A C&EN NEWS FORUM
Risk
_
Assessment of Pesticides
higher than what humans are exposed to. But recently, the value of the animal tests themselves has come under fire. To cast light on some of the issues surrounding pesticide risk, Ever since 1962, when Rachel Carson's book "Silent Spring" C&EN invited prominent figures representing five distinct made the American public aware of dangers to wildlife, synviewpoints to present their contrasting opinions on whether the thetic pesticides have been a focus of controversy. During the risks from pesticides have been grossly exaggerated; what can be early 1970s, scientific research and public concern led to the Environmental Protection Agency's cancellation of agricultur- done to encourage a more balanced public appreciation of their hazards and benefits; and whether there is really a middle al uses of the persistent pesticides, DDT, aldrin, and dieldrin. ground that accepts that some dangers may have been exaggerIronically, EPA restricted these pesticides, not primarily because of their effects on wildlife, but because many people be- ated but recognizes that the issue of health hazards is not a closed book and that much more research is needed. lieved they could cause cancer in humans. Ames and Lois Swirsky Gold of Lawrence Berkeley LaboraThe carcinogenic effects of pesticides on humans have retory contend that regulating low levels of synthetic pesticides mained the primary focus of their regulation ever since, and those effects are usually assessed with rodent studies. It was "by 'risk assessment' using worst-case risk scenarios is not scientifically justified." Instead of concentrating on standard ronot until 1986 that EPA canceled certain uses of a pesticide, dent bioassays, major efforts should be made to study mechadiazinon, solely because of its adverse effects on wildlife. nisms of carcinogenesis, Ames and Gold say. James E. Huff In recent years, much of the concern over pesticides has and Joseph K. Haseman of the National Institute of Environcentered around possible cancer risks from residues in food. The panic over ethylene dibromide residues in grain products mental Health Sciences assert, however, that in the absence of in the winter of 1983-84 and the similar outcry in early 1989 epidemiology data, current rodent bioassays provide valuable over Alar (daminozide) residues in apples and apple products information on the cancer-causing ability of pesticides and other chemical agents. arose over the cancer-causing ability of these pesticides. Will D. Carpenter, a vice president of Monsanto AgriculAt the same time that public concern over pesticide residues in food grew, Bruce N. Ames, professor of biochemistry at the tural, claims that pesticides present only minor risks, which University of California, Berkeley, reported that natural pes- need to be balanced against the enormous benefits they provide for agricultural production and human health. J. P. Myticides account for 99.99% of human pesticide exposure in ers and Theo Colborn of the W. Alton Jones Foundation argue food and concluded from this and other arguments that cancer risks from natural pesticides in food are hundreds or thou- that concentration on the cancer-causing ability of pesticides has diverted attention away from far more important risks sands of times greater than the hazards from synthetic pesticide residues. He also proposed that mitogenesis (cell division) both to humans and wildlife. Douglas D. Campt, James V. is the major cause of the cancers seen in rodent bioassays. Roelofs, and Jeanne Richards of EPA's Office of Pesticide Programs say that the agency has developed a sophisticated proDuring the past two years, Ames and his followers have used these theories to raise questions about the value of rodent car- cess to evaluate and manage pesticides, which, though not cinogenicity studies for assessing cancer risks to humans posed perfect, is basically sound, protective, and steadily improving. Each participant or group of participants submitted a writby pesticides and other chemicals. There has always been controversy over what risk assess- ten contribution, and then was sent copies of the other submissions and given an opportunity to write a rebuttal. Their conment methods should be used to extrapolate human cancer risk from rodent studies, which are usually conducted at doses tributions and rebuttals appear in the following pages. January 7, 1991 C&EN
27
News Forum
(cell division) or selective growth of preneoplastic cells, or both. The concept of promotion, however, has been fuzzy compared with the clearer understanding of the role of mutagenesis in carcinogenesis. The idea that mitogenesis increases mutagenesis helps to explain promotion and other aspects of carcinogenesis. A dividing cell is much more at risk of mutating than a quiescent cell. Mutagens are often thought to be only exogenous agents, but endogenous mutagens cause massive DNA damage (by formation of oxidative and other adducts) that can be converted to stable mutations during cell division. We estimate that the number of DNA hits per cell per day from endogenous oxidants is normally about 105 in the rat and about 104 in the human. This promutagenic damage is effectively, but not perfectly, repaired; the normal steady-state level of 8-hydroxydeoxyguanosine (one out of about 20 known oxidative DNA adducts) in rat DNA has been measured as one per 130,000 bases or about 90,000 per cell. We believe that this oxidative DNA damage is a major contributor to aging and the degenerative diseases associated with aging, such as cancer. Thus, any agent causing chronic mitogenesis can be indirectly mutagenic (and consequently carcinogenic) because it increases the probability of converting endogenous DNA damage into mutations. Nongenotoxic agents, such as saccharin, can be carcinogens at high doses just by causing cell killing, chronic mitogenesis, and inflammation. Therefore, the dose-response curve for such agents would be expected to show a threshold. Genotoxic chemicals, because they hit DNA, are even more effective than nongenotoxic chemicals at causing cell killing and cell replacement at high doses. Since genotoxic chemicals also act as mutagens, they can produce a multiplicative interaction not found at low doses, leading to an upward-curving dose response for carcinogenicity. Furthermore, endogenous rates of DNA damage are so high that it may be difficult for exogenous mutagens to increase DNA damage rates significantly at low doses that do not increase mitogenesis. Therefore, mitogenesis, which can be increased by high doses of chemicals, is indirectly mutagenic, and seems to explain much of carcinogenesis. Nevertheless, potent mutagens induce tumors at moderate doses in the presence of
Cancer prevention strategies greatly exaggerate risks Bruce N. Ames, University of California, Berkeley, and Lois Swirsky Gold, Lawrence Berkeley Laboratory
The attempt to prevent cancer by regulating low levels of synthetic chemicals by "risk assessment/' using worst-case, one-in-a-million risk scenarios is not scientifically justified. Testing chemicals for carcinogenicity at near-toxic doses in rodents does not provide enough information to predict the excess numbers of human cancers that might occur at low-dose exposures. In addition, this cancer prevention strategy is enormously costly, is counterproductive because it diverts resources from much more important risks, and, in the case of synthetic pesticides, makes fruits and vegetables more expensive, thus serving to decrease consumption of foods that help prevent cancer. The regulatory process doesn't take into account that: • The world of natural chemicals makes up the vast bulk of chemicals humans are exposed to. • The toxicology of synthetic and natural toxins is not fundamentally different. • About half the natural chemicals tested chronically in rats and mice at the maximum tolerated dose are carcinogens. • Testing at the maximum tolerated dose frequently can cause chronic cell killing and consequent cell replacement (a risk factor for cancer that can be limited to high doses), and ignoring this greatly exaggerates risks. • An extrapolation from high to low doses should be based on an understanding of the mechanisms of carcinogenesis.
Mitogenesis increases mutagenesis Many "promoters" of carcinogenesis have been described and have been thought to increase mitogenesis 28
January 7, 1991 C&EN
only background rates of mitogenesis. Detailed studies of mechanism are critically important.
Causes of human cancer Brian E. Henderson of the University of California and coworkers have discussed the importance of chronic mitogenesis for many, if not most, of the known causes of human cancer—for example, hormones in breast cancer, hepatitis B or C viruses or alcohol in liver cancer, high salt intake or Helicobacter infection in stomach cancer, papilloma virus in cervical cancer, asbestos or tobacco smoke in lung cancer, and excess animal fat and low calcium in colon cancer. For chemical carcinogens associated with occupational cancer, worker exposure has been primarily at high, near-toxic doses that might be expected to induce mitogenesis. Epidemiologists frequently discover clues about the causes of human cancer, and the resulting hypotheses are then refined by animal and metabolic studies. It appears likely that this approach will lead during the next decade to an understanding of the causes of the major human cancers. Current epidemiologic data point to several risk factors for human cancer: cigarette smoking (which is responsible for 30% of cancer deaths), dietary imbalances, infections, hormones, and occupation. According to the National Cancer Institute's 1987 statistics review, "The age adjusted mortality rate for all cancers combined except lung cancer has been declining since 1950 for all individual age groups except 85 and above/' Although incidence rates for some cancers have been rising, trends in recorded incidence rates may be biased by improved registration and diagnosis. Even though mortality rates for cancers at particular sites can be shown to be increasing (for example, nonHodgkins lymphoma, melanoma) or decreasing (for example, stomach, cervical, rectal), establishing causes remains difficult because of the many changing aspects of our life-style. Life expectancy continues to increase every year. Cancer clusters in small geographical areas are expected to occur by chance alone, and epidemiology lacks the power to establish causality in these cases. It is important to show that a pollution exposure that purportedly causes a cancer cluster is significantly greater than the background of exposures to naturally occurring rodent carcinogens.
Causes of cancer in animal tests Anim il cancer tests are conducted at near-toxic doses—the maximum tolerated dose (MTD) of the test chemical—for long periods of time, which can cause chronic mitogenesis. Chronic dosing at the MTD can be thought of as chronic wounding, which is known to be both a promoter of carcinogenesis in animals and a risk factor fc r cancer in humans. Thus, a high percentage of all chemicals might be expected to be carcinogenic at chronic, near-toxic doses and this is exactly what is found. j\bout half of all chemicals tested chronically at the MT1) are carcinogens. Synthetic chemicals account for 82% of the 427 chemicals adequately tested for carcinogenicity in both rats and mice. Despite the fact that humans eat vastly more
natural than synthetic chemicals, natural chemicals have never been tested systematically. Of the natural chemicals that have been tested, about half are carcinogens, which is about the same as found for synthetic chemicals. It is unlikely that the high proportion of chemicals found to be carcinogens in rodent studies is due simply to selection of suspicious chemical structures. Most chemicals were selected because of their use as industrial compounds, pesticides, drugs, or food additives.
Dietary pesticides: 99.99% natural Daniel H. Janzen of the University of Pennsylvania wrote, "Plants are not just food for animals. ... The world is not green. It is colored lectin, tannin, cyanide, caffeine, aflatoxin, and canavanine." Nature's pesticides are one important subset of natural chemicals. Plants produce toxins to protect themselves against fungi, insects, and animal predators. Tens of thousands of these natural pesticides have been discovered, and every species of plant analyzed contains its own set of perhaps a few dozen toxins. When plants are stressed or damaged, such as during a pest attack, they may greatly increase their natural pesticide levels, occasionally to levels that can be acutely toxic to humans. We estimate that Americans eat about 1.5 g of natural pesticides per person per day, which is about 10,000 times more than they eat of synthetic pesticide residues. Concentrations of natural pesticides in plants are usually measured in parts per thousand or per million rather than parts per billion, the usual concentration of synthetic pesticide residues or of pollutants in water. We estimate that the human diet contains roughly 5000 to 10,000 different natural pesticides and their breakdown products. For example, 49 natural pesticides (and metabolites) are ingested when cabbage is eaten. Only two have been tested for carcinogenicity. Lima beans contain a completely different array of 23 natural toxins that, in stressed plants, range in concentration from 0.2 to 33 parts per thousand fresh weight. None appears to have been tested yet for carcinogenicity or teratogenicity. Many leguminous plants contain canavanine, a toxic arginine analog that, after being eaten by animals, is incorporated into protein in place of arginine. Feeding alfalfa sprouts (1.5 % canavanine dry weight) or canavanine itself to monkeys causes a lupus erythematosuslike syndrome. Lupus in humans is characterized by a defect in the immune system that is associated with autoimmunity, antinuclear antibodies, chromosome breaks, and various types of pathology. The toxicity of nonfood plants is well known: Plants are among the most commonly ingested poisonous substances for children under five years of age. Surprisingly few plant toxins have been tested for carcinogenicity. Among 1052 chemicals tested in at least one species in chronic cancer tests, only 52 are naturally occurring plant pesticides. Among these, 27 are carcinogenic. Even though only a tiny proportion of the plant toxins in our diet has been tested so far, the 27 natural pesticides that are rodent carcinogens are present at levels above 10 ppm in the following foods: anise, apple, basil, brussels sprouts, cabbage, caraway, carrot, cauliflower, celery, cherries, cloves, brewed cofJanuary 7, 1991 C&EN
29
News Forum fee, comfrey herb tea, dill, eggplant, endive, fennel, grapefruit juice, grapes, honey, horseradish, lettuce, mango, mushrooms, brown mustard, nutmeg, orange juice, parsley, parsnip, pear, black pepper, plum, potato, rosemary, sage, heated sesame seeds, tarragon, and thyme. In addition, the following foods contain these 27 natural pesticides at levels below 10 ppm: apricot, banana, broccoli, cantaloupe, cinnamon, cloves, cocoa, collard greens, currants, guava, honeydew melon, kale, lentils, peach, peas, pineapple, radish, raspberries, tea, tomato, and turnip. Thus, it is probable that almost every fruit and vegetable contains natural plant pesticides that are rodent carcinogens. The levels of these 27 rodent carcinogens in the above plants are commonly thousands of times higher than the levels of synthetic pesticides. Caution is necessary in interpreting the implications of ingesting natural pesticides that are rodent carcinogens. It is not argued here that these dietary exposures are necessarily of much relevance to human cancer. What is important in our analysis is that exposures to natural rodent carcinogens may cast doubt on the relevance of far lower levels of exposures to synthetic rodent carcinogens. Particular natural pesticides that are carcinogenic in rodents can be bred out of crops if studies of mechanism indicate that they may be significant hazards to humans.
Residues of pesticides The Food & Drug Administration has assayed food for 200 chemicals, including the synthetic pesticide residues thought to be of greatest importance, and the residues of some industrial chemicals, such as polychlorinated biphenyls. FDA found residues for 105 of these chemicals. The U.S. intake of the sum of these 105 chemicals averages about 0.09 mg per person per day, which we compare with an intake of 1.5 g of natural pesticides. Thus, the average intake of pesticides is 99.99% natural. Other analyses of synthetic pesticide residues are similar. About half (0.04 mg) of this daily intake of synthetic pesticides is composed of four chemicals that are not carcinogenic in rodent tests: ethylhexyldiphenyl phosphate, chlorpropham, malathion, and dicloran. Thus, the intake of known or potential rodent carcinogens from synthetic residues is only about 0.05 mg a day. The cooking of food is also a major dietary source of potential rodent carcinogens. Cooking produces about 2 g per person per day of mostly untested burnt material that contains many rodent carcinogens—for example, polycyclic hydrocarbons, heterocyclic amines, furfural, and nitrosamines—as well as a plethora of mutagens. Thus, the number and amount of total synthetic pesticide residues, including those that are carcinogenic, appear to be minimal compared with the background of naturally occurring chemicals in the diet. Roasted coffee, for example, is known to contain 826 volatile chemicals; 21 have been tested chronically and 16 are rodent carcinogens; caffeic acid, a nonvolatile rodent carcinogen, is also present. A typical cup of coffee contains at least 10 mg (40 ppm) of rodent carcinogens (mostly caffeic acid, catechol, furfural, hydroquinone, 30
January 7, 1991 C&EN
and hydrogen peroxide). Thus, the 10 mg of known natural rodent carcinogens in a cup of coffee (only a few percent of the chemicals have been tested) would be equivalent in amount ingested to a year's worth of synthetic pesticide residues (assuming half of the untested synthetic residue weight turns out to be carcinogenic in rodents). The evidence on coffee and human health has been recently reviewed, and to date it is insufficient to show that coffee is a risk factor for cancer in humans. The same caution discussed above about the implications for humans of natural rodent carcinogens in the diet apply to coffee and the products of cooked food.
