Coltsfoot and Pyrrolizidine Alkaloids: The Liver Risk

If you only read one page about coltsfoot, read this one. Everything else about Tussilago farfara — the two-thousand-year cough tradition, the pleasant slippery tea, the honey-fried Chinese flower buds, the old boiled sweets called coltsfoot rock — sits downstream of a single chemical fact. The plant makes pyrrolizidine alkaloids, and pyrrolizidine alkaloids injure the liver in a specific, well-characterised, cumulative way.

This is not a scare page and it is not a contamination story. Nobody sprayed anything on the plant. The alkaloids are coltsfoot's own manufactured insect defence, built into the leaf and the flower bud, and no amount of washing, organic certification or careful sourcing removes them from ordinary coltsfoot. That is what makes coltsfoot different from a herb with a theoretical concern attached: the hazard is structural. This page explains exactly what the alkaloids are, the chemical trap by which your own liver turns a fairly inert molecule into a reactive one, the vascular disease that results, what has actually been documented in people (including where the record is genuinely uncertain, which we say plainly), and what warning signs and lab tests matter.

Table of Contents

  1. What Pyrrolizidine Alkaloids Are
  2. Senkirkine and Senecionine: Coltsfoot's Two
  3. The Metabolic Trap: The Liver Makes the Poison
  4. Hepatic Sinusoidal Obstruction Syndrome
  5. Why the Damage Is Cumulative and Delayed
  6. Documented Human Cases — and Their Caveats
  7. Genotoxicity and the Cancer Signal
  8. Plant Part, Season, Provenance: Why Content Varies
  9. Does Preparation Change Anything?
  10. Warning Signs and Which Lab Tests Matter
  11. Who Must Never Take It
  12. Evidence Tiers for Every Claim on This Page
  13. Key Research Papers
  14. Connections

What Pyrrolizidine Alkaloids Are

Pyrrolizidine alkaloids — PAs for short — are a family of several hundred related plant compounds built around a double-ring core called the pyrrolizidine nucleus, usually carrying an attached ester side chain. Plants across at least a dozen families make them, and they make them for the same reason a nettle makes a sting: to be unpleasant to eat. Insects and grazing animals that consume PA-rich plants get sick, so over evolutionary time the plants that made more alkaloid were browsed less.

Toxicologists divide the family by one structural detail that turns out to matter enormously. PAs with a double bond at the 1,2 position of the ring — the "1,2-unsaturated" PAs — are the dangerous ones. PAs without it are largely benign. Coltsfoot's alkaloids are in the dangerous group. This is a rare case where a single bond in a diagram predicts whether a plant belongs in a teacup.

The plants that carry 1,2-unsaturated PAs form a rogues' gallery familiar to anyone who reads herbal safety literature: comfrey, butterbur, borage, the ragworts and groundsels (Senecio species), heliotrope, and coltsfoot. Some, like comfrey, have been the subject of decades of regulatory attention. Others, like the Senecio weeds, have caused mass poisonings when their seeds contaminated grain harvests. The whole class has been reviewed repeatedly — the widely cited 1999 survey Pyrrolizidine alkaloids in human diet in Mutation Research laid out the dietary exposure routes, and later reviews such as Pyrrolizidine alkaloids in food: a spectrum of potential health consequences (Food Additives & Contaminants: Part A, 2011) and Pyrrolizidine alkaloids: chemistry, pharmacology, toxicology and food safety (International Journal of Molecular Sciences, 2018) extended it. (Evidence tier: established review literature.)

What every one of those reviews converges on is a three-word verdict: these compounds are hepatotoxic, genotoxic and, for several members of the class, carcinogenic in animals. It is unusual for a plant-constituent class to attract all three descriptors. It is the reason coltsfoot is not simply "a herb with cautions."

Senkirkine and Senecionine: Coltsfoot's Two

Coltsfoot's principal alkaloids are senkirkine and senecionine. Senkirkine is generally reported as the dominant one in Tussilago farfara; senecionine, the alkaloid that gives its name to the whole Senecio group, is usually present in smaller amounts. Both are 1,2-unsaturated. Both belong to the toxic side of the structural divide described above.

