The EU tightened retinol concentration limits for face products in 2023. Many brands reformulated quietly, few explained why, and a wave of bakuchiol-labelled products appeared in the gap. Whether bakuchiol is a meaningful alternative depends on what retinol was actually doing — and the answer is more specific than most product descriptions suggest.
Retinol — vitamin A alcohol — is the most extensively studied cosmetic anti-ageing ingredient with the strongest clinical evidence base in the cosmetics category. Its mechanism is well-understood: retinol is converted in the skin to retinaldehyde and then to all-trans retinoic acid (tretinoin), the biologically active form that binds to nuclear retinoic acid receptors (RAR-α, RAR-β, RAR-γ) and modulates gene expression in keratinocytes and fibroblasts. The downstream effects include increased collagen synthesis, accelerated keratinocyte turnover, reduced melanin production, and improved lamellar lipid organisation in the stratum corneum. These are real, mechanism-driven effects with robust clinical trial support at concentrations as low as 0.1% (Kafi et al., PMID 16651385).
The EU’s modification of retinol’s Annex III status in 2023 — placing face product concentration limits at 0.3% for general use products and 0.05% for products intended for areas around the eyes and on the lips — was driven by the SCCS’s concern about cumulative dietary and topical retinol exposure, particularly regarding teratogenic risk during pregnancy and systemic vitamin A accumulation. The restriction does not remove retinol from EU cosmetics. It caps the concentration at which it can be used in face products — a cap that most well-formulated products were already operating below, given that concentrations above 0.3% have historically produced high rates of irritation without commensurate clinical benefit over lower concentrations.
How retinol works — the retinoic acid pathway
Understanding what the EU restriction affects requires understanding the conversion pathway. Retinol (vitamin A alcohol) itself is not biologically active in the sense that matters for anti-ageing. It must be oxidised in two steps — first to retinaldehyde by retinol dehydrogenases, then to retinoic acid by retinaldehyde dehydrogenases — before it can bind to nuclear receptors and produce gene expression changes. Each conversion step involves enzymatic processing that varies by individual skin type, skin condition, and the presence of co-factors.
This conversion pathway is the key to understanding both the effectiveness question and the alternatives question. Ingredients that enter the pathway at a later stage — retinaldehyde, for example — require fewer conversion steps and are more directly available for retinoic acid production. Ingredients that claim to mimic the pathway from outside it — bakuchiol — operate by different mechanisms entirely and cannot be evaluated on the same scale.
| Ingredient | Pathway position | EU cosmetic status | Efficacy evidence |
|---|---|---|---|
| Retinyl palmitate | 2 steps before retinoic acid | Permitted, no limit (face) | Moderate — slow conversion |
| Retinol | 1 step before retinoic acid | Permitted, 0.3% max (face) | Strong — extensive RCT support |
| Retinaldehyde | Direct precursor | Permitted, no specific limit | Strong — faster conversion than retinol |
| Retinyl retinoate | Ester — converts to retinol + retinoic acid | Permitted — check current INCI list | Emerging — Korean clinical data |
| Tretinoin (retinoic acid) | Active form | Prescription only — not permitted in cosmetics | Gold standard — not cosmetically available |
| Bakuchiol | Outside retinoid pathway | Permitted, unrestricted | Limited — 2 RCTs, small sample sizes |
Sources: SCCS/1576/16; EU Cosmetics Regulation Annex III (2023 consolidated); Draelos ZD et al. J Cosmet Dermatol 2019.
EU Annex III — what the 2023 restriction actually says
The 2023 restriction, implemented via Commission Regulation (EU) 2023/1490, places retinol in Annex III of EU Cosmetics Regulation EC 1223/2009 — the list of restricted substances permitted under specific conditions. The key limits: face products for general use, 0.3% retinol maximum; products for the eye area and lip products, 0.05% maximum. Body lotion and leave-on body products face a separate 0.05% limit with a mandatory warning statement.
