GHK-Cu Benefits for Skin: Why Derms Are Paying Attention

The short answer. GHK-Cu has a large laboratory literature and a very small human one. Four human studies of topical GHK-Cu have been published or presented: a 20-person pilot in 1998, two conference abstracts in 2002, and one randomised split-face trial in 2016. A fifth, the only one indexed in PubMed as a clinical trial, tested it on laser-resurfaced skin in 2006 and found no objective benefit. The best single result is a 24.1% reduction in the volume of one measured wrinkle over eight weeks, against 15.0% for the same serum without the peptide. That is a real effect, modestly sized, in one trial that the paper's own authors credit partly to the delivery system rather than the peptide. Anything you read promising more than that is going beyond the evidence, and until this week some of what you read on this page was doing exactly that.

Corrections made to this page on October 12, 2026

This page previously carried five claims that do not survive checking. We are naming them rather than quietly deleting them, because a page about evidence should be auditable.

1. A study that does not exist. The page described "a 2023 double-blind, split-face study of 60 women (ages 40 to 65)" testing a 0.05% GHK-Cu serum over 12 weeks, reporting that IL-1β and TNF-α fell by 30%. There is no such study. No human trial of topical GHK has ever measured either cytokine. The claim circulates on peptide retail sites, where it is attributed to "Pickart & Margolina reviews"; those reviews cite six human studies and none of them is this one. The description is a garbled version of Badenhorst 2016, which had 40 participants, ran eight weeks, and measured wrinkles.
2. A safety review that does not exist. The page cited "a 2023 safety review analyzing 12 studies with 512 participants" reporting transient redness in 4.2% and mild itching in 2.8%. No systematic or safety review of topical copper peptides was published in 2022, 2023 or 2024. The first systematic review of GHK-Cu appeared in 2026 and pools no adverse event rates at all. The only real adverse event figure in this literature is from Badenhorst 2016: one participant in 40 had a minor transient reaction.
3. A journal that never published the paper. The page said Pickart, Vasquez-Soltero and Margolina published the gene expression analysis in PLOS ONE in 2014. They did not. The analysis appeared in the Journal of Analytical Oncology 3(2), 79 to 87, with Pickart FD and Majnarich J as co-authors. The old reference list repeated the wrong journal.
4. The gene figure had no denominator. "Regulation of over 4,000 genes" is not 4,000 out of a genome. See the section below.
5. Three relative percentages were presented as absolute reductions. The 55.8%, 32.8% and 31.6% figures from Badenhorst 2016 are improvements relative to a comparator, not reductions in anyone's wrinkles. See the section below.

We also dropped the previous heading "It rebuilds collagen better than vitamin C or retinol." A 20-person pilot study with no statistical comparison between treatments cannot support that sentence.

Every human study of topical GHK-Cu, and what it found

This is the whole published human record. It is shorter than the marketing around it suggests.

Study	Design	Participants	What was measured	Result	Weight it carries
Badenhorst et al. 2016, J Aging Sci 4(3):1000166	Randomised, double-blind, split-face, 8 weeks, twice daily	40 women aged 40 to 65; 39 completed	Volume and depth of one identified wrinkle near the eye, by PRIMOS 3D imaging	Wrinkle volume down 24.1% from baseline, against 15.0% for the control serum. Depth down 20.3% against 15.3%	Strongest. Peer reviewed, controlled, blinded. Funded by the product's maker; one author works at the institute that ran it
Leyden et al. 2002, AAD conference abstract	12 weeks, placebo-controlled facial cream	71 volunteers	Skin density and thickness, sagging, fine lines	Reported improvement against placebo	Moderate. A conference abstract. Never published as a full paper, so the methods cannot be read
Leyden et al. 2002, AAD conference abstract	12 weeks, eye cream against placebo and vitamin K	41 volunteers	Periorbital wrinkling, skin density	Reported better than both comparators	Moderate, same limitation
Abdulghani et al. 1998, Dis Manage Clin Outcomes 1(4):136-141	Split-body pilot, 1 month, thigh skin, no placebo arm	20 healthy volunteers, 10 per arm	Procollagen synthesis on biopsy, counted as responders	Increased from baseline in 7 of 10 on copper peptide, 5 of 10 on vitamin C, 5 of 10 on melatonin, 4 of 10 on tretinoin	Weak. The authors called it a pilot. No comparison between treatments was tested. The journal is not indexed in MEDLINE and closed in 2001
Miller et al. 2006, Arch Facial Plast Surg 8(4):252-259	Randomised, blinded evaluators, 12 weeks after CO2 laser resurfacing	13 completers	Erythema resolution, wrinkle and skin quality scores	No benefit on any objective measure. Only patient-reported satisfaction differed (p=0.04)	Notable because it is the only GHK-Cu study PubMed tags as a clinical trial, and it is negative. Copper peptide pages almost never mention it

Two things follow from reading that table rather than a summary of it. First, the evidence for GHK-Cu on normal ageing facial skin rests on one controlled trial and two abstracts. Second, the one trial run in a harder setting, on skin healing from a laser, found nothing an evaluator could see.

