Symphytum: Comfrey, coltsfoot, and pyrrolizidine alkaloids
by Paul Bergner
Medical Herbalism 1(1):1,3-5
Is Symphytum officinale (Comfrey) safe for internal use? And what about Tussilago farfara (Coltsfoot)? Maybe not, according to clinical cases published in the last few years which suggest that the pyrrolizidine alkaloids in these popular traditional herbal medicines may cause serious liver disease if consumed over long periods of time. Germany recently banned some 2500 herbal products with pyrrolizidine alkaloid content, because of these and other cases. Canada recently also banned comfrey root and coltsfoot.
The debate on the toxicity of comfrey and coltsfoot for internal use pits traditional herbalists against regulatory agencies. The traditional herbalists argue that these plants have been widely used for centuries without any observed ill effects. Regulatory agencies point to evidence that pyrrolizidine-alkaloid-containing plants, and comfrey and coltsfoot in particular, have caused liver disease in humans.
The evidence suggests that both plants should be used with caution — especially for long term use, for pregnant and lactating women, and for children. The coltsfoot ban may have been premature, however, based on poor and contradictory evidence.
Most cases of pyrrolizidine alkaloid poisoning in the scientific literature have involved third-world epidemics among people who consumed contaminated grain over a long period. [16,24,25,26] Most clinical cases have involved Senecio, Heliotropium, and Crotalaria species.[2,14,16,21,23,24,25,26] In 1985 and 1987, cases appeared in which comfrey (S. off.) itself was the offending agent.[17,27] Other articles have appeared which analyzed various commercially available comfrey products and showed that they had sufficient levels of pyrrolizidine alkaloids to cause toxicity with long term use.[10,13,19] See Table 2. The case which led to the German banning of herbal products identified consumption of a tea purporting to contain coltsfoot (T. farfara) during pregnancy as the cause of the death of a newborn infant. Although coltsfoot was on the label of the tea, a pyrrolizidine alkaloid characteristic of coltsfoot (senkirkine)  was not present in the tea, and the alkaloid isolated from the tea (senecionine) is usually associated with other plants, such as Petacites spp.  a possible adulterant of coltsfoot.
(S. officinale — leaf) and Coltsfoot (T.
farfara — buds)
have both caused liver tumors when introduced
into the diets of mice. [6,7]
The pyrrolizidine alkaloid constituents of each
plant (symphytine from
comfrey; senkirkine from Coltsfoot) had a
similar effect. Such liver
tumors — never demonstrated in humans — are not
the chief clinical concern
with these plants, however. Pyrrolizidine
alkaloid poisoning causes a liver
disorder in humans called hepatic veno-occlusive
disease. The small and
medium veins in the liver become obstructed,
eventually leading to liver
disfunction, cirrhosis and death. Breakdown
products of the pyrrolizidine
alkaloids appear to bind to tissue in the liver,
causing cell dysfunction
and blockage of the veins by a proliferation of
Other causes of hepatic veno-occlusive disease
can be radiation treatments
to the liver, bone marrow transplantation,
chemotherapeutic drugs, birth
control pills, immunosuppressive therapy (e.g.
azathioprine for bowel inflammations
and organ transplants,)[30,31] other immune
deficiency , and intravenous
administration of a vitamin E preparation to low
birth weight infants.
Hepatic veno-occlusive disease can only be
conclusively diagnosed with
a liver biopsy. In the three cases here
physicians methodically ruled out
causes other than pyrrolizidine alkaloid
poisoning, and verified the diagnosis
by biopsy. [17,27]
CASE 1. 13 year-old boy. Admitted for liver enlargement and abdominal swelling. Three-year history of Crohn’s disease, starting in 1983. Originally treated with prednisolone and sulfasalazine which removed symptoms. Drugs were discontinued at parent’s request, and he was treated with acupuncture and comfrey root tea (S. off.— root) Exact strength and frequency of consumption unknown, but course of treatment was more than 2 years. Flare-up of Crohn’s in 1984 required more prednilosone. June 1986: fever, abdominal pain, and swelling. He was taking prednilosone and sulfasalazine. Had never taken azathioprine. Liver biopsy showed veno-occlusive disease. All other known factors ruled out. 
