Document JoK6rgB6wQov03k9Gyz1adjO

Important Advances in Oncology 1989 J. B. Lippincott Company Philadelphia Cambridge London New York Singapore St. Louis Sydney San Francisco Tokyo TOOSSZSZ BFG39718 -w -- - Edited by Vincent T. DeVita, Jr., m.d. Physician-in-Chief Benno C. Schmidt Chair in Clinical Oncology Memorial Sloan-Kettering Cancer Center New York, New York Samuel Heilman, M.D. Dean and A.N. Pritzker Professor Biological Sciences Division and Pritzker School of Medicine Vice President for the Medical Center The University of Chicago Chicago, Illinois Steven A. Rosenberg, M.D., Ph.D. Chief of Surgery National Cancer Institute Professor of Surgery Uniformed Services University of the Health Sciences School of Medicine Bethesda, Maryland 35 Contributors 25259002 Bruce N. Ames What Are the Major Carcinogens in the Etiology of Human Cancer? Environmental Pollution, Natural Carcinogens, and the Causes of Human Cancer: Six Errors 14a Introduction Arguments that industrial pollutants are present in significant amounts or contribute substantially to cancer rates share several mistaken assumptions. These are discussed in the following sections, as is the danger inherent in diverting our attention away from tobacco and dietary, hormonal, occupational, and viral carcinogens to industrial pollutants. Error 1: Cancer Rates Are Soaring Overall, US cancer death rates are staying at the same levels or decreasing, the major exception being smok ing-related cancer. A recent update from the National Cancer Institute (February 1988)' indicates that "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" (13% decrease overall, 44,000 deaths below expected; 0.1% increase in the over-85 group). The types of cancer deaths that have been decreasing during this period are primarily those of the stomach (by 75%, 37,000 deaths below expected), cervix (73%, 11,000 deaths below expected), uterus (60%, 9,000 deaths below ex pected), and rectum (65%, 13,000 deaths below ex pected). The types of cancer deaths that are increasing are primarily lung cancer (by 247%, 91,000 deaths above expected), which is caused by smoking (as is 30% of U.S. cancer) and non-Hodgkin's lymphoma (by 100%, 8,000 deaths above expected). The overall cancer mortality trends can be seen in the latest plot from the American Cancer Society (Fig. 14a-l). Clearly, changes'in survival rates and incidence rates are also relevant in interpreting those changes.1-2 Incidence rates have been increasing for some types of cancer. Doll and Peto,2 in their definitive study on cancer trends, pointed out that although incidence rates are of interest, they should not be taken in iso lation because of the substantial extent to which trends in recorded incidence rates are biased by im provements in the level of registration and diagnosis, as appears to be the case with breast cancer. Even if particular types of cancer are shown to increase or decrease, establishing a causal relation among the many changing aspects of our lives remains diffi cult. 4 There is no convincing evidence that there is a general increase in cancer related to the conditions of the modem industrial world.2-'12 Life expectancy is steadily increasing in the United States and other industrial countries; infant mortality is decreasing; and, although the statistics are not good, there is no evidence that birth defects are increasing. Thus, the conclusion is that Americans are healthier than they have ever been. 237 25253003 238 What Are the Major Carcinogens in the Etiology of Human Cancer? FIC. 14-1 Cancer death rates by site, United States, 1930-1985. AU figures are age-adjusted. Rate for the population standardized for age on the 1970 U.S. population. Sources of data: National Center for Health Statistics and Bureau of the Census, United States. Rates are for both sexes combined except breast and uterus, female population only, and prostate, male population only. (Reprinted with permission from Cancer Facts and Figures-- 1988, New York, American Cancer Society, 1988) Error 2: Only a Small Number of likely that a sizable percentage of both natural and Chemicals Are Carcinogenic or Teratogenic, and We Can Eliminate Them man-made chemicals will be reproductive toxins when tested at these doses. The world is full of car cinogens and reproductive toxins, and it always has been. The important issue is the human exposure More than 50% of the chemicals tested to date in rats and mice have been found to be carcinogens at the high doses administered,3,15 the maximum tolerated dose (MTD). The exhaustive database ofanimal can cer tests developed by me and my colleagues16,17 listed dose, and, fortunately, this is almost always miniscule. The major preventable risk factors for cancer cau sation, such as tobacco, dietary imbalances,1113,1*20-23 hormones,10 and viruses,24,25 have been discussed by us3-*'26 and others.2,9-12 392 chemicals tested in both rats and mice at the MTD. Of these, 58% of the synthetic chemicals and 45% of the natural chemicals were carcinogens in at least one species.3,15 We concluded that the proportion Error 3: Man-made Chemical Pollutants Are Present in Significant Amounts ofchemicals found to be carcinogens is strikingly high, We have attempted to address the issue of priority a conclusion reached by others on the basis of smaller setting among possible carcinogenic hazards.3 Car compilations. The earlier Innes and colleagues study1* cinogens differ enormously in potency in rodent tests is sometimes cited to support the conclusion that the and comparisons of possible hazards from variou: proportion ofcarcinogens is low. However, that study carcinogens ingested by humans must take this intc used a much smaller dataset (120 chemicals, 11 pos account. Our analysis makes use of an exhaustm itive), and the tests, although appropriate for their to time, used only one species and were less thorough cn than modem tests.15 database of carcinogenic potency (currently 35(X experiments on 975 chemicals)16,17 that analyzes an imal cancer tests and calculates the TDS0, essential!; to cn co Even when one considers that some chemicals are selected for testing based on a high index of suspicion, the large proportion of positive findings is disturbing. the dose of the carcinogen sufficient to cause cance in half of the animals. The TDjo is close to the higl dose (MTD) actually given, and thus involves a min oo From considerations of carcinogenesis mechanisms, imal extrapolation. To calculate our index of possibl it is plausible that a large proportion of all chemicals hazard, we express each human exposure (daily life we test in the future, both natural and man-made, time dose, in mg/kg) as a percentage of the roder will prove to be carcinogens (see "Error 4" below).3 TD50 dose (mg/kg) for each carcinogen. We call th; Large proportions of positives are also reported for percentage HERP (human exposure dose/rodent p< teratogenesis tests. Fully one-third of the 2800 chem tency dose). Because rodent data are all calculate icals tested in laboratory animals have been shown on the basis of lifetime exposure at the indicated dai to induce birth defects at the MTD.19 Thus, it seems dose rate,16'17 the human exposure data are similar BFG39721 Error 3: Man-made Chemical Pollutants Are Present in Significant Amounts 239 Table 14a-l Ranking of Possible Carcinogenic Hazards* Possible Hazard: HERP (%) Daily Human Exposure Carcinogen Dose per 70-kg Person Potency of Carcinogen: TD,, (mg/kg) Rats Mice 0.00 It 0.004f 0.0004f 0.00021 0.0003t 0.00*f 0.6 0.004 2.1 0.0002t 0.00031 0.0004 0.003 0.006 0.003 0.03 0.03 0.06 0.07 0.1 0.1 0.2 0.008 2.8t 4.71 6.2f 1.3 0.0002 0.06t [0.3] (5.6) 114] I6t I7t 5.8 140 Environmental Pollution Ttp wster I L Chloroform. 