Coltsfoot - Is It safe?

The leaves or mature flowers of coltsfoot (Tussilago farfara L) have a long history of use in European traditional medicine for the treatment of respiratory complaints. Even the generic name is derived from tussis, the Latin word for cough. In Chin

With recent developments in analytical chemistry, traces of pyrrolizidine alkaloids (PAs) have been detected in the aerial parts of coltsfoot. Since many PAs are hepatotoxic and may be carcinogenic, this has created concern over the safety of human use of coltsfoot. It is likely that the Australian Government will soon take steps to ban its use.

The likely regulation of the use of coltsfoot highlights some major problems of modern toxicology. These are:

- The lack of sensitivity of biological tests of toxicity compared to a high degree of sensitivity in analytical chemistry. Very low levels of chemical exposure to potential toxicants (natural or synthetic) can now be detected, but the true biologi cal significance of such exposure is not accurately quantifiable. Ultimately the decision about whether such exposure is permissible is made on sociological and economic grounds, not scientific analysis. In other words, a potential risk is assessed again st a perceived benefit.

- The potential for the misuse and over-extrapolation of toxicology data in search of the unattainable goal of absolute safety.

In this context an assessment of the safety of coltsfoot should include objective information about possible benefits as well as toxicology data. Moreover, such data should be scrutinized against similar information for commonly used `safe' foods and drugs for which the benefits of use are perceived to outweigh the potential risks.


Analysis of young coltsfoot flowers demonstrated the presence of 0.015% (150 parts per million) of PAs, and the only alkaloid detected was senkirkine (Culvenor et al., 1976). Tussilagine was later also identified from the aerial parts of coltsfoot (Röder et al., 1981). Although the presence of senecionine has been reported in one study on coltsfoot (Rosberger et al., 1981), this is probably due to adulteration of commercial samples with species of Petasites. Petasites is a common adulterant for coltsfoot (Wagner et al., 1984) and senecionine is the major PA found in this genus (Lüthy et al., 1983).

Coltsfoot also contains mucilage, saponins, and flavonoids including rutin, hyperoside and isoquercetin (Wagner et al., 1984).


Chinese research has demonstrated an antitussive effect in animal experiments (Huang, 1954). German research on the components of a proprietary herbal tea revealed that coltsfoot had the greatest expectorant effect as assessed by the in vitro stimu lation of ciliary transport (Müller-Limmroth and Fröhlich, 1980).

In China, relatively large doses of coltsfoot 18 g/day) were used to treat 36 cases of asthma. Marked improvement was noted in 27 cases (75%) but the herb was not effective at controlling a severe attack. Side-effects, even at these high doses, wer e mild and included nausea, irritability and insomnia (Shao et al., 1964).


Acute Toxicity

The oral LD50 of coltsfoot in mice is extremely high, being 124 g/kg for an aqueous extract and 112 g/kg for an ethanol extract (Wang, 1979). This indicates that coltsfoot has a low acute toxicity. The equivalent toxic dose in a human is about 7,00 0 g, which compares favourably to the typical daily dosage range of 0.6 to 5 g used by Western practitioners.

Senkirkine is hepatotoxic, but tussilagine lacks a 1,2 double bond and is therefore not hepatotoxic (Schoental, 1970). The LD50 of senkirkine in rats is 220 mg/kg (Hirono et al., 1979).

Sub-acute Toxicity

The death of a new-born infant from veno-occlusive disease has been reported in Switzerland. Throughout the pregnancy the mother consumed one cup per day of a herbal tea presumably containing 9% coltsfoot. Veno-occlusive disease commonly occurs fro m sub-acute or acute PA toxicity in humans. Analysis of the herbal tea revealed 0.6 mg/kg (0.6 parts per million) of senecionine (Roulet et al., 1988). The mother denied consumption of any other drugs during the pregnancy.

Whilst a superficial assessment of this study might lead to alarm about the toxic nature of coltsfoot, close examination reveals two problems. Senkirkine, the characteristic PA of coltsfoot, was not detected (this is specifically stated in the stu dy), even though it should have been present at about 10 mg/kg, much higher than the senecionine. Moreover, reliable data indicate that senecionine is not contained in coltsfoot leaves (Röder et al., 1981). Swiss Government authorities (Spang, 1988) have since confirmed that the tea contained Petasites root as well as coltsfoot flowers. (The original study does not mention the presence of Petasites.) The conclusion is inescapable that Petasites was the only source of the senecionine found in the te a.

The second problem relates to the total dose of senecionine, which was no more than 1 mg total during the pregnancy (assuming the amount of herb tea used per cup was 5 g). This represents a PA hepatotoxicity of about two orders of magnitude greater than previously reported, even for neonates (Fox et al., 1978). Clearly, for both these reasons, this isolated case is not relevant to an assessment of the toxicity of coltsfoot, or perhaps even PAs.

