Tobacco Documents Online

100U28310'7

v R The American Council on Science and Health is ani independenteducational assocation promoting scientifically balanced evaluations of chemicals, the environment and human health. The Council is a nonprofit'association exempt from federal ihcometax underSection15011(c) (3) of the InternallRevenue Code. All contributions are tax deductible as provided by law. Ihdividual copies ofthis report are available at a cost of,S2.00! Prices for ten or more copies are available onirequest. Decemtler1978: firstiprinting February 1979: second printing August 1980: third printing,(revised) ,,.

' i This report on cancerin the United States was writteni by Elizabeth M. Whelan, Sc.D!, M.P.H., Terrence Smith, M.P.A. and Melvin A. Benarde, Ph.D. Dr. Whelan is Executive Director of the American Council on Science and Health (ACSH),,Mh Smith is a Research Associate and'Dr. Benard'e is a Director of ACSHand Professor of Epidemiology and Community Medicine at Hahnemann MedicallCollege and Hospital of Philadelpfiia. ACSH gratefully acknowledges the comments and contributions of the following individuals: Norman Borlaug, Ph.D. International Maize and WheaDImprovement Center Philip Cole, M.D., Dr. P.H. University of Alabamaat Birmingham Harry Demopoulos, M.D. New York University Medical Center Merril Eisenbud, Sc.D. New York UhiversityMedica! Center William R. Havender, Ph.D. Universityof,Caiifornia:at Berkeley Victor Herbert, M.D.,J.D. Bronx Weterans:Administration Medical'Center Thomas H. Jukes, Ph.D. University of California atBerkeley Gilbert Leveille, Ph.D. Michigan State University Robert E. Olson, M.D., Ph.D. St. Louis University Medical Center Ruth B. Schwartz, M.S. American Councilion ScienceandlHealth George Silvas, M.D. Continental'Insurance Company Henry F. Smyth, Ph.D. University of, Pittsburgh Fredrick J. Stare, M.D., Ph.D. HarvardiSchool of Public Health John A. Todhunter, Ph.D. The Catholic University of America Richard Wilson, Ph.D. Harvard University

1 J Contents Preface . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . page 2 Part 1: The Statistics . ... . . . . ... . . . . . . . . . . . . . . . . . page 3 Introduction . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . page 3 Cancer Statistics: Terminology . . . . . . . . . . . . . page 3 Cancer Statistics: Sources . . . . . . . . . . . . . . . . . page 3' U.S. Cancer incidence . . . . . . . . . . . . . . . . . . . . . . page 5; U.S. Cancer Mortaiity . . . . . . . . . . . . . ... . . . . . . . .. page 151 An Internatfonal Comparison . . . . . . . . . . . . . . . page 18' Part li: What Causes Human Cancer? ......... page 21 Tobacco . . . .. . . .. . . . . ... . . . . . . .. . . . . . . . . .. .. . . page 21 Diet .......................................... page 22 Alcohol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page 23. lonizing Radiation . . . . . . . . . . . . ... . . . . . . . . . . . . . page 23 Drugs . . ... . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . page 23 Sexual and Reproductive Patterns . . . . . . . . . page 24: Sunlight . . . . , . . . ... . . . . . . . . . . . .. . . .. . . . . . . . . . . page 24 Occupation . .. ..... .. . . . . . .. . .. . . .. .. .. ... . . . . page 24! Cancer and the Environment: An Overview ... page 25 References . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . page 2& Preface Cancer is the term given to a group of related diseases characterized by theunregulated growth and spread of abnormallcells. In 1980, about 785,000 Americans will learn they have cancer, and 405,000 will die of it. Cancer is a highly emotional subject which has also become the center, of politicaCcontroversy. Reports of the number of cancer cases and deaths can be misleading, rnisunderstood and misinterpreted. This report by the American Council oniScience and Health is an attempt to assess objectively the many facts and statistics produced each year by different publicbealth agencies. 2

