Tretinoin

What it is

Tretinoin is all-trans retinoic acid, a naturally occurring metabolite of vitamin A (retinol) that acts on retinoic acid receptors inside cells. Two completely different medicines share this name. The one most people hold is a topical cream or gel, 0.01% to 0.1%, prescribed for acne and, in the Renova 0.02% formulation, for fine facial wrinkles. The other is a 10 mg oral capsule used in hospital to induce remission in acute promyelocytic leukaemia; it carries a boxed warning for embryo-fetal toxicity and differentiation syndrome. They are not interchangeable, and the evidence, doses and risks are entirely separate.

What the research says

For topical use the evidence is good that tretinoin clears acne: a network meta-analysis of 35 trials in 33,472 people found every single topical agent except tazarotene beat placebo for lesion counts, with tretinoin among them, although combination products generally did better. For skin ageing, vehicle-controlled trials from 1988 and 1995 showed real improvement in photoaged facial skin, but the Renova label is unusually blunt that the product does not eliminate wrinkles or reverse photoaging, that only two of five trials demonstrated efficacy even for fine wrinkling, and that in darkly pigmented participants vehicle did better than tretinoin. A large randomised trial found high-dose topical tretinoin did not prevent skin cancer. Skin irritation is near-universal. Oral tretinoin in leukaemia is highly effective but roughly a quarter of patients develop differentiation syndrome, which can kill.

Evidence grade: Well established.

How it works

Drug class: Retinoid (retinoic acid receptor agonist); topical comedolytic for acne and photoageing, and oral differentiating agent for acute promyelocytic leukaemia

Tretinoin is a form of vitamin A that binds retinoic acid receptors in the cell nucleus and changes which genes are switched on, altering how skin cells grow, mature and stick together. In acne it loosens the cells lining the follicle so plugs form less readily and existing plugs are pushed out. In photoaged skin it is not settled whether the benefit comes from receptor activation, from the irritation the drug causes, or both. In acute promyelocytic leukaemia the oral form forces the leukaemic promyelocytes to mature instead of multiplying, after which normal blood cells repopulate the marrow; the label states the exact mechanism in leukaemia is unknown. (Source 1)

Boxed warning

Tretinoin capsules can cause embryo-fetal loss and malformations when administered to a pregnant woman. Advise pregnant women of the potential risk to a fetus. Females of reproductive potential must have a negative pregnancy test before initiating tretinoin capsules. Advise females of reproductive potential to use two effective methods of contraception during treatment with tretinoin capsules and for 1 month after the last dose. Advise males with female partners of reproductive potential to use effective contraception during treatment with tretinoin capsules and for 1 week after the last dose [see Warnings and Precautions (5.1), Use in Specific Populations (8.1, 8.3)].

(Source 2)

What it is used for

• A 2025 network meta-analysis of 35 randomised trials found all single topical agents except tazarotene beat placebo on lesion counts, tretinoin included. Fixed combinations of tretinoin with clindamycin were among the most effective for non-inflammatory lesions, though adding clindamycin did not beat tretinoin alone for those lesions. Evidence: established. (Source 3)
• Only two of the five registration trials demonstrated efficacy, and only for fine wrinkling; none of the five showed a consistent effect on coarse wrinkling, mottled pigmentation, roughness or laxity. The label states the product does not eliminate wrinkles or reverse photoaging. Earlier academic trials of stronger concentrations did show improvement in photoaged skin. Evidence: limited. (Source 4)
• The 2010 Cochrane review of melasma treatments found two trials comparing tretinoin with placebo: participants rated their melasma significantly improved in one but not the other, while objective measures improved in both. Cochrane judged the melasma literature generally poor. Evidence: disputed. (Source 5)
• The Veterans Affairs Topical Tretinoin Chemoprevention Trial randomised 1,131 high-risk patients to 0.1% tretinoin or vehicle for up to 5.5 years and found no reduction in basal cell or squamous cell carcinoma, and no difference in actinic keratosis counts. The only quality-of-life difference favoured the control group. Evidence: not-supported. (Source 6)
• Oral tretinoin is standard induction therapy for APL with the t(15;17) translocation or PML/RARα fusion. In a phase 3 randomised trial, tretinoin with arsenic trioxide gave 100% complete remission and 97% two-year event-free survival versus 86% for tretinoin plus chemotherapy. This use carries a boxed warning. Note that the label's own efficacy section describes the Memorial Sloan-Kettering study as uncontrolled in its text while captioning the same results as a 'Controlled Clinical Trial' in its table; the label is inconsistent on that point. Evidence: established. (Source 7)

