Labeling Biotech Foods: Opening Pandora’s Box?

The controversial subject of labeling of “biotech foods” in the United States recently has been reinvigorated by several occurrences.

These occurrences include the introduction of new legislation requiring the special labeling of foods (including meat and poultry) produced by modern biotechnology (H.R. 3377, 1999); the spillover effects of the European debate about “Frankenstein” foods (Klee, 1999) and requisite European Union (EU) labeling requirements (EU, 1998); a suit against the Food and Drug Administration (FDA) demanding mandatory labeling, among other things (Alliance for Biointegrity v. Shalala, 1998); and three public hearings on the subject of biotech food labeling and other topics (FDA, 1999a). Demands for labeling because of the consumer’s “right-to-know” also have complicated the labeling debate.

These events and other developments can raise, and, indeed, often already have engendered, a host of political, scientific, religious, philosophical, ethical, public policy, and legal questions. This article explores some of the FDA complexities, including legal principles and misconceptions, associated with demands for mandatory labeling or with voluntary labeling initiatives under current U.S. law, the Federal Food, Drug, and Cosmetic Act (FD&C Act).

What Is Biotech?
Discussions about “biotech labeling,” particularly the public debate, often do not involve any definitions or explanations of what actually is covered or required, either in terms of “biotech” or “labeling.” Both words can describe a variety of different technologies or label language.

In a definitional sense, perhaps in contrast to consumer perception, the word “biotechnology” is a collective term, encompassing a number of technologies, some old, others new (Korwek, 1997). The Office of Technology Assessment (OTA, 1984) defined it as processes using living organisms or parts thereof to make or modify products, to improve plants, animals, or to develop microorganisms for specific uses. This definition and others make it clear that the term “biotechnology” can be broad, encompassing a variety of traditional technologies and products thereof, such as bread, cheese, and hybrid corn, as well as newer genetic technologies, such as recombinant DNA (rDNA) techniques.

The use of rDNA methods has sometimes been referred to vaguely as “bioengineering” or “genetic engineering.” The U.S. regulatory agencies have often, but not always, focused only on rDNA techniques, not traditional methods of genetic alteration, such as mutagenesis. Similarly, the EU novel food requirements apply primarily, in the context of the biotech food labeling debate, to food products made by rDNA methods.

The distinction between old and modern biotechnology is still important, though. Historically, most foods derived from plants or animals have not been required to be labeled as to their sources or methods of development. Moreover, nearly all foods today are derived from cultivated plants and domesticated animals that have been genetically modified in some way, if only as a result of conventional breeding practices. These considerations and the lack of a clear understanding of the word “biotechnology” and other terms such as “GMO-free” and “non-GM” often are lost in the labeling debate.

--- PAGE BREAK ---

Key Food Labeling Provisions
It is worthwhile to note that the standards or requirements adopted by one U.S. federal agency are not necessarily relevant to other agencies. FDA is not the only federal agency involved in the labeling of food. Indeed, even within FDA, the Center for Food Safety and Applied Nutrition is responsible for the labeling of human foods, and the Center for Veterinary Medicine is responsible for the labeling of animal feeds. The U.S. Dept. of Agriculture is primarily responsible for the labeling of most meat and poultry products. USDA regulates foods derived from transgenetic livestock and poultry (USDA, 1991). The Bureau of Alcohol, Tobacco, and Firearms is primarily responsible for the labeling of most alcoholic beverages. The Federal Trade Commission typically governs food advertising, such as information provided in the print and broadcast media. Nonetheless, many of the principles discussed here could be used by all relevant government bodies and could apply to both the labeling and advertising of all foods.

Various sections of the FD&C Act provide FDA with different types of labeling authority. Most discussions of labeling fail to recognize this fact. Foods can be mislabeled (misbranded) for a variety of reasons, such as with respect to quality, purity, size, and freshness. They can be also misbranded based on explicit or implied representations. Many of the labeling provisions interweave with one another or are otherwise often related. Indeed, many FDA labeling initiatives undertaken in the past 20 years have involved various statutory provisions, not just one. Examples include labeling of protein hydrolysates to include the identity of the specific food source from which the protein is derived (FDA, 1999b) and labeling of irradiated finished foods (but not irradiated ingredients) to indicate that the food has been irradiated (FDA, 1988).

