More than 2,500 chemical substances are intentionally added to foods to modify flavor, color, stability, texture, or cost. In addition, an estimated 12,000 substances are used in such a way that they may unintentionally enter the food supply. These substances include components of food-packaging materials, processing aids, pesticide residues, and drugs given to animals. An unknown number of naturally occurring chemical contaminants also find their way into food. The most notable of these are the products of mold growth called mycotoxins, which include the aflatoxins. The association of these substances with carcinogenesis is described in Chapters 12, 13, and 14 of the committee's first report (National Research Council, 1982).

The introduction of a new food additive requires the prior approval of the Food and Drug Administration (FDA). This approval can be granted only when the FDA concludes that the manufacturer has submitted sufficient toxicological data to demonstrate the safety of the additive. Long-term studies to evaluate the carcinogenicity of “direct” additives, i.e., those intentionally added to food, may be required when the intended level of usage is high or when possible carcinogenicity is suspected because of the structure or known biological activity of the additive. This same policy applies to “indirect additives,” which are used in food packaging and as food processing aids. However, these substances are generally present in foods at such low levels that a carcinogenicity test requirement would be imposed only if the indirect additive were suspected of being a carcinogen because of its chemical structure or biological activity.

There has been no requirement to perform tests to determine carcinogenicity for most substances added to food. Substances in this category include those “generally recognized as safe” (GRAS), hundreds of flavoring agents, most additives approved before the 1958 Food Additives Amendment (P.L. 85–929) to the Food, Drug, and Cosmetic Act (U.S. Congress, 1958), and additives used at levels considered low by the FDA, except for suspected carcinogens. Furthermore, very little is known about the tumor-promoting activity of the few food ingredients that have been tested for carcinogenicity (National Research Council, 1982).

Of the additives that have been tested, those shown to be carcinogenic when administered orally to laboratory animals are generally prohibited from use. However, there are some exceptions. For example, Congress has passed special legislation (P.L. 95–203) preventing the FDA from restricting the use of the artificial sweetener saccharin, even thought it has been shown to induce tumors in test animals (U.S. Congress, 1977). In addition to saccharin, two other known carcinogens--vinyl chloride and acrylonitrile--may appear at very low levels in food as a result of their application in the manufacture of plastics used in food-packaging materials. According to a recently adopted policy, such chemicals as vinyl chloride and acrylonitrile are considered by the FDA to be “constituents” of food-packaging material rather than additives. Thus, the FDA believes that these chemicals may be exempted from the absolute legal prohibition that applies to carcinogenic additives (U.S. Food and Drug Administration, 1982a, b).

Residues of pesticides that can induce tumors may contaminate foods through their application directly on crops or from other sources of environmental contamination. Chemicals that are intrinsic constituents of foods, such as hydrazines in mushrooms, may also be carcinogenic. Certain unavoidable contaminants in foods, such as aflatoxin B₁ and polychlorinated biphenyls, have been found to be carcinogenic in long-term toxicological studies. Such contaminants are generally permitted in foods only up to levels that the FDA considers the lowest level generally attainable without resulting in severe economic losses or adverse effects on the food supply.

In addition to known carcinogens that may appear in food as natural constituents, contaminants, or additives, there are a number of chemicals in food whose carcinogenic potential has not been adequately assessed but which are suspected carcinogens because of their known mutagenic activity, i.e., they can cause heritable alterations in the genetic material of cells. Systems for determining the mutagenicity of chemicals include tests in bacteria, fungi, mammalian cells in culture, and laboratory animals. Positive results from any of these test systems may be of toxicological significance, because the genetic material, DNA, is similar in all organisms and the mutagenicity of chemicals, even to bacteria, has been correlated with carcinogenicity in animals (National Research Council, 1983).

