-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-02-566						        

TITLE:     GENETICALLY MODIFIED FOODS: Experts View Regimen of Safety 
Tests as Adequate, but FDA's Evaluation Process Could Be Enhanced

DATE:   05/23/2002 
				                                                                         
----------------------------------------------------------------- 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-02-566
     
Report to Congressional Requesters

United States General Accounting Office

GAO

May 2002 GENETICALLY MODIFIED FOODS

Experts View Regimen of Safety Tests as Adequate, but FDA's Evaluation
Process Could Be Enhanced

GAO- 02- 566

Page i GAO- 02- 566 Genetically Modified Foods Letter 1

Results in Brief 2 Background 4 GM Foods Share the Same Types of Health
Risks as Conventional

Foods and Are Evaluated by Tests That Appear Adequate 9 FDA Believes That It
Obtains Necessary Safety Data for GM Food

Evaluations, but FDA?s Overall Evaluation Process Could Be Enhanced 18
Future Changes in GM Foods May Alter Their Composition and

Require New Testing Technologies for Assessing Safety 25 Experts Maintain
That Long- Term Monitoring of GM Foods Is

Neither Necessary nor Feasible 30 Conclusions 32 Recommendations for
Executive Action 32 Agency Comments 33

Appendix I Objectives, Scope, and Methodology 35

Appendix II Potential Use of Metabolic Profiling to Compare a GM Plant to
Its Conventional Counterpart 38

Appendix III Comments from the Department of Health and Human Services 43

Appendix IV GAO Contacts and Staff Acknowledgments 46

Related GAO Products 47

Tables

Table 1: GM Foods for Human Consumption Evaluated by FDA 6 Contents

Page ii GAO- 02- 566 Genetically Modified Foods Figures

Figure 1: Use of Biotechnology to Create a Pest- Resistant Plant 5 Figure 2:
Example of the Regimen of Tests Used for Safety

Assessments of GM Foods 13 Figure 3: Metabolic Profiling of Three GM Mustard
Plant Varieties

in Comparison with Baseline Data from Their Conventional Counterparts 39

Abbreviations

APHIS Animal and Plant Health Inspection Service CFSAN Center for Food
Safety and Applied Nutrition DNA deoxyribonucleic acid EPA Environmental
Protection Agency FAO Food and Agriculture Organization FDA Food and Drug
Administration GM genetically modified PBN Pre- Market Biotechnology Notice
RNA ribonucleic acid USDA U. S. Department of Agriculture WHO World Health
Organization

Page 1 GAO- 02- 566 Genetically Modified Foods

May 23, 2002 The Honorable John E. Baldacci The Honorable John F. Tierney
House of Representatives

Proponents and opponents of modern agricultural biotechnology hold
passionate views about the benefits and risks of using this technology to
produce genetically modified (GM) food. 1 Proponents cite enhanced crop
yields, more environmentally friendly food production, and more nutritious
foods as reasons to move forward. Opponents of biotechnology argue that not
enough is known about the safety of these foods and that they should be more
rigorously controlled.

While confidence in the safety of GM foods is essential to their commercial
success, governments and consumers from different parts of the world have
taken very different positions on their safety and regulation. Some
consumers in Western Europe have shown their opposition to this technology
by destroying GM food crops, and European regulatory entities have not
approved any new GM foods in the past several years. In the United States,
consumers and regulatory agencies, such as the Food and Drug Administration
(FDA), generally support GM foods, with a number of these foods having been
made available for sale in recent years. However, the debate on the safety
of these foods is ongoing and may intensify in the future as genetic
modifications to foods become increasingly complex.

To ensure public confidence in GM foods, the U. S. biotechnology industry
recognized in the early 1990s the need for oversight by FDA, which has
primary responsibility for the safety of most of the nation?s food supply. 2
In response, FDA published guidelines in 1992 to ensure that companies
developing GM foods worked with the agency in assessing the safety of these
foods. As part of the process, companies test new GM foods to

1 Modern agricultural biotechnology is a collection of scientific
techniques, such as genetic engineering, used to modify plants, animals, or
microorganisms by introducing desired traits in them, including
characteristics from unrelated species. For example, traits may be
introduced to facilitate pest management and improve yield or nutritional
value. In this report, we will refer to food derived from genetically
modified plants as GM foods.

2 The U. S. Department of Agriculture?s (USDA) Food Safety and Inspection
Service is responsible for ensuring the safety of meat, poultry, and egg
products. FDA is responsible for all other foods, including seafood.

United States General Accounting Office Washington, DC 20548

Page 2 GAO- 02- 566 Genetically Modified Foods

assess their safety, including their potential health risks, and submit test
data to FDA for evaluation. 3 As of April 2002, FDA has evaluated 50 GM
foods, many of which have subsequently been placed on the market. Currently,
submission of information to FDA is voluntary, but FDA published a proposed
rule in January 2001 that would make this submission mandatory.

In light of the continued debate about GM foods, you asked us to (1)
identify the types of potential human health risks associated with GM foods
and experts? views regarding the adequacy of tests used to evaluate these
risks, (2) describe FDA?s controls for ensuring that companies submit test
data it requests and identify experts? views of the agency?s overall
evaluations of these foods, (3) describe potential changes in future GM
foods and any associated changes in tests to evaluate them, and (4) identify
experts? views on the necessity and feasibility of monitoring the long- term
health risks of these foods.

To conduct this work, we reviewed scientific and technical studies and other
literature and spoke with experts in government, academia, private industry,
and consumer groups. We selected these experts in consultation with
officials from the National Academy of Sciences. As agreed with your
offices, we did not assess the potential environmental risks associated with
GM food production. In addition, since there have been no GM animals
evaluated for commercialization, we did not assess their potential
environmental or human health risks. Also, we did not independently evaluate
FDA?s controls for ensuring it receives safety data. Further details of the
scope and methodology of our review are discussed in appendix I.

GM foods pose the same types of inherent risks to human health as
conventional foods: they can contain allergens, toxins, and compounds known
as antinutrients, which inhibit the absorption of nutrients. Before
marketing a GM food, company scientists evaluate these risks- even though
they are not routinely evaluated in conventional foods- to determine if the
foods pose any heightened risks. While some GM foods have contained
allergens, toxins, and antinutrients, the levels have been comparable to
those foods? conventional counterparts. In evaluating GM foods, scientists
perform a regimen of tests. Biotechnology experts whom

3 In general, risk assessment involves several steps, including identifying
the hazard (type of risk) and assessing the level of exposure to the hazard.
Results in Brief

Page 3 GAO- 02- 566 Genetically Modified Foods

we contacted agree that this regimen of tests is adequate in assessing the
safety of GM foods. While some consumer groups, as well as some scientists
from the European Union, have questioned the ethical or cultural
appropriateness of genetically modifying foods, experts whom we contacted
from these organizations also believe the tests are adequate for assessing
the safety of these foods.

While FDA reports that its evaluation process includes the necessary
controls for ensuring it obtains the safety data needed to evaluate GM
foods, some biotechnology experts state that aspects of its evaluation
process could be enhanced. FDA?s controls include (1) communicating clearly-
through the agency?s 1992 policy statement and subsequent guidance- what
safety data are necessary for its evaluations of GM food safety; (2) having
teams of FDA experts in diverse disciplines evaluate company submissions for
GM foods and request additional safety data, if necessary; and (3) tailoring
the level of evaluation to match the degree of each submission?s novelty,
thereby assuring that staff have time to obtain necessary safety data.
Nonetheless, FDA?s overall evaluation process could be enhanced, according
to some experts, by randomly verifying the test data that companies provide
and by increasing the transparency of the evaluation process- including
communicating more clearly the scientific rationale for the agency?s final
decision on a GM food safety assessment.

In the future, scientists generally expect that genetic modifications will
increasingly change the composition of GM foods to enhance their nutritional
value. For example, one company has modified a type of rice to contain beta-
carotene. In countries where rice is a dietary staple, this rice may reduce
the incidence of blindness caused by vitamin- A deficiency. Current tests
have been adequate for evaluating the few GM foods with relatively simple
compositional changes that FDA has reviewed so far. New testing technologies
are being developed to evaluate the increasingly complex compositional
changes expected. Some scientists view these new technologies as a
potentially useful supplement for existing tests, while others believe that
the technologies will offer a more comprehensive way of assessing the safety
of all changes in GM foods.

Monitoring the long- term health risks of GM foods is generally neither
necessary nor feasible, according to scientists and regulatory officials we
contacted. In their view, such monitoring is unnecessary because there is no
scientific evidence, or even a hypothesis, suggesting that long- term harm
(such as increased cancer rates) results from these foods. Furthermore,
there is consensus among these scientists and regulatory officials that
technical challenges make long- term monitoring infeasible.

Page 4 GAO- 02- 566 Genetically Modified Foods

Experts cite, for example, the technical inability to track the health
effects of GM foods separately from those of their conventional
counterparts. A recent report by food and health organizations affiliated
with the United Nations also expresses skepticism about the feasibility of
identifying longterm health effects from GM foods.

This report contains recommendations to the Deputy Commissioner of Food and
Drugs for enhancing the effectiveness of FDA?s safety evaluations of GM
foods. The recommendations concern the need to randomly verify test data and
increase the transparency of the agency?s safety evaluations of these foods.
In commenting on a draft of this report, FDA agreed with our recommendations
and stated that the recommendations would increase the transparency of, and
public confidence in, FDA?s evaluations of GM foods. FDA also provided
technical comments which we incorporated as appropriate.

Modern agricultural biotechnology refers to various scientific techniques,
most notably genetic engineering, used to modify plants, animals, or
microorganisms by introducing into their genetic makeup genes for specific
desired traits, including genes from unrelated species. For centuries people
have crossbred related plants or animal species to develop useful new
varieties or hybrids with desirable traits, such as better taste or
increased productivity. Traditional crossbreeding, however, can be very
time- consuming because it may require breeding several generations to
obtain a desired trait and breed out numerous unwanted characteristics.
Genetic engineering techniques allow for faster development of new crop or
livestock varieties, since the genes for a given trait can be readily
introduced into a plant or animal species to produce a new variety
incorporating that specific trait. Additionally, genetic engineering
increases the range of traits available for developing new varieties by
allowing genes from totally unrelated species to be incorporated into a
particular plant or animal variety.