The warning required on body products containing 0.05% or more retinol — “Contains retinol. Not recommended for pregnant women and women of childbearing age. Do not use on children under 3 years” — is not required on face products below 0.3%, though brands may include it voluntarily. The restriction does not prohibit retinol in face products. It limits the maximum permitted concentration to a level that most dermatologists and formulators consider sufficient for cosmetic anti-ageing use.
The EU’s retinol restriction did not limit efficacy. Most effective formulations were already below 0.3%. What changed is that higher-concentration products — which were causing more irritation than benefit — can no longer be sold.
Bakuchiol — functional alternative or marketing substitute?
Bakuchiol is a meroterpene phenol derived from the seeds of Psoralea corylifolia, a plant used in Ayurvedic and traditional Chinese medicine. It does not share the retinoid molecular structure, does not enter the retinoic acid conversion pathway, and does not bind to nuclear retinoic acid receptors. The claim that it is a “natural retinol alternative” refers to a different mechanism: bakuchiol activates some of the same downstream transcription pathways as retinoic acid through retinoid-like gene expression modulation, but via a different receptor interaction and with a different potency profile.
The clinical evidence for bakuchiol is thin by pharmaceutical standards but exists. Dhaliwal et al. (PMID 30121949) conducted a 12-week split-face RCT comparing 0.5% bakuchiol twice daily with 0.5% retinol once daily, finding comparable reductions in wrinkle depth and hyperpigmentation, with bakuchiol producing significantly less irritation. This is the primary study referenced in bakuchiol marketing. It is a real study with real findings — but its sample size (44 participants) and single-site design mean it cannot bear the weight of “proven retinol equivalent” claims.
Evidence range: 0.5–1%
pH stability: broad
Tolerance advantage: Unlike retinol, bakuchiol does not produce retinoid dermatitis — the initial irritation, dryness, and peeling that accompanies retinol introduction at effective concentrations. This makes it genuinely useful for retinol-intolerant skin, barrier-compromised skin, and summer use when retinol’s photosensitising properties create additional management requirements.
Honest limitation: The two existing RCTs (Dhaliwal 2019; Chaudhuri & Bojanowski 2014) are too small and too short to confirm equivalence with retinol at matched concentrations across the full range of anti-ageing effects retinol has demonstrated in long-term trials. Bakuchiol is a useful ingredient with real anti-ageing mechanism — it is not a demonstrated retinol equivalent.
Korean retinyl retinoate — the EU legal position
Retinyl retinoate (INCI: Retinyl Retinoate) is an ester of retinol and retinoic acid developed by the Korean cosmetic ingredient company Caregen. On hydrolysis in the skin, it releases both retinol (entering the conversion pathway one step before retinoic acid) and a small amount of retinoic acid directly. The theoretical advantage is a two-pronged delivery: immediate low-level retinoic acid activity combined with sustained retinol-to-retinoic acid conversion.
Its EU legal status is unambiguous: retinyl retinoate is not on Annex III (restricted substances) and is not on Annex II (prohibited substances). It is a permitted cosmetic ingredient under EC 1223/2009, subject to the general safety requirement of Article 3. It is marketed in Korean cosmetics at concentrations typically between 0.05 and 0.1% and is available in products shipped to the EU market with no regulatory barrier. The clinical evidence base is smaller than for retinol — primarily Korean clinical studies with limited external replication — but the mechanistic rationale is sound and the tolerance profile appears significantly better than retinol at equivalent anti-ageing concentrations.
Retinoids in summer — the photosensitivity question
Retinol is photolabile — it degrades on UV exposure, which means morning application reduces its efficacy. It is typically recommended for evening use only, which removes the photosensitivity question at the application step. The more relevant summer concern is that retinol accelerates keratinocyte turnover, producing thinner, more recently renewed epidermal cells with less accumulated sun protection — not immunologically sensitised skin, but structurally less UV-hardened skin. This is the mechanism behind the common observation that skin using retinol burns more easily in summer.
