Document 8VE9VOjN6DbgZE3Xw6qg63vEk

Nature Vot. 28& IS January* J98J 127 Fallacies of lifestyle cancer theories Samuel S. Epstein* & Joel B. Swartzt * Professor of Occupational and Environmental Medicine t Assistant Professor of Occupational and Environmental Medicine, School of Public Health, University of Illinois at the Medical Center, Chicago, Illinois 60680 Self-interest of the chemical industry apart, the lifestyle theory of cancer causation is held to reflect misinformation and `inactive conservatism'. PETO's article1, purporting to be a review of the book The carcinogens, such as vinyl chloride and acrylonitrile. It is Politics ofCancer, is largely a restatement of the lifestyle theory reasonable to anticipate that greater production has been paral of cancer causation. This theory postulates that if you get cancer leled by increased exposure of increasing numbers of both the it is essentially your own fault, and that the causal role of past workforce and the general public, which is likely further to involuntary exposure to environmental and occupational accentuate increasing trends in cancer rates. It must also be carcinogens is trivial. Mot surprisingly, the lifestyle theory has recognized that before the 1976 Toxic Substances Act, which emerged as the major professed basis of the chemical industry's the chemical industry so effectively stalled for so long", there objections to the regulation of its carcinogenic products and were no requirements for testing chemicals before their intro processes1. As an enthusiastic proponent of this theory, Peto duction into commerce (with the exception of special-purpose asserts that smoking-derived and fat-associated cancers legislation for drugs, pesticides and food additives). Thus, the "collectively account for more than half of all cancer deaths.'' As overwhelming majority of industrial chemicals now in use have a corollary, of Peto's emphasis on lifestyle factors, he denigrates never been tested for chronic toxic and carcinogenic effects, let the role of occupational and environmental carcinogens and the alone for ecological effects. need for their effective regulation, claiming that there has been no recent increase in cancer mortality rates other than that due Role of smoking to smoking. We shall demonstrate that there is scant scientific As emphasized in The Politics of Cancer' (p. 178), "Smoking is basis for the lifestyle theory, and that it is in fact contradicted by the single most important cause of lung cancer, as well as of a substantial body of published evidence. cancer at other sites, chronic bronchitis and emphysema, and Cancer rate trends cardiovascular diseases". Less well appreciated by lifestyle advocates is that overemphasis on smoking is widely used to Peto justifies his emphasis on lifestyle factors by dismissing divert attention from occupational causes of lung and other evidence for recently increasing cancer rates, apart from "that cancers. Of the approximately 100,000 annual lung cancer due to the massive effects of smoking on lung cancer". However, deaths in the United States, at least 20% occur in nonsmokers. Ot there is substantial evidence to the contrary. Standardized is relevant that lung cancer death rates in nonsmokers approx cancer death rates, adjusted to the 1940 age structure of the imately doubled12 from 1958 to 1969, an increase maintained total United States population, show a progressive overall since. 