Retinol concentrations: the EU 0.3% cap, what a “1%” means, and how it differs from tretinoin

Since 2024, the European Union has capped how much retinol a cosmetic may contain: a maximum of 0.3% retinol equivalent in facial products and 0.05% in body lotion. It is a rule in force, with specific dates, and it explains why the percentages on European labels stopped going up.

It also explains something more useful for the buyer: the number on the label does not measure what most people think it measures. A product from one family cannot be compared with one from another by its percentage, and there are labels whose number refers not to the active ingredient but to the raw material that contains it.

This page is part of the retinol sub-cluster, which starts at what retinol does for the skin.

What the rule actually says

Regulation (EU) 2024/996, of 3 April 2024, amended the European cosmetics regulation to restrict the use of vitamin A. In Annex III, entry 376, it added retinol, retinyl acetate and retinyl palmitate with these limits:

Type of product	Maximum concentration
Body lotion	0.05% retinol equivalent (RE)
Other leave-on and rinse-off products — the face included	0.3% RE

And with a mandatory warning on the label: «Contains Vitamin A. Consider your daily intake before use».

The two dates worth knowing:

• 1 November 2025. From that day, cosmetics containing these substances and not complying with the conditions cannot be placed on the Union market.
• 1 May 2027. From that day, those already in the chain cannot be made available on the market either.

That eighteen-month gap is why products with different labels may sit side by side on the shelf for a while.

Why the limit was set, and why it is not what it looks like

It is worth understanding the reason, because almost every headline got it wrong.

The problem is not that retinol harms the skin at those concentrations. The reasoning of the scientific committee that advised the Commission runs differently: the vitamin A a cosmetic contributes is low, but it adds to what already arrives through diet and food supplements. In part of the population — around 5% — total vitamin A intake is already high, and there that extra amount can matter.

It is exactly what the label warning says: it is about daily intake, not about the face. And it is why the cap for body lotion is six times lower than for a facial product: it is spread over far more skin.

🚫 What the rule did not do: ban retinol. It has been read that way many times. Retinol is still permitted, authorised and on sale; what there is, is a ceiling.

What «0.3% RE» actually means

This is the part that lets you read a label with judgement.

RE stands for «retinol equivalent», a unit that converts the different forms of vitamin A to a common standard. A product may contain pure retinol, retinyl acetate, retinyl palmitate or a mixture, and each contributes a different amount of vitamin A per unit of weight. The rule does not limit each one separately: it limits the total, converted into retinol equivalent.

The practical consequence: the percentage on the pack and the retinol equivalent may not be the same number.

And why you see labels with higher figures

There are three different explanations, and it is worth knowing which one applies before drawing conclusions:

1. The number refers to the raw material, not to the active ingredient. This is the most frequent and the most confusing. Retinol ingredients are sold to manufacturers dissolved or encapsulated: a raw material may contain, say, 10% retinol and the rest carrier. A product formulated with 10% of that raw material can be labelled in several ways, and only one of them refers to the actual active ingredient. When the big number on the pack does not match what you would expect, this is usually the reason.
2. It is a different market. Cosmetics rules are not global. A product designed for another market may contain concentrations that are not allowed in the European Union, and its European version may be different even though the brand name is the same.
3. It is the transition window. Between November 2025 and May 2027 earlier product may remain in circulation, and it is perfectly legal for it to do so. The next section explains why.

🔑 What to do with this, in one sentence: when the number on a pack surprises you, the useful question is not «is it better?» but «what does that number refer to?» — and the manufacturer is obliged to be able to answer it, because the claims regulation requires every claim to be truthful and demonstrable.

The transition window: why they are still on sale in 2026

This is the part that causes most confusion and that almost nobody explains, because it depends on two legal terms that sound alike and do not mean the same thing. The Regulation itself defines them in Article 2:

Term	What it is, under the rule
Placing on the market	«The first making available of a cosmetic product on the Community market». It is the act of the manufacturer or the importer: bringing the product into the Union.
Making available on the market	«Any supply of a cosmetic product for distribution, consumption or use on the Community market in the course of a commercial activity, whether in return for payment or free of charge». It includes the sale to the final customer.