Section 403 of the FD&C Act is the key labeling provision, since it contains roughly 25 different parts on the misbranding of food. These and other labeling provisions can be divided into whether they relate to safety considerations involving the consumption of food, or to other, non-safety characteristics, such as quality and economic value. Most of the section 403 provisions relate to non-safety matters. As for the labeling provisions pertaining to health consequences, sections 409 and 721 pertain to FDA’s authority to promulgate regulations establishing the “safe use” of food and color additives. Many additive regulations therefore contain limits on the use of an additive, and several contain labeling requirements, based on one or more provisions of the FD&C Act. A notable example is the fat substitute olestra. Not only is it strictly limited in its food uses and levels of use, but also food products containing it must bear special labeling regarding its gastrointestinal effects (FDA, 1999c).

The material facts section of section 201(n) can also relate to safety (and nonsafety) considerations. This section, in conjunction with section 403(a)(1), provides that a food is misbranded if its labeling is false or misleading in any particular and in determining whether the labeling is misleading:

“There shall be taken into account (among other things) not only representations made or suggested by the statement, . . . but also the extent to which the labeling . . . fails to reveal facts material in light of such representations or material with respect to consequences which may result in the use of the [food]. . . .”

Additional sections of the law, other than section 201(n), pertaining primarily to non-safety considerations, include section 402(b), which prohibits the economic adulteration of food, such as where a valuable constituent has been omitted or another ingredient has been substituted in whole or in part; sections 403(b) or 403(i)(1), which prohibit, respectively, the sale of a food under the name of another food or which require foods to be identified by their common or usual name; section 403(c), which requires imitation foods to be identified as such; section 403(g), which requires a food bearing a standardized name to conform to the applicable standard of identity, where a standardized food is one for which a recipe is specified in FDA regulations for making it; and section 403(i)(2), which requires the label of a food to contain an ingredient statement when the food is fabricated from two or more ingredients.

Also, sections 401 and 403(g) allow FDA to establish and enforce standards of identity, including special labeling requirements, to promote honest and fair dealing in the interest of consumers. For example, a number of standards of identity require a label statement of how a food is processed. Identity standards for cheeses require the label statement, “Milk bleached with benzoyl peroxide,” where the milk is so processed (FDA, 1999d). Additional labeling requirements can apply to dietary supplements, and foods with health claims. Other relevant regulations govern nutrition labeling and nutrient content claims, such as “low fat” claims. All of these provisions in various ways and in various contexts could be used to require labeling of rDNA-derived foods that is the same as or different from that for other foods.

--- PAGE BREAK ---

FDA Labeling Policy
In 1992, FDA first addressed and explored the issue of mandatory labeling of foods derived from modern biotechnology, as well as safety considerations (FDA, 1992). In that policy statement, the agency referred to human foods and animal feeds derived from new plant varieties, whether such plants were developed by modern methods of genetic modification or by traditional plant breeding. In discussing the labeling of foods (including food ingredients) derived from new plant varieties, FDA focused on sections 403(i) and 201(n) of the FD&C Act. As mentioned above, the former section requires that foods or food ingredients be described by their common or usual name or, in the absence thereof, an appropriate descriptive term. The latter section requires that the labeling of a food must reveal all facts that are material in light of representations made or suggested by the labeling or with respect to consequences which may result from its use.