'Furthermore, the role of occupational exposure to increase of about 7% from 1935 to 1970 (ref. 4) despite marked carcinogens was not recognized in most of the classic epide reductions of stomach cancer rates for unexplained reasons and miological studies which linked lung cancer with smoking. This of cervix cancer rates for reasons including the frequency of led to overestimation of the contribution of smoking compared elective hysterectomy for non-malignant disease and the success with occupational risks or to their possible interactions. ' I of screening programmes. These trends are consistent with i standardized mortality data for the United States (Table l)1, Table 1 Age-adjusted cancer mortality rates per 100,000 US population tor where they are even more marked in black males, and with crude selected tiles by tea and year 1969-76, and avenge per cent chants' 1 mortality data for the United Kingdom (Table 2)*. The overall rate of increase in US cancer mortality in the 7-year period from 1969 to 1976 (5.5%), adjusted to the 1970 age structure, is Morality rate per 100,000 Avenge % change 1969-76 substantial and comparable with that for the preceding 35 years, Site Sea* 1969 1976 Annual 7-Year 1935 to 1970 (7%). The overall increase in incidence rates is even more marked than mortality rates in the past decade, involving a wide range of organs besides the lung (Table 3)s. Moreover, the increase in incidence for all sites is comparable with that when lung cancer is excluded (Table 4)TM. Reliance on overall age-adjusted incidence or mortality rates alone is simplistic, as such rates can mask steep increases in organ-specific cancers in high risk population subgroups, such as asbestos insulation workers or menopausal women treated with oestrogen replacement therapy. The overall probability, at All sites Stomach Colon Rectum Pancreas Lung WM WF WM WF WM WF WM WF WM WF WM WF 19S.0 129.0 .10.6 S.3 18.7 16.2 6.9 3.9 11.0 6.6 55.0 10.2 210.2 133.8 8.7 4.1 20.7 16.S S.6 3.2 11.0 6.8 66.7 17.8 0.9 OJ -2.9 -3.6 1J 0.0 -3.0 -3.1 0.0 0J 2.6 7.6 7i8< 3.T -17.4 -22.6 10.7 1.9 -18.8 -17.9 0.0 3.0 21.3 74.3 today's death rates, of a person born now getting cancer by the age of 85 is 27% for both men and women; this is increased from the 19% for men and 22% for women born in 1950 (ref. 9). Furthermore, recent cancer rate trends reflect exposures and Melanoma Breast Cervia Uterus WM WF WF WF WF 2.0 1.4 26.2 S.3 4.6 2.6 1.6 27.2 3.9 4.2 4.0 0.8 0J -4.9 -1.7 30.0 14J 3.8 -29.1 10.5 events beginning some 20 or 30 years ago, when the production Prostate WM 19.0 21.0 12 8.7 of synthetic organic chemicals was relatively trivial compared with the present levels. The production of synthetic organic Bladder Kidney WM WF WM 7.1 2.1 4.3 7.5 0.6 5.6. 2.0 -1.4 -4.8 4.5 0.6 4.7 ! compounds in the United States in 1935, 1950 and 1975 was about 1, 30 and 300 billion pounds per annum, respectively10; sharp increases have also been observed for a wide range of derived industrial products such as chlorinated hydrocarbon solvents, plastics and resin materials, and of industrial Leukaemia WF WM WF 2.0 9.4 5.7 2.1 0.7 5.0 9.2 -0.4 -2.1 S2 -1.7 -8.8 For age adjustment the 1970 United Stales population was used as standard. * WM, white male; WF, white female. Cl I Ml InunskfjU 1 Mature Voi. 289 IS January 1981 \ Table 2 Crude cancer mortality rate* per 100.000 population lot selected lites by tea and year. England and Wales, 1971-77, and average par cent change* Mortality rate1 . per 100,000' . Average % change 1971-77 Site All sites Stomach Large intestine and rectum Respiratory Breast Uterus' Prostate Bladder Leukaemia Sex M F M F M F M F F F M M F' M F 1971 26S 215, 30 21; 30i 34 106 22 4Sl IS! . . n\ ,2I 41 710 SIS 1977 Annual 283 ; r- `1:1, 233 iJ 28 . -12 19 -1.7 32 i.i ' 3S V. 0.4 112 0.9 29 5b ` 47 0.7 to : 1.1 ;i9 , lb 12 0.0 S.t 3.7 7b 0.4 > 6.0 Tb': ' 6-Year v 6.8 ' 8.4., -6.7 . -9b 6.7 ' 2.9 S.7 31.8 4.4 6.7 11.8 0.0 24.4 2.9 9.1 Table 2 Age-adjusted -cancer incidence rates per 100.000 US population . (whitea) tar lelected sites by tea anti year. 1969-76, and average per cent change* Incidence rate per 100,000 ' Average % change . 1969-76 , Site All sites Stomach Colon ' Rectum Pancreas Lung Melanoma Breast Cervix' Uterus Ovary Prostate Bladder Kidney.. . Leukaemia . Sex M F M F M F M F M F M F M F F F F F M M F M F M F 1969 .- . 