GENERAL CHEMICAL AND EXPERIMENTAL DATA ON CARCINOGENESIS AND TOBACCO SMOKE polynuclear Aromatic Bydromrbons As criteria for the presence of polynuclear aromatic hydrocarbons in tobacco smoke, the list of J. W. Cook (m) has been widely accepted by tobacco chemists. The Surgeon General's 1964 Report and Cook's paper are in agree- ment with respect to the presence of benzo(a)pyrene (3 : 4benzopy- rene) , dibenz (a,h) -anthracene (12 : 5,6-dibenzanthrac) : benzo(c) phenanthrene (3 : 4benzophenanthrene), and dibenzo ( a,i) pyrene (3,4 : 9,10-dibenzopyrene) , all having carcinogenic activity. Cook considers, furthermore, as identified : Benz (a) anthracene (1,2- benzanthracene) marginal carcinogenic activity; chrysene, benzo(e) pyrene (1,2benzopyrene), questionable carcinogenic activity ; benzo (g,h,i) perylene (1,12benzoperylene) )2 benzo( b) fluoranthene (3,4- benzofluoranthene) carcinogenic (69,106)) and benzo (j) fluoranthene (10,ll~benzofluoranthene) carcinogenic (206). Indeno (l&3-cd) pyrene (2,3-phenylenepyrene) has since been iso- lated from tobacco smoke (@). This polynuclear aromatic hydrocar- bon was found to be carcinogenic (44,59). The following carcinogens, or questionable carcinogens, were isolated by Kiryu and Kuratsune (55) in the smoke of cigarettes smoked by human volunteers: benz (a) anthracene, chrysene, benzo (a) pyrene, benzo (e) pyrene, benzo (b) fluoranthene and benzo (k) fluoranthene. The carcinogenic polynuclear aromatic hydrocarbons are regarded as the major initiating car- cinogens in tobacco smoke. N-Eeterocyclic Aromatic Hydrocarbons The Surgeon General's 1964 Report lists as carcinogenic compounds three N-heterocyclics, dibenz ( a, j ) acridine, dibenz (a,h) acridine and 7 H-dibenzo- (c,g) carbazole. An independent investigation has con- firmed the presence of the first named compound in cigarette smoke WV * N-Nitrosa7nims N-nitrosamines are among the most powerful known animal car- cinogens. Since tobacco smoke contains secondary amines (67, 71) `Dibenzo (a,h)anthracene in the Surgeon General's lQ@ Report should be replaced by dibenz (a,h ) anthracene (Ed). * Benzo(g,h,i)perylene was not tested for carcinogenicity until 1966 and then was found to & inactive (4). and most tobaccos, certainly Burley and Maryland varieties, contain nitrates (64)) tobacco smoke can be considered as a potential environ- ment for the formation of N-nitrosamines. The major nitrates in tobacco are alkaline nitrates. Neurath, et al., isolated three aliphatic N-nitrosamines from the smoke of a cigarette rich in volatile basic components and high in nitrate content. One of them tentatively has been identified as methy- n-butyl-nitrosamine (73). When the particulate matter, "tar," was collected from cigarettes not enriched with basic components or when the smoke particulate matter was collected without aging and not in cold traps, N-nitros- amines could not be isolated from cigarette smoke (79). Since the only other publication concerned with the isolation of nitrosamines in cigarette smoke was based on cold trap collection of "tar," the positive finding of three N-nitrosamines appears questionable (86). In summary, tobacco smoke can be regarded as a potential environ- ment for the formation of N-nitrosamines. However, additional infor- mation is needed to substantiate their presence in tobacco smoke. Pozonium 910 Several investigators (33,35,50,76,99,93,1&?) have found trace amounts of Po210 in tobacco leaf and cigarette smoke. The concentra- tion of Po210 in lung tissue is relatively high (33, 67) as compared to other body tissues and is higher in smokers than in nonsmokers (33, 48 64 66). Lung tumors have been induced experimentally by intratracheal implantation of various radioactive substances. These radioactive sub- stances must, however, be present in the respiratory environment above a certain threshold level and must be in contact with the target organ long enough rto be effective (68,77,88,107). Because Poz'O emits alpha particles, it has been implicated as a lung cancer initiator (@,68,76, 77). More research is needed before definitive conclusions can be made. Until such time, however, PO *lo should be considered as a potential tumor initiator in tobacco smoke. Selenium Selenium has been mentioned as possibly being important in the pathogenesis of human lung cancer (100). Preliminary reports suggest that selenium may be present in some cigarette papers. Because earlier reports (17,34,97) indicated the ingestion of selenium caused cancer of the liver in mice, a recent investigation (101) by the National Can- cer Institute was conduoted, with negative results. So far the earlier reports of the carcinogenicity of selenium have not been substantiated. Additional information is needed on the possible carcinogenicity of selenium and its presence in cigarette smoke before selenium can be indicted as an agent in human cancer. 