FDA made it clear that in either case labeling of foods was not required. The agency explained that it does not consider methods used in the development of plants, such as rDNA, hybridization, and mutagenesis, to be material information within the meaning of sections 201(n) and 403(a)(1). FDA also stated that the newer methods of genetic modification are extensions at the molecular level of traditional methods and that they are used to achieve the same goals as the traditional methods of plant breeding. It further noted that it is not aware of any information showing that foods derived by the new methods “differ from other foods in any meaningful or uniform way or that, as a class, foods developed by the new techniques present any different or greater safety concern than foods developed by traditional plant breeding.” Nonetheless, FDA also noted that consumers must be notified, by appropriate labeling, where a food derived from rDNA or other methods significantly differs from its traditional counterpart, such as where the common or usual name is no longer applicable or where a safety or usage issue exists to which consumers must be alerted, e.g., allergenicity.

In 1993, the agency issued a new notice dedicated to the labeling of foods derived from new plant varieties (FDA, 1993). In that notice, FDA repeated some of its earlier statements regarding the lack of need to label rDNA-derived food differently from other foods. It nonetheless sought data and information on a wide variety of labeling issues. It appropriately asked a variety of important questions about labeling, such as how to define “genetic engineering,” and asked for clarification of the circumstances under which labeling is believed to be necessary, and how such labeling should be accomplished, if required. FDA also asked for information on voluntary labeling, specifically on niche markets for labeled food indicating that it is not derived from the use of rDNA methods. That notice was never formally followed up on. Still, FDA has encouraged voluntary labeling that is truthful and not misleading, in accordance with section 403(a)(1).

In its recent announcement of public meetings (FDA, 1999a), FDA again sought additional information on a variety of public information issues, such as whether FDA’s current policy regarding the labeling of significant changes, including the introduction of allergens, should be maintained or modified, and whether its policy that the name of the food be changed when the common or usual name for the traditional counterpart no longer applies to the new food. The agency also asked whether additional information should be made available to the public about foods derived from “bioengineered” plants and, if so, what information is to be provided and who should be responsible for communicating this information. Finally, public comment was also sought on how additional information should be made available to the public, such as through phone lines, food labels, or the Internet. It is expected that, in response to the public hearings, FDA will modify its 1992 policy, although in the areas of mandatory or voluntary labeling what precisely it will do remains unclear.

--- PAGE BREAK ---

Mandatory Labeling of Modern Biotech-Derived Foods
The above discussion demonstrates that the topic of “labeling” can involve a variety of complex concepts and legal principles and different federal agencies. It also involves questions of where on the food label (e.g., the principal display panel or information panel) and how (e.g., type size, predominance, and language utilized) such information should be provided. Perhaps most important, the FD&C Act currently does not contain any explicit “consumer right-to-know” labeling requirements pertinent to this discussion, although some of the provisions of the statute clearly are aimed at providing consumers with certain types of information. In other words, there is no general, explicit right of consumers to know how their food has been processed or derived, particularly through the use of “genetic engineering,” whatever that term or similar terms are meant to convey or understood to mean. On the other hand, a number of provisions mentioned above, and others, could be used to require certain limited types of labeling of foods derived from the use of modern biotechnology, as FDA has noted in its policy statements.

To the extent that any plant breeding method results in changes in foods that are important to consumer health or the economic value of a food, then labeling may be necessary. For example, if modern (or older) techniques of genetic modification were to result in a food that contained an allergen, then special labeling may be required, a point that FDA acknowledged in both its 1992 and 1993 notices. If an important nutrient is significantly reduced beyond natural variations that occur in the food as a result of any food “production” method, then labeling may be required, at least to prevent the food from appearing to be of more value than it is compared to its conventional counterpart. Such labeling could take a number of forms. A change in the common or usual name of the food could be required; or the use of labeling statements indicating how the food has been changed compared to its traditional counterpart may be needed; or, perhaps, use of the term “imitation” to describe the food may be necessary, although most food companies avoid imitation labeling because of the inferiority it suggests.

