SAFETY ISSUES AFFECTING HERBS:
PYRROLIZIDINE ALKALOIDS

by Subhuti Dharmananda, Ph.D., Director, Institute for Traditional Medicine, Portland, Oregon

Pyrrolizidine alkaloids (see Figure 1) are complex molecules named for their inclusion of a pyrrolizidine nucleus: a pair of linked pyrrole rings. Each pyrrole can be diagramed as five-sided structure with four carbons and one nitrogen forming the ring. Pyrroles are incorporated into the chlorophyll molecule; the biological role of PAs in plants remains unknown.

Pyrrolizidine alkaloids (PAs) are of special interest currently because several of them have been shown to cause toxic reactions in humans, primarily veno-occlusive liver disease, when ingested with foods or herbal medicines. Comfrey, a well-known medicinal herb characterized by U.S. FDA researchers as having been "one of the most popular herb teas in the world," contains PAs that are capable of causing liver damage (10). Earlier alerts about the potential dangers had reduced demand for comfrey markedly, but herb proponents argued that the laboratory studies showing PA toxicity probably did not apply to human use of the whole herb and that no reports of human toxic reactions to comfrey had appeared (28). Soon after, when clinical reports of such human reactions were published, some herbalists argued that other confounding factors were likely the culprit and that, at any rate, modern drugs were far more likely to cause toxic reactions (29). Comfrey has remained commercially available in the U.S., though in 1993 the American Herb Products Association (AHPA) alerted its members to restrict its use to external applications. On July 6, 2001, the U.S. FDA took official action to remove comfrey from all dietary supplements.

The PAs are primarily found in members of three plant families (7):

1. Asteracea family (Compositae): in plants of the Senecioneae subtribe (24 genera, the genus Senecio is prevelant) and the Eupatorieae subtribe (mainly in the genera Eupatorium and Ageratum);
2. Boraginaceae family: in virtually all plants of this family; and
3. Fabaceae family (Leguminosae): in the subtribe Crotalariaceae, mainly in the genus Crotalaria, but also in the genera Chromolaena and Lotononis.

A list of plants for which concerns have been raised about their content of PAs is presented in Table 1. Some are foraged by animals, causing toxic reactions if grazed to excess; others are weeds that grow amongst grains harvested for human use where they can contaminate the food supply. Several of the listed herbs have been used as medicinals, a few of them have been extensively relied upon for many centuries. The principal medicinal genera in current use are Senecio, Borago, Lithospermum, Heliotropium, and Eupatorium. Toxic reactions to some of those medicinal herbs have been reported following long-term ingestion, usually after many months. Over 200 PAs have been identified to date, about half of them deemed toxic.

The current leading expert on pyrrolizidine alkaloids is Dr. E. Röder, at the Pharmaceutical Institute of Bonn University. Much of the information available about which species contain pyrrolizidine alkaloids, which alkaloids are present, and in what amounts come from his research and review efforts. He has published extensive literature reports since 1995 on European (7) and Chinese herbs (9) and on the chemical analysis methodology for detecting these compounds (8). Dr. C.C.J. Culvenor, of the Commonwealth Scientific and Industrial Research Organization in Australia, had earlier pioneered work in the areas of PA structures, toxicity, and both animal and human responses to PA ingestion. Included in his efforts were an evaluation of the structure and toxicity of 62 PAs (25) and an analysis of the amount of PAs consumed by persons who had been reported to suffer adverse effects (22). He published articles on PAs from 1968-1992. Dr. R.J. Huxtable at the Department of Pharmacology of the University of Arizona reported frequently on the pharmacological effects of PAs and their toxic metabolites (26, 31) from 1978-1999, apparently inspired by two cases of pyrrolizidine poisoning of infants in 1977 and 1978 in his home state of Arizona (27).

Table 1: Plants that contain pyrrolizidine alkaloids (PAs).