What the 55.8% figure actually means

Badenhorst and colleagues ran the only randomised controlled trial of a GHK-Cu facial serum. We have read the paper rather than a summary of it, and the numbers are often quoted the wrong way round, including by us until this week.

Change from baseline at 8 weeks	GHK-Cu serum	Control serum	Matrixyl 3000 product
Wrinkle volume	down 24.1%	down 15.0%	down 20.0%
Wrinkle depth	down 20.3%	down 15.3%	down 22.4%

The famous numbers are the gaps between those columns expressed as proportions. GHK-Cu improved wrinkle volume 55.8% more than the control serum did (p<0.001), and 31.6% more than the Matrixyl 3000 product (p=0.004). Depth improved 32.8% more than control (p=0.012). Against Matrixyl 3000 the depth difference was 23.4% and did not reach significance (p=0.058), which is a result the category tends to leave out.

So "GHK-Cu reduces wrinkles by 55.8%" is wrong. The measured reduction was roughly a quarter of the volume of one wrinkle, and the plain serum base on the other side of the same face managed 15% on its own.

Three caveats the paper states and the marketing does not. The product tested was GHK-Cu encapsulated in a lipid nano-carrier, and the control contained neither the peptide nor the carrier, so the trial cannot separate the two; the discussion says plainly that "it was the nano-carrier that impacted significantly on the total wrinkle reduction and not the formulation." The trial was funded by Snowberry New Zealand, which also supplied the serum. And the endpoint was one wrinkle per person, selected at baseline, in a 30 by 40 mm patch near the corner of the eye. None of that makes the result worthless. It does mean the numbers belong to that serum and not to copper peptides in general, including ours.

What the 1998 collagen study really showed

The claim that copper peptide beat vitamin C and retinoic acid on collagen comes from one month of cream on twenty people's thighs. Abdulghani and colleagues split the group in two: ten applied tretinoin to one thigh and vitamin C to the other, and ten applied melatonin to one thigh and a copper-binding peptide cream to the other. Biopsies at baseline and one month were scored for whether procollagen synthesis by dermal papillary fibroblasts had increased.

Seven of ten responded on copper peptide, five of ten on vitamin C, five of ten on melatonin, four of ten on tretinoin. Four points are worth holding onto. Those are counts of people out of ten, so one person moves the figure by ten points. Melatonin tied vitamin C, which the usual retelling drops. Copper peptide and tretinoin were never on the same people, so comparing them crosses two different groups of ten. And the outcome was how many individuals showed any increase, not how much collagen anybody made. The authors reported no statistical comparison between treatments and concluded only that all four creams "enhanced dermal collagen synthesis, although not in all individuals."

The widely repeated version of this study says "70% of women." The paper says twenty healthy subjects and never states anyone's sex, and it ran for one month rather than twelve weeks. Both errors enter the literature through a 2018 review by Pickart and Margolina, which sits next to two citations of the 12-week Leyden trials in women. Loren Pickart discovered GHK-Cu and sells copper peptide skincare, which is worth knowing when his reviews are the source for a figure.

The gene expression number, with its denominator restored

"GHK-Cu regulates over 4,000 genes" is the most repeated claim about this peptide, and it is usually given without the only piece of context that makes it meaningful.

The source is Pickart, Vasquez-Soltero, Pickart and Majnarich (2014), in the Journal of Analytical Oncology 3(2), 79 to 87. The team did not run an experiment. They downloaded three existing gene expression profiles from the Broad Institute's Connectivity Map and reanalysed them. The paper's own sentence is specific: "The number of genes affected by GHK with a percent change greater than or equal to 50% UP or DOWN is 4194/13,424 or 31.2%."

So the figure is 4,194 genes out of the 13,424 represented on the microarray used, which is 31.2% of what was measured. It is not 31% of the human genome. The human genome carries roughly 19,900 protein-coding genes, so 4,194 of them would be about 21%, and the array covered only about two thirds of them to begin with. The two halves of "4,000 genes, roughly 31% of the genome" cannot both be true.

