How skin ages
The dermis is a scaffold of collagen (mostly types I and III) and elastin, built and maintained by fibroblasts. From the mid-twenties, collagen production falls by roughly 1% a year; fibroblasts become fewer and less active, the fibres they made become fragmented and cross-linked, and the scaffold thins. In women the decline steepens sharply at menopause — about 30% of skin collagen is lost in the first five years after it, which is why the change in the late forties and early fifties feels sudden.
That is intrinsic ageing, and it is modest. The larger part of what people see in the mirror is photoageing: ultraviolet light activating the enzymes that chew up collagen and suppressing the ones that build it. A 2013 study of 298 women (Flament et al.) attributed about 80% of visible facial ageing to UV exposure. Smoking is the second-largest external factor, high sugar and glycation third, and sleep deprivation measurably worsens barrier function and the appearance of fine lines within days.
What works — ranked by evidence
| Intervention | Evidence | What was shown |
|---|---|---|
| Daily broad-spectrum SPF 30+ | Strong — randomised trial | Hughes 2013, Annals of Internal Medicine: 903 Australian adults randomised to daily sunscreen or discretionary use for 4.5 years. Daily users showed 24% less skin ageing by microtopography, with no visible ageing at all in some. The single best-proven cosmetic intervention there is. |
| Topical retinoids (tretinoin, and to a lesser degree retinal and retinol) | Strong — randomised trials since 1988 | Weiss 1988, JAMA, and dozens since: tretinoin 0.025–0.1% increases collagen I synthesis, thickens the epidermis, reduces fine lines and pigment. Effects visible at 3–6 months and continuing for years. Irritation is the limiting factor; tretinoin is prescription in most countries, adapalene and retinal are not. |
| Topical vitamin C (L-ascorbic acid 10–20%) | Moderate | Cofactor for collagen synthesis and a UV adjunct; small split-face trials show improvement. Unstable in the bottle. |
| Not smoking; sleep; protein | Strong (cohorts) | Smokers show roughly a decade of extra skin ageing in twin studies; collagen is made of amino acids and a low-protein diet shows. |
| GHK-Cu (copper peptide) | Early — in-vitro and small human studies | A tripeptide that falls with age and in cell studies switches on collagen, elastin and repair genes; small topical trials report thicker dermis and fewer wrinkles. Promising, under-studied, and most products contain too little. |
| Red / near-infrared light (630–850 nm) | Moderate — one good RCT | Wunsch & Matuschka 2014: 136 adults, 30 sessions over 15 weeks; improved skin roughness and increased collagen density on ultrasound versus controls. Mechanism via mitochondrial cytochrome c oxidase. |
| Oral collagen peptides | Moderate but industry-funded | Meta-analyses (de Miranda 2021, 19 trials, 1,125 people) report modest gains in hydration and elasticity at 2.5–10 g a day. Nearly every trial was funded by a manufacturer; the effect, if real, is small and may be a protein effect. |
| Procedures: microneedling, fractional laser, PRP | Moderate | Controlled injury that provokes fibroblasts. Real effects, operator-dependent, repeated. |
What does not work
- "Collagen creams". Collagen is a large molecule that cannot cross the skin barrier. It sits on the surface as a moisturiser.
- Most "peptide" serums. The marketing borrows from GHK-Cu and matrixyl; the concentrations are usually cosmetic and the evidence is a company poster.
- "Detox", "collagen-boosting" drinks, and high-dose biotin for anyone not deficient.
- Stem-cell and exosome creams. Plant stem cells do nothing for human skin; human-derived exosomes in a jar are unregulated and unproven.
- LED masks that are mostly blue. The red and near-infrared wavelengths and a real irradiance are what the trial used; many masks give neither.
- SPF in your foundation. The amount applied is a fraction of the dose the rating assumes.
A routine that reflects the evidence
| When | What | Why |
|---|---|---|
| Morning | Vitamin C serum, moisturiser, SPF 30–50 — two finger-lengths for the face and neck, reapplied if outdoors | The one thing the trial proved. |
| Evening | Cleanse; tretinoin or retinal, starting twice a week and building; moisturiser | Decades of trial evidence. Buffer with moisturiser if it stings. |
| Evening, alternate | GHK-Cu serum on non-retinoid nights | Studied for collagen and repair; avoid mixing with vitamin C in the same application. |
| 3–5× a week | 10–15 minutes of red / near-infrared light at a device that states its irradiance | The 2014 protocol. |
| Always | Sleep 7–8 h, protein at 1.6 g/kg, no smoking, limited alcohol | The scaffold is made of what you eat and repaired while you sleep. |
What to measure
Skin is the rare organ where the measurement is a photograph, and that is the one most people never take properly.
- Standardised photographs: same light, same distance, no make-up, front and both profiles, every three months. Retinoid and light effects take six months and are invisible day to day.
- A dermatologist's mole check yearly from forty, or sooner with a family history; the point of SPF is not cosmetic.
- Labs that show in the skin: ferritin, vitamin D, thyroid, HbA1c (glycation stiffens collagen), estradiol in women around menopause, and zinc if wound healing is slow.
- Sleep duration, because the barrier measurably degrades after a few short nights.
Where peptides and therapies fit
The research library has a page on GHK-Cu covering what it is studied for and the forms in which it is used, and the GLOW page covers the combination most people ask about for skin and repair. Among therapies, red light has its own page with the wavelengths, irradiance and session counts the trials used, and PRP covers the injected procedures. The placement rule is the same as everywhere else on this site: SPF and a retinoid are the base; nothing in the compound or therapy layer substitutes for them, and a protocol that skips them to reach the interesting part is backwards.
Common mistakes
- A €200 serum and no sunscreen. The trial-proven intervention costs the least.
- Starting tretinoin nightly at full strength, peeling for three weeks, and giving up. Twice a week, buffered, building over months.
- Expecting a result in a month. Collagen remodelling is a six-month process; photograph, do not stare.
- Judging a red-light device by its price rather than its stated wavelength and irradiance.
- Oral collagen as a substitute for protein. If it works, it is probably because it is protein.
- Ignoring the neck, hands and ears — the places that give an age away.
Frequently asked questions
What is the single most effective thing for skin ageing?
Daily broad-spectrum sunscreen. In a 4.5-year randomised trial of 903 adults, daily use produced 24% less measurable skin ageing than discretionary use, and UV accounts for about 80% of visible facial ageing. A retinoid at night is the second.
Does GHK-Cu work for skin?
It is studied for switching on collagen, elastin and repair genes in cell studies, and small topical trials report thicker dermis and fewer wrinkles. The evidence is early and most products contain too little to matter. It is a reasonable addition on top of SPF and a retinoid, not a replacement.
Is red light therapy evidence-based for skin?
There is one good randomised trial (136 adults, 30 sessions over 15 weeks) showing improved roughness and increased collagen density on ultrasound at 630–850 nm. Devices vary enormously; the wavelength and the irradiance at the skin are what matter.
Do collagen supplements do anything?
Meta-analyses of mostly manufacturer-funded trials report modest improvements in hydration and elasticity at 2.5–10 g a day. The effect is small and may simply be extra protein. Hit 1.6 g/kg of protein first; collagen on top is optional.
Why does skin change so fast around menopause?
Estradiol maintains collagen, and about 30% of skin collagen is lost in the first five years after menopause. Topical estriol and systemic hormone therapy both slow it, which is a conversation for a physician; retinoids, SPF and protein still do the rest.
Want a plan built on your own numbers?
The skin protocol starts with the two things that work, adds GHK-Cu and red light where they fit, and is written against your labs like every other. €399, follow-up included.
Build my protocol Book a 15-minute call