Ask a search engine whether peptides help rosacea and you will find product pages. Ask PubMed and you find something more interesting: a large literature in which peptides appear constantly, and never as a treatment.
Counted on 2026-09-15: 5,108 records mention rosacea, 217 of them are randomized controlled trials, and 90 join rosacea to peptides. Of those 90, two carry the randomized-trial publication type. Neither tests a peptide.
The two "trials" and what they actually study
| Study | Design | What was compared | Is a peptide the treatment? |
|---|---|---|---|
| Picosse et al., Int J Dermatol 2025 | Comparative, marker-based | Oral isotretinoin versus doxycycline in papulopustular rosacea, read out on immunohistochemical markers | No — the peptides are measured, not given |
| George et al., J Drugs Dermatol 2023 | Randomized | A low-molecular-weight heparan sulfate analog for erythema | No — heparan sulfate is a glycosaminoglycan, a sugar chain, not a peptide |
The first counts peptides as markers of what the drugs did. The second matched the search because the word appears somewhere in its text, and its active is not a peptide at all.
What the other 88 records are about
Two molecules dominate, and both are made by the body.
Cathelicidin, in 65 of the 90. These records name cathelicidin, its fragment LL-37, the kallikrein enzymes that cut it, or antimicrobial peptides generally. The paper underneath them is Yamasaki and colleagues in Nature Medicine in 2007: people with rosacea express abnormally high cathelicidin in facial skin, and the processed forms of the peptide differ from those found in unaffected skin, because stratum corneum tryptic enzyme activity is raised. Inject those rosacea forms into mouse skin, add the enzyme, or delete the gene for the protease inhibitor that restrains it, and inflammation increases.
So in the single largest strand of rosacea research, an antimicrobial peptide is a mediator of the disease. Not an ingredient. Not a treatment. The thing there is too much of, cut the wrong way.
CGRP, in 19 of the 90. Calcitonin gene-related peptide is a neuropeptide that dilates blood vessels and is the target of the migraine biologics. A 2025 study reports differing serum CGRP levels between neurogenic and non-neurogenic rosacea; a 2025 review in Clinical and Experimental Dermatology is titled for the arrival of biologics in rosacea through CGRP inhibition.
And there is human treatment data, which deserves to be reported precisely.
The one human treatment result, stated with its design
Wienholtz and colleagues at the Danish Headache Center and Bispebjerg Hospital gave 30 adults with rosacea 140 mg of erenumab — an antibody against the CGRP receptor — subcutaneously every four weeks for 12 weeks, with no other rosacea treatment allowed.
| Endpoint | Baseline (mean days) | Change | 95% CI |
|---|---|---|---|
| Days with moderate-to-extreme flushing | 23.6 | −6.9 | −10.4 to −3.4 (P < 0.001) |
| Days with moderate-to-severe erythema | 15.2 | −8.1 | −12.5 to −3.7 (P < 0.001) |
The limits belong in the same breath as the numbers. It was open-label, single-group and nonrandomized — every participant knew they were being treated and there was no comparison arm, so the run-in period is the only control. Twenty-seven of 30 completed. Two co-authors were employees of the manufacturer. A third of participants had transient constipation and 13% had transient worsening of flushing. The authors' own conclusion asks for larger randomized trials.
Even taken at face value, notice what the result is: an injected antibody that blocks a peptide receptor for three months. The direction of travel in rosacea therapeutics is subtraction.
There is one population-level record pointing the same way and we will not overstate it: a 2024 JAMA Dermatology research letter examined whether CGRP inhibition is associated with reduced rates of developing acne or rosacea in migraine patients. Its abstract states the question and not the answer, so no effect estimate is quoted here.
What this means for a peptide serum
Nothing in the record says a cosmetic peptide worsens rosacea. Nothing says it helps. The class has not been tested on this condition, for any endpoint, in any design.
What can be said is that the inference people draw from marketing runs against the shape of the literature. A serum's pitch is that adding peptides to skin improves it. In rosacea research, the abundant peptide is the inflammatory one, its abnormal processing is the mechanism, and the intervention with human data works by blocking a different peptide's receptor. Those are three separate molecules and three separate stories, and none of them is "apply peptides to the face".
This is the third time in a fortnight that a peptide search on this site turned out to be measuring something other than an ingredient — the stretch-mark literature is largely brain anatomy and Cushing's syndrome, and the toner literature is largely a cited author's surname. The general rule is worth stating once: in biology, "peptide" is a word for a size of molecule, not a category of product.
The study that would answer it
Rosacea is unusually well set up for a proper test. It has validated endpoints — flushing days, erythema days, investigator assessment — and the 2019 split-face design used elsewhere in cosmetics works here too. A named peptide at a disclosed concentration, the identical vehicle on the other side of the face, twelve weeks, with tolerability recorded on reactive skin as a primary rather than an afterthought.
Until then, the trial count on this question is zero, and a product page that implies otherwise is describing a hope. Our entry on what not to layer with copper peptides covers the little that is actually known about irritation from this ingredient class.
The sharpest contrast is with eczema, where the same molecule is abnormal in the opposite direction — expressed at lower levels in atopic skin than in psoriatic skin — and where, as our count of the eczema trials sets out, the one antimicrobial peptide ever applied to it cut Staphylococcus aureus by 93.5% against vehicle and changed no symptom score at all.
Cathelicidin shows up again one condition over, in a different balance: the antimicrobial-peptide literature in acne runs to 144 records of which 44 are reviews and one is a trial of any design — a mechanism studied far more often than it is tested, where rosacea at least has a direct measurement behind its version of the story.
Sources
- Yamasaki K, Di Nardo A, Bardan A, et al. Increased serine protease activity and cathelicidin promotes skin inflammation in rosacea. Nat Med 2007;13(8):975–80. PMID 17676051.
- Wienholtz NKF, Christensen CE, Do TP, et al. Erenumab for treatment of persistent erythema and flushing in rosacea: a nonrandomized controlled trial. JAMA Dermatol 2024;160(6):612–19. PMID 38630457. ClinicalTrials.gov NCT04419259.
- Thang CJ, Lai J, Garate D, et al. Calcitonin gene-related peptide inhibition and development of acne and rosacea. JAMA Dermatol 2024;160(8):895–98. PMID 38985467.
- Picosse F, et al. A comparative exploration of immunohistochemical markers in patients with papulopustular rosacea undergoing treatment with oral isotretinoin versus doxycycline. Int J Dermatol 2025. PMID 39097930.
- George R, et al. Reduction of erythema in moderate-severe rosacea by a low molecular weight heparan sulfate analog (HSA). J Drugs Dermatol 2023;22(6). PMID 37276169.
- PubMed counts, run 2026-09-15 via the NCBI E-utilities
esearchendpoint:rosacea[tiab]— 5,108;rosacea[tiab] AND "randomized controlled trial"[pt]— 217; the same withAND ivermectin[tiab]— 12;rosacea[tiab] AND peptide[tiab]— 90 records and 2 with the trial publication type; that set with cathelicidin, LL-37, kallikrein or antimicrobial terms — 65; with CGRP terms — 19. PubMed.