Interactions

• Vitamin A supplements (retinol, retinyl palmitate, cod liver oil) (label): Oral tretinoin is itself a vitamin A derivative, so taking vitamin A on top adds to the same toxicity, which can include headache, dry skin and mucous membranes, raised pressure inside the skull and liver enzyme rises. The label instructs avoiding the combination with oral tretinoin capsules. (Source 8)
• Strong and moderate CYP3A inhibitors and strong CYP3A inducers (label): Oral tretinoin is cleared by CYP3A4, so inhibitors raise its blood levels and the chance of adverse reactions, while inducers lower them and may make it stop working. The label advises avoiding both where possible. (Source 9)
• Anti-fibrinolytic agents (tranexamic acid, aminocaproic acid, aprotinin) (case reports): Fatal clots have been reported in people given oral tretinoin, and the label states that adding an anti-fibrinolytic may increase that risk; it advises avoiding the combination. (Source 10)
• Food (oral capsules) (label): The oral capsules are to be taken with a meal. The label is explicit that the food effect has not actually been measured for this product, and rests on the class observation that food increases retinoid absorption. Limit: A caution about the blood-level figures in this same label section, kept as the label prints them: it gives the peak concentration after one week as '138 +/- 139 ng/mL', where the stated standard deviation is larger than the mean, which cannot be right for a measurement that cannot go below zero, and it gives the first-dose figure as '394 +/- 89' with no unit attached. These are the label's own figures and have not been repaired here; do not rely on them as precise. (Source 11)
• Skin products containing high concentrations of alcohol, plus astringents, spices and lime (label): This is about alcohol applied to the skin, not alcohol drunk. The topical label says products with high alcohol concentrations, astringents, abrasive or medicated soaps and strongly drying cosmetics can interact with tretinoin, meaning more irritation; it advises letting the skin settle before starting tretinoin. (Source 12)
• Topical sulfur, resorcinol or salicylic acid (label): These keratolytics stack with tretinoin's own peeling effect, so the label singles them out for particular caution when used alongside topical tretinoin. (Source 12)
• Photosensitising drugs (thiazides, tetracyclines, fluoroquinolones, phenothiazines, sulfonamides) (label): Tretinoin already makes skin more sun-sensitive. The Renova label says the cream should not be used by someone also taking a known photosensitiser because of the possibility of augmented phototoxicity. (Source 13)

Stopping it

• For topical tretinoin on acne, the adverse effects described in the label resolved once the drug was stopped, and the label's own instruction when irritation is bad is to apply it less often, pause, or stop. (Source 14)
• For the wrinkle indication the benefit is not durable: the Renova label records that some of the mitigating effect is lost about 12 weeks after the cream is dropped from the skin-care routine. (Source 15)
• Oral tretinoin for leukaemia is a time-limited course, not a long-term drug: the label directs stopping 30 days after complete remission or after 90 days of treatment, whichever comes first, and stopping altogether if the t(15;17) translocation or PML/RARα fusion is not found. (Source 16)

What goes wrong

Almost everyone using Renova 0.02% on the face got local skin reactions, and in 32% the irritation was severe, forced a pause, or needed a topical steroid. (Source 17)

• Official position, Certainty not rated.
• Size: 339 patients who applied Renova 0.02% to the face in double-blind vehicle-controlled studies.
• Who: Adults treated for fine facial wrinkles.
• How long: Up to 24 weeks.
• Result: Almost all patients reported peeling, dry skin, burning, stinging, erythema or pruritus. In 32% irritation was severe, led to temporary discontinuation, or led to a mild topical corticosteroid. About 7% on tretinoin versus under 1% of controls needed short-term topical corticosteroids; about 4% discontinued.
• Funding: Manufacturer's FDA-approved label (DailyMed version 14, effective 24 February 2026)

In 32% of all study patients, skin irritation was reported that was severe, led to temporary discontinuation of RENOVA (tretinoin cream) 0.02%, or led to use of a mild topical corticosteroid. About 7% of patients using RENOVA (tretinoin cream) 0.02%, compared to less than 1% of the control patients, had sufficiently severe local irritation to warrant short-term use of mild topical corticosteroids to alleviate local irritation. About 4% of patients had to discontinue use of RENOVA (tretinoin cream) 0.02% because of adverse reactions.