In the 1970s, scientists learned how to extract a specific gene from a DNA
strand and insert this gene into a different organism where it would
continue to make the same protein that it did in its original organism.
Scientists have applied this technology to bacteria, plants, and animals.
For example, as shown in figure 1, scientists produced pest- resistant
plants by identifying a gene responsible for pest resistance in an organism,
isolating and copying the gene, and then inserting it into the target
plant?s DNA. The plant was then tested to determine that the transferred
trait Background

Page 5 GAO- 02- 566 Genetically Modified Foods

(transgene) was inherited in subsequent generations and that the
?transgenic? plant grew and functioned as well as the conventional variety.

Figure 1: Use of Biotechnology to Create a Pest- Resistant Plant

Biotechnology offers a variety of potential benefits and risks. It has
enhanced food production by making plants less vulnerable to drought, frost,
insects, and viruses and by enabling plants to compete more effectively
against weeds for soil nutrients. In a few cases, it has also improved the
quality and nutrition of foods by altering their composition. Table 1
summarizes the GM foods evaluated by FDA.

Page 6 GAO- 02- 566 Genetically Modified Foods

Table 1: GM Foods for Human Consumption Evaluated by FDA Modified Attribute

Insect Resistance

Viral Resistance

Herbicide Tolerance

Modified Oil

Plant Reproductive Sterility

Delayed Ripening/ Softening

GM Plant Product-# of Plant Varieties Corn- 8

Tomato- 1 Potato- 4 Cotton- 2

Squash- 2 Papaya- 1 Potato- 2

Corn- 9 Rice- 1 Canola- 8 Sugar Beet- 2 Flax- 1 Cotton a -4 Radish- 1
Soybean- 2

Soybean- 1 Canola- 1

Corn- 3 Canola- 3 Radish- 1

Cantaloupe- 1 Tomato- 4

Total b 15 5 28 2 7 5

a Cotton seed has been used as a protein source in candy. b Fifty products
have been evaluated, as of April 2002. The total number of modified
attributes is 62 because several products were modified with multiple
attributes. Source: GAO analysis of FDA data.

Table 1 shows that the majority of modifications have been aimed at
increasing crop yields for farmers by engineering a food plant to tolerate
herbicides or attacks from pests such as insects and viruses (48 out of 62
modifications). Further, only two food plants have been altered to produce
modified oil: the soybean and canola plants. According to industry
officials, the modified soybean produces healthier oil. They also stated
that the canola plant was modified to have a domestic source for laurate
cooking oil. 4 Because soybean oil is the most commonly consumed plant oil
worldwide, scientists say that the new oil could significantly improve the
health of millions of people.

For three key crops grown in the United States- corn, soybeans, and cotton-
a large number of farmers have chosen to plant GM varieties. In 2001, GM
varieties accounted for about 26 percent of the corn, 68 percent of the
soybeans, and 69 percent of the cotton planted in the United States. These
crops are the source of various ingredients used extensively in many
processed foods, such as corn syrup, soybean oil, and cottonseed oil, and
they are also major U. S. commodity exports. The United States accounts for
about three- quarters of GM food crops planted globally.

4 Laurate canola oil is a form of canola oil that contains lauric acid, a
fatty acid found in tropical oils. As a result, laurate canola oil can
substitute for palm kernel oil which is an imported tropical oil.

Page 7 GAO- 02- 566 Genetically Modified Foods

However, the use of biotechnology has also raised concerns about its
potential risks to the environment and people. For example, some people fear
that common plant pests could develop resistance to the introduced
pesticides in GM crops that were supposed to combat them. Further, some fear
that crops modified to be tolerant to herbicides could foster the evolution
of ?super weeds.? Finally, some fear that scientists might unknowingly
create or enhance a food allergen or toxin. Therefore, as biotechnology was
being developed, U. S. scientists, regulators, and policymakers generally
agreed that GM plants should be evaluated carefully before being put into
widespread use. As a result, the United States published a Coordinated
Framework for Regulation of Biotechnology in 1986. This framework outlined
the regulatory approach for reviewing GM plants, including relevant laws,
regulations, and definitions of GM organisms.

Responsibility for implementing the coordinated framework fell primarily to
three agencies: USDA, the Environmental Protection Agency (EPA), and FDA.
Within USDA, the Animal and Plant Health Inspection Service (APHIS) bears
the main responsibility for assessing the environmental safety of GM crops.
The primary focus of APHIS? review is to determine whether or not a plant
produced through biotechnology has the potential to harm natural habitats or
agriculture. Developers can petition APHIS to exempt a GM plant from
regulation once sufficient and appropriate data have been collected
regarding the potential environmental impact of a GM plant.

To safeguard the environment and human health, EPA is responsible for
regulating genetic modifications in plants that protect them from insects,
bacteria, and viruses. These protectants are subject to the agency?s
regulations on the sale, distribution, and use of pesticides. EPA must
review and grant a permit for field- testing plants with such protectants on
more than 10 acres of land. Prior to commercialization of a GM plant with
such a protectant, EPA reviews the application for approval of the
protectant, solicits public comments, and may seek the counsel of external
scientific experts.

FDA has primary authority for the safety of most of the food supply. The
Federal Food, Drug, and Cosmetic Act establishes the standard for food

Page 8 GAO- 02- 566 Genetically Modified Foods

safety as food being in an unadulterated condition. 5 FDA established its
basic policy regarding the review of GM foods in its 1992 Policy on Foods
Derived from New Plant Varieties. According to this policy, FDA relies on
companies developing GM foods to voluntarily notify the agency before
marketing the foods. 6 Notification leads to a two- part consultation
process between the agency and the company that initially involves
discussions of relevant safety issues and subsequently the company?s
submission of a safety assessment report containing test data on the food in
question. At the end of the consultation, FDA evaluates the data and may
send a letter to the company stating that the agency has no further
questions, indicating in effect that it sees no reason to prevent the
company from marketing the GM food. In 1997, FDA supplemented its 1992
Policy with the current

Guidance on Consultation Procedures, clarifying procedures for the initial
and final consultations.

In January 2001, FDA issued a proposed rule in the Federal Register that
provides further information on these procedures and, more importantly,
would require pre- market notification by companies. Among the reasons that
FDA cited for this change are concerns expressed by consumers and public
interest groups about the limited transparency and voluntary nature of the
current process. FDA also pointed to the growing power of biotechnology to
create potentially more complex safety issues that could require more
stringent regulatory evaluations. FDA, tentatively, expects to finalize this
rule as early as fiscal year 2003.

5 21 U. S. C. sect. 342. A food is deemed adulterated if, among other things, it
contains an added poisonous or deleterious substance that may render the
food injurious to health or if it contains an unapproved food additive.

6 FDA has not routinely evaluated the safety of foods derived from new plant
varieties that were produced by traditional breeding. FDA has generally not
evaluated these foods because plant breeders have longstanding established
and reliable practices for ensuring food safety.

Page 9 GAO- 02- 566 Genetically Modified Foods

All foods, including those from GM plants, pose the same types of inherent
risks to human health: they can cause allergic or toxic reactions, or they
can block the absorption of nutrients. Although some foods from GM plants
have contained allergens, toxins, and antinutrients, scientists agree that
the levels of these compounds have been comparable to those found in the
foods? conventional counterparts. To reach such a finding, each GM food is
evaluated using a regimen of tests. This regimen begins with tests on the
source of the gene being transferred, proceeds to tests examining the
similarity of the GM food to conventional varieties with known allergens,
toxins, and antinutrients, and may include tests on the safety of the
modified protein from the GM food in simulated digestive fluids. At every
phase, test results are compared to the risk levels found in the food?s
conventional counterpart. If the risk levels are within the same range as
those for the conventional food, the GM food is considered as safe as its
conventional counterpart. Despite the limitations of individual tests,
several experts agree that this regimen of tests has been adequate for
ensuring the safety of GM foods.

According to reports from the Organization for Economic Cooperation and
Development, the Codex Alimentarius, 7 and FDA, foods from GM plants pose
three types of risk to human health: they can potentially contain allergens,
toxins, or antinutrients. These risks are not unique to GM foods. People
have consumed foods containing allergens, toxins, and antinutrients
throughout human history. The small percentage of the population with food
allergies (1- 2 percent of adults and 6- 8 percent of children) tries to
prevent allergic reactions by avoiding offending foods. Additionally, people
commonly consume toxic substances in foods, but they usually do so at levels
that are considered safe. People also frequently consume foods containing
antinutrients, such as certain proteins that inhibit the digestion of
nutrients in the intestinal tract, but common food preparation techniques,
such as cooking, break down the antinutrients. Moreover, consumption of a
varied diet, in which a person is exposed to multiple nutrient sources,
mitigates the risk of malnutrition from antinutrients, according to FDA
officials and various academicians.

7 Codex Alimentarius is the joint food standards program for the United
Nations? Food and Agriculture Organization and the World Health Organization
that was established in 1962. Its objectives are to help protect the health
of consumers and facilitate trade through the establishment of international
food standards, codes of practice, and other guidelines. GM Foods Share the

Same Types of Health Risks as Conventional Foods and Are Evaluated by Tests
That Appear Adequate

All Foods Share the Same Three Risks, Which Are Evaluated in GM Foods

Page 10 GAO- 02- 566 Genetically Modified Foods

Because conventional foods contain allergens, toxins, and antinutrients,
scientists recognize that food cannot be guaranteed to pose zero risk. The
primary concern with the genetic modification of food with respect to human
health, state industry officials, is the potential for unintentional
introduction of a new allergen, an enhanced toxin, or an enhanced
antinutrient in an otherwise safe food. For this reason, developers evaluate
GM foods to determine if they are as safe as their conventional
counterparts.