And the note to entry 376 sets a date for each:

• From 1 November 2025, a non-compliant product cannot be placed on the Union market. That is: no new units can come in.
• From 1 May 2027, a non-compliant product cannot be made available on the market either. That is: it cannot be sold, not even what was already inside.

🔑 The practical consequence, put plainly: product that was legally placed on the European market before 1 November 2025 can still be sold until 30 April 2027. It is not a loophole or a concession: it is the window the rule itself grants to run down stock, and it is why in 2026 you still see packs with figures that could not be formulated today.

⚠️ And the flip side, which is the useful part for the buyer. The right question when faced with a pack showing a high figure is not «can this be sold?» but «when did this come onto the European market?». Because a unit that came in after November 2025 without complying with the restrictions is already outside the rules, transition window or no transition window.

🔑 How to check without being a lawyer: look at whether the pack carries the mandatory warning — «Contains Vitamin A. Consider your daily intake before use». Its presence indicates that the manufacturer has adapted the product to the new Annex III entry. And buying through a channel that answers for traceability — a pharmacy, a clinic or an official distributor — is the practical way to avoid having to find out.

⏳ From 1 May 2027 this window closes, and anything non-compliant can no longer be sold anywhere in the Union.

Why more percentage is not more result

This is the point advertising does not explain, and the one that changes the buying decision.

Retinol is not the molecule that acts. To do anything it has to be converted inside the skin, in two steps: first into retinaldehyde and then into retinoic acid, which is the only form that binds to the cell receptor. And that conversion is inefficient: in a study of skin metabolism, treating the tissue with retinol or retinaldehyde produced in the order of 30 to 90 pmol of retinoic acid, whereas applying retinoic acid directly left between 420 and 2,050 pmol.

What that means for the buyer: the bottleneck is not how much retinol you put on, but how much gets converted. Doubling the percentage does not double the available retinoic acid; what it does double, fairly reliably, is the irritation. It is the same reason why the sensible strategy is to start low and build up, not to start with the strongest thing available.

What the brands are doing: where the market is heading

The cap has not held the category back; it has redirected it. And the direction becomes clear once you read carefully who the rule names, because entry 376 does not say «retinoids»: it names three specific substances.

First point: switching ester achieves nothing

Entry 376 covers retinol, retinyl acetate and retinyl palmitate, all three together. And the cap is expressed not as a percentage of the substance but in retinol equivalent (RE), which is the conversion to retinol activity.

🔑 Translated: moving to an ester does not dodge the limit. A retinyl palmitate is weaker than retinol, so more of it would be needed for the same effect — and that is precisely why the rule counts the equivalent and not the grams. It is a limit designed so that it cannot be got around that way. The mandatory labelling warning, moreover, is required for all three.

The real move: retinaldehyde

Here is the fact that explains what has been appearing in launches over the past two years. Retinaldehyde — also called retinal — does not appear in entry 376. Nor does it appear in any other entry of Annexes II or III of the cosmetics Regulation, checked against the consolidated text.

It is a retinoid in its own right: it is one step closer to retinoic acid than retinol in the skin’s conversion chain, so it needs one transformation fewer. And it has its own, recent evidence:

• A 24-week study in 56 women aged 30 to 58, with a split-face design, compared stabilised retinaldehyde at 0.05% against 0.1%. High-frequency ultrasound measured an increase in dermal density across all facial regions, with better elasticity and viscoelasticity by cutometry and a reduction in wrinkles by 3D imaging, more marked with the 0.1%. The authors note that it was well tolerated, with minimal adverse effects.
• Another split-face study in 66 people applied a cream containing retinaldehyde, niacinamide and haritaki extract after a procedure — chemical peel, hyaluronic acid or fractional laser, 22 per group — for three months: by the first month there was already a significant reduction in wrinkles and better firmness. It is the scenario that interests us most here, because it is that of a patient walking out of the consulting room.