1976 Annual > 7-Year 346.6 * < J1 374.0 1.3 7.9 271b ; 30131 2.0 10.9 15.4 12.6 * -2J ' -18b 7.1 , S.6 -3.7 --21.1 34j 36.9 ' lb 7.0 30.6 31.4 0.7 2b 17b 19.4 lb 10.9 11.L ,11.4 . lb , 2.7 12.1 lib -0J -S.0 7b 8.0 0.9 6.7 70.6 77J 1.4 - 10.2 13.3 23.7 8.6 * 78b . M 6b .: 6b - 54b 4.1 6.1 ' 6b 48.8 . 73.9 - sjb ib 13.0 16b 10:6 1 v'-5.9.' " -33J 22.6 3L2 5b 38.1 14.9 13.6 -0.4 8.7 ` 59b 68.6 2J 16.3 23J 26.4 2b 10.9 - ` 6b - -- 7b - , '2b ' 15.9 9.0 /9.6 1 lb ' . 6.7 43 4.8 . IJ 11.6 13J 13.1 . -0b . -0.8 . 8.0 - 7:3 -lb"'"- :-llb The 1970 populetionwaa used eattaadard tor ige*djastaMat.: ., due to smoking than to occupational carcinogens15'1*. "In several instances where the risk of bronchogenic carcinoma has been shown to be increased among occupationally exposed ' groups, there has been an accompanying shift in the distribution , of histologic types: of tumours", away from the small-cel) undifferentiated and squamous cell carcinoma of the lung, the : principal types whose frequency is increased by smoking, in the direction of other types, particularly adenocarcinoma . "This (shift) has been noted among metal miners, uranium miners, copper smelter workers, vinyl chloride polymerization workers, chioromethyl methyl ether production workers, and mustard gas manufacturers" (ref. 16). Possible variations,of smoking patterns fail to account for the marked excess in US lung cancer rates identified in specific occupational exposures, particularly among ethnic minorities and migrants from southern states". A further challenge to the dominant role ascribed to smoking seems to be provided by observations that the risk of lung cancer in certain occupational groups,' such as American Indian uranium miners'*. Swedish zinc-lead miners'*, mustard gas workers10, copper smelters exposed to arsenic1', and chioromethyl methyl ether workers11, ` is about as high' among nonsmokers as smokers, although the latency period is reduced in smokers, suggesting a possible promotional effect of smoking. It appears that the relative risks 1 of lung cancer for smokers as against risks for nonsmokers may have been overestimated, particularly in less than lifetime stu dies'*. Variations in smoking do not account for geographic excesses in lung cancer rates in US males and females, which overall reflect proximity of residence to petrochemical and certain other industries"1*4; there are also data showing asso ciations between levels of atmospheric carcinogens and lung cancer mortality rates". It may be noted that's report14 from Peto's own institution demonstrates that the correlation coefficient between lung cancer and smoking internationally ; explains only one-third as much of the variation as does the correlation between lung cancer and solid fuel consumption (0.4 versus 0.7; r*-0.16 versus 0.49). Overemphasis on the carcinogenic effects of smoking, and ignoring or discounting the role of occupational and other exposures, is extended by Peto and others to cancers of the TaMe4 Changes in US cancer incidence rates bom 1970 to 1975'-* _ ' .. Average %i in inddence ratea, 1970-75 * V - . * ' Cancers of aO anas - - Cancers cfaH sitae except tung . . Thus, "we are unable to say how mud) of the risks attributed, ' to cigarettes is a `pure' tigarette risk and how much is cigarette < times another, possibly on the job hazard"*. Moreover, smoking and occupation are confounded variables, smoking among men being more prevalent in `blue-collar' workers than in profes sional and managerial dasses'*, Occupational causes of lung cancer include asbestos, radon daughters, nickel ores, chromium, arsenic; beryllium, mustard gas, vinyl chloride and bischioromethyl ether, apart from incompletely identified^ carcinogens in a wide range|of|industries such as rubber curing,' tanning, steel (coke ovens); foundries, automobile, and petto-! chemicals. Thus, lung cancer rates in asbestos insulation and . topside coke oven workers are as modi as 10 times greater than general