Document rxBrO9oG6dyavz745V04ZV90

FILE NAME: Smoking (SMOK) DATE: 1956 DOC#: SMOK030 DOCUMENT DESCRIPTION: Journal Article - The Role of Atmospheric Pollution in the Pathogenesis of Pulmonary Cancer C iA ^- The Role of Atmospheric Pollution in the Pathogenesis of Pulmonary Cancer: A Review* P a u l K o t in (University of Southern California, School of Medicine, Ix)s Angeles, Calif.) I. Introduction............................................. 375 A. Lung cancer increase real B. Exogenous environment etiologically implicated II. Historical................................................... 376 A. Nonpulmonary occupational cancer and lung cancers of occupational origin B. Incrimination of atmospheric envi ronment C. Incrimination of tobacco III. Epidemiologic Considerations............... 376 A. Limitations of data B . Variations in lung cancer rates among countries C. Variations in lung cancer rates within countries (urban-rural differences) D. Variations in age peak incidence and sex incidence E. Interpretation of epidemiologic data for selected countries by local inves tigators F. Racial differences in lung cancer rates G. Socio-economic differences in lung cancer rates H. Occupational lung cancer I. Resume IV. Environmental Carcinogenic Agents__ 384 A. Occupational respiratory carcinogens B. Sources of carcinogenic atmospheric pollutants C. Known and suspected carcinogenic agents in the atmosphere D. Biologic demonstrations of carcino gens with atmospheric pollutants V. Discussion................................................ 387 A. Epidemiologic considerations in in terpretation of pathogenesis * Tart of the work reported here is supported by grants from the Field Investigations and Demonstrations Branch, National Cancer Institute, United States Public Health Service, Departm ent of Health, Education, and Welfare, and by a grant from the Ad Hoc Lung Cancer Committee of the American Cancer Society. B. Role of chemical and physical factors in pathogenesis C. Significance of experimental labora tory data in pathogenesis D. A theoretical mechanism for the pathogenesis of lung cancer VI. Conclusion.............................................. 390 INTRODUCTION A. Lung cancer increase real.--A marked in crease in the absolute mortality from cancer of the lung has been demonstrated in various regions of the world during the past several decades (36, 83). Certain epidemiologic aspects of this increase sug gest etiologic association with carcinogenic agents presumably introduced into the external environ ment in the recent past. Of the causal agents ad vanced, atmospheric pollution (12, 55, 97, 98) and cigarette smoking (7,15,28,30) have been regarded as the two most significant. Lung cancer, in common with all neoplasms, appears unlikely to have but a single initiating and promoting agent concerned with its pathogenesis. Any analysis of the atmos phere as a carcinogenic entity must, therefore, in clude reference to its possible role as an adjuvant or associative factor to other potential environ mental carcinogenic sources. Cigarette smoking, the most often accused of these, will be evaluated in the light of its possible association with atmos pheric pollution in the initiation of lung cancer. Sufficient time has elapsed and enough informa tion has been accumulated relating atmospheric pollution to the increase in lung cancer to make a critical review of the data advisable. This report will conclude with a presentation of a suggested theoretical mechanism of human pulmonary car cinogenesis based upon experimental and epidemi ologic studies. Theories of lung cancer pathogenesis, to be valid, should result from a synthesis of responsible epidemiologic, clinical, and pathologic data. Equally significant are supporting data derived from laboratory investigations. B. Exogenous environment etiologically implicat 375 376 Cancer Research ed.--There is no convincing evidence that an in trinsic biological change might be responsible for the emergence of nonoccupational lung cancer from the status of a medical curiosity at the be ginning of the century to a position of major im portance at mid-century. In the absence of such data an exogenous source of the carcinogenic agents must be postulated. These exogenous environmental agents did not become manifest simultaneously throughout the world, as noted by significant differences existing from country to country relative to the time of on set of the increase in lung cancer, the rapidity and intensity of the increase, and variations in the age groups manifesting the peak incidence. These vari ations are accompanied by similarities which point to the exogenous origin of the causal agent or agents. These include a greater frequency of the disease in urban residents (31, 53, 58, 69, 86, 88, 97, 98) and an intensely exaggerated, almost ex clusive increase in males (17, 22, 44, 58, 97). A clinical historical association of heavy cigarette smoking has been reported in a majority of the cases of lung cancer. Sociologic investigations have attempted to associate characteristic histopatho logic patterns of pulmonary neoplasms with spe cific environmental factors. As is evident from the reports of Kirklin et al. (59) and Walters and Price (101), the absolute pathologic classification of lung cancers is a t present exceedingly difficult. The hazards of establishing cell types and cellular origins are especially marked in relation to ana plastic or so-called oat-cell cancer. HISTORICAL A. Nonjmhnonary occupational cancer and lung cancers of occupational origin.---Sir Percival Pott (87) first correlated clinical cancer with carcino genic materials from one source of atmospheric pollution. The local atmospheric pollution inciden tal to the function of chimneys is self-evident. Whether the sweeping of chimneys carried an in creased liability to the development of lung cancer in addition to scrotal cancer is at present undeter minable. B. Incrimination of atmospheric environment.-- Atmospheric contamination as an environmental source of pulmonary carcinogens was first demon strated when Hartig and Hesse (51) identified the pulmonary disease in radioactive ore miners in Schneeberg as lung cancer. There, of course, the atmospheric pollution was a highly localized one, limited specifically to the occupational environ ment of the miners. The epidemiologic and patho logic study of the Schneeberg miners resulted in th e establishm ent of environm ental lung cancer as a definite entity. Concepts of latency, the interval between first exposure to the carcinogenic agent and the clinical manifestation of lung cancer, were initially derived from these studies. An increased risk in the development of lung cancer has since been demonstrated to be associated with occupa tional exposure to nickel (80), chromates (1, 4, 78), and gas-working operations (26). C. Incrimination of tobacco.