Azelaic acid evidence: rosacea, acne and melasma trials

Azelaic acid is a dicarboxylic acid used on the skin. The best evidence is for the pimples and pustules of rosacea, with moderate evidence for acne and smaller, older, often unblinded trials for melasma. For redness without pimples, and for post-inflammatory hyperpigmentation, direct evidence is thin.

Key points

• In the US, 15% gel is FDA-approved for the papules and pustules of mild to moderate rosacea, and 20% cream for mild to moderate inflammatory acne. Both are prescription products [1][2].
• Two 12-week vehicle-controlled trials (664 people) are the base for the rosacea label. The label says efficacy for erythema without papules and pustules “has not been evaluated” [1].
• A 2020 Cochrane-derived review (49 trials, 3,880 people) rated azelaic acid probably less effective than benzoyl peroxide for acne and about equal to tretinoin [5].
• Melasma trials are small or old; a 2023 meta-analysis pooled six trials (673 people) against hydroquinone [8].
• The tier-list line “clears vascular redness & blocks hyperactive melanocytes” is partly supported on redness and could not be verified as a mechanism [12].

What it is and what is sold

Azelaic acid occurs naturally in some grains and in human milk [1]. Pharmaceutical forms are a 15% gel (Finacea, also a 15% foam studied in trials) and a 20% cream (Azelex in the US, Skinoren in the UK) [1][2][3].

Regulatory status, as far as we checked:

Where	Product	Status found	Label date
US (FDA)	Finacea gel 15%	Indicated for inflammatory papules and pustules of mild to moderate rosacea; prescription [1]	11/2020
US (FDA)	Azelex cream 20%	Indicated for mild to moderate inflammatory acne vulgaris [2]	02/2024
UK (MHRA-licensed)	Skinoren 20% cream	“Topical treatment of acne vulgaris”; prescription only medicine [3]	6 Sep 2023

Both FDA PDFs carry the note that the label “may not be the latest approved by FDA” [1][2]. We did not check EMA or national EU registers, BfArM, or the German market, so we do not state the German or EU position.

Over the counter and cosmetic status: we could not verify this from a regulator. We did not open an FDA, SCCS or CosIng page on azelaic acid as a cosmetic ingredient. Secondary pages say that strengths of about 10% are sold as cosmetics in the EU and US, but we did not confirm that, and a DailyMed search returns many labels for 10 to 20% “acne cream” products from overseas companies whose regulatory status we did not assess [4]. Treat any statement about an unprescribed US or EU strength as unverified here. Because cosmetics are not approved before sale, a cosmetic label tells you nothing about trial-tested efficacy.

What the evidence shows

Rosacea

The US rosacea approval rests on two multicenter, randomized, double-blind, vehicle-controlled 12-week trials with identical protocols, 664 adults aged 21 to 86 (333 azelaic acid, 331 vehicle). Mean inflammatory lesions fell from about 17.5 to 6.8 (azelaic acid) versus 10.5 (vehicle) in study one, and from 17.9 to 9.0 versus 12.1 in study two [1]. The published report of the same program enrolled 329 and 335 patients, found lesion reductions of 58% vs 40% and 51% vs 39%, and found erythema improved in 44% vs 29% and 46% vs 28% [6]. The abstracts we read do not state the funder; the label does not either.

Comparator trials:

• A 2003 randomized trial of 251 people compared 15% gel with 0.75% metronidazole gel for 15 weeks. Lesion reduction was 72.7% vs 55.8%, and 56% vs 42% were rated improved for erythema. Neither drug had a clinically appreciable effect on telangiectasia. The abstract names Berlex (azelaic acid) and Galderma (metronidazole) as product makers; it does not state funding [7].
• A 2006 trial of 160 people found 15% gel and 1% metronidazole gel similar [9].
• Older 20% cream trials (116 people vs vehicle; 40 people split-face vs metronidazole) also found lesion benefit and no change in telangiectasia [10].

The 2015 Cochrane review of rosacea interventions (106 trials, 13,631 people for all treatments) rated evidence for topical azelaic acid as high quality [11]. We read only the Cochrane news summary, not the full review, and did not check whether a newer update exists. A 2023 systematic review of 43 RCTs of at least six weeks (search to December 2022) found azelaic acid beat vehicle for erythema and lesion counts at 12 weeks across 20 rosacea studies, and beat 0.75% metronidazole [8].

Acne

The abridged Cochrane review of topical agents for acne (Liu et al., 2020; searches to May 2019) included 49 trials and 3,880 participants. Azelaic acid was probably less effective than benzoyl peroxide on participants’ global assessment (RR 0.82, 95% CI 0.72 to 0.95) and probably no different from tretinoin (RR 0.94, 0.78 to 1.14). Evidence for azelaic acid was moderate quality; adverse events were mild [5].

