Skip to main content

Sources last read

Topical peptide skincare · human evidence

Peptides in skincare: what has actually been tested

Some topical peptide formulas improved measured wrinkles in small human studies. The result belongs to the tested formula—not every peptide serum, and not injectable peptide therapy.

Source: Four controlled human studies and one FDA laboratory report, read 2026-08-16

What matters first

Controlled human studies
4
Wrinkle results
Some tested topical formulas improved them
GHK-Cu after laser
No clear measured advantage
What carries over
The result belongs to the tested formula

“Peptides work” is too broad to help. One study tested 3 ppm pal-KTTKS in one moisturizer. Another tested GHK-Cu after laser treatment and found no clear edge in the measured results. A third used a dissolving patch. The product and the question changed each time.

The claim has four gates

An ingredient name is only the first step

  1. 01Identity

    Peptide named

    The formula names the exact ingredient—not just a “peptide complex.”

  2. 02Formulation

    Formula holds

    The finished product keeps it stable at the dose that was tested.

  3. 03Delivery

    Skin can receive it

    The product gets enough of it to the part of skin behind the claim.

  4. 04Human outcome

    Vehicle beaten

    The full formula works better than the same base without the peptide.

Most packaging can show step 01. A controlled human trial is what begins to answer step 04.

The human ledger

Four studies, four different products

The control and the limit sit beside each result. That stops a good result from becoming a promise about every serum with a similar ingredient name.

93 enrolled

12 weeks

PMID 18492182

Pal-KTTKS moisturizer

Tested
The control moisturizer with 3 ppm palmitoyl pentapeptide (pal-KTTKS)
Compared with
The same moisturizer without pal-KTTKS, used on the other side of the face
Finding
Image tests and expert review found more improvement in wrinkles and fine lines than with the control moisturizer.
Do not carry it past
The study tested one dose in one moisturizer. It enrolled only Caucasian women aged 35–55 and was written by Procter & Gamble researchers.

13 enrolled

12 weeks

PMID 16847171

GHK-Cu after laser resurfacing

Tested
A post-treatment skincare regimen containing copper tripeptide complex (GHK-Cu)
Compared with
A post-treatment regimen without GHK-Cu after the same CO2 laser procedure
Finding
Blind review and computer checks found no clear edge for redness, wrinkles or overall skin quality. Patient satisfaction was higher.
Do not carry it past
Only 13 people finished the study. Healing after laser treatment is not the same as daily use on untreated skin.

21 enrolled

8 weeks

PMID 36909866

AHP-3 vs PPP-4 creams

Tested
Acetylhexapeptide-3 cream or palmitoyl pentapeptide-4 cream, twice daily
Compared with
A placebo cream in a three-group randomized trial
Finding
The authors said palmitoyl pentapeptide-4 did better than acetylhexapeptide-3 and placebo across the tests used.
Do not carry it past
The authors called this an early study. Just 21 people were split among three groups, and longer follow-up is still needed.

52 enrolled

50 completed

29 days

PMID 33911590

AHP-8 microneedle patch

Tested
A dissolving cross-linked hyaluronic acid microneedle patch containing acetyl hexapeptide-8
Compared with
The same microneedle patch without AHP-8 in a split-face trial
Finding
All patch groups improved. The AHP-8 patch improved wrinkles more than the patch alone in the study’s tests.
Do not carry it past
A dissolving patch changes how the peptide is sent into skin. It also had hyaluronic acid, so this result does not prove an ordinary serum works.

FDA laboratory check

The label and the detected ingredient did not always agree

FDA researchers analyzed 77 cosmetic products for six commonly used anti-aging peptides. At least one target was detected in 54.5% of products. Among products that claimed acetyl hexapeptide-8, the test found it in only 33%.

This is not a product ranking. It also does not prove that every failed test means a bad label. It does show why a peptide claim is not the same as a checked formula.

A five-minute product check

What to look for before believing the claim

  1. Find the exact peptide name. “Peptide blend” does not tell you if it matches a human study.
  2. Match the delivery form. A dissolving microneedle patch is not evidence for a dropper serum.
  3. Look for the base control. Skin can improve from the moisturizer, hyaluronic acid or the procedure itself.
  4. Check what was measured. Satisfaction, skin moisture, image tests and wrinkle grades do not mean the same thing.

Topical skincare

A finished cosmetic formula

The evidence belongs to a dose, base, delivery method and measured skin result.

Injectable records

A different route and question

The skin-goal ledger and GHK-Cu record cover the injection and research question on their own pages.

The useful conclusion

Topical peptides have some human evidence, but they are not one group of equal ingredients. The result belongs to the tested formula. A product with the same peptide name still has to show its dose, delivery and full-formula result.

The evidence below comes from human studies and FDA records. It does not rank products or suggest a treatment. The rules behind every row are in How we source.

Start with the question you have

Beginner, everyday-life, evidence and emerging-format questions each have a clear place to start.

Glutathione claims versus human evidence

Detox, energy, skin and healthy-aging claims are separated from tested outcomes.

Skin: the page beside the pitch

The cosmetic clearance is not evidence, and it is doing the selling.

What GHK-Cu is, how it is used, and what has been studied

GHK-Cu is a copper-binding tripeptide complex (research compound).

How strong is each kind of source?

The difference between an approved label, a human study and a compound with no human data.

What each compound is, in one list

Class, approval status and strongest available source for every compound.

7primary sources, each with the day we read it — open the documents used on this page