--'Among the first to describe a correlation between the development of cancer of the lung and factors other than strictly occupational ones was Muller (81) who in 1939 re- ' ported a statistical study purporting to show a ! correlation between smoking and lung cancer. An analysis of the occupational exposure of his sub jects in the light of current environmental cancer knowledge indicates that his conclusions of a posi tive etiologic association are open to question. Fol lowing Muller's original report, numerous studies have been undertaken to ascertain any relation ships existing between lung cancer and a broader spectrum of environmental factors. A majority of recent investigations noted an association between lung cancer and heavy smoking (47, 72, 89, 104, 105). Numerous investigators have by means of comparative retrospective studies adjudged that heavy cigarette smoking embodies an enhanced risk to the development of lung cancer. The con clusion that this association demonstrates a cause-and-effect relationship has been most elo quently advanced by Hammond and Horn (50), who observed, "For reasons discussed, we are of the opinion th at the associations found . . . be tween regular cigarette smoking and death rates from lung cancer reflect cause and effect relation ships." Contrary opinions have been voiced by other students of the problem, who emphasize that the real increase in lung cancer is markedly less than is generally accepted and thus question whether re ported associations between cigarette smoking and lung cancer signify any etiologic implication. Berkson (3) in reviewing the data of Hammond and Horn reported, "M y thesis is only that it is un warranted to conclude from them [data] that a meaningful association already has been proved beyond doubt, as some writers have asserted and as appears to be widely accepted in the United States. Much less do I believe that causation has been established." EPIDEMIOLOGIC CONSIDERATIONS A. Limitations of data.--In evaluating the epidemiologic factors relating to this review, cer tain deficiencies inherent in the available crude data must be considered. K otin-- Atmospheric Pollution and Pulmonary Cancer 377 acy, the interval rcinogenic agent ung cancer, were 5 es. An increased 1 cancer has since ed with occupamates (1, 4, 78), Among the first r. he development her than strictly ) who in 1939 re nting to show a lung cancer. An sure of his sub- mmental cancer usions of a posi- to question. Pol- umerous studies in any relation- r and a broader s. A majority of 'ciation between 47, 72, 89, 104, ve by means of adjudged that es an enhanced ancer. The con- lemonstrates a been most elo- and Horn (50), issed, we are of ^ found . . . be- md death rates effect relation- oiced by other hasize that the `dly less than is ion whether rete smoking and plication. BerkHammond and r that it is uni [data] that a is been proved re asserted and in the United t causation has ATIONS V valuating the | iis review, cer- ft ivailable crude I First, the increase in lung cancer incidence as reported by investigators in different countries in some instances deals with mortality rates derived from vital records, while others report incidence figures obtained from necropsy records. Both sources have limiting factors. James (56) and his co-workers questioned the use of cause-of-death statements on death certificates as he found " . . . the extent of the error in a large number of specific cases raises serious doubts as to the validi ty of the use of cause-of-death data as a basis for epidemiological studies of degenerative diseases." The unjustified use of necropsy data was decried by Gilliam (43) when he directed attention to the . . deplorable and almost universal tendency in the literature of pathology to draw from autopsy data firm epidemiologic conclusions which a t best should be regarded in the same light as clinical impression." This admonition should be tempered by the knowledge that necropsy data have been of great value in the study of certain infectious dis eases. Regardless, the time of introduction of the carcinogenic agent or agents into the atmospheric environment can be arrived at only on the basis of assuming an average period of latency. This inter val represents the period between the first expo sure to the carcinogenic agent and the clinical manifestation or discovery of the lung cancer. Second, great differences exist as to the time when dependable reporting and recording of cancer deaths started in various regions throughout the world. Third, until recently there has been a lack of standardization in the method of reporting and recording cancer deaths. B. Variations in lung cancer rates among coun tries.---While great differences continue to exist in several of the preceding factors, an analysis of available epidemiologic data indicates that prior to the early 1930's records for several countries are incomplete, inconsistent, and in many instances inaccurate. Beginning with this period, however, one majr with some degree of assurance compare death rates from lung cancer with those recorded in the early 1950's. Though similarities in trend exist in all countries, when the two series of death rates as determined by vital statistics are re viewed, a great disparity can be noted in the spe cific rates (83). This disparity in actual incidence rates tends to support the thesis that the environ mental carcinogenic agent differed in the time of its introduction and varied in its intensity from country to country. Charts 1 and 2 record com parative death rates for a series of countries be tween the two periods mentioned above. C. Variations in lung cancer rates within coun tries {urban-rural differences).---Paralleling the dif ferences noted from country to country are the variations exhibited in local geographic areas with in countries. A basic and almost universal obser vation has been the demonstration that urban residence carries with it an increased liability to the development of lung cancer. The urban-rural difference in death rates from lung cancer has been demonstrated by a number of investigators using contrasting methods of study. Stocks (96) in Great Britain used density of population as the reference line for lung cancer rates. Lew (78) demonstrated urban-rural differences in lung cancer as part of a comparative study of death rates in male indus trial policy-holders and males holding ordinary insurance policies with the Metropolitan Life In surance Company. He correlated the 30-50 per cent higher rate in the former group with urban resi dence, low economic level, and industrial or manu facturing occupational environment. In contrast, ordinary policy-holders belonged to the higher in come groups with significantly fewer opportunities for protracted exposure to industrial hazards. Eastcott (38) in an analysis of native New Zea landers and immigrants studied the effect of ur banization on death rates for cancer of various body sites. Of all visceral cancers, those