Aside from ingesting the plants directly, PAs may be consumed by eating honey collected by bees that visit PA-containing plants (mainly species of Senecio) and by drinking milk or eating eggs produced by animals that have consumed PA-containing plants. In honey originating from species of Senecio, the total concentration of PAs was 0.3-3.2 micrograms per kilogram. PAs could be detected in the concentration range of 30-70 micrograms per kilogram in honey from the Alpine foothills of Switzerland. No cases of PA poisoning have been attributed to ingestion of tainted honey or milk (7). Eggs have been implicated as a source of PA toxicity in a case involving use of unregulated grain feed containing some Heliotropium (32).

THE PHYSICAL REACTION TO PYRROLIZIDINE ALKALOIDS

The toxic PAs may cause the following liver reactions when ingested in doses of 10-20 mg: enlargement of liver cells and their nuclei, disturbances of liver cell metabolism resulting in functional losses, areas of cell destruction, and fatty degeneration. Long-term administration of smaller doses, 10 micrograms or less per day, may cause liver cirrhosis with exposure to the most toxic of the PAs. In both acute and long-term responses, veno-occlusive liver disease may occur. Although the main site of toxic reaction to the PAs is the liver, a few PAs act on other organ systems or the nervous system; this may be the result of their long half-life, allowing them to migrate to the other parts of the body prior to further metabolism.

The PAs, which have minimal toxicity in their original form, are metabolized in the liver and can become toxic metabolites, depending on the PA and on the particular condition of the liver enzymes (31). The toxic metabolites, highly reactive smaller pyrroles (the dehydro- form of the alkaloids) resulting from action of microsomal enzymes in the liver, can act locally within the liver cells to cause damage at the chromosome level. If the liver becomes damaged, the pyrrolizidine metabolites can overflow and infiltrate the lung fluids and cause damage there. Pulmonary edema and pleural effusions have been observed, sometimes resulting in fatalities with very high levels of PA contamination of food.

The diagnosis that most often is observed when there has been excessive or prolonged exposure to PAs is veno-occulusive liver disease, involving obstruction of the small veins bringing blood from the liver back to the heart. Early clinical signs of veno-occlusive liver disease include nausea and acute upper-gastric pain, acute abdominal distension with prominent dilated veins on the abdominal wall, fever, and elevated liver enzymes. Jaundice may be present and the gallbladder may be distended and full. Chronic illness from ingestion of small amounts of the alkaloids over a long period proceeds through fibrosis of the liver to cirrhosis, which is indistinguishable from cirrhosis of other causes. Removal of PA exposure will usually alleviate the disorder, but liver cirrhosis is only marginally reversible. Young children are most susceptible to the effects of the PAs, yet are more likely than adults to completely recover.

HISTORY OF HUMAN TOXIC RESPONSES

Reports of human toxic reactions to PAs are mostly limited to cases where grain has been contaminated by seeds of PA-containing plants, causing acute reactions in a substantial number of people. This has occurred, for example, in Uzbekistan and Afghanistan in 1948 and again in 1976, in Jamaica in 1954, and in India in 1975 (drought conditions during those years encouraged growth of the weeds, mainly Senecio, Crotalaria, and Heliotropium, and led to high PA content of flour). Agricultural controls have prevented further occurrences, except in times of disruption by wars: there was a small outbreak of liver disease, with 2 deaths, in northern Iraq in 1994. During the past century, thousands were poisoned central Asia, and other cases have been reported around the world in countries with poorly developed agricultural systems. It has been suggested that the Kwashiorkhor disease, frequently observed with children in Central Africa, is also related to a damage of liver cells by PAs taken in with foods, characterized by the liver losing its ability to synthesize essential proteins.

An extensive review of the medical literature conducted for the current article reveals that there have been only about a dozen formal reports (a few of them then repeated in different journals or with differing details) involving about 20 people with PA-related toxic reactions from taking excessive doses of herbs or taking incorrect herbs. Each of these reports are mentioned here. Some researchers have suggested that the problem is more widespread in certain areas of the world, such as Africa and South America, and simply not reported.