population rates, j ,; Underestimation of the role of such occupational carcinogens has been assisted by the fact that lung cancer mortality rates, based on the Internationa] Classification of Diseases, fail to distinguish pleural | mesotheliomas from lung cancers; there is evidence of substantial under-reporting of mesotheliomas (by about 75%) in high risk groups'4, and even more so in occupa tions, such as automobpe mechanics, where asbestos exposure has not been well'recognized!. There is a further lack of. dis tinction between lung cancers of different histological types, some of which, sudi as adenocarcinomas, are less likely to be Croups White male 1 Non-white male White (emale Non-white lemaie ' Annual 0.9 2J 2b 6.1 S-Year 4.7 11b 11.6 34b ' Annual . 0.9 ` 2.7 lb . 5b 5-Year ` 4.6 14J. 10b 32b Table 9 , International conditions between bread, eotan and bvtt poaaibie aetfetogical varubtas'* Correlation twffdus ;, ' * . Consumption of Gtoas National Total eaarm - ot (at animal proteia Plodact ptodaotao Breast cancer tuodtuo - 0.79 ' Mortality 0b9 Colon cancer (M) Incidence ` 0.74 Mortality ' ' 0.85 Colon cancer (F) Incidence ; , - 0.78 Mortality . 0J1 Liver cancer, (M) Incidence . -0.49 Liver cancer (F) Incidence -Ob9 , 0.77 0J3 0.74 .0.86 0J0 0.S4 -0b9 -0.67 . 0.83 0.72 ; 0.70 040 0J1 0.77 0.68 0.69 082 0.69 -4M2 . 0.67 0b2 - -ObS* -0b3 -0J1* Tiquidcncrg]t\ ' . - " ' ' ,-r- Mature vol ZH9 15 January 19S1 129 bladder and pancreas which are variously characterized as generally far in excess of the dietary differences observed amon g related to or caused by smoking*7,2*. However, the relative risks the various human populations studied22. These experiments for these cancers are several times less in smokers compared invariably failed to adjust the intake in controls of fat soluble with nonsmokers than is the case (or lung.cancer. Excess bladder carcinogens, present in fat as accidental environmental cancer rates have been identified in several occupational cate* contaminants, to reflect variations of fat intake of test animals. gories, including rubber, paint manufacturing and textile dyeing Peto's claim for the causal role of dietary fat in human cancer workers", and among residents in highly industrialized coun overstates the conclusions of those cited as the basis for his ties20, particularly those with large chemical industry claims. Armstrong and Doll2*, for instance, merely suggest that complexes21. Excess pancreatic cancer rates have also been dietary fat levels may influence the incidence of colon and breast reported in various occupations including steel and metal cancers, withoutasserting causality. Doll considers that diet may workers22 and organic chemists. act by modifying the incidence of tumours induced by i Recognition of the important role of occupational exposures carcinogens or by acting as a vehicle for exogenous in lung cancers previously ascribed, exclusively or largely, to carcinogens2*--a suggestion also made in The Politics of Cancer ; smoking in no way detracts from the recognition, emphasized in which Peto dismisses as "implausible". Carrol concludes that j The Politics of Cancer, that the impact of smoking constitutes a "although caloric intake may be a factor in human carcino "national disaster". There is no basis for regarding the smok* genesis, it does not appear to offer a practical approach to the ing/lifestyle and occupational theories as mutually exclusive, problem"27. As recognized by current concepts on the multi- particularly as these exposures may operate interactively. factorial aetiology of cancer, there is a substantial probability Furthermore, lifestyle is a somewhat misleading rubric for that a wide range of influences, diet and other lifestyle factors smoking as it restrictively implies voluntary personal choice. included, modify individual responses to carcinogenic agents. Placing responsibility for personal choice of an