Four further details change how much the number is worth. The profiles came from cultured PC3 prostate cancer cells and MCF7 breast cancer cells, not skin cells. The compound profiled was GHK without copper, and the authors concede they cannot exclude that a copper complex formed in the culture medium. The threshold was a raw 50% fold change with no statistical testing and no correction for multiple comparisons, which at that cutoff will flag a large share of genes for almost any compound. And the gene counts are estimates, derived by dividing probe set counts by an assumed 1.66 probe sets per gene. The split was about 2,686 genes up and 1,508 down.

None of this means the finding is fake. It means it is a hypothesis-generating screen in cancer cell lines, not a demonstration that a face cream reprograms skin.

What the laboratory evidence does support

The cell and animal literature on GHK-Cu is genuinely large, and it is the honest reason to be interested in the molecule.

Maquart and colleagues (1988, FEBS Letters 238(2), 343 to 346) showed that the GHK copper complex stimulates collagen synthesis in cultured fibroblasts at 1 to 10 nanomolar, which is a strikingly low concentration. Copper complexes of GHK reduce TNF-alpha-driven IL-6 secretion in normal human dermal fibroblasts, which is the actual basis for the anti-inflammatory interest in the peptide, and it is a cell culture result rather than a clinical one. In animal wound models GHK-Cu has accelerated wound contraction and granulation, improved angiogenesis, and raised superoxide dismutase, glutathione and ascorbic acid at the wound site. Nerve regeneration work in rats found increased outgrowth and nerve growth factor production when severed nerves were placed in collagen tubes containing GHK.

The copper chaperone argument is also sound and underrated. Free copper ions catalyse oxidative damage. Bound to GHK, the copper's redox activity is quieted in transit, and the complex delivers copper into cells where enzymes such as superoxide dismutase and ceruloplasmin need it. That is a mechanism, demonstrated in cells, and it is a reasonable thing to say. It is not a clinical outcome.

GHK also declines with age in human plasma, from roughly 200 nanograms per millilitre at 20 years old to about 80 by 60. That figure comes from Pickart's own reviews rather than from an independent cohort study, and it is a correlation with ageing rather than a demonstrated cause of anything.

Why the delivery vehicle matters

The Badenhorst trial is a useful lesson in the opposite direction from the one brands usually draw from it. The paper's authors attribute much of the wrinkle result to the lipid nano-carrier rather than to the peptide. GHK-Cu is water soluble and small, around 403 daltons, which gives it reasonable penetration potential, but published work notes that data on the skin permeability of topical GHK-Cu formulations remains thin and that the vehicle substantially affects how much active reaches the dermis.

Grass-fed tallow is a defensible carrier for a specific, checkable reason: its three dominant fatty acids are the same three that dominate human sebum, at similar proportions. Oleic acid runs roughly 40% in tallow against about 41% in sebum, palmitic about 24% against 25%, and stearic about 11% against 12%. What tallow does not have is squalene, wax esters or sapienic acid, all of which sebum contains, and no study has shown that tallow behaves like sebum on skin. We used to print a figure claiming tallow is "roughly 55% identical" to sebum. We removed it across the site because no published comparison could be found behind it.

So the honest version is that tallow is lipid-compatible with the stratum corneum and a sensible base, not that it is a proven penetration enhancer for peptides. Our Blue Balm pairs GHK-Cu with a grass-fed tallow base on that reasoning. It has not been through a clinical trial, and neither has any other copper peptide product on the market except the one Badenhorst tested.

Frequently asked questions

What does GHK-Cu do for your skin?

In the one randomised controlled trial of a GHK-Cu facial serum, eight weeks of twice-daily use reduced the volume of one measured wrinkle by 24.1% from baseline, against 15.0% for the same serum without the peptide. Laboratory work shows it stimulates collagen synthesis in fibroblasts at very low concentrations and damps an inflammatory signal in skin cells. Everything beyond that is extrapolation from cells and animals.

Is GHK-Cu better than retinol?

No, on the evidence for wrinkle reduction. Prescription retinoids have far more and better human trial data. GHK-Cu has a broader range of laboratory activity and far fewer side effects, which makes it useful for skin that cannot tolerate a retinoid. The one head-to-head comparison anyone cites, Abdulghani 1998, had ten people per arm and tested no comparison between treatments.

Does GHK-Cu really affect 4,000 genes?

A 2014 reanalysis of Connectivity Map data found that 4,194 of the 13,424 genes on the array shifted by at least 50% in cultured prostate and breast cancer cells treated with GHK. That is 31.2% of the genes measured, not of the genome, and no statistical testing was applied. It is a screening result, not a finding about skin.

How long does GHK-Cu take to work?

The trials that found anything ran 8 to 12 weeks of twice-daily use. Collagen remodelling is slow, so judging a copper peptide product before two months is premature. The one trial on laser-resurfaced skin found no objective benefit at 12 weeks.

