Only a handful of places left for our September 2026 cohort. Applications close soonApply now →

MSTA
Your future
starts here.

Retinoids

Also known as: retinol, retinoid, tretinoin, retinaldehyde

Retinoids are vitamin A-related compounds used topically in cosmetics and medicines. Prescription retinoids have established roles in acne and photoageing; cosmetic retinol, retinaldehyde and retinyl esters require conversion or formulation-dependent delivery and do not inherit that evidence automatically.

Evidence status

Moderate

Strong for prescription topical retinoids in acne and for tretinoin in photoageing. Smaller, inconsistent and formulation-specific for cosmetic retinol and retinaldehyde — and in the one systematic review of over-the-counter products, eight of nine trials were manufacturer-sponsored.

What retinoids are, and the types available#

Retinoids are a family of vitamin A-related compounds, not a single ingredient. The word covers prescription tretinoin, prescription adapalene and oral isotretinoin on one side, and cosmetic retinol, retinaldehyde and the retinyl esters on the other — and they are not interchangeable names for one active.[13]

The distinction matters because the evidence belongs to specific molecules rather than to the category. A tretinoin trial is not evidence for a retinol serum, and treating the two as equivalent is the most common error in this area of skincare.

Use of retinoids in aesthetic practice#

Prescription retinoids are prescription-only medicines. You can support a client who is using one, and you can recommend that she be assessed for one, but starting or changing it is a prescriber’s decision. Cosmetic retinol, retinaldehyde and retinyl esters are yours to recommend and retail.

Cosmetic concentration limits in Great Britain and the EU

The EU has restricted retinol and retinyl esters in cosmetics to 0.05% retinol equivalents in body lotion and 0.3% RE in other leave-on and rinse-off products, with new products non-compliant from 1 November 2025 and existing stock withdrawn by 1 May 2027.

Those limits apply in the EU and in Northern Ireland. They do not automatically apply in Great Britain. GB operates the assimilated Cosmetic Products Regulation, which changes only by UK statutory instrument; Northern Ireland follows EU cosmetics law under the Windsor Framework.[3, 4]

The distinction is worth confirming before any retail claim is made about compliance, particularly where stock is held across both jurisdictions. This is a live area and the dates above are the operative ones.

Contraindications and cautions#

Pregnancy and planned pregnancy

The oral and topical routes sit differently, and the same MHRA source covers both. Oral retinoids sit under a formal Pregnancy Prevention Programme that women and girls of childbearing potential must be supported by. For topicalretinoids, the MHRA states that systemic exposure is thought to be negligible following application during pregnancy — and that their use is contraindicated during pregnancy as a precaution.[2]

Both halves of that position hold at once. The topical contraindication is real and is observed; it is a precaution against an exposure the regulator expects to be negligible, rather than the same evidential position as the oral programme. In practice the rule is the same whether a client is pregnant, planning or unsure: the decision belongs to the prescriber.

Existing prescription retinoid use

Adding a cosmetic retinoid to a prescribed one is a change to another clinician’s treatment plan. What the client is taking, at what strength, and for how long are consultation questions; the adjustment itself is not an aesthetic practitioner’s to make.

Compromised skin barrier

Dryness, erythema, stinging and peeling are recognised limiting effects of retinoids.[7, 8]i Introducing one to a barrier that is already compromised makes tolerance the first thing to fail, and tolerance is what determines whether the product is still in use several weeks later.

Clinical uses and the evidence behind them#

Acne vulgaris

This is the strongest indication. Topical retinoids are UK first-line for acne in defined combinations, and in a network meta-analysis of 40 trials and 18,089 participants, adapalene with benzoyl peroxide ranked most effective among topical options — 54% self-reported improvement against 35% for benzoyl peroxide alone.[1, 7, 8]

The ranking is clear, and it was drawn from evidence the reviewers themselves graded low-certainty.[8] It remains the best available answer and the one UK guidance reflects. The practical consequence is that the combination matters as much as the molecule.

Photoageing

Tretinoin has the best evidence of any topical retinoid here, and the denominators travel with the outcomes. All eightrandomised vehicle-controlled trials in the meta-analysis, covering 1,361 participants, favoured tretinoin for fine wrinkles (mean difference 0.412, 95% CI 0.233–0.590). Six of themassessed coarse wrinkles, and favoured it there too (0.245, 95% CI 0.119–0.370).[5]

Those effect sizes are real and modest: a measurable improvement in fine lines over months rather than a transformation. The trials behind the figure ran from 16 weeks to two years, which makes months the honest unit for setting an expectation.[5]

Best-evidenced is not the same as best. In a systematic comparison against other topical therapies, comparator agents performed equal to or better than tretinoin in 20 of 25 studies.[6]Tretinoin’s position rests on the depth of its evidence rather than on having outperformed the alternatives.