CASE 2. 49 year-old woman with progressive swelling of the abdomen and extremities over 4 months. Originally diagnosed as Budd-Chiari syndrome, but subsequent workup ruled out known causes. Biopsy confirmed veno-occlusive disease of the small and medium veins. All other known causes of veno-occlusive disease ruled out. Patient was a heavy consumer of herbs, vitamins and natural food supplements, including daily supplements of vitamins C, K, E, A, B-complex, calcium, magnesium, potassium, zinc, iron, lecithin and bovine adrenal extract. Occasional chamomile tea, and 1 qt./day of “Mu-16" tea. Also: 2 capsules of comfrey-pepsin pills with each meal (6 pills per day) for the previous 4 months. The Mu tea and the comfrey pills were analyzed for pyrrolizidine alkaloid content. The Mu tea intake had traces of alkaloids, about 1/10 the level in the comfrey intake. She had consumed about 85 mg. of pyrrolizidine alkaloids over the four months.
CASE 3. 5-day old infant with jaundice, liver enlargement and abdominal swelling. Mother had normal pregnancy, except for rash of unknown origin since 4th month; vaginal bleeding during last 3 days. Delivered by C-section because of unconfirmed suspicion of premature separation of the placenta. Routine investigations for causes of neonatal liver failure were negative. Liver biopsy revealed veno-occlusive disease. Patient died at 27 days of life. Mother had normal liver. Mother acknowledged occasional consumption of cannabis and hallucinogenic mushrooms prior to pregnancy, but strongly denied any such consumption during pregnancy. Investigators ruled out possible mushroom poisoning. Consumed an herbal expectorant tea daily throughout pregnancy. Tea purported to contain 9% coltsfoot (T. farfara) by weight. Isolated senecionine from the sample, but not senkirkine. [calls into question the identification of Tussilago] 
appears from these cases that comfrey was
probably the cause of the disease
in Case 1 and Case 2, but that coltsfoot was not
clearly enough identified
as the offending agent in Case 3. Researchers
tested coltsfoot for pyrrolizidine
alkaloids in 1976, and could isolate only
senkirkine.  The absence of
senkirkine suggests that coltsfoot was not
present, or at least was not
the cause of death in the infant. Senecionine,
one of the most toxic of
the pyrrolizidine alkaloids, is present in a
wide variety of plants, including
Petacites spp. a possible adulterant of
What dose is toxic?
It is not easy to identify a toxic dose of pyrrolizidine alkaloids for several reasons.
1) Pyrrolizidine alkaloids vary widely among themselves in potential toxicity. One alkaloid can be six or more times as toxic as another, and some are not toxic at all. .
2) Plants often contain more than one alkaloid, and proportions may vary from plant to plant and time to time. Comfrey (S. off.) contains at least nine alkaloids  of varying degrees of toxicity.
3) The alkaloid content of plants may vary widely from one plant part to another. A test sample showed that the pyrrolizidine alkaloids in a commercial sample of comfrey root (S. off.) were 10.7 times the level in the leaf. 
4) The content can also vary from season to season. The amount of alkaloids in Russian comfrey (S. uplandicum) leaves can vary at least up to 16 times depending on the time of year and the size of the leaf .
5) The dose may also vary widely depending on age. A two month-old baby died after exposure to 66 mg. of alkaloids over 4 days. A six month old child consumed 70-147 mg alkaloids (from the same Senecio spp.) over 14 days and lived . A 162 mg. dose over 270 days that was fatal to her newborn baby had no toxic effects on the mother in Case 3.
6) There may be differences between one person and another. In a case of four patients in China, closely matched for age and sex, responses to identical doses varied widely. Malnutrition may be a predisposing factor.
Table 1 shows the toxic doses of pyrrolizidine alkaloids in various cases throughout the world. Table 2 shows the amounts of alkaloids in common preparations of comfrey (S. Off.) It is clear from the table that the amount consumed in a comfrey dose could easily be toxic if consumed long enough, or if consumed by a child or pregnant mother.
Table 1: Toxic Doses of Pyrrolizidine Alkaloids
Plant family Pyrrolizidine Time Age Result Reference
or species alkaloids Period
Compositae 630 21 days 27 Mild 11
Compositae 580 19 days 23 Illness 11
Compositae 1350 46 days 26 Illness 11
Compositae 1380 45 days 28 Died 11
Tussilago(?) 162 9 mos. ? No effect 20
Tussilago(?) 62 9 mos. 0 Died 20
Senecio 66 4 days 2 mos. Died 4
Senecio 70-147 14 days 6 mos. Illness 23
Symphytum off. 85 120 days 40 Illness 17
Symphytum off. ? 2-3 years 13 Illness 27
Heliotropium 4000 20 days ? Died 2
Heliotropium ? 1-2 years mixed 25% Died 25
mixed epidemic 16
Table 2: Pyrrolizidine Alkaloids in Comfrey
Dose in 120
Form Daily Dose PA Content days Reference
(S. Off.) (leaf) 6 caps .9 mg 108 mg 10
(root) 6 caps 9.66 1159 mg 10
(leaf) 6 caps .72 mg 85 mg 17
Comfrey root tea (S. Off.)
