Bakuchiol Serum vs Retinol: What the Evidence Says

Aimée's Journal•
A glass dropper of golden face serum in soft natural light

Bakuchiol serum is sold as the pregnancy-safe retinol alternative. What the 2019 trial proved, what it did not, and how to read a 1% label honestly.

Oral isotretinoin is an established human teratogen: Health Canada's pregnancy prevention programme requires written consent, two negative pregnancy tests before a first prescription, monthly testing during treatment and two simultaneous methods of contraception (Health Canada, 2016). Topical retinoids carry nothing close to that risk, yet obstetric guidance still advises against them as a precaution. That gap — a real warning attached to a drug class, applied cautiously to a cosmetic one — is what built the market for the bakuchiol serum.

Bakuchiol is a meroterpene phenol extracted from the seeds and leaves of Psoralea corylifolia, the plant known in Ayurvedic practice as babchi, used in topical cosmetics as a functional analogue of retinol: it produces retinol-like shifts in skin gene expression despite having no structural resemblance to a retinoid (Chaudhuri & Bojanowski, International Journal of Cosmetic Science, 2014). The human trials that carry the category tested it at 0.5%.

What bakuchiol does at the cell level

The 2014 study that opened the category ran retinol and bakuchiol side by side in a full-thickness reconstructed skin model and compared the gene expression profiles. The two overlapped substantially: both upregulated genes for collagen types I, III and IV, and both downregulated matrix metalloproteinase expression — the enzyme family that degrades dermal collagen under ultraviolet exposure and age.

The mechanism behind that overlap has not been fully mapped. Retinoic acid works by binding nuclear retinoic acid receptors directly, which is both why it is potent and why it inflames — efficacy and dermatitis travel down one pathway. Bakuchiol appears to reach a similar transcriptional endpoint by other means, the most plausible explanation for the tolerability difference measured later in human subjects. That is a hypothesis with clinical support, not a settled account.

A point of vocabulary follows. Bakuchiol is not a "natural retinol"; there is no retinol in it and no vitamin A. The phrase used in the primary literature — functional analogue — is far more precise than anything printed on a shelf.

Why retinol became a pregnancy problem

Retinol, retinaldehyde, retinyl palmitate and prescription tretinoin are all vitamin A derivatives, and vitamin A at high systemic dose is teratogenic. That is the entire basis of the concern, and for the oral drugs it is well founded. The European Medicines Agency's 2018 review confirmed that oral retinoids — isotretinoin, acitretin, alitretinoin — require formal pregnancy prevention programmes, while noting that systemic absorption after topical application is negligible and topical products are unlikely to cause fetal harm. The same review nonetheless kept topical retinoids contraindicated in pregnancy, explicitly as a precaution (EMA, 2018).

The epidemiology has since caught up, and it is reassuring rather than alarming. A Nordic cohort study published in the British Journal of Dermatology covered 3.87 million infants born in Denmark, Iceland, Norway and Sweden between 1996 and 2020. Major congenital malformations occurred in 3.3% of infants exposed to topical retinoids in pregnancy — 71 of 2,172 — against 3.0% of unexposed infants, an adjusted risk ratio of 1.1 (95% CI 0.87–1.38). The authors observed no increased risk, offered the finding as reassurance about inadvertent exposure, and held that precaution remains justified given the study's limits (Refsum et al., British Journal of Dermatology, 2026).

Counselling reflects that ambivalence. The MotherToBaby fact sheet on topical tretinoin, updated July 2024, states that because many people have used it during pregnancy without a birth defect the chance of harm is probably low — and then advises avoiding it throughout pregnancy regardless (MotherToBaby, 2024). Anyone handed that guidance at a first-trimester appointment recognises its shape: not a prohibition, a shrug in the direction of caution. It is still enough to empty a bathroom shelf for nine months, and for however long nursing lasts.

Regulators have tightened retinol separately, for an unrelated reason: cumulative vitamin A intake. Commission Regulation (EU) 2024/996 caps retinol, retinyl acetate and retinyl palmitate at 0.05% retinol equivalent in body lotion and 0.3% in other leave-on and rinse-off products, and requires the label line "Contains Vitamin A. Consider your daily intake before use", applying to products placed on the EU market from 1 November 2025 (Regulation (EU) 2024/996). Health Canada's Cosmetic Ingredient Hotlist, revised February 2025, sets its own ceilings — 0.2% total retinol equivalents for full-body leave-on products, 1% for other cosmetics — and added retinaldehyde to the restricted list (Health Canada Cosmetic Ingredient Hotlist, 2025).

Is bakuchiol safe during pregnancy?

No published human trial has tested bakuchiol in pregnancy. What can be said accurately is narrower: bakuchiol is not a vitamin A derivative, so the mechanism behind the retinoid warning does not apply to it. That is the absence of a known risk, not the presence of demonstrated safety. Anyone pregnant or nursing should ask their own provider.

Cleveland Clinic states the position without decoration — bakuchiol has been suggested as less of a pregnancy risk factor than retinol, but more research is needed before that claim can be substantiated (Cleveland Clinic, bakuchiol). Most dermatology and obstetric counselling lands in the same place: plausible, unstudied, discuss it. A brand that tells you otherwise is not reading the literature.

There is one genuine caveat worth naming, and it concerns purity rather than the molecule. Psoralea corylifolia seed also contains psoralen and isopsoralen, phototoxic furocoumarins, and a 2023 letter by Afzal and Sivamani in the Journal of Cosmetic Dermatology warned that bakuchiol reaching the market as a racemic mixture, a whole-plant extract or unrefined babchi oil can carry those compounds with it. Purified bakuchiol, the same authors note, is not photosensitising (Afzal & Sivamani, Journal of Cosmetic Dermatology, 2023). A 2025 review in Frontiers in Pharmacology catalogues hepatotoxicity signals associated with oral Psoralea corylifolia preparations in traditional medicine — a different route at a different dose from a face serum, but a good reason to care whether a formulator is using purified bakuchiol or crushed seed (Frontiers in Pharmacology, 2025).