About a quarter of patients given oral tretinoin for acute promyelocytic leukaemia developed differentiation syndrome, half of them the severe form, and severe cases carried higher mortality. (Source 18)

• Cohort study, Moderate certainty.
• Size: 739 APL patients across two consecutive PETHEMA trials (LPA96 and LPA99)
• Who: Patients with acute promyelocytic leukaemia given ATRA plus idarubicin induction.
• How long: Induction therapy; bimodal onset in the first and third weeks.
• Result: 183 of 739 (24.8%) experienced differentiation syndrome: 93 severe (12.6%) and 90 moderate (12.2%). Severe but not moderate DS was associated with increased mortality. A white cell count above 5 x 10(9)/L and abnormal creatinine predicted severe DS.
• Funding: PETHEMA cooperative group trials; funding not stated in the abstract.

Overall, 183 patients (24.8%) experienced DS, 93 with a severe form (12.6%) and 90 with a moderate form (12.2%). Severe but not moderate DS was associated with an increase in mortality.

Oral tretinoin caused adverse reactions in almost every patient, including headache in 86%, fever in 83%, bone pain in 77% and leukocytosis in 49%, with liver enzyme rises in 50 to 60%. (Source 19)

• Official position, Certainty not rated.
• Size: Patients with APL given 22.5 mg/m2 orally twice daily in the trials supporting approval.
• Who: Adults and children aged 1 and over with acute promyelocytic leukaemia.
• How long: Up to 90 days of induction.
• Result: Adverse reactions at 30% or more: headache, fever, skin and mucous membrane dryness, bone pain, malaise, shivering, upper respiratory tract disorders, dyspnoea, haemorrhage, infections, nausea and vomiting, rash, peripheral oedema, leukocytosis, pain, gastrointestinal haemorrhage, chest discomfort, abdominal pain. Elevated liver function tests in 50% to 60%.
• Funding: Manufacturer's FDA-approved label (DailyMed version 20, effective 20 August 2026)

The most common adverse reactions (≥30%) were headache, fever, skin/mucous membrane dryness, bone pain, malaise, shivering, upper respiratory tract disorders, dyspnea, hemorrhage, infections, nausea/vomiting, rash, peripheral edema, leukocytosis, pain, gastrointestinal hemorrhage, chest discomfort, abdominal pain.

What the evidence supports

In a network meta-analysis of topical acne treatments, every single agent including tretinoin outperformed placebo for lesion counts except tazarotene. (Source 3)

• Meta-analysis, Moderate certainty.
• Size: 35 randomised clinical trials, 33,472 participants, nine topical agents.
• Who: People aged 12 and older with mild-to-moderate acne vulgaris.
• How long: At least 12 weeks of treatment per trial.
• Result: Largest reductions were with adapalene-benzoyl peroxide, clindamycin-benzoyl peroxide and clindamycin-tretinoin (non-inflammatory lesion ratio of means 1.76, 95% CI 1.46 to 2.12; 1.70, 1.44 to 2.02; and 1.87, 1.53 to 2.30 respectively versus placebo). All single agents beat placebo except tazarotene.
• Funding: not stated in the abstract.

All single agents outperformed placebo except tazarotene, which did not significantly outperform placebo for inflammatory and non-inflammatory lesion count reduction.

Both 0.1% and 0.025% tretinoin cream improved photoaging of the face compared with vehicle over 48 weeks, with no efficacy difference between the two strengths but more irritation at the higher one. (Source 20)

• Randomized trial, Low certainty.
• Size: 99 patients completed (0.1% n=32, 0.025% n=35, vehicle n=32)
• Who: Photoaged patients, double-blind, single centre.
• How long: 48 weeks, applied once daily.
• Result: Both concentrations gave statistically significant overall improvement versus vehicle; epidermal thickening 30% and 28% versus an 11% decrease with vehicle; vascularity up 100% and 89% versus a 9% decrease. Erythema and scaling were significantly greater with 0.1% than 0.025%.
• Funding: not stated in the abstract.

Both 0.1% and 0.025% tretinoin produced statistically significant overall improvement in photoaging of the face compared with vehicle; there were no clinically or statistically significant differences in efficacy between the two concentrations of tretinoin. After 48 weeks, 0.1% and 0.025% tretinoin produced similar statistically significant epidermal thickening (by 30% and 28%, respectively) compared with vehicle (11% decrease) and increased vascularity (by 100% and 89%, respectively) compared with vehicle (9% decrease).

In the original 1988 double-blind vehicle-controlled trial, 14 of 15 patients treated on the face improved while none of the vehicle-treated faces did. (Source 21)

• Randomized trial, Very low certainty.
• Size: 30 patients completed.
• Who: Adults with photoaging; within-person forearm comparison plus between-group face comparison.
• How long: 16 weeks.
• Result: All 30 patients improved on the tretinoin-treated forearm but not the vehicle forearm; 14 of 15 treated on the face improved versus none of the vehicle-treated faces, a statistically significant between-group difference. Side effects were limited to irritation of treated skin.
• Funding: not stated in the abstract.