of the lung and bronchus exclusively showed variations a t tributable to exposure of the host to environment. The effect was directly related to the intensity of antecedent exposure. The exclusive factor in the former environment capable of incrimination was urban residence. Mancuso (79), in a study limited to a single, highly populated and industrialized state (Ohio), correlated urban residence with the liability rate to lung cancer by showing th at the observed death rate was greater than expected in the eight industrialized urban counties, and onethird less than expected in the remainder of the state. While it is unlikely th at occupational exposure to a specific carcinogenic atmospheric environment can materially affect nation-wide incidence figures, there is little question th at the increased liability to lung cancer is consistent with the increased in dustrialization. An analysis of lung cancer death rates by states in the central states (82), when re viewed in the light of the degree of their industri alization, shows a consistent positive association between intensity of industrialization and mortali ty from lung cancer. D. Variations in age peak incidence and sex in cidence.--Additional convincing supporting evi dence for variations in the time of introduction of the carcinogenic agent into the atmosphere can be found when the peak age incidence for lung cancer is compared from country to country as shown in Fn s i a n d AND Wales S cotland Fin l a n d 1930- 1932 1949 ` 2 I 12.9 1930 -I932| 10.6 1949 1952 1936 - 193ft 1949 1952 | 13.6 | 49.5 ____________________________________I| 61.4 | 41.4 ]1563 1 29.8 " 1 38.0 SWITZERLAND i929-<93l 1 12.0 1949 1952 [26.1 1 33.5 F s w ZEALAND MemnLANOs '30-'321 7.1 1949 1952 29-'3 lJ n. Z 1949 1952 I 21.6 1 31.5 1 24.5 i 30.-3 Fr a n CD 1949 1952 1 21.7 | 28.2 United S tate& l O ( 1929 - 193)) t949 1952 1m 1 26.1 Denmark ~| 4.5 0 9 3 4 -w ad 1949 | 1<0.1 1952 ~~| 24.6 Ip f l a n o Australia " |5 .4 il9 3 5 -1939) 1949 I IS.I 1952 3 22.2 ~ | <6.3 (1932-1934) 1949 T lg.9 I9S2 I 20.6 Canada 1 4.2 1949 115Z ( l 9 30- (932) i * 1 I t alt 1949 1952 1 11.3 | fi/cRb/AK T I.& 6 9 2 9 - 193l) 1949 J 9 . o 1452 11.5 J apan B3 .3 ( 1949) 3 39 (1 9 5 2 ) C hart 1.---D eath rate for cancer of respiratory system in males (rates per 100,000 deaths). D ata obtained from DirectorConsultant on Health Statistics, World Health Organization. 1 1.4 i from Director- K o t in -- Atmospheric Pollution and Pulmonary Cancer 379 Table 1. In those countries with peaks in the earli er decades of life, the introduction of the carcino genic agent must certainly have been proportion ately sooner or in greater concentrations. Further, Fn iANP An C A/a l p s Ci93o-igsz) ] 9.8 I l 11.3 S cotland j 5.6 Q30-I932) 1940 T9 sz fl I0.fi j 10.9 Fin l a n d g 2 3 C1936-1938) I949 5.4 I9S2 I .5 1 " I 2.0 U9Z9-I93IJ SWITZERLAND 4 , 1 ( 1949 ) 4.1 ( 1962 ) P/EW Z (ALANO 2 . 5 ( 1930- 1932) 1 3. 9 ( 1949 ) I S (1962) f/STNEPLANPS F rance 2 . 2 . <1929- 1931) 4-4 I 1949) 4 .0 ( 1952) 1 5.6 1 0.1 ( 1949) (1952) fl Jn/tep S tates 1.9 ( \ 9 2 9 - 1 9 3 l ) 1 4 . 9 ( 1949) 1 5-4 ( 19 52.) Qcn ma r k ( 1934 -193C.) (1949) 1 0 5 ( 19 62) n 3.2 Zr SLANP (l935-I93t) (1949), 1 73 A u strali R CAKAPA ~1 2.3 1 3.6 L J 4-3 ~1 2 . 0 I 4.0 (, 1 9 3 2 - 1 9 3 4 ) ( 1949) (1952) 6930- 1932) (1949) (1952) differences in urban and rural lung cancer mortality figures may be nothing more than another mani festation of the time of peak incidence. The dose of a carcinogenic agent represents the product of the duration of exposure multiplied by the concen tration of the carcinogen. Rural areas differ from urban areas in that pollutants are present in lower concentrations rather than by their complete absence. TABLE 1 Lung Cancer I ncidence P eaks by Years for Various Countries 1949 19S M a le F e m a le M a le F e m a le Norway Italy England-Wales Finland Netherlands Scotland Ireland Switzerland Japan France Australia Germany United States Canada New Zealand 50-59 50-54 55-59 55-59 55-69 60-64 60-64 60-64 60-64 60-64 60-64 60-64 60-64 65-69 65-69 70 and over 55-59 65-69 60-64 60-64 60-64 60-64 70-74 65-69 65-69 60-64 60-64 65-69 60-69 65-69 60-64 55-59 60-64 60-64 60-64 60-64 65-69 70-74 60-64 55--64 65-69 60-64 60-64 65-69 65-69 70 and over 65-69 65-69 70-74 65-69 60-64 60-64 70-74 65-69 65-69 60-69 65-69 70-74 65-69 70-74 TABLE 2 Ratio of Male to Female withIR e- gard to Death Rates from Cancer of the Respiratory Tract Early IDSO's lvSl England and Wales Scotland Finland Switzerland Netherlands New Zealand United States Denmark Union of South Africa Ireland (Republic) Australia Canada Italy Norway 3.0:1 1.9:1 5.9:1 6.0:1 3.3:1 2.8:1 2.3:1 1.8:1 3.8:1 1.7:1 2.7:1 2.1:1 2.1:1 1.5:1 5.4:1 4.3:1 5.3:1 6.1:1 5.8:1 4.4:1 4.7:1 8 9:1 3.5:1 2.7:1 4.1:1 4.2:1 3.5:1 2.2:1 / talk f l 1.4 L j 4-S (1931) (1949) (1952) A/o r w a K d . 6 ... J 5.4 ( l 9 2 9 - >93l) (1949) ( 1952) b J a pak l.2 ('1949) 2.2 O*'5'*) Chart 2.--D eath rate for cancer of respiratory system in females (rates per 100,000 deaths). D ata obtained from Director-Consultant on Health Statistics, World Health Organization. While there is universal recognition and accept ance of the difference between men and women in liability to the development of respiratory tract cancer, great divergence of opinion exists as to its significance. Changes in the ratio of male to female death rates are shown in Table 2 for the two periods, early 1930's and early 1950's, for various countries. I t will be noted th a t the increase in death rates from lung cancer has been almost ex clusively limited to men and that the rate of cancer for women has been relatively constant within countries. I t is equally apparent that great varia- 380 Cancer Research tion can be noted in the increase in the male rate flecting a true difference in the risk in the develop from country to country. Explanations for this ment of lung cancer, he ascribed the different rates disparity on the basis of better and more readily to a delay of 8 years in the onset of the carcino available diagnostic facilities for men, greater genic influence in provincial to w s and 10 years in susceptibility in men, or an as yet cryptic sex- rural areas. This conclusion followed an analysis of linkage for lung cancer development are unten incidence rates for successive cohorts. The cohort able. Lilienfeld (74), however, very recently sum studies further indicated th at a 2-decade period of marized a preliminary study on nonsexual-site exposure was necessary for the carcinogenic effect cancers, including lung cancer, by stating that his to become clinically manifest. Lie placed the intro findings suggested an endocrine-determined sus duction of etiologic agents into the environment ceptibility to these cancers. There remains as prob during the period between 1900 and 1910. He ably a more valid explanation for the lower inci could find no reason to assume any carcinogenic dence of lung