Ginseng, Pearl, Snow Lotus: Auditing the Traditional Chinese Medicine Actives

The concession comes first, and it is not a small one

Before auditing anything, the position that makes this whole exercise honest.

In the 1970s, researchers in China isolated and tested artemisinin, working from the traditional materia medica — a fourth-century text describing the preparation of Artemisia annua supplied the crucial insight about extraction temperature. The compound became the backbone of modern antimalarial therapy and is credited with saving millions of lives. In 2015 Tu Youyou received the Nobel Prize in Physiology or Medicine for it, and titled her Nobel lecture 'Artemisinin — A Gift from Traditional Chinese Medicine to the World.'

That is not a debating point to be conceded and moved past. It is the reason the reflexive Western dismissal of a two-thousand-year pharmacopoeia as folklore is intellectually lazy. Long-standing traditional use is an unusually productive hypothesis generator: it tells you that a great many humans, over a long period, observed something worth repeating, and that whatever is in the plant was not acutely poisoning them at the doses they used. Given how expensive it is to find bioactive compounds any other way, that is a valuable signal and pharmacology has mined it repeatedly.

And here is where the two things separate. A hypothesis is not a result. Centuries of use cannot, by itself, establish that a specific extract, at a specific concentration, applied topically, outperforms a placebo for a specific modern claim. The tradition points; the trial answers. Everything below turns on keeping those two functions apart — which is exactly what the marketing in this category is designed to blur, and exactly what the reflexive scepticism gets wrong in the opposite direction.

What 'used for centuries' does and does not buy

It is worth being precise about what the heritage claim actually delivers, because the answer is neither nothing nor everything.

What it supports: that the material contains something biologically active; that it is unlikely to be acutely hazardous in the preparation and dose traditionally used; and that it is worth the cost of testing. Those are real.

What it does not support, and this is the whole audit: efficacy for the modern claim being made. And there is a specific failure mode that recurs so often it deserves naming — route of administration. Most traditional use of these materials is ingested: decoctions, teas, powders taken internally. Skincare applies them to the outside of the stratum corneum. Those are not merely different doses; they are different pharmacokinetic universes. A compound that works when 3 grams are absorbed across the gut daily for six months has told you nothing about the same compound at trace concentration on the surface of the skin.

The second failure mode is the mixture problem. Traditional formulations are almost always combinations, prescribed as a whole. Trials of traditional preparations therefore usually test combinations too — which is methodologically appropriate to the tradition and methodologically useless for attributing an effect to any single ingredient on an INCI list.

Hold those two in mind and the rest of this article mostly writes itself.

The strong end: green tea and licorice

Two of these have genuine files, and neither is a C-beauty exclusive — both are in Western products by the thousand, which is itself a useful check on the framing.

**↗Green tea** has the deepest mechanistic literature of the group. The polyphenols, principally epigallocatechin-3-gallate, have been characterised in detail for photoprotective activity: induction of interleukin-12, IL-12-dependent nucleotide excision repair of UV-induced DNA damage, inhibition of UV-induced immunosuppression, anti-angiogenic effects, and prevention of UVB-induced tumour development in animal models under both oral and topical administration. That is a coherent, well-worked mechanism rather than a vague antioxidant hand-wave.

The honest limit: much of the strongest work is murine and mechanistic, and topical human endpoint trials are smaller and less consistent than the mechanism would lead you to hope. Green tea is a well-characterised supporting antioxidant that is cheap, stable enough to formulate, and reasonable to want in a product. It is not a treatment, and no green tea extract replaces a sunscreen.

**↗Licorice root** is the pigment one. Glabridin and related flavonoids inhibit tyrosinase, and mechanistic work published in 2025 mapped Glycyrrhiza glabra extract's depigmenting action onto the CRTC1/MITF signalling pathway — the transcriptional control point for melanogenesis. Worth quoting the framing of that paper, because it is more candid than the marketing: licorice is widely recognised for skin-whitening properties, but the specific active compounds and underlying mechanisms 'remain largely uncharacterised.' In other words, the mechanism is being worked out now, in 2025, for an ingredient sold on the strength of it for decades.

The practical grade: licorice is a legitimate supporting brightener with a real target and a modest effect size, and it belongs alongside a primary agent rather than instead of one. That distinction matters more than it sounds, because ↗stacking more brightening actives does not fade pigment faster — it mostly raises the irritation that drives post-inflammatory pigment in the first place.

Ginseng, and the citation error the whole category makes

↗Ginseng deserves its own section, because it has the single best human trial in this article and that trial does not say what it is used to say.

The study is Cho and colleagues, published in the Journal of Medicinal Food in 2009, from a dermatology department at a Seoul university hospital. Design: randomised, double-blind, placebo-controlled, 24 weeks, in healthy women over 40, with 82 completers. Endpoints were measured objectively, and — unusually and impressively — facial skin biopsies were taken before and after for real-time PCR and immunohistochemistry.

The results were positive and specific. Facial wrinkles significantly improved. Type I procollagen gene and protein expression increased. MMP-9 induction, which drives collagen breakdown, was prevented. Fibrillin-1 fibres were elongated. All of that only in the treatment group, with no changes in elasticity, epidermal water content, erythema, pigmentation or epidermal thickness in either arm.

That is a good study. Biopsy-confirmed changes in procollagen expression are not a soft endpoint.