Senkirkine has one structural quirk worth knowing about, because it undercuts a comforting assumption. It is an otonecine-type alkaloid rather than the more common retronecine type. Retronecine-type PAs are often present in the plant partly as N-oxides — a more water-soluble, somewhat less directly reactive form. Otonecine-type alkaloids such as senkirkine do not form stable N-oxides in the same way, and they are activated by liver enzymes without needing an intermediate reduction step. In practical terms: the presence of senkirkine means you cannot argue that coltsfoot's alkaloid load is mostly in a "safer" N-oxide reservoir. The modern comprehensive review of the plant — A review of the ethnobotanical value, phytochemistry, pharmacology, toxicity and quality control of Tussilago farfara L. (coltsfoot), Journal of Ethnopharmacology, 2021 — catalogues both alkaloids among well over a hundred identified constituents and treats their presence as the plant's central quality-control problem. (Evidence tier: analytical chemistry, well established.)

It is worth being explicit about what the rest of the plant's chemistry does not do. Coltsfoot also contains mucilage, flavonoids such as rutin and hyperoside, and the sesquiterpene tussilagone, which has genuine anti-inflammatory activity in cell studies. None of those compounds neutralise, chelate, block or otherwise protect against the alkaloids. A plant is not a formulation designed for balance; it is a plant. The soothing constituents and the hepatotoxic ones simply coexist in the same leaf.

The Metabolic Trap: The Liver Makes the Poison

Here is the part that surprises people, and it is the key to understanding everything downstream. The alkaloid you swallow is not, itself, very toxic. If PAs stayed in the form they arrive in, they would be a minor curiosity. The damage happens because the liver tries to detoxify them and, in doing so, manufactures something far worse inside its own cells.

The sequence, established in detail across the mechanistic literature, runs like this:

  1. Absorption. The alkaloid is absorbed from the gut and travels in the portal circulation straight to the liver — the organ that receives everything the intestine takes up, at full concentration, before the rest of the body sees it.
  2. Oxidation by cytochrome P450. Hepatic CYP enzymes, principally the CYP3A family, oxidise the alkaloid. The intended outcome is a more water-soluble molecule that can be excreted. For 1,2-unsaturated PAs the actual outcome is a dehydropyrrolizidine — a reactive pyrrolic ester.
  3. Electrophilic attack. That pyrrolic ester is highly electrophilic: it is chemically hungry for electron-rich sites. Inside the cell those sites are proteins, glutathione and DNA. It binds them covalently, forming what analytical toxicologists call pyrrole–protein adducts.
  4. Where the damage lands. The adducts form in and around the hepatocytes closest to the liver's smallest draining vessels, and critically they damage the sinusoidal endothelial cells that line those vessels. Endothelium is thin, metabolically exposed and poor at repair. It gives way first.

This process has a name — metabolic activation — and it explains the organ specificity elegantly. The liver is harmed not because it is passive but because it is the site of the chemistry. The mechanism is laid out in Pyrrolizidine alkaloids — genotoxicity, metabolism enzymes, metabolic activation, and mechanisms (Drug Metabolism Reviews, 2004) and refined in Metabolic activation of pyrrolizidine alkaloids: insights into the structural and enzymatic basis (Chemical Research in Toxicology, 2014). (Evidence tier: mechanistic laboratory work, strongly established and reproduced.)

Two consequences follow that matter to a real person:

Hepatic Sinusoidal Obstruction Syndrome

The disease that PAs cause has two names, and you will meet both. The modern term is hepatic sinusoidal obstruction syndrome (SOS). The older term, still in wide use, is hepatic veno-occlusive disease (VOD). The name changed because pathologists realised the primary injury is to the sinusoids — the tiny, porous capillary channels threading between plates of liver cells — rather than to the larger veins, which become involved later.

Picture the liver's drainage as a river delta running in reverse: thousands of narrow sinusoidal channels collect blood, feed into central veins, and those into progressively larger vessels leaving the organ. When pyrrole adducts damage the endothelium lining the sinusoids, the lining swells, sloughs and detaches. Cellular debris and red cells are swept downstream and lodge where the channels narrow. Fibrin and collagen are laid down. The channels obstruct.

What follows is a plumbing problem with systemic consequences:

Two focused accounts of this pathway are Hepatic sinusoidal-obstruction syndrome: toxicity of pyrrolizidine alkaloids (Journal of Hepatology, 2003) and Hepatotoxicity of pyrrolizidine alkaloids (Journal of Pharmacy & Pharmaceutical Sciences, 2015). (Evidence tier: established clinical pathology, described in human case series.)

