OTC Retinol vs Prescription Retinoids: What the Evidence Shows

Key points

• Retinol and retinaldehyde are precursors that skin can convert to retinoic acid, the active form. Measured conversion is small and was shown mainly in lab models, not as a clinical efficiency figure [1][2][3].
• Trials of retinol and retinaldehyde for photoaging are small, short and often industry-linked. A 2021 review of nine vehicle-controlled OTC retinol trials found four showed no difference and five had major flaws [4].
• Two small head-to-head trials found retinol products comparable to tretinoin on the measures reported, with less irritation [5][6]. Neither settles the question.
• The tier-list label “weak” for OTC retinol is partly supported: weaker per unit than tretinoin on molecular data, but not shown to be ineffective.

What differs between them

All five are vitamin A derivatives. Retinol is an alcohol; retinaldehyde is one step closer to the acid; tretinoin is the acid itself (all-trans retinoic acid). Adapalene and tazarotene are synthetic: adapalene acts directly, tazarotene is a prodrug. The conversion chain, retinol to retinaldehyde to retinoic acid, happens inside skin cells, and each step costs potency. For tazarotene see tazarotene vs tretinoin evidence.

What is measured about conversion in skin

• Human skin in vivo (1995). In a double-blind study, retinol, retinoic acid or vehicle went on normal buttock skin under occlusion for four days. Retinol thickened the epidermis about as much as retinoic acid, with little or no redness, and raised retinoid-binding proteins. Retinoic acid was undetectable or only trace after retinol. The authors concluded retinol may act as a prohormone [1]. The record we opened does not give group sizes or funding.
• Cultured cells and animals (1999). Keratinocytes converted retinaldehyde to retinoic acid, more as they differentiated. In vivo, skin absorbed retinaldehyde and mostly made retinyl esters, a storage form, releasing relatively little retinoic acid [2]. That study used mouse skin and human plasma, not human skin treated in vivo.
• Skin-equivalent model (2024). In a 3-D human skin model, retinol and retinaldehyde each produced about 0.6 to 0.7 micromolar retinoic acid at 2 and 24 hours. Retinyl palmitate produced none detectable [3]. This is a tissue model, not a person.
• Biomarkers in patients (2022). In a randomized trial, a retinol/retinyl ester product raised the retinoic acid-responsive gene CRABP2 but did not significantly change procollagen I [6]. Treatment was daily for 24 weeks.

So: conversion occurs, but we found no study that reports what share of applied retinol becomes retinoic acid in living human skin. Statements such as “X times weaker” are estimates, not measured constants.

Randomized and controlled trials

Study	Arms (concentrations define arms only)	Size, length	Funding / notes
Kafi 2007 [7]	Retinol 0.4% lotion vs vehicle, opposite arms	36 residents, mean age 87; 24 weeks	Record lists NIH and non-US government support; fine wrinkles improved vs vehicle
Kikuchi 2009 [8]	Retinol 0.075% vs vehicle, split face; a second study used 0.04%	57 women, 26 weeks; second study 13 weeks	Funder not given in record; fine wrinkling improved in 27/54 vs 13/54
Creidi 1998 [9]	Retinaldehyde 0.05% vs retinoic acid 0.05% vs vehicle	125 analysed (40/40/45); 44 weeks	Funding not stated in the record. Both actives beat vehicle at week 18; effects smaller at week 44; acid caused more irritation
Draelos 2020 [5]	Retinol serums 0.25 to 1.0% vs tretinoin 0.025 to 0.1%, escalating	45 women aged 35-65; 12 weeks, double-blind	Funding not in record. Equivalent on investigator and subject scores at 12 weeks
Chien 2022 [6]	Retinol with retinyl esters (1.1% total) vs tretinoin 0.02%	20 analysed of 24 enrolled; 24 weeks, double-blind	Funding not in record. No significant difference in photoaging score; erythema 11% vs 64%
Farris 2024 [10]	Retinol 0.1% vs vehicle, six pooled studies	237 vs 234 women; 12 weeks	Single dermatologist rater; pooled analysis; funding not in the record we saw
Monteil 2025 [11]	Retinaldehyde with niacinamide cream vs untreated half face	66 people after cosmetic procedures; 3 months	Funded by Pierre Fabre; four authors affiliated with the funder

Strengths and limits: most arms ran 12 to 26 weeks, rated mild to moderate photoaging, and measured fine wrinkles. Only Creidi ran near a year, and it used retinaldehyde, not retinol. The 2021 systematic review concluded the positive retinol trials should not guide clinical decisions and that high-quality trials are lacking [4]. A branded product is often tested with its own vehicle, so results do not transfer between products.

Head-to-head comparisons with tretinoin are the closest answer to the “weak” question. Draelos and Chien found no significant efficacy difference, with milder irritation for retinol in Chien [5][6]. Both are small. Draelos used step-up concentrations on both sides, so it does not compare equal strengths. A 1999 tolerance study found retinol and retinaldehyde much less irritating than retinoic acid in patch tests (n=6) and, over 44 weeks, retinaldehyde produced less early erythema than retinoic acid in 355 users [12].