Cosmetic retinol and retinaldehyde

Most of the category’s retail volume sits here, and it is the thinnest part of the evidence base. The systematic review of over-the-counter vitamin A cosmetic products found nine trials, and four of them were null.[10]

That is not a reason to withdraw the recommendation. A tolerated cosmetic retinoid in continued use is worth more in practice than a stronger product abandoned early. It is a reason to promise less, and to recognise that where the result matters clinically, the evidence behind a prescription retinoid is different in kind.

Selecting a retinoid#

The conversion chain — retinyl esters to retinol, retinol to retinaldehyde, retinaldehyde to retinoic acid — is often read as a potency ranking, with each step weaker than the last. It does not function that way. Conversion efficiency, formulation, stability, delivery and individual tolerance all sit between the molecule and the result.[12, 13]i

The working consequence is that a well-formulated precursor used nightly will outperform a stronger one used twice and abandoned. Tolerance and adherence are the modifiable variables; nominal strength is not.

Retinol or retinaldehyde

Retinaldehyde sits one conversion step closer to retinoic acid, and some reviews regard it as the more efficient cosmetic retinoid. That is a reasonable position, but it is a review-level view rather than a demonstrated clinical ranking between finished products — a tie-breaker rather than a deciding factor.[12, 13]i

Cosmetic strength relative to prescription tretinoin

The defensible position is that cosmetic precursors deliver less retinoic acid receptor activity than prescription tretinoin, by an amount that varies with formulation and has never been reliably quantified.[no source found] Less tidy than a numerical multiple, and more useful in a consultation, because it locates the decision in the suitability of the product rather than in a league table.

Introduction and build-up

Cautious introduction improves tolerability, so frequency is built up rather than started nightly, and the first weeks are the part of the plan most likely to fail.[7, 8]i Setting out in advance what a normal early response looks like, and what to do when it appears, is what keeps a client on the product past the point at which most stop.

Adverse effects and their management#

Expected and usually transient effects

Dryness, erythema, stinging and peeling are recognised limiting effects. They are not markers of progress. Irritation is an adverse effect, and nothing links its severity to efficacy, so advising a client to continue through it has no support.[7, 8]i

Irritation was more common on tretinoin than on vehicle in the photoageing trials, and the trials varied considerably in how much.[5] It is best treated as an anticipated cost to be managed rather than a stage to be reached.

The reported “purge”

No trial evidence establishes a predictable retinoid-induced flare, and no normal duration for one has been defined. A review that went looking for it found no primary clinical-trial evidence.[9]

The practical position that follows is narrower than calling the purge a myth: a client whose acne worsens on a retinoid should be reassessed rather than advised to wait it out. If it settles, that is a useful observation. If it does not, it is information rather than a phase.

Reducing, stopping and referring

Frequency is reduced where irritation is limiting but the barrier is intact, and stopped where it is not. Referral is indicated where the acne itself has worsened, or where the presentation needs prescription management rather than an adjustment of frequency.

Exposure during an unrecognised pregnancy

A client who used a retinol serum before she knew she was pregnant has not done what the oral warnings describe.[2] The reassurance is worth giving plainly, and the question then belongs with her prescriber or midwife.

Referral and scope boundaries#

The first-line evidence in acne sits with medicines rather than cosmetics, which makes referral a routine part of competent practice in this category rather than a limitation of it. Referral is indicated in four situations:

  • Acne requiring prescription management. The referral carries what was observed, what the client has been using and for how long, and what changed.
  • Pregnancy, planned pregnancy or breastfeeding.
  • An existing prescription retinoid where the client wants to add to or change the regime.
  • Acne that worsens after a retinoid is introduced and does not settle at a reduced frequency.

Barrier support, photoprotection and adherence continue alongside a prescriber. They account for a substantial part of the result and require no prescription.

Mechanism of action#

Retinoic acid is the form that binds nuclear retinoic acid receptors. Cosmetic precursors have to convert to reach it — retinyl esters to retinol, retinol to retinaldehyde, retinaldehyde to retinoic acid — which is why the family behaves so differently from one finished product to the next, and why formulation carries as much weight as the molecule named on the front of the bottle.[13]

Commonly misstated claims#

Four statements circulate widely in the trade. Each is set out below with what the primary literature was found to support.

“Retinol is 20 times weaker than tretinoin.”

The multiple is sometimes given as ten and occasionally as a hundred. A search for the primary source found none — no head-to-head human assay establishing any universal potency multiple.[no source found] The figure traces to review articles that assert it without comparative data.[12, 13]i A single multiple could not be correct in any case, since potency depends on the molecule, the conversion steps, the vehicle, the concentration and the endpoint measured.

Supported statement: cosmetic retinoids deliver less receptor activity than prescription tretinoin, by an amount that has not been reliably measured.

