Need help? Call us on
+44 (0)141 628 8222What should we actually tell patients about electrical stimulation and spasticity? A modality-by-modality audit of the current evidence: the effect sizes, the dose thresholds, the negative trials, and the true state of tSCS.
A question I hear regularly from physiotherapists, and increasingly from case managers, runs something like this: a patient has read that electrical stimulation can help their spasticity, and they want to know whether it is true. What should I tell them?
It isn't easy to give a general answer, because this is a corner of the literature where a confident summary in either direction misrepresents the state of play. We have just published a plain-language guide to spasticity and spasms for patients on our companion site, Spasticity and Muscle Spasms: A Plain English Guide for People Living With Them. This article is its clinical companion: the same evidence base, examined at the level of effect sizes, dose thresholds and trial design, so you can calibrate what you say in the clinic. It also complements our earlier piece on sequencing, Spasticity as the Rate-Limiting Step, which covered how to order tSCS and functional practice within a session. Here, the question is narrower and more fundamental, phrased as: What does the evidence actually show, modality by modality?
Any audit of this literature has to begin with the uncomfortable observation that we do not fully agree on what we are measuring. The 2025 IFCN handbook chapter on spasticity evaluation states plainly that there is no generally accepted definition of the term, which is used as an umbrella term for hyperexcitable stretch reflexes, clonus, spasms, cramps, and co-contraction, each with different underlying mechanisms and implications for therapy.
The measurement tools compound the problem. The Modified Ashworth Scale remains the near-universal primary outcome despite well-known limitations, and the instruments do not always agree with each other. In the UCL meta-analysis discussed below, electrical stimulation produced significant improvements on the MAS but no significant effect on the pendulum test. More recently, the Shepherd Center group reported that stretch-induced spasticity measured biomechanically did not track with electrophysiological testing in the same participants. When two respectable outcome measures move independently, modest pooled effects should be read with humility, in both directions. It seems to me that some real effects will be missed, and some reported effects will not survive better measurement.
For practice, the implication is simply that whatever modality you trial, define the outcome that matters to the patient (night spasms, transfer interference, tolerated stretch, orthosis fit) and measure that, alongside whatever scale you choose to use.
For sensory-level stimulation, the stroke literature is the most developed. Two independent meta-analyses, published seven years apart, both landed on a pooled standardised mean difference of about 0.64 in favour of TENS as an adjunct to physiotherapy, which is a moderate effect. The earlier of the two concluded that benefit was seen when TENS was applied for more than 30 minutes over the nerve trunk or muscle belly in chronic stroke. The 2025 analysis found significant improvement at 100 Hz where lower frequencies did not reach significance, although the direct between-frequency comparison was itself non-significant (p = 0.67) and Egger's test suggested publication bias, so I would resist quoting 100 Hz as established rather than sensible.
The corrective comes from a 2026 network meta-analysis of 38 randomised trials of peripheral somatosensory stimulation after stroke, which put TENS near the bottom of the active modalities with a mean difference of only 0.18 points on the MAS, against 0.45 for NMES. The two figures are not contradictory (different comparators, different pooling), but they illustrate the range: the effect you quote for TENS depends heavily on which analysis you cite.
In spinal cord injury, the direct evidence is a 10-person, double-blind crossover study in which a single 30-minute session of either TENS or FES reduced hip adductor and knee extensor spasticity, with carryover lasting up to 4 hours. In multiple sclerosis, a 2024 systematic review of 32 randomised trials of non-pharmacological interventions rated every included study as methodologically weak, and graded the certainty of evidence for TENS in MS as very low. Worth knowing before a patient spends money on equipment.
For motor-level stimulation, the post-stroke network meta-analysis above places NMES among the better options among peripheral interventions, with a mean difference of 0.45 on the MAS. NICE's 2023 stroke rehabilitation guideline (NG236, recommendation 1.15.6) says clinicians should consider a trial of NMES, FES or TENS for focal spasticity, which is the strongest UK anchor this field has. It is not a recommendation that it works for everyone, but a guideline instruction that a properly conducted trial of therapy is reasonable.
For FES cycling in spinal cord injury, two reviews between them frame the defensible range. A 2021 meta-analysis found a statistically significant reduction in spasticity, with the interesting caveat that a subgroup analysis identified the effect in studies with more than 20 training sessions; meta-regression found no linear dose-response, suggesting a threshold effect rather than a gradient. A 2025 systematic review of sixteen studies (203 participants) nevertheless described the overall evidence as inconclusive, noting that the long-term effects appeared in the moderate-to-high-quality studies using roughly 60-minute sessions, three times weekly, for sixteen weeks, and that optimal frequency, amplitude and pulse width remain unestablished. An earlier review made the quality point bluntly: median PEDro score 3 of 10.