⚠️ And now the caveat, which is not optional. That retinaldehyde is not in Annex III does not mean it is free of control: the general obligation that the product be safe still applies, with its signed safety assessment and its responsible person. And the European scientific committee may turn to it in the future, just as it turned to retinol. What is true today is that it has no number written in an annex; not that it is a free pass.

How they try to make up for the cap: combined formulas

This is the dominant strategy right now, and the commercial argument that goes with it — «we achieve with 0.3% what used to be done with 1%» — deserves reading slowly, because part of it is demonstrated and part of it is not.

What the brands are doing, ordered from most to least supported:

1. Combining retinoids that reach the same place by different routes. Retinol needs two conversions in the skin before it acts; an ester such as hydroxypinacolone retinoate (HPR) binds to the receptor without needing that chain. A 2026 study looked at exactly that mixture — retinol plus HPR — and found that they synergistically activated the TGF-β/Smad pathway and increased the expression of extracellular matrix genes more than retinol alone, with clinical improvement at eight weeks and good tolerance. It is the most direct support there is for the idea of adding rather than raising.

2. Adding actives that work on the same targets by another route. Peptides, niacinamide, plant antioxidants. In that same study, adding silybin further increased collagen and elastin synthesis. The logic is not «boosting retinol» but attacking the same problem from several fronts, which is a different and more defensible thing.

3. Boosters that improve performance and tolerance at the same time. A 2026 study with a bioretinoid obtained from microalgae reports that it synergistically improved the effects of retinol, increasing hydration and reducing irritation. ⚠️ With the caveat that much of that work is in vitro.

4. Improving delivery. Encapsulation, liposomes, nanoemulsions — the next section.

🔑 And now what has to be said precisely, because this is where marketing runs ahead of the data. What these studies demonstrate is synergy compared with retinol alone and, above all, better tolerance. What does not exist is a trial comparing «0.3% with boosters» against «1% retinol» and concluding that they are equivalent. Equivalence in potency with the 1% is a marketing claim, not a published result.

⚠️ And there is an irony worth bearing in mind: as the previous section explains, those «1%» products rarely meant 1% of active ingredient — the number usually referred to the raw material. So a good part of the debate consists of matching a figure that was never what people thought it was.

🔑 So where is the real gain. In tolerance. The main reason a retinol does not work is not that it is weak: it is that it gets abandoned in the third week because of irritation. A formula that can be used four nights a week for eight months delivers more than a potent one that lasts three weeks in the bathroom cabinet. That is the ground on which combined formulas do have data in their favour, and it is also the ground that decides the result.

Encapsulation: it does not change the cap, it changes tolerance

Liposomes, microsponges, gradual-release systems. Reviews of retinoid irritation list encapsulation among the strategies described for improving tolerance, alongside barrier repair and hydration.

🔑 But do not be misled: encapsulating does not raise the legal ceiling. The rule caps retinol equivalent, not the wrapper. An encapsulated retinol at 0.3% is still 0.3%. What changes is how it is released, and therefore how much irritation it causes along the way — which is not a small thing, because abandonment through irritation is the main reason a retinol does not work.

What is not a retinoid

Bakuchiol does not appear in the annexes either and has been studied as an alternative: a trial in 44 patients compared it with 0.5% retinol over 12 weeks, with similar efficacy and less flaking. It is not a retinoid and does not act in the same way, and its literature is far smaller — but it is the route for anyone who cannot use vitamin A, including in pregnancy, where no retinoid is indicated.

And on the medicine side

Outside cosmetics, research is heading in another direction and it is worth naming it to place the whole picture: synthetic retinoids with greater receptor selectivity, inhibitors of retinoic acid metabolism that prolong the activity of what the skin itself produces, and targeted delivery systems — nanocarriers, stimulus-responsive gels — aimed at improving stability and tolerance. This is prescription territory, not shelf territory, and that is why it sits on the next rung of the ladder.