addictive lethal To ascribe causality to any particular modifying factor requires a habit on young teenagers, the fastest growing group of new degree of scientific evidence that has not yet been presented for smokers, seems inappropriate. Failure to control smoking dietary fat reflects a wide range of political and economic constraints, including massive press advertising by the industry which omits Role of occupation the word `death' from the guarded small print warning of Peto associates himself with the insistence by the chemical danger, massive revenues to federal, state and local government industry2 and other lifestyle proponents that occupational from tobacco taxes, federal subsidies to the industry and exposures account for about 5% (refs 38-40) or "a very small unwillingness of governments to increase tobacco taxation or to proportion"41 of all cancers. This view is based on ascribing develop incentives to tobacco farmers to diversify. It is. also given percentages to known or alleged lifestyle factors, including important that the industry has moved to open up massive new smoking, fatty diet and sunlight, leaving a small unaccounted for markets with high*tar cigarettes in less developed countries, residue to which occupational factors are arbitrarily assigned by where the population is poorly informed on the hazards of exclusion. The authors of this simplistic hypothesis compensate smoking. for its tenuous basis by reliance on `educated estimates' and by Role of diet making circular references to each other, often by `personal communication', as the responsible authority. Lifestyle proponents are on less sure ground when they bracket However, there are problems with such `guesstimates'. First, diet, excess fat and ovemutrition with smoking as the causes of they fail to consider the multifactorial aetiology of cancer and the majority of cancer deaths. This claim is based largely on the role of multiple causal agents, such as asbestos and smok international correlations between consumption of total fat and ing42; thus, the summation of known causes of cancer should rates for cancer of the breast and colon2*; however, such cor properly exceed 100%. As one of the lifestyle authors recently relations by themselves are not proof of causality. Similar stressed2*, "there is now strong evidence to suggest that the risk correlations were found, in the same study from Peto's institu of cancer is commonly increased by interaction of two or more tion, between breast and colon cancers and other variables, such factors". Second, current cancer rates reflect exposures 20 to 30 as Gross National Product and consumption of animal protein, years ago, when production levels of occupational carcinogens which also appear to reflect industrialization2* (Table 5). were a small fraction of the present; such estimates should thus Furthermore, "epidemiologically, the case against fat is weak now be adjusted to reflect increasing numbers of workers because there are populations that have a high fat intake and exposed. Third, the authors of these guesstimates failed to little bowel cancer..."21. Of two case control studies on the consider the very limited nature of the data base on exposure to association between diet and breast cancer, one found no occupational carcinogens. Nor have they at any stage protested effect24 and the other found trivial effects of fat and calorie or even commented on the persistent refusal of the chemical intake, concluding that "... recommendations of major dietary industry to make such critical data available. In the absence of modification as a possible preventive measure for breast cancer exposure data, it is even less clear how the Tifestylers' are clearly premature"*5. confidently arrive at their estimate of less than 3%. Equally unconvincing are the studies, cited by Peto as cor Rather than addressing himself to such problems, Peto dis roborative evidence