Liver damage similar to veno-occlusive disease is caused by other agents such as the epilepsy drug valproic acid. It can also occur in some rare inherited disorders of the amino acid metabolism. Budd-Chiari syndrome, for which the cause in up to 75 % of patients is unknown, also resembles veno-occlusive disease. Veno-occlusive disease of an unknown non-toxic cause has been observed in neonates (Etzioni et al., 1987). Although the authors attempted to exclude all other possibilities, their finding t hat the observed pathology was due to herbal tea consumption is not conclusive.



A genotoxic substance has the capacity to cause chromosomal aberrations. This is related to mutagenicity, the capacity to cause genetic mutations, which is usually assessed by a simple in vitro test such as the Ames test. Exposure to genotoxic sub stances can cause cancer, but not all mutagens are carcinogens.

Senkirkine is mutagenic in the Ames test (Yamanaka et al., 1979) and in the fruit fly (Candrian et al., 1984). Chinese hamster cells treated with relatively high concentrations of senkirkine demonstrated genotoxic and mutagenic effects (Takanashi e t al., 1980). However, senkirkine did not cause chromosomal damage in human lymphocytes, even at concentrations of 10-(3)M (a concentration about 100 times higher than in coltsfoot tea). In contrast, heliotrine, a highly toxic PA with a long history of a nimal and human toxicity, induced marked chromosomal aberrations at the lower concentration of 10-(4)M (Kraus et al., 1985). It can be concluded that senkirkine is probably not genotoxic in humans at the normal exposure levels produced by therapy with co ltsfoot.


The carcinogenicity of a substance is determined in long-term feeding experiments in animals. According to the US FDA, for valid results feeding tests must be carried out in at least two rodent species at no more than 5% of the diet of test animals (FDA, 1982). Unfortunately, such data are lacking for coltsfoot.

Senkirkine carcinogenicity was not confirmed in long-term dosing experiments for male ACI rats. The tumours induced by senkirkine were all benign liver cell adenomas, an indication of hepatotoxicity and not necessarily carcinogenicity (Hirono et al ., 1979).

Feeding experiments with coltsfoot have been conducted on only one animal species, inbred ACI rats (Hirono et al., 1976). The test animals were divided into three groups and fed 16% (initially 32%), 8% and 4% coltsfoot in their diet for 600 days. T he incidence of malignant liver tumours in the test groups were 67%, 10% (one rat) and 0% respectively. There were no liver tumours in the control group. In contrast 4% Petasites produced a high incidence of malignant tumours, presumably because of the p resence of the more toxic senecionine (Hirono et al., 1973).

Using the criteria of US FDA, coltsfoot is not carcinogenic, since a feeding level of 4% did not produce malignant tumours, and indeed all the animals survived beyond 445 days (600 days approximates the full lifespan of the ACI rat (Culvenor et al. , 1976). The findings at higher doses may not be relevant, since such doses may overwhelm normal metabolic pathways and detoxification mechanisms. This is supported by the fact that the tumour incidence increases disproportionately with dose.


It is now known that the human being is constantly exposed to low levels of natural and synthetic chemicals, many of which are potentially cancer-inducing. Whilst every step should be taken to minimize this exposure, a zero risk is not realistic, s ince exposure is often incidental to a perceived benefit.

Ames and co-workers have developed an index for ranking carcinogenic risk called the HERP (Human Exposure/Rodent Potency) (Ames et al., 1987). Assuming the very high levels of coltsfoot fed to rats represent a real carcinogenic risk (which is unlik ely), the HERP for a 3 g daily dose of coltsfoot is 0.06%. This compares with HERP values of 0.03% for a peanut butter sandwich, 0.06% for diet cola, and 0.1% for one raw mushroom. It is also more than 100 times lower than the values of 16% and 17% for typical daily doses of the drugs phenobarbital and clofibrate respectively (Ames et al., 1987).


Why coltsfoot should be banned and not peanut butter, or more significantly why potentially high-risk carcinogenic drugs such as clofibrate are used without apparent concern, cannot be answered using science. The grounds for such decisions and prej udices are outside the scope of objective rational analysis.

Clearly the current evidence suggests that the risk to human health from coltsfoot must be very low. Whether it is acceptably low is more a matter for social and political pressures. The Australian Government is probably acting in good faith. To He alth Department officials the rapidly expanding area of natural medicine poses many problems, and they wish to be seen to be acting responsibly. However, there are those in other areas who can see potential gain through damaging the credibility and pract ice of natural medicine. It is clear that the government should withhold regulation of coltsfoot pending new data and a critical analysis of existing information by an appropriately neutral body.


Ames, B.N.; Magono, R.; Gold, L.S. 1987. Ranking Possible Carcinogenic Hazards. Science, 236: 271-80.

Candrian, U.; Lüthy, J.; Graf, U.; Schlatter, C. 1984. Mutagenic Activity of the Pyrrolizidine Alkaloids Seneciphylline and Senkirkine in Drosophila and Their Transfer into Rat Milk. Food and Chemical Toxicology, 22: 223-25.