cancer in women a difference in influence due to atmospheric pollution but rather exposure to the environmental carcinogenic agent correlated the increased development of lung and perhaps to a very minor degree a difference in cancer with heavy cigarette smoking. response to it. For, in fact, where exposure to iden Norway: Kreyberg (68, 69), in reporting the tical environments has been recorded, the liability increasing incidence of lung cancer in Norway, to lung cancer between the sexes is similar (27). In noted a seven-fold increase in men and a four and like fashion, Lew (73) found no difference in the a quarter-fold increase for women in lung cancer lung cancer rates of women policy-holders belong between the years 1930 and 1950. In analyzing the ing to the two insurance groups in his study. phenomenon of lung cancer in Norway, he detect Those ascribing a predominant initiating and ed the first evidence of a rise in approximately the promoting role to cigarette smoking offer as an middle 1940's with a progressive increase occurring explanation the almost universal onset of smoking since then. He concluded from a detailed analysis in women two to three decades later than in men, of his data th at the panorama of lung cancer in and as a corollary they postulate a future rise in crease is not as yet perceptible in true rural dis the lung cancer rates in women. They neglect to tricts in Norway, which continue to have the same consider th at fewer opportunities exist in the in distribution, histological type, and sex ratio as stance of women for daily industrial and urban presented by Oslo a few decades ago. In parallel exposure to carcinogenic air pollutants. Women with Clemmessen, he demonstrated a progressive only to a very limited degree in the past have decrease in lung cancer incidence with decreasing worked in manufacturing installations, have urban community size. He concluded on the basis driven in heavy traffic to and from work, and have of a review of 235 cases that the new development performed heavy manual labor in dirty polluted (lung cancer increase) was as yet not manifest in environments. Rather, they formerly spent most truly rural districts. On the other hand, the in of their time in residential areas within cities or crease was definitely established in all types of semi-rural suburban communities. A note of cau urban settlements. In spite of a decreasing inci tion appears necessary in interpreting any future dence in progressively smaller towns, he noted that rise in female lung cancer rates. Perhaps even the urban predominance was a reflection of an more dramatic than the increase in smoking that essential urban factor other than smoke and fumes has occurred in women during the past several from industry, since towns of identical size had decades has been their entry into the previously similar rates independent of the degree of indus almost exclusively male business, industrial, and trialization. By applying the 20-year exposure or occupational domain. The implications of this are latency period, the onset of the rise may be postu clear. lated as having begun between World War I and E. Interpretation of epidemiologic datafor selectedthe middle 1920's. countries by local investigators.--The interpreta Sweden: Specific death rates for lung cancer tions by investigators of their data, which were from Sweden, as reported by Henschen (52), show freely utilized in compiling Charts 1 and 2 and a pattern of lung cancer increase dissimilar from Tables 1 and 2, are important and will be dis that of either Norway or Denmark. This increase cussed by countries. has been at a slower rate than th at noted in either Denmark: Clemmessen (14, 15, 17) and his co Norway or Denmark, with the over-all incidence workers reported the increase in mortality from figures being more nearly akin to those of the lung cancer as beginning approximately in 1931. Netherlands (60). Worthy of special comment is This increase was most marked in Copenhagen, the fact that the sex ratio in Sweden differs sig and a progressively lower incidence was noted in nificantly from th at of Denmark and Norway. provincial towns and rural areas. Rather than re United States: Dorn (31, 32) noted that the in the developj different rates ' of the carcinoand 10 years in ] an analysis of ts. The cohort ecade period of inogenic effect laced the introie environment and 1910. He >y carcinogenic tion but rather ment of lung lg` . f reporting the er in Norway, and a four and in lung cancer i analyzing the vay, he detect- i roximately the rease occurring tailed analysis ung cancer in- ) true rural dis- I have the same [ d sex ratio as 1 go. In parallel 1 a progressive ,'ith decreasing id on the basis ,v development iot manifest in hand, the in- n all types of 1 ecreasing inci- , he noted that flection of an ioke and fumes itical size had gree of indus- ar exposure or may be postu- rld War I and r lung cancer hen (52), show lissimilar from This increase noted in either r-all incidence those of the d comment is ten differs sigd Norway, oted that the K otin-- Atmospheric Pollution and Pulmonary Cancer 381 mortality from cancer of the lung in the United States in both white and nonwhite populations is greater in males than in females, and the diver gence is becoming more marked with the passage of time. The increase shows a progressive accelera tion from 1930 to 1952. A study of cancer morbidi ty data for the years 1937 and 1947 in nine metro politan centers shows incidence differences and variations in the rate th at can be explained only by environmental variations. The increase in lung cancer in the United States is present in all ages, with the peak age incidence occurring between the ages of 60-65. I t has been postulated th at the increase in lung cancer incidence in the United States became clinically manifest in 1920. By assuming the 20year period of exposure postulated by Clemmesen, we see th at 1900 represents the time of introduc tion of carcinogens into the environment. This parallelism of the time of onset in the United States with that of Denmark is pointed out by Dorn, who further noted th a t the incidence rates for males in Copenhagen in the 1943-47 period are surprisingly similar to those for the United States in the 1947-48 period. A primary difference ob served was th at the maximum incidence in the Danish data occurred between the ages of 55-59, or about 5 years earlier than in the United States. The similarity between Denmark and the United States is even more remarkable in females. Hoffman and Gilliam (53), in studying the geo graphic distribution of lung cancer mortality in the United States, showed th at cancer mortality is greater in towns than in rural areas among all age, race, and sex groups. I t is of interest to note that they used the standardized mortality ratio (S.M.R.) of the Registrar General of England and Wales in reporting their data. Rigdon and Kirchoff (88) in a more limited geo graphic study reported an increased incidence of lung cancer in urban residents. They felt that they had correlated the presence of lung cancer with the availability of medical diagnostic facilities and with low patient-doctor ratios. They summarized their findings in saying th at " . . . in our opinion the data available today do not justify the conclu sions th at the increase in the frequency of cancer of the lung is the result of cigarette smoking. The statement th at carcinoma of the bronchus has actually increased in frequency is in our opinion open to question." England: Stocks (96, 97) showed th at for the years 1920-30 there was a marked increase in cancer of the lung in Great Britain, an increase that was real, progressive, and ever accelerating. The distribution of mortality due to lung cancer showed a positive association with the density of population and maleness. Death rates in the larg est towns were more than twice as high as in the country districts. Subsequent studies in 1946-49 and 1952-54 verified this difference in urban-rural mortality. Curwen, Kennaway, and Kennaway (22) in studying cancer of the lung concluded that " . . . fresh evidence to support earlier findings that mortality from cancer of the male and female lung is positively correlated with population density." Following an analysis over a prolonged period of the same source material as Stocks, they concluded that "the Standard M ortality Ratio (S.M.R.) for cancer of the lung in both sexes, and of the larynx in males, increases with increasing urbanization, that is to say, is greater in the County Boroughs than the Urban Districts, and in the Urban Dis tricts than the Rural Districts. Cancer of the fe male larynx shows exactly the reverse relationship. "These trends apply equally when the figures are analyzed according to the separate regions, but there are differences between the regions, which may or may not be due to differences in degrees of urbanization undetected by the classification we have used." More recently, Stocks and Campbell (98) un dertook a study for the purpose of evaluating the lung cancer death rates among smokers and nonsmokers in relation to air pollution. He summa rized his findings with the statement: "The abso lute urban excess is much the same in each smok ing group, suggesting th at an `urban' factor is added to the effects of smoking." He further noted, "Differences in smoking habits of the populations can account for only a small fraction of the con trast in total rates, and it is estimated that about half the Liverpool deaths of men from lung cancer arise from cigarette smoking and about threequarters of the remaining half are due to a factor which is only slightly present in the rural a re a .. . . " France: Denoix and Gelle (25) report a regular increase in lung cancer as far back as records exist. The evolution has been regular with no accelera tion being demonstrated at any given point. Fur ther, the increase in morbidity is shared by all ages. The over-all death rate from lung cancer in males has increased 30 per cent during the years 1949-52. The actual rate is close to that of the Netherlands and Denmark, all three of which are significantly lower than those for the British Isles. Australia: Fowler (42) reported a uniform in crease in the rate of lung cancer of 7.7/100,000 per year for males and 4.6 for females. He emphasized that, though the rate of increase differs from that of other countries, the exponential pattern is the same. The increase may be presumed to have be gun in 1930, with a fragmentary increase occurring in the previous two decades. 382 Cancer Research New Zealand: Eastcott (38) in analyzing his The foregoing data permit of no conclusion other data was cognizant of the role that immigration to than indicating a need for future intensive studies. New Zealand played in the modification of his G. Socio-economic differences in lung cancer sample. By integrating this factor into his epi rates.---Numerous investigators have been im demiologic data, he concluded that immigrants pressed with the variations in cancer incidence on from Great Britain to New Zealand are affected by the basis of the socio-economic status of different their former environment and th at this effect is population groups. In applying this criterion to related to the length of exposure to that environ patients with lung cancer, Clemmessen and Niel ment. The exclusive effect demonstrated was in sen (16) noted a significant acclivity in the inci relation to lung cancer and was established as prior dence of lung cancer in the male population of the urban residence. He noted, "The chances of dying poorer classes in Copenhagen. Kennawav (57), in of cancer of the lung are 30 per cent higher for all reviewing the data relating to cancer in the publi United Kingdom immigrants, but for those who cations of the General Register Office, could not are 30 years of age or more on entering New discern any influence of social class upon the lia Zealand, the risk is 75 per cent higher. . . . Differ bility to lung cancer. He did, however, emphasize ences in habits of tobacco smoking are unlikely to " . . . the very considerable effect of urban condi contribute to this picture." tions which suggest some carcinogenic factor to Iceland: Dungal (34) in 1950 concluded that the which all classes are exposed." Very recently, rarity of lung cancer in Iceland was associated Cohart (19) in a limited study paralleled the find with a slow rate of increase in the cigarette smoking ings of Clemmessen and Nielsen and concluded, habit throughout his country. He felt that atmos "The incidence of lung cancer was more than 40 pheric pollution would play little, if any, role in per cent greater among the poor than among other any future increase in lung cancer. He predicted socio-economic classes. Unless it is assumed that that a rise in the lung cancer rates would become cigarette smoking is inversely related to socio manifest between 1960 and 1965 if smoking were economic status, an assumption th at probably chiefly responsible for its initiation. In a more re cannot be supported in fact, then it is reasonable cent report (35), he noted a beginning rise in in to conclude th at important environmental factors cidence. Though the number of cases is too few to other than cigarette smoking exist that contribute permit any epidemiologic conclusion, he relates the to causation of lung cancer." Density of popula cases to heavy smoking on the part of older people tion, proximity to industrial installations, atmos in whom "lung cancer is now beginning to pheric pollution, poor socio-economic status, and crop up." increased liability to lung cancer constitute a con F. Racial differences in lung cancer rates.--Afiguration repeated frequently in epidemiologic study of racial differences in the susceptibility to studies on lung cancer. lung cancer is indicated especially in terms of the H. Occupational lung cancer.