An allergic reaction is an abnormal response of the body?s immune system to
an otherwise safe food. Some reactions are life threatening, such as
anaphylactic shock. 8 To avoid introducing or enhancing an allergen in an
otherwise safe food, the biotech food industry evaluates GM foods to
determine whether they are ?as safe as? their natural counterparts. For
example, in 1996 FDA reviewed the safety assessment for a GM soybean plant
that can produce healthier soybean oil. As part of a standard safety
assessment, the GM soybean was evaluated to see if it was as safe as a
conventional soybean. Although soybeans are a common food allergen and the
GM soybean remained allergenic, the results showed no significant difference
between its allergenicity and that of conventional soybeans. Specifically,
serums (blood) from individuals allergic to the GM soybean showed the same
reactions to conventional soybeans.

A toxic reaction in humans is a response to a poisonous substance. Unlike
allergic reactions, all humans are subject to toxic reactions. Scientists
involved in developing a GM food aim to ensure that the level of toxicity in
the food does not exceed the level in the food?s conventional counterpart.
If a GM food has toxic components outside the natural range of its
conventional counterpart, the GM food is not acceptable.

To date, GM foods have proven to be no different from their conventional
counterparts with respect to toxicity. In fact, in some cases there is more
confidence in the safety of GM foods because naturally occurring toxins that
are disregarded in conventional foods are measured in the pre- market safety
assessments of GM foods. For example, a naturally occurring toxin in
tomatoes, known as tomatine, was largely ignored until a company in the
early 1990s developed a GM tomato. FDA and the company considered it
important to measure potential changes in tomatine. Through an analysis of
conventional tomatoes, they showed that the levels of

8 Anaphylactic shock is a severe allergic reaction that can lead to death.
Allergic Reactions

Toxic Reactions

Page 11 GAO- 02- 566 Genetically Modified Foods

tomatine, as well as other similar toxins in the GM tomato, were within the
range of its conventional counterpart.

Antinutrients are naturally occurring compounds that interfere with
absorption of important nutrients in digestion. If a GM food contains
antinutrients, scientists measure the levels and compare them to the range
of levels in the food?s conventional counterpart. If the levels are similar,
scientists usually conclude that the GM food is as safe as its conventional
counterpart. For example, in 1995 a company submitted to FDA a safety
assessment for GM canola. The genetic modification altered the fatty acid
composition of canola oil. To minimize the possibility that an unintended
antinutrient effect had rendered the oil unsafe, the company compared the
antinutrient composition of its product to that of conventional canola. The
company found that the level of antinutrients in its canola did not exceed
the levels in conventional canola.

To ensure that GM foods do not have decreased nutritional value, scientists
also measure the nutrient composition, or ?nutrition profile,? of these
foods. The nutrient profile depends on the food, but it often includes amino
acids, oils, fatty acids, and vitamins. In the example previously discussed,
the company also presented data on the nutrient profile of the GM canola and
concluded that the significant nutrients were within the range of those in
conventional canola.

Companies that may wish to submit new GM foods for FDA evaluation perform a
regimen of tests to obtain safety data on these foods. FDA?s 1992 policy on
safety assessments of GM foods describes the data the agency recommends it
receive to evaluate these foods. Figure 2 provides an example of the regimen
of tests. This regimen usually includes an analysis of

 the source of the transferred genetic material, specifically whether the
source of the transferred gene has a history of causing allergic or toxic
reactions or containing antinutrients;  the degree of similarity between
the amino acid 9 sequences in the newly

introduced proteins of the GM food and the amino acid sequences in known
allergens, toxins, and antinutrients;

9 Amino acids are the ?building blocks? of the proteins in the body. Besides
building cells and repairing tissue, the proteins form antibodies to combat
invading bacteria and viruses, build DNA and RNA, and carry oxygen
throughout the body. Antinutrient Effects

Companies Use a Regimen of Tests to Ensure GM Food Safety

Page 12 GAO- 02- 566 Genetically Modified Foods

 data on in vitro digestibility (i. e., how readily the proteins break down
in simulated digestive fluids)  the comparative severity of individual
allergic reactions to the GM product

and its conventional counterpart as measured through blood (serum)
screening- when the conventional counterpart is known to elicit allergic
reactions or allergenicity concerns remain; and  data on any changes in
nutrient substances, such as vitamins, proteins,

fats, fiber, starches, sugars, or minerals due to genetic modification.
Occasionally, the regimen of tests also includes animal studies for
toxicity. As shown in figure 2, the tests provide evidence at key decision
points to direct which tests are subsequently performed. Tests on the source
of the newly expressed protein, amino acid sequence similarity, and
digestibility are typical for both allergenicity and toxicity assessments,
while serum screening is used only for allergenicity assessment. Also, while
the complete regimen is not necessary for every GM food safety assessment,
companies often perform extra tests in the regimen to corroborate the
results of previous tests.

Page 13 GAO- 02- 566 Genetically Modified Foods

Figure 2: Example of the Regimen of Tests Used for Safety Assessments of GM
Foods

Source: GAO?s analysis of FDA documents. Notes: Figure 2 represents typical
tests undertaken by a company in the safety assessment of a GM food. The
figure is not meant to be a comprehensive illustration that is used in every
safety assessment.

Antinutrients are tested as a subset of toxicity. In addition, they are
often measured with a simple nutrition/ composition profile.

If a company transfers genetic material from an allergenic source and
undertakes serum screening tests, it does not have to go though serum
screening again if in vitro digestibility tests uncover a similarity to an
allergen. At such a point, it would be assessed by amino acid sequence
similarity and in vitro digestibility tests for potential toxicity.

Page 14 GAO- 02- 566 Genetically Modified Foods

If a company reaches the ?Stop (or consult)? decision point, then there are
food safety concerns about allergenicity or toxicity issues. At this point,
FDA?s 1992 policy statement says the company should consult with FDA.
However, this usually means the company will discontinue development of the
GM product because of allergenicity or toxicity concerns. A company may
consult with FDA if there were changes in the nutrition/ composition profile
that were intended.

Animal studies are occasionally conducted as an extra test for potential
toxicity.

Using allergenicity as an example, if a company transfers a gene from a
source that is not an allergen, the company evaluates the amino acid
sequence of the GM protein. If the GM protein has an amino acid sequence
similar to that of known allergens, the company initiates further, more
specific allergenicity testing. The company would undertake in vitro
digestibility tests to see if the GM protein was broken down in simulated
digestive fluids. If there were any concerns about the speed with which the
GM protein was broken down, the company would use serum- screening tests to
support or refute the results of the digestibility tests when serums are
available. If the serum screening yields results showing that the GM protein
does not react with antibodies in serum, then the company concludes the GM
protein does not raise allergenicity concerns. The results from this regimen
of tests provide the weight of evidence necessary to determine the safety of
a GM food.

Examining the source of the transferred genetic material is the starting
point in the regimen of tests for safety assessments. According to a
scientist from a biotechnology company, two principles of allergenicity
assessment underlying the regimen of tests contribute to adequate safety
assessments: scientists (1) avoid transferring known allergenic proteins and
(2) assume all genes transferred from allergenic sources create new food
allergies until proven otherwise. If the source contains a common allergen
or toxin, industry scientists must prove that the allergenic or toxic
components have not been transferred. However, as a practical matter,
biotechnology companies repeatedly state that if the conventional food is
considered a major food allergen, 10 they will not transfer genes from that
source. Accordingly, experts from FDA and the biotechnology industry agree
that the probability of introducing a new allergen, enhancing a toxin, or
enhancing an antinutrient is very small.

The next step involves a comparison between the amino acid sequences of the
transferred proteins of the GM food plant and those of known allergens,
toxins, or antinutrients. If scientists detect an amino acid

10 Major food allergens are dairy milk, eggs, fish, peanuts, shellfish,
soybeans, tree nuts, and wheat. Source of the Transferred

Genetic Material Amino Acid Sequence Similarity

Page 15 GAO- 02- 566 Genetically Modified Foods

sequence in a GM food identical or similar to one in an allergen, toxin, or
antinutrient, then there is a likelihood that the GM food poses a health
risk. Overall, sequence similarity tests are very useful in eliminating
areas of concern and revealing areas for further evaluation.

In vitro digestibility tests are a primary component of all GM food safety
assessments. These tests analyze the breakdown of a GM protein in simulated
human digestive or gastric fluids. The quick breakdown of a GM protein in
these fluids indicates a very high likelihood that the protein is not
allergenic or toxic. Safe dietary proteins are almost always rapidly
digested, while allergens and toxins are not.

If a gene raises allergenicity concerns, a company can include serum
screening tests in its safety assessment of a GM food. Serum screening is
used only for allergenicity assessment. Serum screening involves evaluating
the reactivity of antibodies in the blood of individuals with known
allergies to the plant that was the source of the transferred gene. Antibody
reactions suggest the presence of an allergenic protein. Serum screening
tests are valuable because they can expose allergens whose presence was only
suggested in amino acid sequence similarity tests. 11

Since there are neither abundant, appropriate stored serums nor many
suitable human test subjects, these tests cannot always be used.

Scientists also create a nutritional and compositional 12 profile of the GM
food to assess whether any unexpected changes in nutrients, vitamins,
proteins, fibers, starches, sugars, minerals, or fats have occurred as a
result of the genetic modification. While changes in these substances do not
pose a risk of allergenicity, toxicity, or antinutrient effects to human
health, creating a nutritional and compositional profile further ensures
that the GM food is comparable to its conventional counterpart.

11 Clinical tests to corroborate the results of serum screening include skin
prick tests and double- blind placebo- control food challenges. Skin prick
tests involve scratching the skin of individuals with food allergies to
determine if the new GM product elicits allergic skin reactions. Double-
blind placebo- control food challenges aim to eliminate possible
psychosomatic reactions of allergic individuals by giving them either a test
product or a control product.

12 ?Compositional? refers to the concentration of nutrients and other key
substances in a plant, since the concentration of nutrients- not just the
presence- is important in assessing unexpected changes. Digestibility Tests

Serum Screening Nutritional and Compositional Profile

Page 16 GAO- 02- 566 Genetically Modified Foods

Biotechnology companies occasionally use animal studies to confirm the
results of prior toxicity tests. For the most part, these studies have
involved feeding extraordinarily high doses of the modified protein from a
GM food to mice. The doses of the modified protein are often hundreds to
thousands of times higher than the likely dose from human diets. Scientists
perform these studies to determine if there are any toxic concerns from the
GM food.

