"Visible results in four weeks" is the most common promise in peptide skincare, and it is not sourced to anything. What is sourced is the set of trials themselves — and those trials do not answer the question of how long a peptide takes to work so much as record when each research group decided to look.
On 2026-09-13 we searched PubMed for every randomized controlled trial of a topical peptide in skin ageing, photoageing or wrinkles, and read all sixteen. This page is their clocks.
Every trial, and when it measured
| Year | Journal | What was tested | Treatment period | Measurement points |
|---|---|---|---|---|
| 2026 | J Cosmet Dermatol | Recombinant EGF (oligopeptide-1) serum vs human-derived growth factor serum, 20 women | 12 weeks | Study visits through week 12 |
| 2026 | J Cosmet Dermatol | Injected recombinant collagen III ± a multi-peptide serum, 54 participants, split-face | 12 weeks | Primary endpoint week 4; further outcomes week 12 |
| 2025 | J Cosmet Dermatol | Cyclized hexapeptide-9 vs retinol vs vehicle, both at 0.002% | 56 days | End of treatment |
| 2025 | J Cosmet Dermatol | OS-01 (Peptide 14) formulation vs moisturizer, 60 women aged 60–90 | 12 weeks | Week 12 |
| 2024 | J Cosmet Dermatol | Oral collagen peptides, 85 women (a supplement, not a topical) | 84 days | Baseline, day 28, day 84 |
| 2022 | Med Arch | Postbiotic cream containing peptides, 50 volunteers, split-face | 4 weeks | End of treatment |
| 2021 | Photochem Photobiol | Sunscreens with ascorbyl tetraisopalmitate and rice peptides, 60 women | Daily use study | End of treatment |
| 2020 | J Cosmet Dermatol | Tripeptide/hexapeptide serum as a laser adjunct, 20 subjects, split-face | 14 days before, 60 days after first laser | Days 30 and 60 |
| 2020 | Int J Cosmet Sci | pep_RTE62G, an AI-discovered pea peptide, proof of concept | 28 days | End of treatment |
| 2018 | J Drugs Dermatol | Alpha and beta defensin regimen (serum, cream, mask), 44 women | 12 weeks | Baseline, 6 weeks, 12 weeks |
| 2016 | J Cosmet Dermatol | Zeaxanthin supplement ± a topical containing peptides, 31 subjects | 12 weeks | Weeks 2, 4, 6, 8, 12 |
| 2016 | J Drugs Dermatol | Matrikine-like micro-protein complex vs two growth-factor materials, 133 women | 6 months | 1, 2, 3 and 6 months |
| 2012 | J Dermatolog Treat | Oral carnosine formulation, 42 subjects | 3 months dosing, 1 month follow-up | 4 months total |
| 2010 | Br J Dermatol | Niacinamide/peptide/retinyl propionate regimen vs 0.02% tretinoin, 196 women | 8 weeks, with a 25-subject cohort to 24 weeks | Weeks 8 and 24 |
| 2009 | J Cosmet Dermatol | Products containing Pal-KT and Pal-KTTKS, two split-face studies, 42 and 35 women | 4 weeks | Weeks 2 and 4 |
| 2008 | Br J Dermatol | Lipopentapeptide and lupin-peptide cream under occlusion, 9 volunteers, vs retinoic acid | 12 days | Biopsy at 12 days |
Two things fall out of that table immediately. The shortest study in the category is twelve days and the longest is six months, a twenty-fold spread. And the earliest instrumented look anyone took was week two — which means that on the published record, nothing at all is known about what a peptide does in the first fortnight.
The one curve that exists
Fifteen of the sixteen trials report a result at the end of treatment. One reports the same endpoint at several visits, and it is the closest thing this category has to an answer.
Dreher's 2016 study followed 133 women aged 38 to 65 using a blend of matrikines and matrikine-like peptides twice daily for six months, with an independent blinded evaluator grading wrinkles on a five-point scale at one, three and six months. The share of women improving by at least one point:
| Visit | Periorbital wrinkles | Perioral wrinkles |
|---|---|---|
| 1 month | 28% | 39% |
| 3 months | 65% | 41% |
| 6 months | 81% | 59% |
Read as a whole, that is a slow accrual rather than an onset. At one month roughly three in ten had a gradeable change around the eyes; at six months roughly eight in ten did. Anyone stopping at the four-week mark on the grounds that nothing had happened would have been in the majority — and, on this trial's own numbers, wrong to conclude the product did nothing.
