Agrimony Safety: Iron, Interactions and Dose

Pale small red cells seen in iron deficiency
Pale small red cells seen in iron deficiency
Plant foods rich in non-haem iron, with vitamin C
Plant foods rich in non-haem iron, with vitamin C
Agrimony seed burrs, a key identifying feature
Agrimony seed burrs, a key identifying feature

This is the most practically useful page in the agrimony set, and it is worth saying why up front. Agrimony's benefits are traditional and untested. Its main interaction is established in humans — not for agrimony specifically, but for the class of compounds that gives agrimony its entire character. That asymmetry is the single most important fact about drinking this herb: the best-evidenced thing agrimony does is reduce your absorption of iron and of anything else in your stomach at the time.

That is not an argument against the herb. It is an argument for knowing one thing about timing: in a controlled trial with tea, drinking the tea one hour after an iron-containing meal roughly halved its inhibitory effect on iron absorption, compared with drinking it with the meal (Ahmad Fuzi 2017). The familiar working rule for astringent infusions such as agrimony is a gap of roughly a couple of hours either side of meals, iron supplements and medicines. The rest of this page explains where it comes from, who it matters most for, and what the other cautions are actually worth.


Table of Contents

  1. One Mechanism, Two Columns
  2. Tannins and Non-Haem Iron
  3. Who This Matters Most For
  4. The Timing Fix, and Where It Comes From
  5. Vitamin C as a Partial Counter
  6. Medicines Taken at the Same Time
  7. Gastric Irritation and Constipation
  8. Blood Sugar and Diabetes Medication
  9. Bleeding, Vitamin K and the Anticoagulant Question
  10. Pregnancy and Breastfeeding
  11. Children, and When Diarrhoea Needs a Doctor
  12. Photosensitivity, Allergy and Identity
  13. Numbers We Will Not Give You
  14. Dose and Duration, As Far As They Can Be Stated
  15. The Honest Bottom Line
  16. Key Research Papers
  17. Connections
  18. Featured Videos

One Mechanism, Two Columns

Agrimony's tannins bind proteins and chelate metal ions. That is the whole story, and it produces the benefit column and the hazard column simultaneously — they are not two properties, they are one property described twice.

Benefit column. Tannins precipitate proteins on an inflamed mucous membrane, forming a transient coating that may reduce weeping and blunt the raw feeling. Local, immediate, and the basis of the traditional diarrhoea and gargle uses.

Hazard column. The same binding is indiscriminate. Tannins bind dietary iron in the gut lumen and make it unavailable. They bind proteins and some drug molecules and can reduce their absorption. In quantity they irritate the gastric mucosa they were supposed to soothe. And because large hydrolysable tannins are poorly absorbed, they persist in the lumen doing exactly this — which is the awkward corollary of the argument that makes the mucosal benefit plausible in the first place. The part of agrimony that survives digestion best is the part that causes the interaction.

You cannot keep one column and discard the other by choosing a better product or a gentler dose. Anything astringent enough to be worth drinking for diarrhoea is astringent enough to bind iron. The only lever available is timing, which is why timing gets its own section below and why it is the point this page leads with.

Tannins and Non-Haem Iron

Dietary iron comes in two forms. Haem iron from meat and fish is absorbed by a route that polyphenols barely touch. Non-haem iron — from plants, eggs, dairy, fortified foods and virtually all iron supplements — must be released, kept soluble and reduced before it can be taken up, and that pathway is highly sensitive to what else is in the gut at the time. Polyphenols with the right structure form insoluble complexes with iron in the lumen; the iron is still there, but it can no longer be absorbed.

This is one of the better-established facts in nutritional science, and it was worked out with tea, which is chemically the closest everyday relative of an astringent herbal infusion. Disler and colleagues published The effect of tea on iron absorption in Gut in 1975, showing that tea reduced the absorption of non-haem iron from iron solutions, from bread and from a meal of rice with potato and onion soup, whether or not vitamin C (ascorbic acid) was present. The authors put this down to insoluble iron–tannate complexes forming in the gut. Iron from cooked haemoglobin (haem iron, the form in meat and blood) was not inhibited. Brune, Rossander and Hallberg examined which phenolic structures matter in European Journal of Clinical Nutrition in 1989, establishing that the inhibitory effect is structure-dependent rather than a generic property of "plant chemicals" — galloyl groups in particular. Hurrell, Reddy and Cook published Inhibition of non-haem iron absorption in man by polyphenolic-containing beverages in the British Journal of Nutrition in 1999, extending it across a range of drinks. In that radio-iron study of a bread meal, herb teas were potent inhibitors as well as black tea: compared with water, peppermint tea lowered iron absorption by 84%, pennyroyal by 73%, vervain by 59%, lime flower by 52% and camomile by 47% (black tea 79–94%). Drinks with 20–50 mg of total polyphenols per serving cut absorption by 50–70%, and adding milk made little or no difference. Agrimony itself was not tested. Chung and colleagues' review Tannins and human health in Critical Reviews in Food Science and Nutrition in 1998 remains the standard overview of the class.

