Food for Thought

Janet Raloff
Food for Thought

Honey of a Threat

Sponsor Message

An all-natural, organic food, honey has a benign–if not wholesome–image. Many people consider it a superior alternative to table sugar and corn syrup–two primary sweeteners in the U.S diet.

However, honey has the potential to carry some very disturbing plant poisons to the dinner table, an international trio of scientists reports. The World Health Organization has already identified these toxins–pyrrolizidine alkaloids–as a serious human health threat. In Europe, the toxins are coming under regulation, but only where they occur in herbal medicines.

Honeys can be more potent sources of these chemicals, John A. Edgar, a consulting toxicologist from Melbourne, Australia, and his colleagues note in an upcoming issue of the Journal of Agricultural and Food Chemistry. Moreover, they argue, because honey aficionados sometimes down large quantities of the syrup daily, the potential for unwitting poisoning among this segment of the population might be high.

"We dont want people to stop buying honey," Edgar emphasizes. However, his team argues, if these toxins warrant regulation in herbal medicines, they probably should at least be monitored in foods that, like honey, might be eaten by infants, children, and the elderly–populations believed most vulnerable to pyrrolizidine-alkaloid toxicity.

Newborn fatality linked to Moms tea

The idea that these populations are selectively vulnerable to pyrrolizidine alkaloids stems from findings like the case reported in the March 1988 Journal of Pediatrics. Swiss physicians described a baby born with jaundice, a massively enlarged liver, and other problems. About 1 month later, the baby died.

Though the infant had experienced liver failure, her doctors were stumped as to why. The newborn didnt have hepatitis or any other common liver-damaging diseases.

The first clue to suggest poisoning came at the autopsy, which showed massive blockage of vessels draining blood from the liver, noted Michel Roulet of the University of Lausanne and his colleagues. Upon questioning, the mother reported that she had drunk an herbal tea throughout her pregnancy. She made the tea from dried leaves purchased at her neighborhood pharmacy, where the nostrum is sold as a phlegm-inhibiting cough suppressant. Chemical analyses of this product showed it to be a rich source of senecionine, a pyrrolizidine alkaloid (PA).

The babys condition, known as hepatic veno-occlusive disease, is a hallmark of PA poisoning. In some cases, the alkaloids may also cause cirrhosis of the liver. "Normal detoxification processes in the liver actually activate PAs," Edgar explains, turning these chemicals into potent poisons. Not surprisingly, the liver then becomes their primary victim.

While fatal to the infant, the tea had no noticeable effect on the mother.

Source in many flowers

Some 3 percent of flowering species make PAs, according to the upcoming Journal of Agricultural and Food Chemistry paper. Many of those have a reputation as weeds–such as Patersons curse (Echium plantigineum), tansy ragwort (Senecio jacobaea), and Chromolaena. A few are plants distributed internationally by nurseries and home-garden centers.

The Swiss researchers, in contacting the manufacturer of the herbal tea responsible for the babys poisoning, learned that it contained leaves from 10 plants–including Tussilago farfara. This PA-rich plant–known commonly as coltsfoot or coughwort–constituted some 9 percent of the herbal preparation.

Several other PA-producing herb-garden plants may be plucked for use in salads or home medicines. These include comfrey (Symphytum officinale), Russian comfrey (Symphytum X. uplandicum; formerly S. pergrinum), and borage (Borago officinalis). The first two have a history of folk use as emollients, sedatives, astringents, wound-healing agents, and expectorants. Borage is traditionally recommended for soothing abraded mucus membranes, as a diuretic, and as a stimulant. Borage oil also is a rich source of an omega-3 fat, which some studies suggest can fight inflammation.

A few books on herbal medicine offer stern warnings of the risks posed by ingestion or topical use of PA-containing plants. For instance, Steven B. Karchs Consumer Guide to Herbal Medicine concludes this about comfrey: "The danger of PA exposure has been very clearly demonstrated. Occasional use of a tea made from second growth leaves is unlikely to do much harm, but consumers who buy herbal remedies have no way to be certain which part of the plant they are buying." Since liver failure "is not a nice thing," Karch argues that "[t]aking comfrey . . . would hardly seem to be worth the risk."

Other tomes downplay concern about certain PA-producing plants. For instance, 7 years after the Journal of Pediatrics paper on the Swiss poisoning from coltsfoot, the Herb Society of America Encyclopedia of Herbs and Their Uses pronounced: "The herb contains PAs, which have not proved toxic at low dosages in tests." This book offered no warning of possible fetal risks and concluded that "there is no suggestion that T. farfara should be banned for internal use."

Whos a consumer to believe?

The World Health Organization, for one, Edgar says. After reviewing a huge body of data, it concluded in a 1988 report that PA-tainted foods have triggered several poisoning outbreaks–with high mortality. Moreover, it reported that PAs produce delayed, progressive, chronic effects following even brief, low-level exposures. Finally, the WHO noted that PAs cause mutations, can damage genes, and have triggered cancer in rats.