83 pg (US average) Well wster, 1 L contaminated (wont well in Trichloroethylene, 2800 pg Silicon Valley) Well water. 1 [.contaminated, Woburn Trichloroethylene, 267 p% Chloroform, 12 Mg Tetrachloroethylene. 21 mi Swimming pool. 1 h (for child) Chloroform, 250 pg (average pool) Conventional home air (14 h/d) Formaldehyde, 598 pg Benzene, ISS pg Mobile home air (14 h/d) Formaldehyde, 2.2 mg Pesticide and Other Residues PCBs: daily dietary intake PCBs. 0.2 pg (US average) DOE/DDT: daily dietary intake DDE, 2.2 pg (US average) EOB: daily dietary intake (from grains and grain Ethylene dibromide, 0.42 pg (US products) average) natural Pesticides and Dietary Toxins Bacon, cooked (100 g) Dimcthylnitrosamine, 0.3 pg Dicthylnitrosamine, 0.1 pg Sake (250 ml) Urethane, 43 pg ' Comfrcy herb tea (1 cup) Symphytinc, 38 ut (750 mC of pyrrolitidinc alkaloids) Peanut butter (32 g; one sandwich) Aflatoxin. 64 ng (US average. 2 ppb) Dried squid, broiled in gas oven (54 g) Dimcthylnitrosamine. 7.9 pg Brown mustard (5 g) Allyl isothiocyanate, 4.6 mg Basil (1 t ofdried leaf) Estragole. 3.8 mg Mushroom, one raw (Agaricus bisporus: 15 g) .Mixture of hydrazines, etc. Natural root beer (12 ounces; 354 ml; now banned) Safrolc, 6.6 mg Beer, before 1979 (12 ounces; 334 ml) Dimcthylnitrosamine. 1 Mg Beer (12 ounces; 354 ml) Ethyl alcohol, 18 ml Wine (250 ml) Ethyl alcohol. 30 ml Comfrey-pepsin tablets (nine daily) Comfrcy root. 2700 mg Comfrey-pepsin tablets (nine daily) Symphytinc. 1.8 mg AF-2: daily dietary intake before banning Diet cola (12 ounces; 354 ml) Food Additives AF-2 (furylfuramide), 4.8 Mg' Saccharin. 95 mg Phenacetin pill (average dose) Metronidazole (therapeutic dose) Isoniazid pill (prophylactic dose) Phenobarbital, one sleeping pill Clofibrate (average daily dose) Drugs Phenacetin, 300 ms Metronidazole. 2000 mg isoniazid. 300 mg Phenobarbital. 60 mg Clofibrate, 2000 mg Occupational Exposure Formaldehyde: workers* average daily intake Formaldehyde. 6.1 mg EDB: workers' daily intake (high exposure) Ethylene dibromide, 1 SO mg (119) (-) (-) (119) 101 (119) 1.5 (157) 1.5 1.7 (-) 1.5 (0.2) 0.02 (41) 1.9 0.003 (0.2) 96 (7) (?) (436) (0.2) 9110 9110 626 1.9 29 2143 1246 (542) (150) (+) 169 1.5 1.5 90 941 941 90 (126) 90 (44) 53 (44) (9.6) 13 (51) 0.2 (+) 22 (?) (+) 0.2 (-) 52 20.300 56 0.2 (?) (?) (?) (131) (-) (2137) 506 30 5.5 (?) (44) (5.1) Reprinted with permission from Ames BN. Mapw R. Gold LS: Rankins possible carcinogenic hazards. Science 236: 271-280. 1987. * Pvtemy tifcarcinogens: A number in parentheses indicates a TD* value not used in HERP calculation because it is the less sensitive species: (-) negative in cancer test:<+) positive for carcinofenkity in tesu not suitable for calculating a TD*: (7) not adequately tested for carcinogenicity. TD* values shown are averages calculated by uking the harmonic mean oTthe TD* of the positive tesu in that specks dram the Carcinogenic Potency Database. Results are simitar if the lowest TO* value (most potent) is used instead. For each test the target site with the lowest TD* value has been used. The average TD* has been calculated separately for rats and mice, and the more sensitive species is used for calculating the passible hazard. The database, with references to the source of the cancer tests, is complete for tesu published through 1984 and for the National Toticology Program bioassays through June 1986. We have not indicated the route of exposure or target sites or other particulars ofeach test, although these are reported in the database. Daily human exposure: We have tried to use average or reasonable daily intakes to facilitate comparisons. In several cases, such as contaminated well water or factory exposure to EDB. this is difficult to determine; these are assigned the value for the worst levels found. The calculations assume a daily dose for a lifetime; where drugs are normally taken only for a short period, we have bracketed the HERP value. For inhalation exposures, we assume an inhalation of 9600 L per 8 h for the workplace and 10.800 L per 14 h for indoor air at home. Possible hazard: The amount of rodent carcinogen indicated under carcinogen dose is divided by 70 kg to give a milligram per kilogram of human exposure, and this human dqse is given as the percentage of the TD* dose in the rodent (mg/kg) to calculate the HERP. t HERP from carcinogens thought to be nongenotoxic. 25259005 BFG39722 240 What Are the Major Carcinogens in the Etiology of Human Cancer? expressed, although the human exposure is likely to be less than daily for a lifetime. The HERP values are not risk assessments, because it is impossible to extrapolate to low doses (see "Error 4"), but provide a way to compare possible hazards of exposures so as to put them in perspective and set priorities (see Table 14a-l). This analysis suggests that the amounts of pollution humans ingest from pesticide residues or water pol\ lution appear to be trivial relative to the background I of natural or traditional (e.g., from cooking food) car cinogens.3-26 L Nature's Pesticides Americans ingest in their diet at least 10,000 times more natural pesticides (by weight) than man-made pesticide residues.26 These natural "toxic chemicals" have an enormous variety of chemical structures, ap pear to be present in all plants, and serve as protection against fungi, insects, and animal predators.8,26 Al though only a few dozen are found in each plant spe cies, they commonly make up 5% to 10% ofthe plant's dry weight.26 There has been relatively little interest in the toxicology or carcinogenicity of these com pounds until quite recently, although they are by far the main source of "toxic chemicals" ingested by hu mans. Although most chemicals tested for carcinogenicity in rodent bioassays are synthetic compounds, the proportion ofpositive tests is about as high for