What is tretinoin and what does it actually do for your skin?

What is tretinoin, and is it a retinoid, a retinol, or a steroid?

Tretinoin is a retinoid, not a steroid. It is all-trans-retinoic acid, a first-generation vitamin A derivative, and FDA labeling calls it "a retinoid and a metabolite of naturally occurring Vitamin A" (Retin-A Micro label).

Retinol is a different molecule. We are not comparing its strength with tretinoin here, because the research behind this piece doesn't cover that. More background lives on our tretinoin topic page.

What is it used for? Two things carry FDA labels. Retin-A Micro gel is indicated for acne in people 12 and older, applied once daily in the evening. Renova 0.02% cream is indicated only as an adjunct for "mitigation (palliation) of fine facial wrinkles" alongside skin care and sun avoidance (Renova label, 2014).

A 2025 narrative review covering studies from 2000 to July 2025 found robust trial and systematic-review support only for acne and photoaging. Everything else below is a step down from that.

How does tretinoin work on skin cells?

It binds retinoic acid receptors inside skin cells and changes which genes those cells switch on. Tretinoin activates three nuclear receptors (RARα, β and γ), which alters how epidermal cells grow and mature. RARγ is the abundant one in the epidermis, and the 2025 review ties it to both the benefit and the irritation.

That's the neat version. The FDA label admits it has "not been established" whether the clinical effects come through receptor activation, through irritation, or both.

Why it helps acne: it loosens the buildup of dead cells in follicles and normalizes how keratinocytes multiply and mature, which may help prevent microcomedones, the earliest acne lesions.

Why it helps wrinkles: lab work suggests it activates fibroblasts, increases collagen synthesis and inhibits collagen-degrading enzymes (MMP-1 and MMP-8). Those are mostly mechanistic and preclinical findings. In humans the label says adequate data haven't shown increases in desmosine, hydroxyproline or elastin mRNA. One positive result: a 0.1% cream increased procollagen I propeptide staining on photodamaged forearm skin.

It also appears to reduce epidermal melanin, possibly by inhibiting tyrosinase, which is the proposed basis for its use on dark spots.

Acne: the best-supported use

For acne, the evidence is strong. The 2024 American Academy of Dermatology guideline makes strong recommendations for topical retinoids, including tretinoin, adapalene, tazarotene and trifarotene. It describes them as comedolytic and anti-inflammatory, and says they improve dyspigmentation and help maintain clearance.

Fixed-dose combinations of a retinoid with benzoyl peroxide or a topical antibiotic also get a strong recommendation backed by moderate-certainty evidence.

The label trials (12 weeks, vehicle-controlled) show real but not miraculous numbers. With the 0.1% gel, total lesions fell 45% versus 23% on vehicle in one study and 32% versus 16% in another. With 0.04%, reductions were 40% versus 8% and 35% versus 20%.

An "excellent" global rating was uncommon even on the active drug: 35% versus 11% in the first 0.1% study, and 14% versus 5% in the first 0.04% study.

Timing: the label says improvement may show at 2 weeks, but more than 7 weeks may be needed for sustained benefit. Can it cause acne? It can look that way at first. The label describes an early worsening of inflammatory acne and says that is not a reason to stop if you tolerate the drug. It doesn't call it a purge or say how common it is.

Wrinkles and photoaging: real, but modest

Tretinoin does improve fine and coarse wrinkles compared with vehicle, but the label claims are more conservative than the marketing. A 2025 meta-analysis pooled 8 vehicle-controlled trials with 1,361 patients, followed from 16 weeks to 2 years. Both fine and coarse wrinkles improved, and the results held up in sensitivity analyses.

Two caveats. The effect sizes are in each study's own score units, so don't read them as a percentage fewer wrinkles. And a statistical test hinted at possible publication bias for fine wrinkles. Adverse events were also more likely with tretinoin (odds ratio 3.14).

Head-to-head trials suggest it isn't uniquely powerful. One compared it with 0.3% adapalene in 86 women over 24 weeks and found comparable results. A 173-patient trial found tazarotene 0.1% worked faster than tretinoin 0.05% for fine wrinkling and mottled pigmentation.

The Renova label says the product "DOES NOT ELIMINATE WRINKLES, REPAIR SUN-DAMAGED SKIN, REVERSE PHOTOAGING, or RESTORE MORE YOUTHFUL" skin. That label dates from 2014, but the point stands: this is palliation, not reversal.

Does it tighten skin? There is data on collagen and wrinkles, not on laxity. The same label says no mitigating effect on skin laxity was demonstrated.