Part I: The Statistics Introduction The 1970s witnessed a growingiconcern about the relationship between cancer and the environment. Frequent news articles and magazine features raised disturbing questions about the presence of cancer-causing agents (carcinogens) in our air, water, food'andiworkplaces. For example;,in 1975 newspapersacross1he country carried headlines declaring a dramatic rise inithe national cancer death rate in the course of a few months-a rise large enough to suggest thata cancer time bornb" was indeedlexploding, During 1976 and 1977, national, attention focused on New Jersey, quickly dubbed "Cancer Alley" as aresult of a governnnent study which reported it to have one of the highest, cancer death rates in the country. Television documentaries and magazine features often referred to an "epidemic" of cancern These various reports have led many to conclude that our nation is paying a high price for its sophisticated technological way of life. This section reviews statistical information about cancer in the Unit'ed States. It presents the most recent data on cancer incidence and mortality collected from several sources. Cancer Statistics: Terminology A thorough discussion of cancer statistics rmustibegin with definitions oficancer incidence andicancer mortality. Cancer incidence refers to the number of new cases of cancer diagnosed during a given year. Cancer incidence rates are the number of new cases per year for a fixed population. These rates are usually expressed as the annual number of new cases per 100.000 population. For mostof the U.S., it is not required that each new case of cancer be reported'tola government agency. Thus„exacffigures on the total number of'new cancer cases are not available. Instead, cancer incidence is estimated from an ongoing survey iniseveral states and urban areas which represent the nation as a whole. Cancer mort'ality refers to the number of cancer deaths reported for a given year. Cancer death rates are expressed as the annual number oficancerdeaths per1i00,000 population. Most deaths in the United States are reported and the primary cause ofdeath is usually specified. Because of this broader reporting~system. there is more accurate information on cancer mortality than on cancer incidence. Cancer Statistics: Sources The National Cancer Institute (NCI), Amenican Cancer Society (ACS) and the National Center for Health Statistics (NCHS) regularly publish national'cancer information. However, because each agency uses different methods of analysis and presentation theircancerstatistics are often not comparable. In the past, the National Cancer Institute estimated'cancer incidence on the basis of periodic surveys. More recently, the NCI~has cevelbped an ongoing program that continually reports new cases of cancer in several states and cities. The information collected from this SEER program (Surveillance, Epidemiology and EndlResults)iis then used'to estimate national cancer patterns. 3

0 The SEER'programis currently based on cancer dataifrom eleven geographic locations. These include five entire states (Connecticut. Iowa. New Mexico, Utah, and Hawa i), five large metropolitan areas (Atlanta, New Ortleans. Detroit„San Francisco, and Seattle) and the entire cornmonwealth of Puerto Rico. New Jersey and NewYork have also established cancer registries to monitor new cases andcancer deaths, but these data are not included in the SEER reportingisystem. The American Cancer Society also publishes annual information on cancer incidence. Using data from the SEER prograrn, the ACS estimates the numberofinew cancer cases thatwill occur in each state. These estimates canibe misleading, however, because they are not related to a fixed population. For example, the cover of the ACS report 1980 Cancer Facts and Figures shows the estimated number of new cancer cases that willloccur, in each state in 1980. California leads the country with an estimated 76,000 new cases and is followed by New York with 71,000 cases and Pennsylvania withi49,000. Because these estimates are noUrelated to their state populations. it is not clear if cancer occurs more frequently in these states than in others. When incidence rates are calculated, using state population estimates for 1975*. the order is reversed. Pennsylvania leads, with an incidence rate of 413 new cases per 100,000 population, followed'by NewYork with 393 and California with 358 new casesper100.000, Cancer death rates are published by the National Center for Health Statistics, the American Cancer Society and the National Cancer Institute. Each of these agencies uses,a different standard population forpresenting mortality data. The National Center for Health Statistics collects crude cancer mortality information and adjust6 ittothe age distributioniof the 19401Census population. Because cancer occurs pnmariiy among older persons, it is important to take into account the proportion of eachage group in computing a single death or incidence rate. Age aojustmg permits comparisons of rates from dlff erentlyears without the influences ofia continuously aging population, TheAmerican!Cancer Society uses the,informationcollected by the NCHS and prepares its own analysis of cancer death rates. The ACS also uses these data to estimate the number of cancer deaths,that will occur during the following year. But because these cancer death rates are not age adJWsted. they are not comparable with the~NCHS cancer death rates, In 1975 the National Cancer Institute published a report, U:S. Cancer I1,9ortalitybyCounty., 1950-1969 This study calculated cancer death rates for each county in the continental U S. according to sexandirace. These rateswereageadlusted to tfie1960 Census population„a different standardthan that used by '/ Bureau of Census, County and City Data 8ook 7977, Washingtoni D.C.: 1978. Ci 4