on the experimental effects of diet, which misses recent estimates of the importance of occupational were largely concerned with the influence of fat on the incidence carcinogens in a report by the US Public Health Service42 as of tumours induced by chemical carcinogens and ionizing radia exaggerated, unsound and unreasonable. This report, prepared tion, aad the influence of caloric intake on the incidence of by nine named and internationally recognized experts in cancer spontaneous and induced tumours. Not only were different epidemiology, statistics and carcinogenesis from three federal variables defined in the animal and human studies--per cent fat research agendes, is based on a National Occupational Hazard in the diet and total dietary fat, respectively--but increasing fat Survey which between 1972 and 1974 surveyed nearly 5,000 levels in the animal experiments were associated with increased workplaces chosen to provide a cross-section of industry in the incidence of skin, liver and breast cancers, whereas the reported United States. The report estimated the total number of workers correlations between fat consumption and liver cancer mortality exposed to asbestos, nickel ores, chromium, arsenic, benzene are negative for both men and women (Table 5). Moreover, and petroleum fractions, including aromatics. The excess these experiments often failed to differentiate between varia cancers attributable to each of these carrinogens were derived tions of total dietary fat and caloric intake in test animals and to by multiplying the number of exposed workers by known risk adjust caloric intake in controls to reflect dietary fat variations in ratios and subtracting the "normal inddence" of the cancer. test animals; the magnitude of the variations in fat and caloric The report conduded that "as much as 20% or more" of intake required substantially to influence the incidence of cancers in the near term and future may reflect past exposure induced and spontaneous tumours in experimental animals is to the six cardnogens considered. The uncertainties and limitations in these conclusions, including the possibility that exposures and risk ratios may have been overestimated in some instances, were cieariy stated in the report, as were other considerations including the multifactorial aetiology of cancer, and the role of lifestyle factors and their possible interactions with occupational exposures. The possibility that this government report underestimates rather than overestimates the role of occupational exposures, for several reasons some of which are recognized in the report, has not been considered by its denigrators, including Peto. First, the calculations'in the report ignore the role of radiation and of some ten epidemiologically recognized occupational carcinogens, other than the six considered. Second, the risk ratios considered may be artificially low as they were largely derived from less-than-lifetime epidemiological studies, which may thus underestimate the true risk in view of the long latencies commonly involved. Third, the report does not consider the many statistical and methodological constraints common to most occupational epidemiological studies44 such as relatively small numbers of workers in many locations, changes in exposure patterns over time due to employee turnover, plant shutdown, process and production changes and changes in management, all of which lead to fragmentation of health and exposure records, access to which is often restricted by industry. Fourth, the estimates fail to take account of the many chemicals recognized as carcinogenic in animals for which there are no exposure or epidemiological data. Thus, of 442 chemicals and industrial processes recently evaluated by the International Agency fot Research on Cancer (IARC), epidemiological data are available for only 60 (14%), although evidence of experi mental carcinogenicity was considered to be