In a 2016 German assessor-blinded trial (215 analysed), a benzoyl peroxide 3%/clindamycin 1% gel cut inflammatory lesions by a median 52.6% vs 38.8% with azelaic acid 20% at week 4 [13]. The 2023 systematic review found azelaic acid better than vehicle in 16 acne studies, with 20% reducing more lesions than erythromycin gel [8]. Funding was not in the abstracts.

Melasma and post-inflammatory hyperpigmentation

Two older double-blind trials compared 20% azelaic acid with hydroquinone over 24 weeks. In 155 patients, 73% vs 19% (2% hydroquinone) had good to excellent results [14]; in 329 women, results with 4% hydroquinone were not significantly different [15]. A 2011 open-label trial of 29 women found lower MASI with azelaic acid after two months, and its authors called for larger studies [16]. A 1998 vehicle-controlled trial in darker-skinned patients (phototypes IV to VI, 24 weeks) found greater pigment decrease than vehicle [17].

A 2023 meta-analysis of six RCTs (673 people) found a lower mean MASI with azelaic acid than hydroquinone (mean difference -1.23, 95% CI -2.05 to -0.40) but no difference in several other outcomes; the authors asked for larger, longer studies [18]. The 2023 systematic review found 20% better than vehicle for severity in seven melasma studies [8].

For post-inflammatory hyperpigmentation after acne, we found one 2023 single-blinded RCT: 30 people per arm, 12 weeks, 20% azelaic acid vs 5% tranexamic acid. Scores were comparable, with more side effects early in the azelaic acid group; there was no placebo arm [19]. The systematic review did not report a general PIH finding [8].

Tolerability

For the 15% gel, the label lists burning, stinging or tingling (29%), itching (11%), scaling or dryness (8%) and erythema or irritation (4%) in 788 trial participants. It reports isolated hypopigmentation, advises monitoring people with dark skin, and reports worsening asthma [1]. The UK document lists application-site burning, itching and redness as very common [3]. Systematic reviews found few significant differences in common adverse events versus comparators [8].

Pregnancy and breastfeeding in the labels

The 2020 Finacea label says topical absorption is minimal and fetal exposure is not expected. Oral studies in rats, rabbits and monkeys showed embryofetal toxicity at 162, 19 and 65 times the maximum human dose, with maternal toxicity and no malformations; dermal studies were not done. Lactation: azelaic acid is naturally present in human milk, and infant exposure is not expected [1]. The Azelex label says there are no adequate studies in pregnant women and that the drug should be used in pregnancy only if clearly needed; caution is advised for nursing mothers [2]. The UK Skinoren document likewise cites no adequate studies and advises caution [3]. Labels differ in wording, so these are three regulators’ texts, not one consensus.

Risks and unknowns

• Adverse effects: local irritation is common and usually early; hypopigmentation, vitiligo or depigmentation, asthma worsening and herpes labialis flares appear in label reports [1][2].
• Dark skin: the label says azelaic acid is not well studied in dark complexions [1], although a 1998 trial enrolled phototypes IV to VI [17].
• Not studied or not established: efficacy for erythema without lesions [1]; children (not established in the Finacea label) [1]; people 65 and over (too few enrolled) [1]; skin aging (the 2023 review found no eligible RCTs) [8].
• Evidence quality: many trials are decades old, short (12 to 24 weeks), vehicle-controlled, and funded by makers or unstated. Melasma trials are mostly small and some open-label.
• Gaps we did not resolve: funders for most trials, long-term relapse after stopping, and the US/EU status of non-prescription products.

What the source post claims and how it holds up

The post at [12] (X, @Hoostethics, 1 November 2026) is about tazarotene; its screenshot of a Reddit “starter pack” includes azelaic acid gel. The A-tier wording “clears vascular redness & blocks hyperactive melanocytes” comes from the same account’s 19 October tier list, whose text we took from the project’s source archive; we could not open that post by its own id. See also our tazarotene article and the tier-list check.

“Clears vascular redness”: partly supported. In papulopustular rosacea, erythema improved more than with vehicle (44% vs 29%; 46% vs 28%) [6] and a pooled analysis found less erythema than vehicle and metronidazole [8]. But the label states benefit on erythema without papules and pustules has not been evaluated [1], trials found no telangiectasia benefit [7][10], and no trial we read measured blood vessels. “Clears” overstates modest improvement.

“Blocks hyperactive melanocytes”: not verified as a mechanism; the outcome is partly supported. Trials show lightening in melasma versus vehicle and comparable or better MASI than hydroquinone [8][14][18]. We opened no mechanistic source, and the label’s text describes no melanocyte action, so the “blocks melanocytes” wording could not be verified.

Frequently asked questions

Is azelaic acid approved for melasma? Not in the labels we read. The FDA labels list rosacea (15% gel) and acne (20% cream), and the UK Skinoren label lists acne [1][2][3]. Melasma use rests on trials.

