Palmitoyl pentapeptide-4

A synthetic five-amino-acid peptide attached to a fatty acid, listed on labels as Palmitoyl Pentapeptide-4. The Cosmetic Ingredient Review panel's 2024 final report notes that this single label name covers two different sequences, KTTKS and KTSKS; the version studied as Matrixyl is KTTKS. According to 2023 usage data the ingredient appeared in 239 products, 223 of them leave-on.

Sold as

Applied to the skin (cream, serum, lotion)

Legal status in the EU

A cosmetic ingredient under EU Regulation 1223/2009. No approval is needed before sale — the responsible person must hold a safety assessment and a product information file. The review panel's final report, from a panel meeting on 30 September–1 October 2024 and published 18 November 2024, states that the pentapeptide-4 ingredients are not restricted in any way by EU cosmetics rules. If a product is claimed to treat disease or change how the body works, it can instead fall under the definition of a medicine in Directive 2001/83/EC.

KTTKS is, according to the review panel, a fragment of the precursor to type I collagen, and is regarded as a signalling peptide. In regulatory terms the ingredient is classed as a skin conditioning agent, not a collagen stimulator. The fatty acid group is an anchor meant to help it cross the outer layer of skin; PubChem gives the formula C39H75N7O10 and a weight of 802.1 g/mol.

Part of the bodyWhat happensHow it is knownKind of evidence

The skin where you put it

Almost nothing you would notice. Across every human test on record the peptide sat on the skin without stinging it, reddening it or setting off an allergy. The one reaction reported in any of them was very slight redness in a single person.

Measured in people in four patch tests collected by the Cosmetic Ingredient Review, a United States expert panel that reviews cosmetic ingredients. Two were repeat-insult tests: the same material goes on the same patch of skin many times over three weeks, then, after a two-week rest, once more, to see whether the skin has learned to react. One used 106 people and a formulation containing 0.12% of the peptide, diluted to 15% in water. The other used 51 people and an undiluted trade-name mixture containing 0.01% of it. Neither found irritation or allergy. In a separate single-application test in 10 people, with the patch left on the back for 48 hours, 1 person had very slight redness and the irritation score was 0.10. A fourth patch test, in 11 people and also left on for 48 hours, found no irritation.

Measured in people

Source [02]

Fine lines and wrinkles on the face

This is the whole reason the ingredient exists, and two trials have tested it on its own. In the larger one, the half of the face given the peptide cream had measurably fewer fine lines after three months than the half given the identical cream without it. In the smaller, independent one, the crow's-feet lines at the corner of the eye faded over eight weeks. That second trial had seven women per group and never compared its wrinkle scores statistically, so it cannot tell you whether the cream or the passage of time did it.

Measured in people twice, and neither trial settles the size of the effect. Robinson and colleagues, 93 white women aged 35–55, 12 weeks, with each woman's own face used as her comparison: one side got a moisturiser containing 3 parts per million of the peptide, the other got the same moisturiser without it. Neither the women nor the assessors knew which side was which. The authors were employed by Procter & Gamble, and the published abstract gives no figure for how large the improvement was. Aruan and colleagues, 21 Indonesian women aged 26–55 split into three groups of 7 for 8 weeks, compared the peptide cream against a cream with a different peptide and against a dummy cream. It was funded by a university. Its wrinkle scores were reported as averages only, with no statistical test between the groups, and the authors called it an initial study needing a better design.

Measured in people

Source [01]Source [06]

Open wounds — in rats, not in people

Cuts made on the backs of rats and covered with a patch or a cream containing the peptide healed better than untreated cuts. Under a microscope the new skin over them looked closer to ordinary skin, while the untreated wounds stayed under a crusty scab. Nothing of this kind has been tested on a wound in a person.

Measured only in animals: 42 male albino Wistar rats weighing 200–250 g, in seven groups of 6, followed for 21 days. The groups compared patches carrying 0.1 mg or 1 mg of the peptide and a cream carrying 1 mg against blank patches, blank cream, a commercial wound dressing and untreated wounds. The paper reports healing improved by 63.5% to 81.81% in the treated groups compared with the untreated group. Those are improvements relative to that group, not the share of each wound that closed.

Only seen in animals

Source [07]

How far into the skin it gets

Barely past the surface. The fatty acid on the end is there to help the peptide cross the dead outer layer, and it does — but almost all of it stops in that outer layer and in the one below it. The layer where collagen is actually built receives a fraction of one per cent of what was put on, and none of it passed all the way through into the fluid on the other side.