“Retinol is proven for facial photoageing.”

One study is routinely cited as the proof: a 0.4% retinol lotion producing a fine wrinkle score change of −1.64 against −0.08 for vehicle. The lotion was applied by study staff, up to three times weekly for 24 weeks, to the sun-protected upper inner arms of 36 nursing-home residents with a mean age of 87, of whom 23 completed — and the biopsy subsets were six and four people. It studied naturally aged, non-facial, sun-protected skin, and was excluded from the facial over-the-counter review for that reason. Four of its authors held patent or royalty interests.[11]

Supported statement: a sound study of a different question, and not evidence for facial photoageing.

“Tretinoin triples the risk of side effects.”

The meta-analysis reports an oddsratio of 3.140 (95% CI 1.819–5.419) with substantial heterogeneity (I² 73.3%). Odds and risk are not interchangeable, and an odds ratio is not an event-rate multiplier. Egger’s test also indicated publication bias for the fine-wrinkle outcome, though not for coarse wrinkles.[5]

Supported statement: irritation was clearly more common on tretinoin, and the trials disagreed on the magnitude.

“The evidence for cosmetic retinol serums is strong.”

In the only systematic review of over-the-counter vitamin A products, eight of the nine trials were sponsored by the manufacturer of the product being tested.[10] The same standard applies in the other direction: the review reporting no evidence for a purge is itself a three-page narrative piece whose stated method was a PubMed and internet search, and it is the only support that exists either way.[9]

Supported statement: the cosmetic evidence is largely industry-funded, and a third of it was null.

Areas of remaining uncertainty#

  • The potency relationship between cosmetic precursors and prescription tretinoin. No reliable comparative figure exists, which makes tolerance and formulation the practical basis for selection.
  • Whether a predictable retinoid flare exists, and for how long. Until it is measured, reassessment rather than reassurance is the defensible response.
  • How much of the reported over-the-counter benefit survives the removal of manufacturer sponsorship.
  • Whether early differences between retinaldehyde and retinoic acid persist — one trial found them attenuating by week 44 — leaving the long-run choice between them unsettled.
  • Any evidence-based pre-procedure stop period for topical retinoids. None of the sources reviewed here establishes one, which leaves the device or product instructions for the specific procedure as the operative guidance.

Frequently asked questions#

Should retinoids be stopped before a treatment?

Follow the instructions for the specific device or product. Pausing is common practice, but none of the sources reviewed here establishes an evidence-based stop period, so a general rule would be an invention rather than a finding.

Can retinoids be used with vitamin C?

Yes. The traditional morning-C, night-retinoid split reflects photoprotection logic and tolerance rather than a chemical incompatibility.

How long before a client sees a result?

Months. The photoageing trials ran from 16 weeks to two years and the measured effects were modest, which makes early expectation- setting more useful than later explanation.[5]

References#

Each source is graded by evidence tier. Tier 4 material (manufacturer documents, expert consensus, practitioner experience) is useful for protocol and context, and is never presented as equivalent to independent clinical evidence.