128 phenols Tobacco smoke contains a large number of phenols (107). Several of `them are known to be tumor promoting agents when applied in high concentrations to mouse skin previously treated with a tumor initiator (14). IN VITRO CELLWLAR CHANGES BY TOBACCO SMOKE Lasnitzki (60) extended her studies with tobacco smoke condensate on cultured human fetal lung tissue to include a "highly purified fraction of hydrocarbons" isolated from cigarette smoke condensate. In 33 out of 50 treated lung tissue explants, the epithelium of the bronchi was hyperplastic and sometimes showed squamous changes. These changes were not observed wit.h the untreated controls. Mthough a hydrocarbon-free fraction was weakly active by producing some squamous metaplasia in these explants, these tissue culture tests point strongly to carcinogenic hydrocarbons as the active group in the smoke. The findings with purified carcinogenic hydrocarbons in organ culture (91) support the finding that polynuclear aromatic hydrocarbons are one group of active smoke constituents. Carcinogenic hydrocarbons are also the only group of chemical components that have been demonstrated in vitro to induce malignant conversion of single cells (7,13). In summary, tobacco smoke has been demonstrated in vitro to induce pathological changes in tissue explants. Although such changes may be induced by different smoke constituents, as yet the carcinogenic hydrocarbons are the only agents identified in tobacco smoke which have been shown to induce malignant changes in" tissue cultures. IN Vrvo TUXOR FORMATION BY TOBACCO SMOKE Passive inhalation experiments with tobacco smoke have not yet led to fully established squamous carcinoma in mice (109). This method of application has resulted only in papillomatous growth in the tracheobronchial mucosa of a few hamsters. None of the tumors, however, was found to be invasive (30,121). It appears that passive inhalation may not lead to the induction of scpmmous cell bronchogenic cancer in experimental animals. This conclusion can also be applied to passive inhalation studies in which t.he animals are infected by a virus before long-term smoke exposure (62, 110). The pathological changes seen in the mice were reversible whether or not the animals lvere previously infected with a virus. The hyperplasia and metaplasia seen in mice and rats after passive inhalation appears, at least in part, to be secondary to viral or bacterial infection that is enhanced by exposure to tobacco smoke. The relatively negative findings with pas- 129 sive inhalation experiments probably relate to the relatively small amounts of smoke aerosols that bypass the nasal passages. The defen- sive nature of the upper respiratory tract against airborne irritanti has to be fully appreciated in the evaluation of any passive inhalation study. Active inhalation studies with tracheostomized dogs, as carried out by Rockey, (79,230) and Auerbach ($) , suggest that this approach may lead to the induction of bronchogenic carcinoma. The change in the bronchial epithelium after 1 year of active smoking indicates early pathological changes t.hat may, upon continued smoke exposure, lead to tumors in the bronchi. So far, neither passive nor active inhalation studies have contributed to our knowledge about the nature of the tobacco smoke carcinogens, Studies with the particulate matter, tar, of cigarette, pipe, and cigar smoke, however, have clearly demonstrated that at &he site of applica- tion tumors can be induced. Tumors have been induced on the skin of mice and rabbits, the ears of rabbits, the subcutaneous tissue and hilum of rats and the cervices of mice (.9,21,9& 92, M, 46,.