2 cups 17.0 mg 2040 mg 19
Comfrey leaf (in salad)
1 leaf 32 mg 38.4 mg 13
Robert’s Formula. (traditional anti-ulcer formula containing comfrey rt.)
3 caps .96 mg 115.2 mg. estimate
What of the argument that if comfrey were toxic, it would be known by now after centuries of use? Pyrrolizidine alkaloid poisoning was only discovered in 1954. It is insidious, and can take from 2-13 weeks for onset of symptoms, even after stopping ingestion. In the various epidemics and single cases around the world, none of those who consumed them suspected the plants as the cause. Note that these cases have involved plants used in Ayurvedic medicine , traditional Chinese medicine,[5,9,11] traditional African medicine,  and folk medicines of Jamaica and Mexico [4,23] The Eclectic doctors — careful clinicians — never noted the toxic effects of Senecio spp., now acknowledged by scientists and herbalists alike.[33,34,35]
Researchers in Ethiopia found an herbal healer there prescribing traditional plant remedies to nursing mothers. The herbs he collected contained pyrrolizidine alkaloids, confirmed with analysis, and were later used to induce veno-occlusive disease in rabbits.  The plants are traditionally used to treat stomach troubles and other illnesses. Some of the children of these mothers were diagnosed with kwashiorkor, which has hepatic complications that mimic veno-occlusive disease.
we really be concerned when tons of comfrey root
are consumed each year
in the United States alone, and we can only find
2 cases of comfrey poisoning
in the scientific literature? Possibly so — the
extent of the problem could
be much wider than assumed. Most significant is
that hepatic veno-occlusive
disease has been misdiagnosed as viral hepatitis
or cirrhosis .
The clinical picture of veno-occlusive disease is varied, and can only be confirmed with biopsy, often impractical in acute disease. Many cases may be misdiagnosed by less-than-thorough clinicians. Furthermore, questioning about herb intake is not normal procedure for hepatitis or cirrhosis. The problem is now wider than was originally assumed in third-world countries. Veno-occlusive disease from the consumption of a traditional folk remedy accounts for about 30% of all cirrhosis cases in Jamaica, and incidence fell after an educational campaign there. One author hypothesizes that the high incidence of atypical malnutrition cases in Africa (20% of cases) in children could be due to pyrrolizidine alkaloid poisoning. How many cases of liver disease (one of the ten leading causes of death in the U.S.) might be due to these compounds?
The most serious risk of pyrrolizidine alkaloid poisoning in herbal practice in the U.S. appears to be the use of comfrey root for chronic digestive conditions. Such treatment assumes long-term use, and patients are often weakened from the disease. At least one company has recently removed comfrey root from its Robert’s Formula, a traditional anti-ulcer formula. The lack of warning information about consumption during pregnancy and for small children is also cause for concern.
It should be noted that the Henry Doubleday Research Association (Growers and marketers of comfrey in the United Kingdom) issued a public statement in 1978 which concluded that until further research clarifies the long-term health hazard from comfrey ingestion, “No human being or animal should eat, drink, or take comfrey in any form.” 
Nothing can be said conclusively about coltsfoot. Alkaloids from the plant were apparently not present in the tea that caused the death of the newborn baby in Switzerland. More research needs to be done into the amounts of alkaloids in various mature plant parts and in actual medicinal preparations of coltsfoot. U.S. research into coltsfoot alkaloids used plant buds. Young fast-growing plant parts usually contain much higher levels of alkaloids than mature plants, and it is possible that mature coltsfoot leaves have a greatly reduced hazard. Until such research is done, it remains prudent to avoid coltsfoot during pregnancy, childhood, and for long term-use in adults.
perspective, the public health hazards of
pyrrolizidine alkaloid poisoning
pale when compared with to those of conventional
drugs. Adverse drug reaction
costs the U.S. about 5 billion a year. Ten to
twenty thousand people a
year die from gastric complications of
drugs alone, and recent study shows that overly
aggressive treatment of
high blood pressure with drugs may cause 25,000
heart attacks a year. Nevertheless,
the hypocrisy of regulatory agencies should not
prevent clinicians from
protecting their patients from potential harm.