the NCH& These NCI data also give the death rate for the entire 20-year period'rfor each type of cancer as a single composite figure. While these rates permit intercounty comparisons, they cannot be compared with~either ACS or NCHS annual cancer statistics. U.S. Cancer Incidence' The American Cancer Society estimates that 785,000 new cases oficancer will bediagnosed in 1980. If current rates prevail, one in four Americans now living,will eventually develop cancer,, although manywill die°of other causes. FForthe 24-year period 1947-1971, the composite age adjusted cancer incidence rate declined slightly even though the total number of new cases increased. However„betweeni1971 and 1,976, thernost recent period f'or which data are available; age-adjusted incidence rates have increasedabout 1.3 percent annually. In general, cancer incidence has decreased among women and persons under the age of 45i Incidence has increased since 1947 among men and among~persons 45 and older. Figure 1 shows how the combine6cancer incidencerates have changed for men and women during the period 1947' 1976. Again~ the differences in age adjusting standards for each survey period, require caution in drawing any firm conclusions from this information. ~ace~ ~.',.Se+ S f.S~..... 'RGS E9711 Sr R 947-1u4e'969-. a,es a Pe^'aieS Nbnwnde aa,,,oc ~a.SSe^o'J.TS~ve,man(_-ce,o,1,er.ce an~moea~~!, e.~7s~m;neUne~ o,•.e5. ~~33•1974 -, P:=.,..: . . . ,•h 60~. ~5 '978 Poi acrF S a . . ra-cer ten ..a , onz~ •y. r•.ne U^1co S:; es 1i3Eo,~ 5.. J ~':a~•C.r,~coFinsrt'. _, i6~ ~cyt ~tl0 Sornetypes ofcanceroccur more frequently than others. Among men, lung cancer is now the mostcornmon,form followed by prostate and colon-rectum cancers. Among women, breast,canceris the leader followed by colon+rectum~and uterine cancers. Figure 2 shows estimates ofi the proportion of all cancers thaVoccur, at each major body site. Tabl'e 1 shows the changes in cancer incidence for selected body sites for the period 1947-1976. Incidence rates in this section are taken from the 1947-1948 Second National CancerSurvey(SNCS), the 1969-1971'Third~ National CancerSurvey(TNCS) and'the 1976 data of the SEER program, a

Pigure 2 19&OEsama'ed ~Cance' i nodence orS'eana Si Male Skin 2°b Oiah 5°`0 Lung 22°0CooonRectumr 14'~e Pancreas 3% Prostate 17'.d Uwary 9¢b~ Leukemia & Lymphomas9°'b AILotNer19°'0 SEI• ' Excuding non=melanomaskin cancer and earcmoma insrtu Source ArnencanCancerSocery.1980CancerFaersandFogures, New york 1979 As with all statistical data, several factors will influence the interpretation of these estimates. First, as knowledge aboutcancer increases, changes occur in medical diagnosis. (There are currently more than 100 known types of human cancer.) Past incid'ence surveys may have either under or over-reported certain kinds ofcancer because of these diagnostic cnanges. Second! the survey areas from which data are collected also change with time. For example, two southern cities with large, black populations (Atlanta and NewOrleans)iwene added to the SEER network in 1974'and 11976. As a result, the estimates of cancer incidence among blacks from the past surveys cannot' properly be compared with the newer ddta. Andias noted earlier, the SEER program acjusts its data to a diff erenti standard population thanithe earlleri National Cancer Surveys. These andl other methodological differences suggest that the incidence data for 1971-1976 rnaybe somewhat higher than theywould be if the pre-1i971 techniques were used. Therefore any small irocreases or decreases should be interpreted cautiously. Lung Since 1947 the greatest increase in cancerincid'ence has been for lung cancer. In 1947, white male lncidence was estimated at 30 per 100.000 population. By 1976, this rate had risen to78, an increase ofimore than 160 percenti(Figure3). P3,cr.^~ ~ - _, 6 4la'e5 ^Gerr,°PS 4.4 F erna re 2°0SKn 2'e OraP 27°. B~east 8°b Lung 15% Colpn Rectumi 3°. Pancreas 4', OJary 14°b Wterus 1^, U!hnary , , Leuwemia& Lym;nomas14'o anl o'ner 3 ~ 1t28 j-o anar_... . ,~s.n'ne3~ ._ ,,ar~r~.x F S „d r I'Ve~~,.I1i_p>45IL5 1 Jo9-7Fa 4

