Azelaic acid: the monograph

Azelaic acid is a nine-carbon dicarboxylic acid (nonanedioic acid, C9H16O4, molecular weight 188.22, CAS 123-99-9), a white crystalline solid that is poorly soluble in cold water ^49,38. It is not an exotic molecule: it is a normal dietary constituent of whole grain cereals and animal products, the body can form it endogenously, and normal plasma levels run 20 to 80 ng/mL depending on diet ⁴⁰. The yeast Malassezia, which lives on everyone's skin, produces it naturally ³⁰.

In dermatology it exists in three parallel worlds. It is an FDA-approved prescription drug at 15% (Finacea gel and foam, for rosacea) and 20% (Azelex cream, for acne) ^41,42. It is a cosmetic ingredient in over-the-counter serums, commonly around 10%, which are not drugs and get no FDA review ⁴⁷. And it is a social-media staple credited with everything from fungal acne treatment to DHT blocking, claims this monograph grades against what was actually studied.

The honest summary: at prescription strengths, azelaic acid is one of the best-evidenced topical drugs in dermatology, with high-certainty evidence in rosacea and solid comparative trials in acne and melasma ^6,7. At the strength most people actually buy, the direct evidence is close to zero ⁵². This page keeps those two facts separate, which is the single most useful thing anyone can tell you about azelaic acid.

Sourced fact chips render here: chemistry, FDA status, absorption, half-life, tolerability and this site's own content counts. Every chip carries its claim id and resolves to the numbered references ^40,38.

Azelaic acid is FDA-approved as prescription topical drugs (Finacea for rosacea; Azelex). Over-the-counter azelaic-acid serums are cosmetics, not FDA-approved drugs.

The approved products: Finacea 15% gel (NDA 021470) and Finacea Foam 15% (NDA 207071) for inflammatory papules and pustules of mild to moderate rosacea, Azelex 20% cream (NDA 020428) for mild-to-moderate inflammatory acne vulgaris, and multiple approved generic 15% gels (ANDAs 204637, 208011, 208724, 210549) ^41,43,42,44. All are prescription-only in the United States.

Below the prescription tier, the rules change completely. Cosmetic products and ingredients do not need FDA premarket approval, with the exception of color additives ⁴⁷. A 10% azelaic acid serum is legally a cosmetic: nobody reviewed its safety, effectiveness or manufacturing before it went on sale. There is also no lawful over-the-counter DRUG route for azelaic acid: FDA's OTC acne monograph permits benzoyl peroxide, salicylic acid, sulfur and sulfur-resorcinol combinations, and azelaic acid is not on that list ⁴⁶. Products in FDA's label database registering 10 to 14% azelaic acid acne creams under monograph codes are self-submitted listings, not approvals ⁴⁵.

Compounded azelaic acid formulations sold by pharmacies and telehealth services sit in a third category: compounded drugs are not FDA-approved, and FDA does not verify their safety, effectiveness or quality before marketing ⁴⁸.

Rosacea: the 15% gel and foam are indicated for topical treatment of the inflammatory papules and pustules of mild to moderate rosacea ^38,39. The label is specific about what that excludes: although some erythema reduction occurred in patients who had papules and pustules, efficacy for treating erythema in the absence of papules and pustules has not been evaluated ³⁸. If your rosacea is redness and flushing without bumps, the approved evidence does not cover you.

Acne: Azelex 20% cream is indicated for topical treatment of mild-to-moderate inflammatory acne vulgaris ⁴⁰.

Everything else on this page, melasma and post-inflammatory hyperpigmentation included, is off-label or outside the drug system entirely. Off-label does not mean unstudied: the melasma trials below are real randomized comparisons. It means FDA never reviewed the use, so the claims on this page are graded by the studies themselves, not by an approval.

Here is the fact the skincare market blurs. Of the 17 interventional human studies in our evidence table, 16 tested the prescription strengths of 15% or 20% ^1,16. Exactly one tested the 10% strength the cosmetic market actually sells, and it deserves description: an 8-week, randomized, vehicle-controlled split-face study with 23 completers, testing a COMBINATION of 0.1% retinal plus 10% azelaic acid, measuring skin texture, surface irregularities, redness and pores by imaging ³⁷. It reported real improvements on those cosmetic endpoints. It is not an acne trial, not a rosacea trial, not a melasma trial, and the azelaic acid contribution cannot be separated from the retinal.

