Ingredient glossary
Retinal
On your shelf ✓What Retinal does for your skin, who it suits, and what to watch for.
INCI: Retinaldehyde
In short: Retinal delivers anti-aging support and texture, suits most skin types, and the evidence behind it is moderate. One thing to know: not for pregnancy or breastfeeding; introduce it gently.
- What it does
- Anti aging, texture
- Best for
- Most skin types
- Watch out
- Not for pregnancy or breastfeeding; introduce it gently
- Evidence
- Moderate
- How long
- 3 months+
how it was measured
Measured on crow's-feet wrinkle depth and surface roughness, silicone-replica optical profilometry, versus vehicle; the first visit at which the change was measurable.PMID 9843009PMID 29663701PMID 41729119
The molecule
The Glow take
One of the ingredients I trust most.
When your goal is softening fine lines and keeping skin firm, Retinal delivers, with a solid amount of evidence behind it. Skip it while you are pregnant or breastfeeding, and the rest of the time bring it in slowly.
— Maya
What it does
- anti aging
- texture
- brightening
Retinal scores -1 on barrier support. It is one of only 8 ingredients in our glossary that go below zero there.
A negative score means it asks something of the barrier rather than giving something back, which is why it needs introducing slowly and pairing with ingredients that do support the barrier. It is not a reason to avoid it: the same evidence that shows the irritation also shows the long-term gains. See where it sits against all 388 of them →
moderate evidenceThe one vehicle-controlled trial tested 0.05% for 44 weeks in 125 people and cut measured wrinkle depth and roughness where the vehicle changed nothing.
Why we say this
- The water-loss, hydration and pigment figures come from a separate 40-person trial with no placebo arm, so they are before-and-after readings on people using a cream: texture, TEWL and hydration improved on both 0.05% and 0.1% with no significant difference between them, and a 6.5% drop in melanin index reached significance only in the 0.1% group.
- Most of this work was run or funded by companies selling retinaldehyde.
On a label: works best above 0.05%.
The retinoid family, at a glance
10 retinoid entries in the base, Retinal among them. 7 of the 10 are a hero active.
- Retinyl Acetate3 steps to retinoic acid
- Retinyl Palmitate3 steps to retinoic acid
- Retinyl Propionate3 steps to retinoic acid
- Retinol2 steps to retinoic acid
- Retinal1 step to retinoic acid
- Tretinoin0 steps to retinoic acid
- Adapalene0 steps to retinoic acid
- Tazarotene0 steps to retinoic acid
- Retinyl Retinoateroute not established
- Hydroxypinacolone Retinoateroute not established
How much do you need?
0.05 to 0.1%
The vehicle-controlled trials worked between 0.05 and 0.1%; above the ceiling, more stopped adding effect or started adding irritation.
The trials, and what they used
- Floor 0.05%: 0.05% retinaldehyde cream reduced crow's-feet wrinkle and roughness features on silicone replicas at week 18 where its vehicle changed nothing, 125 patients randomised to retinaldehyde, retinoic acid or vehicle (Creidi 1998).
- Ceiling 0.1%, on tolerance rather than effect: Saurat 1994 applied 0.05, 0.1 and 0.5% for 1 to 3 months and found epidermal thickening rising with dose, but in 229 patients the 1% preparation was tolerated by only 70% and 0.5% only slightly better, while 0.05 and 0.1% allowed years of use.
- Direct comparisons of 0.1 against 0.05%: no difference in texture, water loss or hydration in 40 Korean women over 3 months, only the melanin index favoured 0.1% (Kwon 2018); larger wrinkle reductions with 0.1% in 56 women over 24 weeks, split-face (Deda 2026).
- Read again on 2026-09-21: We compared the activity and tolerance profile of a 0.05% retinaldehyde cream with a 0.05% retinoic acid cream and the retinaldehyde vehicle in patients with photodamaged skin of the face.
- (PMID 9843009)
Who it's for, and who should skip it
Tends to suit
- normal
- oily
- combo
Approach with care
- pregnancy
- breastfeeding
- very sensitive
- compromised barrier
- Clogs poresLow
- Irritation riskModerate
How it fits each skin type
- DrySuits
- CombinationSuits
- OilySuits
- SensitiveTake care
Take extra care if any of these apply to you: pregnancy, breastfeeding, compromised barrier.
Worried about purging? Here is what the trials actually measured →
Where it fits in your routine
Nobody has measured whether Retinal changes how your skin handles sun.
Why we say this
- No human study has measured minimal erythema dose, UV-induced erythema, sunburn cell formation or UV-induced DNA damage in skin treated with cosmetic retinol or retinaldehyde, so we do not know whether either changes how skin responds to UV; the SCCS opinion on Vitamin A (SCCS/1576/16, reaffirmed in the October 2022 revision SCCS/1639/21) names retinol and retinal and records no clinical reports of phototoxic or photosensitising effects, and when the EU restricted retinol in Commission Regulation (EU) 2024/996 it required only a vitamin A intake warning and no sun warning, but neither is a measurement of sun sensitivity.