Animal studies also have the potential to predict allergenicity in humans,
although scientists have not yet identified an animal that suffers from
allergic reactions the same way that humans do. The brown Norway rat has
provided the closest approximation to human allergic reactions to several
major food allergens. However, animal models- as predictors of allergenic
responses in humans- are not scientifically accepted at this time.

Biotechnology experts whom we contacted from a consumer group, FDA, academic
institutions, research institutions, the European Union and biotechnology
companies said that the current regimen of tests has been adequate for
assessing the safety of GM foods. 13 All but one expert considered the
regimen of tests to be ?good? or ?very good? for ensuring the safety of GM
foods for public consumption, and the remaining expert viewed the tests as
?fair.? While the experts noted that individual tests have limitations, most
experts agreed that results from the regimen of tests provide the weight of
evidence needed for scientists to make an accurate assessment of risk. 14

A distinction made by an academician and regulatory officials is that the
available tests do not guarantee absolute safety of GM foods, but comparable
safety. There is no assurance that even conventional foods are completely
safe, since some people suffer from allergic reactions, and conventional
foods can contain toxins and antinutrients. Because they

13 Some consumer groups, as well as some scientists from the European Union,
have questioned the ethical or cultural appropriateness of genetically
modifying foods. 14 According to FDA officials, they are aware of only one
example- involving the Brazil nut- of GM food development where an allergen-
producing gene was transferred. In this case, the regimen of tests used in
safety assessments for GM foods successfully identified the presence of the
allergen- producing gene before the product was ever submitted for FDA
review. The company subsequently discontinued research, development, and
testing of the food. Animal Studies

According to Experts, GM Food Safety Tests Have Been Adequate

Page 17 GAO- 02- 566 Genetically Modified Foods

have been consumed for many years, though, conventional foods are used as
the standard for comparison in assessing the safety of GM foods, and experts
note that the available tests are capable of making this comparison.

While experts agree that the available regimen of tests is adequate for
safety assessments, there are limitations to individual tests. For example,
there are limitations to the acceptability of amino acid sequence similarity
test results, in part because there is not agreement on what level of amino
acid similarity indicates a likelihood of allergenicity and, therefore, the
need for additional testing. Industry scientists assert that as long as
amino acid sequences in a protein are less than 50 percent identical to
those in known allergens, then the protein should not raise concerns. On the
other hand, a scientist associated with a consumer group, as well as a
report from the United Nations? Food and Agriculture Organization, believe a
more conservative level, such as less than 35 percent identical, is
appropriate. Thus, experts from industry and consumer groups suggest that
reaching agreement on this parameter would increase the consistency with
which these tests are applied.

In vitro digestibility tests also have limitations because they can yield
inaccurate results when performed under inappropriate parameters, such as
improper digestive fluid pH levels. 15 If a GM food protein is tested at a
pH level representative of intestine digestion, yet the protein in real life
is digested at a different pH level in the stomach, then the results of the
test are not valid for reaching conclusions on the GM food?s likely effect
in humans. FDA officials note that there is growing acceptance that the
proper pH level for digestive stability tests is the pH level of the human
stomach. As a result, experts from industry and consumer groups suggest that
reaching agreement on the parameters in digestive stability tests- such as
proper pH ranges- would help ensure that they are performed properly.

Information on acceptable testing procedures (including parameters) is
available from a variety of sources. For instance, AOAC International 16
documents standardized tests and test procedures, such as test procedures
for examining nutrient levels in a GM food. Other groups, such as the
American Oil Chemists? Society and the American Association of Cereal

15 The pH level indicates the degree of acidity or alkalinity. 16 This
organization was formerly known as the Association of Official Analytical
Chemists.

Page 18 GAO- 02- 566 Genetically Modified Foods

Chemists also have information on official tests and test procedures.
However, there is no centralized source of information on these procedures.
Although FDA maintains a Web site with guidance for consultations, the Web
site does not contain information about acceptable testing procedures.

According to FDA, it has the necessary controls to ensure it obtains the
safety data needed for its GM food evaluations. In examining a selection of
submissions, we found that companies adhered to FDA?s recommended procedures
for the type of data to be submitted. However, biotechnology experts state
that the agency?s overall evaluation process could be enhanced by randomly
verifying the test data that companies provide and by increasing the
transparency of the evaluation process- including more clearly communicating
the scientific rationale for the agency?s final decision on GM food safety
assessments. FDA believes that making these changes would enhance the
public?s confidence in the agency?s evaluation process.

According to agency officials, FDA has several management practices that, in
aggregate, constitute internal controls. 17 The officials state that these
practices effectively ensure FDA obtains the data necessary for evaluating
the potential risks of GM foods. These practices include:

 communicating clearly what safety data are important to FDA?s evaluations
of GM food safety,  having teams of FDA experts representing diverse
disciplines perform the

evaluations, and  tailoring the level of evaluation to match the degree of
each GM food?s

novelty. One key indication of the effectiveness of these practices is FDA?s
ability to determine when data are inadequate and to specify the additional
data important to a complete evaluation. In the cases we examined when the

17 Standards for Internal Control in the Federal Government (GAO/ AIMD- 00-
21. 3.1, Washington, D. C.: November 1999) defines internal controls as an
integral component of an organization?s management that provides reasonable
assurance that effectiveness and efficiency of operations and compliance
with applicable laws and regulations are being achieved. FDA Believes That
It

Obtains Necessary Safety Data for GM Food Evaluations, but FDA?s Overall
Evaluation Process Could Be Enhanced

FDA Believes Its Management Practices Are Generally Effective in Ensuring It
Obtains the Data Necessary for Its GM Food Evaluations

Page 19 GAO- 02- 566 Genetically Modified Foods

company?s initial submission of data was insufficient, FDA was able to
specify and obtain additional data from the company.

For a GM food, the evaluation process, known as a consultation, generally
lasts between 18 months and 3 years, according to FDA officials. In what FDA
calls the ?initial? phase of the consultation, FDA and company officials
discuss what safety data will be needed for a GM food submission. In the
next or ?final? phase, the company prepares a detailed report summarizing
this data and submits it to FDA. After receiving and evaluating the report,
FDA officials prepare a ?memo to file.? This memo is the formal document in
which FDA summarizes and evaluates everything the company has submitted.
Consultation is complete when FDA determines that it has no further
questions regarding the safety of the GM food and informs the company of
this conclusion in a letter signed by the director of the FDA?s Office of
Food Additive Safety. Receiving such a letter is generally helpful to
companies in marketing their product.

In FDA?s 1992 policy statement and its subsequent 1997 guidance, the agency
clearly states what information companies should submit for FDA to assess
the safety of GM foods. Specifically, the 1992 statement includes several
risk assessment decision trees that provide a step- by- step approach to
testing. FDA recommends that companies follow this approach in their
assessments of GM foods. Using this approach, companies must show whether
any allergens, toxins, or antinutrients have been introduced or enhanced.
FDA?s 1997 guidance builds upon the 1992 policy statement by describing in
more detail the process, procedures, and time frames pertaining to the
initial and final consultations.

FDA officials stated that the principles embodied in their 1992 policy
statement guided the consultations for the 50 GM foods evaluated so far and
that companies have closely adhered to these principles. In examining five
submissions, 18 we found that companies adhered closely to the 1992 policy
statement. For example, a 1996 submission for a GM soybean 19 shows step-
by- step adherence to the allergenicity decision tree established

18 The five submissions reviewed are representative of the three main types
of GM plants evaluated as of April 2002: pest- resistant plants, herbicide-
tolerant plants, and plants having modified oils. Specifically, we reviewed
submissions for a pest- resistant corn; an herbicidetolerant sugar beet and
soybean; and high oleic- acid soybean oil and laurate canola oil.

19 This GM soybean was engineered to contain a substantially higher
percentage of oleic acid. FDA Guidance Describes

Safety Data Needed

Page 20 GAO- 02- 566 Genetically Modified Foods

in the 1992 policy statement. Extensive data submitted by the company
enabled FDA to conclude that it had no unanswered questions about the safety
of the soybean. Later submissions involving an herbicide- tolerant sugar
beet and pest- resistant corn also showed a close adherence to the 1992
policy statement.

Evaluations of GM food safety submissions must include concurrence from
every member of a highly qualified team known as the Biotechnology
Evaluation Team. 20 The 1997 guidance states that the evaluation teams
generally will be composed of a consumer safety officer (who serves as the
project manager), molecular biologist, chemist, environmental scientist,
toxicologist, and nutritionist. The guidance also states that the evaluation
teams may be supplemented with additional expertise on a case- by- case
basis. According to agency officials, these experts are qualified to perform
what is effectively a peer review 21 of each submission.

Consumer safety officers, who generally have doctorates in relevant
disciplines, including molecular biology, cell biology, or immunology, chair
the teams. According to FDA officials, in addition to their scientific
credentials, the consumer safety officers know what is needed for the
administrative record for each submission. This knowledge encompasses the
laws and regulations, such as the Federal Food, Drug, and Cosmetic Act, as
well as specific pertinent procedures, such as FDA?s 1992 policy statement.
According to FDA officials, the combination of scientific and administrative
expertise makes the consumer safety officers effective leaders of the teams.

FDA officials indicated that each member of an evaluation team reviews the
entire file for a given GM food submission. These officials viewed this as
another strength of the evaluation process. In particular, they stressed
that the final evaluation is not a ?piecemeal? evaluation in which, for
example, the toxicologist receives only the toxicological data to review.
Rather, each team member receives and examines all the data that the company
has submitted. Further, team members must document in writing the results of
all key interactions with a company throughout the course of

20 Within FDA, the Center for Food Safety and Applied Nutrition oversees the
Biotechnology Evaluation Teams for consultations on human food. FDA?s Center
for Veterinary Medicine oversees teams for consultations involving GM animal
feed.

21 Peer review is the critical evaluation of data, analysis, or documents by
professional colleagues. It is the traditional method of quality control in
science. GM Food Evaluations Are

Conducted by Multidisciplinary Teams

Page 21 GAO- 02- 566 Genetically Modified Foods

the evaluation; this documentation is then available for the whole team to
evaluate. Lastly, the entire team must concur with the final draft of the
memo to file, which is usually prepared by the consumer safety officer.