Not measured in living skin at all. The one experiment used pieces of skin cut from hairless mice and clamped in a glass cell, with the peptide left on for 24 hours. Afterwards 14.6% of what was applied was still in the skin: 8.3% in the dead outer layer, 5.6% in the layer below it and 0.6% in the deeper layer where collagen is made. The amount that crossed right through was too small to measure. A 2011 review of the peptide called the absence of skin-penetration data surprising, and this experiment was published three years after it.

Only seen in a dish, not in a body

Source [02]Source [08]Source [09]

The cells that build collagen, in a dish

This is where the whole idea came from. The bare five-amino-acid sequence is a scrap left over when the body assembles collagen. When it is dripped onto skin-building cells growing in a dish, those cells make more collagen and more of the mesh that holds skin together. The version with the fatty acid does the same, and the effect appears around the concentration at which the molecules start clumping together into ribbon shapes.

Only in dishes of cells, never in living human skin. The original 1993 experiment used cultured skin-building cells and found the five-amino-acid piece was the shortest fragment that still worked. It raised production of two types of collagen and of fibronectin, one of the proteins in the mesh between cells, without changing total protein production. A 2013 experiment repeated it with the fatty-acid version on human skin and cornea cells. No study has taken a sample of human skin after use of the cream and measured whether collagen actually rose.

Only seen in a dish, not in a body

Source [10]Source [11]

The key study is Robinson and colleagues (2005) in the International Journal of Cosmetic Science: 93 women aged 35–55, 12 weeks, double-blind, placebo-controlled and randomised half-face against half-face, with 3 ppm pal-KTTKS in a moisturiser. It reported a significant improvement in fine lines and wrinkles compared with the same cream without the peptide, judged both by image analysis and by trained assessors. The authors were employed by Procter & Gamble, and no independent replication of the same size was found in this review.

Nobody has published when the first measurable change appears. The 93-woman trial reported a single endpoint, at 12 weeks. The 21-woman trial took measurements at weeks 4, 6 and 8, but reported its wrinkle scores only as a change from the start to week 8. The 196-woman trial of a regimen that contained this peptide among other ingredients measured at 8 weeks and again at 24, and found the wrinkled area still shrinking between the two. Nothing has been measured beyond 24 weeks, and no study has followed anyone after stopping.

What was measuredWhat happenedIn whomHow it is known

Fine lines and wrinkles, measured by image analysis and by expert graders

The side of the face given the peptide cream improved significantly more than the side given the same moisturiser without it, on both measures. The published abstract states that the difference was significant and gives no percentage, so how large the improvement was cannot be read from it.

93 white women aged 35–55, 12 weeks, each woman's left and right sides randomly assigned. Neither the women nor the assessors knew which side had the peptide. Run and written by authors employed by Procter & Gamble.

Measured in people

Source [01]

Area of skin around the eyes covered by wrinkles — for a product containing this peptide among several other active ingredients

The cosmetic regimen cut the wrinkled area by 17% at 8 weeks, against 11% for a prescription vitamin A cream (0.02% tretinoin). That difference did not reach statistical significance (P = 0.06). At 24 weeks the two were indistinguishable (P = 0.77). This peptide was one of several ingredients in the regimen, alongside niacinamide, a second peptide, carnosine and 0.3% retinyl propionate, so none of the result can be attributed to the peptide on its own.

196 women aged 40–65 with moderate to moderately severe wrinkles around the eyes, randomly assigned to one product or the other, 8 weeks for the main comparison and 24 weeks for a self-selected group of 50. Funded by Procter & Gamble Beauty; several authors were company employees and others were paid consultants.

Measured in people

Source [12]

Crow's-feet severity, on a 0-to-4 scale

The average score fell by 0.86 points at 8 weeks in the group given the peptide cream, both at rest and with an expression on the face. The group given a cream with a different peptide, acetyl hexapeptide-3, fell by the same 0.86 points at rest. The authors concluded the peptide cream did better than both the other cream and the dummy cream. But the wrinkle scores were only averaged, never compared statistically, and the trial's three skin-measuring instruments showed no significant change (p greater than 0.05).

21 Indonesian women aged 26–55, split into three groups of 7, applying cream to the eye area twice a day for 8 weeks, with measurements at weeks 4, 6 and 8. Funded by Kristen Krida Wacana University; the authors report no conflicts of interest.

One small study

Source [06]

How well cuts on the back healed, in rats

The paper reports healing improved by 63.5% to 81.81% in the treated groups compared with untreated animals. Those figures are improvements relative to that group, not the share of each wound that closed. The patch carrying 1 mg and the cream carrying 1 mg both healed wounds better than a commercial wound dressing used as the comparison.

