1. What azelaic acid actually is
Azelaic acid is a saturated dicarboxylic acid. It occurs in cereal grains including wheat, rye and barley, and it is also produced on human skin by Malassezia yeasts, which is how it came to dermatological attention in the first place. It is a small, simple molecule with no particular glamour attached to it, which may be why it is consistently underdiscussed relative to its usefulness.
The thing that separates it from almost every other ingredient reviewed on this site is its regulatory position. In the United Kingdom, azelaic acid preparations at the strengths used therapeutically are licensed medicines supplied on prescription, indicated for acne vulgaris and for papulopustular rosacea. That means a regulator has assessed a dossier of efficacy and safety data, a summary of product characteristics exists, and adverse effects are collected through a formal system. Very little in the natural ingredient category has been through anything comparable.
At the same time, the same words appear on cosmetic products sold without prescription. Those products are governed by cosmetics law, which regulates safety and labelling and does not require any demonstration that the product works. The gap between those two regimes is the single most important thing to understand about this ingredient, and it is covered in more depth in supplement regulation in the UK.
2. What it is thought to do
Four mechanisms are usually described, and they are unusually well characterised for an ingredient of this kind.
Antimicrobial activity. Azelaic acid is active against Cutibacterium acnes, the organism implicated in inflammatory acne. Unlike antibiotics it does not appear to drive bacterial resistance, which is a meaningful practical advantage given how much dermatological antibiotic prescribing there has been.
Anti-inflammatory activity. It reduces the production of reactive oxygen species by neutrophils. This is the mechanism most often invoked for the rosacea indication, where the underlying problem is inflammatory rather than infective.
Effects on keratinisation. It appears to normalise the abnormal shedding of cells lining the follicle, which is the process that produces comedones.
Tyrosinase inhibition. It interferes with the enzyme that initiates melanin synthesis, which is the basis of the pigmentation claims. Notably, the effect is reported to be more pronounced on hyperactive melanocytes than on normally functioning ones.
Mechanism is not evidence of benefit, and we say so about every ingredient on this site. It matters here only because it makes the pattern of results coherent: the claims with the best human evidence are the inflammatory ones, and the claims with the weakest evidence are the pigmentation ones, which is what the mechanistic picture would predict.
3. What the human evidence looks like
The shape of this evidence base is different from most on this site, and the difference is worth describing plainly.
For papulopustular rosacea and for inflammatory acne, controlled human trials exist in sufficient number and size to have supported a marketing authorisation, and UK primary care guidance lists topical azelaic acid among the treatment options. That is a stronger position than almost any cosmetic ingredient occupies. It is not the same as saying the evidence is extensive. Much of it was generated by or in collaboration with the companies that hold the licences, trials in this field are typically twelve to fifteen weeks long, and outcome measures rely on lesion counts and investigator assessment scales that are reproducible but coarse.
For redness and flushing, the picture is thinner, and the reason is clinical rather than statistical. Rosacea is not one thing. Guidance distinguishes the inflammatory papules and pustules from persistent central facial erythema, from flushing episodes, and from visible vessels, and treatments that work on one of those do not reliably work on the others. Anyone whose main complaint is background redness is being sold a different problem's solution.
For pigmentation, the human evidence is smaller, shorter, and harder to interpret. Melasma in particular relapses, is driven substantially by ultraviolet exposure and hormonal factors, and produces trial results at three months that frequently do not hold at a year. We have graded it Limited and we would not expect that to change without longer, larger, independent work.
4. The strength question, and why cosmetic versions are not the same product
The licensed UK preparations are considerably more concentrated than the cosmetic ones. Cosmetic products sold as azelaic acid treatments are commonly formulated at around half the licensed concentration or lower, and many do not contain azelaic acid at all in the form studied. Derivatives such as potassium azeloyl diglycinate are used because they are easier to formulate, more stable and less gritty in a cream base. Whether a derivative behaves like the parent molecule on skin is a question that would need to be answered by trials, and it has not been.
There is also a formulation problem that has nothing to do with concentration. Azelaic acid has poor solubility, and how much of it actually crosses the stratum corneum depends heavily on the vehicle it is suspended in. Two products with the same number on the front can deliver quite different amounts to the skin. This is one reason a licensed medicine and a cosmetic sharing a percentage are not equivalent even when the percentage matches.
The decision rule. If the aim is treating diagnosed rosacea or acne, the question is whether a licensed preparation is appropriate for you, and that is a conversation with a GP or a pharmacist. If the aim is general skin tone improvement without a diagnosis, a cosmetic product may still be worth trying, but it should be bought on the basis that it is untested rather than on the basis of trials conducted on a different product.
5. Tolerability, and the one warning worth repeating
The most common adverse effects are local and appear early: stinging, burning, itching, dryness and scaling, usually in the first few weeks and usually settling. That pattern is well documented in the product information for the licensed preparations, which is a document worth reading precisely because it exists.
The warning that gets least attention concerns pigmentation in the other direction. Product information for licensed azelaic acid advises watching for lightening of the skin in people with darker complexions, and any unexpected loss of pigment should be reported to the prescriber. An ingredient that interferes with melanin synthesis can plausibly do so where it was not wanted, and this is a good example of a mechanism cutting both ways.
Rosacea and acne are also conditions where self-diagnosis is unreliable. Persistent facial redness has a long list of possible causes, several of which are not rosacea and some of which matter. If a rash on the face is new, changing or not behaving as expected, that is a reason to be assessed rather than a reason to buy something.
6. Our position
Graded Moderate overall. The inflammatory claims for the licensed preparations sit at Moderate and would sit higher with more independent replication and longer follow-up. The redness and pigmentation claims sit at Limited. Cosmetic strength products sit at Insufficient, not because they are known not to work but because nobody has tested them and the manufacturers are under no obligation to.
The general principle this ingredient illustrates is worth carrying elsewhere. When a molecule exists both as a licensed medicine and as a cosmetic, evidence generated on the medicine is routinely used to sell the cosmetic, and the two are not the same product. That move is made with retinoids, with hydroquinone alternatives and with several of the ingredients graded on this site. It should be noticed every time.