The case of dioxin Cabbage and broccoli contain a chemical whose breakdown products bind to the body's Ah receptor, induce the defense enzymes under the control of the receptors, and possibly cause mitogenesis—just as does 2,3,7,8-tetrachlorodibenzo-p-dioxin, commonly known as dioxin or TCDD, one of the most feared industrial contaminants. TCDD is of great public concern because it is carcinogenic and teratogenic in rodents at extremely low doses. The doses humans ingest are, however, far lower than the lowest doses that have been shown to cause cancer and reproductive damage in rodents. TCDD exerts many or all of its harmful effects in mammalian cells through binding to the Ah receptor. A wide variety of natural substances, such as tryptophan oxidation products, also bind to the Ah receptor, and insofar as they have been examined, they have similar properties to TCDD. A cooked steak, for example, contains polycyclic hydrocarbons that bind to the Ah receptor and mimic TCDD. In addition, a variety of flavones and other plant substances in the diet, such as indole carbinol (IC), also bind to the Ah receptor. IC is the main breakdown compound of glucobrassicin, a glucosinolate that is present in large amounts in vegetables of the Brassica genus, including broccoli (about 25 mg per 100 g portion), brussels sprouts (125 mg per 100 g), and cabbage (25 mg per 100 g). When tissues of these vegetables are lacerated, as occurs during chewing, they release an enzyme that breaks down the gluco'brassicin. The enzyme is quite stable when heated, and cooked vegetables yield most of the indole compounds that raw vegetables do. Therefore, we assume for the following calculation that 20% of glucobrassicin is converted to IC on eating. At the pH of the stomach, IC makes dimers and trimers that induce the same set of detoxifying enzymes as TCDD. IC, like TCDD, protects against carcinogenesis when given before aflatoxin or other carcinogens. However, when given after aflatoxin or other carcinogens, IC, like TCDD, stimulates carcinogenesis. This stimulation of carcinogenesis has also been shown for cabbage itself. Thus, these IC dimers and trimers appear to be much more of a potential hazard than TCDD, assuming that binding to the Ah receptor is critical for toxic effect. ... The Environmental Protection Agency's human "reference dose" (formerly "acceptable dose limit") of TCDD is 6 femtograms per kg of body weight per day.
Defenses that animals have evolved are mostly of a This should be compared with 5 mg of IC per 100 g of general type, as might be expected, because the number broccoli or cabbage. Although the affinity of one major of natural chemicals that might have toxic effects is so indole dimer in binding to Ah receptors is less than large. General defenses offer protection, not only that of TCDD by a factor of about 8000, the effective against natural but also against synthetic chemicals, dose to the Ah receptor from a helping of broccoli making humans well buffered against toxins. would be about 1500 times higher than that of a reference dose of TCDD, taking into account an extra factor These defenses include the following: of 1000 for the several-year lifetime of TCDD in the • Continuous shedding of cells exposed to toxins— body and assuming that the lifetime of the hydrophothe surface layers of the mouth, esophagus, stomach, bic indole dimers is as short as one day. intestine, colon, skin, and lungs are discarded every few days. Another IC dimer has recently been shown to bind to • Induction of a wide variety of general detoxifying the Ah receptor with about the same affinity as TCDD. mechanisms, such as antioxidant defenses or the Phase However, it is not clear whether either IC or TCDD is II electrophile-detoxifying systems. Cells that are exhazardous at the low doses of human exposure. They posed to small doses of an oxidant, such as radiation or may even be protective. It seems likely that many more hydrogen peroxide, induce antioxidant defenses and of these natural "dioxin simulators" will be discovered become more resistant to higher doses of oxidants, in the future. whether synthetic or natural. Natural or synthetic elecCompared with ethanol, TCDD seems of minor intertrophiles induce Phase II detoxifying enzymes that are est as a teratogen or carcinogen. Alcoholic beverages effective against both. are the most important known human chemical terato• Active excretion of planar hydrophobic molecules gen. In contrast, there is no persuasive evidence that (natural or synthetic) out of liver and intestinal cells. TCDD is either carcinogenic or teratogenic in humans, • DNA repair, which is effective against DNA adalthough it is both at near-toxic doses in rodents. If one ducts formed from both synthetic and natural chemicompares the teratogenic potential of TCDD to that of cals and is inducible in response to DNA damage. alcohol (after adjusting for their respective teratogenic Anticarcinogenic chemicals in the diet, such as anpotency, as determined in rodent tests), a daily contioxidants, help protect husumption of EPA's refermans against carcinogens ence dose of TCDD (6 fembut do not distinguish betograms per kg of body tween synthetic and natural weight) would be equivaTesting at near-toxic doses in rodents carcinogens. It has been arlent in teratogenic potendoes not provide enough information gued that synergism betial to a daily consumption tween synthetic carcinogens of alcohol from one threeto predict the excess human cancers that could multiply hazards, but millionth of a 12-oz bottle this is equally true of natuof beer. That is equivalent might occur at low-dose exposures ral carcinogens. to drinking a single beer The fact that defenses are (15 g ethyl alcohol) over a usually general, rather than period of 8000 years. specific, for each chemical makes good evolutionary Alcoholic beverages in humans are a risk factor for sense. The reason that predators of plants evolved gencancer as well as for birth defects. A comparison of the eral defenses against toxins is presumably to be prepared carcinogenic potential for rodents of TCDD with that of to counter a diverse and ever-changing array of plant alcohol (adjusting for the potency in rodents) shows toxins in an evolving world. If a herbivore had defenses that ingesting the TCDD reference dose of 6 femtoagainst only a set of specific toxins, it would be at a great grams per kg per day is equivalent to ingesting one disadvantage in obtaining new foods when favored beer every 345 years. Because the average consumption foods became scarce or evolved new toxins. of alcohol in the U.S. is equivalent to more than one Various natural toxins, some of which have been beer per person per day and five drinks a day are a carpresent throughout vertebrate evolutionary history, cinogenic risk in humans, the experimental evidence nevertheless cause cancer in vertebrates. Mold aflatoxdoes not of itself seem to justify the great concern over ins, for example, have been shown to cause cancer in TCDD at levels in the range of the reference dose. trout, rats, mice, monkeys, and possibly in humans. Similar toxicology Eleven mold toxins out of 16 tested have been reported to be carcinogenic. Many of the common elements, It is often assumed that because plants are part of husuch as salts of lead, cadmium, beryllium, nickel, chroman evolutionary history, while synthetic chemicals mium, selenium, and arsenic, are carcinogenic or clasare recent, the mechanisms that animals have evolved togenic (agents that break chromosomes) at high doses, to cope with the toxicity of natural chemicals will fail to protect us against synthetic chemicals. An example of despite their presence throughout evolution. Selenium this view is the statement of Rachel Carson: "For the and chromium, nevertheless, are essential trace elefirst time in the history of the world, every human bements in animal nutrition. ing is now subjected to contact with dangerous chemiHumans have not had time to evolve into a "toxic cals, from the moment of conception until death." We harmony" with all of the plants in their diet. Indeed, find this assumption flawed for several reasons. very few of the plants that humans eat would have January 7, 1991 C&EN
31
News Forum been present in an African hunter-gatherer's diet. The human diet has changed drastically in the past few thousand years, and most humans are eating many recently introduced plants that their ancestors did not— for example, cocoa, tea, potatoes, tomatoes, corn, avocados, mangoes, olives, and kiwi fruit. In addition, cruciferous vegetables, such as cabbage, broccoli, kale, cauliflower, and mustard, were used in ancient times primarily for medicinal purposes and spread as foods across Europe only in the Middle Ages. Natural selection works far too slowly for humans to have evolved specific resistance to the food toxins in these newly introduced plants. DDT bioconcentrates in the food chain as a result of its unusual lipid solubility. However, natural toxins can also bioconcentrate. DDT is often viewed as the typically dangerous synthetic pesticide because it persists for years. It is representative of a class of chlorinated pesticides. Natural pesticides bioconcentrate if lipophilic. For example, the teratogens from potato, solanine (and its aglycone solanidine) and chaconine, are found in the tissues of potato eaters. Although DDT is unusual with respect to bioconcentration, it is remarkably nontoxic to mammals, saved millions of lives, and has not been shown to cause harm to humans. To a large extent DDT, the first major synthetic insecticide, replaced lead arsenate, a major pesticide used before the modern era. Lead arsenate is even more persistent than DDT, and although natural, both lead and arsenic are carcinogenic. When the undesirable bioconcentration and persistence of DDT and its lethal effects on some birds were recognized, it was prudently phased out, and less persistent chemicals were developed to replace it. Examples are the synthetic pyrethroids that disrupt the same sodium channel in insects as DDT, are degraded rapidly in the environment, and can often be used at concentrations as low as a few grams per acre. Positive results are remarkably common in high-dose screening tests for carcinogens, clastogens, teratogens, and mutagens. About half the chemicals that have been tested, whether natural or synthetic, are carcinogens in chronic, high-dose rodent tests and about half are clastogens in tissue culture tests. A high proportion of positives is also reported for rodent teratogenicity tests: 38% of the 2800 chemicals that have been tested in laboratory animals have been teratogenic in the standard, highdose protocol. It is, therefore, reasonable to assume that a sizable percentage of both synthetic and natural chemicals will be reproductive toxins at high doses. Mutagens may also be common: 46% of 340 chemicals tested for carcinogenicity in both rats and mice and mutagenicity in Salmonella, are mutagens, and mutagens are nearly twice as likely to be carcinogenic than are nonmutagens. Of these 340 chemicals, 70% are either mutagens or carcinogens or both. How much this high frequency of positive results is due to bias in selecting chemicals is not known. Even if selection bias doubled the percentage of positives, which we think is unlikely, the high proportion of positives would still mean that almost everything natural we eat contains carcinogens, mutagens, teratogens, and clastogens. 32
January 7, 1991 C&EN
Thus, testing a random group of natural pesticides and pyrolysis products from cooking should be a high priority for these various tests so an adequate comparison can be made to synthetic toxins. These arguments undermine many assumptions of current regulatory policy and necessitate a rethinking of policy designed to reduce human cancer. Minimizing pollution is a separate issue and is clearly desirable for reasons other than effects on public health. There is a sizable literature on why focusing on worst-case, onein-a-million risks, rather than major risks, impedes intelligent risk reduction. It is by no means clear that many significant risk factors for human cancer will be discovered by screening assays. Dietary imbalances, such as antioxidant and folate deficiencies, are likely to be major contributors to human cancer, and understanding these should be, but is not, a major priority of research. Understanding why caloric restriction dramatically lowers cancer and mitogenesis rates and extends life span in experimental animals should also be a major research priority. More studies on mechanisms of carcinogenesis also should be of high priority. Synthetic pesticides have markedly lowered the cost of vegetables and fruit, thus increasing consumption. Other than giving up smoking, eating more fruits and vegetables and less fat may be the best way to lower risks of cancer and heart disease. • Bruce N. Ames is director of the National Institute of Environmental Health Sciences Center and professor of biochemistry at the University of California, Berkeley. He received a B.A. degree in chemistry from Cornell University in 1950 and a Ph.D. from California Institute of Technology in, 1953. He is a member of the National Academy of Sciences and was formerly on the board of directors of the National Cancer Institute. In 1983, he was the recipient of the General Motors Cancer Research Foundation Prize and in 1985 he received the Tyler Prize for environmental achievement. He has been elected to the Royal Swedish Academy of Sciences and the Japan Cancer Association. His major research interests include identifying agents damaging human DNA and their role in aging and cancer, endogenous oxidants and defenses against them, and mutagenesis and cancer. He has published about 250 scientific papers in these areas. Lois Swirsky Gold is director of the carcinogenic potency project at Lawrence Berkeley Laboratory and also works at the National Institute of Environmental Health Sciences Center at the University of California, Berkeley. She is on the panel of expert reviewers that evaluates rodent carcinogenesis studies for the National Toxicology Program. Gold has directed the development and analyses of Lawrence Berkeley's Carcinogenic Potency Database since 1978. The database provides information on more than 4000 experiments on 1050 compounds. Gold's research has focused on both qualitative and quantitative issues in carcinogenesis and interspecies extrapolation, carcinogenic potency, and the correlation between mutagenic and carcinogenic potency. Gold did her undergraduate work at Goucher College, Towson, Md., and received her Ph.D. in research methodology in political science from Stanford University in 1968.
pesticides has been widespread only during the past several decades, there has been insufficient time for humans to develop some of the adverse long-term health effects, such as cancer, or to develop resistance to diseases that may be associated with pesticide exposures. Numerous and myriad adverse health effects in humans, livestock, and wildlife have been documented as being associated with pesticide exposure. Typically, they involve both the acute and delayed manifestations of toxicity to the nervous and reproductive systems. Because pesticides are designed and selected for their biologic—that is, toxicologic— activity, exposure and toxicity to nontarget species are inevitable and remain significant potential problems. Pesticides have been shown capable of disrupting virtually every major organ system. These adverse effects include altered immune function, mutagenic and teratogenic responses, embryo toxicity and reproductive failure, and an array of neurologic effects. Although these adverse health effects associated with short-term pesticide exposure have long been known, it is only recently—within the past 10 or 20 years—that the adverse health effects of long-term human exposure to pesticides have even begun to be investigated. This article focuses on the evidence linking agrochemical pesticide exposures with cancer. Based on our experience in designing, conducting, and evaluating approximately 200 rodent carcinogenicity studies on a variety of chemicals, including a number of pesticides, we believe there is considerable evidence that exposure to certain pesticides may present real carcinogenic hazards to humans. Our conclusions are based on results of laboratory animal experiments and recent epidemiology studies.