1. Bove KE et al."Vasculopathic hepatoxicity associated with E-Ferol syndrome in low birth-weight infants." JAMA 1985;254:2423-30
2. Datta DV, et al. “Herbal medicines and veno-occlusive disease in India” Postgraduate Medical Journal, August 1978, 54, 511-15
3. Etzioni A. et. al."Defective humoral and cellular immune functions associated with veno-occlusive disease of the liver. J. Pediatr. 1987;110:549-54
4. Fox DW et al. “Pyrrolizidine (Senecio) intoxication mimicking Reye syndrome,” Journal of Pediatrics. December 1978 p.980-2
5. Hou JC [Veno occlusive disease of the liver with report of 2 cases (author’s translation) Chung Hua Nei Ko Tsa Chih 19:187-91 as quoted in reference 11.
6. Hirono et al. “Carcinogenic activity of coltsfoot, Tussilago farfara.” Gan 1976, 67:125-9
7. Hirono et al. “Carcinogenic Activity of symphytum officinale.” J Natl Cancer Inst. 1978 Vol. 61, No.3 p. 865-8
8. Hirono et al. “Induction of hepatic tumors in rats by senkirkine and symphytine.” J Natl Cancer Inst 1979 vol. 63 No. 2 p. 469-71.
9. Hu SY An enumeration of Chinese materia medica. Hong Kong: Chinese University Press. 1980
10. Huxtable RJ et al. “Toxicity of comfrey-pepsin preparations” NEJM 1986 vol. 315, no. 17, p. 1095.
11. Kumana CR et al. “Herbal tea induced hepatic veno-occlusive disease: quantification of toxic alkaloid exposure in adults.” Gut, 1985, 26, 101-104
12. Mattocks AR. “Toxicity of pyrrolizidine alkaloids” Nature 1968 V. 217 p. 723-9.
13. Mattocks AR. Lancet, 1980. November 22.
14. McLean EK “The toxic actions of pyrrolizidine (Senecio) alkaloids.” Pharmacological Reviews, 1970, 22, 429.
15. McGee JO’D. “A case of veno-occlusive disease of the liver in Britain associated with herbal tea consumption. J Clin Path. 1976, 29,788-794.
16. Mohabbat O, et al. “An outbreak of hepatic veno-occlusive disease in north-western Afghanistan” Lancet, 1970, August 7, p.269-71
17. Ridker PM, et al. “Hepatic Veno-occlusive disease associated with the consumption of pyrrolizidine-containing dietary supplements.” Gastroenterology 1985;88:1050-4
18. Ridker PM, McDermott WV “Comfrey herb tea and hepatic veno-occlusive disease.” Lancet, March 25, 1989 p. 657
19. Roitman JN, “Comfrey and liver damage” 1981 Lancet, April 25, p. 944
20. Roulet M, et al. “Hepatic veno-occlusive disease in newborn infant of a woman drinking herbal tea.” J Pediatrics. 1988 March v.112, no.3, P. 433-436
Schoental R. “Herbal medicines to avoid” Nature,
1972, vol. 238 July 14,
22. Smith LW and Culvenor CCJ. “Plant sources of hepatotoxic pyrrolizidine alkaloids.” Journal of Natural Products. 1981. v.44 no.2. p. 129-52
23. Stillman AE et al. “Hepatic veno-occlusive disease due to pyrrolizidine (Senecio) poisoning in Arizona.” Gastroenterology 1977. 73:349-352
24. Tandon RK et al. “Study of an epidemic of veno-occlusive disease in India,” Gut, 1976, 17, 849-855
25. Tandon HD et al. “An epidemic of veno-occlusive disease of the liver in Afghanistan.” Am J Gastroenterology. 1978 70:607-13.
26. Tandon BN et al. “An epidemic of veno-occlusive disease of liver in central India.” Lancet 1976, August 7, p. 271-2
27. Weston CFM, et al. “Veno-occlusive disease of the liver secondary to ingestion of comfrey.” Brit. Med. J., 1987, July, p. 183
28. Henry Doubleday Research Association. Announcement. Observer July 30, 1978 [Reported in Brit Med J 6163:598, 1979]
29. Personal conversation with researcher Christopher Hobbs: an article published in German has speculated that the actual offending plant in the case was Petacites.
30. Merck Manual: Fifteenth Edition. Merck and Company. 1987.
31. Harrison’s Principles of Internal Medicine. McGraw Hill and Company. 1987.
32. Poisondex. Topic: Plants—Pyrrolizidine Alkaloids. Vol 61 (Exp 8/31/1989) Micromedex, Inc.
33. Felter and Lloyd. Kings American Dispensatory. Eclectic Medical Publications, Portland, OR.
34. Felter HW. The Eclectic Materia Medica, Pharmacology, and Therapeutics. Eclectic Medical Publications. Portland, OR35. New Herbal Practitioner. v.4 no.3 1978 p. 24