Silk earns its place against skin on friction and moisture release. Where residual chemistry is the problem instead, OEKO-TEX Standard 100 or GOTS-certified cotton is the better buy, and your own condition decides which of those two sentences applies to you.
What silk does not earn is the layer of claims stacked on top of those two properties. The largest randomised trial of silk garments in childhood eczema followed 300 children for six months and found no clinically important benefit, and no trial at all sits behind the anti-wrinkle promise.
Neither finding makes silk a bad thing to sleep in. Both change what you should pay for it.
The other half of the decision is the dye, the finish and the laundry residue.
Eczema-prone skin, acne-prone skin, contact dermatitis and dry skin do not land on the same fabric. Match the property the condition needs first, then pick the fibre that delivers it: moisture release, low friction, a loose fit, or the least applied chemistry you can verify on a label.
Eczema-prone and dry skin both need a fabric that takes up moisture and gives it back to the air instead of holding it damp against the surface.
Smoothness is the second requirement, because a yarn spun from short staple fibres presents thousands of cut ends to broken skin.
Silk and a fine lightweight cotton both clear that bar, and the finish decides which one to buy.
For acne-prone skin the requirement is low friction, breathability and a loose fit. That is a property list, not a fibre.
The American Academy of Dermatology calls it acne mechanica, where clothing traps heat and sweat and then rubs against heated skin. Its advice is moisture-wicking material next to the body, because pulling sweat away reduces the rubbing.
A light loose wicking layer does the daytime job. Overnight, a smooth filament weave drags less than a brushed cotton.
Reactive and contact-sensitive skin is the one case where the certificate outranks the fibre.
A cotton shirt of unknown dye and finish can carry more of what triggers a reaction than a substance-tested polyester does. “Natural” is the wrong first filter here.
Silk’s measurable case against skin comes down to two behaviours. A continuous filament drags less than a woven cotton of the same weight.
Silk also releases moisture to the air faster than cotton does.
Weight decides whether either behaviour survives the wash. Silk weight is measured in momme, and below 19 momme the fabric thins under weekly laundering, so 19 momme or heavier is the floor for overnight contact.
When you are judging silk quality, momme is the first number to read on a silk pillowcase or sleep mask listing, ahead of the price.
The largest trial found nothing. The CLOTHES trial, reported in 2017, randomised 300 children with moderate to severe eczema, half to silk garments on top of standard care, with blinded nurses scoring severity over six months.
The adjusted ratio was 0.95 (95% CI 0.85 to 1.07, p = 0.43), which the authors called not clinically important. Cost ran to £56,811 per quality-adjusted life year, against an NHS threshold of £30,000.
Smaller studies had pointed the other way, including a 2004 study of 46 children where the silk wearers’ mean SCORAD severity score fell from 43 to 30 in a week.
The tempting explanation is that the big trial used ordinary silk. It did not: its garments were DermaSilk and DreamSkin, the same antimicrobially protected silk the positive studies used.
What separated the results was who held the score sheet. Participants’ own ratings favoured silk for about three months and then faded, which the authors put down to expectation bias, while blinded nurse scores showed nothing throughout.
A 2020 review in the American Journal of Clinical Dermatology reads the cotton and silk evidence as mixed, and both fibres as generally safe.
No trial supports the anti-wrinkle claim. The silk trials in dermatology are in atopic dermatitis, so the claim rests on an extrapolation from reduced friction rather than a measured outcome.
Reduced friction is real. A measured effect on skin ageing is not.
The key difference between two fabrics on a rail is rarely the fibre. What makes a fabric react with skin, rather than simply rub it, is what was left on or put on it after weaving.
A 60/40 cotton-polyester shirt is not automatically worse than pure cotton, because the dye class and the finish decide that.
Disperse dyes, which colour nylon, polyester and acetate, are one of the most common textile allergens. Formaldehyde resins are the equivalent hazard on cotton; they are what makes a shirt wrinkle-resistant.
When the US Government Accountability Office had 180 garments and linens tested in 2010, ten came back over the industry’s voluntary limit, and more than half of those carried durable-press labelling. The United States set no legal limit on formaldehyde in clothing then; the apparel industry counted 13 countries that did.
Silk carries a residue of its own. Sericin, the gum coating the raw filament, comes off during degumming because it can provoke allergic reactions.
Fabric softener is the layer you apply yourself, and it leaves a film that stays in skin contact for hours.
OEKO-TEX Standard 100 tests every component of a certified product against over 1,000 harmful substances. Its own Edition 01.2026 text says the mark is not a quality label. GOTS is the label that certifies organic cotton.
Three lines on a listing do most of the work:
Buy for the property your own skin needs, then check what was done to the fabric before you check the fibre on the label.
For overnight skin contact the choice that holds up is silk at a weight that survives washing. Where residual chemistry is the problem, certified cotton is the better buy, and a substance-tested synthetic beats an uncertified natural fibre for skin that reacts.
None of them is a treatment. The largest randomised trial of silk garments in childhood eczema measured no clinically important benefit, and the money goes wrong when you pay for that outcome instead of for friction and moisture release.
Everything worth checking is printed somewhere: a weight, a certificate scope, a finish claim.