We searched for randomized trials of 10% azelaic acid serums used alone for acne, rosacea or melasma and found none ⁵². That does not prove 10% serums do nothing; concentration-response is plausible and untested territory. It proves that when a serum brand cites azelaic acid's clinical evidence, it is citing trials of a prescription product at 1.5 to 2 times the concentration in a different vehicle. The evidence-borrowing goes one direction: from the pharmacy shelf down.

This is why the concentration column exists in our study table, and why the explorer tool on the dosing pillar lets you filter every study by the strength it actually used ³⁷.

Start with the regulator's candor: the rosacea gel label states the mechanisms by which azelaic acid interferes with rosacea are unknown, and the Azelex label calls its own in vitro data of unknown clinical significance ^38,40. Everything below is mechanism evidence, not proof that a given mechanism drives the clinical results.

Antimicrobial (in vitro, plus human follicular data): azelaic acid is active against Cutibacterium (Propionibacterium) acnes and Staphylococcus epidermidis, plausibly by disrupting bacterial intracellular pH; kill is rapid in acid conditions and weak at neutral pH ^40,25,26. In vivo, a single application of 20% cream produced follicular concentrations in the in vitro inhibitory range ³¹, and acne patients on 20% cream showed 2,500-fold reductions in follicular Micrococcaceae and 44-fold reductions in Propionibacterium species ¹⁰.

Anti-keratinizing (human biopsy): treated skin shows thinner stratum corneum, fewer and smaller keratohyalin granules, and reduced filaggrin, consistent with reduced microcomedo formation ⁴⁰.

Tyrosinase inhibition (in vitro): the origin story. Dicarboxylic acids isolated from Malassezia (Pityrosporum) cultures in 1978 competitively inhibit tyrosinase, the rate-limiting enzyme of melanin synthesis, which is why azelaic acid ended up in pigment disorders at all ²³.

Anti-inflammatory (in vitro): at 20 mM, a concentration the authors state is achievable after 15% topical gel, azelaic acid suppresses UVB-induced IL-1beta, IL-6 and TNF-alpha in human keratinocytes through PPARgamma activation ²⁷.

5 alpha-reductase inhibition (in vitro only): azelaic acid potently inhibits the enzyme in skin homogenates, complete at 3 mmol/l ²⁴. The authors themselves wrote that clinical use depends on in vivo confirmation. Thirty-seven years later, that confirmation does not exist; see the hair-loss question in the FAQ.

No effect on sebum: measured sebum excretion rate did not change in acne patients, so azelaic acid is not an oil-control drug ¹⁰.

All three products follow the same skeleton: wash with a very mild soap or soapless cleanser, pat dry, apply a thin layer twice daily (morning and evening), wash your hands afterward, and let it dry before cosmetics go on top ^38,40. The gel and foam labels add: reassess the diagnosis if there is no improvement after 12 weeks ^38,39. The foam is applied to the entire facial area (cheeks, chin, forehead and nose), shaken well first, using the smallest amount that covers with a thin layer ³⁹.

How much is a thin layer? Two published anchors exist. The foam label's pharmacokinetic study used a mean of 0.94 grams of foam (141 mg azelaic acid) per whole-face application ³⁹. And dermatology's standard measure, the fingertip unit (about 0.5 g, the ribbon from a 5 mm nozzle along the last finger joint), allocates 2.5 FTU for face and neck, with one FTU covering about 286 square centimeters ^32,33. For a face-only application of a 15% gel, that arithmetic lands close to the same 1 gram per application the foam study measured. The application-quantity tool on this site does that arithmetic against your tube size, with every constant cited.

What to avoid while using it, per the labels: alcoholic cleansers, tinctures and astringents, abrasives and peeling agents, and occlusive dressings ³⁸. Keep it out of eyes and mucous membranes ³⁸. The foam is flammable: no fire, flame or smoking during and right after application ³⁹.

Our graded table holds 29 rows: 24 human and 5 in vitro, each with species, design, n, concentration, dose, duration, outcome and effect size ^6,22. The shape of the evidence differs by condition, and the differences are the story: rosacea has the regulator-grade base (two pivotal programs plus Cochrane-rated high-quality evidence), acne has a deep comparative record showing equivalence with standard topicals, melasma has genuine head-to-head wins and a mid-field rank in modern network meta-analysis, and the viral claims (fungal acne, DHT) have laboratory data and no trials.