The same trial also shows why a single number cannot be given. Skin elasticity improved significantly at two months (20%) and stood at 16% at six, so one endpoint was near its ceiling while another was still climbing. "How long do peptides take to work" has a different answer depending on what is being asked to work.
Why the clocks are shorter than the biology
The trials run for weeks. The structure most of them describe changing turns over across years.
The best published estimate for the replacement rate of the human dermal matrix comes from Verzijl and colleagues in 2000, who used the racemization of aspartic acid — a slow, spontaneous chemical change whose extent tracks how long a protein has sat in the tissue — to put the half-life of skin collagen at about 15 years, against 117 years for cartilage collagen. Those were the first quantitative estimates published for either tissue.
This is not an argument that nothing can happen in twelve weeks. The 2008 patch-test study found increased deposition of fibrillin-1 and procollagen I after twelve days under occlusion, so new matrix protein can be laid down quickly and detected in a biopsy. It is an argument about what a visible change at week twelve can be. It is unlikely to be the wholesale replacement of a dermis with a fifteen-year half-life, and much more likely to be some combination of hydration, barrier repair, epidermal thickening and light-scattering — which is also what the endpoints in this set actually measure. The defensin trial's significant histological finding was increased epidermal thickness. The OS-01 trial's were barrier function and hydration. That is the same pattern our count of the human evidence for skin peptides finds across the category: the endpoints that move are the ones nearest the surface.
Where the marketing number comes from
None of the sixteen trials supports "results in four weeks" as a general statement, and two of them complicate it in opposite directions.
The 2010 comparison against prescription tretinoin found the cosmetic regimen significantly ahead at 8 weeks and level by 24 — an early advantage that later disappeared. The six-month matrikine study found the opposite shape, a benefit that kept growing. A brand can therefore find published support for "fast" or for "keeps improving" depending on which trial it reaches for, and both citations would be honest in isolation and misleading as a claim about the category.
What can be said without stretching: week 2 is the earliest anyone has looked, week 12 is the standard design, and the only trial that tracked a responder rate over time found most of its responders arrived after month one.
Format matters here too, and not in the direction the shelf implies. The trials above are all leave-on products used for months; a peptide in a cleanser gets seconds of contact per wash, and the same active sold for arms and legs has no randomized trial at all to put a clock on.
What would settle it
A trial measuring one named peptide at a disclosed concentration against its own vehicle, with the same instrumented endpoint at weeks 2, 4, 8, 12 and 24, in a sample large enough to report a responder rate at each visit. Nothing in this set is that trial. The closest is a six-month study of a proprietary blend graded by eye.
Until then, the useful version of the question is not how long peptides take to work, but what is being waited for. For barrier and hydration measures the published windows are short. For anything described as rebuilding the dermis, the tissue's own clock is the limit, and it is measured in years.
Sources
- Dreher F. A novel matrikine-like micro-protein complex (MPC) technology for topical skin rejuvenation. J Drugs Dermatol 2016;15(4):457–64. PMID 27050701.
- Kaczvinsky JR, Griffiths CE, Schnicker MS, Li J. Efficacy of anti-aging products for periorbital wrinkles as measured by 3-D imaging. J Cosmet Dermatol 2009;8(3):228–33. PMID 19735523.
- Fu JJ, Hillebrand GG, Raleigh P, 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. PMID 20374604.
- Watson RE, Long SP, Bowden JJ, et al. Repair of photoaged dermal matrix by topical application of a cosmetic 'antiageing' product. Br J Dermatol 2008;158(3):472–7. PMID 18070204.
- Verzijl N, DeGroot J, Thorpe SR, et al. Effect of collagen turnover on the accumulation of advanced glycation end products. J Biol Chem 2000;275(50):39027–31. PMID 10976109.
- Taub A, Bucay V, Keller G, et al. Multi-center, double-blind, vehicle-controlled clinical trial of an alpha and beta defensin-containing anti-aging skin care regimen. J Drugs Dermatol 2018;17(4):426–41. PMID 29601620.
- Zonari A, Brace LE, Buhrer LB, et al. OS-01 peptide topical formulation improves skin barrier function and reduces systemic inflammation markers: a pilot 12-week clinical trial. J Cosmet Dermatol 2025;24(4):e70169. PMID 40193112.
- PubMed search, run 2026-09-13: peptide(s) AND (skin aging OR photoaging OR wrinkle*) AND (topical OR cosmetic OR cream OR serum), filtered to randomized controlled trial publication type — 16 records.