What transfers to agrimony and what does not. The direction of the effect transfers with high confidence: agrimony's defining constituents are exactly the galloyl-bearing hydrolysable tannins this literature identifies as inhibitory, and the mechanism is chemical rather than species-specific. Your own tongue confirms the relevant property is present. What does not transfer is the magnitude. Nobody has measured iron absorption from a meal taken with an agrimony infusion. Under the borrowed-evidence rules this site applies, this is a transfer we are willing to make and label, because the shared mechanism is the compound class itself rather than a taxonomic assumption — but the honest statement is agrimony reduces non-haem iron absorption by an unquantified amount, plausibly in the same territory as strong tea.

Who This Matters Most For

For a healthy adult with good iron stores drinking one cup of agrimony for two days of an upset stomach, this interaction is a footnote. For the following groups it is the main event, and for some of them the research raises concerns about using the herb at all:

Two points worth stating explicitly. First, the popular claim that agrimony "contains minerals including iron" and therefore supports iron status has the sign wrong: whatever trace iron a 2 g infusion contributes is negligible, while the same cup's tannins inhibit absorption from everything else consumed near it. Net, an astringent infusion is plausibly negative for iron status. The same arithmetic error appears in popular writing about red raspberry leaf and haritaki. Second, recurrent mouth ulcers — a traditional agrimony gargle indication — are a recognised sign of iron deficiency, which makes long-term astringent use for them a potential way to worsen the cause while treating the symptom.

The tests that measure iron status are ferritin plus a full blood count, with transferrin saturation where interpretation is difficult. See the iron panel.

The Timing Fix, and Where It Comes From

The interaction is luminal and simultaneous. Tannin and iron have to be in the same place at the same time to form a complex. Separate them in time and there is nothing to complex with.

The useful evidence here comes from tea again, and it is more precise than most interaction advice. Ahmad Fuzi and colleagues published a controlled stable-isotope study in the American Journal of Clinical Nutrition in 2017 in healthy UK women, testing whether leaving a one-hour interval between an iron-containing meal and a cup of tea attenuated the inhibition. It did. In 12 iron-replete women, iron absorption was 5.7% with water, 3.6% when tea was drunk with the meal, and 5.7% when the tea came one hour after the meal; the inhibitory effect of tea fell from 37.2% to 18.1%, about half. It is one small study in healthy young women. That is where the familiar "separate by an hour or two" advice comes from: not from folklore, but from a controlled measurement in humans of the exact question a reader is asking.

How the timing evidence is usually applied:

Vitamin C as a Partial Counter

Ascorbic acid keeps iron soluble and in its more absorbable reduced state, and it can override polyphenol and phytate inhibition to a considerable degree. Siegenberg and colleagues demonstrated in the American Journal of Clinical Nutrition in 1991 that ascorbic acid prevents the dose-dependent inhibitory effects of polyphenols and phytates on non-haem iron absorption.

That is a genuinely useful lever, described here accurately rather than oversold:

Medicines Taken at the Same Time

The same binding chemistry applies to drug molecules. Tannins can complex with alkaloids, with basic and amine-containing drugs, and with proteins and peptides, and they can chelate metal-containing preparations — reducing the amount that reaches absorption. This is a class effect of astringent plant material, and it has been documented in general pharmacognosy for a long time.

Two honest qualifications, and then the common working rule.

Qualification one: this has not been studied for agrimony. There is no interaction study of agrimony with any named drug. The inference is from tannin chemistry, and it is a reasonable inference, but "no known interactions" appearing on a product label means no interaction study has been done, not that the herb has been checked and cleared. Those are opposite meanings and the phrase reliably misleads.

Qualification two: the mechanism is timing-dependent, which makes it easy to manage. This is not a metabolic interaction of the kind that persists for days after a dose; it is a physical binding event in the gut lumen that stops mattering once the two substances are no longer together.