--The significance guidance it can provide for future studies. Steiner of occupational respiratory cancers resides not et aL. (95), in a study of the necropsy records at the only in their role as an important group of indus Los Angeles County General Hospital, reported, trial diseases but also as denotations of environ "For present purposes it may be stated that in mental lung cancer truly secondary to atmospheric Mexicans the incidence of lung cancer was as high pollution. They, therefore, should serve as guide- in women as in men, and th at the incidence in the posts for the study of lung cancer in relation to air latter equalled th a t in caucasoid men." Hoffman pollution of a general rather than occupational and Gilliam (53) reported a lower lung cancer rate type. Specific pulmonary carcinogens which have in Negroes and summarized their findings: "The been identified in the occupational environment total rates for the white population are considera include nickel and chromium, as unequivocal ex bly higher than those for the nonwhite, but the amples of inorganic chemicals, and combustion difference is more pronounced among males than and distillation products of coal and petroleum, as among females." In marked contrast, Duchen (33) examples of organic chemicals. The data incrimi in South Africa found no increased incidence in nating radioactive substances have been referred Caucasians when contrasted with the Bantu na to previously. Though there is no unanimity of tive th a t could not be explained by difference in opinion, Doll (29), Weil and his co-workers (104), longevity. Warwick and Phillips (102), in a study Perry et aL. (85), and Bonser and her associates (6) of cancer among the Canadian Indians, detected have reported highly suggestive data associating no differences in incidence they could attribute to increased lung cancer rates with exposure to as race. Finally, Sitbon (92) in Algeria reported no bestos, isopropyl oil, arsenic, and iron, respec racial variations in the incidence in lung cancer. tively. K otin-- Atmospheric Pollution and Pulmonary Cancer 383 Worthy of comment is the paradoxical situation relating to chromates and beryllium. While little question exists as to the increased pulmonary can cer incidence associated with exposure to chro mates, extensive attem pts to induce experimental pulmonary cancers in a broad spectrum of mam malian species have been uniformly unsuccessful (2). Conversely, beryllium, which has but a ques tionable association with increased lung cancer rates in those occupationally exposed, has been used to produce bronchogenic carcinomas in the lungs of rats (99). The experimental corroboration interpretation. Attempts to relate increasing lung cancer rates with increasing tobacco consumption, though falling into a broad and variable pattern, may be considered relatively successful. Compa rable analyses with respect to motor fuel consump tion, increase in asphalt highway mileage, fuel oil sales, and motor vehicle registration show an even more pronounced relationship. I t should be noted that a uniform and perhaps critical deficiency in all these associations is the gross failure to corre late the data with the period of introduction of the carcinogen into the environment rather than with Chart S.--Trends in selected environmental factors, U.S., 1900-1953 (1924-26 = 100). Note: Cigarette consumption per adult reflects entire population rather than th at segment which smokes. D ata obtained through courtesy of D r. E. C. Hammond, of radioactivity as a pulmonary carcinogen has the increase in lung cancer rates. In other words, been reported following the development of bron the role and importance of the latency period are chogenic carcinomas in rats (75). either overlooked or minimized. If the latent As previously mentioned, the cases of occupa tional lung cancer recorded thus far are too few to affect materially nation-wide incidence figures. Nevertheless, neighborhood contamination and general atmospheric pollution with these sub stances is well documented. Additional details re ferring to these factors will be discussed subse quently. period is regarded as being approximately two decades in duration, reference to Chart 3 will show th at not only does the rise in lung cancer follow more closely factors other than tobacco, but in addition the increase in these environmental fac tors is more capable of correlation with the latent period. Hueper (55) has similarly shown that a I. Resume.--An unbiased analysis of the epiparallelism exists between the increased produc demiologic data reviewed here clearly shows that tion of cancer-related chemicals and the rise in at this time the data are capable of more than one lung cancer (Chart 4). 384 Cancer Research ENVIRONMENTAL CARCINOGENIC AGENTS cancers have been identified in nickel workers. As A. Occupational respiratory carcinogens.--Connoted by Hueper (55), "Carcinogenic dusts con siderable basic information has been derived from sisting mainly of coarse particles are mainly ar a detailed study of established respiratory carcino rested in the nares where they cause cancer of the gens in occupational environments. First and most turbinates. The nasal cancers observed among obvious is a broadening of the spectrum of envi copper-nickel matte refinery workers inhaling the ronmental carcinogens capable of producing pul coarse dust of the roasters illustrates this inter monary cancer. relation." The bronchogenic cancers found in Second, opportunities for specific documenta coke-oven and gas-retort workers, and among tion of initial exposure time to these agents and those exposed to chromates, represent instances of their concentrations are possible by referring to occupational exposure to aerosols, dusts, vapors, industrial records or vital statistics. Latency or mists of a particle size sufficiently small to per periods have been shown to vary from less than 10 mit penetration into the arborization of the years to as much as 50 years for the various ac tracheobronchial tree. cepted occupational cancerigenic agents. Regard Failure in this report to detail the association less of the particular agent studied, the average between increased liability to lung cancer in asbes- Chaet 4.