Fourteen of the 15 patients who received tretinoin to the face had improvement in photoaging, whereas none of the vehicle-treated patients' faces improved, a statistically significant difference in response between the two groups.

In a noninferiority trial restricted to low-to-intermediate-risk acute promyelocytic leukaemia, oral tretinoin combined with arsenic trioxide gave complete remission in all evaluable patients and better two-year event-free survival than tretinoin plus chemotherapy. (Source 7)

• Randomized trial, High certainty.
• Size: 156 evaluable (77 ATRA plus arsenic trioxide, 79 ATRA plus chemotherapy)
• Who: Patients with low-to-intermediate-risk acute promyelocytic leukaemia (white-cell count 10x10^9/L or less)
• How long: Median follow-up 34.4 months.
• Result: Complete remission 100% (77/77) versus 95% (75/79), P=0.12. Two-year event-free survival 97% versus 86% (95% CI for the difference 2 to 22 percentage points; P<0.001 for non-inferiority, P=0.02 for superiority). Overall survival also better (P=0.02). Less haematological toxicity and fewer infections but more hepatic toxicity with arsenic trioxide.
• Funding: Associazione Italiana contro le Leucemie and others (non-commercial)

Limit of this finding: The trial was designed to show non-inferiority, not superiority, and enrolled only patients whose white-cell count was 10x10^9 per litre or below. The authors' own conclusion is the cautious one: ATRA plus arsenic trioxide is 'at least not inferior and may be superior' in that lower-risk group. These results do not describe high-risk APL, and the comparison is against another active regimen, not against placebo.

Two-year event-free survival rates were 97% in the ATRA-arsenic trioxide group and 86% in the ATRA-chemotherapy group (95% confidence interval for the difference, 2 to 22 percentage points; P<0.001 for noninferiority and P=0.02 for superiority of ATRA-arsenic trioxide).

What the evidence does not support

Adding clindamycin to tretinoin did not beat tretinoin alone for non-inflammatory lesions, and most single topical agents did not differ from each other. (Source 22)

• Meta-analysis, Moderate certainty.
• Size: 35 randomised clinical trials, 33,472 participants.
• Who: People aged 12 and older with mild-to-moderate acne vulgaris.
• How long: At least 12 weeks per trial.
• Result: Clindamycin-tretinoin versus tretinoin alone for non-inflammatory lesion reduction: ratio of means 1.13 (95% CI 0.94 to 1.36), not significant. No significant difference among most single agents on lesion count reduction.
• Funding: not stated in the abstract.

Limit of this finding: The review's own bottom line goes further than the quoted sentence: it says that for non-inflammatory acne the addition of clindamycin to topical regimens 'is unnecessary and should be avoided'. That is the authors' recommendation, not a measured outcome.

except for clindamycin-tretinoin and clindamycin-BPO, which did not significantly outperform tretinoin (RoM 1.13; 95% CI [0.94; 1.36]) and BPO (RoM = 1.15, 95% CI [0.98; 1.36]), respectively, for non-inflammatory lesion reduction. There was no significant difference amongst most single agents when evaluating lesion count reduction.

High-dose topical tretinoin 0.1% did not reduce basal cell or squamous cell skin cancer in a large randomised trial of high-risk patients. (Source 6)

• Randomized trial, High certainty.
• Size: 1,131 patients randomised.
• Who: Veterans at high risk of keratinocyte carcinoma.
• How long: 1.5 to 5.5 years of daily treatment.
• Result: Basal cell carcinoma at 5 years 53% on tretinoin versus 54% on vehicle (difference 1.0%, 95% CI -6.5 to 8.6%; P=0.3); invasive squamous cell carcinoma 28% versus 31% (difference 3.6%, 95% CI -3.1 to 10.3%; P=0.4). No difference in actinic keratosis counts. The only quality-of-life difference was worse symptoms in the tretinoin group at 12 months.
• Funding: Veterans Affairs cooperative trial (government-funded)

The effects were not significant (P=0.3 for BCC and P=0.4 for SCC). The proportions of the tretinoin and control groups who developed a BCC at 5 years were 53 and 54% and an invasive SCC at 5 years were 28 and 31%. These differences (95% confidence intervals) were: for BCC, 1.0% (-6.5, 8.6%); for SCC, 3.6% (-3.1, 10.3%).