sufficient for 143 (32%)4'. Fifth, the estimates exclude high risk occupations with incompletely defined carcinogens, such as the steel, rubber and tanning industries. Sixth, the estimates do not adequately reflect conditions in small business where exposure levels are likely to be higher than in major chemical companies. Seventh, the report does not reflect major increases in the production of the occu pational carcinogens it considered such as benzene, with the likelihood of recently increasing exposures. Eighth, the study examined only a limited number of sites, excluding cancers such as skin and bladder which are known to be occupationally related. Finally, the estimates neglect the possible role of fugi tive point-source emissions of industrial carcinogens as causes for the excess of overall and organ-specific cancers, including lung, bladder, colon, pancreas and breast, in residents of certain highly industrialized counties. This government report has received extensive support from various expert bodies, such as the Toxic Substances Strategy Committee, whose position has been endorsed by 17 federal t. Faro, R. Nearer 224,297-SOO (1920). I Efwn.St Ttn Fatidn of Ctma (Siam Gob Boofca. Saa Fraactee, 1972. ratree* md tapanea* in Arteer/DeuMtOay, Na York. 1979); qtattPeta rotor to the 1979 edition. S. AIHCAnammtadtd Alternator, ae OSHA't Geaertr Ceneie|ia arpmt Sniffy a Mrate Admraimw (OSKA) Docker Mr. H-090) (Awrtcaa Internal HathIt Council, 24 Fihnnq. 1971). 4. Via*/Sterner Spent/Ripen* C/.1.PMEW.4J. 14J(lP5*h VOatSmtmarfA. US.l (1970). 5. Pollack. E. S a Hera* J. W.S. aaai. Caacerkaat *4,1091 (1990); 1. rm4i in kJanalirr. ItSt-lFfS (UJC. Oin at 7ap,l,hri Caaatea and Sterna. Mn DHl.Ne. 2.1972). 7. 7W Marten/ Cantre Stem; 19*9-/97J (MCI Modal: Ctear forfilter, Epr- drornlngy and Cad Xfte (SEER) Program. 2. Iteaidaraita. M. Omoanraat Safay a Meat* MaMate OSHA Decker 090 (4 April. 1971). 9. ZdaS. M. S Am / Epidrmte 12*. 1 (1*77). 10. Date. DL4 Magee. I. H. Sctear Hi, 11M (1979): aaa alia US larrmaboaal Trade Coaieiiraioe Report* 11. Chrmjtmi Oaaprri ia dr Wriplm S4di Raport at $r Com runn an Oaurmnaai Operation* Haw at Rrpnaaaatea (Sapfrarbar 27.1971). 12. Erntrao* 1. E. J. near Caarrr/art 11.7SS (1979). 12. Sirriinp. T. D. lai J. MM Soaker 2. 4J7 (1971). 14. Keteow. ML* Wapnar. i. C At. J. lad. tdrd. 24. S0211949). 15. Wyndrr. E. L * SleKraan. S. D Cearer Re* 27,4101(1977). 11. Wapanrr. J. X, In/anit. P. F. * Sayte. D. L Eaair. Atl 21,15 (1920). 17. Minrwr. T. F. Jt Sicrtmp. T. D. /. aaaa md An. 17,107 (1975). II. After. V. E.. Giliaat. J. D. * Wapanar, J. X. Aaa. M. V. Arad Sri 271, 210 (1971). 19. Aarhoa. O. * StradeB. L Stand / Warl mair. MM < 41 (1971). 20. Yamada. A. Aria perk. Jap. U. 121 (1912). 21. Pinia. S. S. Anrki rarer. MM 22. 225 (1971). 22. Waaa. W. * Boucar. K.A.J.Anmtd. Ark 124,1129 0975). 22. Meant* J. T. * McKay, F W. AriaiWCanrarAdmlny/W US Caaadar: /959-1949 (Dap, H tW., Waikinftan. DC. 1975); Adar of Cancrr MenaIky fa US Canaan aawaj t/j: Man-WSkrt (Dap, H E W, Waahinfian. O.C. 1971). igencies, and international groups, such as the International Labor Organization, and the L'S and British trades union. The report has also received additional support in the critique of two consultants to the chemical industry's American Industrial Health Council which concluded that",.. the full range (of total cancer attributable to occupational exposure) using multiple classifications may be from 10 to 33% or perhaps higher if we had better information on some other potentially carcinogenic substances___ The annual number of cancer deaths attribut able to asbestos is in the range from 29,700 to S4.000, which corresponds to a percentage range of the total cancer of 7 to 14%-----Any argument over these numbers cannot detract from the fact that asbestos exposure was, as the authors (of the Government report) state, a major public health disaster___ We also believe that