Adapalene: labelled uses

In the US, Differin adapalene 0.1% gel is an OTC drug for acne in people aged 12 and over, and the label asks pregnant or breastfeeding users to consult a doctor [13]. The 0.3% gel is Rx-only for acne at age 12 and over [14]. Neither label lists wrinkles. In the UK, the emc lists Differin 0.1% gel as a prescription-only medicine for mild to moderate acne (revised 30 September 2026) [15]. We did not check other UK adapalene products or German status.

Regulatory status: drug or cosmetic

• US (FDA, page current as of 2022-02-25): cosmetics need no FDA approval before sale. A product intended to remove wrinkles or increase collagen production counts as a drug [16]. The page does not name retinol. Tretinoin and tazarotene are Rx drugs; tretinoin’s only wrinkle label, Renova 0.02%, says it “DOES NOT ELIMINATE WRINKLES” [17].
• EU (SCCS, opinion SCCS/1576/16 adopted 2016-10-06): the committee considered retinol-equivalent levels up to 0.05% in body lotion and up to 0.3% in other leave-on and rinse-off products safe per se. It flagged teratogenicity as a critical endpoint and noted that total cosmetic exposure could reach about 97% of the 5,000 IU/day upper limit if all products held the maximum [18]. We could not open Regulation (EU) 2024/996 or the 2022 revision of the opinion, so we do not state transition dates or label wording.
• UK cosmetics rules, Germany’s national rules, EMA decisions for individual products: not checked.

Pregnancy and safety

EMA said on 2018-03-23 that topical retinoids, including adapalene, tazarotene and tretinoin, should not be used in pregnancy or when planning one, as a precaution; it called systemic absorption “negligible” and harm unlikely [19]. The UK Differin SmPC contraindicates pregnancy and planned pregnancy [15]. The US Renova label says it “should not be used during pregnancy” [17]. The US OTC Differin label advises asking a doctor and says other retinoids have caused birth defects, with no specific evidence of this for topical adapalene [13].

A 2026 Nordic register study followed 3,869,586 infants; 2,172 had first-trimester topical retinoid exposure. Major malformations occurred in 3.3% of exposed and 3.0% of unexposed (adjusted risk ratio 1.1, 95% CI 0.87 to 1.38). It covered births only, and the abstract does not split results by drug [20]. We did not find a comparable pregnancy study of cosmetic retinol.

Verdict on the “weak” claim

@Hoostethics’ topical skincare tier list (2026-10-19), quoted in a post of 2026-11-01, put “standard OTC retinol” in B tier as “weak, inefficient two-step cellular conversion” [21].

Partly supported. The “two-step” description is accurate: retinol must become retinaldehyde, then retinoic acid [1][2]. Lab data show retinaldehyde stored mostly as esters [2] and retinol producing retinoic acid in a skin model [3]. But “weak” implies a clinical shortfall. Two small head-to-head trials found no significant efficacy difference from tretinoin, while the larger body of vehicle-controlled evidence is thin and flawed rather than negative [4][5][6]. The honest reading: no good evidence that OTC retinol equals prescription retinoids, and none that it fails.

Risks and unknowns

• Documented: redness, scaling, dryness and burning from all retinoids, less with retinol and retinaldehyde in the studies above [9][12]; photosensitivity advice appears on the adapalene label [13].
• Populations: pregnancy and planned pregnancy (see above); label evidence in darker skin is small [17].
• Not studied or not found: conversion efficiency in living skin; long-term retinol trials beyond about a year; independent, large trials; retinol-versus-tretinoin comparisons at equal strengths; cosmetic retinol in pregnancy; retinol next to other vitamin A sources in people near upper intake limits.

Frequently asked questions

Does retinol turn into tretinoin in the skin? Retinoic acid is the same molecule as tretinoin. Lab and tissue-model studies show retinol and retinaldehyde can be converted to it, but a 1995 human study found only trace amounts [1][3]. We found no measurement of how efficient the process is in living skin.

Is adapalene available without a prescription? In the US, adapalene 0.1% gel is OTC for acne in people 12 and over, and 0.3% is Rx-only [13][14]. In the UK, the emc lists Differin 0.1% gel as prescription-only [15].

Are retinol and retinaldehyde safe in pregnancy? Authorities restrict prescription and drug topical retinoids in pregnancy as a precaution [15][19]. The EU safety committee names vitamin A teratogenicity as a critical concern for cosmetics [18]. We found no pregnancy outcome study of cosmetic retinol.

Do EU rules limit retinol in cosmetics? The SCCS opinion considered up to 0.05% in body lotion and 0.3% in other products safe [18]. We could not open the implementing regulation, Regulation (EU) 2024/996; our tazarotene article cites it for the same two limits and for the dates after which non-compliant products must leave the market, which we did not confirm here.