natural pesticides as for synthetic chemicals. Because more than 99.99% of the pesticides we ingest are "nature's pesticides,"3-8,26 our diet is likely to be very high in natural carcinogens. Their concentration is usually in parts per thousand or more, rather than parts per billion, as is usual for synthetic pesticide residues or y- water pollution.3 The known natural carcinogens in mushrooms, parsley, basil, parsnips, fennel, pepper, celery, figs, mustard, and citrus oil are no doubt just the beginning of the list because so few of nature's \ pesticides have been tested.3-8-26 For example, a recent analysis27 oflima beans showed an array of 23 natural alkaloids (those tested have biocidal activity) that ranged in concentration in stressed plants from 0.2 to 33 parts per thousand fresh weight. None appears to have been tested for carcinogenicity or teratoge nicity. Man-made Pesticide Residues The intake of man-made pesticide residues from food in the United States, including residues of industrial chemicals such as polychlorinated biphenyls (PCBs), has been estimated by the Food and Drug Adminis tration (FDA). They assayed food for residues of the 70 compounds thought to be of greatest importance.28 The human intake averages about 150 mg/day. Most of this intake (105 Mg) is composed of three chemicals (ethylhexyl diphenyl phosphate, malathion; and chlorpropham), shown to be noncarcinogenic in tests in rodents.3 Thus, the intake ofcarcinogens from res idues (45 Mg/day if all the other residues are carci nogenic, which is unlikely) is extremely small relative to the background of natural substances.3,26 The latest figures from the FDA about actual ex posures do not include every known man-made pes ticide, but represent a reasonable attempt to do so. In a recent National Research Council/National Academy of Sciences (NRC/NAS) report. Regulating Pesticides in Food,29 it is suggested that some of the pesticides not covered by the FDA sampling, partic ularly those used on tomatoes, should have their al lowable limits lowered and presumably should be added to the FDA sampling program. Nevertheless, 1 theestimate of45 Mg of possibly carcinogenic pesticide residues consumed in a day is likely to be a reasonable one, as is our conclusion that the possible hazards from these residues are minimal compared to the background of nature's pesticides. For comparison,3 j there are about 500 Mg of carcinogens in a cup of ! coffee (hydrogen peroxide and methylglyoxal), 185 Mg of carcinogenic formaldehyde in a slice of bread, about 2000 Mg of formaldehyde in a cola, 760 Mg of carcinogenic estragole in a basil leaf, a gram of burnt material from cooking our food, plus nitrosamines (_ formed in gas ovens. An alternative to synthetic pesticides is to raise the level of natural plant toxins by breeding. However, it is not clear that this approach, even where feasible, is preferable. One consequence of disproportionate concern about tiny traces of synthetic pesticide resi dues, such as ethylene dibromide,3 is that plant breeders are developing highly insect-resistant plants, thus creating other risks. Two recent cases are in structive. A major grower introduced a new variety of highly insect-resistant celery into commerce. There was soon a flurry of complaints to the Centers for Disease Control (CDC) from all over the country, because people who handled the celery developed a severe rash when they were subsequently exposed to sunlight. Some detective work revealed that the pestresistant celery contained 9000 parts per billion (ppb) psoralens (light-activated mutagenic carcinogens) in stead of the level of 900 ppb psoralens in normal cel ery.30-31 It is unclear whether other natural pesticides in the celery were increased as well. Solanine andchaconine (the main natural alkaloids in potatoes) are cholinesterase inhibitors that were 25259006 r BFG39723 Error 3: Man-made Chemical Pollutants Are Present in Significant Amounts 241 widely introduced into the human diet about 400 years ago with the dissemination of the potato from the Andes. They can be detected in the blood of all potato eaters. Total alkaloids are present in potatoes at a level of 15,000 ng per 200 g potato, which is only about a sixfold safety margin from the toxic level for humans.3 Neither alkaloid has been tested for carci nogenicity. By contrast, the pesticide malathion, the main synthetic organophosphate cholinesterase in hibitor present in our diet (17 /tg/day), has been thor oughly tested and is not a carcinogen in rodents. Plant breeders produced an insect-resistant potato that had to be withdrawn from the market because of its acute toxicity to humans, a consequence of higher levels of solanine and chaconine. There is a tendency for laymen to think of chem icals as being only man-made, and to characterize them as toxic, as if every natural chemical were not also toxic at some dose. Even a recent NRC/NAS report states that "Advances in classical plant breeding . . . offer some promise for nonchemical pest control in the future. Nonchemical approaches will be en couraged by tolerance revocations if more profitable chemical controls are not available.. . ."29 The report was particularly concerned with some pesticides used on tomatoes. Of course, tomatine, one of the alkaloids in tomatoes, is a chemical too, and was introduced from the New World 400 years ago. It has not been tested in rodent cancer bioassays, is present at 36,000 /ig/100 g tomato, and is orders of magnitude closer to the toxic level than are man-made pesticide resi dues. The Idea That Nature Is Benign The notion that evolution has allowed us to cope with the toxic chemicals in the natural world32 is not compelling7 for several reasons: First, there is no rea son to think that natural selection should eliminate the carcinogenic hazard of a plant toxin that causes cancer past the reproductive age, although there could be selection for resistance to the acute effects of par ticular carcinogens. For example, aflatoxin, a mold toxin that presumably arose early in evolution, causes cancer in trout, rats, mice, monkeys, and probably people, although the species are not equally sensi tive.5,33 Many of the common metal salts are carcin ogens (e.glead, cadmium, beryllium, nickel, chro mium, selenium, and arsenic) despite their presence during all of evolution. Second, it is argued by some that humans, as opposed to rats or mice, may have developed resistance to each specific plant toxin or chemical in cooked food.32 This is unlikely, because both rodents and humans have developed many types of general defenses against the large amounts and enormous variety of nature's pesticides.3'8,26 These defenses include the constant shedding of the surface layer of cells ofthe digestive system, the