Peptides for Skin: What Actually Has Trial Data

"Peptides" is sold as an ingredient. It is a category, and the things inside it do four unrelated jobs by four unrelated mechanisms with four very different amounts of evidence behind them.

Underneath all four sits a problem the marketing never raises. A peptide is a chain of amino acids, and chains are heavy. The skin's outer layer is specifically good at keeping heavy, water-loving molecules out. So before asking whether a peptide works, there is a prior question: can it get to where it would need to work?

For the most famous peptide in skincare, two published studies measured exactly that and disagreed by a factor of 136.

Four Different Things Share One Word

The standard classification splits them by mechanism, and the examples are products you will recognise.

A product listing "peptides" on the front could contain any of these, and they are not interchangeable. One is trying to stimulate synthesis. One is trying to deliver copper. One is trying to do something to a nerve ending. Reading "peptide complex" on a label tells you approximately nothing.

The Problem That Comes First: Size

The rule governing whether any topical gets in is physical, not commercial. A review of peptides in skin senescence states it precisely: for good permeability a molecule should have a molecular weight under 500 Da and a log P between 1 and 3 — small, and balanced between oily and watery.

Then it states the problem with its own subject matter:

Peptides have multiple amide bonds, high hydrophilicity and, in most cases, large molecular weight, so they present difficulties in diffusion through the epithelium. The stratum corneum is very lipophilic and acts as a barrier to hydrophilic compounds — such as peptides.

That is a review of cosmetic peptides saying that cosmetic peptides are the wrong shape to cross skin. It is not a hostile source; it is simply the physics — a review of dermal absorption puts it more bluntly still, listing peptides and proteins among the molecules that do not penetrate autonomously at all. The same constraint governs every other active on the shelf.

Which makes the penetration data the most important evidence in this article, and it is a mess.

"Botox in a Jar", Audited

Acetyl hexapeptide — sold as acetyl hexapeptide-3 or -8 — is the one that built the category. The pitch is that it mimics botulinum toxin topically.

It is not nonsense. A dedicated review of its permeability and efficacy describes a real proposed mechanism: the peptide competes with SNAP-25 for binding to vesicle-associated membrane protein and destabilises formation of the SNARE complex, inhibiting neuronal exocytosis. That is the same machinery botulinum toxin interferes with.

The difference is delivery, and it is not a detail. Botulinum toxin is injected into muscle. The peptide is spread on skin.

For the mechanism to operate, the peptide has to reach a neuromuscular junction. Two studies measured how much crosses:

• One found 30% of the applied amount penetrated a membrane within two hours.
• Kraeling and colleagues found 0.22% of the applied peptide penetrated the stratum corneum, with 99.7% removed from the skin surface and none detected in the receptor solution at all.

Tape-stripping resolved it in the least flattering direction: the highest concentration was found in the outer layers of the stratum corneum — sitting on top, where there are no nerves and no muscle.

The review's own conclusion is the line to keep: delivering the peptide transdermally to muscles by topical application is "likely impossible."

And then the sentence that should end the marketing: "none of the aforementioned in vivo application studies confirmed the peptide's inhibitory effect on muscle contractions and its role in smoothing wrinkles."

The figure that built this ingredient is "up to 30% reduction in wrinkle depth." It traces to a 2002 paper in the International Journal of Cosmetic Science by Blanes-Mira and colleagues — a small, uncontrolled topical proof-of-concept with no placebo arm and no comparator, on a compound commercialised by the organisation the work came from. That primary paper is not openly accessible, and notably, the reviews that cite it do not report its sample size either.

A second figure circulates too: 49% wrinkle depth reduction from 10% acetyl hexapeptide-8 over four weeks of daily use. The review reporting it gives no sample size for that study.

Set against those, another 10% study in the same review found that skin elasticity measured by cutometry did not significantly change, and stratum corneum water content did not significantly increase.

A separate review of the whole literature puts the ceiling on all of it: 10 studies, 302 females and 10 males. Only four of the ten examined acetyl hexapeptide-8 in isolation; five tested it inside multi-ingredient formulations where its own contribution cannot be separated. That is the entire clinical base for the ingredient that defined the category.

The honest verdict is not that it does nothing. It is that the headline percentages come from small uncontrolled studies whose sample sizes often go unreported, that the proposed mechanism has never been confirmed in living skin, and that the review authors themselves suggest a different mechanism may be responsible for whatever is observed — meaning even the believers are no longer sure it works the way it is sold.

Signal Peptides: The Strongest Case

This is where the better evidence sits, and it is still thinner than the shelf implies.