--Rise in annual production or consumption of cancer-related industrial chemicals between 1940 and 1948. Data obtained through the courtesy of D r. W. C. Hueper. latent period invariably falls in a 15-25-year time period (55). This is not unanticipated in view' of our knowledge th a t many of these agents in larger doses are sufficiently toxic so as to produce mor bidity or mortality well in advance of the time necessary' for clinical cancer to develop. As a re sult, exposure to quantitatively small amounts would appear to be the necessary antecedent for any cancer development. Third, the development of neoplasms in selec tive sites of the respiratory tract provides informa tion on the significance of the physical state of the chemical carcinogens. Perhaps foremost among the physical factors is the size of the particulate mat ter of the carcinogen per se or the carrier on which it is adsorbed. Nasal cancers as well as pulmonary tos workers, workers exposed to isopropyl oil, and those in contact with arsenic should not be inter preted as minimizing the importance and validity of the data. Rather, the conclusions to be derived from the data are similar to those already dis cussed. The reader is referred to the comprehen sive and excellent review of Hueper in Recent De velopments in Environmental Cancer (54) for a criti cal study of the newer data relating to occupation al cancer. B. Sources of carcinogenic atmospheric pollu tants.--Of the carcinogenic materials present in the atmosphere, certainly those resulting from the incomplete combustion of organic m atter are most universally distributed. Soot, a clinically recog nized cutaneous carcinogen from the time of Sir Kotin-- Atmospheric Pollution and Pulmonary Cancer 385 percival P o tt (87), is a major component of the smoke formed and emitted into the atmosphere following the partial combustion of solid, liquid, and gaseous fuels. Compounds introduced into the atmosphere from these sources belong primarily to the group of aromatic polycyclic hydrocarbons. A typical carcinogenic representative of this group is 3,4-benzpyrene. Although the gradual shift from solid to liquid or gaseous fuels throughout the world has resulted in quantitative variations in the pollutants emitted into the air, qualitative changes have been minimal. Additional sources of these complex aromatic compounds include the tars and asphalt used for road surfacing. The increase in mileage of asphalt-topped roads has already been referred to in Chart 3. Added to this source of car cinogenic materials is the carbon black introduced into the air secondary to rubber tire wear, tear, and degradation. An ever expanding and increasing source of emission of carcinogenic hydrocarbons into the atmosphere is liquid fuel used in internal combus tion engines. Gasoline engines represent the most universal type of motive power in urban areas, and diesel fuel engines power our buses, trucks, rail road locomotives, and electric power facilities to an ever expanding degree. Not only are large amounts of the known carcinogen, 3,4-benzpyrene, introduced into the air by the combustion of petroleum fuels, but the reaction products of noncombusted gasoline in their aliphatic, nonaro matic, polycyclic state possess experimental car cinogenic potency (63, 67). Supplementing these compounds there are emitted into the atmosphere measurable concen trations of inorganic materials demonstrated to be occupationally associated with increased liability to lung cancer development. Included would be heavy metals, light metals, and inorganic dusts. Quantitation of a known carcinogenic substance obtained from an atmospheric pollution source was first reported in 1949 by Goulden and Tipler (46) who, by means of fluorescence spectroscopy, iden tified 3,4-benzpyrene in a representative sample of chimney sweep's stock in a concentration of 300 mg/kg of soot. C. Known and suspected carcinogenic agents the atmosphere.--Waller (100) sampled the atmos phere at ten different sites in several cities in Great Britain and demonstrated the presence of 3,4benzpyrene in the air. The highest concentration of 4.5 ig/100 cubic meters was obtained in Lon don, and the lowest noted was in Bristol, with 1.3 Mg/100 cubic meters measured there. A mean con centration of 2.6 /xg/100 cubic meters was obtained for all sites stu d ied . The concentration in London was half again as great during the winter as it was during the summer, and during foggy weather the concentration was over 4 times that on clear days. While the amount of benzpyrene retained in the lungs is, of course, impossible to assess at present, he calculated the respiration of approximately 12 mg. of benzpyrene during an average 70-year life span. Blacklock, Kennaway, Lewis, and Urquhart (5), following an analysis of the carbon content of human lungs, estimated th at approximately 16 mg. of benzpyrene may be inhaled during a life time. It should be noted that the primary atmos pheric source of these compounds in Great Britain is from combustion of coal, with as yet undeter mined amounts contributed by vehicular exhausts. Kotin (64) and his associates studied the Los Angeles atmosphere and demonstrated a presence of 0.84 mg. of 3,4-benzpyrene per million cubic feet of atmosphere. In marked contrast with Lon don, this carcinogen could be attributed almost entirely to the exhaust products of gasoline and diesel engines. A detailed study of gasoline and diesel engine exhausts revealed quantities esti mated up to 120 uS- f benzpyrene in 1-minute samples of gasoline engine exhausts and up to 1.7 mg benzpyrene/minute from diesel engine ex hausts (65, 66). In comparing the ratio of pyrene to benzpyrene in the atmosphere with th at at the vehicular exhaust source, Falk and his co-workers (40) noted a reversal of the ratio, with benzpyrene being present in greater concentration than pyrene in the atmosphere. This phenomenon was ex plained through study of the survival of various polycyclic aromatic hydrocarbons in the atmos phere. Compounds were tested in their pure state and while adsorbed on soot and following exposure to washed air and smog. In all instances benz pyrene was significantly more stable than was pyrene. In fact, it was virtually indestructible in the dark. Cooper and Lindsey (20) emphasized the ubiquity of atmospheric pollution following analy sis of 1 kg. of freshly fallen snow in Hertfordshire, England. The snow contained 1-2 p.g. of pyrene and traces of 3,4-benzpyrene and anthanthrene. Cierno and Miller (18) divided the city smoke they collected into three fractions. The presence of in3,4-benzpyrene was detected in one of these fractions. The spectrum of carcinogenic agents present in the atmosphere has very recently been broadened following the report of Kotin and Falk. Following skin painting in C57BL mice and following inhala tion in strain A mice and C57BL mice (63, 67), they found the oxidation products of aliphatic hydrocarbons to be carcinogenic. The implications of the pulmonary neoplasms induced with these 386 Cancer Research agents will be discussed. The chief source of ali cancers in C57BL mice following painting with phatic materials in the atmosphere is unburned extracts of Los Angeles atmosphere and materials gasoline. collected from the exhaust of gasoline engines and Mention should be made of the presence of ar diesel engines (64-66). senic in the atmosphere. The sources would include The preceding representative examples of skin primarily the burning of fuels and secondarily in tumor production in mice, coupled with analytical secticide use and