Luteolin is the flavonoid you meet after you have already read about quercetin and fisetin. It arrives with a specific and genuinely interesting pitch: that it calms mast cells and microglia — the two cell types most often blamed for the diffuse “inflamed brain” complaints of brain fog, long COVID, mast cell activation syndrome and allergy. That mechanism is real, it is well documented, and it has been worked on for two decades. The question this page answers is a different one: has anyone shown that swallowing luteolin does anything to a person? The answer is uncomfortable, and almost nobody selling luteolin says it out loud.
What luteolin actually is, and why the flavone/flavonol distinction matters less than you’d think
Luteolin is a flavone: 5,7,3′,4′-tetrahydroxyflavone. Quercetin and fisetin are flavonols, which carry an additional hydroxyl group at the 3-position of the C ring. Put concretely, luteolin is quercetin with one hydroxyl removed, and apigenin is luteolin with one more removed. These are near neighbours, not different classes of thing.
Supplement copy sometimes leans hard on that 3-hydroxyl, as if it explained why luteolin is the “right” flavonoid. It doesn’t, in either direction. It matters in specific assays — when a senolytic screen ranked ten flavonoids, the flavonol fisetin was the clear winner and luteolin showed only weak activity[18] — and it is irrelevant in others, where luteolin and quercetin behave almost identically. Structure tells you where to look. It does not tell you what will happen in a person.
Dietarily, luteolin comes from celery, parsley, thyme, green and sweet peppers, artichoke, perilla and chamomile tea. It is a minor part of what anyone eats. In a food-frequency analysis of nearly 117,000 US health professionals, total flavonol plus flavone intake averaged roughly 20–22 mg per day, of which quercetin alone contributed about three-quarters — leaving luteolin and apigenin sharing a small remainder.[2] A 100–300 mg capsule is therefore not “more celery.” It is a dose two orders of magnitude above habitual intake, and it should be judged like one.
The mast-cell story: the best-supported part of the file
Mast cells are the immune cells that sit in tissue — skin, gut, airway, and around blood vessels in the brain — and release histamine, tryptase, cytokines and vascular mediators when triggered. The clinical problem is that there is essentially one licensed “mast cell blocker,” disodium cromoglycate (cromolyn), and it is poorly absorbed orally and a weak inhibitor of histamine release from human mast cells.[4] That gap is why luteolin gets attention at all.
The head-to-head is direct. In a cultured human mast cell line stimulated through IgE or through IL-33, equimolar luteolin was significantly more potent than cromolyn at inhibiting release of histamine, tryptase, MMP-9 and VEGF — and luteolin additionally suppressed IL-1β, IL-6, CXCL8 and TNF, where cromolyn had no effect at all.[4] Separate work shows luteolin blocking both the classical allergic route (FcεRI) and the pseudo-allergic/drug-triggered route (MRGPRX2) by damping calcium signalling.[5] Two different activation pathways, one molecule, consistent results.
That is a real finding and it is why people with mast cell activation syndrome, chronic urticaria or histamine-intolerance symptoms keep landing on luteolin. It is also, precisely, a result in a dish. No trial has given luteolin to people with mast cell activation syndrome and measured whether they felt better.
Microglia, neuroinflammation and “brain fog”
Microglia are the brain’s resident immune cells, and they drift toward a persistent pro-inflammatory state with age. Luteolin’s neuro-story rests on interrupting that. The cleanest animal demonstration: aged mice fed a luteolin-supplemented diet for four weeks. In aged mice on control chow, roughly 23–25% of isolated microglia stained positive for MHC class II, IL-1β and IL-6, versus under 3% in adult mice — and dietary luteolin cut that age-associated activation back down.[7] Reviews of luteolin in neuroinflammation and neurotrauma catalogue the same pattern across models: reduced glial activation, reduced cytokine output, preserved neurons.[6]
This is a coherent, mechanistically satisfying story and it is the engine behind every “luteolin for brain fog” post you have read. It is also entirely rodent and cell work. The gap between “luteolin reduces microglial MHC II staining in a 24-month-old mouse” and “luteolin clears your brain fog” is the whole article.
The human evidence — and the attribution problem that swallows most of it
Here is where luteolin diverges sharply from quercetin, which at least has a meta-analysis of its own for blood pressure. Nearly every human trial that gets cited for luteolin gave people luteolin plus something else.
It is worth walking through the three biggest ones, because each is more interesting than its headline.