Pyrrolizidine alkaloids became an issue of safety concern in relation to herbal medicine about 20 years ago, when an herb enthusiast consumed comfrey (Symphytum officinale) on a daily basis and suffered liver damage (2). Three additional cases of human toxic response to comfrey were eventually reported (a possible fifth one has been mentioned, see reference 31), including one death attributed to the effects of the herb. The four widely-reported cases, all occurring in the 1980s, were outlined in a presentation by the National Institute of Medical Herbalists (London), to the Department of Health in January 1993 (29):

• Case 1 (11): A 49-year-old woman in the U.S. was admitted to a hospital with progressive swelling of the abdomen and extremities over the preceding four months. Veno-occlusive disease was eventually diagnosed, allegedly caused by chronic exposure to PAs consumed in a comfrey powder, estimated at a minimum of 85 mg of PAs per day over the previous six-month period. She was a heavy consumer of herbs, vitamins and natural food supplements, and she drank three cups of chamomile tea per week and for the six months before admission had consumed one quart a day of a herbal tea known as Mu-16. In addition, for the 4 months before admission, she had taken two capsules of "comfey-pepsin pills" with each meal. The authors conclude "To our knowledge, this is the first report of veno-occlusive disease in any human after the use of a preparation claiming to be made from comfrey."
• Case 2 (12): A 13-year-old boy in England was admitted to a hospital with symptoms that were found to be caused by veno-occlusive disease. He had been suffering from Crohn's disease for three years and had been treated with prednisolone and sulphasalazine, which removed symptoms. At his parents' request, the drugs were discontinued and he was treated with acupuncture and comfrey root prescribed by a naturopath. Exact quantities and frequency are unknown. When admitted to the hospital he was taking prednisolone and sulphasalazine. The authors concluded that "the only possible causal factor in this patient was comfrey."
• Case 3 (13): A 47-year-old white non-alcoholic woman in the U.S. began to feel unwell in 1978 with vague abdominal pain, fatigue and allergies. A homeopathic doctor recommended comfrey tea. She consumed as many as ten cups per day in addition to taking comfrey pills by the handful, which continued for more than one year. Four years later, serum aminotransferase levels were twice the upper limit of the normal range, and four years after that she had further signs of liver disease.
• Case 4 (14): A 23-year-old man in New Zealand presented with veno-occlusive disease and severe portal hypertension and subsequently died from liver failure. He had eaten comfrey leaves for some time before his illness. The man presented with a three-month history of initial influenza-like symptoms followed by continued malaise and night sweats. Three weeks before admission he noticed peripheral edema and abdominal distension. For four years prior to this illness he had been living in a commune and had eaten a predominantly vegetarian diet. He had a striking binge-type eating pattern whereby he would eat large quantities of a particular food for days and weeks on end. In the one to two weeks before the onset of symptoms he ate four to five steamed young comfrey leaves as a vegetable every day. The authors suggest that the patient's protein deficient diet could have played a contributory role; they attributed comfrey as a possible cause due to the temporal sequence of events. In a separate review of potential risk to consuming comfrey published in the Australian Medical Journal (15), the author declined to consider this case in his report because "there is some controversy surrounding this case."

Though not of special notice at the time, because the herbs were not so widely used, cases of veno-occlusive disease associated with medicinal herbs had been reported just a few years earlier. In the U.S., a two-month-old infant in Arizona experienced veno-occlusive liver disease after being treated with gordolobo yerba, one of the most popular of Mexican herbs, usually obtained from Gnaphalium, but sometimes substituted by the similarly appearing Senecio longilobus. The case, involving portal hypertension and extensive hepatic fibrosis, was first reported in 1977 (19); the child died of the disease. An additional case of an infant administered Senecio longilobus as gordolobo yerba, used to make herbal cough medicine, involved a 6-month-old female. She had acute hepatocellular disease, ascites, portal hypertension, and a right pleural effusion in 1978. She improved with treatment. However, after 6 months, a liver biopsy revealed extensive hepatic fibrosis, progressing to cirrhosis over 6 months (27). It was found that one supply house in Arizona routinely stocked Senecio as the source of gordolobo. The U.S. FDA provided the following additional report from the same time period without references to documentation: an American woman who had used a medicinal tea for 6 months while visiting Ecuador developed typical hepatic veno-occlusive disease, with voluminous ascites, centrilobular congestion of the liver, and increased portal vein pressure. The patient completely recovered within one year after ceasing to consume the tea.