My reading, which I offer as clinical experience (of delivering FES Cycling systems over 20 years) alongside the evidence rather than as proof, is that the threshold finding matches what we see in practice. Patients rarely report meaningful changes in spasticity in the first fortnight, and those who train consistently beyond a month usually do. If you are using FES cycling for spasticity as well as for exercise benefits, set expectations for a long enough duration to cross the threshold identified in the literature, and communicate this to the patient at the start.
One finding cuts across all of this and deserves more attention than it gets. The UCL meta-analysis of 29 studies on spinal cord injury found that producing a visible muscle contraction was not necessary for an anti-spastic effect: sensory-level modalities also reduced spasticity. The authors' interpretation, that activation of afferent fibres is the common active ingredient, fits the mechanism work in tSCS discussed below, and it has two practical consequences. First, patient tolerance need not be sacrificed to intensity when spasticity is the target. Second, the ubiquitous patient anecdote, "I bought a TENS unit and it did nothing", tells you a lot about ignorance of parameters and application, not about whether stimulation can help them.
Transcutaneous spinal cord stimulation is where the claims currently outpace the trials.
The foundational work is coherent. The Vienna group showed that 30 minutes of tSCS at 50 Hz (1 ms biphasic pulses, paraspinal electrode over T11 to T12, abdominal returns, intensity around 90% of the posterior root-muscle reflex threshold) reduced MAS scores, clonus and spasms in spinal cord injury, with all measures improved at two hours and large effect sizes; their title says temporary, and they meant it.
In 2024 the same group published mechanistic work in Cell Reports Medicine showing that deficient presynaptic and postsynaptic spinal inhibition underlies spinal spasticity and that tSCS transiently restores both towards normative levels. As mechanisms in this field go, that is unusually well demonstrated. An exploratory single-arm study in MS reported similar effects lasting around two hours, with clinically assessed hypertonia still improved at 24 hours.
Then the controlled trials. In 2025, the first randomised sham-controlled crossover trial of this exact protocol for spasticity, in progressive MS, reported an effect size of 0.25 on the primary MAS outcome with p = 0.12: more responders under real stimulation (10 of 16 versus 7 of 16), but not significant, and negligible effects on gait. Two disclosure points belong in any clinical summary of that trial. First, its authors include people connected to SensorStim Neurotechnology, the Berlin company behind the Stim2Go device we supply; a company-adjacent group publishing a result that does not flatter the technology is a point in the literature's favour, and it should be cited, not buried. Second, a follow-on case series of repeated sessions (twice weekly for four weeks) found moderate-to-large effect sizes that again did not reach significance in 9 participants, with effects persisting at 1 week. As researchers like to say, more research is needed!!
The spinal cord injury picture is similarly mixed at trial level: a randomised sham-controlled study in subacute SCI found no significant spasticity change; a 17-person crossover found no group-level effect, with benefit confined to participants with high baseline quadriceps spasticity and one dual-site condition significantly increasing soleus spasticity in those who started with little. That last finding is the clearest argument I know for individual assessment and measurement over standard settings. The systematic reviews concur: evidence for effectiveness is graded as limited, and, by GRADE, very low certainty for clinical outcomes, with moderate certainty for safety, with adverse events confined to skin irritation and tingling.
So the clinical summary I suggest for tSCS is this. It is mechanistically well-argued, safe, with reproducible short-term effects in single-session SCI studies, and has not yet been demonstrated to beat sham in a controlled trial. A patient who asks deserves both halves of that sentence. Our own experience with tSCS across spinal cord injured clients and those with MS has been positive, and I hold that personal experience as stated.
Pulling the audit into practice:
If you would like to talk through the evidence, see the protocols run, or discuss the patients in your caseload for whom spasticity is the presenting obstacle, we are glad to help; supporting clinicians in the practical application of stimulation is a large part of what we do.
https://www.anatomicalconcepts.com
https://denervatedmuscle.com
Articles on this site are researched and drafted with the help of AI tools, then checked, edited and approved by me. Every citation is verified against the original source before publication.
Have questions about this article?
Get in TouchContinue reading about similar topics
Ask in your own words about FES cycling. Answers are drawn from both the patient guide and the clinical evidence review.
AI-generated answers based on the book. Not a substitute for professional medical advice.