Table 1iEstiinated Cancer Incidence Rates pec100,000 Population by, Race and Sex„Selected Body Sites, SNCS (1947 48)'„TNCS (1969-71)", SEER (1976)'"' White Nonwhitet Cancer Site Total Males Females Males Femalas All Sites Combined SNCS 288.9 283.7 305:0 225,6 273 2 TNCS 277.7 309.0 256!8 3302 2311.5 SEERI 374 0 301.2 452.1 280.1 _ Lung SNCS 17.6 29;5 6.7 23 3 4'.9 TNCS 392 68.0 14.9 77 9 13.5 SEER 77.8 23.7 112.8 25.6 Colon SNCS 23 8: 238 26.0 13.7 11.9, TNCS 26.4 290 248 22 9 236, SEER 36.9 31.4 37 8 32.4 Rectum SNCS 166 20 7 1319 11.4 12 3 TNCS 12 2 16.0 96 130 7.6 SEERI 19.41 11_4 12 6 8 6 Breast! SNCS 0:9 716 0.2 50.4 TNCS 0!8' 733 0.7 53.7 SEER 83 5 66.7 P rostate SNCS 37.4 43 8 TNCS 45 2 68.6 SEER 686 1079 Bladder SNCS TN'CS 11.3 11.7 17.2 21 .3 7, 1~ 5.6 48 9 8 56 3.5 SEER 26.4 7:3 13.2 61 Uterus Corpus and Uterus NOS# SNCS 22 9: 25 0 TNCS 23.2' 1311 SEER 31 2' 15.2 Meianoma SNCS 27 26 33 06 04' TNCS 4.3 4.8 4 7 0 8 0.8 SEEP. 68 6 1 1 3 1 1 Pancreas SNCS 7 1 8 9 5.6 9.9 4.1 TNCS 8 7 10 7 6.5 14_1 82' SEER 11.5 80 177 11.1 Kicney- SNCS 4_0 5.2 29 4 8~ 25 TNCS 5.7 82 3.8 6 9! 3.3 SEER 96 48 9.0, 38 Leukemia SNCS 7.6 9;0 6 9' 92 30 TNCS 8 41 11,0 6.7 87 5.3 SEER 13.1 71 9 2 8.1i UterireCervix SNCS 38 3 74.6 TNCS 1 5 1 31 3 SEER 10 6 2614 Stomach SNCS 252 324 17.8 38I6 18.9! TNCS 9 1 1211 58 18',6 7.9' SEER 126 58 2013 9.2 SNCS Second National CancerSurvey. 1947-1948;,rates adjusted to age d.stribution of 1950:Census population "TNCS! Third National Cancer Survey. 1969-1971. rates adjusted to age distribution of,1950,Census population. "'SEER Surve uance. Epidemiology and End Results, 1976, rates adjusted to agecistnbution of 1970 Censuspopulation. t: Nonwhite rates for SEER program are for bVack populationionl~y, $ Not otherwise specifred Source: Devesa, S S. and D L Silverman. Canoerincidenceandlmortalitytrenesin the Urnted States; 1935-19741 J. Matl: CancerInstif. 60(3)545. 1978' PoliacK, ESand JI W Horm. Trends in cancer incidence and mortalrtyin the United States, 1969-76, J IVatl, Caneerlnstrt„ 64(5)1091, 1980; ..e

. Anieven greater increase has been noted among nonwhite males. InI1947; non,vhite,rnale incidence was estimated at 23 per 100,000: by 1976 it had increased to 113per 100.000, or more than 500 percent These large increases are beiieved to be due to the effects of cigarette smoking, although some of the increase for nonwhites may be explained by a lack ofi medical care in the past. Nonwhit'es hadifewer contacts with the health care system in 1,947than theyddafter legislation of Medicaid ini 1965. This suggests that:1the number of new cases identified in 1947 was lower thanot should have been. Lung cancer among women has also steadily increased since 1947. The American Cancer Society estimates that by 1983, lung cancer will replace breast cancer as the leading cause oficancer deaths among women. Among white females, lung cancer incidence increased from 7 in 1947 to 24 per 100,O001irn 1976. Among nonwhite females, incidence rose from 5 to 26 per 100,000 in 1976. Atipresent„lungicancer incidence is increasing at'a faster, rate for females than f'or males: Indeed, a substantial proportion ofithe overall increase inicancer incidence during the 1970s can be explained by the dramatic surge in female lung, cancer cases This increase is also believed due to the increase in female cigarette smoking following World War II. Colon Colon cancer incidence among white males increased from 24 per 100.000 in1'947 to37n1 976+Figure 4). Among nonrn+hitemales. incidence increased from 14 in,1947 to,38'per 100.000 ini 1976: Nonwhite femaleincidence also increased during this period from 12to 32 per 100,000. .~ M5 ec oF tm3es No~a.haa . . ., ___. ..._. . C Sa^:~ ., ... . ~._ . ~ . < . - . . . .c U 97E Colbn cancer incid'enceamong white femalesdecreased~slightly during the Deriod 1947-1971ifrom 26 to 25 per 100,000. Since 1971, this incicence has increased to 31 per 100,000.