& 61,74, 82?,8t9,84,107,108). Only relatively few investigators have been concerned with the nature of chemical carcinogens in tobacco smoke (47, 84, 107). Al- though t,he acidic and nicotine-free basic portions of tobacco tar had been found to have weak tumorigenic activity, the only fraction shown to have induced significant numbers of tumors is fraction B of the neutral portion (2 percent of the whole condensate) (I#`). This B fraction was further fractionated into three subfractions from which only B, was shown to have tumorigenic activity (47). The B, frac- tion equals 0.6 percent of the tar and combines all aromatic hydro- carbons with three to seven rings including the carcinogenic ones. This can be considered as evidence that in in zri~u studies, the poly- nuclear aromatic hydrocarbons are the major carcinogens in tobacco smoke. Although these compounds alone can account for only a small portion of the tumorigenic activity of tobacco tar, they are, neverthe- less, the only identified carcinogens and tumor initiators in tobacco smoke shown by experimentation to be biologically active. Their tumorigenic effect is enhanced by the presence of tumor-promoting agents in &he smoke. TUMOR -I?FUXK~~NG AGENTS IN TOBACCO PRODUCIS In the experimental setting, the tumorigenicity of tobacco smoke condensate cannot be solely explained by the presence of known car- cinogens. In assays on mouse skin and rat subcutaneous tissue, the known carcinogens must be enhanced by other components such IIQ: tumor-promoting agents. In fact, it has been demonst.rated that to- 130 bacco extract and tobacco smoke condensate can act as promoters to mouse skin previously treated with tumor-initiating carcinogenic 1,olycyclic aromatic hydrocarbons (10, 19, 96, 107). Although some tumor-promoting activity of tobacco "tar" can be explained by some phenols and carboxylic acids, additional tumor promoters in tobacco products remain to be isolated and identified. It is important, however, that a significant decrease of the poly- nuclear aromatic hydrocarbons in tobacco "tar" leads to a significant decrease of t.he overall activity of the %ar" on mouse skin (9,&?, 168, 109). In summary, experimental studies have demonstrated that the par- ticulate matter of tobacco smoke, "tar:' is tumorigenic. Some poly- nuclear aromatic hydrocarbon-carcinogens have been identified as con- tributing significantly to the overall tumorigenic activity of tobacco smoke condensates in the experimental setting. LUNG CANCER MORTALITY DATA ' The annual number of deaths in the United States from cancer of the lung (International Classification of Diseases, Codes 162,163) rose from 18,313 deaths in 1950 to 45,338 in 1964 ($4). In this 15-year period, deaths from lung cancer totaled 467,442. During this same time. period the death rate for cancer of the lung almost doubled, a rise from 12.2 deaths per 100,000 population in 1950 to 24 deaths per 100,000 population in 1964. (The corresponding age-adjusted rate has also nearly doubled, therefore the increase in the death rate cannot be attributed to the changing age composition of the population.) The lung cancer mortality in the male population increased from 19.9 deaths per 100,000 population in 1950 to 41.4 in 1964, while in the female population the deaths increased from 4.5 to 7.1 per 100,000 population over the same time period. The mortality experience of the individual male cohorts during 1949-64 (fig. 1) shows that at any given age the risk of dying from lung cancer was almost always higher for the more recently born cohort. Within each cohort, the death rate for lung cancer increased steadily to the end of the life span. Figure 2 shows the death rate for women by cohort groups and age at death. One can see the increasing death rate slope for each more recently-born cohort, starting with cohort F-those women who were 26-30 years old in 1930. This corresponds to the time when smoking became increasingly popular among women. `AU death rates throughout this chapter are per 100,000 population unless otherwise indicated. 131 / fir, ID 0.9 - a* - 0.4 - 03 //, , , , , , 90 9* 90 99 40 49 90 99 90 A ..- 0.9 0.9 04 I I I I I a3 99 10 79 .O 99."4 a.&. FIWBE l.-Qancer of the lung among men, by birth cohort and age at death; 1@40,1964,1969, and 1964. Fmuu Z.