Dark spots, melasma and acne scars: how strong is the evidence?

For melasma it's moderate, for other dark spots it's thin, and for pitted acne scars it's essentially missing. A melasma systematic review rates tretinoin 0.05% or 0.025% cream as grade B, level 1 evidence. A 2024 update says hydroquinone and tretinoin should be offered as first-line treatments, with triple-combination cream suggested for moderate-to-severe or stubborn cases and strict sun protection with tinted sunscreen.

For post-inflammatory hyperpigmentation, the 2025 review cites only smaller trials and prospective studies. The Renova label goes further the other way, saying no effect was demonstrated on mottled hyperpigmentation or sun spots (lentigines). Those two positions don't line up, so treat the sun-spot claims cautiously.

We found no effect sizes for melasma, and no data on whether irritation can worsen pigmentation in darker skin. If you're prone to dark marks, ask a clinician how to ramp up slowly.

Acne scars: the 2025 review mentions tretinoin only for hypertrophic scars, keloids and post-burn scars, and rates that evidence low. We found nothing supporting pitted scars. What's reasonable: it is used for active acne, and the AAD notes retinoids improve the dark marks acne leaves behind.

Stretch marks, hair, rosacea, molluscum, pores: hype or help?

Mostly hype, or at least unproven. Here is the map.

Stretch marks. A network meta-analysis ranked topical tretinoin the worst performer for effectiveness and satisfaction against options like fractional CO₂ laser, radiofrequency, PRP and microneedling. A topical-therapy review found only 11 studies and "a distinct lack of evidence" for every topical.

StatPearls notes possible benefit for early red stretch marks over 6 months versus placebo. Mature white ones aren't supported. And stretch marks often appear during pregnancy, when tretinoin should be avoided.

Hair growth. The 2025 review lists alopecia areata and androgenetic alopecia only as conditions where smaller trials "indicate potential efficacy." We couldn't find doses, sample sizes or any society endorsement. Call it preliminary and off-label.

Rosacea and molluscum. Off-label, and we found no guideline-level evidence either way.

Blackheads, pores and milia. Blackheads are plugged follicles, and clearing them fits tretinoin's comedolytic action in acne. We found no evidence that it shrinks pores or removes milia, so don't expect either.

Why is tretinoin prescription only, and is it safe?

The sources we reviewed don't state why it's prescription-only. Our inference is that the label's irritation, photosensitivity and pregnancy concerns explain it.

Common effects. The most frequent reactions are irritation, burning, redness, peeling, dryness and itching. Using more than directed will not work faster or better, and larger amounts increase irritation. The label advises avoiding UV and sunlamps and using SPF 15 or higher plus protective clothing. Use caution with eczema, avoid it on sunburn, and avoid other photosensitizing drugs such as tetracyclines, thiazides and fluoroquinolones.

Who should not use it. The older Renova label says don't use it if you are pregnant, trying, or at high risk of pregnancy. Newer labels say observational studies haven't established a drug-associated risk of major birth defects or miscarriage but can't rule out risk, and they note case reports of birth defects resembling those from oral retinoids. Practically, the guidance is avoidance; talk to your doctor.

There are no human-milk data, so the label advises the smallest area for the shortest time, kept off the nipple and areola. Safety under age 12 isn't established.

Cancer. In a Veterans Affairs trial, 1,131 high-risk veterans (mean age 71) used 0.1% tretinoin or vehicle for 1.5 to 5.5 years. There was no effect on basal cell carcinoma (53% vs 54% at five years) or squamous cell carcinoma (28% vs 31%). So no cancer increase, and no protection either.

One wrinkle: a commentary on that trial called a higher mortality in the tretinoin group "inexplicable, surprising, but significant." We didn't see follow-up analyses, and those participants were older, high-risk and applying strong cream twice daily, so they aren't a model for cosmetic use.

Mouse studies at high doses showed tumors, but the labels say the significance is unclear because doses exceeded the maximum tolerated. In hairless mice, tretinoin plus UV may enhance UV tumor formation. The Renova label also notes atypical changes in melanocytes and keratinocytes and more dermal elastosis in some people using 0.05% for over 48 weeks, with unknown significance.

Hair loss: we found no evidence that tretinoin causes it, and nothing on eyebrows or lashes.

Are tretinoin's effects permanent?

No. The Renova label says patients "may lose some of the mitigating effects" after 12 weeks off the drug. For acne, the AAD describes retinoids as the way to maintain clearance, which implies ongoing use.

Long-term follow-up data are thin, so treat tretinoin as maintenance, not a one-time fix.