detoxification of alkylating agents by glutathione transferases, the active excretion of hydrophobic toxins out of liver or intestinal cells, numerous defenses against oxygen radicals, and DNA excision repair. The fact that defenses usually are general, rather than specific for each chemical, makes good evolutionary sense and is sup ported by various studies. Experimental evidence in dicates that these general defenses are effective against both natural and synthetic compounds,34 since basic mechanisms of carcinogenesis are not unique to ei ther. Third, the human diet has changed drastically in the last few thousand years, and most of us are eating recently introduced plants (coffee, potatoes, tomatoes, and kiwi fruit) that our ancestors did not eat. Fourth, the argument that plants contain anti carcinogens which protect us against plant carcino gens is irrelevant: plant antioxidants, the major known type of ingested anticarcinogens, do not dis tinguish whether oxidant carcinogens are synthetic or natural in origin, and thus help to protect us against both. Fifth, it has been argued that synthetic carcin ogens can be synergistic. However, this is also true of natural chemicals and is irrelevant to the argument that synthetic pesticide residues in food or water pol lution appear to be a trivial increment over the back ground of natural carcinogens. Dioxin Compared to Alcohol and Broccoli Common sense suggests that a chemical pollutant should not be treated as a significant hazard if its possible hazard level is far below that of common food items. Dioxin (TCDD) is a substance of great public concern, because it is an extremely potent car cinogen and teratogen in rodents, yet the doses hu mans are exposed to are very low relative to the ef fective level in rodents. TCDD can be compared to alcohol, as an example. Alcohol is an extremely weak carcinogen and teratogen, yet the doses humans are exposed to are very high relative to the effective dose in rodents (or humans). Indeed, alcoholic beveragesT7 are the most important known human teratogen, and-U the effective dose level of alcohol in humans (in mil ligrams per kilogram) is similar to the level that causes birth defects in mice. By contrast, there is no con vincing evidence that TCDD is carcinogenic or ter atogenic in man, although it is in rodents. If one <compares the teratogenic potential of TCDD to that of alcohol for causing birth defects, after adjusting 25259007 242. What Are the Major Carcinogens in the Etiology of Human Cancer? for their potency in rodents,3 then a daily consump tion of the Environmental Protection Agency (EPA) "reference dose" (formerly "acceptable dose limit") ofTCDD, 6 pg/kg per day, is equivalent in teratogenic potential to the amount ofalcohol ingested daily from 1/3000 of a beer, the equivalent of drinking one beer (15 g ethyl alcohol) over a period of 8 years. A daily slice of bread or glass of orange juice contains much more natural alcohol than this. Alcoholic beverages are clearly carcinogenic in man (at a daily dose ofabout 5 drinks), although only one of several tests on ethyl alcohol in rats was pos itive.22,23 This test should be replicated as confirma tion that ethyl alcohol is the active ingredient, al though the evidence for that is fairly strong.22 A comparison of the carcinogenic potential of TCDD with that of alcohol, adjusting for potency in rodents, shows that the equivalence for the TCDD reference dose of 6 pg/kg per day is l beer every 5 months. Since the average per capita consumption of alcohol in the United States is equivalent to more than one beer per day, the great concern over TCDD at levels in the range ofthe reference dose seems unreasonable. The assumption of a worst-case linear dose-re sponse, often used for carcinogens, is not plausible for TCDD, yet extrapolations to man using such as sumptions have generated great concern. TCDD binds to a receptor in mammalian cells, the Ah re ceptor, and the evidence suggests strongly that all of TCDD's effects are mediated through this binding.35 Moreover, a wide variety of natural substances bind to the Ah receptor, and, as far as they have been ex amined, they have all of the properties ofTCDD. A cooked steak contains polycyclic hydrocarbons, which bind to the Ah receptor and mimic TCDD. In ad dition, our diet contains a variety of flavones and other substances from plants, which bind to the Ah receptor. The most interesting of such substances is indole carbinol (IQ, which is present in large amounts in broccoli (500 mg/kg), cabbage, cauliflower, and other members of the Brassica family.34 The two substances induce the same set of enzymes.37 When given before aflatoxin or other carcinogens, IC pro tects against carcinogenesis, as does TCDD.38 How ever, when it is given after aflatoxin or other carcin ogens, IC is a strong promoter of carcinogenesis, as is TCDD.3' This stimulation of carcinogenesis has also been shown for cabbage itself.40 When IC is ex posed to acid pH (equivalent to that of the stomach), it is converted to a series of dimers and trimers that are similar to TCDD in size and shape, bind to the Ah receptor, and induce the set of TCDD-inducible enzymes, thus mimicking TCDD.37,41 The 360 pg/ day TCDD EPA reference dose should be compared with 50 million pg of IC per 100 g of broccoli (one portion); the affinity of the indole derivatives in bind ing to the Ah receptor is less by a factor of about 8000, suggesting that the broccoli portion might be roughly 20 times the possible hazard. Although these IC derivatives appear to be much more of a possible hazard than TCDD, it is not clear whether, at these low doses, either represents any danger. Another study42 also shows that when sunlight ox idizes tryptophan, a normal amino acid, it converts it to a variety of indoles (similar to the broccoli IC dimers), which bind to the Ah receptor and mimic the action ofTCDD. It seems likely that many more of these "natural dioxins" will be discovered in the future. Water Pollution The possible hazards from carcinogens in contami nated well water (e.g., the Santa Clara, or "Silicon," Valley, in California, and Woburn, Massachusetts) should be compared to the possible hazards of ordi nary tap water (see Table 14a-l).3 Of the 35 wells shut down in Santa Clara Valley because of a sup' posed carcinogenic hazard--low traces of trichloro ethylene--only two were of a possible hazard greater than ordinary tap water. Well water is not usually chlorinated and therefore lacks the 83 ppb chloroform present in average chlorinated tap water.3 Water from the most polluted well had a relative hazard that was orders of magnitude less than that for the carcinogens in an equal volume of cola, beer, or wine, or many natural carcinogens in our daily diet. The consump