1. Use caution in dosage and duration of usage and consider using other expectorants before using coltsfoot.
To evaluate the safety of any medicine, the effectiveness must be considered. Coltsfoot has been considered effective because of its mucilage content but there are other useful demulcent plants such as marshmallow. It has been shown to have an antioxidant action and an antiseptic action, but an in vitro study of the antibacterial qualities of Siberian medicinal plants found that the results for coltsfoot were similar to those for yarrow Achillea millefolium, another member of the Asteraceae family. Elecampane is an Asteraceae, which is an effective expectorant and contains sesquiterpenes and may be fully equivalent to coltsfoot immature flower buds, which are widely used in China.
2. Coltsfoot should not be used in children under the age of 18 or in pregnancy and lactation.
Coltsfoot contains pyrrolizidine alkaloids (PAs), which are widespread plant toxins and their incidence in medicinal plants has been reviewed extensively. The compounds are not part of the medicinal benefit of a plant but are toxins whose ingestion should be avoided. Hepatotoxicity of medicinal plants is of concern for regulators but in this case it is a known and thus predictable mechanism rather than an idiosyncratic response.
The structure of different PAs varies and the most toxic compounds are unsaturated PAs, in particular diesters with a macrocyclic ring. The main PAs in coltsfoot are of this type: senkirkine, which is an otonecine-type PA, and senecionine, which is a retronecine-type PA. It also contains saturated PAs, such as tussilagine, which are safe. Senecionine is perhaps the most researched PA as it is found in Senecio species where usage has been linked to deaths. Young children are particularly susceptible and use of Senecio species has been banned in Britain (MHRA 2008), therefore despite its long usage, there are serious concerns about the use of coltsfoot as a medicinal plant.
These concerns are based on studies of the metabolism of PAs. Metabolism of PAs in the liver is primarily by carboxylesterases, which hydrolyze the alkaloids into the corresponding necine base and necic acid. As with all detoxification in the liver, the aim is to produce water-soluble compounds but the rate of metabolism varies widely between individuals. Where PAs are not fully metabolized by carboxylesterases, another phase I reaction occurs, which is metabolism by cytochrome p450 3A4. This leads to an activated product, a pyrrole, which is then detoxified if it conjugates with glutathione. This reaction is catalyzed by glutathione S-transferases. If this step does not occur and the pyrrole is not detoxified, it forms adducts with DNA and RNA which leads to cell death. Cell death also occurs because of crosslinks made between pyrroles and proteins in the cytoskeleton. This leads to characteristic hepatic injury and veno-occlusive liver disease, which was first described in lamaica. A review summarizes research. Other comprehensive reviews are Bertram et al (2001) Prakash et al (1999), Chojkier (2003) and a short summary is given by Denham (1996). Chou & Fu (2006) were able to isolate DNA adducts derived from dehydropyrrolizidines in the livers of female rats one day after they had been fed coltsfoot flowers for 3 days. The rats were given 10 times the recommended dose but the importance of this study is that it shows that these adducts can occur with coltsfoot. Bone (1990) argues that, as sene-cionine is either not present in coltsfoot or found at very low concentrations, then it is relatively safer than Petasites species. However, it has since been shown that otonecine-type and retronecine-type PAs have been found to form the same type of dehydropyrrolizidine-derived DNA adducts.
Guidance on the safe limits for content of pyrrolizidine alkaloids in dried herb material was set in Germany in 1992. The limits are 1 microgram daily for up to 6 weeks where there is a positive Commission E monograph and 0.1 microgram daily where there is no supporting monograph. An exception was made for coltsfoot teas, where the limit was set at 10 micrograms daily. The limits continue to be used and referred to by other sources.