Every number in this section resolves through the claims ledger to a fetched source, and the citation manifest ships with the page so the trail is machine-checkable ⁶.

The pivotal base: two identical vehicle-controlled phase III trials (664 subjects, 12 weeks) found mean inflammatory lesion reductions of 57.9% versus 39.9% and 50.0% versus 38.2%, with investigator global success in 61% versus 40% and 61% versus 48% ^38,1. Erythema in patients with papules and pustules improved more than vehicle too (44% versus 29%, 46% versus 28%), with no serious treatment-related adverse events ¹. The foam repeated the pattern in 1362 subjects: IGA success 32.1% versus 23.4% and 43.4% versus 32.5% ^39,4.

Against the standard comparator, 15% azelaic acid gel beat metronidazole 0.75% gel in a 251-patient randomized trial: 72.7% versus 55.8% lesion reduction, more erythema improvement, and continued gains after metronidazole plateaued at week 8 ². In a 40-patient split-face trial, 20% cream equaled metronidazole cream on lesions with better global-improvement ratings ³. The price of that efficacy is tolerability: 19.4% adverse reactions versus 7.1% with metronidazole in the active-controlled trial ³⁸.

The syntheses agree with unusual force. Cochrane 2015 (106 studies, 13,631 participants) rated azelaic acid's superiority over placebo as HIGH quality evidence, participant-assessed RR 1.46 ⁶. The 2019 GRADE update (152 studies, 20,944 participants) kept azelaic acid and ivermectin as the two high-certainty topicals for papules and pustules ⁷. A 2025 network meta-analysis ranked azelaic acid 20% first for IGA improvement (OR 8.54) while flagging azelaic acid 15% as the only topical with a significant adverse-event odds ratio (1.95) ⁸. Best in class and most stinging in class, in the same dataset: that trade-off is the honest headline.

Scope limit, straight from the label: none of this covers erythema without papules and pustules ³⁸. If your rosacea is flushing and background redness only, this drug's approved evidence does not speak to you.

The 20% cream's controlled program: better than vehicle in moderate inflammatory acne (92 patients, 3 months), equal to tretinoin 0.05% in comedonal acne (289 patients, 6 months) with fewer local side effects ⁹, and only just short of oral tetracycline over 6 months ¹¹. A separate placebo-controlled program documented the mechanism in patients: inflamed lesions fell within a month, follicular bacteria fell up to 2,500-fold, and sebum did not change ¹⁰.

The 15% gel matched both modern standards in two randomized blinded trials: median inflamed-lesion reductions of 70% versus benzoyl peroxide 5% (351 patients) and 71% versus clindamycin 1% (229 patients) over 4 months, with irritation between the two comparators (less than BPO, more than clindamycin) ¹². A vehicle-controlled trial in 60 patients showed a three-fold greater severity reduction than placebo ¹³.

For adult female acne, 15% gel was non-inferior to adapalene 0.1% over 9 months with significantly less dryness and scaling, and continuing past 3 months held results better than stopping ¹⁴. The European program overview adds two clinically important negatives: no teratogenicity signal and no induction of P. acnes resistance, which is why azelaic acid keeps its place as antibiotic-sparing therapy ¹⁵.

The head-to-head record against hydroquinone is genuinely strong and genuinely old. In the largest trial (329 women, 24 weeks, double-blind, both arms with sunscreen), 20% azelaic acid produced good or excellent results in 65% with NO significant differences versus hydroquinone 4% on any efficacy measure, and no allergic sensitization or ochronosis ¹⁶. Against the weaker 2% hydroquinone it was clearly superior: 73% versus 19% good-to-excellent in 155 patients ¹⁷. A small 2-month open trial even favored azelaic acid over hydroquinone 4% on MASI, though its authors flagged the open design and called for larger studies ¹⁸.

Now the ranking honesty. A 2021 network meta-analysis of 59 randomized trials put azelaic acid 8th of 14 melasma therapies, behind lasers, triple combination cream, topical vitamin C and oral tranexamic acid; combination therapy beat monotherapy in 87% of studies comparing them ²⁰. A 2019 systematic review gave its strong recommendations to cysteamine, triple combination therapy and tranexamic acid, not azelaic acid ²¹. Equal to hydroquinone 4% in its best trial, mid-field in the modern league table: both are true, and a page that tells you only one of them is selling something.