In England, veno-occlusive disease was blamed on long-term consumption of large amounts of Paraguay tea (yerba mate, Ilex paraguensis), from which small amounts of pyrrolizidine alkaloids were reportedly isolated (18). However, later studies did not find these alkaloids in Ilex; the patient may have had a contaminated source. Veno-occlusive disease was said to be caused by medicinal herb use in India in a report published in 1978 (16). Three patients with the disease had taken Heliotropium eichwaldii. The authors of the medical report suggested that: "The medicinal use of the herb may possibly be responsible for a significant proportion of acute and chronic liver disease in India." However, such a conclusion seems unjustified; in India, Heliotropium species are mainly used in external therapies, and only rarely being taken in teas or other internal forms. In 1985, there appeared a report linking ingestion of an unidentified herb from India, used as a treatment for psoriasis, to three cases of liver damage, with one death. Pyrrolizidine alkaloids were identified, with mean dose estimated at 30 mg/day. The herb may have been a species of Heliotropium, as these species are used as remedies for various skin diseases. A case of Heliotropium induced veno-occlusive disease was reported in England in 1986 (17); the species ingested was H. lasiocarpum.

A new concern arose with use of another popular herb: coltsfoot (Tussilago farfara). Coltsfoot was widely used for coughs in Europe and America, and has also been used by the Chinese for the same purpose. There was a report from Switzerland in 1988 of liver damage leading to death of a newborn due to consumption of coltsfoot tea by the mother during pregnancy (20). The adverse reaction was attributed to the herb's pyrrolizidine alkaloids, and that herb was then withdrawn from the market in Europe in 1992. However, later investigation indicated that the coltsfoot was probably not a contributor to the infant's death (23). Still, other PAs may have been responsible in this case: the tea mixture included Petasites (pestilence wort), a source of toxic PAs.

In 1995, three more reports appeared in the medical literature indicating liver damage from ingestion of PA-containing herbs. An infant in Austria who was given a tea that was understood to contain peppermint and coltsfoot (not a commercial product) developed veno-occlusive disease that was, fortunately, reversible (5). An analysis showed that the coltsfoot was again a substitute, this time from Adenostyles xalliariae (common name alpendost). It was estimated that the child consumed at least 60 micrograms PAs per kilogram of body weight each day over a 15 month period. It was noted that an elderly patient taking a tea of Senecio vulgaris for two years continuously suffered from veno-occlusive liver disease (3) and that a younger woman taking Senecio tephrosioides as a cough remedy on an occasional basis but for many years also suffered from this liver disease (4).

To sum up the cases reported thus far: there are four (possibly five) incidents related to comfrey ingestion in adults, two incidents of Senecio ingestion in infants and two incidents of Senecio ingestion in adults; four cases of Heliotropium ingestion in adults, two cases involving coltsfoot substitution by other PA-containing herbs in infants, and five other cases involving adults with unknown PA-containing herbs (three in China, one during a visit to Ecuador, and one using a tea product from South America). Although some cases of liver damage may have been due to something other than the herbs, in most instances, other sources of the disorder were not evident or were deemed improbable.

DOSE AND DURATION OF USE TO CAUSE HEPATOTOXIC REACTIONS

The dose and duration of exposure to the toxic PAs that have been associated with liver damage in humans was estimated by Culvenor (22). For heliotrine, an intake of 4-10 mg/kg per day for 3-7 weeks led to liver necrosis and veno-occlusion; the combination of crotanine and cronaburnime at a dose of less than 1 mg/kg per day for several months led to the same disorders. Retrorsine and riddelline, among the most toxic of the PAs caused liver necrosis, liver fibrosis, and cirrhosis within 2 weeks of intake from 0.7-1.5 mg/kg per day. In one of the reports of an unidentified herb from India, the PA ingestion rate was 30 mg/day, and in one of the infants, the reported PA intake was 0.06 mg/kg per day. Thus, the range of toxic doses in humans appears to be in the range of about 0.1-10 mg/kg per day. In most instances of PA liver toxicity in adults, the daily intake was several milligrams or hundreds of milligrams per day. However, it has been suggested by the World Health Organization in 1989 that the lowest intake rate of PAs that reportedly caused veno-occlusive disease in a human was just 0.015 mg/kg of body weight per day, based on use of comfrey. For a 70 kg adult, that would correspond to 1 mg total per day.