The manufacturer's own label states that Renova 0.02% does not eliminate wrinkles, repair sun-damaged skin or reverse photoaging, and that many vehicle-group patients achieved the desired effect with skin care and sun avoidance alone. (Source 23)

• Official position, Certainty not rated.
• Size: 324 evaluable on Renova 0.02% and 332 on vehicle across five trials.
• Who: Adults with fine facial wrinkling using comprehensive skin care and sun avoidance.
• How long: 24 weeks.
• Result: Two of the five trials demonstrated efficacy for fine facial wrinkling. No two of the five demonstrated efficacy for coarse wrinkling, mottled hyperpigmentation, tactile roughness or laxity.
• Funding: Manufacturer's FDA-approved label (DailyMed version 14, effective 24 February 2026), a regulatory position rather than independent analysis.

Limit of this finding: The label says it also collected patients' own ratings of fine wrinkles at 24 weeks, but it prints that result only as a picture (figure1.jpg) and gives no numbers for it anywhere in the label text. So the self-rated result cannot be checked, and nothing here should be read as reporting what that chart shows. The numbers the label does print in words are the ones quoted above.

Two of the five trials provided adequate demonstration of efficacy for mitigation of fine facial wrinkling. No two of the five trials adequately demonstrated efficacy for mitigation of coarse wrinkling, mottled hyperpigmentation, tactile skin roughness, and laxity.

In the one registration trial done in darkly pigmented skin, fewer people improved on tretinoin than on vehicle. (Source 24)

• Randomized trial, Very low certainty.
• Size: 107 participants, single centre.
• Who: Darkly pigmented, mostly African-American adults with Fitzpatrick skin types IV-VI.
• How long: 24 weeks.
• Result: Minimal or mild improvement in fine facial wrinkling in 43% on vehicle plus comprehensive skin protection versus 29% on Renova 0.02% plus the same programme. The label notes this may reflect the small size of the study.
• Funding: Reported in the manufacturer's FDA-approved label (DailyMed version 14, effective 24 February 2026)

A single-center study (N=107) in darkly pigmented, mostly African-American, subjects with Fitzpatrick Skin Types IV-VI demonstrated minimal or mild improvement in fine facial wrinkling in 43% of patients using Vehicle + CSP* compared to 29% of subjects using RENOVA (tretinoin cream) 0.02% + CSP*. Although fewer darkly pigmented subjects improved with RENOVA (tretinoin cream) 0.02% than with vehicle, these findings may reflect the small size of this study.

A meta-analysis of first-trimester topical retinoid exposure did not detect an increase in major congenital malformations, but says plainly that it lacked the power to justify use in pregnancy. (Source 25)

• Meta-analysis, Low certainty.
• Size: 654 exposed pregnancies and 1,375 unexposed controls.
• Who: Pregnant women with inadvertent first-trimester exposure to topical retinoids.
• How long: Pregnancy outcomes.
• Result: Major congenital malformations OR 1.22 (95% CI 0.65 to 2.29); spontaneous abortion OR 1.02 (0.64 to 1.63); stillbirth OR 2.06 (0.43 to 9.86); elective termination OR 1.89 (0.52 to 6.80); low birthweight OR 1.01 (0.31 to 3.27); prematurity OR 0.69 (0.39 to 1.23). The authors state the statistical power is not adequate to justify use in pregnancy.
• Funding: not stated in the abstract.

This meta-analysis, including a total of 654 pregnant women who were exposed to topical retinoids, and 1375 unexposed control pregnant women, did not detect significant increases in rates of major congenital malformations [OR 1·22, 95% confidence interval (CI) 0·65-2·29], spontaneous abortions (OR 1·02, 95% CI 0·64-1·63), stillbirth (OR 2·06, 95% CI 0·43-9·86), elective termination of pregnancy (OR 1·89, 95% CI 0·52-6·80), low birthweight (OR 1·01, 95% CI, 0·31-3·27) or prematurity (OR 0·69, 95% CI 0·39-1·23). No significant heterogeneity was detected among the studies for the evaluated outcomes. The present meta-analysis ruled out a major increase in the rates of major congenital malformations, spontaneous abortions, low birthweight and prematurity. This result may be used primarily in reassuring women who were inadvertently exposed to topical retinoids during their pregnancy. However, the statistical power is not adequate to justify the use of topical retinoids during pregnancy.

Where the evidence is mixed

For melasma, tretinoin beat placebo on participants' own ratings in one trial but not in a second, in a Cochrane review that judged the whole melasma literature poor. (Source 5)

• Systematic review, Low certainty.
• Size: 20 studies, 2,125 participants across 23 treatments; two trials of tretinoin versus placebo.
• Who: People with epidermal, dermal or mixed melasma.
• How long: Varied by trial; pooling was not possible because of heterogeneity.
• Result: Participants rated melasma significantly improved with tretinoin versus placebo in one trial (RR 13, 95% CI 1.88 to 89.74) but not the other; objective measures improved in both. Adverse events were mostly mild and transient: skin irritation, itching, burning and stinging.
• Funding: Cochrane review, independent.