Post-inflammatory hyperpigmentation, the dark marks acne leaves in brown and black skin, has exactly one direct azelaic acid study: a 16-week open-label baseline-controlled pilot of 15% gel that reduced both acne and PIH ¹⁹. Promising, uncontrolled, and thin: that is the entire direct PIH evidence base. The tyrosinase-inhibition rationale is real but preclinical ²³.

DHT and hair loss. The claim that azelaic acid blocks DHT traces to one 1988 in vitro study: potent 5 alpha-reductase inhibition in skin homogenates, whose authors conditioned everything on in vivo confirmation ²⁴. Our search for azelaic acid trials in androgenetic alopecia returns no randomized monotherapy trial: reviews, a case report, a minoxidil combination study, and one pilot in a different disease ⁵⁰. That pilot randomized 31 people with patchy alopecia AREATA (autoimmune, not androgenetic) to 20% azelaic acid or anthralin and found similar regrowth with no vehicle arm ³⁶. If you are buying azelaic acid to protect your hairline, you are betting on a 37-year-old test-tube result.

Fungal acne. Malassezia folliculitis trials of azelaic acid: none found. Our searches return in vitro work, reviews and case reports, not treatment trials ⁵¹. The irony is that the Malassezia connection is real, just backwards: the yeast PRODUCES azelaic acid, and the molecule's entire dermatologic career began when researchers isolated tyrosinase-inhibiting dicarboxylic acids from Malassezia cultures in 1978 ^30,23. Plausible pharmacology, zero clinical trials: that is the grade.

Skin aging. The 2023 systematic review looked for eligible randomized trials of azelaic acid for skin aging and found none ²².

For most users, yes, within the label's boundaries, and the pharmacokinetics explain why: about 4% systemic absorption, plasma levels that stay inside the normal dietary range even with twice-daily prescription gel, and a molecule your body already makes and eats ^40,38. The 15% gel lists NO contraindications; Azelex is contraindicated only in component hypersensitivity ^38,40.

Safe does not mean comfortable. Roughly one user in three of the 15% gel gets burning, stinging or tingling, and the label sections below quantify every reaction the trials and postmarketing reports documented ³⁸. Pediatric safety is not established for any of the products ^38,40.

The 15% gel's numbers, from 788 trial subjects: burning, stinging or tingling 29%, itching 11%, scaling or dry skin 8%, redness or irritation 4% ³⁸. Most reactions were mild (16% of subjects) or moderate (9%); severe burning affected 4% ³⁸. In the active-controlled trial the gel caused more overall adverse reactions than metronidazole (19.4% versus 7.1%) ³⁸.

The foam runs milder in its trials: application site pain 6.2% versus 1.5% for vehicle, itching 2.5%, dryness and redness under 1% ³⁹. Azelex reports its most common reactions (pruritus, burning, stinging, tingling) in only 1 to 5% of patients, attributes early stinging partly to the molecule's low pH, and notes it commonly subsides with continued use ⁴⁰.

Context for the sting: irritation clusters in the first few weeks and usually settles ^38,40. In dermal safety studies the gel was more irritating than its own vehicle, and no phototoxicity or photoallergy was seen ³⁸.

Darker skin: isolated cases of hypopigmentation (light spots) have been reported, and the labels say plainly that azelaic acid has not been well studied in patients with dark complexions; monitor for early pigment changes and report them ^38,40. Weigh that against the PIH data above: the population with the most to gain is the one with the least dedicated safety study.

Asthma: worsening of asthma has been reported with azelaic acid formulations, prominently enough that the gel label gives it a dedicated warning ³⁸. Postmarketing reports also include hypersensitivity reactions: angioedema, eye and facial swelling, dyspnea, urticaria, wheezing ³⁸.

Eyes and mucous membranes: keep it away from them; accidental eye exposure has caused iridocyclitis. Rinse thoroughly with water on contact and get care if irritation persists ³⁸.

Rarities on the record: vitiligo depigmentation, small depigmented spots, hypertrichosis, keratosis pilaris-type reddening, and flare of cold sores ³⁸. The foam adds a physical hazard: it is flammable and pressurized ³⁹.