Is azelaic acid better than hydroquinone for melasma? A 2023 meta-analysis of six trials found a lower mean MASI score with azelaic acid but no difference on other outcomes, and called for larger studies [18].

Is it safe in pregnancy? The labels cite no adequate human studies. The FDA gel label expects minimal fetal exposure; the UK and Azelex labels advise caution [1][2][3]. Pregnant readers should ask their prescriber.

Does it work as well as benzoyl peroxide for acne? The Cochrane-derived review found it probably less effective than benzoyl peroxide and probably similar to tretinoin on global assessment [5]. For related topics see niacinamide.

Sources

1. FDA, “FINACEA (azelaic acid) Gel, 15% prescribing information” (revised 11/2020). https://www.accessdata.fda.gov/drugsatfda_docs/label/2020/021470s015s016lbl.pdf
2. FDA, “AZELEX (azelaic acid) cream 20% label” (revised 02/2024). https://www.accessdata.fda.gov/drugsatfda_docs/label/2024/020428Orig1s027lbl.pdf
3. Electronic Medicines Compendium, “Skinoren 20% Cream SmPC” (6 Sep 2023). https://www.medicines.org.uk/emc/product/285/smpc
4. NLM DailyMed, search results for “azelaic acid” (read 2026-11-08). https://dailymed.nlm.nih.gov/dailymed/search.cfm?labeltype=all&query=azelaic+acid
5. Liu H et al., “Evidence-based topical treatments (azelaic acid, salicylic acid, nicotinamide, sulfur, zinc, and fruit acid) for acne: an abridged version of a Cochrane systematic review”, J Evid Based Med (2020), doi 10.1111/jebm.12411. https://pubmed.ncbi.nlm.nih.gov/33034949/ (abstract read via Europe PMC)
6. “Efficacy and safety of azelaic acid (15%) gel as a new treatment for papulopustular rosacea: two vehicle-controlled phase III studies” (2003). https://pubmed.ncbi.nlm.nih.gov/12789172/ (abstract via Europe PMC)
7. “A comparison of 15% azelaic acid gel and 0.75% metronidazole gel in papulopustular rosacea: a randomized trial” (2003). https://pubmed.ncbi.nlm.nih.gov/14623704/ (abstract via Europe PMC)
8. “A systematic review to evaluate the efficacy of azelaic acid in acne, rosacea, melasma and skin aging”, J Cosmet Dermatol (2023). https://pubmed.ncbi.nlm.nih.gov/37550898/ (abstract via Europe PMC)
9. “Efficacy and safety of once-daily metronidazole 1% gel compared with twice-daily azelaic acid 15% gel in rosacea” (2006). https://pubmed.ncbi.nlm.nih.gov/16706244/
10. “Double-blind comparison of azelaic acid 20% cream and its vehicle in papulo-pustular rosacea” (1999), https://pubmed.ncbi.nlm.nih.gov/10598760/ ; “azelaic acid 20% cream vs metronidazole 0.75% cream” (1999), https://pubmed.ncbi.nlm.nih.gov/10365928/
11. Cochrane, “Featured review: Interventions for rosacea” (15 Jun 2015; CD003262.pub5). https://www.cochrane.org/about-us/news/featured-review-interventions-rosacea
12. @Hoostethics on X (1 Nov 2026). https://x.com/Hoostethics/status/2116976569034092805 (partial text read through the fxtwitter mirror because X blocked access; tier-list wording from project source archive) ; original topical skincare tier list of 2026-10-19: https://x.com/Hoostethics/status/2112214865779806696
13. Schaller M et al., “Benzoyl peroxide 3%/clindamycin 1% vs azelaic acid 20% in mild-to-moderate acne” (2016). https://pubmed.ncbi.nlm.nih.gov/26915831/
14. “Double-blind comparison of azelaic acid and hydroquinone in the treatment of melasma” (1989). https://pubmed.ncbi.nlm.nih.gov/2528260/
15. “The treatment of melasma. 20% azelaic acid versus 4% hydroquinone cream” (1991). https://pubmed.ncbi.nlm.nih.gov/1816137/
16. “Comparative study of 20% azelaic acid and hydroquinone 4% cream in melasma” (2011). https://pubmed.ncbi.nlm.nih.gov/22151936/
17. “Azelaic acid 20% cream in the treatment of facial hyperpigmentation in darker-skinned patients” (1998). https://pubmed.ncbi.nlm.nih.gov/9829446/
18. Albzea W et al., “Azelaic acid versus hydroquinone for melasma: systematic review and meta-analysis” (2023). https://pubmed.ncbi.nlm.nih.gov/37457606/
19. Sobhan M et al., “20% azelaic acid cream versus 5% tranexamic acid solution for post-inflammatory hyperpigmentation in acne” (2023). https://pubmed.ncbi.nlm.nih.gov/37213446/

This article is information, not medical advice.