised, externally powered system hold up over two years?
What it found
At 24 months, 39 of 46 assessed participants (85%) had at least 50% pain reduction and mean pain scores fell 67% from baseline. These results combine the original treatment group with participants who crossed over from conventional medical management; they are not a two-year randomised comparison.
Limitations
The control period lasted only three months. At 24 months, 46 of 68 participants in the modified intention-to-treat group were assessed, with no continuing randomised comparator; attrition and missing data limit certainty. The study was not double-blinded and several authors disclosed manufacturer ties. Therapy requires a wearable external disc.
What The Evidence Does Not Answer
The honest gaps
There is no adequately powered trial telling you which manufacturer's system will work best for you specifically. There is no reliable way, before a trial, to predict who will be a responder. Long-term data beyond five years is thin for every platform, and the oldest data comes from technology nobody implants any more.
Explant rates — how many people eventually have the device removed, and why — are reported inconsistently and are almost certainly higher in real-world practice than in trial cohorts. Anyone who quotes you a single confident figure is over-reading the data.
This is precisely why the trial period exists. It substitutes a week of your own experience for a statistical prediction, which is a far better test than any published responder rate.
Evidence informs a decision. It does not make one.
These summaries are educational. They are not a substitute for a discussion with a physician who knows your history, and none of them should be read as a promise about what a given therapy will do for you.
Put the evidence in context
Read the professional literature review, then discuss these findings in light of your history, imaging and goals with your own clinician.