Not whole collagen. It is far too large to get through the skin’s outer layer. Smaller fragments of collagen are less clear-cut, and a few laboratory studies report that some get partway through, but nothing we found shows that this turns into a visible effect on a person’s skin. What collagen on the surface can plausibly do is moisturise. Eating collagen is a different route with a different, and also contested, evidence base.
This is the evidence companion to Collagen Stick or Wand? Two Different Products, which explains what topical collagen products are.
How big is collagen?
Size is the whole argument, so here are the numbers, with where each comes from. A dalton is a unit of molecular mass.
| What | Approximate size | Source |
|---|---|---|
| Native collagen molecule | about 300,000 daltons (300 kDa) | 2020 review in Antioxidants |
| Hydrolysed collagen | about 1,000 to 10,000 daltons | The same review |
| Fish-scale collagen peptides tested in one diffusion-cell study | about 1,300 to 4,500 daltons | Chai et al., 2010 |
| Proposed size limit for absorption through skin | under about 500 daltons | Bos and Meinardi, 2000 |
Worked through: whole collagen is around 600 times the proposed limit. Hydrolysed fragments, the form used in supplements and cosmetics, are roughly 2 to 20 times it. Our guide to what “hydrolysed collagen” means on a UK label explains how they’re made.
The 500 dalton guideline
A review in Experimental Dermatology (Bos and Meinardi, 2000) proposed that a compound generally needs to weigh under about 500 daltons to be absorbed through skin, because larger molecules can’t get through the stratum corneum, the skin’s outer layer. The authors reached this by looking at what was already known: nearly all common contact allergens, most topical drugs, and most drugs given through the skin sit below that weight.
It is a rule of thumb drawn from existing compounds, not a measured law of nature, and the next section shows why it isn’t the end of the discussion. But it is the starting point, and by that standard whole collagen isn’t close.
What the lab studies on smaller fragments found
The 2020 review of collagen hydrolysates (Aguirre-Cruz et al.) says plainly that native collagen can’t penetrate the stratum corneum. For hydrolysed collagen the picture is messier. Two studies it summarises:
- Fish-scale peptides (Chai et al., 2010). Peptides made from tilapia scales were separated into fractions of roughly 1,300, 2,000, 3,500 and 4,500 daltons and tested in a Franz-type diffusion cell. The larger fractions showed higher cumulative penetration than the smaller ones, and the authors report that the peptides reached the epidermis and dermis. The abstract doesn’t say what skin was used.
- Deer-sinew hydrolysate (Zhang et al., as summarised in the review). About 8% of a hydrolysate of 5,000 to 13,000 daltons was reported to penetrate mouse skin in a laboratory test.
Two cautions. First, every size tested here is still at least two and a half times the 500-dalton guideline, so these results don’t fit it neatly, which is a reason to treat them carefully rather than to dismiss them. Second, they are laboratory models. A fragment that crosses a test membrane isn’t the same as a cream on a face producing a change someone could see, and nothing we found shows that.
Two other papers are worth knowing about. A 2024 study in the Journal of Cosmetic Dermatology (Lee et al.) tested collagen packed into liposomes, tiny fat-based capsules designed to carry ingredients. In artificial membranes and a laboratory 3D skin model, the liposome version stayed in place better after washing and was associated with higher collagen levels in the model. That tells us delivery systems are being researched, not that ordinary collagen products have them, or that the effect carries over to people. And an early radiotracer study (Huc et al., 1981) tracked labelled collagen applied to skin and reported radioactivity in the skin, with most of what it found in deeper skin collagen sitting in peptide chains. It is too old, and the abstract is too thin on detail (it doesn’t name the species or the preparation), to settle anything.
What topical collagen does on the surface
The research that does exist points to the skin’s surface. The 2020 review describes hydrolysed collagen as having good moisturising properties at the stratum corneum. It also summarises small human studies, with fewer than about 60 participants and durations from hours to a few months, in which products containing collagen were followed by gains in skin hydration or elasticity.
Those findings come with a limit we can’t resolve from here. The review’s summary doesn’t tell us whether those studies compared the collagen product against another moisturiser. If they didn’t, a hydration gain can’t be credited to collagen specifically, because plenty of ingredients raise surface hydration. We’re also relying on the review’s account of these studies rather than the original papers.
Eating collagen is a different route
When collagen is swallowed it goes through digestion, not the skin barrier. Hydrolysed collagen is broken into small peptides, and one collagen-derived dipeptide, prolyl-hydroxyproline (Pro-Hyp), was detected in human blood one to two hours after ingestion in a 2005 study (Iwai et al.). Laboratory work has shown that Pro-Hyp stimulates skin cells called fibroblasts. We cover that chain of evidence in Does eating collagen reach your skin?.
That evidence is about the oral route only and doesn’t transfer to a cream or balm. It is also described by its own authors as preliminary, and our evidence roundup sets out where it is mixed. The honest comparison isn’t that one route is proven and the other isn’t. It’s that the two routes face different questions: one has to cross the skin barrier, the other has to survive digestion and reach the skin through the blood.
What this means for a product label
“Contains collagen” on a topical product describes an ingredient, not a delivery. Seller claims such as “visibly reduces wrinkles” are the seller’s claims, and we report them without repeating them as fact. In Great Britain, cosmetics regulation says claims must not suggest characteristics a product doesn’t have, and the one authorised UK/EU collagen-related claim, that Vitamin C contributes to normal collagen formation for the normal function of skin, belongs to food law and doesn’t cover topical products. We cover that in more detail in our guide to collagen sticks and wands.
We haven’t tested any topical product and don’t recommend one. For a skin concern, a pharmacist or GP is the right person to ask. Nothing here is medical advice.
The honest bottom line
Whole collagen can’t be absorbed through skin, by any size argument we found. Smaller fragments might get partway in laboratory models, with results that sit awkwardly against the size rule and haven’t been shown to matter for a visible effect on people. What collagen can plausibly do on the surface is moisturise. And eating collagen is a separate question with its own mixed evidence, covered in the guides above.
Sources: Aguirre-Cruz et al., “Collagen Hydrolysates for Skin Protection: Oral Administration and Topical Formulation”, Antioxidants 2020; Bos and Meinardi, “The 500 Dalton rule for the skin penetration of chemical compounds and drugs”, Experimental Dermatology 2000; Chai et al., “Effects of sizes and conformations of fish-scale collagen peptides on facial skin qualities and transdermal penetration efficiency”, Journal of Biomedicine and Biotechnology 2010; Lee et al., “Liposome-assisted penetration and antiaging effects of collagen in a 3D skin model”, Journal of Cosmetic Dermatology 2024; Huc et al., International Journal of Cosmetic Science 1981; Regulation (EC) No 1223/2009 on cosmetic products as it applies in Great Britain, Article 20.