Sources

1. Kang S et al., “Application of retinol to human skin in vivo induces epidermal hyperplasia… without measurable retinoic acid levels or irritation”, J Invest Dermatol (1995). Abstract read via Europe PMC; PubMed blocked us. https://pubmed.ncbi.nlm.nih.gov/7561157/
2. Sorg O et al., “Metabolism of topical retinaldehyde”, Dermatology (1999). Abstract via Europe PMC. https://pubmed.ncbi.nlm.nih.gov/10473954/
3. Kim JE et al., spectral and mass characterization of retinoid conversion to retinoic acid in a 3-D human skin equivalent model, Eur J Pharm Sci (2024). Abstract via Europe PMC. https://doi.org/10.1016/j.ejps.2024.106784
4. Spierings NMK, “Evidence for the efficacy of over-the-counter vitamin A cosmetic products in the improvement of facial skin aging: a systematic review”, J Clin Aesthet Dermatol (2021). https://pubmed.ncbi.nlm.nih.gov/34980969/
5. Draelos ZD, Peterson RS, “A double-blind, comparative clinical study of newly formulated retinol serums vs tretinoin cream in escalating doses”, J Drugs Dermatol (2020). https://pubmed.ncbi.nlm.nih.gov/32574009/
6. Chien AL et al., “Biomarkers of tretinoin precursors and tretinoin efficacy in patients with moderate to severe facial photodamage: a randomized clinical trial”, JAMA Dermatol (2022). https://pubmed.ncbi.nlm.nih.gov/35675051/
7. Kafi R et al., “Improvement of naturally aged skin with vitamin A (retinol)”, Arch Dermatol (2007). https://pubmed.ncbi.nlm.nih.gov/17515510/
8. Kikuchi K et al., “Improvement of photoaged facial skin in middle-aged Japanese females by topical retinol”, J Dermatolog Treat (2009). https://pubmed.ncbi.nlm.nih.gov/20078381/
9. Creidi P et al., “Profilometric evaluation of photodamage after topical retinaldehyde and retinoic acid treatment”, J Am Acad Dermatol (1998). https://pubmed.ncbi.nlm.nih.gov/9843009/
10. Farris P et al., “Efficacy and tolerability of topical 0.1% stabilized bioactive retinol for photoaging: a vehicle-controlled integrated analysis”, J Drugs Dermatol (2024). https://pubmed.ncbi.nlm.nih.gov/38564380/
11. Monteil C et al., “Enhancing facial rejuvenation outcomes with a novel retinaldehyde-based cream: a comparative randomized intra-individual study”, J Cosmet Dermatol (2025). https://pubmed.ncbi.nlm.nih.gov/41312593/
12. Fluhr JW et al., “Tolerance profile of retinol, retinaldehyde and retinoic acid under maximized and long-term clinical conditions”, Dermatology (1999). https://pubmed.ncbi.nlm.nih.gov/10473963/
13. FDA via DailyMed, “Differin (adapalene) gel 0.1%”, OTC label (updated 2022-12-07). https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0739d631-171b-42a8-bd55-0022b8df2d8a
14. FDA via DailyMed, “Differin (adapalene) gel 0.3%”, Rx label (revised 2023-12). https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=a0031324-92a6-4c11-90e2-c2818f7278ec
15. electronic Medicines Compendium, “Differin 0.1% w/w Gel” SmPC (revised 2026-09-30). https://www.medicines.org.uk/emc/product/921/smpc
16. FDA, “Wrinkle Treatments and Other Anti-aging Products” (current as of 2022-02-25). https://www.fda.gov/cosmetics/cosmetic-products/wrinkle-treatments-and-other-anti-aging-products
17. FDA via DailyMed, “Renova (tretinoin) cream 0.02%” (updated 2026-02-24). https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=9f0a0e8d-1051-48a4-80c5-1bd260dd9c36
18. European Commission, SCCS opinion SCCS/1576/16 on vitamin A (retinol, retinyl acetate, retinyl palmitate), adopted 2016-10-06. https://health.ec.europa.eu/publications/vitamin-retinol-retinyl-acetate-retinyl-palmitate_en
19. European Medicines Agency, “Updated measures for pregnancy prevention during retinoid use” (2018-03-23). https://www.ema.europa.eu/en/news/updated-measures-pregnancy-prevention-during-retinoid-use
20. Refsum E et al., “Topical retinoid use in women of reproductive age and risk of major congenital malformations in exposed pregnancies: a Nordic cohort study”, Br J Dermatol (2026). https://pubmed.ncbi.nlm.nih.gov/41365815/
21. @Hoostethics on X, post of 2026-11-01 (id 2116976569034092805) quoting the 2026-10-19 tier list; text taken from the project’s saved copy (fxtwitter mirror), X itself not opened. https://x.com/Hoostethics/status/2116976569034092805 ; original topical skincare tier list of 2026-10-19: https://x.com/Hoostethics/status/2112214865779806696

This article is information, not medical advice.