Autism spectrum disorder. The study that put luteolin on the map for parents was a 26-week open-label trial in 50 children aged 4–10, using a formulation of 100 mg luteolin (from chamomile), 70 mg quercetin and 30 mg rutin per capsule in olive kernel oil, dosed at one capsule per 10 kg of body weight. Forty completed. Adaptive-behaviour age-equivalent scores rose by 8.4 months in communication, 7.2 in daily living skills and 8.0 in the social domain (P < 0.005), with 26.6–34.8% reductions on aberrant-behaviour subscales.[8] Read that carefully: the trial ran for six months and reported gains of roughly eight months of age-equivalent development — a margin over ordinary maturation that is small enough to be explained by test–retest practice effects and parental expectancy in an unblinded study with no control group. The paper also reports transient increased irritability in 27 of 50 participants, which is a notable tolerability signal in children. A companion analysis of the same cohort found serum IL-6 and TNF fell significantly over the 26 weeks, most in the children whose behaviour improved most[9] — suggestive, and equally uncontrolled.
Post-viral smell loss. A multicentre double-blind trial randomised 185 patients with persistent post-COVID olfactory dysfunction to PEA-LUT plus olfactory training or training plus placebo, reporting 92% improved versus 42% and a mean TDI recovery of 12.8 versus 3.2.[12] Impressive — but note the arms were 130 versus 55, an unusual split for a randomised trial. When Dresden’s smell and taste clinic ran an independent version in 50 patients, the PEA-LUT group did improve significantly on discrimination and total score while controls did not, yet when the analysis asked about clinically meaningful improvement, the difference between groups disappeared.[14] Two results, one enthusiastic and one sober, and the sober one came from the group that defines olfactory testing.
Cardiometabolic. A six-month placebo-controlled trial of Altilix — a Cynara cardunculus extract of chlorogenic acid and luteolin, 150 mg/day — in 100 people with metabolic syndrome reported weight −2.4%, waist −2.8%, HbA1c −0.95% and fatty liver index −21.8%.[16] It also reported carotid intima-media thickness falling 39.5% in six months, a magnitude that is not biologically plausible for an arterial wall and should make you discount the whole readout rather than celebrate it. Chlorogenic acid is also the better-studied of the two ingredients.
Exercise. Twelve physically active men in a double-blind crossover took mango-leaf mangiferin combined with a peanut-husk extract standardised to 95% luteolin, at 50 or 100 mg of luteolin. Sprint performance, muscle oxygen extraction and brain oxygenation all improved.[17] The study was part-financed by the ingredient manufacturer, which also participated in the experimental design, and mangiferin was the co-active. Twelve men, two ingredients, one sponsor.
What has actually been tested in humans, and as what
| Claim you'll see | What was actually given to people | Verdict |
|---|---|---|
| Mast-cell stabilisation / MCAS | Nothing — human data is cultured mast cells only | Untested in people |
| Brain fog, neuroinflammation | PEA 700 mg + luteolin 70 mg (10:1) | The combination works; luteolin's share unknown |
| Post-COVID smell loss | PEA-LUT + olfactory training | Real pooled effect; independent replication found no clinically meaningful gap |
| Autism symptoms | Luteolin + quercetin + rutin, open-label, no control | Uninterpretable as a luteolin effect |
| Metabolic syndrome / liver | Chlorogenic acid + luteolin extract | Combination only; one implausible endpoint |
| Exercise performance | Mangiferin + luteolin, n = 12, sponsor-funded | Combination only |
| Senolytic / anti-aging | Nothing in humans | Weak even in the cell screen |
| Symptom relief, luteolin alone vs placebo | Luteolin alone, crossover, 21 veterans | Negative at both doses |
| Prostate cancer under surveillance | Luteolin 50 mg/day alone, n = 5, single arm | No control group; 2 of 5 progressed |
The two times luteolin was given on its own
Both are small, and neither helps the sales pitch.
The first is the only published placebo-controlled test of luteolin by itself we could find. As part of an NIH-funded screening programme for Gulf War Illness, 21 male veterans each ran a month of baseline, a month of placebo, a month of a lower dose and a month of a higher dose, for up to three different botanicals. Resveratrol beat placebo on symptom severity at both doses (P = 0.035 and P = 0.004). Luteolin did not, at either dose (P = 0.718 and P = 0.492), and neither did fisetin.[11] It is a screening study, not a definitive trial; the design was pseudo-randomised and the sample tiny. But it is the closest thing that exists to a clean question, and the answer came back no — in a design where a positive answer was clearly attainable, because another arm produced one.
