Two papers, from 1996 and 2002, name fisetin as an inhibitor of thyroid enzymes. Neither is obscure, both are indexed on PubMed, and neither appears in any fisetin marketing material. If you go looking for thyroid safety information on fisetin you will mostly find silence, which is not the same as reassurance.
The findings are real. The interpretation is where the work is, and it comes down to arithmetic that most coverage of dietary flavonoids and the thyroid skips entirely.
Key Takeaways
- A 1996 structure-activity study of 13 dietary flavonoids found most were potent thyroid peroxidase inhibitors, with IC50 values across the panel spanning 0.6 to 41 micromolar, and named fisetin among the more potent compounds [1].
- For fisetin, kaempferol, naringenin and quercetin, the inhibition was consistent with mechanism-based inactivation, the kind that is reversed only by the cell building new enzyme [1].
- In a 2002 study of thyroid type 1 deiodinase, fisetin was one of the weakest inhibitors tested at 70 micromolar, against 11 for baicalein and 13 for quercetin [2].
- A 1000 mg oral dose of unformulated fisetin produced a peak human plasma concentration of 9.97 ng/ml, which is roughly 0.035 micromolar [3], about 2,000-fold below the deiodinase figure.
- No study has ever measured TSH, free T4 or thyroid antibodies in humans taking a fisetin supplement. The concentration gap is an argument, not a clearance.
What the 1996 study found
Thyroid peroxidase, usually abbreviated TPO, is the enzyme that catalyzes thyroid hormone biosynthesis. It iodinates tyrosine residues on thyroglobulin, and without it the thyroid cannot make T4. It is also the antigen that anti-TPO antibodies target in Hashimoto’s thyroiditis, which is why it appears on thyroid panels.
A 1996 paper in Chemical Research in Toxicology screened 13 commonly consumed flavonoids for TPO inhibition. Most were potent, with IC50 values across the whole panel ranging from 0.6 to 41 micromolar. Fisetin was grouped with kaempferol, naringenin and quercetin as the more potent compounds, all of which share a resorcinol moiety in their structure [1].
That structural detail carries the mechanistic conclusion. Resorcinol and its derivatives are known mechanism-based inactivators of TPO, sometimes called suicide substrates: the enzyme processes the compound and is permanently disabled in the act. The authors were explicit about why this distinction matters, writing that suicide substrate inhibition in vivo would be reversed only by de novo protein synthesis, and would therefore be long-lasting, whereas reversible inhibitors and alternate substrates produce temporary effects that metabolism and excretion clear [1].
The abstract does not report fisetin’s individual IC50, only the range across the panel and fisetin’s membership in the potent group, so the precise number for fisetin cannot be quoted from it.
The deiodinase result runs the other way
Type 1 iodothyronine deiodinase, D1, converts T4 into the active hormone T3 and clears reverse T3. A 2002 study in Food and Chemical Toxicology measured D1 activity in murine thyroid microsomes against a panel of flavonoids and reported individual IC50 values [2].
All of them inhibited D1, but the potencies differed by roughly sevenfold across the set: baicalein at 11 micromolar, quercetin at 13, catechin at 17, morin at 55, rutin at 68, fisetin at 70, kaempferol at 72 and biochanin A at 77 [2].
Fisetin was second from the bottom. On this enzyme it is a weaker inhibitor than green tea catechin and roughly five times weaker than quercetin, a compound sold in far larger volumes and frequently stacked with fisetin in senolytic protocols. That is a genuinely useful piece of context, and it inverts the intuition that a compound named in a thyroid inhibition paper must be the worrying one in the cabinet.
The concentration gap
Every number above comes from an enzyme assay: purified or microsomal enzyme, a known concentration of flavonoid, a controlled buffer. The question that determines whether any of it applies to a person is what concentration a swallowed capsule actually produces.
The first human pharmacokinetic study of fisetin, published in 2022, answered that with a randomized double-blind crossover in fifteen healthy volunteers. A single 1000 mg dose of unformulated fisetin produced a peak plasma concentration of 9.97 ng/ml and an area under the curve over 12 hours of 12.67 [3]. At fisetin’s molecular weight of about 286, 9.97 ng/ml is approximately 0.035 micromolar.
Against the D1 figure of 70 micromolar, that is a gap of roughly 2,000-fold. Against the low end of the TPO range, 0.6 micromolar, it is roughly 17-fold. Both gaps point the same way.
The same study also tested an encapsulated hydrogel formulation delivering 192 mg of fisetin, which reached a peak of 238.2 ng/ml, more than 23 times higher, with a 26.9-fold larger area under the curve [3]. That works out to roughly 0.83 micromolar, still far below the deiodinase figure but no longer below the bottom of the TPO range. Enhanced-bioavailability products are the ones where this question is least settled, which is the opposite of how they are usually marketed.