reduction of exposure to carcinogens in the course of employment can certainly be expected to affect major reductions in the frequencies of occurrence of cancer and is one of the most promising applications of preventive medicine"41. The American Industrial Health Council failed to release this critique until the record of the recent Occupational Safety and Health Administration hearings on regulation of occupational carcinogens closed. Finally, there is no basis whatsoever for recent unsubstan tiated allegations by Peto and others that all or most authors of the government report have disowned or rejected it or its conclusions (K. Bridbord, M. Schneiderman and A. Upton, personal communication). It should be further emphasized that this 50-page report was prepared as a government document specifically for inclusion in public hearing records, and not for submission to a scientific journal. Conclusions Cancer is a disease of multifactorial aetiology to which occupa tional exposure and smoking can contribute importantly, some times interactively. There have been substantial recent increases in cancer rates which cannot be accounted for by smoking alone. Smoking is the major lifestyle factor of importance in cancer, and evidence for the causal role of other lifestyle factors, particularly diet, is slender. The role of lifestyle factors has been exaggerated, by those with an economic or intellectual invest ment in this theory, by largely excluding involuntary exposures to carcinogens and minimizing the role of occupational carcinogens. These considerations further illustrate the primary thesis of The Polities of Cancer: cancer is essentially a prevent able disease which requires intervention and regulation at several levels, particularly the occupational and smoking. Failure to prevent cancer reflects major political and economic constraints which have hitherto been largely unrecognized or discounted. 24. BM. W. J.a at Seine 191,51 (1977). 25. Canter. B. Cte MIA Ampart 22,17 (1972). 24. Arawront. B. K. * Dell. R. /. 1. Cam IS, 6)7 (1975). 27. Celt, F, Monte. X. R_ Halting. R * FriadaU. G. H. Men Cagi /. AM 224,129 (1971). . 22. Wyadtr, E.L.Mabu*H.K~Marach.N. k Fortner, J.O. Cancer 21, *41 (1972). 29. Celt, F. * Goldman, R. Aritani at High Mat at Conor. Am Appmtt) Cttttr Ftriif) at* Cm*/ (a*. Fninan, J. F. Jr) 147-124 (Aadaaik, Nat, York. 1975). 20. Bio* W. J. * Fraamaai. J.Z ate Geacw/aat 41,1017 (1972). 21. Heew. JL * Fraian*ni.J. Eaair. Re*9.194(1975). 22. Milham. J. la Orctptitea/ Cattatgtatnt led* Wapt tar. J. a State. U.) 242-249(Haw Vert Academy o/ Sdenea* 1974). 22. t w ;- --- a- rr fr-r<-frVrfT and Canto/ (a*. Frawneai. J. F. Jr) 201-224 (Academic. Near York. 1975). 24. FMKpa. R. L Caacar An. 25.2512 (1975). 25. Mdte. A. B. cedt Am./Epidrmaet 197.499(1972). 24. DoQ, R. Career Rat. 45,2475 (1920). . . -- 27. Cared. K. K. Cater An. 25,2274 (1975). 22. Hvsteea, J. * Mak. C.I.J. atm Gaarrr Jan. 42.129111979). 29. Wyader, E. L * Oort. G. B. Art. / Fu*i MIA 44, 259 (1977). 40. Maafh, T. H. Srtecr 205,1242 (1979). 41. Doll. R. Maruar 245,529 (1977). 42. StKkoff. 1. J. * Hamate*. C. E. /. Am. mrd. An. 242.452 (1979). 42. Bndbord, X n a/. Frontier tf At /rate of Grata m the Umitrd From Rater* a Otcapantasl From (National Cater, literal*. Natioaal Inatiniic at Eartremaaaa) Haahh Sdtaati, tad Naiienal latutalt lor Ozupadonal Salay aad Health, 15 Septem ber. 1972). 44. Siellmaa, J. M. * StaUman, S. D. HAei AaperatealGaarree AiateaaNr ae Oticepanea; Steaace)tadSatie/ Coaiteroate (Aiecncart Sotety lor /mtimaOraoiety. Qiicaae. 4-7 Mate. 1920). 45. Report at a* (ARC Warkmq Group Ctata An. 44,1 (1920). 44. Foak Chtmictlt aad PrMtc Meatam iReport to lbe PrtaiOtni at the Teak Sataiaaaaa Skaiasy Cemaiintt at rha Couacil on Eevirensanial Oualrty, May. 19501 47. Satloaee. R. A. * Derma. T. A Ctvittl Aenrar of Badtard. tC. er at prepare* tar the Amanata ladaatrial Htahk Couacil (1979k ___