SCCS Opinion on Vitamin A SCCS/1576/16 (6 October 2016) and its revision SCCS/1639/21 (adopted 24-25 October 2022), section on human data and phototoxicity/photosensitisation; Commission Regulation (EU) 2024/996 of 3 April 2024, Annex III entry 376; CIR Expert Panel re-review book for Retinol and Retinyl Palmitate (September 2012) and the reopened draft report (June 2013); Health Canada Cosmetic Ingredient Hotlist entries for Retinal and for Retinol and its esters; Antille and colleagues, Journal of Investigative Dermatology 2003 (retinyl palmitate human UVB challenge); Sorg and colleagues, Photochemistry and Photobiology 2005 (hairless mice); Smit and colleagues, Acta Dermato-Venereologica 2000 and Slade and colleagues, Photodermatology Photoimmunology and Photomedicine 2009 (retinoic acid). Three independent searches of PubMed via E-utilities, Europe PMC full text and ClinicalTrials.gov located no human study of either endpoint for cosmetic retinol or retinaldehyde, in either direction.
Give it three months or more — that is when the change first became measurable.
What was measured, and in whom
- Measured on: crow's-feet wrinkle depth and surface roughness, silicone-replica optical profilometry, versus vehicle.
- The only vehicle-controlled trial is Creidi et al. 1998: 125 patients on 0.05% retinaldehyde, 0.05% retinoic acid or the retinaldehyde vehicle, with a silicone replica of the left crow's feet taken at baseline, week 18 and week 44 and read by optical profilometry.
- Wrinkle and roughness features were significantly reduced with retinaldehyde at week 18; at week 44 the reduction was less pronounced, and the vehicle produced no significant change at any assessment point.
- The result therefore dates to week 18, not to the end of the study, and it does not keep growing with time on treatment.
- Because there is no assessment between baseline and week 18, the true onset could be earlier and simply was not looked for.
- The other retinaldehyde trials cannot date onset at all: the 40-woman 3-month study and the 56-woman 24-week study both compare 0.1% against 0.05% with no vehicle arm, which is where the often-quoted TEWL, hydration and 6.5% melanin-index figures come from.
- Most of this work was run or funded by companies selling retinaldehyde; 0.05% is consumer strength.
Creidi et al., J Am Acad Dermatol 1998;39(6):960-5, 'Profilometric evaluation of photodamage after topical retinaldehyde and retinoic acid treatment', PMID:9843009 - abstract read: replicas at baseline, weeks 18 and 44; significant at week 18, less pronounced at week 44, vehicle unchanged throughout. Concentration comparisons without a vehicle arm, not used for timing: PMID:29663701, PMID:41729119.
Light, and how to keep it
Retinal breaks down in light. The packaging is doing real work: opaque, and closed between uses.
how it was measured
Thus, irradiation (>300 nm) of retinal and retinol in methanol produces free radicals, as does irradiation of RP in dimethylformamide (13).PMID 16705812 · read in full
what the source asks of the bottle: opaque
Plays well with
- ceramides
- niacinamide
- panthenol
Retinal, answered
Is Retinal safe during pregnancy?
No. Retinal is best avoided during pregnancy and breastfeeding. If you are pregnant or trying, leave this one out and ask your doctor about a safe alternative.
Who should avoid Retinal?
Approach Retinal with care if any of these apply to you: pregnancy, breastfeeding, very sensitive, and compromised barrier. Its irritation risk is moderate, so introduce it slowly and patch test first.
What concentration of Retinal actually works?
Research tends to show Retinal starts pulling its weight above 0.05%. If a brand prints the percentage, that is the number to look for.
Does Retinal go off?
It breaks down in light, and that is the part that has been measured. Thus, irradiation (>300 nm) of retinal and retinol in methanol produces free radicals, as does irradiation of RP in dimethylformamide (13). So the packaging is not decoration: opaque, and closed between uses. Heat, air and simple age matter too, and we have not sourced numbers for those, so treat this as one reason a product weakens rather than as a shelf life.
During pregnancy
Leave Retinal out while you are pregnant or breastfeeding. It sits on the list of ingredients usually held back during this time.
When in doubt, check with your doctor or midwife. They outrank any website, including this one.When it is worth booking →
Commonly confused with
- RetinolRetinal sits one conversion step closer to the active form than retinol, so it tends to work a little faster.
Readers asked about Retinal
Where Retinal sits
On the map
- careful withAdapaleneAmmonium lactateBenzoyl PeroxideBetaine SalicylateBromelainand 23 more
- fine withNiacinamide
Compare Retinal with
42 comparisons built, the 1 most looked-up above.
Can you mix Retinal with
19 pairings built.
This, for every label you own
One ingredient is easy. A full label is where it gets confusing. The Ingredient Decoder reads any product and tells you the honest truth, ingredient by ingredient.
Sources & how we reviewed this
- PMID 29663701abstract read, not the full text
- PMID 9843009abstract read, not the full text
- PMID 36220974read in full
- PMID 41729119read in full
What changed on this page
- : the pH it needs
- : the effective dose from trials, the highest dose tested
- : light stability