Cupping therapy

Start with what suction physically does. A cup at roughly -20 to -50 kPa lifts skin and superficial fascia into the dome, stretches the tissue radially, and deforms capillaries and small venules until some of them rupture. Blood extravasates into the dermis. The circular marks are ecchymosis. They are a bruise, produced deliberately and reproducibly.

What follows is ordinary bruise chemistry. Extravasated haemoglobin is degraded by heme oxygenase-1 into biliverdin, then bilirubin, releasing iron and carbon monoxide, which is why the marks fade through purple to green to yellow on a predictable schedule. Some researchers have proposed that this HO-1 induction is mildly cytoprotective and could be the real mechanism. It is an interesting hypothesis with no clinical endpoint attached to it.

Now the mechanical claim. When cupping has been modelled properly (Tham and colleagues' biomechanical analysis is the standard reference), the stresses concentrate in skin and the most superficial layers, and fall off steeply with depth. Deep fascia and muscle see very little. Decompressing structures below the superficial layer is not supported by the mechanics of a suction cup on skin.

There are defensible mechanisms here, and they are neurological rather than structural: counter-irritation, diffuse noxious inhibitory control (a strong benign stimulus in one place dampens pain processing more broadly), sustained mechanoreceptor input competing with nociceptive input, and a therapeutic ritual with high expectancy value. Those are Mind-Body Wing mechanisms. They are real. They are not tissue repair.

A word on respect, because it is warranted. Within traditional Chinese medicine cupping is internally coherent and has been practised for over two millennia, embedded in a theory of stagnation and flow that predates capillary physiology by many centuries. Judging it against qi is a category error. Judging the modern marketing against physiology is entirely fair, because those claims are made in physiology's own language and asked to be believed on physiology's terms.

The systematic review literature (Kim et al. 2011 on pain; Bridgett et al. 2018 specifically in athletes) reaches a consistent conclusion: trials are small, randomisation is often unclear, outcome assessment is rarely blinded, and there is no credible sham, because you cannot fake a circular bruise on a person's back.

Bridgett's review of cupping in sport found eleven studies with wide heterogeneity in technique, dose, and population, and no consistent effect on performance or on objective recovery markers. Where positive findings appear they are short-term and pain-report based.

This is why the evidence-quality axis scores low rather than the effect axis carrying the whole judgement. The problem is not only that cupping fails to show effects. It is that the studies as designed largely cannot detect the difference between the treatment and the theatre.

The specific claim of toxin release deserves its own line: it has never been operationalised. No candidate molecule, no assay, no before-and-after measurement in any published trial. A mechanism that cannot be measured is not a mechanism, it is an image.

Where cupping could honestly land is as a short-duration analgesic with a strong ritual component, which is a legitimate thing to offer as long as it is what you say you are offering.