Exposure to certain pesticides may pose real carcinogenic risk James E. Huff and Joseph K. Haseman, National Institute of Environmental Health Sciences
Synthetic chemical pesticides are currently essential for our way of life in modern society. According to the Environmental Protection Agency, in 1988 more than 1 billion lb of pesticides and related products were used in the U.S.: herbicides (680 million lb), fungicides (132 million lb), insecticides (268 million lb), and other related chemicals (70 million lb). This quantity corresponds to more than 4 lb of pesticides for each person in the U.S. While pesticide usage provides benefits for crop protection, evidence is increasing that the benefits are getting smaller, as more pests become resistant (together with the decimation of the natural insectivores), and as more lethal pesticides are used to kill fewer of the target pests. For example, the Nov. 10, 1989, issue of Science reported that although the amount of pesticides used on U.S. crops has increased greatly since World War II, the annual crop losses to insects have increased from 7% in the 1940s to 13% in the 1980s. Moreover, nearly 450 species of insects, ticks, and mites have become resistant to some or all pesticides. Further, according to Jack R. Plimmer of the U.S. Department of Agriculture, 97 to 99% of pesticide spray applications never reach the target population. As the cycle continues, the opportunity for continued and widespread exposure to pesticides becomes more likely. Pesticides are a particularly interesting class of chemicals, unique in that they are specifically designed to kill living organisms. In addition to having desired effects on target organisms, this diverse group of agrochemicals also has toxic effects on nontarget organisms, including wildlife and humans. Further, since the use of chemical
Value of animal studies The value of laboratory animal studies is unmatched in the annals of physiology, biochemistry, pharmacology, medicine, toxicology, and cancer. The progress made in these disciplines would have been impossible without animal studies. Research utilizing laboratory animals is fundamental to the discovery of new drugs and other beneficial chemicals, as well as to the identification of potential hazards these agents may pose for public health. Clinical and epidemiologic studies obviously offer the most direct means for assessing the human health January 7, 1991 C&EN
33
News Forum thesis that "everything causes cancer" when adminrisks associated with such exposures. However, for istered to animals at relatively high exposure levels chronic diseases such as cancer, the long time interval for long periods of time. Therefore, when chemically or latency period that almost always exists between exposure to the agent and the onset of clinically recognizrelated carcinogenic effects occur in these studies, able disease may mean that human experience with the the results should be taken seriously from a public agent to determine its full toxicological potential may health viewpoint. be insufficient. In other instances, adequate historical Laboratory animal pesticide studies have been very exposure data may be lacking. In the case of pesticides, consistent with respect to the development of cancer there is a further problem: People are typically exposed between sexes and between species. Within a single to a variety of different pesticides, and it is difficult for species (rats or mice), males and females are either both epidemiology studies to identify specific pesticides that positive or both negative for carcinogenicity about 90% may be associated with observed carcinogenic effects. of the time. When rats and mice are compared, both For these reasons, data generated from animal experispecies give positive or negative results for carcinogements often form the primary basis for the identificanicity in about 75% of the studies. tion and evaluation of possible human health risks. High exposure levels and cancer The vast majority of the scientific community in the Bioassays are usually conducted at relatively high exU.S. and abroad endorses the scientific value of laboraposure levels to optimize the sensitivity of the studies tory animal studies and supports the view that in the for detecting carcinogenic effects. These concentrations absence of extensive epidemiology data, long-term rofor the most part are above the environmental levels to dent studies are currently the best method available for which people are exposed, and some scientists consider evaluating a chemical's carcinogenic potential. These these "high" exposures to be controversial. The levels scientists agree with the conclusions of the Internationused, however, are often in the range of workplace exal Agency for Research on Cancer (IARC) that in the posures. Ethylene dibromide (EDB), l,2-dibromo-3absence of adequate data on humans, it is biologically chloropropane (DBCP), methylene chloride, phenaceplausible and prudent to regard agents for which there tin, 1,3-butadiene, and benzene are examples of chemiis sufficient evidence of carcinogenicity in experimencals for which the levels tal animals as if they preused in lab animal tests sented a carcinogenic risk were similar to those in the to humans. Those who reworkplace. ject the scientific value of Rodent carcinogenicity studies identify animal studies in the overUse of high exposure levall evaluation of a chemichemicals that cause cancer in lab animals els in rodent carcinogenicical's carcinogenic potential ty studies is widely recogand help to decide which will most likely are often quite vocal, but nized in the scientific comsuch extreme views do not munity as essential, since a pose significant human health risks represent the mainstream limited number of aniscientific thinking on this mals—generally 50 per exmatter. perimental group—are serving as surrogates for a large human population—10 For eight of the 54 known human carcinogens, the million farmers and farm workers, for example. The exevidence of carcinogenicity was first obtained in laboposure levels selected for long-term cancer studies are ratory animals. Furthermore, all chemicals known to generally below those that cause observable toxic efinduce cancer in humans that have been studied under fects in shorter term studies. Even so, some scientists adequate experimental protocols and conditions also have hypothesized that the high exposure levels may cause cancer in laboratory animals. Consequently, produce carcinogenic effects that would not be seen at when a pesticide or other chemical is found to be clearlower doses. ly carcinogenic in a well-designed laboratory animal In our experience, however, NTP bioassays consisstudy, we believe that an appropriate scientific retently yield results indicating that cancer incidences are sponse is to limit or eliminate human exposure, while increased in animals exposed to levels below the top at the same time carrying out additional studies to betexposure concentration (the maximum tolerated dose, ter understand the underlying biological mechanisms or MTD). In the pesticide studies we evaluated, 122 difof such a response. ferent site-specific carcinogenic effects were observed Over the past 20 years, the National Cancer Instifor the 31 positive chemicals. For 91% of these effects, tute (NCI) and the National Toxicology Program there was a numerical (and in most cases a statistically (NTP) have carried out nearly 400 long-term rodent significant) increase in these same tumor types at expocarcinogenicity studies, generally using male and fesure levels below the top exposure level. Similar results male rats and mice. Uniquely, these data are reported for other chemical carcinogenesis studies have been reand peer reviewed in public session. Seventy-two ported by David Hoel and coworkers at the National pesticides have been evaluated, with only 31 found Institute of Environmental Health Sciences [Carcinogento cause cancer in one or more of the experimental esis, 9, 2045 (1988)]. These results do not support the groups. This is true even though animals were exview that carcinogenesis mechanisms at high exposure posed to relatively high levels of pesticides for up to levels produce peculiar results not found at lower levtwo years. Clearly, these results do not support the 34
January 7, 1991 C&EN
els. Rather, they indicate that in most cases, increases in tumor rates are present at lower doses, and the limiting factor in detecting significant tumor increases at these levels is the inherent insensitivity of a bioassay using a relatively small number of animals. One might ask: What about the site-specific carcinogenic effects where tumor incidence did not increase below the highest exposure level? For each of the pesticides producing these responses, there are at least two other types of cancer that showed increased tumor rates at lower exposure levels. Thus, not one of the 31 positive agrochemicals is a "high-dose-only" carcinogen.
while target organ toxicity may be associated with certain carcinogenic effects, it is an oversimplification to assume that toxicity can account for all or even most of the varied cancers found in these studies. The many different sites and types of carcinogenic effects observed, coupled with the diverse structures of the pesticides, make it difficult or impossible to identify specific underlying mechanisms of action or to make global or generic statements about structure-activity relations.
Ranking pesticide carcinogenicity
Pesticides show notable differences in terms of the strength of the carcinogenic effects observed in rodent Toxicity and carcinogenicity studies. For example, EDB and DBCP produce from six to nine different types of cancers, with nearly a 100% Recently, Bruce N. Ames of the University of Califortumor response even in the low-exposure groups. nia, Berkeley, and others have claimed that chemical Moreover, the exposure levels in the inhalation studies toxicity is, in essence, responsible for carcinogenicity. that produce striking carcinogenic effects are quite low The toxic effects are said to be most pronounced at high for both DBCP (0.6 and 3 ppm doses) and EDB (10 and exposure levels, leading to carcinogenic effects unique 40 ppm). For both chemicals, exposure concentrations to these levels. At lower exposure levels, toxic effects such as these have been reported in the workplace. are less frequent or absent, and thus it is argued that DBCP is also found to cause infertility in laboratory anincreases in carcinogenicity would not occur at these imals, as well as in male agricultural workers and pestilevels. cide manufacturers. Clearly, the experimental results Pesticide studies offer little evidence to support this and human data lead to the conclusion that exposures theory. As noted above, carcinogenic effects were gento such chemicals should be minimized. erally not limited to the top exposure level, but were Certain other pesticides also frequently observed at produce less pronounced lower levels where there is carcinogenic effects that are no evidence of toxicity. limited to a single sexFurthermore, these data Epidemiology studies should serve to species group and/or to a suggest that even lower single tumor type. Alremind us that long-term exposure to doses may also have prothough the potential hazduced carcinogenic effects pesticides can cause insidious diseases and ards associated with such had they been included in chemicals cannot be disthe studies. Moreover, if may be associated with cancer in humans missed and require further toxicity were the cause of study, these pesticides carcinogenicity, we would would appear to pose less expect to see carcinogenic cause for concern than multisite, multispecies carcinoeffects associated with toxicity on an organ-by-organ gens. Such differentiations are quite valuable to the (site-by-site) basis. However, in chemical carcinogeneregulatory agencies and perhaps would allow them to sis studies it is not unusual to find carcinogenicity concentrate their efforts on those chemicals most likely without toxicity in the organ in which the cancer deto pose significant human health risk. veloped. We also observe severe toxicity developing in organs Classifying carcinogens where cancer does not occur. Many such examples were The NTP Annual Report on Carcinogens was mandocumented by Hoel and coworkers in their evaluation dated by Congress in 1978 and lists those chemicals in 1988 of the possible correlation between toxicity and known or anticipated to be human carcinogens. Eleven carcinogenicity in long-term rodent studies. A recent of the 31 pesticides causing cancer in rodents in NCI/ evaluation by Robert R. Maronpot and his colleagues at NTP studies are listed in the NTP Fifth Annual Report NIEHS on another set of chemicals has confirmed these on Carcinogens. These 11 chemicals, which "may reafindings—that is, certain chemicals cause toxicity and sonably be anticipated to be [human] carcinogens," are no cancer and others cause cancer and no toxicity. the pesticides producing the most striking carcinogenic Similar results are seen in our pesticide studies. For effects in laboratory animal studies: EDB, DBCP, 2,4,6example, in 10 of these studies, evidence of carcinogetrichlorophenol, mirex, 1,2-dichloroethane, sulfallate, nicity is limited to the mouse liver, but for only three nitrofen, toxaphene, 1,3-dichloropropene, 1,4of these pesticides is liver toxicity reported. Conversedichlorobenzene, and 3-chloro-2-methylpropene. Use ly, the pesticide monuron produces considerable liver of the first seven of these pesticides has been either toxicity in male mice, yet does not increase the incicanceled or suspended. The use of toxaphene has been dence of liver neoplasms. severely restricted, and 1,3-dichloropropene is currentOf course, target organ toxicity may occur together ly under special review by EPA. Records for the final with carcinogenicity. For example, 1,4-dichlorobenzene two pesticides are unavailable. shows toxicity at all sites of carcinogenicity. However, January 7, 1991 C&EN
35
News Forum IARC is an agency of the United Nations' World Health Organization. It convenes workgroups of experts in chemical carcinogenesis to prepare definitive evaluations of published data from animal and human cancer studies. The resulting monographs IARC publishes (Volume 48 is the latest) reflect a consensus view of the experts on the evaluation of carcinogenic risks to humans. Ten of the 11 pesticides noted above are considered by the agency to be Category 2 carcinogens— that is, chemicals "probably carcinogenic" (EDB) or "possibly carcinogenic" (the other nine pesticides) to humans. The 11th pesticide, 3-chloro-2-methylpropene, has not yet been evaluated by IARC. The classification criteria used by both the NTP Annual Report on Carcinogens and IARC generally require a carcinogenic response in at least two species before a chemical is considered to be a possible human carcinogen. Most of the pesticides showing lesser or more limited evidence of carcinogenicity in laboratory animal studies have not been classified by either organization as being possibly carcinogenic to humans. Thus, rodent carcinogenicity studies not only allow chemicals producing carcinogenic effects in laboratory animals to be identified, but also help in deciding which chemicals will most likely pose significant human health risks. While a relatively small number of the pesticides evaluated in laboratory animals have shown evidence sufficient for them to be considered possible human carcinogens, for those that have been so designated, the experimental animal evidence for carcinogenicity is quite strong. These effects are generally not limited to the top exposure level and are often not associated with target organ toxicity. Further, in some instances, the doses producing carcinogenic effects in animal studies are within the range of reported human exposure levels. An immediate public health concern should arise for chemicals producing such effects.
Recent epidemiological evidence Given the difficulties inherent in carrying out good epidemiology studies of exposures to chemicals and associated diseases, when adequate studies do appear, they merit particularly close attention. Within the past few years, several well-designed epidemiology studies have indicated that there are links between cancer and exposure to pesticides. Sheila H. Zahm and her colleagues at NCI report that 2,4,-D, a chlorophenoxy herbicide present in agent orange (a defoliant used in Vietnam during the war there) appears to increase the incidence of nonHodgkin's lymphoma in a large, population-based casecontrol study of Nebraska agricultural workers [Epidemiology, 1,349 (1990)]. Further, Linda M. Brown and coworkers at NCI report that pesticide exposures are related to an increased incidence of leukemia in farmers from Iowa and Minnesota [Cancer Research, 50, 6585 (1990)]. These two investigations are not isolated examples. Each provides confirming evidence of similar carcinogenic effects reported in earlier epidemiological investigations. Other epidemiology studies indicate increased cancer 36
January 7, 1991 C&EN
risks associated with pesticide exposure. For example, Michael C. R. Alavanja and coworkers at NCI reported excess risks for developing non-Hodgkin's lymphoma, leukemia, and pancreatic cancer in workers employed in flour mills, where pesticides are used frequently [Journal of the National Cancer Institute, 82, 840 (1990)]. An epidemiology study conducted by Jonathan D. Buckeley of the University of Southern California and other scientists from the Children's Cancer Study Group found that a significantly increased risk of acute nonlymphocytic leukemia in children was associated with paternal and maternal pesticide exposure [Cancer Research, 49,4030 (1989)]. Further, the Aug. 15,1990, issue of Pesticide & Toxic Chemical News reports that Rohm & Haas has found a significant association between cancer of the pancreas and exposure to DDT and other pesticides at its Philadelphia plant. Unfortunately, worldwide trends in cancer mortality are increasing in the older age groups. Devra Lee Davis of the National Research Council has documented that all forms of cancer mortality except lung and stomach are increasing in persons over age 54, and that these increases cannot be attributed solely to changes in diagnosis or to increased access to heialth care [Lancet, 336, 474 (1990)]. Interestingly, the specific types of tumors mentioned as showing clear increases include some of the same cancers (for example, non-Hodgkin's lymphoma) that have independently been shown to be associated with pesticide and other chemical exposures. The results of these various epidemiology studies should serve to remind us that long-term exposure to pesticides can cause insidious diseases and may indeed be associated with the development of cancer in humans. Thus, due caution must be exercised in the manufacture and use of these chemicals, and in some cases, appropriate regulatory action may be needed to reduce or eliminate human exposure. Although there are recognized benefits associated with pesticide usage, the health risks must also be recognized and dealt with properly. Based on our experience in the area of carcinogenicity evaluation, we believe there are compelling reasons to be concerned about the dangers of unbridled or incautious exposure to certain pesticides. Many well-conducted laboratory animal studies have shown striking carcinogenic effects, and increasing numbers of scientific reports link the induction of cancer to human exposures to pesticides. The aim of toxicology always has been the prevention of injury. This remains the guiding principle of most scientists in the field of experimental carcinogenesis. The emphasis on prevention and finding means of choosing less hazardous ways of working with chemicals, including the development of appropriate regulations and controls, should receive even more attention in the next decades. • James E. Huff is senior toxicologist and associate director for risk evaluation in the Division of Biometry and Risk Assessment at the National Institute of Environmental Health Sciences, where he has worked since 1979. Before that he was a
News Forum scientist at the International Agency for Research on Cancer in Lyons, France. Huff did his undergraduate work at Philadelphia College of Pharmacy & Science and received a Ph.D. degree in biophysics and pharmacology from Purdue University in 1968. He has published about 200 scientific papers on pharmacology, toxicology, and carcinogenesis. Since 1986, he has been a member of the Toxicology Information Program Committee at the National Research Council. He has received three technical communication awards from the Society for Technical Communication. Joseph K. Haseman is a research mathematical statistician in the Division of Biometry and Risk Assessment of the Na-
tional Institute of Environmental Health Sciences. He has been actively involved in the National Toxicology Program's long-term rodent carcinogenicity studies since 1980 and is responsible for the statistical integrity of these studies. He has published 140 scientific papers and is a fellow in the American Statistical Association, an honor accorded less than 5% of its membership. In 1983, he received the National Institute of Health's director's award for advice and service to the application of statistical methods for the National Toxicology Program. Haseman received a B.S. degree from Davidson College, North Carolina, in 1965 and a Ph.D. in biostatistics from the University of North Carolina, Chapel Hill, in 1970.
Insignificant risks must be balanced against great benefits
Will D. Carpenter, Monsanto Agricultural Co.
