Retinoid 'Purging' vs. Irritation: What Trial Data Say and How to Tell the Difference

Few phrases in skincare carry as much reassurance as 'it has to get worse before it gets better.' Brands, forums and well-meaning friends use it to explain why a new retinoid leaves skin red, flaky and bumpy. But the published research draws a sharper line between two things that often get lumped together: a true flare of existing blemishes, and plain irritation. Knowing which one you are looking at changes what you should do next.

What 'purging' is supposed to mean

The popular story goes like this: retinoids speed up skin cell turnover, which pushes hidden congestion to the surface all at once, so skin looks worse for four to six weeks before clearing. It is a tidy explanation, and it is mostly repeated by brand blogs rather than primary research. A review of the evidence published by a skincare company concludes that a distinct 'purging' phase is not established, and that there is no validated way to tell purging from an adverse reaction. That source sells skincare itself, so it is worth checking against the underlying trial data.

What the trial data actually show

The most-cited analysis pooled three multicenter, double-blind Phase 3 trials with 4,550 participants in total, using a gel combining clindamycin and 0.025% tretinoin. Researchers counted how many people had a 10% or greater increase in inflammatory lesions at week 2. The numbers are more nuanced than either camp suggests:

• In people with mild baseline blemishes, 0.025% tretinoin alone had a flare rate of 15.4%, versus 8.7% on the vehicle (a base with no active).
• In people with severe baseline blemishes, the highest flare rate, 23.8%, appeared in the vehicle group, meaning it occurred without any retinoid at all.
• The combination gel had the lowest flare rates, 7.6% in mild and 8.5% in moderate groups.

The author's reading was that some of the early worsening may reflect the natural ups and downs of skin, or a delay before benefits show, and not only a drug effect. A separate review in the Journal of Drugs in Dermatology looked at trial data and likewise found no primary evidence of retinoid-induced flaring, though both papers come with the caveat of heavy pharmaceutical involvement. So the honest summary is: a small flare signal appears in one subgroup, and bumps can appear on skin that is using nothing at all.

Irritation is the better-documented effect

What trials do document consistently is irritation. The original alcohol-based tretinoin formulations caused redness, scaling, burning and itching, mostly in the first few weeks. In a 21-day split-face study of 45 participants, one side of the face received a clindamycin and tretinoin gel and the other side received either 0.1% tretinoin or 0.1% adapalene, once daily. The combination product produced significantly less redness, scaling, itching, burning and stinging than 0.1% tretinoin, with only a non-significant trend against 0.1% adapalene. That study was run by authors affiliated with the maker of the combination gel, so treat the product comparison cautiously. The broader takeaway is the useful part: strength and formulation both change how irritating a retinoid feels, and irritation is dose-dependent.

A practical way to tell the two apart

Because there is no validated test, use pattern clues. They are heuristics drawn from clinician guidance, not proof:

1. Where it shows up. Bumps in the spots where you usually see congestion are more consistent with an ordinary flare. Bumps in new places, such as the cheeks of someone who rarely gets them, point toward irritation.
2. What it feels like. Stinging, burning, tightness and visible flaking signal irritation. Blemishes alone do not.
3. Which way it is heading. Mild dryness that settles after a few weeks is common. Redness that spreads, or discomfort that gets worse week over week, is a reason to scale back.

If skin looks swollen, blistered or intensely red, or you notice itching that spreads, stop the product and talk to a healthcare professional.

Slow down instead of pushing through

The advice to 'just push through the purge' is the part with the weakest support. A more defensible approach is to lower the dose of the exposure. Common clinician guidance is to start two nights a week, move up only when skin looks calm, apply to fully dry skin, and pair it with a plain moisturizer. Expect benefits to take months, not days. Starting with a lower-strength product is also supported by the tolerability differences between formulations seen in comparative studies.

What about buffering with moisturizer?

Many people layer moisturizer around the retinoid to soften the sting. A poster presented at the 2025 American Academy of Dermatology meeting tested this ex vivo on skin explants from a single donor, measuring gene-expression markers of retinoid activity over 48 hours. Moisturizer applied either before or after the retinoid (an 'open sandwich') left activity comparable to the retinoid alone. Moisturizer on both sides (a 'full sandwich') cut that activity roughly three-fold. The study was sponsored by a consumer-health company, used one donor, and was a conference poster rather than a peer-reviewed paper, so it is suggestive rather than settled. Still, it hints that one layer of moisturizer is a reasonable trade, while stacking thick layers on both sides may blunt the very thing you are applying the retinoid for.

The takeaway

Early bumps on a new retinoid are not automatically a good sign, and they are not automatically a bad one. The research supports three working rules: expect some dryness, do not assume every bump is 'purging,' and treat stinging or spreading redness as a signal to ease off. Start at two nights a week, use one layer of plain moisturizer, track whether the changes stay in your usual trouble spots, and give a stable routine several weeks before judging how skin looks. If things are escalating instead of settling, the evidence favors reducing frequency over enduring it.