Salicylic Acid Is in Ten of Our Products. Using Them Together Is the Problem.

A patch on your chin overnight. Body wash in the shower. Scalp scaler on Sunday, a hair tonic every night before bed, and the anti-hairloss shampoo you've been on since spring. Five products, one ordinary week, and every one of them has salicylic acid in it.

We wanted a number. How many milligrams of salicylic acid does that routine actually put on a person in a day? We spent a while on it and came up empty, because not one of those ten products prints its salicylic acid percentage anywhere, and no rule requires them to. The EU's safety committee published the only serious attempt at the question we could find, in June 2023. The FDA has not published one at all.

Our own planning note for this piece said nine products. We recounted the catalog before writing and it came back ten. The miscount turned out to be the useful part.

We said nine. We counted again and it's ten.

Here are all ten, re-pulled from the ingredient data rather than copied from our own earlier note.

The tenth is the hair dye shampoo, and we missed it for a reason that should bother you slightly. Nine of these products record salicylic acid with the risk grade mid and the function label "BHA Exfoliant." The GDS entry records it as low, with the function written as "Scalp / pH." Same molecule, same catalog, different words, entered by hand on a different day. That's why our own count was wrong: we had been filtering on the flag, not the ingredient.

Position matters here more than it looks. INCI order runs by descending concentration down to 1%, after which the order is free. Salicylic acid at 46 of 46 in the COSNORI tonic is almost certainly a fraction of a percent. At 6 of 7 in a patch with seven total ingredients, it could be anything.

Every ceiling in the book caps one product at a time

Commission Regulation (EU) 2019/1966 rewrote Annex III entry 98. Salicylic acid: 3.0% maximum in rinse-off hair products, 2.0% in other products, with a separate 0.5% limit when it's doing preservative duty under Annex V. Not permitted for children under three. Not permitted in anything that could put it in the lungs. In the United States, 21 CFR 333.310 is two words and two numbers — "Salicylic acid, 0.5 to 2 percent" — and unlike some monograph actives it carries no combination requirement at all.

Read entry 98 again and notice what the subject of the sentence is. It's the product. It's always the product.

A regulator who wanted to cap a person's daily exposure would have to know how many salicylic acid products that person owns, in what formats, on how much skin, for how long. No agency in Brussels, Silver Spring or Cheongju collects that. So the ceilings hold ten times over in our catalog and mean something slightly different from what most people assume they mean.

How much of it actually gets through

The Cosmetic Ingredient Review's 2019 amended assessment is the document most of the industry leans on, and it is more reassuring than we expected. Roughly 10% of applied salicylates stay in the skin. A representative product at 2% or below delivers about 20% of the systemic exposure you get from swallowing one 81 mg baby aspirin. In peel studies — 30% salicylic acid, far above anything cosmetic — plasma came back at 0.81 µg/mL against 56.4 µg/mL for oral aspirin, a ratio of about 50 to 1.

Measured dermal absorption in the same document ranges from 34.5% to 50.1% depending on conditions, which is a wide enough band to make any single-number estimate a bit of theatre.

Davis and colleagues put 2% salicylic acid on faces for fourteen straight days in 1997, across normal, acne-prone and aged skin, and found the skin condition made no difference to absorption. The vehicle did — a hydroalcoholic base pushed more through than a cream. Their conclusion was that people without a gross skin disorder run minimal systemic risk from routine 2% products, with an 81 mg aspirin producing at least five times the exposure.

Then there is the SCCS opinion adopted on 6–7 June 2023, which is the only place we found anyone modelling the stacking question directly. It runs a probabilistic aggregate across product categories, assuming 60% dermal absorption for most and 100% for oral care and lipstick, and arrives at a systemic exposure dose of 0.45 mg per kg of body weight per day. The developmental no-observed-adverse-effect level it checks that against is 75. Divide one into the other and the margin is about 167 times.

The multiplier nobody models

Benfeldt, Serup and Menné ran microdialysis probes under human skin in 1999 and measured salicylic acid penetration against deliberately damaged barriers. Acetone, mild damage: 2.2 times more got through. One percent SLS, enough to raise a mild dermatitis: 46 times. Two percent SLS: 146. Tape-stripping: 157. Penetration tracked TEWL and erythema readings closely enough that you can predict one from the other.

Which reframes the whole question.

Ten products at legal concentrations on intact skin is a rounding error. The same ten on skin you have already over-exfoliated is a different arithmetic entirely, and the thing that damaged the barrier may well have been the products themselves.

The documented harm, when it happens, follows that pattern. Madan and Levitt collected every topical salicylic acid toxicity case they could find between 1966 and 2012 and got twenty-five, including four deaths. Toxic plasma threshold around 35 mg/dL; the reported cases ran from 252 to 2,234 mg/L. Every one involved application across 40% to 80% or more of the body, usually onto psoriasis or ichthyosis, or onto a child. Not one came from stacking ordinary cosmetic-strength products, and we looked hard for one.

There is a second assumption in the rulebook that nobody has tested. Rinse-off hair products get 3.0% where leave-on gets 2.0%, on the reasoning that ninety seconds of contact delivers less than eight hours of it. That is obviously true in direction. We could not find a single pharmacokinetic study comparing the two head to head, which means the gap between those two numbers is a judgement written into law rather than a measurement anyone has taken.

On whether any of this is buying you clear skin, the evidence is thinner than the safety literature. The 2020 Cochrane review of topical benzoyl peroxide swept up 120 trials and 29,592 participants and rated the certainty predominantly low to very low. Salicylic acid appears in it mainly as a comparator whose self-reported improvement outcomes the trials did not report.

What our screening found

The practical takeaway

Count leave-on, not total. A shampoo at 3.0% that spends ninety seconds on a scalp is not the same event as a tonic at a fraction of a percent that stays there all night, and the ceiling numbers will mislead you about which is which. Of our ten, three are leave-on.

Then watch the barrier rather than the bottle count. Stinging that wasn't there last month, tightness after cleansing, a flush that takes longer to settle — those are the signs that the multiplier has switched on, and they arrive well before anything a blood test would catch. Back off to one salicylic acid product until they're gone.

And if you are treating skin across most of your body, or treating a child, this stops being a cosmetics question. Every one of the twenty-five toxicity cases in the literature came from that territory. Take it to a clinician instead of to us.

For how we grade individual ingredients in the first place, our screening method is written up here, the ten ingredients we flag most often puts salicylic acid in context against the rest of the catalog, and if you only want the patch question, we compared plain hydrocolloid against salicylic patches separately. The products that cleared our screen are listed on our awards page.