SOS is not unique to plants — it is also a recognised complication of certain chemotherapy regimens and of stem-cell transplant conditioning — which is precisely why clinicians recognise the syndrome. It is also why a herbal cause is easy to miss: the picture looks like a transplant or chemotherapy complication in someone who has had neither, and the diagnosis depends on a clinician thinking to ask what the patient has been drinking. Herbal-cause SOS is closely related but distinct from Budd–Chiari syndrome, where the obstruction sits in the large hepatic veins rather than the sinusoids.

Why the Damage Is Cumulative and Delayed

The single most under-appreciated feature of PA toxicity is its arithmetic. This is not a threshold poison that is harmless below a line and dangerous above it. Each exposure produces some quantity of reactive pyrrole; some of that binds irreversibly to tissue; and the binding persists. Adducts have been detected in blood long after exposure ended, which is the analytical basis for using them as exposure biomarkers.

Three practical consequences follow, and each one defeats an intuition people reasonably have:

(Evidence tier: established from human outbreak epidemiology and adduct-biomarker studies; the specific dose needed to cause harm in an individual is not known and varies with liver health, co-exposures and genetics.)

Note carefully what that last parenthesis does not say. It does not say a small dose is safe. It says nobody can tell you where your own line is, which is a different and less reassuring statement. Because these alkaloids are also genotoxic, safety assessors generally decline to set a "no effect" level at all and instead work from the principle of driving exposure as low as reasonably achievable.

Documented Human Cases — and Their Caveats

Coltsfoot's human record is thinner than its notoriety suggests, and honesty requires saying both halves of that sentence. The class-level evidence in humans is extensive; the coltsfoot-specific case reports are few, and the most famous one carries a real identification caveat.

The 1988 newborn case

The case that shaped coltsfoot's reputation was published in The Journal of Pediatrics in 1988 under the title Hepatic veno-occlusive disease in newborn infant of a woman drinking herbal tea. A mother drank a herbal tea regularly through pregnancy; her newborn developed hepatic veno-occlusive disease and died. Pyrrolizidine alkaloids were identified as the cause, and the tea was one intended to contain coltsfoot.

Two lessons come out of it, and the second is the one usually omitted.

First, and not in doubt: PAs cross the placenta and can destroy a developing liver. A fetus has immature metabolic and repair capacity and no way to refuse the exposure. Nothing about this case is disputed at the level of "a baby died of PA-induced veno-occlusive disease traced to a herbal tea drunk in pregnancy."

Second, and important for accuracy: when the plant material was examined, the identification of the herb was questioned. The material sold or used as coltsfoot appeared to have been misidentified or admixed — the usual suspects being other PA-bearing composites such as Adenostyles alliariae, whose leaves resemble coltsfoot's. So the case is unambiguous proof of PA hepatotoxicity in a fetus, and unambiguous proof that loose herbal material is unreliable, but it is not a clean demonstration that authentic Tussilago farfara at a normal tea dose killed an infant. We flag that because the caveat is part of the honest record, and because people who defend coltsfoot raise it — correctly.

What the caveat does not do is exonerate coltsfoot. Authentic coltsfoot contains senkirkine and senecionine; the misidentification question concerns which PA-containing plant delivered the dose, not whether the dose was PA. And a hazard that includes "you may be sold a more toxic look-alike" is a worse hazard, not a lesser one. (Evidence tier: single case report, cause of injury established, botanical attribution uncertain.)

Adult veno-occlusive disease

Hepatic veno-occlusive disease has also been reported in adults following consumption of PA-containing herbal preparations, including preparations described as coltsfoot. Separately, a fetal case of veno-occlusive disease attributed to dietary pyrrolizidine alkaloids was reported in the European literature in the early 2000s, and a reversible infant case following a PA-containing herbal tea was reported in European Journal of Pediatrics in the mid-1990s — showing that the injury is not invariably fatal when exposure stops early. (Evidence tier: individual case reports.)

The class-level human evidence, which is not thin at all

Where the human evidence becomes overwhelming is at the level of the alkaloid class. Mass poisonings from PA-contaminated grain and from PA-containing traditional medicines have produced hundreds of cases of veno-occlusive liver disease and many deaths, documented across several continents and reviewed in Hepatic veno-occlusive disease associated with toxicity of pyrrolizidine alkaloids in herbal preparations (Netherlands Journal of Medicine, 2010). Those episodes are the reason nobody in toxicology treats 1,2-unsaturated PAs as a theoretical concern. (Evidence tier: human outbreak epidemiology, established.)