-Cancer of the lung among women, by birth cohort and age at death : 1!349,1954,1959, and 1984. In the female population the greatest percentage increase-(116 per- cent) over the 15-year period, 1949-64, occurred in the 35-44 year age group. The next highest percentage increase was noted in the age group 45-54 years. The death rate from lung cancer among women, 25 years and over, rose steadily with advance in age for each year during 195&64, and the cohort experience shows that these death rates continued to increase for each cohort to the end of the life span Hammond's (4.0) prospective study provides extensive information about the lung cancer mortality experience of both men and women in relation to cigarette-smoking history as presented by mortality ratio * and by death rates per 100,000 person-years. (Table 1). TABLE 1 .-Lung cancer mortality ra&os andoh% rates * of amokem by sex am! specijfc age fpwup8 Mortality ratios- _ _- _____ _ __ ___ 2. 17 7. 84 1 1.76 11.69 Death rates _____ ________-___-_ * (7)15 2 (12)87 f (17)30 ' (~3)262 lC.om *Num E t tedfmmap .tablel9. ninpsrPmt eawiudi&ed&hratetorncm!mokm. SOTJBIX: Hammond. E. C. [tablea 24 and 2J epp. table 19 (/o)]. Tables 2 and 3 below show the relationships of number of cigarettes smoked per day, degree of inhalation, and age smoking began, to lung cancer mortality ratios and death rates for males and females, respectively. Generally, mortality ratios and death rates increase with increasing amount of cigarettes smoked and degree of inhalation, and with a longer lifetime history of smoking. Table 3 shows the relatively lower lung cancer mortality among women as contrasted to men, but reveals, for the most part, the same relationship to amount smoked, degree of inhalation, and age when smoking began. Table 4 illustrates the fact that cessation of cigarette smoking is associated with a decline in lung cancer death rates. o The mortal&y ratio is the ratio of the death rate of smokers to that af non- smokers-the mortality mtb of nonsmokers always being one, by definition. 134 TABLE 2.-Lung cancer (men). Number of o?tmS, and &ge-stQndard&d death rat&S and mdtiy rash, by current number of c&are&s smoked per day, Mee of inhalahn, and age began smoking, by age at start of study 1 - c~ent numbex of cigarettea a day: * t(le _____ _____--__--__--__--____ 10 to Ill ________-____-___-__------ 20 to 39 __________________________ 40 plus--. _ - _---_---_----_------- Degree Of fllh8htiOll: None or S&ht _____--__--________ M&rate~-~ _ ______ _ ___ __ ___ _.___ D-p- _ _____ ______________ _ __ _ ___ *4ge began elgaretta Qnou 25 or older _______________________ 20 to 24 ____________________------ 15 to 19 ____________-___-___------ Laps than 16 __--___-_--___---_--- Never smoked regularly ___-___--__-- Q 16 la8 !&I 1Q 114 55 5 a1 112 a5 11 - Current numbsr of c@ratted B day: 1 to Q _____________--___-_-------- -_---_-. 10 to 19 -_--__--__---_------------ __-----. 20 to a9 -_--__--__---_------------ _------. 40 plus _______-________-___------ __-----. Degrea of inhalation: None or slight ___________________ _______. Mien--.-------.--------.---- _______. Deep. _ __ _ _ _ ___ _ _______-___-___-- ----__-. Age began cigarette 8moHng: 2.5 or older _______________________ _______. 20 to m -___ ._--___-__-_- _ -------- -_ --__-. 15 to 18 ---__-__-_--__---_-------- ------ -. Less than 15 _____________________ _______. Hum- berof k&ha 12 67 2ls 50 a7 1Tf 72 12 72 175 s7 27 19 11 I I 2.5 E% "2 ieaths -- w m a2 90 881 15Q a2 201 l!20 102 all ls8 141 175 20 a9 110 118 816 155 1Ol 188 49 12 .-__---- 4.60 .---_--_ 7.48 ._--_--_ la 14 ._-__- -- 16.61 I 1 I I I I 0.17 ____-__ a.63 _---_--_ haa aw -_--___ a77 ________ 9.62 9.87 -_----- 18.82 ______-_ 17.M 7.57 __----_ ._ 17.47 ______-_ 29.04 4.7b ____--_ .- la&l -------- 7.65 a4a _______ ,_ 11.72 ______-_ 1.5.88 0.00 ------_ taDa 1 I ______-_ 26.24 2.77 __--_._ aal ______-_ aa6 5.89 -_---_- ,- 11.11 _-__---- 1211 8.71 ___--_ _ 13.05 ---_---_ lQ.57 12.80 ___--_. ._ 15.81 ---_--_- 16.76 - ._---- -- a42 .; ------ 11.44 ._------ 14. a1 _ - - _ _ _ _ - 3.21 --__--__ 9.72 _ _ - _ _ _ - - 12 81 - _ _ - _ _ _ _ 15.10 1 Mortality ratioa are based on de&h ratea carried out tn 1 more shnlticant llgum thanshown. Sonam Hammond. E. C. [table 20 (&I)]. 