Methods of plant breeding may also be relevant to meeting established food standards. Roughly 250 food standards exist, covering a variety of foods, including milk, cheeses, bakery products, fruit juice, fruit preserves and jellies, cereal, macaroni, and vegetable products. These standards, which often contain “source” requirements, could affect the labeling of foods developed though the use of rDNA methods. For example, the standard of identity for vanilla extract (FDA, 1999e) specifies that it “is the solution in aqueous ethyl alcohol of the sapid and odorous principles extractable from vanilla beans.” If vanilla extract were to be obtained from “engineered” microorganisms instead of vanilla beans, it may not meet the vanilla bean “source” requirement of the standard of identity. Therefore, it may not be able to be called “vanilla extract” and may have to be labeled as “imitation” or with some other language to clarify that it is not “vanilla extract” as defined by FDA regulation.

Generally, the economic adulteration provision of section 402(b) has been used to prevent the use of deceptive common or usual names for foods or food ingredients. A product that is labeled as “orange juice,” a standardized food, but that lacks orange juice and contains natural and artificial orange flavor with significant reductions in key vitamins and minerals probably would be found to be both adulterated and misbranded. As in the vanilla bean example, a different common or usual name of the food which accurately characterizes the product could be needed to differentiate the food from the standardized food named “orange juice.” Flavor labeling requirements may also be triggered (FDA, 1999f), requiring labeling of the fact that artificial flavor is added.

--- PAGE BREAK ---

Mandatory Labeling Based on Material Information
One key provision that is relevant to the labeling of whole foods or food ingredients is section 201(n), pertaining to material information that must be disclosed. FDA stated in its 1993 notice that historically the materiality concept in section 201(n) is limited to information about the attributes of the food itself. In other words, the food itself must somehow be changed in a material way before section 201(n) is triggered.

A relevant example is food irradiation labeling. Since FDA concluded that irradiation changes the organoleptic properties of whole foods and that consumers may assume that such foods were unprocessed without special labeling, the fact that a food has been irradiated was therefore deemed to be material information that must be disclosed under section 201(n). In the context of labeling of “genetic engineering,” however, particularly rDNA techniques, FDA has noted that plant breeding methods “are applied at an early stage of development of new plant varieties and are not processes applied to the finished food.” Thus, as noted above, FDA has stated that the method of development of new plant varieties is not material information that is required to be labeled.

The manner in which rDNA methods are characterized could be relevant to the discussion of how section 201(n) applies to labeling of foods derived by rDNA methods. For example, if rDNA methods were not characterized as plant breeding or plant development techniques, but instead as “sources” of food, then precedent exists for other “source” labeling. Such labeling has been required in isolated circumstances where the source has a significant effect on the products’ composition or formulation or if, for religious or cultural reasons, the agency decides that the source is a material fact. These reasons have been articulated as the basis for FDA’s requiring the labeling as ingredients of the specific vegetable source of protein hydrolysates, i.e., “hydrolyzed soy protein.” On the other hand, the “sources” of foods in the context of this rDNA discussion are “plants”, or particular types of plants, such as corn or soy. These different plant “sources” do not necessarily have a significant effect on a food’s composition or formulation. Moreover, FDA has consistently stated that a product’s method of development through rDNA is not material information.

Voluntary Labeling with “Absence of” or “Avoidance” Claims
It seems likely that, in the near future, foods derived from “genetically engineered” sources will be different from other rDNA-derived foods available today. These newer foods will have obvious consumer value, perhaps containing lower levels of saturated fat or higher levels of vitamins. Such foods will be promoted for their beneficial characteristics and, thus, like other foods with similar label claims, will be subject to a variety of FDA regulations, including those governing health claims, dietary supplements, and nutrient content claims.

Unfortunately, however, major food crops today, such as corn and soy, have altered characteristics, such as insect tolerance or herbicide resistance, that are of agronomic value primarily to the grower. Such input traits that are of value to farmers, or even food processors, obviously can involve different food labeling considerations than those output traits that are of primary value to the consumer. A growing trend today in food from plants with rDNA-derived input traits is not to label such foods differently from other foods, but instead to label their traditional counterparts as having not been derived from the use of “genetic engineering.” These claims about the absence of the use of rDNA, similar to nutrient content claims about the absence of sodium or fat, are sometimes called avoidance claims. They are meant to convey information to the consumer that helps them to avoid other foods which are derived from the use of rDNA or which do contain sodium or fat.

