Neuromuscular electrical stimulation and the knee

Electrodes on the thigh, a dial, and the quadriceps contracts whether or not the person can make it. That is neuromuscular electrical stimulation, and the reason it exists in knee medicine is arthrogenic inhibition: after injury or surgery the nervous system stops fully recruiting the muscle around a painful joint, and no amount of willing recruits it back quickly.

It is worth separating from TENS, which uses the same kind of box at parameters chosen to modulate pain rather than to contract the muscle. The two get sold together and tested together, and they are different claims.

The claim that reaches cartilage is entirely indirect and it has two links: the stimulator strengthens a quadriceps, and a stronger quadriceps changes how the joint is loaded. This entry covers what has been measured. Neither link has been measured to the cartilage.

After knee surgery

This is the use with real support, and every endpoint in it is muscle.

Pooled across nine trials and 691 patients after total knee replacement, stimulation raised quadriceps strength within the first month and the difference was still there at a year, with smaller effects on pain at one to six months and on the WOMAC at three to four months. The authors of that pooling are careful to say that many outcomes did not reach the minimal clinically important difference and that the clinical benefit remains to be confirmed.

The trial that established the pattern started stimulation 48 hours after surgery, twice a day at maximum tolerable intensity, in 66 patients. At three and a half weeks the treated group was stronger in the quadriceps and hamstrings, performed better on function tests and had more active extension. By 52 weeks the gap had narrowed and still favoured stimulation. Its own limitations are worth knowing: treatment volume was not matched between arms, testers were not blinded, and some patients reached the stimulator's maximum output and might have tolerated more.

The most granular look comes from muscle biopsies after ligament reconstruction. Surgery shrank single muscle fibres and cut their contractility across all fibre types; stimulation reduced that atrophy in fast-twitch fibres and preserved contractility in slow-twitch ones. At six months there was no detectable difference in whole-muscle strength. That is the shape of the evidence at the cellular level.

Across eight post-surgical trials, the effect on quadriceps strength ranges from slightly against stimulation to strongly for it. The parameter recipe those reviewers offer — first two postoperative weeks, 50 hertz or above, maximum tolerable intensity, biphasic current, large electrodes — is a description of which protocols happened to work, not the result of anyone randomizing a parameter.

In knee osteoarthritis

The cleanest test is a four-arm trial: a hundred people, twenty-five each to TENS, interferential current, Aussie current and placebo, over four weeks. None of the three waveforms beat placebo on pain intensity, disability, pressure pain threshold or any functional outcome. Most measures improved over the four weeks in every arm, placebo included, which is the reason uncontrolled improvement counts for so little here.

The two most recent syntheses agree with it. In the review that isolates what a device adds to exercise, stimulation did not separate from a placebo device with exercise or from exercise alone on pain or function, and two trials found no quality-of-life difference against exercise alone; certainty for every stimulation comparison was very low. In the network that ranks nine interventions, electrical stimulation did not beat exercise on any scale.

Against that sits one positive pooling, from 2012: nine randomized trials and one controlled trial, 409 people, with WOMAC function and WOMAC pain both favouring stimulation and walking tests not moving. Its authors grade the body of evidence level D — inconsistent — write that the role of stimulation here is ambiguous, and note that one included trial produced the same gains after sham stimulation. It searched to January 2011 and has never been repeated with the placebo-controlled trials published since.

A more recent trial is positive and shows why the placebo arm matters. Seventy-five women were split between stimulation, exercise and both; both together beat either alone on knee flexion range, thigh girth and vastus medialis thickness, and stimulation alone gave the lowest pain at twelve weeks. There is no placebo-stimulation arm in it, so nothing separates the current from the electrodes and the attention, and with no nominated primary outcome the winner changes with the scale being read.

Patellofemoral pain

A Cochrane review of eight trials and 345 participants reaches no conclusion: insufficient and inconclusive evidence, and uncertainty about whether a multi-session programme with exercise differs from exercise alone at the end of treatment or at a year. Its one placebo comparison — a single fifteen-minute session in twenty-two people — was downgraded three levels, partly for indirectness, because one session is not what anyone actually does.

Safety

Adverse events are barely reported anywhere in this literature. The Cochrane review found no data at all on muscle fatigue or discomfort in any of its eight trials. The four-arm placebo trial reports none, and the forty-trial adjunct review found adverse events in one trial, none of them stimulation events.

What would change this

The 2012 pooling repeated with the placebo-controlled trials that have appeared since, which is a paper that could be written now. A trial that randomizes a stimulation parameter rather than describing the ones that worked. And a trial in the people this is most often proposed for — those who cannot exercise — because the trials so far recruit people who can, and the interesting question about a device that contracts a muscle for you is what it does for someone who has no other way to load it.