The possibility that we might see an increase in the number of allergic reactions to food as a result of genetic engineering has a powerful emotional appeal. Parents of a severely allergic child, already conditioned to scrutinize food labels and to watch for warning symptoms, are understandably concerned that an additional and unpredictable danger may have been introduced into the food they buy. Even people who have never experienced an allergic reaction may worry that they are being exposed to new substances for which there is little track record of safety or harm.

However, there is no evidence so far that genetically engineered foods are more likely to cause allergic reactions than are conventional foods. The genetic engineering process itself does not create allergens. The nature of the genes that are chosen for transfer will determine whether allergens are introduced into the engineered host plant. The Food and Drug Administration reviews proposed transgenic foods and compares possible allergens to a checklist of characteristics that have been found to be associated with allergenicity. In several years of testing dozens of proposed transgenic crops, only two potential problems have been uncovered: a soybean that was withdrawn from development and the now-famous StarLink corn.

Source: FDA

The soybean, developed by Pioneer Hi-Bred, was modified by the introduction of a gene from the Brazil nut. The intention was to improve the nutritional quality of the soybean by adding the amino acid methionine, which is low in soybean and abundant in the Brazil nut. Because allergies to nuts are common, the use of a nut as the donor of the transgene raised eyebrows among regulators. The transgenic soybean was tested (Nordlee et al., 1996) and people who had allergic reactions to Brazil nuts also had allergic reactions to the genetically engineered soybean. It appeared that the gene chosen to improve nutritional quality was one of the genes that trigger allergic reactions. Pioneer had intended to market the soybean only for animal feed, but the difficulty of keeping animal feed separate from human food during harvesting, transport, and storage became a consideration. Pioneer decided not to ask for approval to market the soybean. It was never approved by the government and was never grown commercially or sold in stores. More information on this is available at http://www.comm.cornell.edu/gmo/issues/brazilnut.html.

StarLink, developed by the company Aventis, was originally intended to be sold as an all-purpose corn, but concern that it might be allergenic led to its approval only for animal feed. Although the 2001 finding is that StarLink corn is probably not allergenic, the scientific debate continues, with a panel of experts still unable to reach a consensus about the adequacy of the evidence. Some of the points of debate are presented below.

• Data provided by Aventis suggested that StarLink protein was degraded by wetting and heating, so that commercially processed foods and foods prepared in the home by mixing and baking would contain only low levels of the intact protein. The advisory panel suggested that higher levels of the protein might still be present and capable of causing allergic reactions, but might be undetectable via current tests, if wetting and heating changed the shape of the molecule.
• Although the Centers for Disease Control and Prevention did not find evidence that the StarLink protein had caused allergic reactions in people who reported adverse health effects from eating corn-based products, the scientists on the panel said the CDC's test might not be sufficiently sensitive.
• Physicians have not reported seeing cases of possible allergenicity to StarLink corn. But the panel members suggested that the government should be more aggressive in collecting information about cases that may have gone unreported.

Readers with an interest in this issue are encouraged to review the report (http://www.epa.gov/scipoly/sap/2001/july/julyfinal.pdf) by the FDA's panel of independent scientists concerning the validity of the tests.

Although the issue of allegenicity remains unresolved, the panel noted that the government's estimate in early 2001 of about 0.4% StarLink in the nation's human food supply of corn was probably an overestimate and that aggressive actions by the government and by Aventis had reduced the amount to less than 0.125% by the summer of 2001.

Testing for allergenicity is a complex process that utilizes both test-tube reactions and responses in live subjects. People and animals can be used as live subjects, and researchers are considering various animals, including rats, dogs, pigs, guinea pigs, and mice, that might serve as useful substitutes for tests on people. Animals may not always be good predictors of human allergic reactions. Before the research on the Brazil nut soybean (Nordlee et al., 1996), other researchers (Melo et al., 1994) had tested the same Brazil nut protein in rats and mice and decided that it was not a major allergen. Improvements in testing for allergens will probably be spurred by the increased demand for such testing in genetically engineered crops.

Just as some people every year discover that they have developed an allergy to common foods such as wheat or eggs, or to newly popular foods such as kiwifruit, some people may develop allergies to transgenic foods despite regulatory efforts to weed out potential problems. This is undeniably serious for people who develop allergic responses, but there is currently no evidence that transgenic foods pose more of a risk than conventional foods do.

More information about allergenicity in transgenic foods is available from the Pew Initiative on Food and Biotechnology at http://pewagbiotech.org/buzz/display.php3?StoryID=12 and from the Royal Society's 2002 report "Genetically modified plants for food use and human health--an update" (http://www.royalsoc.ac.uk/files/statfiles/document-165.pdf).