- Three things have real evidence. Daily sunscreen, a retinoid, and not smoking. Everything else on the shelf ranges from modest to unproven
- There is an actual randomised trial on sunscreen and aging. 903 adults, 4.5 years: the daily-sunscreen group showed no detectable increase in skin aging, and 24% less than the group using it at their discretion
- Most of what you're seeing isn't "aging." It's accumulated UV. The clearest illustration is a truck driver in the New England Journal of Medicine whose window side aged dramatically more than his other side over 28 years
- How much sunscreen you use matters more than which one. SPF is measured at 2.0 mg/cm² in the lab. Most people apply a fraction of that, so the real protection is well under the number on the bottle
- Retinoids are the only topical that rebuilds rather than protects. Tretinoin has seven randomised trials behind it; over-the-counter retinol is weaker but genuinely supported
- Collagen supplements are more positive than sceptics expect and weaker than the ads imply. The main meta-analysis drew a published complaint about the trials it pooled, and the 2026 review declined to pool at all
- Your 40s is the last decade where prevention beats correction. Damage already done is harder and more expensive to reverse than damage not yet accumulated
Somewhere in your early forties the mirror changes its mind. The lines that used to appear when you smiled stop leaving when you stop. Skin that bounced back from a bad night takes three days. Pigmentation you'd never noticed arrives on your cheekbones as if it had been there all along — which, in a sense, it had.
The skincare industry has an answer for this, and the answer costs about $180 billion a year. Most of it is not supported by anything you could show a sceptical reader. But a small part of it is supported extremely well — better than most things in nutrition, better than most supplements — and that part is worth knowing precisely, because it's cheap and it works and almost nobody does it properly.
What's actually happening
Skin ages two ways at once, and they are not the same process.
Intrinsic aging is the clock: collagen synthesis slows, cell turnover lengthens, fat pads shift and thin. It happens to skin that has never seen the sun, on the inside of your upper arm, and there is very little you can do about it.
Extrinsic aging is everything else — overwhelmingly ultraviolet light, plus smoking and pollution. It's what produces the deep wrinkling, the leathery texture, the broken capillaries and the uneven pigment that people actually mean when they say someone looks older than their age. And it is almost entirely modifiable.
The single most persuasive piece of evidence for that distinction isn't a trial at all. In 2012 the New England Journal of Medicine published a photograph of a 69-year-old man who had driven a delivery truck for 28 years. The left side of his face — the window side — was deeply furrowed and thickened. The right side looked like a normal man in his sixties. Same person, same genes, same age, same diet, same sleep. One variable.
The controlled version of that observation: a study of 298 women aged 30 to 78, sorted by a dermatologist into 157 "sun-seeking" and 141 "sun-phobic," found markedly more wrinkling, sagging, pigmentation irregularity and vascular damage in the sun-seekers across nearly every measured sign.
UV damage is cumulative and largely silent for decades. It's typically in your forties that the accumulated total crosses the threshold where it becomes visible all at once — which is why the change can feel abrupt when the cause wasn't.
It's also the last decade where prevention outperforms correction by a wide margin. Damage you haven't accumulated is free to avoid. Damage you have is slow, partial and expensive to undo.
The three things with real evidence
Daily sunscreen
This is the one people assume is marketing and isn't. In Nambour, Australia, 903 adults under 55 were randomised into four groups in a factorial design: daily broad-spectrum sunscreen or discretionary use, crossed with beta-carotene supplements or placebo. Skin aging was graded from skin microtopography by assessors blinded to allocation, over four and a half years.
The daily-sunscreen group showed no detectable increase in skin aging over the study period — and 24% less aging than the discretionary group. That is a randomised controlled trial, in humans, with a hard endpoint, published in Annals of Internal Medicine. There is nothing else in skincare with a pedigree like it.
The beta-carotene arm, incidentally, did nothing. Which is its own useful finding.
In practice: every morning, on face, neck and the backs of your hands, regardless of weather or season. The hands are where people notice their age first and protect least.A retinoid
Retinoids are the only topical category that meaningfully rebuilds rather than merely protects — they increase collagen production and normalise cell turnover, and they've been studied for forty years.
Tretinoin, the prescription form, has a systematic review of randomised controlled trials behind it: seven RCTs, concentrations from 0.025% to 5%, durations from three to 24 months, all reporting improvement in photoaging, all reporting good tolerability. Four of the seven studied women exclusively.