In summary, FDA officials told us that the expertise of the Biotechnology
Evaluation Team members coupled with the multiple reviews of information
enables the team to adequately evaluate safety assessments and determine if
and when more data is needed.

According to agency officials, FDA?s practice of varying its level of
evaluation based on the degree of novelty of the GM food submission allows
it to devote resources where they are most needed, thus assuring that
Biotechnology Evaluation Teams have time to obtain necessary safety data.
FDA?s evaluation of one company?s GM tomato provides an example of a
detailed evaluation of a novel submission that went through both the initial
and final consultations. Specifically, the Biotechnology Evaluation Team
requested extensive detail from the company on the modification of the
tomato, which involved the insertion of one gene to delay ripening and
another gene to show that this trait was transferred. FDA?s documentation of
its evaluation presented background information on these modifications, a
point- by- point evaluation of the company?s food safety assessment, and
FDA?s conclusion that the tomato was not significantly different from
conventional tomatoes.

By contrast, FDA officials stated that evaluations of company submissions
for GM foods similar to GM foods previously evaluated by the agency (such as
a virus- resistant squash and various herbicide- tolerant corns) required
fewer agency resources because these submissions skipped the initial
consultation and proceeded to the final consultation. In fact, FDA?s 1997
guidance states that a company might skip the initial consultation and go
directly to the final consultation by submitting its final report. According
to FDA officials, this skipping often occurs when a company has made
multiple submissions for similar GM foods involving only minor variations
from one case to the next. Having once gone through the full consultation
process for a specific genetic modification, such a company is familiar with
the kinds of safety information that FDA expects and thus can proceed
directly to preparing a final report for similar cases. FDA?s documentation
of its evaluation of such submissions can be less detailed.

According to FDA officials, in cases in which the agency determines that the
data submitted by a company are insufficient, the company has always
cooperated with FDA by performing additional tests and/ or submitting the
data needed. FDA officials described three types of situations where they
FDA Tailors Its Evaluation to

Match the Degree of Novelty in Each Submission

FDA Believes It Obtains Complete, Consistent Data

Page 22 GAO- 02- 566 Genetically Modified Foods

have requested additional data and companies have responded: (1) the absence
of a reliable or ?validated? method for performing a test; (2) reliance on a
prevailing scientific ?assumption? that, when tested at FDA?s request, was
proven incorrect; and (3) inconsistent or incomplete data in the final
reports.

The first situation involved the lack of a reliable method for testing
tomatine, a naturally occurring toxin in tomatoes. The company that
encountered this problem was inexperienced in analytical chemistry, and the
laboratory with which it was working did not have an acceptable method. In
evaluating the measurements of tomatine submitted by the company, FDA
officials found these data unconvincing. As a result, FDA officials
suggested that the company find a more appropriate method. In response, the
company obtained a suitable method from another laboratory and later
provided FDA with new data that the agency found convincing.

The second situation is illustrated by FDA?s evaluation of a GM tomato
altered to delay ripening. In this submission, the company assumed that only
a certain segment of DNA was transferred. FDA asked the company to prove the
accuracy of this assumption. Testing by the company then revealed that
additional DNA had been transferred. This discovery led to more thorough
analysis of the genetic modifications, including additional efforts to
ensure that the transfer of extra DNA did not cause unintended changes.

In the third situation, FDA noted discrepancies in the data in final reports
involving GM cotton, rice, and canola and requested the relevant companies
to correct the information, which they did.

Biotechnology experts state, and FDA agrees, that its overall evaluation
process for assessing the safety of GM foods could be enhanced by

 verifying the GM food- related test data that companies provide, and 
increasing the transparency of the evaluation process.

Biotechnology experts from consumer groups and academia state that FDA?s
evaluation process could be enhanced if the agency validated companies? test
results on proposed GM products by reviewing raw data (e. g., the actual,
unverified test results). Further, FDA believes that occasional reviews of
the raw data developed by companies would further enhance the credibility
of, and public confidence in, the overall safety data FDA Evaluations of GM

Foods Could be Enhanced FDA Could Validate Company Safety Data

Page 23 GAO- 02- 566 Genetically Modified Foods

that companies submit. In addition, we believe occasional data verification
by a federal agency is necessary to (1) identify the risk of the agency?s
receiving faulty data from external sources and (2) ensure that no one agent
is allowed to control every key aspect of a safety assessment. 22

FDA officials stated that they do not believe it is necessary for the agency
to routinely review raw data for two reasons. First, the risk of incurring
criminal penalties for deliberately submitting false data to FDA provides a
significant degree of deterrence. Second, FDA?s evaluation process
constitutes a peer review of the safety data that will generally detect any
problems. 23 However, these officials added that an occasional review of raw
data, performed on a random basis, would further help ensure the reliability
of FDA?s evaluation of these foods, and thus enhance public confidence in
the agency?s evaluation process.

Officials from a major biotech company described three types of GM food
safety data developed for each submission and available for FDA?s review:
(1) raw data, (2) refinements and comprehensive interpretations of the raw
data, and (3) summaries of these interpretations. According to these
officials, FDA has reviewed the summaries, and in some instances the
comprehensive interpretations, but has not reviewed the raw data. These
officials note, and FDA officials concur, that nothing prevents FDA from
reviewing these raw data. In general, these raw data are readily available
from companies. The company officials also note that EPA has occasionally
reviewed raw data in its safety assessments of GM plants regarding their
environmental effects. Moreover, FDA officials stated the agency reviews raw
data in its safety assessments of new drug applications.

Experts from consumer groups and academia have stated that the transparency
of the agency?s evaluation process for GM foods could be enhanced if FDA
described more clearly the scientific rationale for its safety decisions in
its memo to file. FDA agrees. Guidelines issued by the

22 U. S. General Accounting Office, Internal Control Management and
Evaluation Tool

GAO- 01- 1008G (Washington, D. C.: August 2001). 23 Among other factors not
directly linked with internal controls, FDA officials pointed out that the
majority of requests for access to GM food submissions through the Freedom
of Information Act come from competitors, who, according to these officials,
would likewise be quick to detect a problem with the data. Scientists,
consumer groups, and other interested parties have also asked for access to
these submissions. However, the data reviewed by these groups does not
include the raw data. FDA Could Increase the

Transparency of Its Evaluations

Page 24 GAO- 02- 566 Genetically Modified Foods

Office of Management and Budget on the quality of information disseminated
by federal agencies state that transparency is important in reviews of
technical information and that these reviews should be conducted in an open
and rigorous manner. 24 Yet critics have stated that FDA?s current memos to
file do not adequately communicate the scientific rationale for the
decisions. Some consumer groups have pointed out the brevity of some of the
memos and described them as ?perfunctory? summaries of company data that
provide little or no insight into FDA?s evaluation of the data. Likewise,
the Council for Agricultural and Science Technology, a group of universities
and companies established to provide a more scientific basis for analyzing
and prioritizing agricultural issues, stated that FDA does not adequately
clarify in its memos to file the basis for its decisions on GM food
submissions. Our review of memos to file for the 50 GM food products
evaluated by FDA as of April 2002 confirms that these memos do not clearly
explain the scientific rationale for FDA?s decisions.

In response to these concerns, FDA officials note that the memos to file had
originally been created for FDA?s internal use rather than as public
documents. Thus, they were not designed to provide detailed rationales of
FDA?s decisions on GM food submissions. In addition, FDA officials said that
some memos are brief because they record decisions on GM foods that are very
similar to previously evaluated GM foods. However, FDA officials acknowledge
that FDA could do more to inform the public of the basis for their
decisions. For example, FDA could include comments in the memos to file that
better reflected the context of the evaluation (for instance, its similarity
to previous evaluations), the adequacy of the tests performed by the
company, and the level of evaluation provided by FDA. For those memos to
file on submissions for GM foods that are similar to GM foods previously
evaluated, FDA could make reference to earlier, similar submissions having a
more detailed memo to file.

24 Guidelines for Ensuring and Maximizing the Quality, Utility, and
Integrity of Information Disseminated by Federal Agencies (Office of
Management and Budget, Washington, D. C.: January 2002).

Page 25 GAO- 02- 566 Genetically Modified Foods

Scientists expect future GM foods to include modifications of plant
composition that may enhance the nutritional value of these foods but may
also increase the difficulty of assessing their safety. While current tests
have been adequate for evaluating the small number of relatively simple
compositional changes made so far, some scientists believe that new testing
technologies under development may be needed to assess the safety of these
more complex GM foods. Scientists have diverging views on the potential role
of these new technologies: some view them as a useful supplement to existing
tests, while others view them as a new, more comprehensive way to assess the
safety of all changes in GM foods. However, the lack of technical standards
for these new technologies and proof of their reliability prevents their
current use.

Until now, most genetic modifications of plants have been aimed at
increasing or protecting crop yield. These modifications have generally
focused on the portions of plants, such as cornstalks, that are not consumed
by humans. However, many scientists believe that the current wave of yield-
related modifications will expand to include a new wave of genetic
modifications involving compositional changes in the foods to enhance their
nutritional value. For example, ?golden? rice is a GM food under development
that was modified to contain beta- carotene, a precursor of vitamin A. 25
Golden rice may help to reduce the incidence of blindness in countries where
rice is a dietary staple and malnutrition is common. Also under development
are compositional changes that will increase the levels of vitamin E in
foods. Plants are the primary source of this vitamin, which is believed to
have cancer- preventing properties, but plants generally contain it in
relatively low concentrations. A gene controlling vitamin E production was
transferred recently to a member of the mustard plant family, which
subsequently exhibited a nine- fold increase in this vitamin. According to a
recent report, 26 incorporation of this gene into major crops such as
soybeans, canola, and corn is probably not far in the future.