42 male albino Wistar rats weighing 200–250 g, in seven groups of 6, followed for 21 days.

Only seen in animals

Source [07]

The expert panel concluded in its 2024 final report that these ingredients are safe as currently used and at current concentrations. The highest reported level in leave-on face and neck products was 0.0012%. Skin irritation is uncommon at those levels. The open question is how much effect there is and how much gets in: in a lab experiment on hairless mouse skin, 14.6% of the applied pal-KTTKS was found still in the skin, of which 0.6% in the deeper layer, while the amount that made it all the way through was below the limit of measurement. The researchers concluded the peptide did not pass through the skin.

EffectWhat it feels likeHow often

Slight redness where the cream was applied

The redness was described as hardly visible, and the overall irritation score for the test material was 0.10, which the report called well tolerated. The patch had been left on the back for 48 hours before the skin was read.

1 person out of 10 in a single-application irritation test, and nobody in the two larger repeat-insult patch tests of 106 and 51 people.

Source [02]

Allergy that builds up with repeated use

Both of the repeat-insult tests were negative, and two laboratory screening assays also predicted no allergy. Both human tests used more of the peptide than any product on sale contains.

Not seen in either of the two tests designed to find it: none in 106 people and none in 51 people.

Source [02]

Stinging in the eye

The only eye tests are laboratory ones and they do not agree. A trade-name mixture containing 0.01% of the peptide scored 6.0 on a test that uses the membrane inside a hen's egg as a stand-in for an eye, where a detergent scored 12.0; it was classed as moderately irritating. A different formulation scored 4.25 on the same test and was classed as slightly irritating. A human corneal cell model and a test in 3 rabbits found no irritation at all. No trial has asked people using an eye product whether it stung.

Not measured in people. 31 of the 239 products reported to contain the peptide are used around the eye.

Source [02]

Itching, stinging or burning reported during a trial

No serious side effects occurred in any group. With 7 women per group these counts mean very little on their own. They are quoted because this is the only trial that asked directly and reported the answers separately for each cream.

None of the 7 women given the peptide cream for 8 weeks reported any complaint. On the dummy cream, 2 of 7 reported itching and 2 reported burning. On the cream with a different peptide, 2 of 7 reported stinging, 1 itching and 1 burning.

Source [06]

These are the amounts the studies below gave their participants, and they are here as facts about those studies. They are not a recommendation, not a starting point and not a range to pick from. What a person should take, if anything, is a question for a doctor who knows them.

Robinson (2005), fine lines and wrinkles in women with sun-aged facial skin — one side of the face against the other

3 parts per million of the peptide (0.0003%) in a moisturiser, applied to one side of the face

12 weeks

Source [01]

What people report

• Whether it raises collagen in living human skin — no study has taken a skin sample from a person using it and measured.
• How much of it reaches the deeper layer of human skin; the only penetration figures come from pieces of skin cut off a mouse.
• How large the wrinkle improvement in the 93-woman trial actually was — the published abstract reports that it was significant and gives no number.
• Whether an independent group can reproduce that trial at the same size; the only independent trial had 7 women per group and never compared its wrinkle scores statistically.
• What the peptide contributes on its own to the mixed products it is usually sold in — the 196-woman regimen trial never published how much of it the cream contained.
• How much of the peptide was in the cream used in the 21-woman trial; that trial never states it.
• Whether anything it does lasts, or reverses, after a person stops using it; no study has followed anyone past the last day of treatment.
• What it does over more than 24 weeks — no trial has run longer.
• Whether the two sequences that share the label name, KTTKS and KTSKS, behave the same on skin.
• What it does in pregnancy or to a developing child, and whether it has any effect on cancer risk — the 2024 review found no studies of either.
• How often it stings the eye in real use, in the 31 eye-area products that contain it — the only eye data are laboratory tests that disagree with each other.

Does it actually do anything to wrinkles?

Nobody can say how much. The 93-woman, 12-week trial found the treated half of the face significantly better than the untreated half; the authors worked for Procter & Gamble, and the published abstract gives no figure for the size of that improvement. The 21-woman, 8-week independent trial reported the peptide cream doing better than a dummy cream, but never compared its wrinkle scores statistically and had 7 women per group. A 2026 systematic review of peptides for skin ageing included 19 randomised trials covering 1,341 people. Only 2 of them tested a peptide applied to the skin, covering 105 people between them, and neither of those tested this peptide.

Source [01]Source [06]Source [13]

How much of it is actually in the products on sale?

Very little. Of 239 products reported to contain it in 2023, 223 were leave-on. The highest amount reported in any leave-on face, neck or eye product was 0.0012% — twelve parts per million. The highest anywhere was 0.0035%, in a single hair conditioner. The cream in the 93-woman trial contained three parts per million.