The full ladder: from the cream to the prescription

Ordered by their distance from the target:

Molecule	Conversions it needs	Status
Retinyl esters (palmitate, acetate)	3	Cosmetic
Retinol	2	Cosmetic
Retinaldehyde	1	Cosmetic
Retinoic acid (tretinoin)	0 — acts directly	Prescription medicine

The first three rungs share a legal concentration ceiling. The fourth is in another category: in Spain, tretinoin is marketed as a medicine at 0.25 and 0.5 mg per gram — that is, 0.025% and 0.05% — requires a prescription, and its summary of product characteristics authorised by the Spanish Medicines Agency recognises two indications: mild to moderate acne vulgaris and, as an adjuvant, the reduction of fine wrinkles, mottled hyperpigmentation and facial roughness.

⚠️ Look at those numbers again. The medicine works at 0.05% and the cosmetic can reach 0.3%: six times the percentage, and yet the medicine is clearly more potent. It is the proof that comparing percentages between different molecules tells you nothing.

When they have been compared directly on human skin, retinol and retinoic acid both produced an increase in epidermal thickness and in the expression of type 1 and type 3 collagen genes, with retinol inducing similar changes but of smaller magnitude. They do the same thing, at a different speed and with a different tolerance profile — and with one difference that is not minor: one is bought and the other is prescribed.

How to choose a concentration, in practice

With all of the above, the criterion becomes fairly simple:

• If you have never used retinoids, start with the lowest concentration in the range in front of you. The skin’s adaptation is the limiting factor, not the strength of the product.
• If you have tolerated a concentration well for months and no longer notice any change, it makes sense to go up a step — of concentration or of molecule. In that order.
• If you are after real potency, the next step is not a cosmetic with a higher percentage: it is a medical consultation. Go up a rung, not up a number.
• And if your skin is constantly irritated, the problem is not solved by changing brand. Go down, repair the barrier and build back up slowly.

The dermocosmetic products we dispense are in the retinol and retinoid shop, and several ranges are graded precisely so that you can step up without jumps. Which step is right in each case is decided in consultation, looking at the skin and at what is already being used.

Frequently asked questions

Has retinol been banned in Europe?

No. What there is, is a concentration limit: 0.3% retinol equivalent in facial products and 0.05% in body lotion, with a mandatory warning on the label. Retinol is still permitted and on sale. The «ban» headline went round a great deal in 2024 and it is wrong.

Why is the limit lower for body lotion?

Because the reason for the rule is total vitamin A intake, not damage to the skin. A body lotion is spread over a far larger area than a face cream, so at the same concentration it would contribute considerably more. Hence a cap six times lower.

I bought a retinol before all this — do I throw it away?

The rule is addressed to those who manufacture and sell, not to someone who already has it at home, and the limit responds to the sum of vitamin A from all sources, not to an acute risk. If you have doubts about your own case — for instance, if you take supplements containing vitamin A — that is exactly the conversation worth having in consultation.

Does a European 0.3% work worse than a higher percentage from another market?

Not necessarily, and the comparison is rarely as direct as it looks. The molecule, the stability of the formula, the vehicle and the packaging all matter, and frequently the number on the pack refers to the raw material and not to the active ingredient. Besides, the clinical evidence for retinol was obtained with regimens lasting months: consistency counts for more than the decimal point.

How long can products with more than 0.3% be sold?

Until 30 April 2027, provided they were legally placed on the European market before 1 November 2025. The Regulation distinguishes between «placing on the market» — the first making available, which is a matter for the manufacturer or the importer — and «making available on the market», which is any subsequent supply, including the sale to the customer. Each has its own date. From 1 May 2027 nothing non-compliant can be sold.

Is retinaldehyde better than retinol?