271-394 ~7-10 185 TABLE 3.-Lung cancer ( women). Number of o2uthq age-standardiz,?d death ratea, and mortality ratios, by type of .smoking (lijetiime h&tory), current number of cigar& smoked per day, degree of inhalation, and age began smoking, by age at start of study l Type of smokfng (lifetime bI6tor~) Never smoked regularly. ____ ____ __ __ ____. HlstoryofcigarettesmoHng~..... _____ ___ Current mmber of dgarettes a day: ltolQ--------..-.......~------------- ZOpltEL.. ~~~~~~-~_~~~~~~~~ _ _--___ ____ Degree of inhalation: None or dIgbte ____ ._ ____ _____ _____ ___ Modernteordesp __-----__ _ ___________ Age began SmoHng: Iorolder _______________ _ _________ ___ Lessthan _______------ _ -________ ___ Never smoked regularly _____- ---_ _________ HIstory of cigarette smoking __________ _ ___ cnrrent number of cigareta B dny : 1tolQ _____-________ _ ________ _ ________ !mPlaa ---------_____.__ _._-__ ________ Degree of inhalatbn: Noneordfgbt ______________ _ _______ __ Moderateor deep ___________ _ ______ ___ 43 bagea -ohsI: loroldsr.---..-.-------------------. Les9tbanl~_~~~~___~_________ _ ______ - Age 40-64 Age m-74 All sges. 40-74 1 Mortality ratloe are based on death ratea carrkd out to 1 more &nfficant &we than shown. Sowx: Hammond, E. C. [table 23 (@)I. TABLE 4.-Lung cancer (men). Age-standurdized d&h rates and mor- t&ity ratios for ex-cigarette smokers wG!h a history of cigarette smoking only, by jormer number of cigareti smoked per day, and years since last cigarette smoking. Death rates for current cigarette smoker8 with a history of cigarette smoking on4y. Men who never smoked regularly are shown for comparison. Men ageo? 60-69. U,&,r 1 yMU _-______________ l~lyealti ___------________- 5 tlJ 9 years -_--_______-___-_- 10p1usyem ---------_------- Total ex-smokem. _____ current clmrettesmokers~~- 22.808 Never smoked regal~~ly.-~~- 1 computed from source. SOURCE: Hammond, E. C. [table 21 (JO)]. The Dorn study (49) of U.S. veterans provides additional informa- tion on the relationship of dosage to mortality ratios and d&h rates for males who smoked cigarettes only (table 5). TABLES .-Lung cancer nwrtul~y ratios and d&z+% rates for US. veterana by age, type, and amount of smoking I Nnmber of cfgmttea~dey 0 l-9 10-33 21-39 Current cigarette smokers only: Age 4s to 54 __-_----___--_ __---- __------ _----- -___--_- Agesstoo4 __-_-__-____-- 10 I.00 70 Agetlsto74. -____________ a0 1.00 133 A3rraphu -____----__--__ 46 To&l. _ __________ ______ Ex-clgerette smokers only ____ ______ ________ ______ 'DR,De&hrate;MR,Morhlitgmtio. SOOBQC: U.H. vetamns study [app. table A (#I)]. rO+ DR MR I -___-- 23.93 ____-_ 8.34 187 The mortality ratios of the Dorn (~$9) study can be compared with those of the Canadian vebrans study, in table 6 : TABLE &-Lung cancer mortd&!~ ratios for Canadian veteran by age, type, and amount of smoking Current cigarette smokers only: Age30b49 ______________________ 1.00 Age5Oto69 __________ - ___________ 1.00 Age 7Oplut3--- ____________________ 1. 00 Total ___________________________ 1. 00 1-Q 10-20 21+ 2. 47 4.15 4.08 10. 71 26.92 26. 33 12. 15 9. 43 24.53 la 00 16.41 17.31 Ex-cigarette smokers only total __________ 0. 06 Bower: Camdlsn Pamionm study I(S), Table 8.1 and 8.21. From the data shown in table 2 mortality ratios of 17.47 and 29.84 may bbe noted for smokers of 40+ cigarettes per day, age 55-69 and 70-84, respectively. The Dorn (-69) study (see table 5) similarly shows mortality ratios of 33.80 and 23.20 for smokers of 40+ cigarettes per day, age 55-64 and 65-74, resp&ively. The Canadian study (see table 6) shows mortality ratios of 26.83 and 24.53 for smokers 50-69 and 70 years of age and older respectively who smoked over 20 cigarettss per day. There is rather close agreement among the three large prospec- tive studies for the general range of mortality ratios observed in heavy smokers. From the data supplied by the Doll and Hill survey of British physicians (98, f29) a mortality ratio of 31.86 can be calcu- lated for all smokers of more than 25 cigarettes per day, as com- pared to a mortality ratio of approximately 8, for smokers of 1-14 cigarettes per day (sea table 8). There is relatively little risk of lung cancer associated with pipe or cigar smoking, probably ,&cause smoke from these sources is rarely inhaled. "Mixed smokers," i.e., smokers of cigarettes, pipes, and/or cigars, have less risk than do smokers of cigarettes only, also suggest- ing that they may smoke fewer cigarettes or inhale less tobacco smoke than do smokers of cigarettes only (see tables 7 and 8). 