Exposure to PAs can vary markedly when using any given herb. Determination was made of PA content of comfrey roots and leaves by Couet and his colleagues (21); the roots had a range of 1400-8300 ppm, while the leaves had from 15-55 ppm. In an evaluation of 300 comfrey root samples in Germany, the PA range was found to be 450-5990 ppm (30). An evaluation of commercial comfrey products (10), showed that the PA content varied markedly: none detected in 2 products, to a range 0.2-220 ppm among 8 other products tested with detectable levels, and one with 1520 ppm (a comfrey root product). To reach a 1 mg per day dose, just 0.7 grams of herb at 1,520 ppm would be needed, about the amount that would be found in 3 of the 250 mg capsules, indicating that this product would be too toxic to consume on a regular basis. On the other hand, products with no detectable PAs and those with less than 1 ppm might be entirely safe. Further, method of preparation is also important. A decoction of root and leaf samples of comfrey lowered the available PAs by 75-95%. Different species of comfrey contain different types and amounts of toxic PAs.

POTENTIAL CARCINOGENICITY WITH LONG-TERM EXPOSURE TO PAS

A concern that has arisen during the past two decades is that persons who take an herbal preparation regularly for many months or years might experience a cumulative effect from the PAs. Although these ingredients do not accumulate in the liver or elsewhere in the body, the prolonged use of amounts that cause no acute reactions may increase the chances for developing liver disease, including liver cancer. Laboratory studies indicate that some of the PAs have a high mutagenic capacity. This determination is made through tests in bacteria or fruit flies, but the results may imply a carcinogenic potential for humans. According to one study using fruit flies, the mutagenic potential of 16 PAs can be rated as follows (7):

senkirkine > monocrotaline > seneciphylline > senecionine > 7-acetyl intermedine >
heliotrine > retrorsine > 7-acetyllycopsamine > symphytine > jacoline > symlandine >
intermedine > indicine > lycopsamine > indicine N-oxide > supinine

That is, senkirkine has the highest mutagenic potential in the test system, while supinine had no detectable mutagenic activity. Indicine N-oxide (from heliotrope) is an antitumor agent currently investigated for treatment of brain tumors; formerly it was found useful in treating leukemia in infants, except that it produced liver damage as a side effect in some cases (thus, it was withdrawn from further consideration). Senkirkine is found in coltsfoot, petasites, and other herbs that were routinely recommended in Switzerland and Germany as remedies suited for long-term administration.

Despite the findings in laboratory studies, including mutagenic activity, adducts of pyrrolizidine alkaloids and their metabolites in tissue of test animals, and induction of cancer in animals (some PAs cause hemangioendothelial sarcomas in rats), the potential carcinogenic effects on humans have been questioned. On the one hand, the frequent occurrence of primary liver tumors in the natives of Central Africa and South Africa has been ascribed to the consumption of traditional medicinal plants of the genera Crotalaria, Cynoglossum, Heliotropium and Senecio. On the other hand, in a recent review of PAs in the human diet (2), the authors concluded that "while humans face the risk of veno-occlusive diseases and childhood cirrhosis, PAs are not carcinogenic to humans."