Can you use GHK-Cu with other actives?

Generally yes. GHK-Cu is stable at neutral pH and is well tolerated alongside vitamin C, hyaluronic acid, niacinamide and retinoids. The theoretical concern about using it with a low-pH vitamin C serum at the same moment is about formulation chemistry rather than safety. See copper peptides with retinol for the detail.

What concentration of GHK-Cu is effective?

Nobody has established one in humans. Products are typically formulated between 0.05% and 0.5%. Cell studies find activity in the nanomolar range and the collagen response in Maquart's work peaked at 1 nanomolar rather than rising with dose, so higher is not reliably better.

Are there side effects?

In the only controlled trial, one of 40 participants had a minor transient reaction and stopped applying the product. There is no pooled safety analysis of topical copper peptides in the literature, despite figures to that effect circulating online. Our own page on copper peptide side effects covers what is and is not documented.

References

1. Badenhorst, T., Svirskis, D., Merrilees, M., Bolke, L., & Wu, Z. (2016). Effects of GHK-Cu on MMP and TIMP expression, collagen and elastin production, and facial wrinkle parameters. Journal of Aging Science, 4(3), 1000166. Randomised, double-blind, split-face, 40 women aged 40 to 65, 8 weeks. Funded by Snowberry New Zealand Ltd with Callaghan Innovation; trial run at Dermatest GmbH, where one co-author is affiliated.
2. Leyden, J., Stephens, T., Finkey, M. B., & Barkovic, S. (2002). Skin care benefits of copper peptide containing facial cream. American Academy of Dermatology 60th Annual Meeting, New Orleans, February 22-27, 2002, p. 68. Conference abstract, 71 volunteers, 12 weeks.
3. Leyden, J., et al. (2002). Skin care benefits of copper peptide containing eye creams. American Academy of Dermatology 60th Annual Meeting. Conference abstract, 41 volunteers, 12 weeks.
4. Miller, T. R., Wagner, J. D., Baack, B. R., & Eisbach, K. J. (2006). Effects of topical copper tripeptide complex on CO2 laser-resurfaced skin. Archives of Facial Plastic Surgery, 8(4), 252-259. Randomised with blinded evaluators, 13 completers, 12 weeks. No objective benefit found.
5. Abdulghani, A. A., Sherr, A., Shirin, S., Solodkina, G., Tapia, E. M., Wolf, B., & Gottlieb, A. B. (1998). Effects of topical creams containing vitamin C, a copper-binding peptide cream and melatonin compared with tretinoin on the ultrastructure of normal skin: a pilot clinical, histologic, and ultrastructural study. Disease Management and Clinical Outcomes, 1(4), 136-141. Pilot, 20 healthy volunteers, 10 per arm, thigh skin, 1 month. Not indexed in MEDLINE.
6. Pickart, L., Vasquez-Soltero, J. M., Pickart, F. D., & Majnarich, J. (2014). GHK, the human skin remodeling peptide, induces anti-cancer expression of numerous caspase, growth regulatory, and DNA repair genes. Journal of Analytical Oncology, 3(2), 79-87. In silico reanalysis of three Broad Institute Connectivity Map profiles from PC3 and MCF7 cancer cell lines at 1 micromolar GHK. Source of the 4,194 of 13,424 genes figure.
7. Maquart, F. X., Pickart, L., Laurent, M., Gillery, P., Monboisse, J. C., & Borel, J. P. (1988). Stimulation of collagen synthesis in fibroblast cultures by the tripeptide-copper complex glycyl-L-histidyl-L-lysine-Cu2+. FEBS Letters, 238(2), 343-346.
8. Pickart, L., & Margolina, A. (2018). Regenerative and protective actions of the GHK-Cu peptide in the light of the new gene data. International Journal of Molecular Sciences, 19(7), 1987. Review. Pickart discovered GHK-Cu and has commercial interests in copper peptide skincare. This review is the origin of the "70% of women" and "12 weeks" errors about Abdulghani 1998.
9. Pickart, L., Vasquez-Soltero, J. M., & Margolina, A. (2015). GHK peptide as a natural modulator of multiple cellular pathways in skin regeneration. BioMed Research International, 2015, 648108.
10. Pickart, L. (2012). The human tri-peptide GHK and tissue remodeling. Oxidative Medicine and Cellular Longevity, 2012, 324832.

Related reading

• GHK-Cu copper peptides: the complete guide
• GHK-Cu before and after: what the research shows
• Copper peptide vs retinol
• What is GHK-Cu?
• Shop the copper peptide range