In 1992, the German government responded by setting strict limits on the concentrations of PAs it would tolerate in herbal medicines. Regulations limit the alkaloids concentrations to 1 microgram per daily dose of a drug in medicines that will be ingested. Such products are to be used for no more than 6 weeks annually–and never by children or pregnant women. For herbal preparations that might be consumed regularly, Germany cut tolerable PA concentrations lower still, to 0.1 microgram per daily dose.

Various countries in the European Union have adopted or are now considering adoption of similar limits, Edgar notes–and the EU has just created a committee to investigate the development of such limits to be applied consistently throughout its member nations.

What about foods?

PA regulations have focused on herbal medicines. The irony, Edgar says, is that studies have demonstrated that the same troublesome poisons can turn up in foods, such as honey, at concentrations that exceed Germanys regulated limits for herbal drugs.

For instance, the scientists note one study that found PA concentrations of roughly 4 micrograms per gram in a honey. This amount is not detectable by taste. Bees appeared to have picked up the toxins when they visited tansy ragwort, a weed common throughout much of the world, including the western United States.

Considering that some infants consume up to 32 g of honey per day and adults up to 90 g daily, there is potential for high PA intake.

Identifying the general source of the nectar collected by bees wouldnt necessarily flag PA contamination. Edgar says that a honey might be labeled as coming from a particular type of flower, such as clover, when that flower represents only 45 percent of nectar the bees collected. However, even if 95 percent of the nectar came from the labeled source, the honey still might carry traces of PAs.

Just a few of the bees in a hive going to a tansy ragwort, Edgar says–were talking one bee in a thousand–could spike the resulting honey with PA concentrations that would be banned under German law for herbal drugs. Since apiculturists dont tether their bees, theres no way to keep rogue buzzers from harvesting nectar at any flower nearby.

And PA-producing weeds can often be found in and amongst crop plants, especially in developing countries. Indeed, Edgar notes, many Asian mass-poisoning outbreaks occurred when local crops were harvested together with adjacent PA-rich weeds, which tainted the batches of grain sent on to market.

A plea for research

At present, no one knows how big or small a threat the PAs in honeys pose. The reason, Edgar laments, is that nobody has investigated the issue on more than a cursory, academic level.

Perhaps a half dozen studies have detected PAs in honey, he notes–but only because PA researchers had reason to believe bees had visited plants producing the alkaloids.

Contributing to his concern about the issue are several reports in publications aimed at beekeepers about the potential advantages of encouraging their colonies to harvest nectar from borage, a plant increasingly grown as a source of a reputedly healthy vegetable oil that has a specialized market.

For instance, in a 1997 article in The Beekeepers Quarterly, a British publication, Lester Quayle notes that borage cultivation has been developed in England and the expertise exported to the world. He goes on to note that Although there seems to be no recommended ratio of hives per acre, [one beekeeper] feels that the more the merrier.

A beekeeper himself, Quayle notes that borage honey is startlingly clear,. . . slow to crystallize and ideal for cut comb and chunk honey. He adds that the health benefits attributable to borage oil might help to build a big following for the plants honey.

A 1999 report in Bee World notes that borage is so attractive to bees that it is commonly grown by people to invite bees into their garden. With this intentional nectar harvesting from borage, Edgar says, one has to worry about potential for PA contamination of that honey. Yet, to date, there doesnt appear to have been any investigation of PAs in this nectar or the honey that results.

Edgars team argues that researchers should look into various types of honey. All three authors of the new Journal of Agricultural and Food Chemistry review have spent parts of their careers investigating human PA-poisoning incidents. Two of the authors, Edgar and Russell J. Molyneux of the Agriculture Departments Western Regional Research Center in Albany, Calif., have also researched the toxins effects on livestock and foods made from tainted animals.

They teamed up for the new review paper, Edgar told Science News Online, in the hope that it would prompt more monitoring and more awareness of the potential problem. Personally, Id also like to see this issue more widely debated in the food-safety risk-assessment community.

PAs are only one of many potent trace toxicants lacing the food supply. However, as with aflatoxin on peanuts and fumonisins on corn flakes, these near ubiquitous poisons can be kept to very low concentrations by routine monitoring–and a culling of any highly contaminated batches. Edgar says that theres even the prospect of one day deliberately breeding lower production of PAs among plants destined for our gardens.


John A. Edgar

c/o CSIRO Livestock Industries

PB 24


Victoria 3220

Further Reading

Raloff, J. 2002. Honey may pose hidden toxic risk. Science News 161(May 18):317. Available to subscribers at [Go to].

______. 1998. The color or honey. Science News 154(Sept. 12):170-171. Available at [Go to].

______. 1997. Sphinx of fats. Science News 151(May 31):342-343.

______. 1992. Liver cancer: Homing in on the risks. Science News 142(Nov. 7):308.

Schatzki, T.F., and W.F. Haddon. In press. Rapid, non-destructive selection of peanuts for high aflatoxin content by soaking and tandem mass spectrometry. Journal of Agricultural and Food Chemistry.

Wu, C. 1997. Chemical buzz. Science News 151(May 24):324-325. Available at [Go to].

More from Science News