RESPONSE OF GOVERNMENT AND INDUSTRY AGENCIES

As recent studies revealed the presence of pyrrolizidine alkaloids in several plants used in herbal medicines, concern has arisen about exposure to these compounds and how to limit that exposure. Bulletins have been issued by the U.S. Food and Drug Administration (FDA) and Centers for Disease Control (CDC) on this subject. In Belgium it is proposed that the limit for pyrrolizidine alkaloids in herbs be set at 1 ppm, a very small amount (1 microgram per gram of herb). The German Health Administration has set a standard for use of the herb Petasites such that the daily dose should not contain more than 1 microgram of PAs, with limited duration of administration of six weeks total per year; for tea mixtures including PAs, a total limit of 10 micrograms per day is permitted for up to six weeks. The American Herb Products Association (AHPA) has issued the following recommendation (adopted July 1996): "AHPA recommends that all products with botanical ingredients which contain toxic pyrrolizidine alkaloids bear the following cautionary statement on the label: For external use only. Do not apply to broken or abraded skin. Do not use when nursing." The Consumer Healthcare Products Association established a voluntary program for manufacturers with the same recommendation (adopted March 2001).

Based on the Belgium recommended limit for PAs in herbs at 1 ppm, it would require ingestion of 1 kilogram (2.2 pounds) of an allowed herb to yield 1 mg (the smallest amount cited as toxic for comfrey ingestion), so the limit set would appear to be quite safe, in terms of any potential for acute reactions. Normal daily intake of herbs is usually only a few grams, so there is a large margin of safety; a 10 gram dose would yield a maximum of 10 micrograms of PAs from an allowed herb. The German recommended limit, just 1 to 10 micrograms per day, is consistent with the Belgium proposal based on PA content. In Europe, efforts are underway to develop coltsfoot and other desirable medicinal plants that are naturally low in PAs and will then meet the standards that have been set. In addition, extract manufacturers have sought methods of removing the PAs from their finished products.

Drastic restriction of PAs, such as avoidance of any amount in an herb intended for internal use as suggested by AHPA, may not be justified. There is a great diversity of alkaloid structures among the pyrrolizidine group. Some of the pyrrolizidine alkaloids, such as farfugine and tussilagine, are considered non-toxic. Although the AHPA recommendation specifically mentions "toxic PAs," the burden of proof as to whether the PAs are to be considered toxic would fall on the seller of the herb material, a burden unlikely to be undertaken. The range of amounts of PAs in plants (and different parts of the same plant) is also quite large: from less than 0.001% to 0.1.2%, more than a 1,000-fold range among samples with reported levels. This variability in amounts and toxicity makes it difficult to rationally suggest that no pyrrolizidines, including the toxic ones, are acceptable in herbs used medicinally as internal remedies. As test methods improve, tiny amounts of these compounds may be found in very popular herbs. Thus, for example, echinacea (in the Asteracea family), one of the most commonly used herbs in America and Europe, has been shown to contain small amounts of PAs. These are thought to be non-toxic (based on the chemical structure), but they remain suspects in potential liver toxicity from long-term use (over 8 weeks) of the herb (33).

It is potentially problematic to deem any amount of any such alkaloid a contraindication for using the herb at all, especially as test methods improve and miniscule amounts of a wider range of PAs are revealed. PAs have been detected in plant families other than the three mentioned above, including the Apocynaceae, Celastraceae, Euphorbiaceae, Orchidaceae, Poaceae, Ranunculaceae, Rhiophoraceae, Santalaceae, Sapotaceae, and Scrophulariaceae, and have been suggested to be present in up to 3% of the species of flowering plants (24). Therefore, numerous medicinal herbs may fall under restriction from oral consumption if there is zero tolerance for presence of PAs. Further, the implications of such a restrictive policy for the future are that any time a natural compound is found to have some level of toxicity, all medicinal herbs containing that compound could be eliminated from use, greatly limiting the ability to rely on a natural health care approach.

