PEMF Mats and Vibration Plates: Small Trials, Big Prices

The full story

There is a particular kind of longevity product that arrives with a citation list, and the citation list is the product. Pulsed electromagnetic field mats and whole-body vibration plates both belong to it. Both have genuine randomised, placebo-controlled literature behind them - which puts them well ahead of most of the recovery aisle - and both have effect sizes, sample sizes and follow-up windows that sit awkwardly beside a price tag in the high three or four figures.

The interesting thing is that they fail in different ways. The mat has a small pile of trials that mostly cannot agree with each other. The plate has a large pile of trials that agree quite well, and what they agree on is narrower than the sales copy.

The mat: eleven trials, none of them large

Start with the exact product category, because the trick in this space is citing research on a device that is not the one being sold. A systematic literature review in Bioelectromagnetics in 2012 looked specifically at whole-body PEMF mat devices - the home-use mats, not surgical bone stimulators - and found only 11 randomised, sham-controlled, double-blind trials in existence. Sample sizes ran from 12 to 71. Observation periods topped out at 12 weeks. Magnetic flux densities spanned 3.4 to 200 microtesla, a range broad enough that calling these the same intervention is generous. The authors concluded the evidence for therapeutic effects of whole-body PEMF devices was insufficient and their use could not be recommended, and noted that long-term adverse effects had not been studied at all.

That last clause is not a scare story. It is a description of a knowledge gap, and it is the correct thing to know about a device you are considering lying on daily for years.

Two meta-analyses, one intervention, opposite answers

PEMF for osteoarthritis is where the real trials live, and the field cannot make up its mind. A 2020 systematic review and meta-analysis in Phys Ther reviewed 16 randomised placebo-controlled trials and pooled 15, and reported PEMF beating placebo on pain with a standardised mean difference of 1.06, plus smaller but significant effects on stiffness and physical function. Quality of life did not reach significance. PEMF parameters did not influence symptoms, and the evidence was limited to short-term effects.

Now the same intervention, a year earlier. A 2019 meta-analysis in J Rehabil Med of 8 randomised trials in 421 knee osteoarthritis patients found improved physical function, but no significant advantage on WOMAC total score, on WOMAC pain, on visual analogue scale pain, or on WOMAC stiffness. Four pain and symptom outcomes, four p-values sitting between 0.06 and 0.15, none of them crossing the line. That is a null result on pain from the same literature that produced one of the largest pain effect sizes you will see in a device meta-analysis.

A 2021 meta-analysis in Cartilage, 13 studies and 914 patients, explains part of the discrepancy in a way worth internalising. PEMF did reduce knee osteoarthritis pain overall, with a modest standardised mean difference of -0.41, and did not significantly improve activity scores. But the benefit appeared only in placebo-controlled comparisons and vanished when PEMF was compared against alternative conservative treatments such as physiotherapy. Longer follow-up correlated with smaller differences between treated and control patients.

Read that carefully: the device beats a sham and does not beat physiotherapy, and the gap shrinks the longer you look. Every consumer claim you have seen for these mats rests on the first half of that sentence.

The plate beats sitting still

Whole-body vibration has better internal consistency. A 2023 systematic review and meta-analysis in Arch Phys Med Rehabil covering 18 randomised trials in adults with a mean age of 65 or older found improvements across the board: knee strength, explosive power, sit-to-stand performance, timed up-and-go, walking speed and walking endurance, with standardised mean differences between roughly 0.43 and 0.72. For a device you stand on, that is a respectable spread of functional outcomes.

A 2012 systematic review in Disabil Rehabil found much the same thing in 16 randomised trials, with knee isometric strength up 18.30 Nm, muscle power up 10.44 W and Tinetti balance scores up 4.5 points - versus non-exercising controls. And then it did the comparison almost nobody quotes. Against conventional exercise, the only outcome on which vibration showed a significant advantage was femoral-neck bone mineral density.

That is the sentence the category cannot escape. The plate outperforms sitting still. It does not outperform training.

What the evidence does not show

It does not show that vibration helps the people most often sold on it. A 2021 meta-analysis in BMC Geriatr, 26 studies and 1,191 participants with sarcopenia, found whole-body vibration produced no significant improvement in knee extension strength or in gait speed, while resistance training and mixed training improved both. Sarcopenia is the exact indication in the brochure, and vibration is the one modality in that analysis that came up empty.

It does not show that vibration prevents falls. A 2012 meta-analysis in Maturitas of 13 trials in older adults found significant improvements in Tinetti total score, Tinetti body balance and timed up-and-go, but not in Tinetti gait score, and the authors state directly that the effect on other balance and mobility outcomes and on actual fall rate remains inconclusive. Improving a balance score is not the same as falling less often, and only one of those is what anyone actually wants.

And the bone-density claim is far weaker than the marketing. A 2010 meta-analysis in Osteoporos Int of 8 randomised trials found a small but significant increase in hip areal bone mineral density in postmenopausal women, 0.015 g/cm2, with no significant change in spine areal density, none in tibial trabecular volumetric density, and no effect whatsoever in young adults. A 2023 meta-analysis in the same journal reviewed 23 studies and pooled 20 in postmenopausal women and found improvements in lumbar-spine and trochanter density across several subgroups - but only two parameter combinations reached high-quality evidence: about 30 Hz at about 0.3 g, and a high cumulative dose of roughly 7,000 minutes at low magnitude. Both only for lumbar spine. Bone responds to vibration, in specific women, at specific settings, at specific sites, by a small amount.

How to read the spec sheet

Two numbers on a vibration plate box matter more than the rest: frequency in hertz and magnitude in g. The only high-quality bone evidence sits near 30 Hz and 0.3 g, or at very low magnitude accumulated over thousands of minutes, which is a serious time commitment measured in years of standing rather than weeks. A plate advertising far higher magnitudes is not advertising a stronger version of the evidence, it is advertising something the evidence has not tested for bone.

For PEMF mats, the useful question is what the mat is being compared with. If a claim rests on beating a sham, ask what happened against physiotherapy, because that comparison has been run and the benefit did not survive it. And check the flux density against the 3.4 to 200 microtesla range of the trials, because outside it you are extrapolating from nothing.

None of this means either device is a fraud. It means the plate is a mild exercise substitute with real functional gains for people who are not otherwise training, and the mat is a short-term symptom device whose pain literature contradicts itself. Both are legitimate purchases at the right price. Neither is a four-figure purchase, and the gap between those two statements is the entire business model.