The second is a single-arm phase I study: five men with low-to-intermediate-risk prostate cancer on active surveillance took 50 mg of oral luteolin daily for six months. No adverse events were recorded, and on protocol biopsy two patients showed a favourable response, one was stable, and two progressed and went on to radical surgery.[19] The paper concludes luteolin is “safe and effective with regard to oncological outcomes.” With no control arm, five patients and a 40% progression rate, that conclusion is unsupported by its own data. The useful content is the safety observation, not the efficacy claim.
Bioavailability, and why “enhanced absorption” claims need a receipt
Luteolin has the standard flavonoid absorption problem, and it was characterised early. Passing through intestinal mucosa, luteolin is converted to glucuronides; the main conjugate circulating in blood is the monoglucuronide of the unchanged aglycone, with free luteolin and its monoglucuronide both detectable in human serum after ingestion.[3] So some luteolin does escape conjugation — but what mostly circulates is a metabolite, not the molecule that was so impressive against cromolyn in a dish.
That same 1998 work is, quietly, the intellectual root of the “enhanced absorption” market: luteolin dissolved in propylene glycol was absorbed faster than luteolin in carboxymethyl cellulose, in rats.[3] Vehicle matters. That is why the autism formulation was suspended in olive kernel oil,[8] and why liposomal luteolin is now sold at a premium — the mast-cell paper itself closes by recommending luteolin “especially in liposomal form to increase oral absorption.”[4]
We looked specifically for the study that would justify the premium: a human pharmacokinetic comparison of liposomal or otherwise “enhanced” luteolin against plain luteolin. We could not find one. The nanocarrier and liposome literature for luteolin is cell-line and tumour-model work, and reviews covering luteolin’s pharmacokinetics and clinical trials describe the delivery field as a pharmaceutical research programme rather than a solved problem.[1] Until someone publishes plasma curves in people, “10× absorption” on a liposomal luteolin label is a formulation hypothesis with a price attached.
Is luteolin a senolytic? Not really
Luteolin gets pulled into senolytic conversations by proximity — it is a flavonoid, and two flavonoids (fisetin, quercetin) are the field’s stars. The screening study that made fisetin famous tested a panel of ten flavonoid polyphenols against senescent mouse and human fibroblasts. Fisetin was the most potent; luteolin showed only weak activity at a dose where quercetin was ineffective, and it was fisetin — not luteolin — that went on to reduce senescence markers across tissues and extend median and maximum lifespan in wild-type mice dosed late in life.[18] Luteolin has never been shown to clear senescent cells in a human being.
Grade it honestly: weak, and weaker than the alternatives in its own family. If senescent-cell clearance is what you are after, luteolin is not the molecule, and even fisetin’s human case is thinner than its reputation.
Metabolic and other signals
Low-dose dietary luteolin improved diet-induced obesity and insulin resistance in mice,[20] and reviews list antioxidant, anti-inflammatory and anticancer activity across a very large preclinical catalogue.[1] This is the “does everything in a dish” profile shared by most dietary flavonoids, and it is the least informative part of any flavonoid file. The volume of preclinical papers is not a volume of human proof; if anything, a compound with 5,000 positive cell papers and one negative human crossover should make you more suspicious of the cell papers, not less.
Dose, forms and cost
Commercial luteolin capsules are typically 100–300 mg per day. The doses actually used in human research bracket that range widely and were chosen with little to go on: 50 mg/day in the prostate study,[19] 50–100 mg in the exercise crossover,[17] 70 mg inside the PEA-LUT sachet,[13] around 100 mg per capsule in the autism formulation (so roughly 300 mg/day for a 30 kg child),[8] and 300 mg twice daily in the largest randomised trial of luteolin alone.[23] There is no dose–response curve in humans for any outcome, so nobody — including us — can tell you the right dose.
On forms: most supplement luteolin is a botanical extract, and the source varies more than labels suggest. Published trials have used luteolin from chamomile,[8] from Chrysanthemum at 98% purity,[24] and from peanut husk at 95% purity[17] — the last of which is worth a direct question to the manufacturer if you have a peanut allergy, because the botanical origin is rarely printed on the bottle. Plain luteolin powder is a commodity and prices like one; the branded combination products — PEA-LUT sachets, the olive-oil-suspended flavonoid blends, liposomal versions — cost substantially more per month, and as shown above the human evidence attaches to the combination rather than to the luteolin in it.