tion of tap water is only about 1 or 2 liters/day, and the animal evidence cited3 provides no good reason to expect that chloroform in water or current levels of man-made pollution of water would pose a signif icant carcinogenic hazard. The trace amounts of chemicals found in polluted wells should be a negligible cause of birth defects, when compared to the background level of known teratogens such as alcohol. Most agents causing birth defects would also be expected to be harmless at low doses. Important risk factors for birth defects in hu mans include age of mother, alcohol ingestion, smoking, and rubella virus. Air Pollution A person inhales about 20,000 liters of air in a day. Thus, even modest contamination of the atmosphere results in inhalation of appreciable doses of a pollut ant. Indoor air pollution is, in general, considerably more of a health hazard than outdoor air pollution, partly because of cigarette smoke, formaldehyde, benzene, and radon.3,43 The most important indoor 25259008 BFG39725 Error 4: Extrapolating Risks Without Understanding Mechanisms of Carcinogenesis 243 lir pollutant is radon gas.43 Radon is a natural radio active gas that is present in the soil, gets trapped in houses, and gives rise to radioactive decay products that are known to be carcinogenic in humans.43 It has been estimated that one million homes in the United States have a level of exposure to products of radon decay higher than that received by today's ura nium miners. Two particularly contaminated houses had a risk estimated to be equivalent to receiving about 1200 chest x-rays a day. Approximately 10% of the lung cancer in the United States has been ten tatively attributed to radon pollution in houses.43 Many of these cancers may be preventable, because the most hazardous houses can be identified and modified to minimize radon contamination.43 General outdoor air pollution is a small risk relative to indoor air pollution or to the pollution inhaled by a smoker: a person breathing Los Angeles smog for a year inhales the same amount of burnt material that a smoker does in one day (two packs).3-26 It is difficult for epidemiologists to determine cancer risk from outdoor air pollution because smoking and ra don exposure must be accurately controlled. Cooking Food The cooking of food generates a variety of mutagens and carcinogens. The total amount of browned and burnt material eaten in a typical day is at least several hundred times more than that inhaled from severe outdoor air pollution.26 Nine heterocyclic amines, isolated on the basis of their mutagenicity from pro teins or amino acids that were heated in ways that reproduce cooking methods, have now been tested; all have been shown to be potent carcinogens in ro dents.44-45 Many others are still being isolated and characterized.44,45 Three mutagenic nitropyrenes present in diesel exhaust have been shown to be car cinogens,46 but the intake of these carcinogenic ni tropyrenes from grilled chicken is estimated to be much higher than that from air pollution.44,45-47 Gas flames generate NO2, which can form both the carcinogenic nitropyrenes23-24 and the potently car cinogenic nitrosamines in food, such as fish, cooked in gas ovens. It seems likely that food cooked in gas ovens may be a major source of dietary nitrosamines and nitropyrenes. Occupational Exposures Occupational exposures to chemicals are often sig nificant.3-48 The potential carcinogenic hazards to US workers has been ranked using the PERP index (analogous to the HERP index,3 except that the Oc cupational Safety and Health Administration per mitted exposure levels replace actual exposures).48 The PERP values differ by more than 100,000-fold. For 12 substances, the permitted levels for workers are greater than 10% ofthe rodent TDjo dose. Priority attention should be given to reduction ofthe allowable worker exposures that appear most hazardous in the PERP ranking. Error 4: Extrapolating Risks Without Understanding Mechanisms of Carcinogenesis Rodent Carcinogenesis Tests It is prudent to assume that if a chemical is a carcin- ~1 ogen in rats and mice at the MTD, it is also likely to I be a carcinogen in humans at the MTD. However, until we understand more about mechanisms of car cinogenesis, we cannot reliably predict risk to humans at low doses, often hundreds of thousands of times below the dose where ah effect is observed in rodents. Thus, quantitative risk assessment currently is not scientifically possible.3,4,7 Carcinogenesis Mechanisms and the Dose-Response Curve The study of mechanisms of carcinogenesis is a rap idly developing field and is essential for rational risk assessment. Both mutations and cell proliferation (i.e., promotion) are required in carcinogenesis.3,49,50 There is an enormous spontaneous rate of damage to DNA from endogenous oxidants, which we have discussed in relation to cancer and aging.51-52 There is also a basal spontaneous rate for cell proliferation in some organs, e.g., colon, but not others, e.g., liver.10-'1 Thus, increasing either mutation or cell proliferation should be carcinogenic. Additional complications are that several mutations appear nec essary for carcinogenesis, and there are many layers of defense against carcinogens. These considerations suggest a sublinear dose-response relationship, which is consistent with both the animal and human data,3-7 and indicate that multiplicative interactions will be common in human cancer causation. Admin--) istering chemicals in cancer tests at the MTD com- 1 monly causes cell proliferation and inflammatory reactions.3-49-50-53 Inflammatory reactions with release \ of oxygen radicals from phagocytic cells are equivalent J to irradiating the tissue. Ifa chemical is nonmutagenic and its carcinogenicity is due to cell proliferation re sulting from near-toxic doses, one might commonly expect a threshold.3 49 30 25259008 BFG39726 244 What Are the Major Carcinogens in the Etiology of Human Cancer? The fact that high doses of a chemical cause tumors these may be acting as promoters and therefore may does not necessarily mean that small doses will. Most not be of interest at doses much below the toxic dose.3 chemicals may, in fact, be harmless at low levels. A Thus, the common water pollutants, such as tri list of carcinogens is not enough. The main rule in chloroethylene (TCE) and perchloroethylene (PCE), toxicology is that the "dose makes the poison:" at are unlikely to be of public health significance because some level, every chemical becomes toxic, but there [a] the amounts we are exposed to in pollution are are safe levels below that. A scientific consensus trivial relative to the background of natural carcin evolved in the 1970s that we should treat carcinogens 1 ogens, and [b] the evidence is that they are likely to differently--that we should assume that even low be acting as promoters, not as DNA-damaging car doses could possibly cause some harm, even though I cinogens, and therefore should be ignored at low con we do not have the methods to measure effects at low ^ centrations. levels. This idea evolved because most carcinogens appeared to be mutagens (agents that damage the DNA). The precedent of radiation, which is both a mutagen and carcinogen, gave credence to the idea Error 5: Storks Bring Babies, and Pollution Causes Cancer and that there could be effects of chemicals even at low Birth Defects doses. Some recent work on radiation, however, sug gests that low doses may be of no harm or even pro The number of storks in Europe has been decreasing tective. 