  1. National Institute for Health and Care Excellence. Acne vulgaris: management. NICE guideline NG198.Tier 1Supports: UK first-line acne options, which include topical adapalene and topical tretinoin in defined combinations alongside benzoyl peroxide, topical clindamycin and oral tetracyclines.
  2. Medicines and Healthcare products Regulatory Agency. Oral retinoid medicines: revised and simplified pregnancy prevention educational materials. Drug Safety Update.Tier 1Supports: Published 19 June 2019. Covers BOTH routes, and the difference matters. ORAL: women and girls of childbearing potential taking oral retinoids for dermatological conditions must be supported by a Pregnancy Prevention Programme — oral isotretinoin for severe acne, oral acitretin for severe psoriasis, oral alitretinoin for chronic severe hand eczema; the Programme's requirements were unchanged by this update. TOPICAL: 'Systemic exposure is thought to be negligible following application of topical retinoids (topical adapalene, alitretinoin, isotretinoin, tazarotene, and tretinoin) during pregnancy', and 'use of topical retinoids is contraindicated during pregnancy as a precaution'. So the topical contraindication is a precaution against negligible expected exposure, not the same evidential position as the oral programme.
  3. European Commission. Commission Regulation (EU) 2024/996 of 3 April 2024 amending Regulation (EC) No 1223/2009 as regards vitamin A and other substances.Tier 1Supports: Restricts retinol and retinyl esters to 0.05% retinol equivalents in body lotion and 0.3% RE in other leave-on and rinse-off products, with new products non-compliant from 1 November 2025 and existing stock withdrawn by 1 May 2027.
  4. Office for Product Safety and Standards. Cosmetic Products Enforcement Regulations 2013: guidance for Great Britain and Northern Ireland.Tier 1Supports: Great Britain operates the assimilated Cosmetic Products Regulation and does not automatically adopt EU amendments; Northern Ireland follows EU cosmetics law under the Windsor Framework.
  5. Huang HY, Lee LT. Tretinoin for Photodamaged Facial Skin: Systematic Review and Meta-Analysis of Randomized Controlled Trials.Tier 1Supports: Eight randomised vehicle-controlled trials, 1,361 participants in total, follow-up 16 weeks to two years. DENOMINATORS DIFFER BY OUTCOME: all eight trials assessed fine wrinkles; only SIX assessed coarse wrinkles (Olsen et al. and Weinstein et al. excluded). Fine wrinkles mean difference 0.412 (95% CI 0.233–0.590, P<0.001); coarse wrinkles MD 0.245 (95% CI 0.119–0.370, P<0.001). Adverse events are reported as an ODDS RATIO: OR 3.140 (95% CI 1.819–5.419) with substantial heterogeneity, I² 73.261%. An odds ratio is not a risk ratio and not a 3.14-fold event rate. Publication bias by Egger's test was indicated for the fine-wrinkle outcome (P<0.001) and not for coarse wrinkles (P=0.457). Competing interests: none.
  6. Siddiqui Z, Zufall A, Nash M, et al. Comparing Tretinoin to Other Topical Therapies in the Treatment of Skin Photoaging: A Systematic Review.Tier 1Supports: Comparative photoageing evidence. Comparator agents performed equal to or better than tretinoin in 20 of 25 studies — tretinoin is the best-evidenced option, not a demonstrated best performer.
  7. Kolli SS, Pecone D, Pona A, Cline A, Feldman SR. Topical Retinoids in Acne Vulgaris: A Systematic Review. American Journal of Clinical Dermatology. 2019;20(3):345–365.Tier 1Supports: Efficacy of topical retinoids in acne and their recognised local tolerability effects.
  8. Stuart B, Maund E, Wilcox C, et al. Topical preparations for the treatment of mild-to-moderate acne vulgaris: systematic review and network meta-analysis. British Journal of Dermatology. 2021;185(3):512–525.Tier 1Supports: Network meta-analysis of 40 trials and 18,089 participants. Adapalene with benzoyl peroxide ranked most effective among topical options — 54% self-reported improvement against 35% for benzoyl peroxide alone. Certainty of evidence was rated LOW on the CINeMA framework.
  9. Yentzer BA, McClain RW, Feldman SR. Do topical retinoids cause acne to 'flare'? Journal of Drugs in Dermatology. 2009;8(9):799–801.Tier 4Supports: Found no primary clinical-trial evidence for a predictable retinoid-induced acne flare. A three-page narrative review whose stated method was a PubMed and Google internet search — thin support, and stated as such.
  10. Spierings NMK. Evidence for the Efficacy of Over-the-counter Vitamin A Cosmetic Products in the Improvement of Facial Skin Aging: A Systematic Review. Journal of Clinical and Aesthetic Dermatology. 2021;14(9):33–40.Tier 1Supports: Nine trials of over-the-counter vitamin A cosmetic products, four of which were null. EIGHT OF THE NINE were manufacturer-sponsored.Funding / interest: Not the review itself, but its finding: eight of the nine included trials were sponsored by the manufacturer of the product tested.
  11. Kafi R, Kwak HS, Schumacher WE, et al. Improvement of naturally aged skin with vitamin A (retinol). Archives of Dermatology. 2007;143(5):606–612.Tier 2Supports: NATURALLY AGED, NOT PHOTOAGED, AND NOT FACIAL. 0.4% retinol lotion applied by study staff up to three times weekly for 24 weeks to the sun-protected upper inner arms of 36 nursing-home residents of mean age 87. Fine wrinkle score change −1.64 versus −0.08 for vehicle. This trial was excluded from the facial over-the-counter review at source 10.
  12. Creidi P, Vienne MP, Ochonisky S, et al. Profilometric evaluation of photodamage after topical retinaldehyde and retinoic acid treatment. Journal of the American Academy of Dermatology. 1998;39(6):960–965.Tier 2Supports: 125 participants comparing retinaldehyde and retinoic acid on photodamage by profilometry. Differences between arms attenuated by week 44.
  13. Sorg O, Antille C, Kaya G, Saurat JH. Retinoids in cosmeceuticals. Dermatologic Therapy. 2006;19(5):289–296.Tier 4Supports: The cosmetic retinoid conversion pathway. NOTE: this review states retinaldehyde 'seems to be the most efficient cosmeceutical retinoid', so it must NOT be cited as rebutting a ranking of retinaldehyde above retinol — it supports one.

Continue learning with MSTA#

Skinipedia is written by the educators at MSTA, the medic-led skincare training academy in Liverpool.