The limit of 10 micrograms for coltsfoot teas or infusions is based on the argument that the level of extraction into teas is low and thus the full content of pyrrolizidine alkaloids will not be available. This may reflect methods of analysis in 1992, not the actual rate of extraction. It has actually been argued that otonecine-type PAs such as sen-kirkine may be more soluble in water than other PAs as they exist in an ionized form as well as the usual non-ionized form. This is supported by a study referred to by Westerdorf (1992), where up to 80% of the senkirkine in a tea from young shoots of coltsfoot was extracted into the aqueous phase. One part per million (ppm) equals 1 microgram/gram so we can compare the concentrations given above with the dose recommended by Commission E. Bartkowski et al (1997) decocted 11 samples of coltsfoot leaf in water for 15 minutes and the mean concentration of senkirkine was 40 ppm, i.e. 40 micrograms per gram. This means that the total daily dose of coltsfoot leaves as a tea would have to be 250 mg to be within the limits. Lebada et al (2000) found a mean concentration of 18 micrograms/gram, which would give a dose of 500 mg to be within the limits. Extraction is lower at room temperature: a study of extracts from one sample of coltsfoot leaf found that there were 2.9 ppm of senkirkine in a sample stirred in water at room temperature for 30 minutes and 9.3 ppm in a sample refluxed in water for 15 minutes.
The recommendation of teas makes another assumption that has since been questioned. Part of the total concentration of PAs in plant material is present as the corresponding N-oxide. These are water soluble and thus may be excreted via the bowel without being absorbed. However, this assumes that the water soluble N-oxides of senecionine found in the plant material are not reduced to PAs in the acid conditions in the stomach. There is some evidence from an animal study that N-oxides can be absorbed and reduced to pyrroles in the liver. PAs are difficult to analyze and it has been argued that determination of the total content, including N-oxides, is important in any evaluation of safety. To conclude, teas and infusions will be safer than more concentrated extracts but there may not be the margin of safety that has been assumed.
3. Coltsfoot should be used short term unless there are overriding reasons for continuing its use in a particular patient.
Many commentators consider that although there is a dose below which there is no frank liver disease, there is no safe dose as genotoxicity and thus changes leading to cancer may be occurring without any symptoms.
The only feeding study published is now over 30 years old and is discussed by Bone (1990). Four groups often rats were fed different doses of coltsfoot flower buds. A dose-dependent incidence of hemangioendothelial sarcoma in the liver was found and the group fed the lowest dose did not develop tumours. Bone (1990) discusses the relevance of feeding studies on animals and argues that they are generally not relevant as the doses are not equivalent. However, there have been more recent studies of the mechanisms of carcinogenicity for PAs. It is outside the remit of this book to take this discussion further as there is a strong argument that there are many other carcinogenic factors in life and that orthodox drugs are used which are known to be carcinogenic. It can equally be argued that a low daily dose of coltsfoot is completely safe as the liver would be able to metabolize any PA.
4. Coltsfoot should not be taken alongside Hypericum perforatum or any medications for the treatment of epilepsy.
Hypericum perforatum has been shown to induce cytochrome p450 3A4. Induction of cytochrome p450 3A4 has been found to increase the rate of metabolism of pyrrolizidine alkaloids to dehydropyrroles, which are the active toxins. Coltsfoot should not be taken alongside any medications for the treatment of epilepsy because some epilepsy medications induce cytochrome p450s. 5. Positive identification of the plant material must be carried out.
We found two adverse event reports putatively involving coltsfoot. In 1988 a case was reported of a baby who died at 38 days who was suffering from venoocclusive disease. Analysis of the tea which his mother had drunk throughout pregnancy identified both Tussilago farfara and root of Petasites hybridus. The main pyrrolizidine alkaloid in butterbur Petasites hybridus is senecionine with its isomer integerrimine. The concentration varies widely between different samples but is highest in the rhizome.
A boy of 18 months from southern Tyrol was admitted with vomiting, diarrhoea, subfebrile temperature, distended abdomen and abdominal pain. On examination the liver was enlarged, 3 cm below the costal margin. Biopsy showed haemorrhagic congestion of zones II and III of the acinus. Fortunately, after 6 weeks of therapy there was remission of the symptoms, no signs of portal hypertension and the child made a full recovery. Subsequent ultrasound evaluation was normal. Further investigation found that since he was 3 months old, the child had been consuming up to 500 mL daily of a tea. Collection and examination of the herb material showed that the herb had been collected in error. It was not coltsfoot but alpendost Adenostyles alliariae. It was found to contain a high concentration of seneciphylline and its corresponding N-oxide such that the authors estimate that the boy could have consumed 60 micrograms/kg of body weight per day.