PYRROLIZIDINE ALKALOIDS IN CHINESE HERBS

The commonly used Chinese herbs that are currently known to contain pyrrolizidine alkaloids are these (24):

• Zicao: obtained from Lithospermum erythrorhizon and Arnebia euchroma of the Boraginaceae. Total PA yield of Lithospermum erythrorhizone was 0.02%, consisting mainly of intermedine, and myoscorpine; from Arnebia euchroma the total PA yield is only 0.0006%, comprised mainly of angeloylretronecine.
• Kuandonghua: obtained from Tussilago farfara of the Asteracea. It contains tussilagine, isotussilagine, senecionine, and senkirkine. In Japan, Petasites japonicus has been used as kuandonghua.
• Qianliguang: obtained from Senecio scandens of the Asteracea. Senecio scandens and tablets from its extract were officially listed in the Chinese Pharmacopoeia 1977, indicated for bacterial diarrhea, enteritis, conjunctivitis, and respiratory tract infections. It is still used in composition formulas for treatment of rhinitis, ulcerative colitis, and burns. It contains the PAs senecionine and seneciphylline.
• Peilan: obtained from Eupatorium fortunii and E. japonicum of the Asteracea.

In an evaluation of 20 herbs of the Asteracea, only species of Eupatoreum were found to contain pyrrolizidine alkaloids (1), with content ranging from less than 5 ppm for E. chinensis, to 65-95 ppm for E. fortunei, and 172-422 ppm for E. japonicum. The PA-free herbs (less than 1 ppm) included species of Arctium, Artemisia, Aster, Atractylodes, Carthamus, Centipeda, Chrysanthemum, Circium, Eclipta, Inula, Saussurea, Taraxacum, Tussilago, and Xanthium. The reported absence of PAs in Chinese samples of Tussilago is consistent with findings of negligible amounts in some European samples. In a study of three Chinese herbs, two species of Eupatoreum (E. japonicum and E. cannabinum) were shown to contain PAs and a species of Crotalaria (C. assamica) was found to contain one PA, monocrotaline (6). Other Chinese herbs with PAs include various species of Senecio used in folk medicine (e.g., S. argunensis and S. integrifolius), Emilia sonchifolia (Asteraceae), and Crotalaria sessiliflora. Heliotropium indicum (Boraginaceae) is a folk remedy in Taiwan, used for treatment of lung diseases and sore throat.

In order to meet the Belgium standard for PAs, the alkaloid content of an herb would have to be no more than 0.0001%, which is lower by far than that for Arnebia euchroma a Chinese herb with very low levels of PAs. To meet the German standard of a maximum daily dose of 10 mg of PAs, an herb containing 0.02% of these compounds (as reported for Lithospermum erythrothrizone) would be limited to just 50 mg per day, an amount far less than is likely to be used in Chinese herbal formulas. Therefore, elimination of the herbs with established PA content from Chinese herb prescribing would be necessary to meet Western standards.