The common working rule is a gap of about two hours between agrimony and any oral medication. The concern is greatest where the drug has a narrow therapeutic window or where absorption is already fussy:

Gastric Irritation and Constipation

Concentrated tannin is a gastric irritant. Nausea, stomach discomfort and vomiting are the classic effects of overdoing an astringent, and they are dose-related. This produces the awkward situation described on the diarrhoea page: the substance being taken to settle a stomach is, at higher concentration, a reason for a stomach to be unsettled. Nobody has measured where the crossover sits, which means doubling the dose when a mild infusion is not working moves toward the irritant range rather than away from it.

Constipation is the expected consequence of overshooting an antidiarrhoeal astringent, and it is worth anticipating rather than being surprised by. It is also a reason the traditional use is short-term: an astringent taken continuously has nothing left to correct.

For anyone with active gastritis, peptic ulcer disease, reflux oesophagitis or an inflammatory bowel condition in flare, a strong astringent infusion is a particular concern — the mucosa is already compromised, and "tightening" it is not the same as healing it. For that setting the demulcent approach is the traditional alternative, and it is at least mechanistically gentler: see marshmallow root for the gut lining and slippery elm.

Blood Sugar and Diabetes Medication

This is the one place where an animal finding earns a genuine clinical caution, and it is worth being clear about why it belongs in the safety column rather than the benefits column.

Gray and Flatt reported actions of agrimony on hyperglycaemia, cellular glucose metabolism and insulin secretion in the British Journal of Nutrition in 1998, following earlier work by Swanston-Flatt and colleagues in Diabetologia in 1990 on traditional plant treatments in normal and streptozotocin-diabetic mice, in which agrimony reduced hyperglycaemia. This is among the better-characterised pharmacological actions attributed to the herb.

The asymmetry in how such data should be read is deliberate and defensible. As a benefit claim, mouse hypoglycaemic activity is worth very little: no human trial exists, no dose translation exists, and the poor absorption of the herb's dominant constituents makes a systemic mechanism hard to account for. As a safety signal, the same data is worth attending to, because the cost of a false alarm (an extra glucose check) is trivial while the cost of a missed additive hypoglycaemia in someone on insulin or a sulfonylurea is not. When the evidence is thin, the asymmetry of consequences decides which way to lean.

The data are from mice and cells only: agrimony added to the diet (62.5 g/kg) and drinking water (2.5 g/l) countered the hyperglycaemia of streptozotocin-diabetic mice, and a water extract stimulated insulin secretion 1.9–3.8-fold in a pancreatic beta-cell line (Gray and Flatt 1998). For people on insulin or other glucose-lowering medication, an added glucose-lowering effect is therefore a theoretical concern that has not been tested in humans. See type 2 diabetes.

The same reasoning applies, more weakly, to blood pressure. Preclinical work has suggested hypotensive activity; the human relevance is unknown; for people on blood-pressure medicines it is a point of awareness, with no human data either way.

Bleeding, Vitamin K and the Anticoagulant Question

This section contains a real contradiction and a figure we are declining to give, so it needs care.

The contradiction. In Traditional Chinese Medicine, agrimony — as Agrimonia pilosa, Xian He Cao — is classified among the haemostatic, "stop bleeding" herbs. Western caution lists warn about additive effects with anticoagulant and antiplatelet drugs, which is a warning about increased bleeding. These point in opposite directions and both circulate widely. Candidate reconciliations: the two claims may concern different species; a local contact effect on capillary oozing in an open wound is genuinely different from a systemic effect on the coagulation cascade; and the Western warning may be a category inference applied to polyphenol-rich plants generally rather than an observation about agrimony. We cannot resolve it and will not pretend to. Yarrow carries the identical contradiction, which suggests the problem is in how astringent and haemostatic traditions get translated into interaction lists rather than in either plant.

The vitamin K question. Herbal reference works commonly list vitamin K among agrimony's constituents, and this is offered as a theoretical reason for caution with warfarin, since vitamin K opposes warfarin's action. We are reporting that as a theoretical and unquantified consideration, and here is why we will not do more with it: we have not found a quantitative analysis of vitamin K1 content in Agrimonia eupatoria, and printing a figure on inference would be a guess. A green leafy plant containing some phylloquinone is unremarkable — so does spinach, and so does broccoli — and the relevant question is how much reaches you in a water infusion, phylloquinone being poorly water-soluble and largely bound in chloroplast membranes. The plausible answer is "not much," but plausible is not measured.

