Fisetin 98% (HPLC) · CAS 528-48-3

Fisetin (CAS 528-48-3; 3,3′,4′,7-tetrahydroxyflavone) is a naturally occurring flavonol. The molecule itself has been known since the nineteenth century and is firmly in the public domain, so nothing about the structure differentiates one supply from another. What differentiates them is the plant source, the purification route, and the analytical package that travels with the batch.

Two plant sources dominate commercially. In China the material is extracted mainly from the heartwood and stems of the smoke tree, Cotinus coggygria Scop., in which fisetin is a minor constituent: a granted Chinese process patent (CN116023358B) states that fisetin accounts for roughly 1% of the raw material, with fustin and sulfuretin at 0.2–0.5% each. Independent phytochemical work on wild-grown C. coggygria heartwood found an ethanol extract composed of fustin 22.1 ± 0.8%, sulfuretin 14.4 ± 0.4%, butin 9.3 ± 0.4%, fisetin 5.0 ± 0.4%, taxifolin 2.0 ± 0.3% and butein 0.8 ± 0.3%. In Korea and Japan the source is Rhus verniciflua or R. succedanea — the same family as poison ivy — which has to be detoxified: the Korean requirement for detoxified R. verniciflua extract is fustin above 13.0%, fisetin above 2.0%, and urushiol not detected.

Two consequences follow, and both are supply-chain facts rather than marketing. First, fisetin's nearest neighbour in the raw material is fustin, its dihydro analogue, which has almost the same polarity and UV behaviour. That single fact is why the cost of a batch climbs steeply between 90% and 98% purity, and why a semi-synthetic route that dehydrogenates the abundant fustin into fisetin is commercially attractive (US8859014B2 covers exactly that conversion). Second, Anacardiaceae raw material carries its own laccase and polyphenol oxidase, and the granted process patent above is explicit that this enzyme degrades phenolic flavonoids in exposed material — the loss happens during comminution, drying and storage, before extraction even begins. That is why the same raw material yields very differently at different plants, and it is one of the first questions we put to a supplier.

There is no pharmacopoeial monograph for fisetin. Our review found no fisetin monograph in USP, EP, ChP or JP. In practice that means the specification is contractual: whatever your QA and our technical team agree, written down, with the test methods attached. We would rather say that plainly than let a buyer assume a compendial standard exists behind the number 98.

The best available human pharmacokinetic study is a randomised, double-blind crossover trial in 15 healthy volunteers (J Nutr Sci 2022;11:e82). A 1,000 mg dose containing 982 mg of unformulated fisetin produced a plasma C_max of 9.97 ng/mL and an AUC_0–12 of 12.67 ng·h/mL, with fisetin quantifiable for only about two hours (T_max 0.88 h, t½ 1.14 h). A hybrid-hydrogel formulation containing 192 mg of fisetin — roughly a fifth of the dose — produced C_max 238.2 ng/mL and AUC_0–12 341.4 ng·h/mL, that is 23.9× and 26.9× higher on a fifth of the dose. The authors also record that below a 1,000 mg dose, fisetin levels were too low to quantify at all.

Two further pieces come from animal work. In rats dosed 30 mg/kg intravenously, the plasma AUC ratio of free fisetin to its glucuronide to its sulfate was 1:6:21, and in bile 1:4:75, with P-glycoprotein-mediated biliary excretion (J Agric Food Chem 2018). In mice, clearance is biphasic and a methylated metabolite, geraldol, is formed (Biochem Pharmacol 2011). Reported aqueous solubility figures diverge by more than an order of magnitude — 10.45 µg/mL in one source, 0.151 mg/mL in another — and we quote both rather than pick the flattering one. logP is around 3.2 and fisetin is generally classified BCS IV.

Put the two halves together and the arithmetic is uncomfortable but useful: 9.97 ng/mL is about 35 nM, while in vitro work in this field typically uses 5–20 µM, two to three orders of magnitude higher. We publish this because it changes what a buyer should ask for. Unformulated 98% powder is the correct, honest starting material for a formulator who intends to build a delivery system around it — solubilisation, cyclodextrin complexation, lipid or hydrogel systems — and it is not a shortcut to exposure. Where a supplier implies otherwise, the primary literature is one click away in the reference list below. That is the entire point of linking it.

Fisetin is an easy compound to assay badly. The two failure modes we see are co-elution with structurally adjacent flavonoids and measuring the parent compound alone when the conjugates carry most of the exposure. Both are avoidable, and both belong in the method you agree with us before the first batch ships.

The research literature on fisetin is active and genuinely contested, and we summarise both sides because your regulatory and R&D teams need both. Preclinical work published in eBioMedicine in 2018 put fisetin forward as a senotherapeutic candidate in mice. A 2017 paper from the same field found that fisetin behaved as a senolytic in senescent HUVECs but not in senescent IMR90 human lung fibroblasts or primary human preadipocytes — cell-type specificity that has not been resolved since. A 2025 mouse study reported effects from intermittent oral dosing in aged but not young animals. A 2024 review concludes that the mechanistic case is strengthening while human tissue exposure still has to be demonstrated.

On the clinical side, the largest completed randomised trial — knee osteoarthritis, 74 patients randomised, 20 mg/kg for two days repeated after 28 days — reported no evidence of significant benefit for pain, function or cartilage health, with no significant safety signal (OARSI 2025). A dedicated human pharmacokinetic and food-effect study is under way at the time of writing. We make no claims in either direction: we supply a characterised raw material, and we would rather hand you the negative result than let you find it after you have built a brand on the positive ones.