In addition to increasing nutrients in GM foods, scientists are working to
reduce the presence of allergens, toxins, and antinutrients. For example,

25 The human body converts beta- carotene into vitamin A if the body is
deficient in vitamin A. 26 Pew Initiative on Food and Biotechnology, Harvest
on the Horizon: Future Uses of Agricultural Biotechnology, Washington, D.
C., September 2001. Future Changes in GM

Foods May Alter Their Composition and Require New Testing Technologies for
Assessing Safety

Experts Foresee Compositional Changes

Page 26 GAO- 02- 566 Genetically Modified Foods

scientists have genetically modified wheat, one of the major allergenic
foods, to stimulate a gene that diminishes wheat?s allergenic properties.
Scientists are also seeking ways to reduce toxic substances, such as
alkaloids in potatoes, by inserting genes that block their production.
Preliminary findings have indicated that GM potatoes produced fewer of these
alkaloids. Likewise, some plants, especially cereals and legumes, are
nutritious foods but contain varying amounts of antinutrients. Genetic
modifications are being explored to reduce these antinutrients.

If adopted, FDA?s proposed rulemaking mandating the testing of all GM foods
prior to commercialization will represent a timely response to this new wave
of GM foods. For example, the preamble to the rule notes that some of the
new ingredients in GM foods will significantly differ from ingredients that
have a history of safe use. The rule also notes that products derived from
this advanced biotechnology will present more complex safety and regulatory
issues than those seen to date. The proposed rule concludes that
nontraditional strategies for evaluating food safety will become the norm as
the use of biotechnology expands. FDA officials explained that
?nontraditional strategies? could include new technologies under development
such as those described in the next section.

Some scientists believe that testing technologies being developed but not
yet widely applied to GM foods may be useful in assessing the safety of
compositional changes and detecting unintended effects. 27 In contrast to
current tests that examine the human health effects of transferred genes and
other relevant components on a highly selective basis, the new technologies
will examine essentially all of the components- such as DNA, proteins, and
metabolites 28 -in conventional and GM plants simultaneously to detect any
differences. These new technologies include

27 Unintended effects involve the accidental creation of an allergen,
enhancement of a toxin or antinutrient, or unintentional alteration of the
nutritional profile of a food. Unintended effects can also involve changes
in crop yield and growth rates.

28 Metabolites are small molecules in living cells. These small molecules
include vitamins (with potential benefits for human health) and alkaloids (a
major source of toxicity in plants). Their ?smallness? is defined in
contrast to generally much larger molecules such as proteins. New Testing
Technologies

May Be Useful in Addressing Compositional Changes

Page 27 GAO- 02- 566 Genetically Modified Foods

 gene chips that use thousands of droplets of DNA on glass chips to
identify gene sequences and determine the expression level or abundance of
the genes;  proteomics which can analyze up to 100,000 proteins
simultaneously; and  metabolic profiling that can analyze the 2,000 to 3,
000 metabolites in

people and 3,000 to 5,000 metabolites in plants. In essence, these new
technologies combine huge increases in automated computing power with
traditional testing technologies to identify differences between
conventional and GM foods in ways that would have been impossible even a few
years ago.

A university scientist further explained the contrast between the current
and new technologies by noting that traditional tests focus on known toxins
and nutrients in a ?targeted? approach, whereas new technologies use a ?non-
targeted? approach to increase the chance of detecting unintended effects of
genetic modifications such as the creation of a toxin. According to this
scientist, the latter approach has particular applicability to second-
generation plants with extensive modifications, which may be more likely to
have unintended effects. For example, a scientist with a consumer group
stated that the new technologies may be useful in detecting unintended
effects that traditional tests, such as those for digestibility, are not
likely to identify. Other scientists expressed the need for caution and
additional information to determine the potential role of these new
technologies.

Gene chips consist of grids of thousands of droplets of DNA on small glass
surfaces. The chip- based DNA can bind with the DNA or RNA being tested to
determine which genes are present or are being activated. Used in
conjunction with DNA and RNA databases under development at various
universities and other research institutions, this testing technique has
yielded insights into areas such as the ripening process of tomatoes and its
relation to toxins and nutrients. The major advantage of gene chips over
conventional testing techniques is that they allow small- scale analysis of
thousands of genes at the same time in a precise and quantitative manner.
According to a university scientist, researchers are determining the extent
to which this technology may be effective in assessing GM food safety.

Proteomics is a biotechnology technique used to identify many proteins
simultaneously in a given organism. Using chemical analyses and computers,
proteomics goes beyond plant studies focusing on DNA and RNA, which do not
provide information on the actual creation of the proteins. Proteomics has
been introduced successfully in medical Gene Chips

Proteomics

Page 28 GAO- 02- 566 Genetically Modified Foods

disciplines such as oncology, where it has helped to identify proteins
associated with cancer, but it has not yet been used to evaluate the safety
of GM foods for two reasons. First, there are a large number of proteins
that need to be analyzed in any given plant. Second, the function of
proteins in a plant may change depending on their interaction with different
cells and tissues. According to a university scientist, researchers are
working to expedite the analysis of proteins in plants.

Metabolic profiling uses chemical analyses and computers to obtain a
simultaneous, detailed look at all of the small molecules (metabolites) in a
given GM plant to determine the extent to which these molecules have changed
in comparison to a conventional plant, if at all. According to scientists at
one company involved in developing metabolic profiling, this technique can
determine whether a specific, intended change in a small molecule has been
achieved. It can also identify any unintended changes in other small
molecules- changes such as increased alkaloids, which are a major source of
toxicity in plants. If the profiling finds no unintended changes in these
molecules, then it offers a reasonable certainty that the genetic
modification has not led to any changes with potentially adverse health
consequences. In general, metabolic profiling has not yet been used
commercially. However, scientists working with this technique believe that
it may play a potentially important role as a safety screening tool for
companies developing complex, compositionally altered GM foods in the
future. In addition, scientists state that it shows promise in the health
care field in assessing the safety of future new drugs.

Despite progress in developing and applying gene chips, proteomics, and
metabolic profiling, technical limitations currently prevent their use to
assess the safety of GM foods. Biotechnology experts told us that internal
standards must be developed for the methods and chemicals used in these new
technologies and that the reliability of these technologies must be proven.
For example, in gene chip testing, experts state that standardization of the
thousands of genes represented on the chips is essential to improve the
quality of this technology. Further, experts state that the chemical
analysis used in proteomics needs to be enhanced to improve its reliability.

Beyond these technical challenges, however, lies a more fundamental problem.
Because these new technologies are more sensitive, they may identify a flood
of differences between conventional and GM food products that existing tests
could not detect. Not all of these differences will stem from genetic
modification. Some of the differences will stem Metabolic Profiling

Challenges to Adopting New Testing Technologies

Page 29 GAO- 02- 566 Genetically Modified Foods

from the tremendous natural variations in all plants caused by factors such
as the maturity of the plants and a wide range of environmental conditions,
such as temperature, moisture, amount of daylight, and unique soil
conditions that vary by region of the country. For example, there can be a
tenfold difference in the level of key compositional elements, such as
nutrients, depending on the region in which soybeans are grown. Thus,
according to a biotechnology company expert, it will be difficult to
differentiate naturally occurring changes from the effects of deliberate
genetic modifications.

Industry and university scientists have expressed strong concerns about the
problem of interpreting the potential significance of these differences.
They believe that the new technologies will be of limited value unless
baseline data on the natural variations of nutrients and other compositional
values for each of the major food crops can be developed. However, experts
disagree on the difficulty of developing this baseline. Some experts,
including those at FDA, assert that developing the baseline will be
difficult because of the extreme sensitivity of plants to environmental
variations. Other experts, especially those pioneering the new techniques,
state that a baseline can definitely be established in the next few years.

Some companies have started to respond to the need for baseline information.
New developments in technology have begun to provide an encyclopedic
database on natural variations in plants and on the variations resulting
from deliberate genetic modification. For example, using metabolic
profiling, one company has analyzed approximately 150 characteristics, such
as the size and rate of growth, of individual plants. The company has also
examined about 12,000 genes in one species of plant- a member of the mustard
family- and analyzed the consequences of eliminating or stimulating
particular genes. About one million mustard plants of this type have been
analyzed in this line of research. Even with the development of baseline
data and the detection of differences, scientists will still need to
evaluate the significance of these differences for human health. Appendix II
provides more information regarding advancements in the development of
baseline information and the experimental use of metabolic profiling to
assess the safety of GM foods.

Page 30 GAO- 02- 566 Genetically Modified Foods

Scientists and federal regulatory officials we contacted generally agreed
that long- term monitoring of the human health risks of GM foods through
epidemiological studies 29 is not necessary because there is no scientific
evidence suggesting any long- term harm from these foods. These scientists
and officials also stated that it would be very difficult, if not
impossible, to develop a process for monitoring the long- term health risks
of GM foods because of the technical challenges in developing such a system.
A recent report by the United Nations also expresses skepticism about the
feasibility of identifying long- term health effects from GM foods.

The scientists and federal regulatory officials generally agreed that
because there is no scientific evidence that GM foods cause long- term harm,
such as increased cancer rates, there is no plausible hypothesis of harm.
Researchers need such a hypothesis in order to know what problem to search
for, test, and potentially measure. For example, in the Framingham Heart
Study of Massachusetts, researchers hypothesized that there were biological
and environmental factors that contributed to cardiovascular disease. Using
this hypothesis, researchers were able to design a study that established a
relationship between the levels of cholesterol and the risk of heart
disease. The resulting effort, comprising more than 10,000 participants over
two generations (more than 50 years), developed groundbreaking information
on the major risk factors associated with heart disease, stroke, and other
diseases. For example, researchers found that a lifestyle typified by a
faulty diet, sedentary living, or unrestrained weight gain exacerbated
disease risk factors and influenced the occurrence of cardiovascular
problems.

