The risks of scientific revolutions: manipulating the genetic properties of plants has been heralded by proponents as a boon for the environment and a weapon against famine. But practice so far has produced unpredictable outcomes that urge a cautious approach.

Katherine Barrett, who recently completed her PhD in botany at the University of British Columbia, where she studied Canadian policies for agricultural biotechnology and now works for the Science and Environmental Health Network, a non-government organization based in Halifax, writes that in mid-1998, scientist Dr. Arpad Pusztai announced to the British media that certain genetically engineered potatoes damaged the organs and immune systems of rats. Intense controversy followed this announcement. The Rowett Institute, where Pusztai carried out his research, disputed the experiments, saying that they were poorly designed and inappropriately reported. The institute later suspended Pusztai.

In May 1999, Dr. John Losey and colleagues from Cornell University reported that pollen from Bt corn (engineered to contain a pesticide derived from bacteria) could kill caterpillars of monarch butterflies. Many people saw the monarch study as further evidence that genetically engineered crops pose unknown and under-investigated hazards. Others, however, argued that the study's conclusions were premature, and that laboratory experiments were not a good predictor of events in the field.

In November 1999, members of the biotechnology industry, the U.S. government and the scientific community took the unusual step of holding a conference to debate the significance and soundness of Losey's research. As with Pusztai's rat experiments, the monarch study continues to be highly scrutinized, contested and divisive.

What do these events tell us about the safety and the "real" risks of genetically engineered foods? They tell us, foremost, that these are not simple questions. Developing genetically engineered organisms and predicting their fate in the environment raise questions complex enough to keep researchers busy for a long time. However, as the rat and monarch studies show, scientific questions are closely tied to political stakes, social values and the standards of professional conduct.

Real risks, writes Barrett, are a combination of these factors: There are no easy answers, and no convenient sound bites. It is essential to keep this in mind when tallying what we know, what we believe, and most importantly, what we do not know, about the safety of genetically engineered foods.

Although research and development of genetically engineered plants began in the early 1980s, these crops have been grown on commercial scales and marketed in Canada only since 1995. Our experience with large-scale, uncontained release and human consumption of genetically engineered foods is therefore limited. What have we learned about safety and hazards during this time?

Three broad categories of risks are generally recognized: environmental, public health, and social-economic. Evidence for specific hazards within each of these categories is, says Barrett, rapidly growing, and keeping track of the information and the inevitable debate can be a full-time job. The list below is a sample of risks for which we now have documented evidence.

Environmental Risks
Gene escape: Genes engineered into crops may "escape" to other related plants through pollen. For example, research has shown that genes for herbicide tolerance can be transferred from canola to related weedy plants.

Effects on non-target organisms: The monarch study suggests that genetically engineered crops may have unintended effects on "non-target" plants and animals. Earlier studies demonstrated harm to ladybird and lacewing beetles, but broad-scale testing for non-target effects has not been done.

Pesticide-resistant insects: Extensive use of Bt crops favors insects that can survive the usually fatal dose of Bt pesticide. Populations of healthy, resistant insects mean that the Bt crops as well as Bt sprays are rendered useless. This situation has proven difficult to understand and control, despite farm management plans recently promoted by the Canadian government.

Health Risks
Allergic response: Preliminary research aimed at splicing a brazil nut gene into soybeans showed that a single new protein engineered into a plant could produce allergic responses in humans.

Toxicity: Similarly, Arpad Pusztai's rat studies suggest that genetic engineering could introduce new toxins into our food supply. Health Canada may regulate genetically engineered foods containing known allergens and toxins. However, genetic engineering enables the creation of completely new proteins that may cause unprecedented health effects. This risk can only be assessed through direct laboratory testing of all genetically engineered foods.

Antibiotic resistance: Many genetically engineered crops contain genes that code for resistance to antibiotics as well as genes that code for the intended new characteristic. These antibiotic resistance genes may be inadvertently transferred to micro-organisms, making them, and the diseases they cause, more difficult to control. While conclusive tests have not been done, the potential consequences of increased antibiotic resistance have prompted the British Medical Association to call for a ban on the use of these genes in all genetically engineered foods.

New pathogens: Research on genetically engineered plants containing viral genes has shown that these genes can combine in the plant with other viral strains to produce new pathogens. Because pieces of viral genes are used in many genetically engineered foods, there is growing speculation that widespread use of these crops may result in new human, animal and plant pathogens.

Social and Economic Risks
Many of the social and economic consequences of genetic engineering fall under the broad heading of "food security" -- that is, our ability to ensure a healthful, affordable and diverse food supply at the level of communities and individuals. Of course, this is often the crux of the debate: Biotechnology advocates claim genetically engineered food is necessary to feed the world, while critics highlight the potentially devastating effects of increased corporate control, patents on life forms, and replacement of traditional crops and agricultural knowledge.

Barrett says that detailed studies of the economic impacts of genetically engineered crops (for example, on yield and chemical use) have been published only in the past year, and the results are variable. In some cases, yields were significantly reduced, but in other cases use of chemical pesticides was also reduced. As might be expected, the performance of genetically engineered crops seems to depend on local conditions such as the number of weeds and insect pests, weather and soil types. Few of the genetically engineered foods on the market aim to improve nutritional quality.

We have evidence that genetically engineered foods may sometimes be harmful. However, scientific conclusions are rarely absolute, and a great deal of uncertainty remains. Perhaps the most urgent question when facing such uncertainty is: How can we make better decisions?

Barrett says we might first examine the current regulatory process. To date, decisions about the safety of genetically engineered food in Canada have fallen mainly to the federal government and the biotechnology industry -- a very narrow range of interests and expertise. The complexity and potential stakes of genetic engineering demand a process that is more open and accessible to the Canadian public.

Second, we might step back from the details of current debate to raise some larger questions. What kind of food and agriculture do we want and need? Does genetic engineering take us closer to, or further from, these goals? What alternatives to genetically engineered food are available or could be made available with comparable time and research effort?