It is one rung closer to retinoic acid, so it needs one conversion fewer. That makes it, in principle, more potent at the same concentration — and by the same logic, somewhat more demanding on the skin. It is not «better»: it is a different step, and which step suits you depends on your own tolerance. It is worth knowing that retinaldehyde does not appear in the Annex III entry that caps retinol, which explains a good part of the launches of recent years.

And what about «encapsulated retinol»?

Encapsulation is a way of protecting the active ingredient and releasing it gradually, and reviews of retinoid irritation list it among the strategies for improving tolerance. It still counts towards the legal cap, because the rule limits retinol equivalent, not the wrapper.

How do I know whether a product complies with the rule?

Those that comply carry the mandatory warning: «Contains Vitamin A. Consider your daily intake before use». Its presence indicates that the manufacturer has adapted the product to Annex III. Buying through a channel that answers for traceability — a pharmacy, a clinic or an official distributor — is the practical route.

Sources

• Commission Regulation (EU) 2024/996 of 3 April 2024 amending Regulation (EC) No 1223/2009 as regards the use of Vitamin A, Alpha-Arbutin and Arbutin and certain substances with potential endocrine disrupting properties in cosmetic products. Text at the BOE, in Spanish. — Annex III, entry 376: retinol, retinyl acetate and retinyl palmitate, with maxima of 0.05% retinol equivalent in body lotions and 0.3% RE in other leave-on and rinse-off products. Mandatory warning: «Contains Vitamin A. Consider your daily intake before use». Footnote (*3) to the entry, verbatim: «From 1 November 2025 cosmetic products containing that substance and not complying with the conditions shall not be placed on the Union market. From 1 May 2027 cosmetic products containing that substance and not complying with the conditions shall not be made available on the Union market.» 🔑 All three — retinol, retinyl acetate and retinyl palmitate — share that same footnote, and the cap is expressed in retinol equivalent, not as a percentage of the substance. Verified against the official text on 27-08-2026; footnote and full entry verified against EUR-Lex on 28-08-2026, and against the English text on 17-09-2026.
• Regulation (EC) No 1223/2009 — Article 2, definitions. Consolidated text at EUR-Lex. — Verbatim, points (g) and (h): «making available on the market» means any supply of a cosmetic product for distribution, consumption or use on the Community market in the course of a commercial activity, whether in return for payment or free of charge; «placing on the market» means the first making available of a cosmetic product on the Community market. These are the two definitions on which the transition window depends. Verified on 28-08-2026.
• Comparative Evaluation of Topical Stabilized Retinaldehyde 0.1% vs. 0.05% on Skin Biophysical and Biomechanical Parameters. Skin Research and Technology, 2026. PMID 41729119. — Split-face design in 56 women aged 30 to 58, 24 weeks, with high-frequency ultrasound, an Antera 3D camera and a cutometer. Results: increased dermal density across all facial regions, better elasticity and viscoelasticity, and a significant reduction in wrinkle parameters, more marked with the 0.1%. Well tolerated, with minimal adverse effects. ⚠️ Study with no independent placebo group: it compares two concentrations against each other.
• Enhancing Facial Rejuvenation Outcomes With a Novel Retinaldehyde-Based Cream: A Comparative Randomized Intra-Individual Study. Journal of Cosmetic Dermatology, 2025. PMID 41312593. — Single-centre, controlled, randomised split-face study over three months in 66 people who had undergone a procedure — chemical peel, hyaluronic acid or fractional laser, n = 22 per group — with assessment starting after re-epithelialisation. The cream (retinaldehyde, niacinamide and haritaki extract) significantly reduced wrinkles at one month and improved firmness. ⚠️ It is a formula with three actives: the result is not attributable to retinaldehyde alone.