At self-service gasoline stations throughout America, men, women, and teenagers routinely fill the tanks of their cars, apparently unconcerned about the health warnings that are posted on the pumps: "Long-term exposure to gasoline vapors has caused cancer in laboratory animals. Avoid prolonged breathing of vapors." Have motorists concluded that the risk of contracting cancer from such limited exposure to fumes is negligible? Are the savings they realize from pumping their own gas worth some minimal risk that they might get cancer? One would like to think that these people believe the axiom "the dose makes the poison," and that the amount of gasoline they are being exposed to is safe. Or, perhaps because gasoline is a common, very familiar substance, their fear about it is diminished. Whatever reasoning leads people to judge the risk of gasoline fumes worthwhile has not been applied to the consumption of food. The level of chemicals and pesticides that sometimes are detected in our food is a minute fraction of the dosage that could cause cancer in laboratory animals. But the fear of food contamination
is increasing. Half the people responding to a 1989 Roper poll said they regarded pesticide residues on food as a "very serious" problem. This concern is in part psychological. We know what gasoline is, how it is used, and what would happen to our world if we suddenly found ourselves having to live without it. People do not seem to feel the same way about what pesticides are, how they are used, and most important, what life would be like in a world without them. The fear of pesticides has been exacerbated by frenzied publicity over celebrated cases, such as the "60 Minutes" report on the apple spray Alar (daminozide). By any rational standard, Alar poses no risk to human health. Someone would have to ingest the equivalent of 28,000 apples a day for life to duplicate the dosage that caused the animal cancer that aroused the national panic. Yet Alar has been banned, and many other beneficial applications of chemicals and pesticides have been eliminated in the current climate of fear. As growers continue to lose products that help them increase yields and control pests, we must come to grips with the reality that lower production and higher food costs to consumers may dramatically affect society's ability to meet its citizens' essential needs. Why are the same consumers who accept minimal risk to save money on gasoline willing to pay more for food rather than submit to what many noted scientists have determined is an insignificant risk from pesticide residues? Perhaps Will Rogers said it best: "It ain't the January 7, 1991 C&EN
37
News Forum The Delaney clause is not compatible with today's things people don't know that get them in trouble. It's sophisticated laboratories, which can detect parts per all the things they know that ain't so." billion. And even more sophisticated analytical equipHow ironic it is that panic about our food supply is ment is now detecting parts per trillion, so now what? occurring at a time when life expectancy has never A policy of hypercaution would dictate that we add anbeen higher and when the U.S. food supply has never other measure of safety and ban applications at that been in better shape. In December 1989, the U.S. Food level of microrisk. & Drug Administration concluded that pesticides pose little threat to food safety. FDA's survey of 18,114 food John Parnell, the Department of Agriculture's depuproducts found 96% of them to be free of pesticides or ty secretary, had this to say about the problem: "Sciento contain pesticide residues well within the legal limtific advances that allow pesticide residues to be detectits. Most of the 4% in noncompliance were technical vied at very low levels in food have surpassed our sociolations, such as food products with residues of pestiety's wisdom to determine how significant those cides not approved for that crop. Only 1% of the foods findings are. As a result, perceptions of problems—not that were tested contained pesticide residues above lethe problems themselves—have tended to drive public gal limits. policy on this issue. These concerns must be put in proper perspective." California, a state that monitors food more closely Similarly, Michael J. Boskin, President Bush's chief than the federal government does, found in a 1989 economic adviser, warns: "There is no such thing as study that less than 1% (0.71%) of produce contained ilzero risk, and the effort to reach such an unrealistic legal residues. Again, two thirds of the violations ocgoal can have enormous costs in economic and human curred because no tolerance level had been established terms." for the specific commodity on which the chemical had been applied. Some estimates indicate that food production would decline 40 to 50% without pesticides. The risk of starvaAt the same time our food supply is at its safest level, tion and malnutrition, not only worldwide but in this cancer mortality rates are holding steady or declining, country, becomes increasingly real under that scenario. with the notable exception of skin cancer and lung canScience editor Daniel E. Koshland Jr.'s editorial titled cer, which is directly linked to smoking. Two facts about cancer mortality "Scare of the Week" put it rates are very interesting: this way: "The public must According to the National recognize that a risk-free soCenter for Health Statisciety is not only impossible, Rather than debate the risk of pesticides tics, rates for stomach canbut intolerably expensive. versus the benefit of pesticides, we should cer, which is directly relatAt some point, the real daned to food, have decreased ger of too much pesticide consider the risk of pesticides versus the the most (75%) over the must be balanced against past 50 years, and rates for the value to poor people of risk of not having them liver cancer, the type of cheaper fruit. . . . We cannot cancer laboratory mice get be complacent about real when they ingest massive threats, but we must redoses of chemicals, have not increased. member that to be alive is to be at risk." I have no argument with conducting an investigaBut the excellent report card on food safety and cantion into the real presence of pesticide residues. Stacer rates is not enough in today's climate. People are tistically valid testing shows pesticide levels are subbeing led to believe that they must have zero risk. Instantially less than U.S. and state regulatory levels, deed, the Delaney clause of the federal Food, Drug & which have built-in tolerances based on thousandCosmetic Act, when interpreted literally, implies that fold safety factors. any pesticide can be banned from certain uses if it has Also, many foods naturally contain carcinogenic been linked to tumors in test animals, no matter how chemicals. The most potent carcinogen known, aflamarginal the risk may be to consumers. toxin B-l, is found in peanuts and corn. Bruce N. Such a policy prohibits many applications of benefiAmes, professor of biochemistry at the University of cial products used to control pests. Indeed, 4500 specifCalifornia, Berkeley, has concluded that the amount ic applications are threatened, 1000 of them termed "esof natural carcinogens people consume each day is sential" to agriculture. The policy also threatens the 10,000 times greater than their consumption of synthree-pronged goal of those of us who work in the area thetic pesticides. of food production: to provide good, wholesome food, Ames stresses that the human body is easily capable in adequate quantity, at a reasonable cost. of warding off low doses of toxins and that it does not That goal is like a milking stool, to use a farm analodistinguish between natural and synthetic toxins. I'm gy. If one of the three legs is longer, or if two of the not saying we should walk off and leave the pesticide legs are shorter, then someone is going to slide off. If issue alone. We should continue to strive for improvegrowers lose their ability to ward off pests or fertilize ments in our products. I am saying it is time for pertheir plants, they will supply less food and prices will spective. be unreasonably high. It's the consumer who will be Rather than debate the risk of pesticides versus the sliding off the stool if the "wholesomeness" leg bebenefit of pesticides, we should consider the risk of comes needlessly long. 38
January 7, 1991 C&EN
pesticides versus the risk of not having them. It is becoming clearer and better documented that toxins produced by the pest itself or produced by a plant to defend itself against a pest are equally or more toxic than the pesticide residues that were put on to protect the crop in the first place. Instead of debating with insufficient data as to what that risk is, I would challenge those with resource capabilities and those with persuasive capabilities to see to it that data are generated to learn what risk people are subjected to when they are asked to eat fruits and vegetables that have been invaded with disease or contaminated with insect parts or weed seeds. If they are truly objective in their desire to achieve the three-pronged goal—wholesome food, adequate quantities, at a reasonable price—they would be willing to use the same rigorous types of assessment that are being used on pesticides. If we look at medical records and the scientific literature, by far the most damaging threat to the food supply has been bacterial contamination. For every single case of documented, scientifically based pesticide contamination, there are literally 100,000 more cases of real problems from bacterial, fungal, or insect contamination. Frank E. Young, deputy assistant secretary for health/science and the environment at the Public Health Service, told the New York Times: "We are seeing a marked increase in microbial contamination and a constantly diminishing contamination by pesticides in our food supply/' As a result, he said, there has been an explosion of disease caused by foodborne microorganisms. At the same time, he observed, fear of chemicals has obscured the "low and decreasing risks associated with . . . pesticides." We are finding that natural does not always mean better. As Parnell has said, "Chemicals applied to agricultural commodities can, in fact, safeguard the public from naturally occurring health threats." From protection against such natural threats, to conserving available energy, to sheer yield, there are all sorts of reasons that pesticides need to be with us. Fifty percent of the pest control money spent in the world today is spent to control weeds. Those who would have us revert to the labor-intensive farm system that existed before the modern agricultural age ought to look at Jean-Francois Millet's painting "The Man with a Hoe" to get insight into where we've come from. In the painting, a peasant stands hunched over a shorthandled hoe, his labor behind him, an endless field ahead of him. He looks to the heavens, a portrait of total despair—more animal than man. Upon seeing the painting in 1899, Edwin Markham, the American labor poet, described the laborer as "stolid and stunned, a brother to the ox." "O masters, lords and rulers in all lands, how will the Future reckon with this Man?" Markham wrote. We have come far from those days when back-breaking, unfulfilling labor fed the world's people. Most of us in this country are now four generations removed from the days when America had an agrarian culture. Today we spend less than 12% of our income on food. Prior to 1950, consumers spent 24% of their
dollars on food. The public enjoys that fact, yet it does not know and cannot appreciate the enormous complexity of the agricultural system, from production to processing to shipment to the marketing availability of their food. And we who work toward contributing to the food supply can't expect that we are going to turn most of the public into extremely knowledgeable people about their food supply. That's just not going to happen. But what we can expect is that the innocent, unknowing public will continue to receive good, wholesome food, in adequate quantities, at a reasonable price. We can expect that organizations and people committed to crop protection will fulfill their obligation to continue to revise and enhance products, to strive for lower residues, to achieve lower toxicity, to create products that will break down more quickly so as to be more innocuous to mankind and the environment. And we can expect enlightened public policy will properly assess the risk versus benefit and risk versus risk. The weight of evidence is beginning to sink in. Last September, Environmental Protection Agency chief William K. Reilly called for a new system of pollution control based on provable risks to health. "I think it is time to start taking aim before we open fire," he said. Reilly called for a "broad and robust national dialogue" about the future of environmental regulation. He said discussions should take place "in the kitchens of American homes, in school classrooms, in the halls of Congress, the board rooms of industry, and the conference rooms of our vigorous environmental groups." I personally welcome such a dialogue. I want the American public to feel as comfortable about eating grapes and apples as they are about refueling their cars. The future of the world's food supply is too valuable to be trusted to uninformed public opinion and perceptions whipped to a panic. • Will D. Carpenter is vice president and general manager of the new products division for Monsanto Agricultural Co. His responsibilities include worldwide product research efforts for herbicide discovery; crop protection and improvement; plant sciences technology; formulation, metabolism, and residue technology; product development; licensing; and market introduction of new products. Carpenter received a B.S. degree in agronomy from Mississippi State University in 1952 and a Ph.D. degree in plant physiology from Purdue University in 1958. In 1983, Carpenter was named a society fellow of the Weed Science Society of America, the highest honor bestowed by this society. He served as chairman of the Environmental Management Committee of the Chemical Manufacturers Association from 1982 to 1984 and since 1979 as CMA's representative to the U.S. government in chemical warfare disarmament negotiations. Carpenter serves on the board of directors of the Industrial Biotechnology Association and as adviser on biotechnology to the House Science & Technology Subcommittee on Natural Resources, Agriculture, Research & Environment. In 1979, he was program chairman for the National Agricultural Chemicals Association. He is currently a member of the commercial panel of arbitrators for the American Arbitration Association. January 7, 1991 C&EN
39
News Forum
Blundering questions, weak answers lead to poor pesticide policies J. P. Myers, W. Alton Jones Foundation, and Theo Colborn, W. Alton Jones Foundation and World Wildlife Fund/Conservation Foundation
Ask blundering questions. Provide weak answers. Generate bad policies. That is the pathway along which public policies on pesticide regulation have trod. That is where they are likely to remain headed, unless the research and policy communities step back to take a clearer view of possible pesticide risks to health and the environment. The questions posed in this C&EN news forum typify the problem. We are asked to comment on whether pesticide risks to consumers have been grossly exaggerated or whether there really is "a middle ground." We could respond by providing our own interpretations of where the literature stands on those risks. Increasingly, that literature contains studies measuring contamination loads delivered in consumer goods and determining how those loads compare with dose-responses in one vertebrate or another. The literature even has a burgeoning line of clever, if not insightful, analyses demonstrating that natural carcinogens are likely to be present at higher risk levels than are pesticides. But while we could focus our comments on those narrow issues, we won't, because that would simply help perpetuate the inappropriate narrowness of the debate. The questions asked are grossly wrong in at least two areas. First, until the notion of risk is broadened beyond cancer to include other equally serious health effects, the question of pesticide safety remains an open chasm. Were we looking for conspiracy we might argue that the pesticide industry is content to let the cancer paradigm persist. It's one the industry has a remote chance of winning; at least by delay and by default. 40
January 7, 1991 C&EN
Second, public attention focuses on pesticide residues in food, ignoring side-stream exposures that take place throughout the pesticide delivery system: commuters driving through pesticide fogs, kids on treated lawns and golf courses, and horrific exposures of agricultural workers and spray-plane mechanics. Regulations currently governing the pesticide industry encourage this food focus. Meanwhile, wildlife and humans are constantly exposed to powerful chemicals capable of subtly, and not so subtly, undermining their health and their ability to function normally.