34,34*,33,33* for decades. At the same time, the European birth The idea that most of the classical carcinogens were rate has also been decreasing. We would be foolish mutagens that damaged DNA (about 90% in our to accept this high correlation61 as evidence that storks studies),36,37 along with work on oncogenes,3* rein bring babies. The science ofepidemiology tries to sort forced the mutagen-carcinogen connection. How- out from the myriad chance correlations those that evervin recent years there has been a change in the are meaningful and involve cause and effect. How picture. About half of all chemicals tested in animals ever, it is not easy to obtain convincing evidence by are carcinogens, but only about half of these appear epidemiologic methods because of inherent meth- to be mutagenic. It is now standard in cancer tests to odologic difficulties.9 There are many sources of bias be rigorous about giving the MTD of the chemical in observational data, and chance variation is also an for the lifetime ofthe animal, and this may be a factor. important factor. For example, because there are so It seems quite reasonable that nonmutagens cause many different types of cancer or birth defects, by cancer, and mutagens in part cause cancer, because chance alone one might expect some of them to occur dosing at the MTD accelerates the promotional step at a high frequency in a small community. Toxicology of carcinogenesis.39,60 provides evidence to help decide whether an observed Promotion, or cell proliferation, can also be accel correlation might be causal or accidental. erated by viruses, such as the human carcinogenic There is no convincing evidence from epidemiology hepatitis B viruses, a major cause of liver cancer or toxicology that pollution is a significant source of around the world,23 or human papilloma virus 16 birth defects and cancer. For example, the epidemi (HPV-16), a contributor to cancer of the cervix.24 ologic studies of Love Canal, dioxin in Agent Or Both cause chronic cell killing and consequent cell ange,62,63 Contra Costa County refineries,64,65 Silicon proliferation. Promotion can also be induced by hor Valley,66 Woburn,3,8 and the use of DDT provide no mones, which cause cell proliferation. Hormones ap convincing evidence that pollution was the cause of pear to be major risk factors for certain human can human harm in any of these well-publicized expo cers, such as breast cancer, and appear to only increase sures. Even in Love Canal, where people were living cell proliferation.10 The promotional step of cancer next to a toxic waste dump, the epidemiologic evi causation can also be accelerated by chemicals. Al dence for an effect on public health is equivocal. cohol, for example, causes cirrhosis of the liver, lead Analysis of the toxicology data on many of these cases ing to cancer. The classical chemical promoters, such suggests that the amounts of the chemicals involved as phenobarbital and tetradecanoyl phorbol acetate, were much too low relative to the background of nat would be expected to be, and are, carcinogens when ural and traditional carcinogens to be credible sources tested in thorough animal tests at the MTD.60 There of increased cancer risk to humans.3 A comparative is increasing evidence to show that low doses of pro analysis of teratogens using a HERP-type index ex moters are not active.49,30 It seems likely, therefore, pressing the human exposure level as a percentage of that a high percentage of all chemicals, both man the dose level effective in rodents would be of interest made and natural, will cause cell proliferation at the (see Error 3), but this has not been done in a system MTD and be classified as carcinogens, but most of atic way. 25259010 References 245 Environmental exposure to TCE, PCE, trichloroethane, ethylene dibromide (EDB), and other pollut ants is thousands of times lower than the exposure to these same agents in the workplace.3,48 Thus, if parts per billion ofthese pollutants were causing can cer or birth defects, one might expect to see an effect in the workplace. The studies on these chemicals to date do not provide any evidence for a causal asso ciation,33 although epidemiologic studies are inher ently insensitive. Historically, cases of cancer due to workplace exposure resulted mainly from exposures to chemicals at very high levels. For example, the permissible and actual EDB levels for workers were shockingly high (see Table 14a--1). (I testified in Cal ifornia in 1981 that our calculations showed that the workers were allowed to breathe in a dose higher than the TDjo in rats.) California lowered the permissible worker exposure by more than 100-fold. Despite the fact that the epidemiology of EDB in highly exposed workers does not show any significant effect, the un certainties in our knowledge make it important to have strict rules because workers can be exposed to extremely high doses. Error 6: Technology Is Doing Us In Modem technologies are almost always replacing older, more hazardous technologies. The reason that billions ofpounds ofTCE (one of the most important industrial nonflammable solvents) and PCE (the main dry-cleaning solvent in the United States) are used is that they-have low toxicity and are not flammable. Is it advisable to go back to the age when industry and dry cleaners used flammable solvents and fires were frequent? Eliminating a carcinogen may not always be a good idea. For example, EDB, the main fumigant in the United States before it was banned, was present in trivial amounts in our food: the average daily intake was about one-tenth of the possible carcinogenic haz ard ofthe aflatoxin in the average peanut butter sand wich, a trivial risk in itself (see Table 14a-1).3 Elim ination of fumigation results in insect infestation and subsequent contamination of grain by carcinogenproducing molds. This might result in a regression in public health, not an advance, and would also greatly increase costs. The proposed alternatives, such as ir radiating food, could