138 TABLE 7.-Lung cancer mod&y ratios by type and amount smokea? ~~~iwx: U.S. veterans study [app. table A W)l. TABLE 8.- Lung cancer death rater by type of smoker and amount smoked ~vsce: Study of British phyzddmu [tables 22 and 24 (ts)l. TABLE 9.- Lung cancer death raks for ez-smokers of c&are&a by length of time stopped smoking somcs: Study of British Physicfans [table 25 C?41. The preceding studies show appreciably lower mortality ratios and death rates from lung canCer with the cessation of cigarette smoking (see tables 4,5,6,7,8,9). This lower risk is evident irrespective of the quantity of cigarettes formerly smoked. The Doll and Hill study (a) of British physicians is of particular interest in respect to ex-smokers. Over the M-year period of the study (1951-61) 29 percent of the smokers of cigarettes only, had signifi- cantly decreased (one-half pack cigarettes or more) their smoking (in- cluding those who stopped) and 5 percent had switched to pipes and/or cigars. While the overall lung cancer mortality of men over age 25 in England and Wales had increased 22 percent over this lo-year period, that for the physician group decreased 7 percent. Since the total physician group is involved in these figures, we can compare this population group to the entire population of England and Wales where there was no general decrease in amount of smoking. This can be thought of as a controlled cessation experiment and the beneficial effects of stopping or decreasing the amount of smoking become quih evident. Wicken (A&?), in a retrospective study of lung cancer mortality in Northern Ireland during the period 1960-62, reported the following results (Table 10) : TABLE IO.-Lung cancer mortality ratios a?w! de&h rate.9; by sea?, age 86 and over, by type and amount of smoking, N&m Irelund, 1860-M NOD smokers "s:z l-10 ll-!a !a+ --- Male: Mortality ratios- _ _ __ _ __ Death ratea ____________ Female: Mortality rat&- _ __ ___ _ Death ratea ____________ 1.00 4.83 9.33 21.2 18 87 168 383 1.00 2 27 6. 72 19. 0 11 25 74 210 ~OUBCE: Wiekeu, A. J. [(Icn), Table 17. I -- 5. 22 2. 27 94 41 -__-----------__ -_----------we__ Wicken also analyzed the proportion of lung cancer deaths which would have occurred if the lung cancer mortality rates of the least susceptible groups had been applied to the whole population of North- ern Ireland, and found that males would have had only 18 percent of the lung cancer mortality if none smoked and that if they lived in truly rural areas they would have only 10 percent of the mdrtality. Thus, the difference percent-may be attributable to the urban or suburban residence factor, possibly air pollution. If no females smoked, they would have had only 65 percent of the total female lung cancer mortality, and 53 percent if they lived in truly rural areas. Thus, for females, the difference of 12 percentage points might be attributed to the urban environment. The magnitude of these differ- ences depends on the prevalence of lung cancer in the various sub- groups of the particular population studied. HISTOPATHOLOGY OF LUNG TUHO~ Classification of lung cancer by histologic type was discussed in the Surgeon General's 1964 Report with the conclusion that the squamous, undifferentiated, and oat-cell carcinomas were far more frequently found in smokers than in nonsmokers, while adenocarcinoma was rela- tively more frequent in nonsmokers, especially women. Changes in the bronchial mucosa resulting from the inhalation of cigarette smoke in- cluded loss of cilia, basal cell hyperplasia, and the appearance of atypical cells with irregular hyperchromatic nuclei. These changes, it was concluded, were related to the premalignant process of the de- 140 velopment of invasive carcinoma. Auerbach (6) has more recently reported on a study of the pathology of the trabheobronchia] trees of 339 men who died from causes other than lung cancer and of 63 men who died from lung cancer. Up to 55 cross-sections of the tracheo- bronchial tissue were studied in each case. The 339 non-lung