Industry use data puts the maximum reported concentration of Houttuynia cordata extract in leave-on face products at 1.5%, and most formulas use a small fraction of that. The published research is almost entirely cell cultures plus one oral mouse study. Nobody has measured whether it crosses human skin from a finished cosmetic.
Heartleaf arrives in our studio in a particular way: a bag of toners bought over three days in Hongdae, and when we read the labels, four of the five have heartleaf on them. That is one ingredient four times a day, layered, at a strength nobody has disclosed.
What is heartleaf, and what is it doing in your toner?
Heartleaf is Houttuynia cordata — 어성초 in Korean, "fishy-smell herb" — a creeping wetland plant native to East Asia, used for centuries in Korean herbal practice, mostly taken internally. In cosmetics it is a water or alcohol extract of the leaves, stems or aerial parts, sold to you as a calming ingredient for reactive, redness-prone and acne-prone skin.
It is not niche. The Cosmetic Ingredient Review's 2025 safety assessment counted 1,032 reported uses of Houttuynia Cordata Extract, 836 of them leave-on. Six INCI names exist for this one plant, and one, Houttuynia Cordata Water, is registered for fragrance use only — worth knowing when "heartleaf water" sits near the top of a list.
In Seoul you will meet it as toners and essences, cleansing foams and sheet masks, and — at a pharmacy rather than a cosmetics shop — dried 어성초 to brew. Only the first makes a claim about your face.
How much Houttuynia cordata is actually in your product?
At most 1.5%, and usually far less. The CIR review is the only place I know of that publishes real numbers, and they are startlingly low: reported leave-on concentrations span 0.00008% to 1.5%, the maximum being in face and neck products; eye lotions, 0.0008% to 0.0027%. For the other five heartleaf INCI names, no concentration data exists at all.
Now compare the research. The most thorough recent laboratory work — a 2025 Scientific Reports study on heartleaf essential oil — ran its cell assays at 25 to 500 µg/mL and formulated its test gel and cream at 2% w/w of the oil. Those are not translatable units: you cannot convert a concentration in a culture dish into a percentage in a toner, and that incomparability is the real finding, not a detail.
Does any of it get into your skin?
No one has published a measurement of how much heartleaf extract crosses human stratum corneum from a cosmetic product. That is the honest state of it.
What is known is the shape of the gap. The 2025 Scientific Reports work was explicit that it "was confined to in vitro assays" — microbial cultures, cell-free enzyme assays, and four cell lines. No animals, no human skin, no permeation study. A 2021 review of Centella asiatica makes the same admission about heartleaf's better-studied sibling: cutaneous absorption of its topical actives "has yet to be" measured.
The one efficacy study I could read in full was oral. Park and colleagues (2023) gave BALB/c nude mice 30 mg/kg/day of heartleaf extract by mouth for four weeks, then induced atopic dermatitis with DNCB; n = 5 per group, n = 3 for IgE. Heartleaf alone brought scratching from 42.4 to 22.8 bouts per 30 minutes, SCORAD from 15.4 to 12.0, and epidermal thickness from 104.8 to 88.3 µm against healthy controls at 26.16 µm. The effect was real in that model — and heartleaf was the weakest of the three herbs tested. Eating a plant extract also tells you nothing about applying it.
Unknown is nearly everything else: which of the extract's dozens of compounds does the work, whether it survives formulation, and whether any reaches living epidermis. "Houttuynia Cordata Extract" specifies no solvent, plant part or standardisation, so two identical ingredient lists can be chemically different products.
Why does the eye area get a thousand times less of it?
Because formulators treat thin skin as a different exposure problem, and the use data shows it plainly: eye lotions report heartleaf at 0.0008% to 0.0027%, between 500 and 1,900 times lower than the 1.5% maximum reported for face and neck.
That gap is not timidity. Eyelid skin is the thinnest on the face, it flexes thousands of times a day, and the same amount of product is a far higher dose per unit of barrier there — or on the neck, or the corners of the mouth — than across a cheek. Which is why an ingredient that is unremarkable on your cheekbone can sting at the inner corner of your eye. The ingredient did not change. The exposure did.
Know your own skin before you buy anything
Ingredients have to be introduced one at a time, at a pace your barrier can currently take — and what it can take today is not a fixed property of your skin type. It depends on the state you are in this week.
Think of someone midway through treatment for a serious illness. They will not recover faster by training harder; the body is already spending everything it has on repair, and more demand is simply more demand. What it needs back is rest. Compromised skin works the same way. When the barrier is broken — stinging water, tight flaking, a flush that will not settle — the first move is subtraction. Take things away, keep cleansing gentle and a bland moisturiser on, and wait for the skin to stop reacting before you add anything meant to act on it.
None of this is a diagnosis or a treatment plan. Skin that is painful, spreading, weeping or not improving belongs with a dermatologist, not a toner.