Source [02]

Does it reach the layer of skin where collagen is made?

A very small amount does, in the only experiment that has looked. Pieces of skin cut from hairless mice, with the peptide left on them for 24 hours, held 14.6% of what was applied afterwards. Only 0.6% of it had reached the deeper layer where collagen is built, and the amount that crossed all the way through was too small to measure. The bare peptide without its fatty acid was not detectable in any layer at all. Nobody has measured this in living human skin.

Source [02]Source [08]

Two products both list it — are they the same thing?

Not necessarily. The label name Palmitoyl Pentapeptide-4 covers two different amino-acid sequences, written KTTKS and KTSKS. The one that was tested on 93 women, and the one the cell experiments used, is KTTKS. Several of the safety tests in the 2024 review were run on KTSKS instead. A label cannot tell you which one is in the jar.

Source [02]

Does it act like a hormone?

There is one laboratory signal and it is weak. A formulation containing 0.12% of the peptide produced an oestrogen-like response in genetically modified yeast cells, at the top of the range of concentrations tested. In the companion testosterone test on the same material, the cells were dying at the two highest concentrations. It produced no anti-oestrogen and no testosterone-like response. A second yeast test on the same material found no oestrogen-like response at any concentration, and the expert panel counted that among its reasons for not needing reproductive studies.

Source [02]

Is anyone running a trial on it now?

No. Searches of the United States clinical trials register in August 2026 for "palmitoyl pentapeptide" and for "Matrixyl" both return zero studies — none recruiting, none completed, none registered at all. Everything published on this peptide in people was run and reported outside the trial registration system.

Source [14]Source [15]

Each letter represents one amino acid.

1. [01]Robinson LR m.fl. Topical palmitoyl pentapeptide provides improvement in photoaged human facial skin. Int J Cosmet Sci 2005;27(3):155-60 (2005)
2. [02]CIR slutrapport: Safety Assessment of Myristoyl Pentapeptide-4, Palmitoyl Pentapeptide-4, and Pentapeptide-4 as Used in Cosmetics (Final Report, 18 november 2024) (2024)
3. [03]PubChem: Palmitoyl pentapeptide-4 – C39H75N7O10, 802,1 g/mol, IUPAC-namn motsvarande hexadekanoyl-Lys-Thr-Thr-Lys-Ser
4. [04]Regulation (EC) No 1223/2009 on cosmetic products (EUR-Lex, CELEX 32009R1223) (2009)
5. [05]WADA: World Anti-Doping Code International Standard – Prohibited List 2026 (i kraft 1 januari 2026) (2026)
6. [06]Aruan RR et al. Double-blind, randomized trial on the effectiveness of acetylhexapeptide-3 cream and palmitoyl pentapeptide-4 cream for crow's feet. J Clin Aesthet Dermatol 2023;16(2):37-43 (2023)
7. [07]Kachooeian M et al. Matrixyl patch vs Matrixyl cream: a comparative in vivo investigation of Matrixyl effect on wound healing. ACS Omega 2022;7(28):24695-704 (2022)
8. [08]Choi YL, Park EJ, Kim E, Na DH, Shin YH. Dermal stability and in vitro skin permeation of collagen pentapeptides (KTTKS and palmitoyl-KTTKS). Biomol Ther (Seoul) 2014;22(4):321-7 (2014)
9. [09]Abu Samah NH, Heard CM. Topically applied KTTKS: a review. Int J Cosmet Sci 2011;33(6):483-90 (2011)
10. [10]Katayama K et al. A pentapeptide from type I procollagen promotes extracellular matrix production. J Biol Chem 1993;268(14):9941-4 (1993)
11. [11]Jones RR et al. Collagen stimulating effect of peptide amphiphile C16-KTTKS on human fibroblasts. Mol Pharm 2013;10(3):1063-9 (2013)
12. [12]Fu JJ et al. A randomized, controlled comparative study of the wrinkle reduction benefits of a cosmetic niacinamide/peptide/retinyl propionate product regimen vs. a prescription 0.02% tretinoin product regimen. Br J Dermatol 2010;162(3):647-54 (2010)
13. [13]Nukaly HY et al. Oral and topical peptides for skin aging: systematic review and meta-analysis of randomized controlled trials. Front Med 2026;13:1618306 (2026)
14. [14]ClinicalTrials.gov API v2, search for "palmitoyl pentapeptide" — 0 studies returned (2026)
15. [15]ClinicalTrials.gov API v2, search for "Matrixyl" — 0 studies returned (2026)