What the anticoagulation literature does establish is that for warfarin patients, consistency of vitamin K intake matters more than avoidance — erratic intake destabilises international normalized ratio (INR) control more reliably than a steady moderate intake does. That principle is more useful than any specific warning about agrimony: for someone on warfarin, steady use or no use is less disruptive to international normalized ratio control than on-and-off use.

Combining the strands: for someone on an anticoagulant or antiplatelet drug, agrimony introduces an unquantified variable in a setting where control depends on keeping variables few. Surgical teams commonly ask for herbal products to be stopped a couple of weeks before elective surgery, and that practice covers agrimony as much as any herb.

Pregnancy and Breastfeeding

Three independent reasons converge against medicinal doses of agrimony in pregnancy, which is unusual and makes the concern firm despite thin data:

  1. The iron interaction lands hardest here. Iron requirements rise substantially in pregnancy, deficiency is common, and supplementation is routine. A tannin infusion is working directly against that.
  2. Traditional sources themselves advise against medicinal amounts in pregnancy. Traditional caution is weak evidence for a benefit and reasonable evidence for a hazard, because the incentive runs the other way — a tradition has no reason to invent a restriction on its own remedy.
  3. There is no safety data. No reproductive toxicology, no human pregnancy exposure data, and no data in lactation. Under the traditional-use framework this herb is registered within, "acceptable safety record" refers to the population that historically used it — which is not a systematic study of pregnancy outcomes.

An occasional culinary-strength cup that has already happened is not the exposure these concerns are about; they concern regular medicinal use, tinctures and concentrated extracts. For breastfeeding, nothing is known: there are no data on agrimony in lactation.

Children, and When Diarrhoea Needs a Doctor

There are no paediatric data for agrimony at any dose. That absence is not academic, because childhood diarrhoea is precisely where diarrhoea is dangerous, and the danger is dehydration, not stool consistency.

This is the point at which the astringent story fails clinically rather than evidentially. An astringent, even working exactly as advertised, changes the appearance and frequency of the stool. It does nothing about the fluid and electrolytes already lost. A child who looks better because the stool has firmed while remaining dehydrated is a child whose most useful warning sign has been switched off. Oral rehydration solution addresses the thing that harms; an astringent addresses the thing that alarms.

Diarrhoea is treated as a matter for medical care rather than a herb when it comes with any of the following:

Photosensitivity, Allergy and Identity

Photosensitivity. Secondary herbal sources fairly consistently note increased skin sensitivity to sunlight with agrimony. We have not been able to trace this to a primary case series or a mechanistic study, and we are reporting it as it stands: a caution carried in the secondary literature whose original source we cannot verify. That is a weaker claim than "agrimony causes photosensitivity," and it is the accurate one. In practice the reported effect would show as unusual sunburn in someone using agrimony regularly.

Allergy. Agrimony is a Rosaceae plant, the family that includes apple, pear, cherry, almond, strawberry and raspberry. Allergy to Rosaceae fruits is common, and cross-reactivity to a Rosaceae herb is at least conceivable. Rash, itching, swelling or wheeze after use are the signs of an allergic reaction.

Identity of the material. This is a quieter risk than any of the above and worth a sentence. Agrimonia procera is used interchangeably in commerce, and the East Asian Agrimonia pilosa is a different plant with a different traditional profile. Loose herb bought unlabelled or from an unverified source cannot be matched to any literature at all — and an adulteration caveat clears the plant, not the bottle. If you cannot establish which species and which plant part you have, the safety record you are relying on is not a record of the material in your cupboard.

Numbers We Will Not Give You

Stating refusals openly is more useful than a confident figure that turns out to be inherited from another website. On this page we are declining to assert:

  1. A tannin percentage for the dried herb. Secondary sources quote figures that disagree, the content varies with harvest and plant part, and an analytical method was developed specifically to standardise it — which is what you build when the content moves.
  2. Milligrams of tannin per cup. This would require both a content figure and an extraction efficiency, and neither is available.
  3. A percentage reduction in iron absorption. The tea literature gives a range that varies with meal composition and dose; no agrimony measurement exists. Picking a number from the tea studies and attaching it to agrimony would be exactly the borrowed-precision error this site tries to avoid. The tea and herb-tea figures quoted above belong to those drinks.
  4. A vitamin K content. Discussed above. Listed as a constituent in secondary sources; not quantified in anything we can cite.
  5. A maximum safe daily dose or a toxic dose. No formal toxicology at repeated high doses exists. The traditional-use safety record covers traditional amounts and says nothing about concentrated extracts taken daily for months.
  6. A treatment duration. Traditional practice is short-term, but the specific limits quoted in various places are not derived from data.

