Explained · 8 min read

Why GHK-Cu injections burn and what week three really is

The sting has three untested explanations and the week-three ‘crash’ has one, and it’s a rat. I read the actual papers so you don’t have to take either story on faith.

In this article ▾
  1. Why the injection burns in the first place
  2. The boring answer nobody has measured
  3. The “copper uglies” story everyone repeats
  4. Where “week three to five” actually came from
  5. The dose curve nobody argues about
  6. Topical evidence exists. Injected evidence doesn't
  7. A normal reaction, and where that stops being your call
  8. The bottom line

This is the written version of GHK-Cu Injections: The Burn Nobody Warns You About — watch it instead if you’d rather.

Somewhere tonight, someone is standing in a bathroom pre-loading a painkiller and an antihistamine before they inject a copper peptide, because a comment section told them that's how you survive it. That is not a protocol. It's a person negotiating with a chemical they don't understand — and I don't blame them, because almost nobody explains why GHK-Cu hurts, or what's actually happening a few weeks in when their skin looks worse instead of better. I'm a Chartered Chemical Process Engineer, and I take no money from anybody who sells peptides, so let's look at what the underlying studies actually say about both.

Why the injection burns in the first place

Start with the one thing that's genuinely solid: it isn't a true allergy. An allergic reaction needs a first exposure to sensitise the immune system before it can react — it physically cannot hurt the first time you meet a substance. GHK-Cu stings on shot one, every time. Whatever is happening, it isn't that.

Beyond that, there are three competing explanations, and none of them has been tested directly in a human. One is mast-cell degranulation — immune cells releasing histamine without a true allergic trigger — an idea borrowed from copper's general effects on immune cells rather than from any study of injection pain itself. Another is free copper ions irritating tissue directly. The third, and the one I'd bet on, is the least dramatic: the properties of the solution itself.

The boring answer nobody has measured

There's a published review of what actually makes a subcutaneous injection hurt, and the causes are unglamorous: a pH that doesn't match body tissue, the wrong salt concentration, too much buffering, too much volume delivered too fast. A reconstituted GHK-Cu vial is, chemically, an unbuffered peptide sitting in bacteriostatic water with nothing added to bring it in line with your tissue.

Separately, GHK-Cu's own stability window — the pH range across which the compound stays intact on the shelf — runs from about pH 4.5 to 7.4. That's a shelf-life figure, not a measurement of what any given reconstituted vial's actual pH sits at, and the two shouldn't be conflated. Nobody has published that measurement either, for a peptide thousands of people inject every week.

These numbers are study facts, not recommendations Where this article quotes a figure, it is reporting what a specific study measured, in a specific system, by a specific route. That is not a suggestion for anyone reading it. This article names no vendor and no source, recommends nothing, and is not medical advice.
Free guide · goes with this article How to Read a Peptide Study Like an Engineer

The eight questions to ask of any study before you believe its headline. Comes with the other three free guides, for one email.

Get the free guides

The “copper uglies” story everyone repeats

Then there's the thing people are actually frightened of. Around week three or four, some people look in the mirror and their skin looks worse — duller, crepier, older. The explanation that circulates everywhere is that GHK-Cu is breaking down old collagen before laying down new collagen: demolition before construction, so you push through.

It's a tidy story, and the collagen literature doesn't actually support it. The enzyme that cuts intact, structural collagen is MMP-1, interstitial collagenase — the only demolition crew that matters. In the key study everyone is indirectly drawing on, MMP-1 was measured, and it didn't move. What rose instead was MMP-2, a gelatinase that clears up fragments that are already broken — the cleanup crew, not the wrecking ball. In the same experiments, GHK-Cu simultaneously raised TIMP-1 and TIMP-2, the natural brakes on those enzymes. Both directions moved together. The authors' own word for it is "remodeling," and a separate in-vivo study found the tissue accumulating net collagen and connective-tissue matrix throughout, not losing it.

Where “week three to five” actually came from

So where does the timeline come from? There is exactly one paper in the whole copper-peptide literature reporting a late rise in one of these enzymes, and it's a rat. Specifically: a rat with a stainless-steel wound chamber implanted under the skin, receiving direct injections, with MMP-2 rising on days 18 and 22 of the experiment.

Days eighteen to twenty-two got rounded up to "week three." Somewhere along the way, a rodent wound-healing timeline measured in days got transplanted onto human faces measured in weeks — while quietly dropping the facts that it was the wrong enzyme (MMP-2, not MMP-1), the wrong model (an open surgical wound in a rat, not intact human skin), and a paper that also reported the tissue gaining collagen the whole time. Three category errors stacked on top of one another, repeated until they read as fact. There is no peer-reviewed report of a human copper-peptide "purge" phase — not weak evidence, none.

The dose curve nobody argues about

Here's the part that gets less attention than it deserves. The collagen-stimulating effect that everyone is chasing was measured in human fibroblasts, and it's a bell curve peaking at around one nanomolar. Push the concentration higher and the effect tails off rather than climbing further. A reconstituted vial at a typical strength sits roughly a million times past that published optimum — which is a genuinely enormous distance down a curve that has never been mapped in living human tissue at those concentrations.