Now the two facts that appear in the abstract and almost never in the citation. First, the intervention was oral — 3 grams per day of an extract-containing preparation, taken by mouth for six months. Second, it was a mixture: red ginseng together with Torilus fructus and Corni fructus. So the trial demonstrates that ingesting a specific three-herb preparation daily for half a year is associated with measurable dermal collagen changes. It demonstrates nothing whatsoever about a ginseng extract sitting in a serum, and it cannot attribute the effect to ginseng even in the oral form.

This is the most common evidence error in the entire botanical skincare category, and it is not confined to Chinese brands — Korean, Japanese and Western products cite the same class of oral or in vitro work for topical claims constantly. When a product page says ginseng is 'clinically shown' to improve wrinkles, this is very often the study meant. Read what was administered, and how.

The thin end: pearl, snow lotus, tremella, ganoderma

Here the honest grades drop, and it is worth stating them individually rather than dismissing the group.

Pearl powder is chemically unremarkable — largely calcium carbonate with a conchiolin protein fraction and free amino acids. The best available work is a 2022 study in which pearl extract gel was applied topically to shaved mouse skin before UV irradiation; it reported reduced UV-induced wrinkling, laxity and dryness, raised catalase, glutathione peroxidase and superoxide dismutase activity, lowered malondialdehyde, and suppressed IL-1β, IL-6, PGE2, TNF-α and COX-2. That is genuine, careful preclinical work with a plausible mechanism. It is also a mouse model, and there are no controlled human trials to set beside it. The correct grade is preclinical, and the correct posture is interest without expenditure. Note that pearl is not a C-beauty tell either: it turns up in ↗Herborist's T'ai Chi mask and equally in ↗La Mer's cleansing foam, which should tell you something about how the ingredient is being used commercially.

Snow lotus — Saussurea involucrata — has essentially in vitro data only, and carries a second problem the evidence discussion tends to omit: it is a scarce high-altitude plant, so a sourcing question sits alongside the efficacy one.

Tremella, the snow mushroom, is the case where the modest claim is the true one. It is a polysaccharide humectant, and the human work that exists is cosmetic-science-scale: a 2021 study in 20 volunteers found a gel containing 10% snow mushroom extract significantly out-performed its placebo on skin moisturisation across a three-hour assessment window. That is a real result for a real, narrow claim. Tremella is a decent humectant — you will find it doing that job in products like ↗Glow Recipe's Plum Plump Hyaluronic Cream — and nothing establishes it as an anti-ageing active, which is how it is usually sold.

Ganoderma (reishi) sits where snow lotus does: antioxidant and polysaccharide activity in vitro, no meaningful topical human data.

And a cross-tradition note, since it comes up in the same conversations: ↗turmeric and curcumin have a far larger literature than any of the above and a notorious problem underneath it — curcumin's bioavailability is poor enough that positive in vitro findings translate badly, and topically it stains. A big literature is not the same as a usable ingredient.

Standardisation: why two 'ginseng' products are not the same input

There is a structural problem beneath every grade above, and it is the reason botanical evidence is so much harder to apply than molecular evidence.

Niacinamide is niacinamide. It has one structure, one molecular weight, and a percentage on a label means something. A botanical extract is not a molecule — it is a variable mixture of hundreds of compounds whose composition depends on the species used (Panax ginseng, Panax quinquefolius and Panax notoginseng are different plants with different ginsenoside profiles), the part of the plant, the age at harvest, the growing region and soil, the processing (red ginseng differs from white ginseng by a steaming step that chemically alters the ginsenosides — the trial above used red), the extraction solvent, the extraction ratio, and whether the extract was standardised to a marker compound at all.

An INCI list captures none of this. 'Panax Ginseng Root Extract' is compatible with a standardised, characterised extract at a meaningful concentration and with a trace of a cheap commodity extract included so the word can appear on the box. Both are the same six words.

So even where a trial is positive and the route is right, transferring the result to a product requires that the product contains something resembling the tested material — and in almost every case, you cannot tell whether it does. The ingredient decoder will resolve what an INCI entry is and where it comes from; what it cannot do is tell you the concentration or the standardisation, because the label does not contain that information and no regulation requires it to.

How to actually use this

The practical conclusion is less exciting than either camp would like, which is usually a sign it is right.

Buy the product for its backbone. Choose on the components whose evidence and dose you can actually assess — the retinoid, the vitamin C, the ceramides, the UV filters, the acids — and treat a heritage botanical as a bonus if it is present. A well-constructed Chinese moisturiser with ceramides and niacinamide is a good moisturiser for the same reasons a French one is, and the ginseng in it is not what is doing the work.

Be sceptical where the botanical is the pitch and the price. If the story on the box is the plant, and the plant is the reason the product costs what it costs, that is the moment to check what the evidence actually is and what was administered in it.

And apply that scepticism symmetrically, which is the entire point of this audit. Pearl powder in a luxury Western cleansing foam is the same claim structure as pearl powder in a Chinese mask, resting on the same preclinical file, and it should get the same grade in both places. If your reaction to a heritage botanical changes depending on which country's brand is selling it, the thing being evaluated is not the ingredient.

Finally: thin evidence is not disproof. Almost nothing in the weak tier of this article has been tested and failed. It has mostly not been tested — which is a statement about research funding and commercial incentives, not about the plants. Snow lotus may well turn out to contain something worth having. Somebody has to run the trial first, and until they do, the correct price to pay for the possibility is low.