So the fair summary is this: coltsfoot has a small number of case reports, one of which is famous and botanically contested, sitting inside a very large and uncontested body of human evidence that its alkaloids cause the disease in question. That is a much stronger position than "one disputed case," which is how the risk is sometimes dismissed.

Genotoxicity and the Cancer Signal

Liver failure is the acute worry. DNA damage is the long one, and it changes how the risk should be reasoned about.

The same reactive pyrroles that bind proteins also bind DNA, forming DNA adducts and producing cross-links. That is the definition of a genotoxic agent, and it has been demonstrated for PAs across standard assay systems. The consequence for risk assessment is decisive: for a genotoxic carcinogen, assessors generally do not assume a safe threshold exists. Instead of "keep below X," the operating principle becomes "keep as low as reasonably achievable." Every regulatory position on PAs flows from that choice.

For coltsfoot specifically there is direct animal evidence, and it is unusually direct because it tested the whole herb rather than an isolated compound. The 1976 study Carcinogenic activity of coltsfoot, Tussilago farfara L., published in Gann (the Japanese cancer-research journal), fed coltsfoot to rats, which developed liver tumours. Whole plant material, dietary route, tumours in the target organ the mechanism predicts. (Evidence tier: animal carcinogenicity study.)

The parallel case of comfrey has been worked out in more depth — Metabolism, genotoxicity, and carcinogenicity of comfrey (Journal of Toxicology and Environmental Health, Part B, 2010) reviews it — and international cancer-classification bodies have placed certain individual pyrrolizidine alkaloids in the "possibly carcinogenic to humans" category. Note the honest limit here: there is no epidemiological study showing that people who drank coltsfoot tea developed liver cancer. Such a study would be nearly impossible to conduct. The cancer concern is mechanistic and animal-based, not epidemiologically demonstrated in coltsfoot users, and we are not going to overstate it. What it does mean is that the risk cannot be dismissed as dose-dependent-and-therefore-avoidable by drinking a little.

Plant Part, Season, Provenance: Why Content Varies

People sometimes read that coltsfoot's alkaloid content is variable and hear reassurance. It is the opposite of reassurance. Variability means unpredictability, and unpredictability is exactly the wrong property for a substance with no assumed safe threshold.

Known sources of variation include:

Analytical surveys of retail herbal teas and honeys have repeatedly found measurable pyrrolizidine alkaloids in products that were never labelled as containing PA plants at all — the alkaloids arriving via weed seeds and plant fragments harvested alongside the intended crop, or via bees foraging on PA-bearing flowers. That work is what put PAs on food-safety agendas rather than herbal-safety agendas alone. Its relevance here is that a coltsfoot drinker's PA exposure may not be coming from coltsfoot alone. (Evidence tier: analytical survey work, well replicated.)

Does Preparation Change Anything?

A reasonable question, and one worth answering directly, because several folk beliefs about it are wrong.

(Evidence tier: analytical chemistry for the extraction points; absence of evidence — not evidence of absence — for honey-frying, which simply has not been shown to remove PAs.)

Warning Signs and Which Lab Tests Matter

If you or someone you know has been taking coltsfoot — or comfrey, or borage, or an unidentified herbal tea — these are the symptoms that warrant medical attention rather than watchful waiting:

The blood tests a clinician will reach for are the standard liver function tests: the transaminases ALT and AST, alkaline phosphatase, bilirubin, albumin and prothrombin time or INR. GGT often accompanies them. In sinusoidal obstruction the pattern typically shows raised transaminases and bilirubin with the clinical triad of tender hepatomegaly, ascites and weight gain; imaging with Doppler ultrasound assesses flow, and definitive diagnosis has historically required biopsy.

Two practical points that make a real difference to an outcome:

  1. Tell the clinician about the herb, by name, unprompted. Herb-induced liver injury is routinely missed because nobody mentions the tea. Bring the packet if you have it — identification of the actual plant material can matter.
  2. Normal liver tests today do not license continuing. Because the injury is cumulative and can be clinically silent until advanced, a clean panel is a snapshot, not a permission slip.

There is no antidote. Management of PA-induced sinusoidal obstruction is supportive — stopping the exposure, managing fluid and portal pressure, and in severe cases considering transplantation. That absence of a rescue treatment is itself an argument for prevention.