Preoccupation with cancer In the mid-1950s, when DDT and a number of other pesticides were being spread lavishly across the North American and European continents, reports of eggshell thinning, loss of reproduction, and albinism in bird species began to appear regularly in scientific journals and the popular press. These anomalies were ignored by the guardians of public health for years. A common argument was that the economic benefits from the use of pesticides and other chemicals outweighed the risks. A federal judge used this argument in June 1972, in his decision not to ban DDT even though he knew DDT and other persistent pesticides were rapidly accumulating in the environment. Only when the Environmental Protection Agency Administrator, William Ruckleshaus, was advised that DDT was a possible human carcinogen and was found in blood and fat tissues of nearly all Americans was it banned. Until Rachel Carson's book "Silent Spring," little organized criticism emerged against an industry that seemed to be making the world a better place to live. Those, including Carson, who insisted that pesticide use could lead to cancer in humans had no clear-cut reason for their arguments initially other than the fact that cancer was becoming the major health concern, and people thought pesticides could be one cause. True, cancer was induced in lab animals following the administration of some pesticides. But while there were extensive reports on wildlife mortality and reproductive problems during that period, there were no reports of cancer in the wildlife literature to support the cancer hypothesis. Wildlife impacts were occurring at dosage levels far below what it takes to induce cancer. Cancer has dominated the public health discussion ever since. It has not only dominated that discussion, it
has defined and driven the regulatory process. Perhaps we shouldn't be surprised—cancer is a very large concern to the white, 40- to 60-year-old males who write public health regulations. They die from it, their friends die from it. As we both age, we are concerned about cancer, also. But tragically, this morbid love affair with cancer has blinded that same regulatory process to issues equally as large, if not potentially far more pervasive and insidious, across a much broader and more representative fraction of the American people. Even as pesticide manufacturers carefully screen their products for overt, gross, and dire health effects, including cancers, gene toxicity, major malf ormities, and direct morbidity, they ignore impacts on immuno- v competency, neurotoxicity, and endocrine function. These, in fact, are where the wildlife indicators are blaring warning signals. Unfortunately, the current screening process designed to safeguard the American public from the health impacts of pesticides does not protect us from these more insidious threats. Many questions remain unresolved on this issue. The demonstrable impacts on wildlife, however, are so pervasive and horrific—and their extrapolation to human health and development so plausible—that the burden of proof lies clearly in the hands of those who develop, produce, and use pesticides, not those who would prevent their application. New testing procedures should be imposed that require, at a minimum, multigenerational studies that demonstrate no risk of developmental abnormalities in the offspring of dosed parents. That won't be cheap nor will the testing process be fast. But it will be necessary to determine the extent of subtle behavioral, cognitive, and sexuality impacts. And frankly, when all the health end points of pesticides are calculated into the risk equation, including cancer, we won't be surprised to learn that often the economic and social benefits simply do not compare favorably with the total health costs our society incurs. The hard evidence that opened our eyes to broader issues of pesticide health risks came from wildlife studies: first a few hints from the Channel Islands offshore of Santa Barbara, then a deluge from the Great Lakes. Both these regions share a common geographic feature: water, in one case the Los Angeles Bight, in the other the Great Lakes themselves, that had been used as a chemical dumping ground. In the mid-1970s, biologists began noting distorted sex ratios in California gulls breeding on the Channel Islands. While the initial interpretations sought vainly for arcane evolutionary principles that would lead not just to skewed sex ratios, but also to the observed female-female pairing, insightful experiments by wildlife toxicologist Michael Fry at the University of California, Davis, established definitively that the skewed sex ratios were a result of DDT's interfering in the development process. Fry showed that birds that would normally have matured as males (sex in birds is determined chromosomally) were in fact feminized. The Great Lakes impacts are even more dramatic, involving not just sex determination but a wide range of abnormalities affecting behavior, physiology, and morphology. Most important, of the 16 top predators in the
Great Lakes basin known to suffer reproductive problems and population instability, only bottom-dwelling fish and the Beluga whale at the mouth of the St. Lawrence River are experiencing cancer. An association between cancer and polyaromatic hydrocarbons (petroleum products and their derivatives, not pesticides) has been demonstrated in a number of these cases. Wildlife around the Great Lakes is devastated. In many areas, bald eagles, lake trout, common terns, black terns, and mink and otter are close to extirpation. This dramatic impact is not the result of cancer, but a suite of anomalies that are most often found in the offspring of animals exposed to pesticides, by-products of pesticide manufacture, and industrial products. The xenobiotics most often found in wildlife tissue so far include the pesticides DDT and its metabolites and analogs, chlordane and lindane congeners, dieldrin, and hexachlorobenzene; the industrial by-products, dioxin and furan congeners, octachlorostyrene, and hexachlorobenzene; and the industrial residuals, polychlorinated biphenyl (PCB) congeners. Many share structural similarities that often lead to similar effects in dosed laboratory animals that range from low birth weight, loss of weight, wasting that often leads to early death, cleft palate, immune suppression, abnormal sexual development, behavioral changes, and liver, thyroid, and kidney damage. These chemicals have not all been found unequivocally to be human carcinogens. Many do not effect changes in DNA structure. Many are not acute toxicants. They are, however, developmental toxicants with impacts that are often more subtle than gross malformities. Less acute, unfortunately, does not mean nothing to worry about. What is emerging from the wildlife literature in ways that suggest plausible extrapolations to humans is that very low dosage levels applied at sensitive points in the development process cause problems that affect the offspring's ability to function normally throughout life. Increasingly, the issue of health costs of pesticides will revolve around impacts of this nature—that is, as functional teratogens. In looking at functional teratogenicity, it is important to bear in mind that even as offspring health is clearly affected, little or no visible evidence of harm to the mother may be evident as a result of her exposure to the chemicals. To date, researchers have linked maternal exposure to toxic chemicals with adverse health effects in offspring in nine wildlife species in the Great Lakes region. In each case, exposure during the earliest stages of development has been demonstrated. Joseph L. and Sandra W. Jacobson of Wayne State University have found similar links with people, also. For example, skull circumference, gestation period, birth weight, and cognitive, motor, and behavioral competency of newborn babies are adversely affected by the mothers' lifetime consumption of contaminated Lake Michigan fish prior to pregnancy. The affected children showed measurable impacts up to the age of four (the most recently published results), and at that age some had begun to demonstrate difficulties in social behavior. It is still unclear from these studies which of the compounds in Lake Michigan are the root problem. Neither January 7, 1991 C&EN
41
News Forum The fact that wide-ranging impacts on natural popuPCBs nor pesticides in the contaminated fish can be lations occur without conspicuous effects on the indieliminated as the causative agent. viduals receiving the initial exposure is crucial to why The issue goes beyond the Great Lakes and the Chancurrent regulatory screening is inadequate. What fracnel Islands. Dose a female sparrow hawk, just prior to tion of the compounds chemically similar to commating, with dicofol and methoxyclor—DDT analogs and compounds still used widely on certain U.S. pounds known to induce cross-generational problems crops—and her male offspring develop ovarian tissue have been screened adequately? How many are in use? and grow up behaviorally subordinate and reproducWith this suite of issues enveloping the question of tively incompetent for life. One might argue that the pesticide safety, to maintain that current testing procecompounds are special or that sparrow hawks are irreldures protect the public's health becomes a posture of evant. That reasoning, however, would ignore the biowillful irresponsibility. Why haven't the regulators agchemical reality that chlorinated hydrocarbons mimic gressively pursued this line of research with DDT and vertebrate hormones when they intrude into the verteits analogs? Why aren't the endocrine effects of pestibrate development process. Sparrow hawks belong to cides more carefully screened—especially since many one set of vertebrates—birds, and humans to another— pesticides are growth hormone regulators? mammals. Although the roles that estrogens and androAlternative exposure pathways gens play in bird development are significantly different from their roles in mammalian development, the As we suggested in our introductory comments, the processes are sufficiently similar to expect that comnarrow focus on human pesticide exposure via food pounds that mimic estrogen and androgen may also represents another strange distortion within the regudistort the development of human babies. latory process. Not that this isn't important, but it has led attention away from what may be far greater risks The estrogenic action of DDT was reported as early as in the pipeline that takes pesticides from production 1950 by H. Burlington and V. F. Lindeman of Syracuse through application and ultimately to the consumer's University. They found male cockerels exposed to DDT plate, or indeed, to other endpoints in pesticide appliexperienced decreased testicular growth and a reduccation. Industry and government agencies continue to tion in development of secondary sex characteristics. William H. Bulger and negotiate "safe" tolerance David Kupfer of Worcester levels for pesticides in food, Foundation for Experimeni g n o r i n g complaints of tal Biology in Shrewsbury, those caught in the pipeline Until the notion of risk is broadened Mass., noted in 1983 that or hit by the sidestreams of although many compesticide use. The May/ beyond cancer to include other equally pounds are only weak esJune edition of the EPA serious health effects, the question of trogens, the frequency Journal devoted most of its with which they are enspace to assuring the public pesticide safety remains an open chasm countered might result in that all is well—its food is doses sufficient to elicit regenerally free of pesticide sponses. They also pointed residues. The journal did out that "estrogenic action is not restricted to one strucnot discuss exposure and resultant risk from municipal tural class of compounds." and countrywide application across the nation of knockdown pesticides, such as malathion and the synthetic In the meantime, how many other species of vertepyrethroids, for mosquitos. These are often applied brates, including humans, have experienced the same twice a week during summer months when windows demasculinization, perhaps not as dramatically as are open and when people spend a considerable growing ovaries, but in more subtle ways? amount of time outdoors. Nor did the journal consider Pesticides in use have been selected for their extreme the effects of drift to individuals living in close proximeffectiveness at low doses. Researchers now realize that ity to agricultural areas dependent on monocultures, a simple shift in the endocrine budget of vertebrates such as fruit, berries, corn, wheat, and soybeans. during early stages of development can release a casTolerance levels established by EPA under the Food, cade of biochemical changes that can ultimately lead to Drug & Cosmetic Act for pesticides in food do not take effects not manifested until adulthood. Brain imprintinto consideration side-stream exposure. Cumulative ing with sex hormones during the earliest stages of deexposure, encountered temporally and spatially, in velopment is extremely important for an animal to many instances might be far more hazardous to human reach sexual maturity. Timing, in this case, is an imporhealth than food residues. Estimating exposure to pestitant aspect of toxicity. Minor dosages at sensitive times cides via the food pathway using market basket studies in the development process can have major effects, is a difficult task. Estimating side-stream exposure to while massive dosages later may have none. agriculture and nonagricultural applications is even The endocrine and immune systems are closely conmore daunting. For analogy, look to the difficult chalnected. Recent studies indicate that some widely used lenge of testing for latent health effects on children pesticides, such as metribuzin and aldicarb, have imfrom second-hand cigarette smoke. munotoxic effects as well—more evidence that both enDrift from aerial spraying over vast land areas puts docrine and immune effects should be considered many at risk to both high concentrations of these matewhen estimating risk from pesticide exposure. 42
January 7, 1991 C&EN
rials and also to low-dose, continual exposure. In some instances, the pesticide is applied directly to residential areas, such as spraying for medfly control in Los Angeles. Similarly, misting twice a week in residential areas with mosquito insecticides has become standard practice in many areas. Individuals driving through agricultural areas risk repeated, incidental exposure to aerial applications of a vast number of chemicals. Commuters in these areas are especially at risk since most applications take place in the early morning or evening when wind is at a minimum. Vast expanses of U.S. forest and rangeland are sprayed with herbicides to discourage growth of what is considered to be undesirable understory. Applications along roads for weed control put hikers, joggers, pets, and children at increased risk as they use these thoroughfares. Many homeowners carelessly overapply chemicals to their yards, disregarding the labels on the containers and their neighbors' concern. Apartment dwellers are often exposed to monthly pesticide eradication applications whether they request it or not. Golf courses are deluged with sprays. Schools are treated chemically for insects. The list goes on endlessly: more compounds, and more powerful compounds, more pathways for exposure. And each of these compounds is usually met with inadequate review. Reviews are often based not on direct tests for functional teratogenicity, but instead on analogies with earlier compounds, which themselves were put through minimal scrutiny at best. Yet even as the lists of compounds in use grow and as this growth outruns our understanding of what the compounds' true effects might be, new medical problems arise that may also have their origins in pesticide exposure. Chemical sensitivity, a condition driven by exposure to chemical compounds that are foreign to living organisms, is an increasing problem in the U.S. It has been described as a condition with multiple adverse health end points, the result of reaching a threshold from exposure to a number of chemicals. What role have pesticides played in triggering an onset of chemical sensitivity? The increasing number of individuals experiencing chemical sensitivity perhaps in itself is an indication of the inadequacy of pesticide regulation.
Bad policy The current regulatory system is the result of a classic interplay between unequal warriors in the halls of Congress: an industry touting flashy products of great potential benefit and armed with lobbying resources commensurate with the dollars at stake versus an environmental community equipped with strong knowledge of the risk issues on which the science has advanced but lobbying resources of tragically smaller proportions. Science enters on both sides: demonstrating benefits, proving risks, and, less constructively, ridiculing uncertain concerns, no matter how plausible. Congress could not ignore the plus sides of pesticides for agriculture and health (through disease control), nor could it ignore the demonstrable risks of unregulated pesticide use. The less advanced the science of risk, however, the more compelling the lure of economic benefit. As a result, Congress left gaping holes in the
application of existing laws through which the industry has shoveled its products, good and bad. Although strong regulatory mechanisms have resulted to deal with potential carcinogenicities, the crossgenerational, functional teratogenic impacts remain poorly controlled, if at all. The law does allow EPA to require additional tests as science indicates. But the sheer volume of registrations overwhelms EPA's resources. The backlog of pesticides awaiting ^registration is a mountainous deterrent to effective regulation—on the order of 50,000 compounds, including at least 1100 active ingredients. This backlog—-coupled with EPA discretion on testing for functional teratogenicity, along with, ironically, a fine levied against EPA if it fails to process the reregistration application in a timely fashion—means far too few compounds are ever tested for teratogenic impacts. The pesticide industry functions under the "what we don't know won't hurt us" strategy that is abetted by these regulatory weaknesses. Shrewd lobbying by industry allows this to thrive. Until regulators require testing beyond gross and conspicuous birth malformities; until they extend acute lethal-dose measurements to longterm exposure studies that include neurotoxicity, immunotoxicity, and adverse endocrine effects; until they follow high-dose maternal toxicity studies through three and four generations; until they include low- to moderate-dose exposure studies when screening chemicals for release in the environment; until regulators assess the cumulative effect of multiple exposure; until all these tests are included in the regulatory process and their results are used to guide application of pesticides in the environment—until all of the above, grave threats to public health will remain in place. • /. P. Myers is director of the W. Alton Jones Foundation, which supports work to protect the environment and to reduce the likelihood of nuclear warfare. He was trained in biology and psychology at Reed College, and in zoology at the University of California, Berkeley, where he received a doctorate in 1979. His research in the behavioral ecology of migratory birds and in wetland conservation has resulted in more than 50 scientific articles. He has held research positions at the Bodega Marine Laboratory of the University of California, Davis, and the Academy of Natural Sciences of Philadelphia, where he was associate curator of systematics until 1987. From 1987 to 1990, he was senior vice president of science at the National Audubon Society, with responsibilities for research in wildlife populations, ecosystem management, global warming, energy and the environment, and impacts of biotechnology, among others. He is on the board of directors of the National Audubon Society. Theo Colborn is a fellow at the W. Alton Jones Foundation and a senior fellow at the World Wildlife Fund/Conservation Foundation. Colborn earned an M.A. degree in freshwater ecology at Western State College of Colorado and a Ph.D. in zoology at the University of Wisconsin, Madison, in 1985. She then worked for two years as a Congressional fellow and science analyst at Congress' Office of Technology Assessment. She has done research in Great Lakes wildlife toxicology and ecology, the effects of Great Lakes' toxic chemicals on human health, and aquatic insect biomonitoring for cadmium. January 7, 1991 C&EN
43
News Forum
Douglas D. Campt, James V. Roelofs, and Jeanne Richards, Environmental Protection Agency
Public awareness and concern about pesticides is at an all-time high. In the past few years, the nation has witnessed two instances involving widespread public worry—if not panic—over the possible dangers of pesticide residues in food. In the winter of 1983-84, alarm about residues of ethylene dibromide (EDB) in grain led many states to pull cake mixes and other packaged foods from store shelves. Similarly, in the early spring of 1989, panic over Alar (daminozide) led school lunch programs to dump their apples and citizens across the nation to avoid apple products for months. The level of concern has, at least for the moment, died down—but polls indicate that most Americans still worry about the dangers of pesticide residues in food. Meanwhile, public concerns have intensified over other types of pesticide exposures as well—from exposures in homes and offices or on lawns, parks, and golf courses to worker exposures to pesticides in agriculture and other occupations. Threats to wildlife and contamination of ground and surface waters are also growing concerns. These concerns have prompted legislative and regulatory activity across the nation. Several states have passed aggressive groundwater protection measures in reaction to the discovery that pesticides, usually at low levels, have leached into drinking water supplies. A number of states and localities have initiated programs requiring pesticide applicators to notify neighbors before spraying in buildings or on lawns. Last year, Californians placed on their ballot an initiative that would have established the most stringent pesticide measure yet seen in this country—a measure that could have changed the way farming is done in the state that produces nearly a quarter of the nation's food. Although defeated, "Big Green" garnered considerable support and became a major issue in the state gubernatorial race. Meanwhile, at the federal level, the recently
Pesticide regulation is sound, protective, and steadily improving
44
January 7, 1991 C&EN