The reasoning transfers to fragrance. The useful question in pregnancy is never "is it natural" but "what is the exposure, and what is known about it". The pregnancy-safe fragrance guide works through that logic for perfume, and the longer piece on what pregnancy-safe really means on a label takes apart the marketing term itself.

Does bakuchiol actually work like retinol?

In a 12-week randomised, double-blind trial of 44 adults, bakuchiol 0.5% applied twice daily and retinol 0.5% applied once daily produced statistically indistinguishable reductions in wrinkle surface area and hyperpigmentation. Retinol users reported more facial scaling and stinging. That is one small, well-run trial — not a body of evidence.

The study (Dhaliwal et al., British Journal of Dermatology, 2019) remains the strongest single piece of evidence the category has, and its limits are worth reading precisely. Forty-four participants is a small sample. Twelve weeks is short for a photoageing endpoint that develops over decades. And the comparator was cosmetic retinol at 0.5%, not prescription tretinoin — a substantially more powerful molecule with a substantially longer evidence base.

What the wider literature does not show

A 2024 systematic review in the Journal of Drugs in Dermatology assembled every human clinical trial of topical bakuchiol it could find — fifteen studies across ageing, acne and post-inflammatory hyperpigmentation — and found that twelve were unblinded, open-label trials with no control group. Ten used combination formulations, which makes it impossible to attribute any result to bakuchiol itself. Four did not state the concentration used. Heterogeneity across designs and outcomes was severe enough that a meta-analysis was not feasible, and the authors named a high risk of bias in the reported outcomes (Fanning et al., Journal of Drugs in Dermatology, 2024).

The honest grade, then, in three parts. Bakuchiol is well supported as better tolerated than retinol. It is plausibly supported as effective on fine lines and pigmentation, on the strength of one good trial and a long tail of weak ones. It is not supported as equivalent to a prescription retinoid, and anyone selling it as the third thing is ahead of the data.

Why 1%, when the trials used 0.5%

Aimée de Mars formulates its SÉRUM ANTI-ÂGE — a 30 ml facial oil, $71 CAD — with bakuchiol concentrated at 1% — twice the concentration used in the Dhaliwal trial.

Higher is not automatically better. No published trial has established a dose-response curve for bakuchiol above 0.5%, and implying one would be inventing evidence that does not exist. The defensible reading is narrower: 1% is a disclosed number rather than a position in an ingredient list, and disclosure is not a trivial virtue in this category — four of the fifteen trials in the 2024 systematic review did not specify the concentration their own test products contained. If a consumer cannot verify the dose, neither the efficacy claim nor the tolerability claim means anything.

What else is in the bottle

The serum's two supporting materials are Rosa damascena and immortelle (Helichrysum italicum), both long-standing materials in Provençal distillation. The evidence for each is preclinical and should be read as such. Rosa damascena essential oil inhibits tyrosinase, the rate-limiting enzyme in melanin synthesis, in enzymatic and cell-culture work — a mechanistic rationale for a pigmentation claim, not a clinical demonstration of one (Rosa × damascena essential oil, anti-tyrosinase activity, Frontiers in Pharmacology, 2024). Helichrysum italicum oil improved wound contraction and collagen deposition in a diabetic-rat model, again mechanistic rather than clinical (Helichrysum italicum topical formulations, wound healing, Pharmaceuticals, 2021).

Naming that distinction is the point. A face oil built on one molecule with a good randomised trial behind it and two botanicals with preclinical support is a defensible formulation. Calling the whole thing clinically proven would not be.

How to use a bakuchiol serum

  1. At night, on clean skin. The published trials dosed once or twice daily; an oil-based serum layers more comfortably under sleep than under sunscreen and makeup.
  2. Three to five drops, pressed rather than rubbed. Warm the oil between the palms first; pressing avoids the mechanical irritation that gets misattributed to the active.
  3. After water-based steps, before anything occlusive. Oils sit late in a layering sequence; an essence applied afterwards will not penetrate through one.
  4. Use it as a massage medium. An oil is the correct vehicle for manual facial work, which is where this article meets the maison's other subject — see the Kobido method and the Kobido self-massage sequence for a technique that does not stretch the skin.
  5. Patch test if your skin is reactive. Inner forearm, 48 hours. Better tolerated than retinol is a population statistic, not a personal guarantee.
  6. Give it twelve weeks. The head-to-head trial measured its endpoints at twelve weeks; judging a photoageing active at three tells you about its texture, not its effect.
  7. Keep the sunscreen. Nothing in the bakuchiol literature substitutes for photoprotection: the serum addresses the consequence, sunscreen addresses the cause.

Where this leaves the question

The traffic behind the phrase "bakuchiol serum" is not really about a plant phenol becoming fashionable. It is about a large group of people — pregnant, nursing, trying to conceive, or simply intolerant of retinoid dermatitis — were told to stop using the one anti-ageing active with a real evidence base and given nothing to replace it. Bakuchiol is the most credible answer on offer, and it is credible exactly to the degree that its limits are stated rather than hidden.

Aimée de Mars formulates without phthalates, synthetic musks or ethanol of undocumented origin, and states its actives at concentration rather than at "contains" — the bakuchiol in the anti-ageing serum is 1%, printed as a number. The rest of the face and body range sits in the skincare collection. What none of it can do is answer a question your obstetrician or midwife has not yet been asked; that conversation stays where it belongs.

Sources

Hero photograph by Arup Kumar Das on Unsplash.

pregnancy skincare
Share

Related Articles