Where the reassurance stops
Four things keep this from being a closed question.
Plasma concentration is not thyroid tissue concentration. The thyroid actively concentrates iodide and is a metabolically distinctive organ, and TPO sits at the apical membrane inside the follicular cell rather than in circulation. Neither paper measured what fraction of a circulating flavonoid reaches that compartment.
Mechanism-based inactivation accumulates in a way reversible inhibition does not. If an enzyme molecule is permanently disabled each time it processes fisetin, then a low concentration sustained daily for months is a different exposure from a single blood draw at peak, and the peak figure understates it.
Nobody has measured the outcome that matters. No study has reported TSH, free T4, free T3 or thyroid antibodies in humans taking fisetin. The entire safety argument here is inferential, built from enzyme assays and a pharmacokinetic curve rather than from a thyroid panel.
And the 1996 authors themselves raised the long-horizon concern, noting that the central role of hormonal regulation in thyroid growth suggests chronic consumption of flavonoids, especially suicide substrates, could play a role in the etiology of thyroid disease [1]. That is a hypothesis rather than a finding, and it is theirs rather than an inference added here.
The dietary baseline nobody mentions
Fisetin occurs naturally in strawberries, apples, persimmons and onions, and the flavonoids in the 1996 panel are staples of ordinary eating: quercetin from onions and capers, catechin from tea, naringenin from citrus, kaempferol from leafy greens. Anyone eating vegetables has a continuous low-level flavonoid exposure and always has. Our article on fisetin in food covers how little of it dietary intake actually supplies.
That baseline is the reason this is a question about supplemental doses rather than a reason to avoid fruit. Population-scale flavonoid consumption has not produced a visible thyroid signal, and the 1996 authors framed their own findings as predicting differences in hazard among compounds rather than a blanket dietary warning [1].
What to do with this if you have a thyroid condition
The reasonable move is monitoring rather than avoidance or ignorance. If you have Hashimoto’s, Graves’ disease, a history of thyroid surgery or radioiodine, or you take levothyroxine, that is a conversation to have with the clinician managing it before starting fisetin, and a reason to have a thyroid panel on file from before you start so a later one means something.
The pulse-dosing schedules used in senolytic protocols, covered in our article on fisetin cycling, involve two or three consecutive high-dose days separated by weeks off. Whether intermittent high exposure or continuous low exposure is the safer pattern for an enzyme subject to mechanism-based inactivation is not something the existing data can answer.
None of this is a reason for a person with normal thyroid function and no risk factors to panic about a strawberry, or about a standard fisetin dose. It is a reason for the subset of people with thyroid disease to treat this as an open question and get it checked rather than assume the silence in the marketing means the question was asked and answered.
Common questions
Does fisetin cause hypothyroidism?
No study has shown that in humans, because no study has measured it. Fisetin inhibits thyroid peroxidase and type 1 deiodinase in enzyme assays [1][2], at concentrations far above what a standard oral dose produces in plasma [3]. That is a reason for people with existing thyroid disease to monitor, not evidence of a clinical effect.
Is fisetin worse for the thyroid than quercetin?
On type 1 deiodinase, no. Fisetin’s IC50 was 70 micromolar against quercetin’s 13, making fisetin roughly five times weaker [2]. Both were listed among the more potent thyroid peroxidase inhibitors in the 1996 panel, which did not report their individual values in the abstract [1].
Should I stop fisetin before a thyroid blood test?
There is no evidence that fisetin interferes with thyroid assay chemistry, which is a separate question from whether it affects thyroid physiology, and neither has been studied. Tell whoever ordered the test what supplements you take, which is good practice regardless.
Do enhanced-absorption formulations change this?
They raise the exposure that the whole argument rests on. The formulated product in the human study reached roughly 0.83 micromolar at peak from a 192 mg fisetin dose [3], more than twenty times the unformulated figure. Nothing about that is known to be harmful, and it does mean the concentration-gap reasoning is thinnest exactly where the products are strongest.
References
- Inhibition of thyroid peroxidase by dietary flavonoids. Chem Res Toxicol (1996). PMID 8924586
- Inhibition of thyroid type 1 deiodinase activity by flavonoids. Food Chem Toxicol (2002). PMID 12065212
- Enhanced bioavailability and pharmacokinetics of a novel hybrid-hydrogel formulation of fisetin orally administered in healthy individuals: a randomised double-blinded comparative crossover study. J Nutr Sci (2022). PMID 36304817
These statements have not been evaluated by the Food and Drug Administration. This information is not intended to diagnose, treat, cure, or prevent any disease. Content is for informational purposes only and is not medical advice; consult a qualified healthcare provider before starting any supplement. As an Amazon Associate we earn from qualifying purchases.