passed 1990 Farm Bill will create a variety of programs aimed at reducing the effects of pesticides on the environment and promoting the use of integrated pest management and sustainable agriculture, which emphasize minimizing pesticide applications. Is all this activity a misdirected overreaction to risks that—compared with other things in life—are relatively small? Have the risks of pesticides been grossly exaggerated and public fears grown out of proportion to the actual situation? This is not a simple question. What we can say with assurance is that the Environmental Protection Agency does not believe our nation faces an imminent public health crisis because of our use of pesticides. For example, the American food supply, which is regulated by the U.S. Department of Agriculture and the Food & Drug Administration as well as by EPA's pesticide regulations, is undoubtedly among the safest and most regulated in the world. While EDB and Alar were both removed from the market because EPA believed their use on food posed unacceptable lifetime cancer risks, it was not a realistic portrayal of those risks to frighten people into believing they or their children were likely to get cancer from eating another muffin or apple. It is also important to take into account the major benefits pesticides offer our society. Pesticides contribute significantly to the efficient production of food and fiber, to improved public health through the control of disease-carrying pests, and to the overall quality of life by reducing aggravations caused by a wide variety of pests. Pesticides also perform a myriad of little known functions. For example, pesticides stop algae from
believe the system is sound, protective, and steadily improving. Clearly, however, not everyone agrees with us. A large number of people seem to have serious misgivings about pesticides—and to lack confidence that the federal regulatory system is adequately addressing them. In spite of some clear-cut incidents of exaggerated fears, this widespread concern can hardly be characterized as simply overreaction. Perhaps the question we should be asking, then, is not whether pesticide risks have been exaggerated, but why pesticides have become so alarming to so many people. First, it is important to acknowledge that pesticides are not harmless—and that an EPA registration (a license to growing in jet fuel, kill market) is meant to ensure only that the chemical should germs in s w i m m i n g pose no unreasonable risks if used according to label dipools, disinfect food rections. Unfortunately, it is all too easy to find instancpreparation areas, and es in which people have seriously underestimated pestisterilize surgical instrucide risks. Accidents, carelessness, failure to follow the ments and medical devices. label, and other factors lead to several thousand cases of reported pesticide poisonings every year. At the same time, it is Pesticides can pose a variety of potential hazards. quite true that pestiCertain common insecticides, for example, are known cides can pose real risks to people and the environto affect the human nervous system, causing acute ment. It would be difficult to argue, for example, that a symptoms ranging from mild—headaches, nausea, dizfarm worker who has daily, direct contact with pestiziness—to seizures and death in the most extreme cases cides has no basis for concern. The simple facts that pesticides are designed to be toxic to living organisms, of exposure. and that they are used at the rate of over a billion Epidemiological studies have linked a small number pounds a year in our country, demand that we scrutiof pesticides to cancer or reproductive damage in peonize their risks and closely regulate them. ple with significant exposures, typically in the workplace. Where evidence has demonstrated a human efThis is precisely why EPA has developed a sophistifect—such as with arsenic and l,2-dibromo-3-chlorocated process to evaluate pesticides and manage them propane (DBCP)—EPA has banned or severely in such a way as to prevent unreasonable risks to either restricted these compounds. However, more than 60 humans or the environment. Before letting any new pesticide on the market, EPA evaluates an enormous registered pesticides are thought to be carcinogenic amount of scientific data. We are concerned about dibased on animal studies, and others have been linked etary risk, farm worker exposure, ecological risks (such to reproductive, developmental, or neurotoxic effects as toxicity to fish and in laboratory animals. Sevbirds), and the potential eral pesticides have been to leach into groundwatargeted for regulatory acter. We require data on tion because of their harmIt would be difficult to argue acute (short-term) and ful effects on birds and othchronic (long-term) toxicthat a farm worker who has daily, er wildlife or on sensitive ity, reproductive effects, ecosystems. direct contact with pesticides birth defects, physical Thus, it is only reasonable and chemical properties, that people should have a has no basis for concern environmental fate, and healthy respect for the very residue chemistry. When real risks involved in using a chemical is identified as pesticides. Indeed, in the potentially risky, we may ask for a number of addition1987 EPA report "Unfinished Business," which comal studies. pared different environmental risks, top agency officials ranked pesticides as a relatively high risk for adSimilarly, when new concerns arise about a pesticide verse health effects to both people and nontarget wildthat is already in use, EPA subjects that pesticide to an life. In a recent update of this report, members of the intensive reevaluation of both risks and benefits. Again, we look closely at the gamut of health and enviagency's external Science Advisory Board, in the report ronmental risks, we consider the availability of alterna"Reducing Risk: Setting Priorities and Strategies for Entive methods for controlling the pest, and we evaluate vironmental Protection," also ranked pesticides as a relthe effectiveness of risk-reducing application methods. atively high health concern for people with occupaUltimately, if we judge the pesticide's risks to be unactional exposures and at least potentially high for conceptably high, we take it off the market. sumers, owing to the large size of the exposed EPA would be the first to admit that our pesticide population. regulatory program is not perfect. But on the whole, we It is readily apparent, however, that the concern over January 7, 1901 C&EN
45
News Forum EPA announces a pesticide regulatory decision, critics point to the uncertainties inherent in our decisionmaking and charge that our approach is not sufficiently protective. EPA's response is that we try to balance Inadequate data the risks and benefits according to the best information available and use our best judgment to make a reasonOne of the central issues for pesticide regulation is able decision. that many of these chemicals—registered years or even decades ago—lack studies now considered necessary to Finally, at least some of the current anxiety about meet current scientific standards. Under 1988 amendpesticides must be attributed to the fact that, during the ments to the basic pesticide law (the Federal InsectiAlar incident in particular, the public lost confidence cide, Fungicide & Rodenticide Act), EPA is requiring that EPA could act quickly and decisively to remove a manufacturers to submit studies to fill in the missing or risky pesticide from the market. Part of the delay stems inadequate data at an unprecedented rate, in order to from provisions in the federal pesticide law that re"reregister" nearly 25,000 pesticide products based on quire EPA to analyze thoroughly both a pesticide's risks about 600 different active ingredients. The target date and its benefits—a process that takes months or years— for reregistration decisions on these chemicals and before moving to cancel or suspend the chemical when products is 1997. potential new risks are discovered. This means there can be a substantial gap between the time a potential The insufficiency of the databases does not mean we risk problem is raised and the time EPA can act on it. know nothing about these pesticides or have reason to be alarmed that their use continues. For many of these EPA has long urged Congress to amend the law conchemicals, particularly those used on food crops, EPA cerning the cancellation/suspension process. Last year, does have fairly substantial data. Moreover, as new studpartly in response to the Alar incident, our case was ies come in, if new risks are revealed, the agency has the taken up by President Bush in his Food Safety Plan, regulatory mechanisms in place to manage and reduce which proposes a package of legislative changes, inthose risks. But the situation clearly demands our concluding a streamlining of suspension and cancellation. tinuing attention. The presence of scientific data gaps for Although Congress has not yet acted on this proposal, so many currently marketEPA is optimistic that it ed pesticides is an open inwmammm——^ma——mm will, given the level of pubvitation for anxiety about lic concern about food saferisks as yet undiscovered. ty. Concern over pesticides goes beyond If we divert our attention Completing the reregiswhat we know about their risks, and much from the specific instances tration program for older in w h i c h pesticide risks pesticides should thus go a of the concern seems based on what we have clearly been oversimlong way toward addressplified and exaggerated, ing some of the confusion don't know about pesticides and look closely at these and concern about pestifour issues—real risks, data cide risks. However, we gaps, scientific uncertainwill still be confronted ties, and lengthy regulatory procedures—we must acwith a set of uncertainties that are not likely to be reknowledge that there are legitimate causes, not for solved soon, if ever. These involve the more subtle and alarm, but for rational concern. These issues, in fact, are long-term impacts of pesticides. a central focus of EPA's continuing efforts to improve There is much we do not know, for example, about our pesticide regulatory program. We may never rehow long-term, low-level exposure to toxic substancsolve all the uncertainties, nor will we ever satisfy eves affects human cells and the disease process. We eryone with our regulatory decisions. But there are have no way to account accurately for people's cumusteps we can take to increase our understanding of peslative exposure to chemicals from different sources, ticide risks, to reduce those risks, and to deal with the or for the fact that these chemicals may react togethremaining uncertainties in a planner that protects the er, sometimes causing synergistic effects. We are only public's interests and earns its confidence. beginning to understand how pesticides affect complex ecosystems. Improving the science To account for these uncertainties, EPA takes a cauOne of EPA's top priorities is improving the science tious, protective approach to risk assessment and reguthat supports our pesticide regulatory decisions. In the latory decision-making. We use risk models that are coming years, we will be devoting extensive resources based on "conservative" assumptions designed to reto our reregistration program for old pesticides. As we flect a reasonable "worst-case" scenario and, therefore, receive new studies on these chemicals, we will be likely to err on the side of safety and public health proscrutinizing health and environmental effects only retection. We continually update our risk-assessment cently recognized as concerns, as well as revisiting canmethods as scientific knowledge advances, and we subcer risks and established tolerances, or residue limits, ject our assessments to rigorous internal and external for pesticides on food. At the same time, we will be peer review. Indeed, our risk assessments are often critworking to improve our information on pesticide expoicized as being overprotective, at the expense of ecosures in the workplace, in homes, and through the diet. nomic concerns. Almost invariably, however, when
pesticides goes beyond what we know about their risks. Much of the concern seems based on what we don't know about pesticides.
46
January 7, 1991 C&EN
world. And we continue to fund the National Pesticide Telecommunications Network, a toll-free telephone service available to the public, which provides information on specific pesticides; this service gets over 35,000 Reducing risks calls a year. There is no question that pesticides and the policies A second priority for our pesticide program is reducfor regulating them will ing risks both through regucontinue to be a lively toplatory actions dealing with ic of public debate for some specific, identified hazards, years to come. The debate and by encouraging the deWe are developing a "safer chemicals no doubt will become heatvelopment of low-risk pest ed at times, and this should control methods. The phipolicy to expedite the registration of new not be surprising. Pestilosophy of "pollution precides have become controproducts that could replace some of the vention" will increasingly versial in our society preshape our pesticide policies riskier pesticides currently in use cisely because the concerns and programs. EPA will they raise involve unrecontinue to support resolved questions concernsearch and education on ing both science and social values. What we have ahead low-pesticide methods and to encourage the developis an opportunity to make this debate one that is intelment of biological and other less toxic products through ligent and informative, and one that serves to strengthincentives built into the pesticide registration program. en the public trust in the government's commitment to We are developing a "safer chemicals" policy to expedite protecting health and the environment. • the registration of new products that could replace some of the riskier pesticides currently in use. We will be taking major steps to reduce pesticide exposures among farm workers and implementing preventive approaches Douglas D. Campt has worked with the Office of Pesticide to protect endangered species and groundwater. If we Programs at the Environmental Protection Agency since 1970, succeed in getting the legislative changes outlined in the when he joined the registration division, which reviews and President's Food Safety Plan, we will have clear authorregisters new pesticide products. After serving as associate diity to subject all pesticides used on food to a consistent rector for registration and director of the registration division, "negligible-risk" standard—the approach recommended Campt was appointed director of the Office of Pesticide Proby the National Academy of Sciences as a sound and grams in 1986. Before joining EPA, Campt served as a plant practical way to minimize dietary cancer risks. quarantine inspector for the U.S. Department of Agriculture's And as we gain an understanding of both hazards and exposures, we will use this knowledge to increase the accuracy of our risk assessments.
Risk communication One of EPA's long-term goals is to educate, inform, and involve the public in our regulatory process. EPA has learned by experiences such as those with EDB and Alar that the public's perception of risks is not simply a matter of misunderstanding scientific information. The public, the news media, and for that matter, nearly everyone who is not directly involved with pesticide regulation are generally mystified by EPA's regulatory process. If EPA expects people to have confidence that their interests are being protected, we clearly need to do a better job of explaining our system and encouraging informed participation in the process. This goal is increasingly reflected in the pesticide program's activities. For example, we have made a point of speaking at journalists' conventions in an effort to give the news media a better background on our program so they can report more accurately on specific regulatory actions. We interact frequently with the states, who shoulder much of the burden of risk communication. We are reaching out to the farming community as well, notifying grower groups of impending regulatory actions and encouraging them to track their use of certain pesticides and to contribute their expertise to our decision-making process. We have initiated a program to train all pesticide program staff in risk communication, so that we will all have a better sense of the concerns and information needs of the outside
plant quarantine division and worked in USDA's pesticide regulation division. Campt is a charter member of the Senior Executive Service of the federal government. His recent awards for outstanding public service include EPA's silver medal for superior service and the Administrator's award for equal opportunity in recognition of his major contributions to the goals of EPA's affirmative action program. Campt was educated at North Carolina Central University and Howard University, where he earned a B.S. in biology. James V. Roelofs has been a member of the policy and special projects staff of the Office of Pesticide Programs at EPA since 1979. This office is responsible for preparation of Congressional correspondence, speeches, testimony, and a wide variety of policy and liaison functions. At various times, Roelofs has also served as a special review project manager and as a special assistant to EPA's assistant administrator for pesticides and toxic substances. Roelofs recently rejoined the policy and special projects staff as head of the section responsible for special reviews and groundwater policy. He received an M.A. in English from Yale University. Jeanne Richards is a member of the policy and special projects staff of the Office of Pesticide Programs, which she joined in 1989. Before this, she worked as a reporter for the Environmental Health Bulletin, a publication of the Public Health Foundation; as a research assistant for the National Academy of Sciences' Board on Environmental Studies & Toxicology; and as a research associate for the Environmental Policy Institute's Agricultural Resources Project. She earned a B.A. in English at Duke University. January 7, 1991 C&EN
47
News Forum
Rebuttals Bruce N. Ames and Lois Swirsky Gold
Natural plant pesticides pose greater risks than synthetic ones A wide range of views on scientific and regulatory issues related to pesticide residues and cancer is presented in this issue. We discuss here four main points in response to the commentaries of the other authors: • The necessity of comparing results for natural chemicals to results for synthetic pesticide residues when attempting to assess hazards, and of using the same standards for both. • The importance of distinguishing effects in rodent studies at high doses from hypothetical effects at the very low exposure levels of pesticide residues. • The importance of adding measurements of mitogenesis to rodent tests, so as to better understand the mechanisms involved in positive carcinogen bioassays, rather than using available measurements of toxicity that do not address mechanisms directly. Regulatory efforts should, therefore, take mechanisms into account. • The necessity in disease-prevention strategies of prioritizing and concentrating on important factors. Synthetic pesticides have been described by a few authors as having unique properties that confer a particularly hazardous status upon them. Yet the chemicals that plants produce to defend themselves—that is, nature's pesticides—have these same properties as well. Plants contain natural toxins that are harmful to insects, birds, and mammals and toxic to the major organ systems of their predators. The vast proportion of pesticides in the diet are natural, as we have discussed at length. In the following quotes, if the reader replaces the words "pesticides" or "synthetic pesticides" with the words "natural pesticides" the statements are equally true. James E. Huff and Joseph K. Haseman describe synthetic pesticides as "unique in that they are specifically designed to kill living organisms." Also, they say: "Because pesticides are designed and selected for their biologic—that is, toxicologic—activity, exposure and toxicity to nontarget species are inevitable and remain significant potential problems." And finally, they say: "Pesticides have been shown capable of disrupting virtually every major organ system." Because these statements apply to nature's pesticides, one can't just assume that every synthetic pesticide is a potential time bomb, while every natural chemical is likely to be harmless. We have pointed out that the extensive, general, and inducible defenses of humans (and rats) do not distinguish between synthetic and natural chemicals. Therefore, one would expect, and indeed one finds, in animal cancer tests a similar rate of positive results (about half) for natural as for synthetic chemicals. Huff and Haseman have shown a similar rate for synthetic pesticides tested by the National Toxicology Program (NTP). In both weight and number, the vast bulk of chemicals 48
January 7, 1991 C&EN
that humans consume are natural. For example, on average each day in our food we consume about 1500 mg of natural plant pesticides and about 2000 mg of chemicals produced in the cooking process, compared with 0.09 mg of synthetic pesticide residues. The exposure consumers receive from pesticide residues is tiny compared with the doses used in rodent studies. J. P. Myers and Theo Colborn, as well as Huff and Haseman, point out the large volume of synthetic pesticides used in the U.S. Yet the amount applied to crops is not relevant to the quantity that actually finds its way to the consumer as pesticide residues in foods. These residues are usually hundreds of thousands times lower than the maximum tolerated doses used in rodent tests for carcinogenicity. Huff and Haseman say that bioassay doses are "for the most part above the environmental levels to which people are exposed." This is quite an understatement for pesticide residues. We agree that some human exposures are at high doses. Those exposures in the workplace or from chronic use of medicinal drugs are examples. For these cases, it is prudent to assume that at exposures close to the near-toxic doses in rodent tests, a rodent carcinogen may well be a carcinogen in humans. For example, on a milligram per kilogram per day basis, the high worker exposures to ethylene dibromide (EDB) were about equal to the tumorigenic doses in rodent studies, while average consumer intake of EDB from grain residues was hundreds of thousands of times lower. Epidemiology is a very blunt instrument for measuring small risks. The epidemiological studies discussed by the other authors that suggest possible associations between synthetic pesticide exposure and human cancer all involve occupational exposures. Even then, the results are often inconclusive.