Limit of this finding: The review covered 20 studies and 2,125 people but could not combine their results: it states that 'Statistical pooling of the data was not possible due to the heterogeneity of treatments', so each figure comes from a single small trial. The tretinoin-versus-placebo risk ratio of 13 has a confidence interval running from 1.88 to 89.74, which is so wide that it tells you almost nothing about the size of any benefit.

In two studies where tretinoin was compared to placebo, participants rated their melasma as significantly improved in one (RR 13, 95% CI 1.88 to 89.74) but not the other.

Where the research disagrees

Whether topical tretinoin meaningfully improves sun-damaged skin, or mainly fine wrinkles under a sun-protection programme

• Griffiths and colleagues, 1995 Archives of Dermatology double-blind vehicle-controlled trial, 48-week double-blind vehicle-controlled randomised trial in 99 photoaged patients, with histology: Both 0.1% and 0.025% tretinoin produced statistically significant overall improvement in photoaging of the face compared with vehicle (Source 20)
• Bausch Health, in the FDA-approved Renova 0.02% label (DailyMed version 14, effective 24 February 2026), Regulatory position based on five vehicle-controlled trials, 324 treated and 332 vehicle participants, of which only two demonstrated efficacy and only for fine wrinkling: RENOVA (tretinoin cream) 0.02% DOES NOT ELIMINATE WRINKLES, REPAIR SUN-DAMAGED SKIN, REVERSE PHOTOAGING, or RESTORE MORE YOUTHFUL or YOUNGER SKIN. (Source 4)

How much

• Reference intake: There is no reference intake for a prescription drug; the strength and how often it is used are set by the prescriber. The topical label (DailyMed version 19, effective 31 December 2023) records as a regulatory position that the gel or cream is applied once a day at bedtime, using enough to cover the affected area lightly. The oral capsule label is a different product entirely. (Source 26)
• Upper limit: No upper limit in the nutritional sense exists. The marketed topical strengths are, as a label position, 0.01% and 0.025% gel and 0.025%, 0.05% and 0.1% cream; Renova is 0.02% cream. The oral capsule label states a dosage of 22.5 mg/m2 twice daily, stopped after 90 days or 30 days after remission, whichever is first. (Source 27)
• Studied: The 48-week photoaging trial applied 0.1% or 0.025% tretinoin cream, or vehicle, once daily to 99 patients. (Source 28)
• Studied: The Veterans Affairs chemoprevention trial gave 1,131 patients topical 0.1% tretinoin or matching vehicle for 1.5 to 5.5 years. (Source 6)
• Studied: The oral trials gave tretinoin capsules 22.5 mg/m2 twice daily for up to 90 days, or 30 days after complete remission. (Source 29)

A common belief, and what the research shows

The belief: Tretinoin is one drug, and the warnings you read about it, including the boxed warning, apply to the cream you put on your face.

What the research shows: They are two different medicines. The boxed warning belongs to the 10 mg oral capsule used for acute promyelocytic leukaemia, which states: “Differentiation Syndrome, which can be life-threatening or fatal, occurred in about 26% of patients with APL who received tretinoin capsules.” The topical gel and cream have no boxed warning at all. The pregnancy picture also differs: oral tretinoin is a potent teratogen, while a meta-analysis of 654 exposed pregnancies “did not detect significant increases in rates of major congenital malformations” after first-trimester topical exposure, though that analysis adds “the statistical power is not adequate to justify the use of topical retinoids during pregnancy”.

Questions and answers

What is it?

Tretinoin is all-trans retinoic acid, a vitamin A derivative. Chemically it is one molecule, C20H28O2, molecular weight 300.44. It reaches people in two very different forms: a topical gel or cream of 0.01% to 0.1% for acne, and 0.02% cream for fine facial wrinkles; and a 10 mg oral capsule used in hospital for acute promyelocytic leukaemia. The same molecule, entirely different uses and risks. (Source 27)

What does it do in the body?

Tretinoin is the body's own active form of vitamin A. It binds retinoic acid receptors in the cell nucleus and changes gene activity, which alters how skin cells grow, mature and stick together. In acne this loosens the cells lining the follicle so plugs form less easily and existing comedones are extruded. On photoaged skin it is still unsettled whether the benefit comes from receptor activation, from irritation, or both. (Source 30)

Is it good or bad for you?