Dose and Duration, As Far As They Can Be Stated

What can be said, with its status labelled:

The Honest Bottom Line

Agrimony is a low-hazard herb with one interaction worth genuinely respecting, and the awkward thing about it — the reason this page exists — is that the interaction rests on better evidence than any of the benefits. Tannins reduce non-haem iron absorption; that is established in humans for the compound class, and agrimony's whole identity is that compound class. Magnitude for agrimony specifically has never been measured, and we are not going to invent it.

The lever is timing, and timing is nearly free. In the tea trial a one-hour gap halved the inhibition; the common two-hour gap from meals, iron and medicines works on the same principle, because the binding is luminal and simultaneous rather than metabolic and persistent. A gargle that is spat out removes it entirely.

Beyond that: concentrated tannin irritates the stomach it was meant to settle; animal and cell data make an added glucose-lowering effect a theoretical concern for people on glucose-lowering medication, read as a caution rather than a claim; in pregnancy three separate reasons converge against medicinal doses; for people on an anticoagulant it adds an unquantified variable where control depends on keeping variables few; small children have no data at all, and for them rehydration is the intervention that matters; and persistent, bloody or systemically unwell diarrhoea is a medical problem rather than an astringency problem.

Used that way — briefly, away from other things, in an adult who is otherwise well — agrimony has a long safety record and a coherent rationale. Used casually and indefinitely alongside supplements and prescriptions, it is a slow, quiet interaction with no benefit data to justify it.

Back to Table of Contents

Key Research Papers

Named in prose with journal and year, each linked to its PubMed record.

  1. Disler PB and colleagues (1975) — The effect of tea on iron absorption, Gut. The foundational human demonstration. PubMed PMID: 1168162
  2. Brune M, Rossander L, Hallberg L (1989) — iron absorption and phenolic compounds: importance of different phenolic structures, European Journal of Clinical Nutrition. Establishes that the effect is structure-dependent, which is what licenses the transfer to agrimony's galloyl-bearing tannins. PubMed PMID: 2598894
  3. Hurrell RF, Reddy M, Cook JD (1999) — inhibition of non-haem iron absorption in man by polyphenolic-containing beverages, British Journal of Nutrition. PubMed PMID: 10999016
  4. Ahmad Fuzi SF and colleagues (2017) — a controlled stable-isotope trial in healthy UK women showing that a one-hour interval between an iron-containing meal and tea attenuates the inhibition, American Journal of Clinical Nutrition. The source of the timing advice. PubMed PMID: 29046302
  5. Siegenberg D and colleagues (1991) — ascorbic acid prevents the dose-dependent inhibitory effects of polyphenols and phytates on non-haem iron absorption, American Journal of Clinical Nutrition. PubMed PMID: 1989423
  6. Chung KT and colleagues (1998) — Tannins and human health: a review, Critical Reviews in Food Science and Nutrition. The standard overview of the compound class, benefits and hazards together. PubMed PMID: 9759559
  7. Benvenga S and colleagues (2008) — altered intestinal absorption of levothyroxine caused by coffee, Thyroid. The model case for "separate it by hours." PubMed PMID: 18341376
  8. Gray AM, Flatt PR (1998) — actions of agrimony on hyperglycaemia, cellular glucose metabolism and insulin secretion, British Journal of Nutrition. Animal data, read here as a caution. PubMed PMID: 9797650
  9. Swanston-Flatt SK and colleagues (1990) — traditional plant treatments for diabetes in normal and streptozotocin-diabetic mice, Diabetologia. PubMed PMID: 2210118
  10. Granica S and colleagues (2013) — HPLC standardisation of polyphenols in Agrimoniae eupatoriae herba, Journal of Pharmaceutical and Biomedical Analysis. Why no tannin percentage appears on this page. PubMed PMID: 24001902
  11. Paluch Z and colleagues (2020) — review of the therapeutic effects of Agrimonia eupatoria, Physiological Research. PubMed PMID: 33646008
  12. Topic searches — tannins and non-haem iron · vitamin K intake and INR stability · herb–drug interactions with anticoagulants · tannins and drug absorption · Agrimonia eupatoria

Back to Table of Contents

Back to Table of Contents

Connections

Back to Table of Contents