Current cancer policy is neither scientifically sound nor useful, confusing regulators and the public as to what is important All chemicals are "toxic chemicals" at some dose. However, the fact that half the chemicals tested in rodents are carcinogens at high doses indicates that understanding the mechanisms of carcinogenesis is critical to predicting human risk at low doses. Chemicals differ in mechanisms of carcinogenesis, and information on both mutagenicity and mitogenicity is necessary to determine which of them might have significant effects at low doses. Rodent bioassays provide virtually no information about low doses because they are conducted at the maximum tolerated dose and at half the maximum tolerated dose. Both doses are high and relatively close to one another in comparison with low-dose human exposures. It is a rare chemical that is tested across a range of doses. We have discussed at length the importance of
mitogenesis in increasing the probability of mutation (and cancer), and the fact that at near-toxic bioassay doses, many chemicals can induce mitogenesis through cell killing and cell replacement. Analyses of nonneoplastic lesions observed at the end of two-year bioassays that do not measure mitogenesis are inadequate to assess cancer mechanisms. Huff and Haseman attempt to address the question of whether positive results of bioassays at high doses may be unique to high doses. They examine synthetic pesticides tested by NTP, and cite the 1988 article by Hoel and coworkers as evidence against a role for toxicity in carcinogenesis [Carcinogenesis, 9, 2045 (1988)]. Our view is that it is not toxicity that is important, but chronic mitogenesis in cells that are not discarded. In addition, mitogenesis can occur without toxicity. The data discussed in Hoel's article do not include any measurements of mitogenesis. Moreover, the measures of toxicity used in his analysis are crude. For example, an increase in tumors only at the maximum tolerated dose but not at half that dose is one of his measures of toxicity. This ignores the fact that half the maximum tolerated dose is itself a high dose, one that might also produce toxic effects, such as mitogenesis. Hoel's second measure of toxicity is nonneoplastic lesions in the target organ. However, assessing which le : sions may actually represent a toxic response is largely subjective, particularly when it is done with only routine histopathology at the end of an experiment. Another critical complexity is that mutation through mitogenesis is not of interest in cells that are discarded from, for example, epithelial tissues or killed (from apoptosis). In contrast, more pertinent, recent studies of mitogenesis at NTP by Michael L. Cunningham and coworkers compare two pairs of mutagenic isomers, only one of which is carcinogenic, and show that mitogenesis is increased in each case only by the carcinogenic isomer. Such studies of mechanism need to be incorporated into prechronic studies and rodent bioassays so maximum scientific information can be obtained from these tests. Risk assessment models that include the effects of mitogenesis make more biological sense than those that do not. Strategies to prevent disease should prioritize and concentrate on important factors. Several authors have described the prevailing "prudent" policy with respect to low-dose extrapolation. Current policy attempts to protect the public from hypothetical, worst-case risks of one cancer death in a million from industrial pollutants and pesticide residues at levels that are about 380,000 times below the maximum tolerated doses that induce tumors in rodent tests, while ignoring the natural world. This is done at enormous cost. We believe such policy is neither scientifically sound nor useful. It confuses regulators and the public as to what is important and diverts resources from more important risks. The policy of testing primarily synthetic chemicals one by one for their "carcinogenic potential," when about half of both natural and synthetic chemicals test positive and when the mechanism of carcinogenesis for many chemicals may be unique to high doses, is not likely to provide much information that will be useful in
cancer prevention. Instead, more emphasis is needed on the mechanisms of carcinogenesis, and on other factors, such as dietary imbalances, hormones, and viruses, as causes of human cancer. Other than giving up smoking, probably the most important thing Americans can do to lower cancer rates is to eat more fruits and vegetables. Pesticides, by cutting costs of such foods, encourage this. Myers and Colborn are very uncritical about the evidence on cause and effect when synthetic pesticides are involved. Tiny doses of synthetic pesticide residues in foods are particularly unlikely to be a potential hazard for human birth defects and other diseases because of threshold effects. From their discussion of DDT, one would never know that it saved millions of lives and replaced lead arsenate. We remain unpersuaded that current pesticide use is having a devastating effect on the environment. We agree with Myers and Colborn that pollution control is desirable, but cancer risks at parts per billion levels shouldn't be a surrogate for environmental concerns. D
Joseph K. Haseman and James E. Huff
Arguments that discredit animal studies lack scientific support The opinions presented in this C&EN forum represent extremely divergent views concerning the possible health effects of pesticide exposures. On the one hand, J. P. Myers and Theo Colborn emphasize that pesticide exposures have been linked to major health effects other than cancer and conclude that a serious problem exists, particularly for those receiving side-stream exposures from the pesticide delivery system. In contrast, Will D. Carpenter believes that the traces of pesticides present in fruits and vegetables pose little or no health problem and that public fears regarding this matter are irrational and illogical. Bruce N. Ames and Lois Swirsky Gold claim that synthetic pesticides pose little health risk and discount the scientific value of rodent studies that have indicated certain pesticides may cause cancer. Douglas D. Campt, James V. Roelofs, and Jeanne Richards adopt somewhat of a middle-ground position, explaining the Environmental Protection Agency's approach to the evaluation and regulation of pesticides. Our view is that exposures to certain pesticides may represent an important public health hazard. Although concerns regarding the potential risks of pesticide residues on agricultural products should not be minimized, we are more concerned about the farmers, occupationally exposed workers, pesticide applicators, weekend gardeners, and others who may be repeatedly exposed to much higher levels of pesticides and therefore are at greater risk. These populations are quite large (for example, 10 million individuals, including family members and hired workers, are engaged in farming in the U.S.), and, as we note in our article, several studies have January 7, 1991 C&EN
49
News Forum already linked these types of exposures with increased incidences of cancer. Further, in October 1990, the International Agency for Research on Cancer (IARC) working group concluded that occupational exposures in spraying and application of insecticides "are probably carcinogenic to humans/' As noted by Myers and Colborn, pesticides can also have significant behavioral, reproductive, and neurological effects in livestock, wildlife, and experimental animals at exposure concentrations well below those that produce cancer in laboratory animals. Certain of these effects have been observed in humans as well. Moreover, the World Health Organization estimates that pesticides cause from 500,000 to 1 million poisoning cases a year, and possibly 10,000 deaths. Thus, any overall evaluation of pesticide exposure and human health must take into account these varied and important adverse effects. Perhaps one reason current scientific interest seems to focus on cancer is that it is dreaded by the public to the extent that evidence of carcinogenicity tends to drive the regulatory process. Further, in contrast to other major health problems, such as heart disease, there is little evidence that continuing efforts at prevention and treatment are having much effect on cancer incidence. Although Carpenter and Ames and Gold maintain that human cancer rates are not increasing, recent studies, such as that by Devra Lee Davis of the National Research Council and coworkers, indicate otherwise [Lancet, 336, 474 (1990)]. Moreover, these evaluations have taken into account changes in diagnoses and increased access to health care, often cited as confounding factors in such analyses. The 1990 National Cancer Institute "Cancer Statistics Review" reports that overall ageadjusted cancer incidence rose 15% from 1973 to 1987
Abandoning animal studies to help evaluate a chemical's carcinogenic potential is an extreme position that runs counter to mainstream scientific thinking and that this increase does not merely reflect an increase in smoking-related lung cancer, but instead is due to a number of different site-specific neoplasms. For example, the incidence of non-Hodgkins lymphoma, a specific cancer that has been linked to exposure to herbicides and pesticides, increased more than 50% between 1973 and 1987, and the death rate associated with this cancer increased about 100% between 1949-51 and 1984-86. Even Ames and Gold admit that the incidence of this particular cancer "can be shown to be increasing." These findings, considered together with the results of epidemiological and laboratory animal studies, indicate it is likely that certain pesticides do indeed cause cancer in humans, particularly in those individuals receiving high exposures. The criticism of laboratory animal studies by Ames 50
January 7, 1991 C&EN
and Gold illustrates a troubling trend that extends beyond the immediate issue of possible human health risks associated with pesticide exposures. Recently, a barrage of articles in the lay press by a small but vocal group of individuals has sought to discredit the value of laboratory animal studies for assessing potential cancer risks to humans. Many of these criticisms are from individuals with little or no direct experience in chemical carcinogenesis. Previously, such attacks had been directed primarily toward the uncertainties associated with quantitative risk estimation—for example, extrapolating risks from high to low doses and from laboratory animals to humans. However, more recent criticism has focused on qualitative risk estimation—that is, the scientific value of the animal studies themselves. For example, the Oct. 15, 1990, issue of Business Week reports that Ames and Gold have concluded that the use of the animal studies to predict cancer risk to humans is a "bankrupt approach that should be abandoned." If this accurately reflects their views, it is unfortunate, because the abandonment of animal studies to help evaluate a chemical's carcinogenic potential is an extreme position that runs counter to the mainstream scientific thinking of experts in this area. The arguments presented by Ames and Gold to discredit animal studies (for example, that carcinogenic effects are observed only at high dose and/or are due to increased cell proliferation related to toxicity) are not new nor are they convincing; they are based primarily on speculation and lack supportive scientific data. As we indicate in our article, most carcinogenic effects observed in rodents are not limited to the high dose, nor are they strongly correlated with target organ toxicity. Further, based on what we currently know concerning mechanisms of action, most carcinogenic effects in rodents should be considered relevant to humans. Thus, most experts agree that in the absence of adequate human data, animal studies are the best method currently available for determining carcinogenic hazards to humans. If we lack human data and abandon animal studies, we have no other scientifically accepted alternative for assessing a chemical's carcinogenic potential. Although Ames and Gold are highly critical of animal studies, they use animal data and risk assessment methodology when such information serves their purposes—as, for example, when they attempt to compare the teratogenic risk associated with low levels of dioxin with the risk of alcohol in beer. Such calculations, including comparisons of relative cancer risk based on Ames' frequently-cited HERP (human exposure/rodent potency) indexes, involve the same "bankrupt" animal studies and risk extrapolations that Ames and Gold criticize in other contexts. Ames and Gold's view that synthetic pesticides pose little or no carcinogenic hazard because "dietary pesticides are 99.99% all natural" is misleading for several reasons. First, this conclusion is based on their assumption that each day we consume 1.5 g of "natural pesticides" and 0.09 mg of "synthetic pesticides." These figures ignore exposures and ingestion of synthetic pesticides from sources other than residues on food, such as the high exposures received by farmers and by workers
in the pesticide industry. Moreover, even if Ames and Gold's pesticide consumption figures are accurate, comparisons based strictly on the relative weights of pesticides consumed are of little scientific value because they do not take into account carcinogenic potency. For example, according to Ames, a major "natural pesticide" consumed by humans is caffeic acid. However, caffeic acid is a relatively weak rodent carcinogen; synthetic pesticides such as mirex, DDT, and aldrin are carcinogenic at one two-thousandth of the dose required for caffeic acid to produce carcinogenic effects. Clearly, any meaningful comparisons of natural and synthetic pesticides must take into account carcinogenic potency, as well as the amount of pesticide consumed. Further, less than one third of the "carcinogenic natural pesticides" cited by Ames are considered by IARC or the National Toxicology Program to be possible human carcinogens. Some have not even been evaluated directly in rodent studies, and their carcinogenicity is apparently inferred by Ames and Gold. Clearly, this whole issue of natural versus synthetic pesticides is fraught with major problems and should not divert attention from the very real issue of evaluating potential hazards associated with any substance used as a pesticide. In conclusion, we believe that certain pesticides, many of which have well-documented adverse reproductive, neurological, or behavioral effects, may also cause cancer in humans. Epidemiology studies have confirmed such a link, and laboratory animal studies have also identified pesticides potentially carcinogenic for humans. To protect public health, we strongly support the continued use of animal studies for evaluating the carcinogenic potential of pesticides and other chemicals. •
WfflD. Carpenter
Myers and Huff present biased views of pesticide risks The article by James E. Huff and Joseph K. Haseman and that by J. P. Myers and «Theo Colborn both fail to present balanced views of the risks and benefits of pesticide usage. Huff and Haseman are respected scientists at the National Institute of Environmental Health Sciences. Unfortunately, much of what these scientists report also reflects the distinctive biases of that program. What they report is neither wrong nor right; it merely represents opinion presented as fact. The article by Bruce N. Ames and Lois Swirsky Gold offers a much different viewpoint regarding the risks and benefits of pesticides. Pesticides have been used for more than 40 years, which means that adequate time has elapsed to observe the effects of such use. Despite Huff and Haseman's extrapolations regarding cancer, the American Cancer Society reports that, with the exception of lung cancer, cancer mortality rates in the U.S. have declined or re-
mained the same over the past 50 years—a period that coincides with the development of modern crop protection chemicals. A recently published study in the British medical journal Lancet by Devra Lee Davis of the National Research Council and coworkers states that cancer mortality rates
Pesticides have been used for more than 40 years, which means that adequate time has elapsed to observe the effects of such use are rising for persons 55 and older. Critics of the study, including renowned epidemiologist Sir Richard Doll of Oxford University, attribute these statistics to improved screening procedures; to people living longer; and to ingestion of preservatives such as mercury and arsenic, which were used in food early in the 20th century. Interestingly enough, the Lancet article also reports declining death rates from cancer among younger people—particularly in the U.S. It suggests that the declining cancer rates could be due to improved diet and better food storage, among other things. Less than 0.5% of lethal poisonings in the U.S. are due to agricultural chemicals. Since 1900, average life spans have increased from 45 years to 75 years. And what we have found is that the cost of food to American consumers has dropped from 24% to 14% of disposable income. Even with today's pesticides, pests continue to destroy 30% of crops annually in the U.S. Contrary to Huff and Haseman's opinion, crop loss would be much greater if pesticides were not used, 1 lb of soil would be lost for each pound of corn and soybeans produced, and the amount of land necessary to be plowed to feed a hungry and malnourished world population would have to increase—probably at the expense of wildlife preserves and tropical rainforests. I don't believe this is acceptable for our growing world needs. The benefits of pesticide use are not readily apparent to the average consumer. Only 53% of pesticides are used in agriculture and forestry. The rest are used for water purification, disinfectants, homes and gardens, sterilization, and medicinal purposes—particularly where bacterial infections are present. For example, diseases spread by such insects as fleas, cockroaches, and mosquitoes have been well documented over thousands of years. Diseases such as malaria and bubonic plague have not been completely eradicated, but the use of pesticides to control insects spreading these often fatal illnesses has saved millions of lives. Those who criticize herbicide usage in forestry and roadside applications don't readily understand the alternative: that young trees can't hope to compete against useless brush, or that roadside weeds and brush harbor deer, family pets, or children who are at risk from road traffic. January 7, 1991 C&EN
51
News Forum I prefer not to contemplate the cockroach in my bathroom hopping from upstairs to downstairs to crawl over food in the kitchen, spreading the risk of salmonella and other harmful bacteria. "We promise according to our hopes, and perform according to our fears, " said Francois Due de la Rochefoucauld. What we as a society must continue to do is to regulate pesticides strictly and appropriately based on risk and benefit, not fear and prejudice. Much of what Myers and Colborn have to say is based on fear and prejudice. Two pages of rebuttal comments are not sufficient to address the large number of inaccuracies and inadequacies concerning pesticide usage, effects, and registration readily apparent in their article. The pejorative, frankly exaggerated tone heightens the hyperbole they express in their article, which relies on 20-year-old myth. Myers and Colborn are primarily ignorant of the state of pesticide registration today. I disagree strongly with Myers and Colborn's logic that pesticide oversight lies in a narrow paradigm of cancer risk and includes few requirements beyond the scope of cancer. The Environmental Protection Agency requires exhaustive studies to assess a pesticide's potential for health and environmental effects for mammals, fish, birds, and other organisms. These include assessments for birth defects, reproductive effects, multigenerational effects, and neurotoxic and immunologic effects. Environmental fate and behavior in soil, plants, water, and animals are also subject to intensive investigation. Toxicology plays a large role in determining the status of a candidate active ingredient. The cost of registering a pesticide today is escalating above $40 million. Many of the anecdotal examples of wildlife impact from chemicals proffered by Myers and Colborn, by their own admission, may not relate directly or at all to