I want to be precise about what that does and doesn't mean. Nobody has shown that higher concentrations flip the effect into destroying collagen — that would be the same overclaiming crime with different paperwork, and I'm not going to commit it just because it would make a tidier ending. What's true is narrower and less satisfying: the effect people are hoping for was characterised at a concentration nothing being injected resembles. If you want the actual arithmetic behind reconstituting a vial and where a typical strength lands on that curve, I've laid it out separately in the reconstitution and shelf-life piece.

Topical evidence exists. Injected evidence doesn't

The evidence base for injecting this compound is thinner than most people assume, and thinner still than the confidence online suggests. The one controlled human trial of topical GHK-Cu — thirteen patients following laser resurfacing — came back negative on every objective measurement instruments could take, and positive only on how patients said their skin felt. That's the best human data that exists for the topical form.

For the injected form, there is no published human trial at all. No pharmacokinetic data, no safety study, nothing establishing what a "typical" dose even means. If you're newer to this compound and want the grounding I didn't cover here, I wrote the beginner's version separately in the GHK-Cu and Epitalon explainer, and if you want to see how it stacks up against other compounds people reach for on appearance, that's in the looks-peptides ranking.

A normal reaction, and where that stops being your call

This is the part I care most about getting right, so I'll say plainly what I'm not doing: I'm not giving you a self-treatment checklist. I'm reporting what injection-site literature — vaccine-safety grading criteria and public health injection-reaction guidance — actually describes, so you know roughly what "ordinary" looks like and, more importantly, when a reaction has left that category.

The pattern that literature describes for an expected local reaction is redness confined to a modest area, tenderness that peaks within a day or two, and the whole thing fading within about a week. The more useful test isn't size on its own — it's direction. With an ordinary reaction, as any redness spreads, the tenderness generally eases off. With a genuine infection, it spreads and the pain gets worse as it does. The grading criteria used in vaccine trials treat things like skin breaking down or turning dusky or dark as their most severe category, alongside fever or feeling systemically unwell.

None of that is a rule for you to apply to yourself at home. It's a description of what the documents say a normal reaction looks like and what a clinician looks for when it isn't one. If a reaction is spreading with worsening pain, comes with fever or feeling properly unwell, shows red streaking, forms a lump with fluid moving inside it, or the skin itself starts to break down or discolour — that judgement belongs to a doctor, not to a reader of a video or a blog post, and it belongs to them today, not after you've tried to wait it out.

One of these has never been measured. The other is a rat, in days, wearing a human costume.

The bottom line

GHK-Cu's collagen biology is real and, at the right concentration, genuinely interesting — the remodelling data holds up. But the two things people actually experience when they inject it are both explained with a confidence nobody has earned. The burn has three candidate causes and zero of them are confirmed in a human being. The "week three crash" isn't weak evidence for a purge phase; it's a rat's wound chamber, measured in days, mistranslated onto a human face measured in weeks.

If you're at the point of wondering whether you even know what's in the vial to begin with, that's a separate, solvable problem — I keep four free guides for reading a label and a lab report without needing a chemistry degree. Everything else here is just what the papers say. Keep safe, keep sceptical.

Before you buy anything How to Read a Peptide Study Like an Engineer is free — the eight questions to ask of any study before you believe its headline. It comes with the other three free guides for one email. If you want the full method behind them, that’s the Field-Guide Series.

Sources

  1. Siméon et al. 1999, J Invest Dermatol 112:957-64 — the rat subcutaneous wound-chamber study; MMP-1 unchanged, MMP-2 rising at days 18 and 22. The paper the whole “week 3 to 5 purge” story is misread from — https://doi.org/10.1046/j.1523-1747.1999.00606.x
  2. Siméon et al. 2000, Life Sci 67:2257-65 — MMP-2, TIMP-1 and TIMP-2 all rising together; the authors’ own word is “remodeling,” not demolition — https://doi.org/10.1016/S0024-3205(00)00803-1
  3. Maquart et al. 1988, FEBS Lett 238:343-6 — the bell-shaped dose response in human fibroblasts, peaking at 1 nanomolar — https://doi.org/10.1016/0014-5793(88)80509-X
  4. Maquart et al. 1993, J Clin Invest 92:2368-76 — net collagen and GAG accumulation in vivo; the tissue gains collagen, it does not lose it — https://doi.org/10.1172/JCI116842
  5. Miller et al. 2006, Arch Facial Plast Surg 8:252-9 — the only controlled human trial of topical GHK-Cu: negative on every objective endpoint, positive only on patient satisfaction — https://doi.org/10.1001/archfaci.8.4.252
  6. Usach et al. 2019, Adv Ther 36:2986-96 — pH, osmolality, buffer strength and volume as the documented causes of subcutaneous injection pain — https://doi.org/10.1007/s12325-019-01101-6
  7. Badenhorst et al. 2016, Pharm Dev Technol 21:152-60 — GHK-Cu’s stability window of pH 4.5 to 7.4 — https://doi.org/10.3109/10837450.2014.979944
  8. FDA/CBER, Toxicity Grading Scale for Preventive Vaccine Clinical Trials — the erythema size bands and the necrosis threshold — https://www.fda.gov/media/73679/download
  9. NCIRS, “Injection site reactions” — the direction-not-size discriminator, and the peak-at-24-to-48-hours, resolves-within-a-week course — https://www.ncirs.org.au/fact-sheets-faqs/injection-site-reactions

Educational and research purposes only — not medical advice. Peptide Corner does not recommend any vendor, source, or dose. Keep safe, keep skeptical.