Bakuchiol does not share this photosensitivity issue — it is photostable and can be used morning and evening without UV-related concerns. This is a genuine practical advantage for summer use in Berlin, where high UV days in June and July make morning retinol avoidance a meaningful inconvenience. Retinyl retinoate is also more photostable than retinol, which is one of its clinically claimed formulation advantages.
Selected for EU-compliant retinoid concentration, clinical evidence base, and summer tolerance profile.
Frequently Asked Questions
Did the EU ban retinol in 2023?
No. The EU restricted retinol concentration in face products to a maximum of 0.3% (down from a previously unspecified limit, though most products were already below this). Retinol remains fully permitted in EU cosmetics. The restriction was driven by concerns about cumulative vitamin A exposure, particularly regarding teratogenic risk in pregnancy. Products formulated at 0.1–0.3% retinol — the range with the best evidence-to-irritation ratio — are not affected in practical terms.
Is bakuchiol as effective as retinol?
The honest answer is: probably not at matched concentrations for the full range of retinol effects, but the evidence is insufficient to state this definitively. The primary RCT (Dhaliwal et al., 2019) found comparable wrinkle and pigmentation results at 0.5% bakuchiol versus 0.5% retinol over 12 weeks, with better tolerance for bakuchiol. However, the study is small (44 participants) and short. Retinol has decades of trials behind it; bakuchiol has two. Bakuchiol is a genuinely useful ingredient with real anti-ageing mechanism — the claim that it is equivalent to retinol is not yet supported by the evidence.
Can I use retinol in summer?
Yes, but with two adjustments. First, use retinol in the evening only — it is photolabile and degrades on UV exposure, so morning application wastes the ingredient. Second, be consistent with broad-spectrum SPF 50+ during retinol use, as the accelerated keratinocyte turnover retinol produces creates a thinner, less UV-adapted epidermal surface. These are management steps, not reasons to pause retinol entirely. Pausing retinol in summer means losing several months of collagen-stimulating activity — the UV risk is manageable with adequate sunscreen.
What is retinyl retinoate and is it available in the EU?
Retinyl retinoate is a Korean-developed ester of retinol and retinoic acid that delivers both components on hydrolysis in the skin. It is not restricted under EU Cosmetics Regulation Annex III and is legally permitted in EU cosmetics. It is available in Korean skincare products sold through EU retailers. Its clinical evidence base is smaller than retinol’s, but its tolerance profile and photostability are better, making it an interesting option for summer anti-ageing use. It is not a prescription retinoid — it is a cosmetic ingredient with modest direct retinoic acid delivery.
What concentration of retinol is actually effective?
Clinical evidence supports efficacy at concentrations as low as 0.025% (Kafi et al., 2007), with 0.1% showing clear anti-ageing effects and 0.3% showing stronger effects with more irritation. Concentrations above 0.3% show diminishing returns relative to the irritation increase and are not superior to prescription tretinoin, which works via a different (direct) mechanism. For most people, 0.1–0.3% retinol with proper barrier support produces the best efficacy-to-tolerance ratio — and this falls within the EU’s current permitted range.
References
- Kafi R, et al. Improvement of naturally aged skin with vitamin A (retinol). Arch Dermatol. 2007;143(5):606–12. PMID 16651385.
- Dhaliwal S, et al. Prospective, randomized, double-blind assessment of topical bakuchiol and retinol. Br J Dermatol. 2019;180(2):289–96. PMID 30121949.
- Commission Regulation (EU) 2023/1490 — retinol restriction in Annex III. Official Journal of the European Union.
- SCCS Opinion SCCS/1576/16 on retinol safety. European Commission SCCS.
- INCI Decoder — retinoid ingredient database. incidecoder.com
- Chaudhuri RK, Bojanowski K. Bakuchiol: a retinol-like functional compound. Int J Cosmet Sci. 2014;36(3):221–30.
** Affiliate disclosure: Links above are affiliate links. JJJOZ earns a small commission on qualifying purchases at no cost to you. Product selection is based on formulation criteria only — not commercial relationships with brands.
Leave a Reply