What we as a society must continue to do is to regulate pesticides strictly and appropriately based on risk and benefit, not fear and prejudice pesticide exposure. One of the extrapolated examples they use, DDT, has been banned for more than 15 years—so the system apparently is working! It is easy to use inflammatory language, as Myers and Colborn do, to attempt to scare people into believing these products are dangerous and have little benefit. It is more difficult to write a rational, logical, attributable criticism of pesticides that focuses on risk-benefit analysis. For example, Myers and Colborn say that schools are sprayed with insecticides. That is true. What they don't say is that schools are sprayed with insecticides to control insects, which spread bacteria and disease. The largest pesticide application, chlorine used in purifying water supplies, prevents typhoid and other life-threatening diseases. These uses have been thoroughly evaluated and have been determined to be safe. 52
January 7, 1991 C&EN
The current U.S. regulatory system governing the use of pesticides is the strictest in the world, and our food supply is certainly the safest in the world. I take issue with the authors' assertion that industry has huge lobbying war chests, while the environmental community does not. That is certainly not the case. While the environmental community may have strong emotions regarding risk, factual data and knowledge of risk and benefit is documented and discussed by both industry and government to evaluate pesticide use. Myers and Colborn conclude that regulatory weaknesses exist in the following areas: birth malformities, neurotoxicity, immunotoxicity and adverse endocrine effects, chronic maternal toxicity, environmental fate, multiple exposures, and more. Not only are Myers and Colborn wrong, but they have made these allegations with absolutely no knowledge of toxicology or regulatory requirements. I suggest they contact the regulatory agencies or check recent literature before writing future articles on pesticide issues. EPA regularly proposes new testing procedures now under development to provide even more comprehensive evaluations as the state of science develops. We in industry don't think the regulatory system is perfect. However, when we are critical, we also take the re^ sponsibility to improve the testing and regulatory system. The privilege of dissent bears the responsibility of offering alternatives. Many users of chemical products who need new chemical products, because some older formulations have been eliminated, find themselves limited in controlling pests crucial to growing crops. The risk we face in making the stakes so high that companies are unwilling to pursue research and development is to continue the registrations of older chemical products, even though many new products are superior to older ones. The bleak picture Myers and Colborn paint of agricultural chemical companies does not reflect the companies I am familiar with. Monsanto itself, like most firms in our industry, openly acknowledges the need for improvement in environmental areas. To that end we have made a pledge that states: • Reduce all toxic and hazardous releases and emissions, working toward an ultimate goal of zero effect. • Ensure that no Monsanto operation poses any undue risk to our employees and our communities. • Work to achieve sustainable agriculture through new technologies and practices. • Ensure groundwater safety. • Keep our plants open to our communities and involve the communities in plant operations. • Manage all corporate real estate, including plant sites, to benefit nature. • Search worldwide for technologies to reduce and eliminate waste from our operations, with the top priority being to not make waste in the first place. And I'll add one additional pledge: Monsanto will, place its name only on products that we can ensure are of premium quality—both in effectiveness and for consumers desiring a safe, wholesome, nutritious food supply and safe environment to live in. •
J. P. Myers and Theo Colborn
Immense unknowns characterize the assessment of pesticide risks When we agreed to participate in this C&EN forum, we anticipated our comments would be the exception rather than the rule, for we fully intended not to present a traditional argument. We wanted to advance the readers' perspective beyond contemporary thinking about pesticides. Now, with an opportunity to read the other contributions, we see that our expectations were correct. Although we recognize substantial differences between our comments and those of other contributors, we are pleased that they share our concerns about pesticide hazards in addition to cancer. Our hope is that, as concerns have become broader, research on pesticides and related contaminants has reached a pivotal point, and from here on, the same respected researchers who have struggled with the pesticide-cancer link for so long will now direct their energies toward quantifying the effects of pesticides and industrial contaminants on immune, endocrine, and nervous system function. In our article we pointed out that preoccupation with cancer has perhaps blinded us to other more subtle health effects of pesticides. We were thus intrigued to note that Bruce N. Ames and Lois Swirsky Gold suggest that "cancer prevention strategy" is diverting resources from "much more important risks." Despite this, Ames and Gold do not describe what they perceive the more important risks to be. The rest of their article is devoted largely to cancer, in which they make very useful comments about the relative risks of artificial versus natural sources of carcinogens. We agree with Ames and Gold concerning the carcinogenic risk of dioxin (TCDD), but want to point out that TCDD has other effects they did not mention. They failed to list the other conditions that have been linearly associated with doses of TCDD after binding to the Ah (aryl hydrocarbon) receptor. Following Ah binding, TCDD and a number of other coplanar organochlorine chemicals, including some polychlorinated biphenyls (PCBs), induce enzyme activity that is associated with wasting and immunotoxicity in all lab animals and teratogenicity, porphyria, and organ damage in rats, mice, guinea pigs, and hamsters, depending upon the species. In addition, literature is accumulating on the estrogen- and antiestrogenlike effects of TCDD, as well as its depleting effects on Vitamin A (a key morphogen in development) that may or may not be mediated via the Ah receptor. Most important, Richard E. Peterson and coworkers at the University of Wisconsin, Madison, recently found that by feeding female rats a single meal of TCDD during pregnancy (0.064, 0.16., 0.64, and 1.0 ng per kg of body weight) their male offspring were feminized (the more TCDD given, the greater the effect). Peterson points out that many of the TCDDinduced effects were not fully manifested until the rats reached adulthood.
Ames and Gold believe that we have made progress through the development of more potent, new-generation synthetic pyrethroids requiring only a few grams per acre. Although this advance may have some beneficial consequences, we must remain vigilant about hazards from pyrethroids to nontarget species for several reasons. First, pyrethroids have the capacity to bioaccumulate, and physiologically they exert an estrogenic effect. Second, only exquisitely small shifts in hormone balance are needed to turn on or off a cascade of events in developing organisms. Finally, timing of exposure during development is as critical as the presence of the hormone. Douglas D. Campt, James V. Roelofs, and Jeanne Richards of EPA begin their article with a reassuring tone conveying the message that all is well, or nearly so. They include some encouraging comments about efforts to improve the science that underpins risk assessment and about reducing existing risks directly by pushing for "pollution prevention"—that is, "low-risk pest control methods" that presumably involve reduced exposures. Unfortunately, the history of pesticide use is laced with a pattern of diminishing benefits and underestimated risks, and changing this pattern lies squarely on the shoulders of EPA. Forgive us for being skeptical. The authors state that before registering a product, EPA looks at a full array of endpoints that include acute and chronic toxicity, reproductive effects, birth defects, physical and chemical properties, environmental fate, and residue chemistry. But they acknowledge that thousands of products in current use, which involve some 600 active ingredients, have yet to be tested thoroughly. In the past pesticides have not been routinely tested for immunological, neurological, or endocrine disruption/What must be remembered is that EPA has
Under present regulations, a genuine independent review of the health effects of a pesticide never takes place before registration not released its final guidelines for neurotoxicity testing to date and is still developing a rat model for immunological screening. Admittedly, in the past five years interest in the field of immunotoxicology has nurtured research to develop a battery of tests to screen products for effects on various components of the immune system. Because of the complexity of the immune system, this has not been an easy task. However, guidelines for a tiered testing system using mice and, soon, rats will be available. Concern has been expressed that these testing procedures detect only the gross effects of immunotoxicants and that sensitive, more subtle effects will go undetected. Unfortunately, EPA is limited by law to use the information provided by the manufacturer when registering a pesticide. This information is proprietary—in January 7, 1991 C&EN
53
News Forum other words, not available to the public. The laws were amended in 1978 to make this information more available. However, the obstacles to gaining access (a Freedom of Information Act request is required), the limited amount of material that is released for review, and the restrictions governing the use of the data thwart disclosure. Under present regulations, a genuine independent review of the health effects of a product never takes place before registration. It is not surprising, therefore, that the public is uneasy about pesticide exposure. Campt and his coauthors acknowledge that much of the concern about risks from pesticides arises from what is not known, especially about human disease, impacts of cumulative exposure, synergistic effects, and effects of pesticides on complex ecosystems. These are profound concerns that will yield only to the most sophisticated and tenacious of scientific investigations. The notion is absurd that all of the products requiring reregistration will be examined sufficiently by the year 1997, as required by law. This does little to alleviate legitimate worries, even concerning the narrow issues directly mandated by narrow interpretations of the Federal Insecticide, Fungicide & Rodenticide Act, much less in these broader unknowns touched on briefly here. Although the point of Will D. Carpenter's article is that the dangers of poorly regulated chemicals are overstated, ironically, in many ways his article supports our fundamental argument that the focus on cancer and pesticides in food has diverted attention from other more important health effects. His bald denials concerning risk simply fuel the debate between the different stakeholders who purport and deny the carcinogenic risk of pesticides. As a result, the subtle health effects we may experience in future generations could be overlooked. James E. Huff and Joseph K. Haseman's fine contribution to the forum reinforces our argument that functional deficits are difficult to recognize and may take generations before they are detected. They state that too little is known to determine the "full toxicological potential" in humans from exposure to widely used agents, but that the very nature of pesticides—that is, compounds designed to kill living organisms—guarantees that impacts occur on both target and nontarget organisms. In conclusion, the net impact of the several authors' contributions to this forum should be to convey to readers the immense unknowns that characterize the field of environmental toxicology (and specifically, pesticide risk), as well as the need for skepticism about the depth and focus of current government regulation. We close by summarizing two articles that have appeared since we wrote our original article. Although the articles focus on PCBs, they help to make our point that research in the field of functional teratogenicity needs to be encouraged. In our original commentary we described the subtle differences noted in children of women who ate Lake Michigan fish prior to their pregnancy, compared with children of women who ate no Great Lakes fish. Since then, Hugh A. Tilson and coworkers of the National 54
January 7, 1991 C&EN
Institute of Environmental Health Sciences have reported that humans show transplacental neurotoxicological effects at one ten-thousandth of the PCB dose that produces the same effects in rodents. To arrive at this conclusion they used available data to calculate a reference dose following EPA's "Proposed Amendments to the Guidelines for Health Assessment of Suspect Developmental Toxicants." Another study suggesting that PCBs may have adverse neurological effects is reported by Joseph L. Jacobson and coworkers of Wayne State University. They found an association between contemporary breast milk concentrations of PCBs and "reduced activity" in exposed offspring. PCBs were first produced in the 1920s. After 65 years, we still do not fully comprehend the toxicity of these chemicals. This surely is equally true for a wide range of compounds now in use. •
Douglas D. Campt, James V. Roelofs, and Jeanne Richards
Some experts view pesticide risks with too narrow a focus This forum does an excellent job of capturing the diverse nature of the pesticide risk debate. While differing sharply in their central views, the articles all raise valid points—indeed, their recommendations often match current directions in the Environmental Protection Agency's pesticide regulatory program. On the other hand, the articles also demonstrate how easy it is to oversimplify or distort this complex debate by narrowing the focus to a single issue or set of criticisms. We find much to agree with, for example, in Bruce N. Ames and Lois Swirsky Gold's recommendations for research on the mechanisms of carcinogenesis. EPA has projects under way now to investigate biological mechanisms and metabolic and pharmacokinetic factors that affect cancer risk. We are committed to incorporating such research into our risk assessments once it has been peer reviewed and accepted by the scientific community. We also agree with Will D. Carpenter's judgment that our food supply is safe and that pesticides have been important in ensuring efficient and abundant agricultural production. Finally, we acknowledge the validity of J. P. Myers and Theo Colborn's point that food safety and cancer risks are not the only issues of concern regarding pesticides. Other types of human health effects, as well as pesticide effects on wildlife and the environment, also need to be considered in an adequate regulatory program. This said, we also have fundamental disagreements with each of these three articles. To begin, we take issue with Ames and Gold on two points. First, we disagree with the implication that naturally occurring toxins should be held up as a standard for judging pesticide risks. To use the alleged risks of natural carcinogens as an argument for dismissing syn-
thetic pesticide risks is neither a valid nor a responsible approach. Unlike naturally occurring toxins in food, pesticide use is something regulation can control. Where risk seems high, we can reduce application rates, lengthen preharvest intervals, or look for safer chemical or cultural alternatives to reduce residue levels on food. At this time, natural toxins are not subject to this type of regulatory control. EPA's job is to reduce risk through careful control of pesticides, not to excuse a risk from pesticides simply because larger risks may be posed by other sources for which EPA lacks a legislative mandate. We also disagree with Ames and Gold's views on the validity of our current cancer risk assessment procedures. We would defend our approach as both good science and good public policy. Our colleagues at the National Institute of Environmental Health Sciences have ably discussed the scientific basis for our current procedures. To this we would add that, as a matter of public policy as well as science, EPA embraces a conservative, health protective approach. We acknowledge that our risk assessment methods are likely to overestimate rather than underestimate actual cancer risks. But we would defend this approach as a reflection of social values, which are highly risk averse when it comes to health, especially regarding cancer risks from involuntary exposure. EPA welcomes scientific questioning as a way of improving our risk assessment methods. On a case-by-case basis, where the science has been rigorous, the pesticide program has, in fact, incorporated new methods and new types of information in our risk assessments. As a general rule, however, in the absence of a full understanding of the mechanisms of carcinogenesis and acceptance of new risk assessment methods by the scientific community, we have purposely chosen to take the conservative path. Carpenter's article, while arguing a view similar to that of Ames and Gold, raises another set of issues. In this case, the article contains a number of misleading statements that serve to deemphasize risk and overemr phasize the benefits of pesticides. In interpreting the Alar (daminozide) risk numbers, for example, Carpenter states that a person would have to eat 28,000 apples a day to duplicate the dose given to test animals. Such a statement sidesteps an important point—that scientists generally agree that low-level exposures may pose cancer risks and that EPA's risk estimates for Alar were extrapolated from animal studies to estimate the cancer risk people might face from actual Alar levels found in products in grocery stores. Similarly misleading is Carpenter's reference to estimates of the dramatic food production declines that would occur should all pesticides be banned. Because EPA has never contemplated such a sweeping action, this is a fairly meaningless straw man. Indeed, we doubt EPA could ban all pesticides under current laws and policies, which require the agency to weigh the benefits of pesticides against their risks in reaching regulatory decisions. Finally, Myers and Colborn are quite right to point out that noncancer and nondietary risks are important
factors in a pesticide regulatory program. But they fail to acknowledge that there is some basis for the high priority given to food safety. Protection of the food supply is a deeply held societal concern—and this is not surprising, given the number of people who may be exposed to any contamination and the fact that this exposure is involuntary. More important, Myers and Colborn imply that EPA has done nothing about other types of risks, human or ecological. This is simply untrue. As discussed in our article, the pesticide program requires a large battery of tests on human health effects, often including tests for cancer, neurotoxicity, and developmental and reproductive effects. The program also considers worker/applicator and home-owner risks and, where enough information is available, risks associated with other routes of exposure, such as drift, runoff, and groundwater contamination. We can, and frequently do, institute regulatory measures to reduce risks identified through such tests.
To use the aDeged risks of natural carcinogens as an argument for dismissing synthetic pesticide risks is neither a valid nor a responsible approach But Myers and Colborn are most misleading in their implication that EPA has entirely ignored wildlife effects. For more than a decade, the pesticide program has required multigeneration reproduction tests for birds, fish, and aquatic invertebrates. EPA has banned or restricted several pesticides because of their acute toxicity to wildlife—-including diazinon, tributyltin compounds, and compound 1080—and has proposed to cancel carbofuran. Since 1985, we have made wildlife and ecological effects an increasing focus of EPA's pesticide regulatory program; and this trend will undoubtedly continue. Evaluating pesticide risk is a complex task involving a broad spectrum of science and policy issues and social values. This forum illustrates both the controversial nature of these issues and the diverse views held by experts on pesticide risk. What is important is that we continue to talk and to work together to develop sensible policies for pesticide use that allow society to benefit from these products without jeopardizing public health and the environment. •
Reprints of this C&EN news forum are available in black and white at $10 per copy. For orders of 50 or more, subtract 3 0 % from the total order cost. On orders of $50 or less, please send check or money order with request. Send orders to: Distribution, Room 210, American Chemical Society, 1155—16th St., N.W., Washington, D.C. 20036.
January 7, 1991 C&EN
55