• An Innovative Serum With Retinol, Hydroxypinacolone Retinoate, Peptides, and Silybin Improves Mild Photoaged Facial Skin in Middle-Aged Chinese Women. Journal of Cosmetic Dermatology, 2026. PMID 41450017. — Mechanism studied by transcriptomic profiling in a co-culture of keratinocytes and fibroblasts and by protein expression in UV-exposed ex vivo skin; clinical efficacy assessed at 8 weeks in middle-aged women with mild photoageing. Key result: retinol + HPR synergistically activated the TGF-β/Smad pathway and increased the expression of extracellular matrix genes compared with retinol alone; adding silybin increased collagen and elastin synthesis. Clinical improvement in wrinkles, elasticity, hydration, barrier function and pigmentation, with good tolerance. ⚠️ Much of the work is mechanistic; the clinical phase is short and in a specific population.
• Bioretinoids from microalgae: Boosting retinol performance and tolerability. International Journal of Cosmetic Science, 2026. PMID 40905519. — A microalgal bioretinoid compared with retinol and bakuchiol. In vitro it outperformed both in cell proliferation, melanin reduction and matrix modulation; in vivo, results comparable to retinol, and it synergistically boosted the effects of retinol, improving hydration and reducing irritation. ⚠️ The weight of the work is in the in vitro part.
• The Next Generation of Skin Care: Transforming Retinoid Therapeutics. Cells, 2025. PMID 41227299. — Review of the state of retinoid development: synthetic retinoids with greater receptor selectivity (RAR-γ agonists such as trifarotene), inhibitors of retinoic acid metabolism that sustain endogenous activity, and nanocarriers and stimulus-responsive gels to improve stability, bioavailability and tolerance. ⚠️ This is medicine territory, not cosmetics: it is cited to place the whole picture.
• Regulation (EC) No 1223/2009 of the European Parliament and of the Council of 30 November 2009 on cosmetic products. Text at the BOE, in Spanish. — The framework rule that the previous regulation amends. Annex III lists the substances that may only be used subject to restrictions.
• Commission Regulation (EU) No 655/2013 of 10 July 2013 laying down common criteria for the justification of claims used in relation to cosmetic products. Text at the BOE, in Spanish. — Criteria of truthfulness and evidential support: if a product is claimed to contain a specific ingredient, that ingredient must be present, and every claim must be supported by adequate and verifiable evidence.
• AEMPS — CIMA. Summaries of product characteristics for Retirides 0.25 mg/g and 0.5 mg/g cream (tretinoin). cima.aemps.es. — A medicine subject to medical prescription; 0.25 and 0.5 mg of tretinoin per gram. Authorised indications: mild to moderate acne vulgaris and, as an adjuvant, the reduction of fine wrinkles, mottled hyperpigmentation and roughness of the facial skin.
• Bailly J, Crettaz M, Schifflers MH, Marty JP. In vitro metabolism by human skin and fibroblasts of retinol, retinal and retinoic acid. Experimental Dermatology, 1998;7(1). PMID 9517919. — After treatment with retinol and retinaldehyde, 30-90 pmol of retinoic acid were detected in the epidermis and dermis, against 2,050 and 420 pmol when retinoic acid was applied directly. ⚠️ An in vitro study: it supports the order of magnitude of the conversion, not an equivalence of clinical dose.
• Kong R, Cui Y, Fisher GJ, et al. A comparative study of the effects of retinol and retinoic acid on histological, molecular, and clinical properties of human skin. Journal of Cosmetic Dermatology, 2016;15(1). PMID 26578346. — Both increased epidermal thickness and the expression of type 1 and type 3 collagen genes; retinol induced similar changes with a tendency towards smaller magnitude.
• Narsa AC, et al. Comprehensive Review of Strategies to Reduce Retinoid-Induced Irritation. Dermatology Research and Practice, 2024. PMID 39184919. — Lists encapsulation, nanoparticles and cyclodextrin complexes among the published strategies for reducing retinoid irritation.

Book your consultation

If you are comparing two retinols and do not know whether the difference in percentage means anything, the short answer is that it almost never means what it appears to. In consultation we look at the skin, review what you are already using and decide which step is the right one — which is sometimes one lower than the one you had in mind.

Clinical confirmation by the surgeon

Last reviewed: 17 September 2026. This content is for information only and does not replace an individual assessment by a healthcare professional.