REFERENCES

1. Zhao XL, Chan MY, and Ogle CW, The identification of pyrrolizidine alkaloid-containing plants, American Journal of Chinese Medicine 1989; 17 (1-2): 71-78.
2. Stickel F and Seitz HK, The efficacy and safety of comfrey, Public Health and Nutrition 2000; 3(4A): 501-508.
3. Ortiz CA, et al., Veno-occlusive liver disease due to intake of Senecio vulgaris tea, Gastroenterology and Hepatology 1995; 18(8): 413-416.
4. Tomioka M, et al., Hepatic veno-occlusive disease associated with ingestion of Senecio tephrosioides, Review of Gastroenterology (Peru), 1995; 15(3): 299-302.
5. Sperl W, et al., Reversible hepatic veno-occlusive disease in an infant after consumption of pyrrolizidine containing herbal tea, European Journal of Pediatrics, 1995; 154(2): 112-116.
6. Edgar JA, et al., Pyrrolizidine alkaloid composition of three Chinese medicinal herbs, American Journal of Chinese Medicine 1992; 20 (3-4): 281-288.
7. Röder E, Medicinal plants in Europe containing pyrrolizidine alkaloids, Pharmazie 1995; 50: 83-98.
8. Röder E, Analysis of pyrrolizidine alkaloids, Current Organic Chemistry 1999; 3: 667-676.
9. Röder E, Medicinal plants in China containing pyrrolizidine alkaloids, Pharmazie 2000; 55: 711-726.
10. Betz JM, et al., Determination of pyrrolizidine alkaloids in commercial comfrey products, Journal of Pharmaceutical Science 1994; 83(5): 649-653.
11. Ridker PM, et al., Hepatic veno-occlusive disease associated with consumption of pyrrolizidine alkaloid-containing dietary supplements, Gastroenterology 1985; 88: 1050-1054.
12. Weston C, et al., Veno-occlusive disease of the liver secondary to ingestion of comfrey, British Medical Journal 1987; 295: 183.
13. Bach N, et al., Comfrey herb tea-induced hepatic veno-occlusive disease, American Journal of Medicine 1989; 87: 97-99.
14. Yeong ML, et al., Hepatic veno-occlusive disease associated with comfrey ingestion. Journal of Gastroenterology and Hepatology 1990; 5(2): 211-214.
15. Abbott PJ, Comfrey: assessing the low-dose health risk, Medical Journal of Australia, 1988; 149 (11-12): 678-682.
16. Datta DV, et al., Herbal medicines and veno-occlusive disease in India, Postgraduate Medical Journal 1978; 54(634): 511-515.
17. Culvenor CCJ, et al., Heliotropium lasiocarpum identified as cause of veno-occlusive disease due to a herbal tea, Lancet 1986; 1(8487): 978.
18. McGee J, et al., A case of veno-occlusive disease of the liver in Britain associated with herbal tea consumption, Journal of Clinical Pathology 1976; 29(9): 788-794.
19. Stillman AS, Hepatic veno-occlusive disease due to pyrrolizidine (Senicio) poisoning in Arizona, Gastroenterology 1977; 73(2): 349-352.
20. Roulet M, et al., Hepatic veno-occlusive disease in newborn infant of a woman drinking herbal tea, Journal of Pediatrics, 1988; 112(3): 433-436.
21. Couet CE, Crews C, and Hanley AB, Analysis, separation, and bioassay of pyrrolizidine alkaloids from comfrey, Natural Toxins 1996; 4: 163-167.
22. Culvenor CCJ, Estimated intakes of pyrrolizidine alkaloids by humans, Journal of Toxicology and Environmental Health 1983; 11: 625-635.
23. Spang R, Toxicity of tea containing pyrrolizidine alkaloids, Journal of Pediatrics, 1989; 115(6): 1025.
24. Tang Weici, Mutagenic and carcinogenic constituents of Chinese herbal medicines, Abstracts of Chinese Medicine 1995; 6(2): 227-271.
25. Culvenor CC, et al., Hepato- and pneumotoxicity of pyrrolizidine alkaloids and derivatives in relation to molecular structure, Chemical and Biological Interactions 1976; 12(3-4), 299-324.
26. Cooper RA and Huxtable RJ, The relationship between reactivity of metabolites of pyrrolizidine alkaloids and extrahepatic toxicity, Proceedings of the Western Pharmacology Society 1999; 42: 13-16.
27. Huxtable RJ, Herbal teas and toxins; novel aspects of pyrrolizidine poisoning in the U.S., Perspectives in Biology and Medicine 1980; 24(1): 1-14.
28. Blumenthal M, Speaking of herbs, Health Foods Business 1980 (April): 19, 94.
29. Whitelegg M, In defense of comfrey, report to the Department of Health, London, 1993.
30. Denham A, Using herbs that contain pyrrolizidine alkaloids, European Journal of Herbal Medicine 1996; 2(3): 27-38.
31. Huxtable RJ and Cooper RA, Pyrrolizidine alkaloids: Physiochemical correlates of metabolism and toxicity, in Tu AT and Gaffield W (editors), Natural and Selected Synthetic Toxins: Biological Implications, 2000 American Chemical Society, Washington D.C. pp: 100-117.
32. Edgar JA and Smith LW, Transfer of pyrrolizidine alkaloids into eggs: Food safety implications, in Tu AT and Gaffield W (editors), Natural and Selected Synthetic Toxins: Biological Implications, 2000 American Chemical Society, Washington D.C. pp: 118-128.
33. Miller LG, Herbal medicinals: Selected clinical considerations focusing on known or potential drug-herb interactions, Archives of Internal Medicine 1998; 158(20): 2200-2211.
    November 2001

    
    Figure 1: Chemical structures of various PAs.