Culvenor, C.C.; Edgar, J.A.; Smith, L.W. 1976. The Occurrence of Senkirkine in Tussilago farfara. Australian Journal of Chemistry, 29: 229-30.

Etzioni, A.; Benderly, A.; Rosenthal, E.; Shehadah, V.; Auslander, L.; Lahat, N; Pollack, S. 1987. Defective Humoral and Cellular Immune Functions Associated with Veno-occlusive Disease of the Liver. Journal of Pediatrics, 110: 549-54.

Food and Drug Administration (US). 1982. Toxicologic Principles for the Safety Assessment of Direct Food Additives and Color Additives Used in Foods.

Fox, D.W.; Hart, M.C.; Bergeson, P.S.; Jarrett, P.B.; Stillman, A.E.; Huxtable, R.J. 1978. Pyrrolizidine (Senecio) Intoxication Mimicking Reye Syndrome. Journal of Pediatrics, 93: 980-82.

Hirono, I.; Haga, M.; Fuji, M.; Matsura, S.; Matsubara, N.; Nakayama, M.; Furuya, T.; Hikichi, M.; Takanashi, H.; Uchida, E.; Hosaka, S.; Ueno, I. 1979. Induction of Hepatic Tumors in Rats by Senkirkine and Symphytine. Journal of the National Canc er Institute, 63: 469-71.

Hirono, I.; Mori, H.; Culvenor, C.C.J. 1976. Carcinogenic Activity of Coltsfoot, Tussilago farfara L. Gann, 67:125-29.

Hirono, I.; Shimizu, M.; Fushimi, K.; Mori, H.; Kato, K. 1973. Carcinogenic Activity of Petasites japonicus Maxim, a Kind of Coltsfoot. Gann, 64: 527-28.

Huang, Q.Z. 1954. Chinese Medical Journal, 40: 849. Cited in: H. Chang and P.P. But, eds., Pharmacology and Applications of Chinese Materia Medica, Volume 2. World Scientific Publishing Co., Singapore, 1987.

Kraus, C.; Abel, G.; Schimmer, O. 1985. Studies on the Chromosome Damaging Effect of Some Pyrrolizidine Alkaloids on Human Lymphocytes in Vitro. Planta medica, 89-91.

Lüthy, J.; Zweifel, U.; Schmid, P.; Schlatter, C. 1983. Pyrrolizidine Alkaloids in Petasites hybridus L. and Petasites albus L. Pharma Acta Helvetica, 58:98-100.

Müller-Limmroth, W.; Fröhlich, H.-H. 1980. Action of Some Phytotherapeutic Expectorants on Mucociliary Transport. Fortschr. Med., 98:95-101.

Röder, E.; Wiedenfeld, H.; Jost, E.-J. 1981. Tussilagine -- A New Pyrrolizidine Alkaloid from Tussilago farfara. Mitt. Geb. Lebensmitt. Hyg., 74:4.

Rosberger, D.F.; Resch, J.F.; Meinwald, J. 1981. The Occurrence of Senecionine in Tussilago farfara. Mitt. Geb. Lebensmitt. Hyg., 72:4.

Roulet, M.; Laurini, R.; Rivier, L.; Calame, A. 1988. Hepatic Veno-occlusive Disease in Newborn Infant of a Woman Drinking Herbal Tea. Journal of Pediatrics, 112: 433-36.

Schoental, R. 1970. Hepatotoxic Activity of Retrorsine, Senkirkine and Hydroxysenkirkine in Newborn Rats, and the Role of Epoxides in Carcinogenesis by Pyrrolizidine Alkaloids and Aflatoxins. Nature, 227: 401-402.

Shao, C.R., et al. 1964. Shanghai Journal of Traditional Chinese Medicine, 10. Cited in: H. Chang and P.P. But, eds., Pharmacology and Applications of Chinese Materia Medica, Volume 2. World Scientific Publishing Co., Singapore, 1987.

Spang, R. 1988. Toxicity of Tea Containing Pyrrolizidine Alkaloids. Journal of Pediatrics, 115:1025.

Takanashi, H.; Umeda, M.; Hirono, I. 1980. Chromosomal Aberrations and Mutation in Cultured Mammalian Cells Induced by Pyrrolizidine Alkaloids. Mutation Research, 78: 66-67.

Wagner, H.; Bladt, S.; Zgainski, E.M. 1984. Plant Drug Analysis. Springer-Verlag, Berlin.

Wang, J.M. 1979. Acta Pharmaceutica Sinica, 14: 268. Cited in: H. Chang and P.P. But, eds., Pharmacology and Applications of Chinese Materia Medica, Volume 2. World Scientific Publishing Co., Singapore, 1987.

Yamanaka, H.; Nagao, M.; Sugimura, T. 1979. Mutagenicity of Pyrrolizidine Alkaloids in the Salmonella/Mammalian Microsome Test. Mutation Research, 68:211-16.

The British Journal of Phytotherapy.


By Kerry Bone

Share this with your friends