Safety and the interaction surface
Luteolin is a normal component of food and short-term human exposure looks benign: no adverse events in six months at 50 mg/day,[19] no major adverse effects over 26 weeks in the paediatric formulation study — though transient irritability in over half of participants is not nothing.[8] There is no long-term human safety data at supplement doses, in anyone.
- CYP inhibition — real in vitro, probably not at achievable doses. Luteolin inhibited eight human CYP isoforms in liver microsomes, most potently CYP1A2 and CYP2C8, and inhibited acetaminophen sulfation but not glucuronidation. But the authors’ own conclusion is the useful part: the IC₅₀ values exceed customary in-vivo human exposure at tolerable supplemental doses.[21] Poor bioavailability cuts both ways — it limits the benefit and it limits the interaction risk. That said, this was a microsome study, not a clinical interaction study, and gut-wall concentrations after a 300 mg capsule are higher than plasma concentrations.
- Thyroid peroxidase — a class-level flag, not a luteolin finding. A structure–activity study of 13 commonly consumed flavonoids found most were potent inhibitors of thyroid peroxidase, the enzyme that makes thyroid hormone, with IC₅₀ values from 0.6 to 41 µM; the most potent — fisetin, kaempferol, naringenin, quercetin — acted by mechanism-based (suicide) inactivation, which in vivo would be long-lasting.[22] Note who those four are: flavonols and a flavanone, not luteolin. This is a reason for anyone with thyroid disease or on levothyroxine to mention high-dose flavonoid supplements to their prescriber, and it is not evidence that luteolin suppresses thyroid function in people. No human study has tested that.
- Pregnancy, breastfeeding, children. No adequate data at supplement doses. The only paediatric exposure of any size is the open-label autism formulation, which is a combination product and reported irritability in more than half of participants.[8] Not a setting for self-experimentation.
- The regulatory baseline. Luteolin is a dietary supplement ingredient with no FDA approval for any indication. It is not a treatment for autism, long COVID, mast cell activation syndrome, allergy or cancer.
Who this plausibly makes sense for — and what would change our mind
We can only reason from mechanism here, and we will say so rather than dress it up.
- Histamine- or mast-cell-driven symptoms, as an experiment. This is the only use where the mechanism is specific enough to be worth a personal trial. Luteolin outperformed the licensed mast-cell blocker across every mediator in cultured human mast cells and hit two independent activation pathways.[4][5] Nobody has tested that in patients. If you try it, treat it as an n-of-1 with a defined stop date, not a therapy.
- Post-viral smell loss. The evidence that exists is for PEA-LUT alongside olfactory training, and the independent replication found the added benefit small.[14][15] If you want the intervention that was actually studied, buy the combination product; buying plain luteolin is not the same trial.
- Brain fog, longevity, senescence. No. The senolytic claim is weak in luteolin’s own family screen,[18] and the neuroinflammation case stops at rodents.[6][7] If the goal is a flavonoid with at least one human meta-analysis behind it, quercetin is the better-documented choice, modest as its effect is.
What would change our mind. Not another mouse study, and not another combination product. Two specific readouts already exist and are simply unreported. A quadruple-masked, randomised, placebo-controlled trial of luteolin alone at 300 mg twice daily for 12 weeks in 85 people with schizophrenia — measuring global psychopathology, cognition and oxidative stress — completed in January 2026 with no results posted.[23] A randomised, double-blind, placebo-controlled crossover trial of 250 mg of luteolin twice daily for two weeks on memory in 44 healthy adults completed in May 2024, also with no results posted and no publication.[24] Those are the first properly controlled tests of luteolin on its own at a realistic supplement dose. When they report, this page changes — in whichever direction they point. Until then, the honest position is that luteolin is a mechanistically excellent candidate that nobody has yet shown to work in a person.
This article is research information, not medical advice. Luteolin is sold as a dietary supplement and has no FDA approval for any indication; it is not a treatment for autism, long COVID, mast cell activation syndrome, allergy or cancer, and nothing here should be used to delay or replace evaluation for those conditions. The human evidence described above is overwhelmingly from combination products, and no long-term safety data exists for luteolin at supplement doses in anyone. If you take thyroid medication, take drugs metabolised by CYP1A2 or CYP2C8, are pregnant or breastfeeding, or are considering giving a flavonoid supplement to a child, talk to a licensed clinician first.