--- PAGE BREAK ---

The provision mentioned above pertaining to the fact that a food is misbranded if it is false or misleading in any particular would be especially relevant to avoidance claims such as “GM-free” or “GMO-free,” or “non-GM.” These terms or claims, which have been prevalent in Europe, recently have been proliferating in the U.S., as part of the development of niche markets similar to those for organic foods. Since these terms can convey a variety of different meanings to different people, they can be misleading, if not false, depending on what consumers understand them to mean. How claims might be understood by consumers is not necessarily based on what the reasonable person might infer, but what the “ignorant, unthinking or credulous” person understands (U.S. v. An Article, 1969).

In any event, these types of “-free” or “non-” claims can engender confusion based on their literal meanings. For example, few foods are themselves genetically modified (GM); rather, such foods are derived from plants or animals that are genetically modified, whether by traditional or modern biotechnology methods. Therefore, without at least qualification, it can be misleading, if not false, to suggest that foods with such GM claims are themselves not genetically modified since most foods are not so modified anyway. An equivalent labeling scenario would be to label celery as “fat-free,” which is prohibited because celery is normally “fat-free.” It is the plants (or animals) they are derived from that are genetically modified. If such claims can be shown to imply instead that the foods are derived from plants or animals that are not genetically modified, the claims still can be viewed as misleading, if not false. Most foods have been derived from sources that are genetically modified through cultivation or domestication. Similarly, with regard to the claim “GMO-free,” few foods, except products like yogurt, contain “organisms.” It therefore can also be misleading, if not false, to suggest that a food, which ordinarily would not contain organisms, is “organism-free,” regardless of whether the acronym “GM” accompanies such a claim.

Related to this discussion is the extent of FDA restriction that is permissible on the voluntary use of these catch phrases or terms. This topic involves the Constitutional First Amendment guarantee of free speech, particulary of commercial speech. In a recent suit against FDA involving its regulation of health claims for dietary supplements, the court in Pearson v. Shalala (1999) decided that FDA was required, under the Commercial Speech Doctrine, to consider whether the inclusion of appropriate disclaimers would negate the potentially misleading nature of health claims. This doctrine essentially embodies the preference for disclosure of information over out-right suppression or banning of a claim. The central issue was whether FDA could prohibit potentially misleading claims, rather than allow the speech with adequate qualifications through the use of disclosures or disclaimers. The court held that, in some cases, disclaimers could cure the misleading effect of such claims, as long as the claims were also not inherently misleading, in which case they were not subject to First Amendment protection.

In the context of “GMO-free” and similar types of claims, the Pearson case might suggest that while such claims, standing alone, are potentially misleading, depending on their meanings as understood by consumers, their misleading nature might be cured by appropriate disclosures or explanatory language. Thus, their use might not necessarily be unlawful, if they were appropriately qualified to explain what they mean. On the other hand, they could be viewed as inherently misleading and therefore unlawful; they would then have to be worded differently to convey what is actually meant.

Another significant issue associated with the use of these types of avoidance claims is whether they constitute implied superiority claims, in terms of food quality or safety. Of potential relevance to this discussion is FDA’s 1994 interim guidance on the labeling of milk and milk products in cows that have not been treated with recombinant bovine somatotropin, rBST (FDA, 1994). Citing sections 403(a) and 201 of the FD&C Act, FDA stated that claims such as “rBST-free” may imply a compositional difference between milk from treated and untreated cows rather than a difference in the way the milk is produced. Moreover, without proper context, such statements could be interpreted to imply that milk from untreated cows is safer or of higher quality than milk from treated cows. The agency therefore required that a disclaimer be used in conjunction with such “free” claims, stating that “No significant difference has been shown between milk derived from rBST-treated and non-rBST-treated cows.” By this line of reasoning, a similar disclaimer might therefore also be required for “-free” and “non-” type claims that are otherwise appropriately worded or qualified, such as “No significant difference has been shown between foods that are derived from plants genetically modified by newer recombinant DNA techniques and those that are derived from plants by traditional plant breeding methods.”