possibly be more hazardous than EDB, as well as more expensive. Similarly, modem pesticides replaced more hazardous substances such as lead arsenate, one of the major pesticides before the modem era. Lead and arsenic are both natural, highly toxic, and carcinogenic. Pesticides have in creased crop yields and brought down the price of foods, a major public health advance. Every living thing and every industry "pollutes" to some extent. How much does society wish to spend to get the last part per billion of TCE out of the wells in Silicon Valley, or to remove PCE from dry-cleaning plants? We are currently spending enormous amounts of money trying to eliminate lower and lower levels of pollution; one estimate is about $80 billion an nually.7 The fact that scientists have developed meth ods to measure parts per billion (one part per billion is equivalent to one person in all of China) of carcin ogens and are developing methods to measure parts per trillion does not mean that significant pollution is increasing, or that the pollution found is a cause of human harm. Conclusion j'Everyone knows that spending all of one's effort on I trivia without focusing on important problems is i counterproductive. If we divert too much of our ati tention to traces ofpollution and away from impor\ tant public health concerns such as smoking (400,000 i deaths per year), alcohol (100,000 deaths per year), \ unbalanced diets (e.g., too much saturated fat and j cholesterol), acquired immunodeficiency syndrome, I radioactive radon in our homes, and high-dose oc cupational exposure, we do not improve public : health, and the important hazards are lost in the con tusion. It is the inexorable progress of modem tech nology and scientific research that will continue to provide the knowledge resulting in steady progress to decrease cancer and birth defects and lengthen life span. The author thanks Lois Gold and David Freedman for helpful discussion and criticisms. This work was sup ported by Outstanding Investigator Grant CA-39910 from the National Cancer Institute and by National Institute of Environmental Health Sciences Center Grant ES01896. References 1. National Cancer Institute: 1987 Annual Cancer Statistics Re view Including Cancer Trends: 1950-1985. NIH Publication No. 88-2789 Bethesda, MD, National Institutes of Health. 1988 2. Doll R, Peto R: The Causes ofCancer. Oxford, England, Oxford University Press, 1981 3. Ames BN, Magaw R, Gold LS: Ranking possible carcinogenic hazards. Science 236:271-280, 1987 4. Ames BN, Magaw R, Gold LS: Response to letter Risk as sessment. Science 237:235, 1987 5. Ames BN, Magaw R, Gold LS: Response to letter Carcino genicity of aflatoxins. Science 237:1283-1284, 1987 25259011 BFG39728 246 What Are the Major Carcinogens in the Etiology of Human Cancer? 6. Ames BN, Gold LS, Magaw R: Response to letter: Risk as sessment. Science 237:1399-1400, 1987 7. Ames BN, Gold LS: Response to letter. Paleolithic diet, evo lution, and carcinogens. Science 238:1634, 1987 8. Ames BN, Gold LS: Response to technical comment: Carci nogenic risk estimation. Science 240:1045-1047, 1988 9. Higginson J: Changing concepts in cancer prevention: Limi tations and implications for future research in environmental carcinogenesis. Cancer Res 48:1381-1389, 1988 10. Henderson BE, Ross R, Bernstein L: Estrogens as a cause of human cancer The Richard and Hinda Rosenthal Foundation Award Lecture. Cancer Res 48:246-253, 1988 11. Lipkin M: Biomarkers of increased susceptibility to gastroin testinal cancer New application to studies ofcancer prevention in human subjects. Cancer Res 48:235-245, 1988 12. Peto R: Epidemiological reservations about risk assessment In Woodhead AD, Shcllabarger a. Pond V, Hollaender A (eds): Assessment of Risk from Low-Level Exposure to Radia tion and Chemicals, pp 3-16. New York, Plenum, 1985 13. Yang CS, Newmark HL: The role of micronutrient deficiency in carcinogenesis, CRC Crit Rev Oncol Hematol 7:267-287, 1987 14. Pence BC, Buddingh F: Inhibition ofdietary fat-promoted colon carcinogenesis in rats by supplemental calcium or vitamin Dj. Carcinogenesis 9:187-190, 1988 15. Gold LS, Bernstein L, Magaw R, Slone TH: Interspecies ex trapolation in carcinogenesis: Prediction between rats and mice. Environ Health Perspect (in press) 16. Gold LS, Sawyer CB, Magaw R, Backman GM. de Veciana M, Levinson R, Hooper K, Havender WR, Bernstein L, Peto R, Pike MC, Ames BN: A carcinogenic potency database of the standardized results of animal bioassays. Environ Health Perspect 58:9-319, 1984 17. Gold LS, Slone TH, Backman G, Magaw R, Da Costa M, Ames BN: Second chronological supplement to the carcino genic potency database: Standardized results ofanimal bioassays published through December 1984 and by the National Tox icology Program through May 1986. Environ Health.Perspect 74:237-329, 1987 18. Innes JRM, Ulland BM, Valerio MG, Petrucelli L, Fishbein L, Hart ER, Pallotta AJ, Bates RR, Falk HL, Gart JJ, Klein M, Mitchell 1, Peters J: Bioassay of pesticides and industrial chemicals fortumorigenicity in mice: A preliminary note. JNC1 42:1101-1114,1969 19. Schardein JL. Schwetz BA, Kenal MF: Species sensitivities and prediction of teratogenic potential. Environ Health Perspect 61:55-67, 1985 20. Reddy BS, Cohen LA (eds): Diet, Nutrition, and Cancer A Critical Evaluation, Vols I and 11. Boca Raton, FL, CRC Press 1986 21. Joossens JV, Hill MJ, Geboers J (eds): Diet and Human Car cinogenesis. Amsterdam, Elsevier Science Publishers, 1986 22. Ames BN: Review ofevidence for alcohol-related carcinogen esis. Report for Proposition 65 Meeting, Sacramento, CA, De cember 11, 1987 23. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Alcohol Drinking. Lyon, France, International Agency for Research on Cancer (in press) 24. Peto R, zur Hausen H (eds): Banbury Report 21. Viral Etiology of Cervical Cancer. Cold Spring Harbor, NY, Cold Spring Harbor Laboratory, 1986 25. Yeh F-S, Mo C-C, Luo S, Henderson BE, Tong MJ. Yu MC: A seriological case-control study ofprimary hepatocellular car cinoma in Guangxi, China. Cancer Res 45:872-873, 1985 26. Ames BN: Dietary carcinogens and anticarcinogens: Oxygen radicals and degenerative diseases. Science 221:1256-1264, 1983 27. Harborne JB: The role of phytoalexins in natural plant resis tance. In Green MB, Hedin PA (eds): Natural Resistance of Plants to Pests. Roles ofAlletochemicals. ACS Symposium 296, pp 23-35. Washington DC. American Chemical Society, 1986 28. Gartrell MJ, Craun JC, Podrebarac DS, Gunderson EL: Pes ticides, selected elements, and other chemicals in adult total diet samples, October 1980-March 1982. J Assoc Off Anal Chem 69:146, 1986 29. National Research Council, Board on Agriculture: Regulating Pesticides in Food. Washington, DC National Academy Press, 1987 30. Berkley SF, Hightower AW, Beier RC, Fleming DW, Brokopp CD, Ivie GW, Broome CV: Dermatitis in grocery workers as sociated with high natural concentrations of furanocoumarins in celery. Ann Intern Med 105:351-355, 1986 31. Seligman PJ, Mathias CGT, O'Malley MA, Beier RC, Fehrs LJ, Serrill WS, Halperin WE: Phytophotodermatitis from celery among grocery store workers. Arch Dermatol 123:1478-1482, 1987 32. Davis DL: Paleolithic diet, evolution, and carcinogens. Science 238:1633-1634, 1987 33. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Overall Evaluations of Carcinogenicity: An Up dating of IARC Monographs Volumes 1-42, Supplement 7. Lyon, France, International Agency for Research on Cancer 34. Jakoby WB (ed): Enzymatic Basis of Detoxification, Vols 1 and 2. New York, Academic Press, 1980 35. Knutson JC, Poland A Response of murine epidermis to 2,3,7,8-tetrachlorodibenzo-p-dioxin: Interaction of the Ah and hr loci. Cell 30:225-234, 1982 36. Bradfteld CA, Bjeldanes LF: High-performance liquid chro matographic analysis of anticarcinogenic indoles in Brassica oleracea. J Agric Food Chem 35:46-49, 1987 37. Bradfield CA, Bjeldanes LF: Structure-activity relationships of dietary indoles: A proposed mechanism of action as modifiers of xenobiotic metabolism. J Toxicol Environ Health 21:311323, 1987 38. Dashwood RH, Arbogast DN, Fong AT, Hendricks JD, Bailey GS: Mechanisms of anti-carcinogenesis by indole-3-carbinol: Detailed in vivo DNA binding dose-response studies after di etary administration with ailatoxin Bl. Carcinogenesis 9:427432, 1988 39. Bailey GS, Hendricks JD, Shelton DW, Nixon JE, Pawlowski NE: Enhancement ofcarcinogenesis by the natural anticarcin ogen indole-3-carbinot. JNCI 78:931-934, 1987 40. Birt DF, Pelting JC, Pour PM, Tibbels MG, Schweickert L, Bresnick E: Enhanced pancreatic and skin tumorigenesis in cabbage-fed hamsters and mice. Carcinogenesis 8:913-917, 1987 41. Bradfield C, Bjeldanes L: Personal communication, 1988 42. Rannug A, Rannug U, Rosenkranz HS, Winqvist L, Westerholm R, Agurell E, Grafstrom A-K; Certain photooxidized de rivatives of tryptophan bind with very high affinity to the Ah receptor and are likely to be endogenous signal substances. J Biol Chem 262:15422-15427, 1987 43. Nero AV Jr Controlling indoor air pollution. Scientific Amer ican 258:42-48, 1988 44. Sugimura T, Sato S, Ohgaki H, Takayama S, Nagao M, Wakabayashi K: Overview. Mutagens and carcinogens in cooked food. In Knudsen I (ed): Genetic Toxicology of the Diet, pp 85-107. New York, Alan R Liss, 1986 45. Sugimura T: Studies on environmental chemical carcinogenesis in Japan. Science 233:312-318, 1986 46. Ohgaki H, Hasegawa H, Kato T, Negishi C, Sato S, Sugimura T: Absence of carcinogenicity of 1-nitropyrene, correction of previous results, and new demonstration of carcinogenicity of 1,6-dinitropyrene in rats. Cancer Lett 25:239-245, 1985 25253012 BFG39729 47. (Cinouchi T, Tsutsui H, Ohnishi Y: Detection of 1 -nitropyrene in yakitori (grilled chicken). Mutation Res 171:105-113, 1986 48. Gold LS, Backman GM, Hooper NK, Peto R: Ranking the [totential carcinogenic hazards to workers from exposures to chemicals that are tumorigenic in rodents. Environ Health Perspect 76:211-219, 1987 49. Pitot HC, Goldsworthy TL, Moran S. Kennan W, Glauert HP, Maronpot RR. Campbell HA: A method to quantitate the rel ative initiating and promoting potencies of hepatocarcinogenic agents in their dose-response relationships to altered hepatic foci. Carcinogenesis 8:1491-1499, 1987 50. Farber E: Possible etiologic mechanisms in chemical carcino genesis. Environ Health Perspect 75:65-70, 1987 51. Adelman R, Saul RL, Ames BN: Oxidative damage to DNA: Relation to species metabolic rate and life span. Proc Natl Acad Sci USA 85:2706-2708, 1988 52. Richter C, Park J-W, Ames BN: Normal oxidative damage to mitochondrial and nuclear DNA is extensive. Proc Natl Acad Sci USA 85:6465-6467, 1988 53. Swenberg JA, Richardson FC, Boucheron JA. Deal FH, Belin sky SA. Charbonneau M, Short BG: High- to low-dose extrap olation: Critical determinants involved in the dose response of carcinogenic substances. Environ Health Perspect 76:57-63, 1987 54. WolffS, Afzal V, Wiencke JK, Olivieri G, Michaeti A: Human lymphocytes exposed to low doses ofionizing radiations become refractory to high doses of radiation as well as to chemical mutagens that induce double-strand breaks in DNA. Int J Ra dial Biol 53:39-18, 1988 54a.Yalow RS: Biologic effects oflow-level radiation. In Burns ME (ed): Low-Level Radioactive Waste Regulation: Science, Pol itics, and Fear, pp 239-259. Chelsea, ML Lewis Publishers, Inc, 1988 55. Ootsuyama A, Tanooka H: One hundred percent tumor in duction in mouse skin after repeated (J irradiation in a limited dose range. Radiation Res 115:488-494, 1988 55a.Kondo S: Mutation and cancer in relation to the atomic-bomb radiation effects. Jpn J Cancer Res (Gann) 79:785-799, 1988 References 247 56. McCann J, Choi E, Yamasaki E, Ames BN: Detection ofcar cinogens as mutagens in the Salmonella/microsome test. Assay of 300 chemicals. Proc Natl Acad So' USA 72:5135-5139, 1975 57. McCann J, Ames BN: The detection ofcarcinogens as mutagens in the Salmonella/microsome test: Assay of 300 chemicals: Discussion. Proc Natl Acad Sci USA 73:950-954, 1976 58. Stowers SJ, Maronpot RR, Reynolds SH, Anderson MW: The role of oncogenes in chemical carcinogenesis. Environ Health Perspect 75:81-86, 1987 59. Butterworth BE, SlagaTJ (eds): Banbury Report 25. Nongenotoxic Mechanisms in Carcinogenesis. Cold Spring Harbor, NY, Cold Spring Harbor Laboratory, 1987 60. Iversen OH (ed): Theories ofCarcinogenesis. Washington, DC, Hemisphere, 1988 61. Sies H: A new parameter for sex education. Nature 332:495, 1988 62. Lathrop GD, Machado SG, Karrison PG, Grubbs WD, Thomas WF, Wolfe WH, Michalek JE, Miner JC, Peterson MR, Ogerskok RW: Air Force Health Study. Epidemiologic Investigation of Health Effects in Air Force Personnel Following Exposure to Herbicides. First Follow-Up Examination Results. Brooks Air Force Base, TX, US Air Force, 1987 63. Gough M: Dioxin, Agent Orange: The Facts. New York, Plenum Press, 1986 64. Austin DF, Nelson V, Swain B, Johnson L, Lum S, Flessel P: Epidemiological Study of the Incidence of Cancer as Related to Industrial Emissions in Contra Costa County, California, NTIS Publication No. PB84-199785. Washington, DC, US Government Printing Office, 1984 65. Smith AH, Waller K: Air Pollution and Cancer Incidence in Contra Costa County: Review and Recommendations. Report Prepared for the Contra Costa County Department of Health Services, 1985 66. California Department of Health Services, Epidemiological Studies and Services Section: Pregnancy Outcome in Santa Clara County, 1980-1985. Berkeley, CA, California Depart ment of Health Services, 1988 BFG39730