Azelaic acid is one of the few actives dermatology reviews actually recommend in pregnancy: a peer-reviewed review of acne treatment in pregnancy names topical azelaic acid or benzoyl peroxide as baseline therapy ³⁵. Here is what the primary documents say underneath that recommendation, because the two labels frame it differently.

The modern gel label reasons from exposure: azelaic acid is minimally absorbed, maternal use is not expected to result in fetal exposure, and the animal embryofetal toxicity seen orally occurred only at maternally toxic doses 19 to 162 times the human dose, without malformations ³⁸. The older Azelex label is more conservative with the same data: no adequate and well-controlled studies in pregnant women, so use only if clearly needed ⁴⁰. Both are honest; neither is a human trial, because there are none.

Breastfeeding: azelaic acid is naturally present in human milk already, and with under 4% absorption the labels expect no meaningful change in milk levels, while Azelex still advises caution ^38,40. For contrast, the melasma alternative hydroquinone is the one a pregnancy-safety review singles out for relatively high systemic absorption ³⁴. Decisions here belong with your clinician; this page's job is to show you what the documents actually say.

Purging, the idea that a treatment makes skin worse before better, is not a measured endpoint in any azelaic acid trial we cite. What the trials and labels do document in the early weeks is sensory irritation: burning, stinging and tingling that clusters at the start and usually settles ^38,40. If azelaic acid has a retinoid-style comedonal purge, no controlled study has quantified it, and we will not pretend otherwise.

Timelines, from the documents rather than from influencers: acne improvement in the majority of patients within 4 weeks per the Azelex label, with trials running 3 to 6 months ^40,9. Rosacea trials ran 12 to 15 weeks, and the label says to reassess the diagnosis if 12 weeks brings no improvement ³⁸. Melasma head-to-heads ran 24 weeks, with one small open trial showing separation from hydroquinone by month 2 ^16,18. Adult-acne maintenance data suggest staying on it: stopping at 3 months lost ground versus continuing to 9 ¹⁴.

The labels list no drug-drug interactions; none of the three has a drug interactions section at all ^38,40,39. That is an absence of studied interactions, not a certificate that combinations are safe.

What the labels DO restrict is product pairings: avoid alcoholic cleansers, tinctures and astringents, abrasives and peeling agents, and occlusive dressings while using azelaic acid ³⁸. In routine terms, that puts high-strength exfoliating acid toners and physical scrubs on the caution list during treatment, on label language rather than interaction studies. Popular pairing questions (niacinamide, vitamin C, retinoids) have no data in the labels or in our searches: no compatibility studies exist either way, and the schedule tool on this site says exactly that instead of inventing rules.

About 4% of a topical dose is absorbed systemically; 3 to 5% of the applied dose penetrates the stratum corneum and up to 10% reaches viable epidermis and dermis ⁴⁰. The topical half-life is about 12 hours against 45 minutes orally, meaning skin entry is the bottleneck, not clearance ⁴⁰.

Because azelaic acid is a dietary constituent with endogenous plasma levels of 20 to 80 ng/mL, systemic exposure from treatment disappears into the noise: twice-daily 15% gel produced plasma levels of 42 to 63.1 ng/mL, inside the range measured in vehicle-treated subjects ^40,38. The foam PK study (21 patients, whole-face dosing at a measured 0.94 g per application) reached steady state by day 5 with mean peak levels of 51.8 ng/mL, again overlapping baseline ³⁹.

The clinically interesting delivery is local: follicular casts sampled after a single 20% cream application held azelaic acid at concentrations comparable to the levels that inhibit acne bacteria in vitro ³¹. Elimination is mainly unchanged in urine ⁴⁰.

Gel: 50 g tube, stored at 25 C with excursions 15 to 30 C. Azelex cream: 30 g or 50 g tube, protect from freezing, store on its side at 15 to 30 C. Foam: pressurized 50 g can; flammable; no fire, flame or smoking during and right after application; do not puncture or incinerate; keep below 120 F (49 C) ^38,40,39.

Rendered from the faq[] array; answers ship with inline citation anchors and 320-character answer summaries for the machine layer.

All 52 citations render as a numbered list in n order; inline cite tokens resolve to these numbers. Fetch dates and verification method for every source live in the claims ledger; the structured citation manifest ships alongside this page ⁶.