--- PAGE BREAK ---

A further complexity with such “-free” or “non-” claims is whether they mean different things and, if so, what level, if any, of adventitious de minimis presence of material from rDNA sources is allowable before a food product cannot be labeled with “-free” or “non-” type claims or terms of similar import. This adventitious material might occur during cultivation (e.g., from cross-pollenization of rDNA-derived plants with other plants), harvest, transport, storage, and processing. The EU has established an adventitious level of 1% per ingredient, whereby foods containing rDNA-derived materials above this de minimis threshold would have to be labeled as containing “genetically modified” ingredients (EU, 2000).

With regard to “-free” type claims, which seem to suggest a zero threshold, whereas “non-” type claims seem to suggest some allowable undefined level of rDNA-derived materials, some precedent exists. FDA has stated, for example, that “non-alcoholic” and “alcohol-free” are not synonymous. “Alcohol-free” may be used only when the product contains no detectable alcohol, whereas “non-alcoholic” beverages could actually contain traces (less than 0.5% by volume) of alcohol (FDA, 1995). On the other hand, FDA regulations on the use of “free” in regard to nutrient content claims concerning the lack of fat, sodium, and sugar allow 0.5 g per reference amount and per labeled serving, because this threshold level involves insignificant dietary amounts. Perhaps a preferable approach to dealing with “-free” or “non-” claims on labels of foods that are not rDNA-derived is to indicate the percentage of the food which does not contain rDNA-derived materials, such as, perhaps, “99% free from recombinant DNA-derived ingredients.”

Another complication associated with the use of thresholds and “-free” type claims is the necessity for validated, standardized testing methods, with appropriate reference materials and sensitivity, to detect the amount of adventitious contamination. Since such validated methods are not yet widely available commercially for all foods and for different types and levels of contamination, certification schemes may be relevant. Such certifications might be used to verify and document the source of the food or food ingredients throughout the food chain. Since certifications might not necessarily always cover the unavoidable presence of rDNA-derived materials that occurs in all foods, they still may need to be used in conjunction with testing.

Such dual schemes that involve both testing and certification requirements to ensure that a food has not been rDNA-derived or to verify the level of rDNA-derived materials that are present not only might be costly, but also might invite the creation of further systems, e.g., to reconcile a positive test result with documented certification that a food is not derived from a recombinant source.

A Sound Policy So Far, But Where Do We Go From Here?
Demands for labeling “biotech foods” differently from other foods can belie the complexity of the labeling debate. This view especially may be relevant to such demands based on the consumer’s right-to-know. This basis for requiring labeling can be simply too good to be true. The FD&C Act has been carefully worded, and FDA’s interpretations of this law in other food labeling contexts often have been cautiously applied to limit the types and amount of information required on a food label. Imagine the precedent that could be set if FDA were to change its position on the rationale and need for mandatory labeling of “biotech foods.” The legal floodgates could be opened for consumers to demand, and for FDA to require, a variety of endless label statements related to how all foods are derived or processed. FDA’s approach to mandatory labeling thus far has been sound from a variety of legal and other perspectives. It remains to be seen, however, whether the agency’s mandatory labeling position will change in light of all the legal and other pressures now being brought to bear on it, internationally and otherwise. In the area of voluntary claims, such as “GM-free” pertaining to avoidance of rDNA-derived products, these types of claims also will present challenges to FDA, both in terms of its interpretations of the FD&C Act, as well as in terms of its enforcement resources.

by EDWARD L. KORWEK
The author is a Partner in the law firm of Hogan & Hartson, L.L.P., 555 13th St., N.W., Washington, DC 20004.

Edited by Neil H. Mermelstein,
Senior Editor

References