The case against some of this
1. I implied heartleaf calms inflammation. The strongest number behind that is a nitric oxide inhibition IC50 of 1.176 µg/mL in LPS-stimulated mouse macrophages — and the 2025 authors state no positive control was included, so there is nothing to scale it against. The mouse model that did show reduced inflammation used oral dosing. The Centella review shows where this pattern ends: deep mechanism work, sparse trials, no measurement of whether the active reaches skin. Narrowed: heartleaf formulas may reduce how reactive skin feels for some people; I will not claim they treat inflammation.
2. I implied its antibacterial activity is relevant to breakouts. In the same study, heartleaf essential oil gave an 11.6 mm inhibition zone against Cutibacterium acnes while ampicillin gave 24.1 mm on the same plate; its antioxidant capacity was similarly modest, a DPPH IC50 of 13.25 mg/mL against Trolox's 0.008 mg/mL. And agar is not a sebum-filled follicle. The Centella review records a parallel failure: a methanol extract showed little antibacterial activity against acne bacteria despite promising mechanism data. Narrowed: no acne claim from me.
3. I implied heartleaf is a gentle, low-risk ingredient. The CIR found no dermal irritation or sensitisation studies for heartleaf-derived ingredients in the published literature, and none were submitted — the panel is formally requesting that data, comments due 14 February 2026. Separately, 28-day oral rat studies found deaths at 500 and 1,000 mg/kg/day with liver and kidney changes (NOAEL 250 mg/kg/day). Oral toxicology is irrelevant to a toner, but it establishes that the plant is not inert. Narrowed: heartleaf's reassuring record comes from long use at very low concentrations, not from safety testing — a weaker thing than "gentle."
Is heartleaf the same thing as cica?
No. Cica is Centella asiatica, a different plant with different chemistry and a larger clinical literature. Korean products often pair them, which makes any result hard to attribute to either.
Should I start heartleaf and a new acid in the same week?
Not if you want to learn anything. Introduce one at a time and give it a fair run. If both go on at once and your skin reacts, you cannot know which to stop.
Is the pharmacy version stronger than the cosmetic one?
It is a different product, not a stronger one. Dried 어성초 is a herb for brewing, with no topical dosing guidance attached. Please do not make it into a DIY toner.
At Laurel Studio we read the skin in front of us, not the shelf behind it: barrier state first, then what to add and in what order. If you are in Hongdae with a bag of toners and no plan, a 1:1 consultation is the fastest way to find out which two you should actually be using.
Sources
- Cosmetic Ingredient Review. Safety Assessment of Houttuynia cordata-Derived Ingredients (Scientific Literature Review, December 2025) — concentration of use, frequency of use, oral toxicology, data gaps. Full document read.
- Sangthong S. et al. "In vitro bioactivities and formulation stability of Houttuynia cordata essential oil for cosmetic applications." Scientific Reports 15:37263 (2025) — all in vitro; cell lines and microbial cultures; no human or animal testing; no permeation study. Full text read.
- Park S. et al. "Phenolic changes in a combined herbal extract of Lithospermum erythrorhizon, Houttuynia cordata, and Spirodela polyrhiza and alleviation of DNCB-induced atopic dermatitis in BALB/c mice." Food Science and Biotechnology (2023), PMC10766889 — animal study, oral dosing, n = 3–5 per group. Full text read.
- Not read: Jung J. et al. "In vitro skin permeation and anti-atopic efficacy of lipid nanocarriers containing water soluble extracts of Houttuynia cordata." Drug Development and Industrial Pharmacy 40(10), 2014 — paywalled; bibliographic record only, no findings used here.
- Not read: two papers whose titles claim human testing of topical heartleaf formulations — a 17% heartleaf cream study in Pharmacognosy Magazine (2024) and a facial toner clinical efficacy evaluation (2023). Neither full text was accessible, so neither is counted as evidence in this article.
Sources for the case-against section
- Sangthong S. et al., Scientific Reports 15:37263 (2025) — authors' stated absence of a positive control in the anti-inflammatory assay; ampicillin comparator zone of 24.1 mm vs 11.6 mm for the oil against C. acnes; DPPH IC50 of 13.25 mg/mL vs Trolox 0.008 mg/mL. Full text read.
- Park KS. "Pharmacological Effects of Centella asiatica on Skin Diseases: Evidence and Possible Mechanisms." Evidence-Based Complementary and Alternative Medicine (2021), PMC8627341 — sparse clinical evidence despite extensive mechanism data; cutaneous absorption of topical actives not yet measured; weak antibacterial activity against acne bacteria. Full text read.
- Cosmetic Ingredient Review (December 2025) — no published dermal irritation or sensitisation studies located; data formally requested, comments due 14 February 2026; 28-day oral rat NOAEL 250 mg/kg/day with mortality at higher doses. Full document read.