Most of the studies that have been conducted to identify mutagens in foods have utilized bacteria (Salmonella typhimurium) as the target organism in the initial screening. Positive results in the bacterial assay generally lead to further testing in other systems. Substances that are negative upon initial screening are only rarely investigated further. Since different mutagenicity test systems may give different results with a given test chemical, the use of the Salmonella assay alone in screening for mutagenicity could lead to a failure to identify mutagens or carcinogens in foods. Therefore, it is important to use other genetic tests in addition to the Salmonella assay in the initial screening of foods and food components for mutagenic activity. Mutagens in foods identified by any one test system should be assessed for mutagenic activity in a variety of in vitro and in vivo mutagenicity test systems, in vitro transformation assays, and carcinogenicity tests in vivo. It would be ideal to supplement laboratory tests for mutagenicity with systems that could be used to assess mutagenic damage to human cells in vivo. The detection of chromosome aberrations in peripheral lymphocytes is the most widely used of such methods, but its apparent insensitivity limits its usefulness. Other tests, both for chromosome damage and for more subtle chemical changes in the DNA (i.e., gene mutations), are in various stages of development and may become suitable for application to populations consuming different diets. The development of such methods for detecting mutagenic effects in human cells in vivo is an important area for continued research.

There are several different sources of dietary mutagens. For example, intrinsic components of certain foods may be mutagenic. Into this category fall caffeine, other methylxanthines, and methylglyoxal (Kasai et al., 1982) in coffee as well as flavonoids in a wide variety of plants used for food. Other mutagens may be present in foods as naturally occurring contaminants such as aflatoxin B₁, as unintentional contaminants such as industrial chemicals or pesticides, or as intentionally used additives such as nitrites. In addition, mutagens may enter food during various food-processing techniques. For example, the smoking or charcoal-broiling of meat will result in the deposition of mutagenic polynuclear aromatic hydrocarbons such as benzo [a] pyrene; the cooking of some foods can result in the formation of potent mutagens, some of which are the products of the pyrolysis of amino acids; and nitrosamines can be formed during the frying of bacon that contains nitrite (National Research Council, 1982).

The significance of the presence of mutagens in food with respect to cancer risk is largely unknown. Some of the mutagens, such as aflatoxin B₁ and certain polynuclear aromatic hydrocarbons, are known to be carcinogenic. For others, such as nitrite and caffeine, long-term feeding studies in laboratory animals have failed to demonstrate carcinogenic activity, although endogenous reactions of nitrite with amines in the gastrointestinal tract can produce carcinogenic nitrosamines. The results of animal studies on the widely distributed flavonol quercetin are conflicting. Most mutagens in foods have not been adequately assessed for carcinogenic activity in animals. Only further research will enable us to decide whether significant health benefits might be derived from reducing the levels of mutagens that naturally occur in foods or of those that appear during cooking or processing of foods.

Dietary components may be converted to mutagenic (potentially carcinogenic) chemicals in vivo. The reaction between nitrite and amines to form nitrosamines is an example of such a reaction, as mentioned in Chapter 7.

Regulatory agencies regard chronic feeding studies in whole animals as the only definitive method for establishing the carcinogenicity of a chemical in foods. Thus, positive mutagenicity data are regarded only as an indication of the need for additional testing for carcinogenicity. When carcinogenicity in laboratory animals is established, a chemical is generally regarded and treated as if it were known to be carcinogenic in humans. However, there are no satisfactory methods for establishing, or even estimating, the magnitude of the cancer risk that may be associated with a given level of human exposure to a substance known to be carcinogenic in animals. Furthermore, the cancer risk associated with particular food additives cannot generally be determined through epidemiological studies, because the use of these additives is so widespread that the reliable identification of unexposed controls would not be feasible. Therefore, federal regulatory agencies have generally adopted the prudent policy of attempting to restrict the presence of known carcinogens in food to the lowest feasible levels, including outright banning of most carcinogenic food additives. The actual health benefit of this policy cannot be determined, however, since satisfactory methods of quantitative risk estimation for carcinogens do not currently exist.