Without a plausible hypothesis such as that used in the Framingham study,
most scientists we contacted said that epidemiological studies on GM foods
would not provide any useful information. Two of these scientists also noted
that the primary ways in which foods might cause long- term harm are through
(1) proteins that remain stable during human digestion, thereby retaining
the potential to exert adverse effects such as a toxic reaction, and (2)
detrimental changes in nutrients and other food components. However, for all
50 GM food plants reviewed by FDA as of April 2002, the genetically modified
proteins in those foods that potentially

29 Epidemiological studies assess various factors influencing the
occurrence, distribution, prevention, and control of disease, injury, and
other health- related events in a defined human population. These studies
often run for 20 years or more, involve thousands of people, and cost
millions of dollars. Experts Maintain That

Long- Term Monitoring of GM Foods Is Neither Necessary nor Feasible

There Is No Plausible Hypothesis Suggesting Long- Term Harm from GM Foods

Page 31 GAO- 02- 566 Genetically Modified Foods

could be cause for concern have been shown in tests to be rapidly digested.
Further, the two GM food plants reviewed that produced modified oils-
soybean and canola- had nutritional profiles that were similar to or better
than their conventional counterparts. As discussed previously, the soybean
oil was modified to be more nutritious than conventional soybean oil. The
canola oil was modified to contain a higher level of laurate, which would
allow it to substitute for imported tropical oils, such as palm kernel oil.
However, industry determined that the total intake of laurate in the diet
would not change significantly by substituting the improved canola oil for
the tropical oil. Accordingly, industry officials stated, and FDA officials
concurred, that long- term studies of health effects of this oil would not
be needed.

Scientists and federal regulatory officials also stated that there are
substantial technical challenges that make long- term monitoring of the
health effects of GM foods virtually impossible. The challenges cited
include the following:

 Conducting long- term monitoring would require both an experimental group
that has consumed GM foods and a control group. The control group would
consist of people who could confirm that they do not eat GM foods. In
countries such as the United States, where labeling is not required for GM
foods, reliably identifying such control groups would be virtually
impossible.  Even if GM foods were labeled in the United States, it would
be very

difficult to separate the health effects of GM foods from those of their
conventional counterparts, since to date there has been very little
nutritional difference between these foods. Further, over long periods of
time, there would be practical challenges in feeding both the experimental
and controls groups diets comprising large amounts of GM food, such as
soybeans or corn, and their conventional counterparts.  Since the long-
term human health effects of consuming most foods are not

well understood, there is no baseline information against which to assess
health effects caused by GM foods.  Changes in human food consumption
patterns, specifically the addition

and removal of various foods, add new variables to the diet and compound the
difficulty of conducting long- term monitoring. The fairly recent
introduction of the kiwi fruit (to which some individuals are allergic) and
the reduction of the use of cotton seed (to which some individuals have also
been allergic) as a protein source in candy or breads illustrate the
challenges in monitoring food consumption patterns when conducting a
Technical Challenges Make

Long- Term Monitoring Infeasible

Page 32 GAO- 02- 566 Genetically Modified Foods

20- to- 30 year epidemiological study. A report issued in June 2000 by the
United Nations? Food and Agriculture Organization and World Health
Organization supports the scientists? and regulators? views about the
infeasibility of identifying long- term health effects from GM foods. 30 The
report states that, in general, very little is known about the potential
long- term effects of any foods, and that identification of such effects is
further confounded by the great variability in the way people react to
foods. 31 The report also states that epidemiological studies are not likely
to differentiate the health effects of GM foods from the many undesirable
effects of conventional foods, which according to scientists include the
effects of consuming cholesterol and fats. Accordingly, the report concludes
that the identification of long- term effects specifically attributable to
GM foods is highly unlikely.

Given the challenges to long- term monitoring, federal regulatory officials,
as well as some U. S. and European scientists, state that the best defense
against long- term health risks from GM foods is an effective pre- market
safety assessment process.

Biotechnology experts believe that the current regimen of tests has been
adequate for ensuring that GM foods marketed to consumers are as safe as
conventional foods. However, some of these experts also believe that the
agency?s evaluation process could be enhanced. Specifically, FDA could
verify companies? summary test data on GM foods, thus further ensuring the
accuracy and completeness of this data. In addition, the agency could more
clearly explain to the public the scientific rationale for its evaluation of
these foods? safety, thereby increasing the transparency of, and public
confidence in, FDA?s evaluation process. By addressing these issues, FDA?s
assurance to consumers that GM foods are safe could be strengthened.

To enhance FDA?s safety evaluations of GM foods, we recommend that the
Deputy Commissioner of Food and Drugs direct the agency?s Center for Food
Safety and Applied Nutrition to

30 FAO/ WHO (2000b) Safety Aspects of Genetically Modified Foods of Plant
Origin. Report of a Joint FAO/ WHO Expert Consultation on Foods Derived from
Biotechnology (Geneva, Switzerland, May 29 -June 2, 2000).

31 According to scientists we contacted, this especially includes foods
containing allergens. Conclusions

Recommendations for Executive Action

Page 33 GAO- 02- 566 Genetically Modified Foods

 obtain, on a random basis, raw test data from companies, during or after
consultations, as a means of verifying the completeness and accuracy of the
summary test data submitted by companies; and  expand its memos to file
recording its decisions about GM foods to

provide greater detail about its evaluations of the foods, including the
level of evaluation provided, the similarity of the foods to foods
previously evaluated, and the adequacy of the tests performed by the
submitting companies.

We provided FDA with a draft of this report for review and comment. In its
written comments, FDA stated it believes that its current process for
evaluating bioengineered foods provides appropriate oversight but agreed
that enhancements can be made. Specifically, concerning the need to randomly
review raw safety data, FDA agreed that occasional audits would provide
additional assurance to the public that pre- market decisions about
bioengineered foods are based on sound science and that safety and
regulatory issues are resolved prior to commercial distribution. Concerning
the expansion of its memos to file, the agency agreed that providing greater
detail on its decisions about the safety of GM foods would enhance public
understanding and confidence in the evaluation process. The agency noted
that actions in its proposed rule- titled Premarket Notice Concerning
Bioengineered Foods (66 FR 4706, January 18, 2001)- are relevant to our
recommendations. FDA explicitly states it will evaluate whether to adopt
occasional audits as it evaluates comments on its proposed rule. Since FDA
officials told us that some of its proposed rule changes in the Federal
Register have taken years to implement, we believe that the public?s
interests would be served by implementing our recommendations separately
from the proposed rule approval process.

FDA also had general comments about the terms and definitions used in
discussing agricultural biotechnology. FDA stated that our draft report
avoided many of the pitfalls in terminology and in general was written in a
manner that will be understandable to the public. However, the agency
believes the use of terms such as ?Genetically Modified Food? in the title
and ?GM food? in the text can be misleading and such foods are more commonly
referred to as bioengineered foods. While perhaps the scientific community
refers to these foods as bioengineered, the lay public is more familiar with
the term genetically modified foods. Accordingly, we have continued to use
the term genetically modified, which is defined on page one of our report.
Agency Comments

Page 34 GAO- 02- 566 Genetically Modified Foods

Separately from its written comments, FDA provided us with some technical
changes, which we incorporated into the report where appropriate. FDA?s
written comments are presented in appendix III.

We performed our review from July 2001 through May 2002 in accordance with
generally accepted government auditing standards. (See app. I for our
objectives, scope, and methodology.)

We are sending copies of this report to congressional committees with
jurisdiction over food safety programs, the Deputy Commissioner of Food and
Drugs, the Director, Office of Management and Budget, and other interested
parties. We will also make copies available to others upon request. In
addition, the report will be available at no charge on the GAO Web site at
http:// www. gao. gov.

If you or your staff have any questions about this report, please call me at
(202) 512- 3841. Key contributors to this report are listed in appendix IV.

Lawrence J. Dyckman Director, Natural Resources and Environment

Appendix I: Objectives, Scope, and Methodology

Page 35 GAO- 02- 566 Genetically Modified Foods

Representatives John Baldacci and John Tierney asked us to (1) identify the
types of potential human health risks associated with genetically modified
(GM) foods and experts? views on the adequacy of tests used to evaluate
these risks, (2) describe the Food and Drug Administration?s (FDA) controls
for ensuring that companies submit test data it requests and identify
experts? views on the agency?s overall evaluations of these foods, (3)
describe potential changes in future GM foods and any associated changes
needed in tests to evaluate them, and (4) identify experts? views on the
necessity and feasibility of monitoring the long- term health risks of these
foods.

In addressing our review objectives, we interviewed representatives from U.
S. consumer groups, academic and research institutions, federal regulatory
agencies, and the biotechnology industry. We also E- mailed a set of
questions to experts representing a variety of positions on biotechnology
issues. We selected these experts in consultation with officials from the
National Academy of Science?s National Research Council. These experts
included scientists from the Center for Science in the Public Interest, the
Union of Concerned Scientists, the Biotechnology Center of the University of
Illinois, the Health Sciences Center of Tulane University, FDA, the Aventis
Corp., the DuPont Corp., the Monsanto Corp., and Paradigm Genetics, Inc. In
addition, we analyzed reports, policy documents, or issue papers from the
Center for Science in the Public Interest, the Consumer Federation of
America, the Union of Concerned Scientists, the Council for Agricultural
Science and Technology, the National Academy of Sciences, the Pew Initiative
on Biotechnology, the Environmental Protection Agency, FDA, the
Biotechnology Industry Organization, the Institute of Food Technologists,
the Codex Alimentarius, 1 and the National Institute for Quality Control of
Agricultural Products at the Wageningen University and Research Center of
the Netherlands. We did not assess the potential environmental risks
associated with GM food production. In addition, since there have been no GM
animals evaluated for commercialization, we did not assess the potential
environmental or human health risks associated with them.

To identify the types of potential health risks of GM foods, we analyzed and
synthesized information from the interviews, E- mail question responses, and
documents regarding these risks. To identify tests

1 Codex Alimentarius is the joint food standards program for the United
Nations? Food and Agriculture Organization and the World Health Organization
that was established in 1962. Appendix I: Objectives, Scope, and

Methodology

Appendix I: Objectives, Scope, and Methodology

Page 36 GAO- 02- 566 Genetically Modified Foods

commonly used by industry to assess GM food safety, we examined several FDA
evaluations of GM food. In examining these evaluations, we also analyzed how
FDA addresses any potential limitations in these tests and what guidance FDA
provides to industry regarding scientifically acceptable tests. In our E-
mail questions, we also asked the experts to describe any limitations to
these tests, and then analyzed and synthesized their responses, particularly
regarding test- specific limitations and suggestions for improving the
tests. In addition, we asked whether there were any limitations to FDA?s
guidance on acceptable tests. We then synthesized their responses, including
suggestions for improving FDA?s guidance.