It depends entirely on which product and which use. Topical tretinoin for acne is well supported: in a meta-analysis of 35 trials in 33,472 people all single agents including tretinoin beat placebo. For fine facial wrinkles the benefit is small and the label itself says the cream does not reverse photoaging. For preventing skin cancer it does not work: a 1,131-patient randomised trial found no reduction in basal or squamous cell carcinoma. Irritation is near-universal. Oral tretinoin saves lives in leukaemia but about a quarter of patients get differentiation syndrome, which can be fatal. (Source 6)

How do you get more of it?

Tretinoin itself is prescription-only in both its topical and oral forms. There is no supplement or food that delivers it as a drug. It is worth knowing that the body makes small amounts of tretinoin from dietary vitamin A: the label describes it as an endogenous retinoid metabolite of vitamin A, and applying a tretinoin product raises skin levels above those physiological concentrations. Topical absorption into the bloodstream is low, under 2% in one study of 0.05% cream, so a cream does not meaningfully raise whole-body retinoid levels. (Source 1)

If it is harmful, what reduces it?

For the topical products, the remedy for irritation is to use less, use it less often, pause, or stop; the label records that all adverse effects of tretinoin were reversible on stopping. For the oral capsule, a severe adverse effect is managed differently: at the first sign of differentiation syndrome the label directs immediate high-dose dexamethasone, 10 mg intravenously every 12 hours, with haemodynamic monitoring, and withholding the drug for moderate or severe cases. (Source 31)

Why might someone be low in it or missing it?

Nobody is deficient in tretinoin as a medicine. Reasons a person might not be using it include never having been prescribed it, having stopped because of irritation, which made about 4% of Renova trial participants discontinue, pregnancy or planned pregnancy, eczematous skin, where the label says severe irritation has been reported, or taking a photosensitising drug such as a thiazide or tetracycline, which the Renova label says should not be combined with the cream. For the oral capsule, the label directs stopping if the t(15;17) translocation or PML/RARα fusion is not found. (Source 17)

Which whole foods contain it or feed it?

No whole food contains tretinoin as such. Foods rich in vitamin A, liver, oily fish, eggs, dairy, and in provitamin A carotenoids, carrots, sweet potato, dark leafy greens, supply the retinol the body converts in small amounts to retinoic acid; this is a nutritional pathway, not a way of dosing the drug. Food matters in the other direction for the oral capsule, which is taken with a meal: the label states the food effect has not actually been measured, only that food increases retinoid absorption as a class. (Source 11)

What happens if you do not have it?

Not using topical tretinoin means acne that might have cleared does not, nothing more; it is a treatment, not a nutrient. For fine wrinkles the loss is small, and the label notes many people in the vehicle groups got the effect they wanted from sun avoidance and emollients alone. In leukaemia the stakes are different: oral tretinoin is part of the regimen that has made acute promyelocytic leukaemia curable, with 97% two-year event-free survival when combined with arsenic trioxide. (Source 4)

How can you test for it?

There is no blood test for topical tretinoin and none is needed; response is judged by looking at the skin, and the label notes benefit may take more than six weeks. For the oral capsule, monitoring is extensive: liver function tests at baseline and during treatment, because 50 to 60% of patients develop raised results, plus pregnancy testing before starting with a sensitivity of at least 50 mIU/mL, and confirmation of the t(15;17) translocation or PML/RARα fusion before continuing. White cell counts are watched because leukocytosis occurred in 49%. (Source 32)