Palmitoyl pentapeptide-4 is the best-known. The mechanism is to boost production of collagen types I and III and to interfere with collagen degradation. Critically, it is palmitoylated — a fatty acid chain attached to make a hydrophilic peptide more lipophilic, which is a direct engineering answer to the penetration problem above.

The clinical anchor is a double-blind, placebo-controlled, randomised study of 93 women with facial wrinkles over 12 weeks, which found improvement in fine lines and overall skin appearance. At 3% it has been shown safe, non-irritating and non-sensitising.

That is a properly designed trial — blinded, placebo-controlled, randomised, and three times the size of the entire isolated-ingredient evidence base for acetyl hexapeptide. It is also reported, in the review carrying it, as "improvement" without an effect size. A direction without a magnitude is better than an unblinded percentage, and it is not enough to compare against anything.

Carrier and Enzyme-Inhibitor Peptides

Carrier peptides — copper tripeptide-1, GHK-Cu — have a well-characterised mechanism: modulating matrix metalloproteinase expression and stimulating collagen, elastin and glycosaminoglycan production. The clinical evidence cited is an eight-week study in which wrinkle depth became significantly smaller than vehicle alone, and a study combining GHK-Cu with light therapy, which has the same confounding problem as the exosome trials run alongside a procedure. Mechanism strong, human outcome data light.

Enzyme-inhibitor peptides — soy, silk, rice — are the thinnest. The review's own words: this type "shows promising results, but not enough clinical studies are available to confirm their benefits in anti-aging products." The specifics are consistent with that: soy peptides reduced UVB-induced erythema but had no impact on pigmentation spots or hydration; rice peptides reduced melanin content and tyrosinase activity in laboratory conditions.

What the Evidence Base Actually Looks Like

Three structural problems, stated by the reviewers themselves.

Most of it is in vitro. The senescence review notes that most bioactivity studies use in vitro assays because they are a cost-effective way to assess preliminary activity — and that there remains a significant shortage of in vivo clinical trials to confirm efficacy.

The trials that exist are small and heterogeneous. The authors attribute variable results to inclusion criteria differing by skin phenotype, age, wrinkle depth, photoageing signs, dermatological disorders and prior treatments. Which is a polite way of saying the studies are not comparable.

And much of it is commercially entangled. The defining peptide in the category was characterised in a paper on a compound sold by the organisation behind it. That is not unusual in cosmetic science, and it is the reason to read who ran a study before reading what it found — exactly as with the patent that set the rules for the vitamin C category.

In fairness, the senescence review declares no external funding and no conflicts of interest, which is why it is the source carrying most of this article's weight.

Where Peptides Sit Against a Retinoid

The comparison people actually want.

A retinoid has decades of controlled trials, a well-established mechanism acting through nuclear receptors, a molecular weight that comfortably clears the penetration threshold, and a known, predictable irritation profile. Peptides have a small number of trials, a penetration problem their own reviewers describe as a difficulty, and excellent tolerability.

Which produces the honest positioning: peptides are the gentle option, not the strong one. If you cannot tolerate a retinoid, a palmitoylated signal peptide is a reasonable thing to try, with modest expectations and a blinded 93-person trial behind it. If you can tolerate a retinoid and you are choosing between them for wrinkles, the evidence is not close.

And for photoageing specifically, neither is the first line. The only intervention with a randomised controlled trial behind a long-term ageing outcome is daily sunscreen — which beats everything in this article on evidence and costs less than most of it.

Conclusion

Four unrelated mechanisms share one marketing word, and a product saying "peptides" could be any of them.

Signal peptides have the best case — a double-blind, placebo-controlled trial of 93 women, and a palmitoyl group deliberately added to solve the penetration problem. The review reports improvement without an effect size, which is a direction without a magnitude.

Carrier peptides have a strong mechanism and light human data, some of it confounded with a procedure. Enzyme-inhibitor peptides do not have enough clinical studies to confirm anything, by their reviewers' own account.

And the neurotransmitter-inhibitor peptide that built the category has a delivery problem its own literature calls likely impossible. Two studies of how much crosses the skin returned 0.22% and 30%. The highest concentration found by tape-stripping was in the outermost layers. No in vivo study has confirmed the effect on muscle contraction that the entire pitch depends on. The whole clinical base is 312 people across ten studies, only four of which tested the ingredient on its own.

None of that makes peptides useless. It makes them a gentle, well-tolerated category with one properly designed trial behind its strongest member, sold at prices that imply something much more settled.

If a product's case rests on a wrinkle-depth percentage, ask how many people were in that study. Surprisingly often, nobody reports it.