Over-the-counter retinol is weaker per application but genuinely supported. A randomised, double-blind, vehicle-controlled trial applied 0.4% retinol or vehicle to opposite arms three times weekly for 24 weeks and found significantly greater improvement in fine wrinkling with retinol, alongside increased collagen production on biopsy. Worth knowing the limits of that trial: the participants averaged 87 years old and it was done on arms, not faces.
In practice: start over-the-counter, two or three nights a week, a pea-sized amount for the whole face, on dry skin. Expect irritation for a few weeks. Build up slowly rather than quitting.Not smoking
The most memorable demonstration is a pair of identical twins reported in Archives of Dermatology: same genes, two decades of shared upbringing, later the same job at the same latitude with well-matched sun exposure — but one had a roughly 52-pack-year smoking history and visibly more severe skin aging than her non-smoking twin.
That's a single case report rather than a trial, and it should be read as an illustration rather than proof. But it controls for the confounder that wrecks most observational skin research, and the mechanism — impaired dermal blood flow, degraded collagen and elastin — is well described.
That's the list. Three items, two of which cost under $30 a month. If you did only these and nothing else, you would be doing most of what the evidence supports, and more than the overwhelming majority of people with a bathroom shelf full of serums.
Build the routine the evidence supports
Five questions, then an ordered routine with the reason for every step — and, more usefully, an explicit list of what to skip for your situation. Ranked by evidence, not by margin.
The application problem nobody mentions
Here is the detail that undoes most sunscreen use.
SPF is not a property of the bottle. It's measured in a lab at a standardised application thickness of 2.0 mg per square centimetre of skin. That is a great deal more product than it sounds — for a face and neck it's roughly a third of a teaspoon, applied as an opaque layer, which almost nobody does because it looks like a layer.
Research examining the relationship between application thickness and actual SPF found what you'd expect: apply less, get less. The practical recommendation that came out of it was blunt — apply sunscreen twice. Once, then again a few minutes later. It's the cheapest possible upgrade to the highest-evidence step in your routine, and it costs nothing but thirty seconds.
An SPF 50 applied at a quarter of the tested thickness is not giving you SPF 50. Buying SPF 100 to compensate is the wrong fix — using enough of the SPF 30 you already own is the right one.
The maybes
These aren't dismissals. They're things with real but smaller or shakier evidence than the three above — worth buying if you enjoy them and can afford them, not worth buying instead of the basics.
Vitamin C, topical
Plausible mechanism, decent support for tone and pigmentation, and a genuine role as antioxidant backup to sunscreen. It's also notoriously unstable, which is why formulation and packaging matter more than concentration, and why there's so much room to overpay. A supporting act.
Niacinamide
Well tolerated, helps barrier function, oil and redness. Worth knowing where its strongest evidence actually sits: oral nicotinamide has systematic review and meta-analysis support for preventing non-melanoma skin cancers and actinic keratoses. That is a more impressive finding than anything in its anti-aging file — and it's a different intervention from the serum, which the marketing tends not to clarify.
Oral collagen
The most-searched supplement in this category, and the evidence is genuinely mixed rather than simply absent.
A 2021 meta-analysis pooled 19 studies covering 1,125 participants aged 20 to 70 — 95% women — and found favourable effects on skin hydration, elasticity and wrinkles. It promptly attracted a published comment specifically challenging the heterogeneity of the trials it had combined.
A 2026 systematic review took 25 randomised trials, ranging from 13 to 236 participants, and found improvements over placebo in hydration (10 of 15 trials), elasticity (10 of 13) and wrinkle measures (9 of 10) — while stating plainly that the clinical evidence "remains inconsistent," and choosing to synthesise the results descriptively rather than pool them. That choice is a signal in itself.
So: probably something, mostly measured by instruments rather than mirrors, in small and frequently industry-adjacent trials. If you want to take it, the case is defensible. Just don't take it instead of sunscreen.