References

1. Bausch Health US LLC / DailyMed (U.S. National Library of Medicine). RENOVA (tretinoin cream) 0.02% - FDA prescribing information (DailyMed SPL, version 14). 2026. Read the source
2. Par Health USA, LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN capsule - FDA prescribing information (DailyMed SPL, version 20). 2026. Read the source
3. Journal of the European Academy of Dermatology and Venereology : JEADV. Efficacy of topical treatments for mild-to-moderate acne: A systematic review and meta-analysis of randomized control trials.. 2025. PMID 38943431, DOI 10.1111/jdv.20154. Read the source
4. Bausch Health US LLC / DailyMed (U.S. National Library of Medicine). RENOVA (tretinoin cream) 0.02% - FDA prescribing information (DailyMed SPL, version 14). 2026. Read the source
5. The Cochrane database of systematic reviews. Interventions for melasma.. 2010. PMID 20614435, DOI 10.1002/14651858.CD003583.pub2. Read the source
6. The Journal of investigative dermatology. Tretinoin and the prevention of keratinocyte carcinoma (Basal and squamous cell carcinoma of the skin): a veterans affairs randomized chemoprevention trial.. 2012. PMID 22318383, DOI 10.1038/jid.2011.483. Read the source
7. The New England journal of medicine. Retinoic acid and arsenic trioxide for acute promyelocytic leukemia.. 2013. PMID 23841729, DOI 10.1056/NEJMoa1300874. Read the source
8. Par Health USA, LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN capsule - FDA prescribing information (DailyMed SPL, version 20). 2026. Read the source
9. Par Health USA, LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN capsule - FDA prescribing information (DailyMed SPL, version 20). 2026. Read the source
10. Par Health USA, LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN capsule - FDA prescribing information (DailyMed SPL, version 20). 2026. Read the source
11. Par Health USA, LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN capsule - FDA prescribing information (DailyMed SPL, version 20). 2026. Read the source
12. Padagis US LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN gel / TRETINOIN cream - FDA prescribing information (DailyMed SPL, version 19). 2023. Read the source
13. Bausch Health US LLC / DailyMed (U.S. National Library of Medicine). RENOVA (tretinoin cream) 0.02% - FDA prescribing information (DailyMed SPL, version 14). 2026. Read the source
14. Padagis US LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN gel / TRETINOIN cream - FDA prescribing information (DailyMed SPL, version 19). 2023. Read the source
15. Bausch Health US LLC / DailyMed (U.S. National Library of Medicine). RENOVA (tretinoin cream) 0.02% - FDA prescribing information (DailyMed SPL, version 14) - Clinical Trials section, the two paragraphs that follow the self-assessment figure. 2026. Read the source
16. Par Health USA, LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN capsule - FDA prescribing information (DailyMed SPL, version 20). 2026. Read the source
17. Bausch Health US LLC / DailyMed (U.S. National Library of Medicine). RENOVA (tretinoin cream) 0.02% - FDA prescribing information (DailyMed SPL, version 14). 2026. Read the source
18. Blood. Differentiation syndrome in patients with acute promyelocytic leukemia treated with all-trans retinoic acid and anthracycline chemotherapy: characteristics, outcome, and prognostic factors.. 2009. PMID 18945964, DOI 10.1182/blood-2008-07-168617. Read the source
19. Par Health USA, LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN capsule - FDA prescribing information (DailyMed SPL, version 20). 2026. Read the source
20. Archives of dermatology. Two concentrations of topical tretinoin (retinoic acid) cause similar improvement of photoaging but different degrees of irritation. A double-blind, vehicle-controlled comparison of 0.1% and 0.025% tretinoin creams.. 1995. PMID 7544967. Read the source
21. JAMA. Topical tretinoin improves photoaged skin. A double-blind vehicle-controlled study.. 1988. PMID 3336176. Read the source
22. Journal of the European Academy of Dermatology and Venereology : JEADV. Efficacy of topical treatments for mild-to-moderate acne: A systematic review and meta-analysis of randomized control trials.. 2025. PMID 38943431, DOI 10.1111/jdv.20154. Read the source
23. Bausch Health US LLC / DailyMed (U.S. National Library of Medicine). RENOVA (tretinoin cream) 0.02% - FDA prescribing information (DailyMed SPL, version 14). 2026. Read the source
24. Bausch Health US LLC / DailyMed (U.S. National Library of Medicine). RENOVA (tretinoin cream) 0.02% - FDA prescribing information (DailyMed SPL, version 14). 2026. Read the source
25. The British journal of dermatology. Pregnancy outcomes following first-trimester exposure to topical retinoids: a systematic review and meta-analysis.. 2015. PMID 26215715, DOI 10.1111/bjd.14053. Read the source
26. Padagis US LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN gel / TRETINOIN cream - FDA prescribing information (DailyMed SPL, version 19). 2023. Read the source
27. Padagis US LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN gel / TRETINOIN cream - FDA prescribing information (DailyMed SPL, version 19). 2023. Read the source
28. Archives of dermatology. Two concentrations of topical tretinoin (retinoic acid) cause similar improvement of photoaging but different degrees of irritation. A double-blind, vehicle-controlled comparison of 0.1% and 0.025% tretinoin creams.. 1995. PMID 7544967. Read the source
29. Par Health USA, LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN capsule - FDA prescribing information (DailyMed SPL, version 20). 2026. Read the source
30. Padagis US LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN gel / TRETINOIN cream - FDA prescribing information (DailyMed SPL, version 19). 2023. Read the source
31. Par Health USA, LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN capsule - FDA prescribing information (DailyMed SPL, version 20). 2026. Read the source
32. Par Health USA, LLC / DailyMed (U.S. National Library of Medicine). TRETINOIN capsule - FDA prescribing information (DailyMed SPL, version 20). 2026. Read the source