To describe FDA internal controls for ensuring that companies submit safety
test data requested by the agency, we interviewed FDA officials and reviewed
agency documents about the functions of these internal controls,
specifically (1) FDA?s 1992 Policy on Foods Derived from New Plant Varieties
and its 1997 Guidance on Consultation Procedures that describe what safety
data companies should submit; (2) the qualifications and roles of the FDA
Biotechnology Evaluation Teams responsible for evaluating these submissions;
and (3) FDA?s practice of matching its level of evaluation to the degree of
novelty of the GM food submitted. Further, we compared the safety data
specified in FDA?s 1992 policy with data provided by companies in five GM
food submissions and analyzed the extent of the companies? adherence to
FDA?s recommended procedures for safety assessments. 2 We contacted
officials at the Department of Health and Human Services? Office of
Inspector General to determine if they had reviewed FDA?s internal controls.
(They had not.) We did not, however, independently verify the adequacy of
FDA?s internal controls. To identify experts? views on the agency?s overall
evaluations of GM foods, we interviewed consumer groups, industry officials,
and other experts, analyzed their views and concerns- including any
suggestions for improving FDA?s evaluation process- and reviewed related
literature. For each concern identified with the process, we obtained FDA?s
response and then determined the extent to which FDA?s response effectively
addressed the concern or suggested a need for additional action by FDA.
Further, we examined Office of Management and Budget and GAO guidance and
policies relevant to these concerns.

2 The five submissions reviewed are representative of the three main types
of GM plants evaluated as of April 2002: pest- resistant plants, herbicide-
tolerant plants, and plants having modified oils.

Appendix I: Objectives, Scope, and Methodology

Page 37 GAO- 02- 566 Genetically Modified Foods

To describe the potential changes in future GM foods and associated changes
needed in the tests to evaluate them, we interviewed scientists and
regulators on the likely changes in GM foods and new testing approaches
under development. We also focused several of our E- mail questions on this
topic and analyzed the responses. In addition to E- mail respondents, we
contacted experts from biotechnology companies concerning research on new,
more complex GM foods as well as new testing approaches that may supplement
or replace existing tests. We synthesized these respondents? and experts?
views on likely changes to GM food and the value and challenges of using
these new testing approaches. Further, we reviewed the relevant scientific
literature for discussions of anticipated changes in GM foods and
information on specific tests under development. We also met with scientists
developing one of these new testing approaches to understand its potential
value for assessing GM food safety.

To identify the views of experts on the necessity and feasibility of
monitoring the long- term health risks of GM foods, we asked respondents to
our E- mail questions for an assessment of whether such an effort is
necessary or feasible and then analyzed their responses. Further, we
reviewed a variety of documents concerning the necessity and feasibility of
long- term monitoring, including a recent joint United Nations? Food and
Agriculture Organization and World Health Organization report, as well as a
recent report by the National Institute for Quality Control of Agricultural
Products at the Wageningen University and Research Center of the
Netherlands. We also discussed the topic with other regulatory officials
connected with monitoring food safety. In particular, we discussed whether
the long- term effects of GM foods could be separated from other factors
that may influence human health.

Finally, we submitted a draft of this report for technical review by
scientists from industry, academia, and a consumer group, and we
incorporated their comments as appropriate.

We conducted our review from July 2001 through May 2002 in accordance with
generally accepted government auditing standards.

Appendix II: Potential Use of Metabolic Profiling to Compare a GM Plant to
Its Conventional Counterpart

Page 38 GAO- 02- 566 Genetically Modified Foods

Metabolic profiling could be used as a safety- screening tool for GM foods.
Specifically, as shown in figure 3, special software has allowed one company
to graph the metabolic profile of one variety of mustard plants and analyze
the effects of genetic modifications. In the figure, the vertical axis in
each graph provides a list of different small molecules, or metabolites, in
mustard plants from this variety. The horizontal axis measures variation or
deviation from the metabolite levels in this conventional variety. The
vertical line in the middle of each graph represents the average value for a
range of small molecules, or metabolites, in this conventional variety. In
this example, the company analyzed thousands of conventional plants from
this variety to come up with a range of naturally occurring metabolite
levels. The company then used the averages of these ranges to generate the
vertical line in the middle of the graphs. The points plotted with squares
represent the levels of small molecules in GM mustard plants. Points
appearing to the right of the center vertical line indicate increased levels
of specific small molecules, while points appearing to the left indicate
decreased levels. The graphs in figure 3 illustrate three scenarios: graph
(a) shows a GM mustard plant with small molecule levels nearly identical to
its conventional counterpart; graph (b) shows a GM mustard plant with a few
easily measurable decreases; and graph (c) represents a GM mustard plant
with many significant differences from the small molecule levels of its
conventional counterpart. Appendix II: Potential Use of Metabolic

Profiling to Compare a GM Plant to Its Conventional Counterpart

Appendix II: Potential Use of Metabolic Profiling to Compare a GM Plant to
Its Conventional Counterpart

Page 39 GAO- 02- 566 Genetically Modified Foods

Figure 3: Metabolic Profiling of Three GM Mustard Plant Varieties in
Comparison with Baseline Data from Their Conventional Counterparts

Note: Standard deviation is a measure of the probability that an observed
value will be different from the average value.

Graph (a)

Appendix II: Potential Use of Metabolic Profiling to Compare a GM Plant to
Its Conventional Counterpart

Page 40 GAO- 02- 566 Genetically Modified Foods

Graph (b)

Appendix II: Potential Use of Metabolic Profiling to Compare a GM Plant to
Its Conventional Counterpart

Page 41 GAO- 02- 566 Genetically Modified Foods

Source: Paradigm Genetics, Inc.

If baseline data on normal ranges of variation, such as those developed for
the mustard plants, can be made available for all GM food crops, companies
might use this type of testing to develop safety data. For example, in graph
(a), the absence of significant changes in the small molecules would
strongly indicate that no significant changes had resulted from the genetic
modification. Hence, a change in the risk of allergenicity, toxicity, or
antinutrients would be very unlikely. In the case represented by graph (b),
the software could determine which small molecules have changed. Then,
traditional testing techniques such as toxicity testing, could be used to
determine if the altered small molecules would have any Graph (c)

Appendix II: Potential Use of Metabolic Profiling to Compare a GM Plant to
Its Conventional Counterpart

Page 42 GAO- 02- 566 Genetically Modified Foods

effect on human health, plant growth, or crop yield. In the case shown in
graph (c), scientists would probably not proceed with development and
commercialization of the GM food in the absence of extensive evaluations for
allergens, toxins, or antinutrients, due to the significant differences in
small molecules between it and its conventional counterpart.

Appendix III: Comments from the Department of Health and Human Services

Page 43 GAO- 02- 566 Genetically Modified Foods

Appendix III: Comments from the Department of Health and Human Services

Appendix III: Comments from the Department of Health and Human Services

Page 44 GAO- 02- 566 Genetically Modified Foods

Appendix III: Comments from the Department of Health and Human Services

Page 45 GAO- 02- 566 Genetically Modified Foods

Appendix IV: GAO Contacts and Staff Acknowledgments

Page 46 GAO- 02- 566 Genetically Modified Foods

Lawrence J. Dyckman, (202) 512- 5138 James R. Jones, Jr., (202) 512- 9839

In addition to the individuals above, Nathan J. Anderson, Dennis S. Carroll,
Kurt W. Kershow, and Cynthia C. Norris made key contributions to this
report. Appendix IV: GAO Contacts and Staff

Acknowledgments GAO Contacts Staff Acknowledgments

Related GAO Products Page 47 GAO- 02- 566 Genetically Modified Foods

International Trade: Concerns Over Biotechnology Challenge U. S.
Agricultural Exports GAO- 01- 727. Washington, D. C.: June 15, 2001.

Biotechnology: Information on Prices of Genetically Modified Seeds in the
United States and Argentina GAO/ T- RCED/ NSIAD- 00- 228. Washington, D. C.:
June 29, 2000.

Biotechnology: Information on Prices of Genetically Modified Seeds in the
United States and Argentina GAO/ RCED/ NSIAD- 00- 55. Washington, D. C.:
January 21, 2000. Related GAO Products

(360117)

The General Accounting Office, the investigative arm of Congress, exists to
support Congress in meeting its constitutional responsibilities and to help
improve the performance and accountability of the federal government for the
American people. GAO examines the use of public funds; evaluates federal
programs and policies; and provides analyses, recommendations, and other
assistance to help Congress make informed oversight, policy, and funding
decisions. GAO?s commitment to good government is reflected in its core
values of accountability, integrity, and reliability.

The fastest and easiest way to obtain copies of GAO documents at no cost is
through the Internet. GAO?s Web site (www. gao. gov) contains abstracts and
fulltext files of current reports and testimony and an expanding archive of
older products. The Web site features a search engine to help you locate
documents using key words and phrases. You can print these documents in
their entirety, including charts and other graphics.

Each day, GAO issues a list of newly released reports, testimony, and
correspondence. GAO posts this list, known as ?Today?s Reports,? on its Web
site daily. The list contains links to the full- text document files. To
have GAO e- mail this list to you every afternoon, go to www. gao. gov and
select ?Subscribe to daily E- mail alert for newly released products? under
the GAO Reports heading.

The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:

U. S. General Accounting Office 441 G Street NW, Room LM Washington, D. C.
20548

To order by Phone: Voice: (202) 512- 6000 TDD: (202) 512- 2537 Fax: (202)
512- 6061

Contact: Web site: www. gao. gov/ fraudnet/ fraudnet. htm E- mail: fraudnet@
gao. gov Automated answering system: (800) 424- 5454 or (202) 512- 7470

Jeff Nelligan, managing director, NelliganJ@ gao. gov (202) 512- 4800 U. S.
General Accounting Office, 441 G Street NW, Room 7149 Washington, D. C.
20548 GAO?s Mission

Obtaining Copies of GAO Reports and Testimony

Order by Mail or Phone To Report Fraud, Waste, and Abuse in Federal Programs

Public Affairs
*** End of document. ***