Who Must Never Take It

For some groups, the risk–benefit question does not need weighing at all — and this applies regardless of any "PA-free" label, because a label is not a verified assay:

Evidence Tiers for Every Claim on This Page

Because this site labels evidence rather than asserting it, here is the honest ledger for the claims above:

Read that list twice, because the shape of it is the argument. The harm side is anchored in established mechanism and human disease. The benefit side is anchored in tradition and plausibility. Those are not the same weight of evidence, and that asymmetry — not alarm — is the reason this page reads the way it does.

Key Research Papers

Every link below runs a live PubMed topic search rather than pointing at a single record, so the results stay current as new work appears.

  1. A review of the ethnobotanical value, phytochemistry, pharmacology, toxicity and quality control of Tussilago farfara L. (coltsfoot), Journal of Ethnopharmacology, 2021 — the modern comprehensive review of the plant, including its alkaloids and the quality-control problem they create. Search PubMed
  2. Hepatic veno-occlusive disease in newborn infant of a woman drinking herbal tea, The Journal of Pediatrics, 1988 — the landmark fatal infant case attributed to pyrrolizidine alkaloids in a coltsfoot-intended herbal tea. Search PubMed
  3. Carcinogenic activity of coltsfoot, Tussilago farfara L., Gann, 1976 — rats fed coltsfoot developed liver tumours; direct whole-herb animal carcinogenicity evidence. Search PubMed
  4. Pyrrolizidine alkaloids — genotoxicity, metabolism enzymes, metabolic activation, and mechanisms, Drug Metabolism Reviews, 2004 — how hepatic enzymes convert PAs into reactive pyrroles that bind protein and DNA. Search PubMed
  5. Metabolic activation of pyrrolizidine alkaloids: insights into the structural and enzymatic basis, Chemical Research in Toxicology, 2014 — which structures activate, and via which enzymes. Search PubMed
  6. Hepatic sinusoidal-obstruction syndrome: toxicity of pyrrolizidine alkaloids, Journal of Hepatology, 2003 — the vascular pathology of PA liver injury. Search PubMed
  7. Hepatic veno-occlusive disease associated with toxicity of pyrrolizidine alkaloids in herbal preparations, Netherlands Journal of Medicine, 2010 — the human clinical record from herbal PA exposure. Search PubMed
  8. Hepatotoxicity of pyrrolizidine alkaloids, Journal of Pharmacy & Pharmaceutical Sciences, 2015 — mechanisms and clinical presentation together. Search PubMed
  9. Pyrrolizidine alkaloids in human diet, Mutation Research, 1999 — dietary exposure routes and the genotoxic consequence. Search PubMed
  10. Pyrrolizidine alkaloids in food: a spectrum of potential health consequences, Food Additives & Contaminants: Part A, 2011 — from acute liver disease through to cancer risk. Search PubMed
  11. Pyrrolizidine alkaloids: chemistry, pharmacology, toxicology and food safety, International Journal of Molecular Sciences, 2018 — a broad modern synthesis of the class. Search PubMed
  12. Metabolism, genotoxicity, and carcinogenicity of comfrey, Journal of Toxicology and Environmental Health, Part B, 2010 — the parallel PA herb, worked out in more depth than coltsfoot. Search PubMed
  13. Analytical surveys of pyrrolizidine alkaloids in commercial teas, herbal drugs and honey — repeatedly finding measurable PAs in products not labelled as PA plants, via weed admixture and bee foraging. Search PubMed
  14. Pyrrole–protein adducts as biomarkers of pyrrolizidine-alkaloid exposure — the analytical basis for detecting past exposure after symptoms appear. Search PubMed

External Resources

Connections

Safety Note and Disclaimer

This page is health education, not medical advice, and it exists to document what is known about a plant — including what is claimed for it — so that you can decide for yourself with the risks in plain view. Nothing here should be read as a recommendation to take coltsfoot. On the contrary: because coltsfoot's cough benefit is unproven and interchangeable with that of several safe herbs, while its alkaloids are established liver toxins with no assumed safe threshold, this site's honest reading is that internal coltsfoot is not a sensible choice for anyone. If you are pregnant, breastfeeding, giving something to a child, have any liver condition, drink heavily, or take prescription medicines, do not take it at all. If you have been taking coltsfoot — or any PA-containing herb — and have right-upper-abdominal pain, abdominal swelling, jaundice or unexplained fluid weight gain, seek medical attention promptly and tell the clinician exactly what you have been taking. A cough lasting more than about three weeks, or accompanied by blood, breathlessness, chest pain or persistent fever, needs a clinician rather than any herb.

Back to Table of Contents