What the evidence doesn't support
- Antioxidant supplements for skin aging. The Nambour trial randomised people to beta-carotene alongside its sunscreen arms. The sunscreen worked. The beta-carotene did nothing. That's what a clean negative result looks like
- "Anti-aging" cleansers. Seconds of contact time. Physically incapable of delivering what the label claims
- Price as a proxy for efficacy. There is no evidence relationship between what a moisturiser costs and what it does. The active ingredients that work are cheap and off-patent
- Collagen in a face cream. The molecule is far too large to cross the stratum corneum. Whatever the cream does, it isn't delivering collagen into your dermis
- Stacking more actives. Past a certain point additional products buy irritation, and irritated skin looks worse and heals slower. Most people's routines would improve by subtraction
One thing that changes late in the decade
Most women reach the start of the menopause transition in their mid-to-late forties, and estrogen has real effects on skin — collagen density, hydration, wound healing. If your skin changes noticeably faster around then, that's a known physiological shift rather than a failure of your routine, and it's a subject large enough to deserve its own article rather than a paragraph here.
What doesn't change is the list. Sunscreen and a retinoid remain the two things with the best evidence before, during and after that transition. The rest of the shelf stays optional.
And the unglamorous background truth: skin is an organ with a blood supply, and it reflects what the rest of you is doing. Sleep, glucose control, not smoking and cardiovascular fitness show up on a face over a decade far more reliably than anything in a bottle. Our pieces on sleep and on what your lipid panel actually means are, in a slightly indirect way, also skincare articles.
Two products, used properly, beat a shelf full of almost anything.
The skincare industry is built on the idea that this is complicated. It mostly isn't. There is one intervention with a randomised controlled trial and a hard endpoint behind it, and it costs about twelve dollars. There is one topical category that rebuilds rather than protects, and the over-the-counter version works. There's a third item, which is simply not smoking.
What's left over is preference. Vitamin C is reasonable. Niacinamide is pleasant and its best evidence is for something else entirely. Collagen probably does a little, measured by machines. None of it is worth buying at the expense of the two things that are.
The reason your forties matter is arithmetic. UV damage accumulates silently for decades and then becomes visible more or less at once, and prevention is dramatically cheaper than correction. The version of your face that exists at sixty is still substantially negotiable right now. At seventy it will be much less so.
References
Every figure above traces to one of these. Links go to the published record.
| Sunscreen and skin aging | Hughes et al., randomised trial, 903 adults, Nambour Australia, 4.5 years, Ann Intern Med 2013. PMID 23732711 |
| The truck driver | Unilateral dermatoheliosis, N Engl J Med 2012. PMID 22512500 |
| Sun exposure and visible aging | Flament et al., 298 women aged 30–78, sun-seeking vs sun-phobic, Clin Cosmet Investig Dermatol 2013. PMID 24101874 |
| Sunscreen application thickness | SPF measured at 2.0 mg/cm²; double application recommended, Clin Exp Dermatol 2012. PMID 23050556 |
| Tretinoin for photoaging | Systematic review of 7 RCTs, 0.025–5%, 3–24 months, Int J Womens Dermatol 2022. PMID 35620028 |
| Retinol for aged skin | Randomised, double-blind, vehicle-controlled, 0.4% retinol, 24 weeks, Arch Dermatol 2007. PMID 17515510 |
| Smoking and skin aging | Identical twin case report, Arch Dermatol 2007. PMID 18087005 |
| Oral collagen | Meta-analysis, 19 studies / 1,125 participants, Int J Dermatol 2021 PMID 33742704 · heterogeneity comment PMID 34196407 · systematic review of 25 RCTs, Eur J Clin Nutr 2026 PMID 42342959 |
| Nicotinamide | Skin cancer and actinic keratosis chemoprophylaxis, systematic review and meta-analysis, J Cutan Med Surg 2022. PMID 35134311 · photoprotection review, Exp Dermatol 2019 PMID 30698874 |
| Topical vitamin C | Systematic scoping review, Dermatol Surg 2026. PMID 41052296 |
- This article is educational and is not medical advice. It was written by a layperson who reads the primary literature carefully, not by a dermatologist, and it has not been clinically reviewed
- Tretinoin is prescription-only in the United States. Nothing here is a route around that — obtaining it is a conversation with a licensed clinician who can assess your skin
- Pregnancy, breastfeeding, rosacea, eczema and photosensitising medications all change what is appropriate. Check with your own doctor
- A mole or lesion that is new, changing, bleeding or asymmetric is a reason to see a doctor promptly. No skincare routine addresses that