cancer cases included 65 men who had never smoked cigarettes and 274 men jvho had smoked in various amount. Figure 3 shows that only 1.3 percent of the slides from those who never smoked regularly have 60 percent or more atypical cells, whereas 76 percent of the slides of those smoking more than two packs a day had 60 percent or more atypical cells. (See figs. 3 and 4). r PERCENT OF SLIDES WITH MSJONS SHOWING 603 OR MORE ATYPICAL CELLS 76.8 t 34.9 /.4, 14.7, `I .,,3, I Novrr Smoked Iortality ratio--- _ _ _ _- _ 1 3.27 8.45 13.62 18.85 7.28 10.33 Death rates: Age 55 to f34--- ___ _ 1 __---- 4 5 20 4 3 Age65to74 _______ - _____ 7 13 17 -----__- 12 20 Age 75 to 84-- _____ 13 _---_____-______________________________ SOUROR: U.S. Vetemm shdy [app. table A WI. The Doll and Hill study reported their data in terms of cancer of the larynx or trachea (see tables 14 and 15) for relationships with type and amounts of tobacco smoking. The Canadian study did not provide separate data on cancer of the larynx. No additional information has become available, since the Surgeon General's 1964 Report, relating the several forms of smok- ing, i.e., cigarettes, cigars, and/or pipes, to specific laryngeal cancer sites (intrinsic versus extrinsic larynx). The study previously referred to (69) which analyzed the develop- ment of second sites of cancer after cure of a. primary oral cancer, reports that of 37 smokers who stopped smoking, none developed cancer of the larynx but that four of 65 continuing smokers developed cancer of the larynx. Although small numbers arB involved, beneficial aspects of smoking cessation are suggested. Additional epidemiological evidence supports the previous con- clusion that cigarette smoking is a significant factor in the causation of cancer of the larynx. CANCER OF THE ESOPHAGUS The Surgeon General's 1964 Report concluded: "The evidence on the tobacco-esophageal cancer relationship supports the belief that an association exists." However, the Committee at that time noted that there was not adequate data on which to base a decision as to whether the relationship was causal. The National Center for Health Statistics (94) reports that from 149 1950 to 1964 the mortality from cancer of the esophagus rose about 8 percent in the male population and 9 percent in the female popula- tion. In 1964, males had a death rate for esophageal cancer that was 3.7 times higher than the female rate. The greatest relative increases were in the age groups under `65 years, especially the age group 3544 years. MORTALITY DATA FROX THE LARGE PR~SPECI~VE STUB=: The Ham- mond (40) study reports the following death rates and mortality ratios for males in the age groups 45-64 and 65-79 who have a history of smoking regularly: TABLE 18.--E8opha@z.l cancer nwrta.ltiy rat& and deau ratm for mule Cigar@ 8??der8, by Sp&fi age grvup8 I b@- I see 66-n Mortalityratios------ _______ -_-_--_-- __________ 4. 17 1.74 Deathrates------------------------------------ 1 (1) 4 ' (4) 7 1 Nombem in parentheses IndIe& death rates of penrom who have never smoked -3'. Souacn: Hammond. E. C. [t&b 24 (4O)l. The Dorn study (49) reports the following mortality ratios and death rates in rslation to number of cigarettes smoked per day plus other forms of smoking: TABLE 19.~Esophageal cancer mortdity ratios and death ratea for U.S. veterans, by age, type, and amount of smoking Number of c&a&tea per day Pipe $rg yg Pipe 0 14 10-20 21-m 40+ ---- ----- Mortality ratios- _ _ 1. 00 1. 76 4. 71 11.50 7.65 4.05 6.33 1. 99 Death ratea: Age 55 to64--- 1 2 5 14 9 5 8 ----_- Age 65 to 74-w 3 ______ 16 25 10 20 23 18 Age 75 to 84--w 45 -------------------------- 41 ------ 72 Scmac~: U.S. Vetwane study [app. table A (@)I. The Canadian veterans study did not give separate information about deaths from esophageal cancer. Autopsy studies of smokers as compared with nonsmokers, spe- cifically observing the pathological changes in esophageal tissue, have been performed by Auerbach (3). A microscopic study was made of 12,598 sections of esophageal autopsy tissue from 1,268 men, who died from causes other than esophageal cancer. The smoking histories were recorded ,but not known to the person examining the slides. The find- ings were strikingly similar to the abnormalities generally accepted as 150