metallurgical sources. Goulden studies previously mentioned, provide direct bio (45) and his co-workers measured the arsenic con logic evidence of carcinogenic materials belonging tent of the atmosphere at eight sites in England to the aromatic polycyclic hydrocarbon group of during the winter and summer. They computed compounds in several pollution sources and in the from their data th at approximately 0.5 mg. of atmosphere per se. The traditional carcinogenicity arsenic as arsenic trioxide would be respired during of these compounds is well known. In 1955 Kotin the course of 1 year. While this amount is low, its and his co-workers reported the successful produc possible role in association with other known car tion of skin tumors in mice using aromatic poly cinogens should be kept in mind. cyclic hydrocarbon-free atmospheric extracts (61). The concentration of metals in the atmosphere The samples used for the tumor production con has been determined by investigations of the Stan sisted of oxidation products of aliphatic hydro ford Research Institute (93) for Los Angeles and carbons formed in the atmosphere in accordance by Chambers and his co-workers (13) for several with the theory as developed by Haagen-Smit (48, other American cities. The amounts present, espe 49) and since confirmed by others (103). He postu cially of chromium and nickel, are minimal. As in lated a primary photochemical reaction between the case of arsenic, however, possible additive or oxides of nitrogen (a product of internal combus synergistic effects with hydrocarbon carcinogens tion engine exhaust) and organic molecules: alco should not be forgotten. hols, aldehydes, ketones, acids, and hydrocarbons, Perhaps the most controversial of atmospheric both as emitted from vehicular exhaust and as in factors of theoretical carcinogenic significance are troduced into the atmosphere through the volatili those relating to ionizing radiation. In a compre zation of uncombusted gasoline. Ozone forms as a hensive study of radioactive material in the a t result of a radical chain reaction. While less is mosphere carried out in London, in Manchester, known about the reaction products other than and in the country a t Rothamsted, Dawson (24) ozone, the ozone itself is known to react spontane concluded that no considerable difference existed ously with unsaturated molecules which also are between urban and rural districts. Day-to-day present in gasoline engine exhaust and unburned variations were marked and could be related to the gasoline. The reaction products consist of an activity of the atmosphere in terms of wind veloci aerosol which was used for skin painting. ty: "The more stationary the air, the greater the The products of this subsequent reaction (ozone activity." At all times amounts present were ex plus hydrocarbon) occurring in the atmosphere are ceedingly small in comparison with the lowest more completely understood than those of the concentration considered harmful to man. primary free radical reaction. This reaction is in D. Biologic demonstrations of carcinogens withstantaneous and produces an ozonide which in the atmospheric 'pollutants.---Biological demonstration presence of moisture gives rise to a variety of of carcinogenicity with materials which are fre peroxides. These compounds result in aldehydes quent sources of air pollution was successfully and acids. The peroxides react subsequently with undertaken by Passey (84), who in 1922 reported aldehydes and acids, producing peracids which the production of experimental cancers in mice then react with unsaturated hydrocarbons to yield following painting with ether extracts of household epoxides. A simplified schematic representation of chimney soot. Campbell (11) similarly produced these reactions is shown in Chart 5. I t should be skin tumors in mice following painting with tars noted, however, th at many additional side reac extracted from chimney soot. Leiter (70, 71) and tions take place simultaneously, reducing the yield his co-workers reported the development of sub of these compounds. The tumor yield with these cutaneous sarcomas in mice injected with tars compounds was less than that observed with a t extracted from the atmospheric dusts of several mospheric extracts containing aromatic polycyclic American cities. The extracted tars were injected hydrocarbons. in amounts varying from 21 to 71 mg. suspended The broadening of the spectrum of carcinogens in 0.25 cc. of tricaprylin. More recently, Kotin and in the atmosphere with these agents is of special his co-workers reported the production of skin significance in th at their entry into the environ- K otin-- Atmospheric Pollution and Pulmonary Cancer 387 jnent is compatible with their action as initiating or promoting agents in the increasing incidence of cancer of the lung. Attempts to induce or increase the yield of pul monary tumors by inhalation experiments were first reported by Campbell (8-10), who exposed mice in inhalation chambers to resuspended sweep ings of dust from tarred roads. In addition to the development of cutaneous tumors, he noted a high er incidence of pulmonary tumors in his test mice than in his controls. Seelig and Benignus (90, 91), in one of two ex periments, used chimney soot as an inhalant for Buffalo strain mice and reported an 8 per cent pri mary' pulmonary tumor yield in contrast with a 2 of exposure and ending with the 52d week. At all intervals the number of tumor-bearing mice, the number of multiple tumor-bearing mice, and the total tumors in the test chamber were significantly greater than those in the control chamber. Kotin further has demonstrated the production of pul monary tumors in C57BL mice exposed to a simi lar atmosphere of ozonized gasoline.1 Bronchogenic cancers, apparently of the type seen clinically, have thus far been produced ex perimentally only in rats. Vorwald (99i exposed rats for over 1 year to an atmosphere containing soluble and insoluble beryllium salts and noted the development of true bronchogenic neoplasms. Dutra (37) produced osteogenic sarcomas in ex- H1 i; ' i .1 . /MDERUNeo Co m p o u n d s SUSPR-CTCD O r C/4&C/A/QOfAS/C/ Tv Chart 5.--Schematic representation of oxidation reactions of aliphatic compounds per cent yield in their controls. In a second experi perimental animals exposed to an atmosphere con ment they adsorbed gas-work ta r onto carbon taining beryllium oxide. Lisco and Finkel (75) ob black for inhalation by C57BL mice. No pulmo served neoplastic changes in the bronchial epi nary' tumors were produced in the test mice. thelium in rats exposed to an aerosol of radioactive > McDonald and Woodhouse (77) exposed mice cerium. The carcinogenic powers of ionizing radia of apparently indifferent strains to dust obtained tion for tissues other than the lung are well known. from city thoroughfares and to dust collected from the air-purifying system of a hospital adjoining an DISCUSSION industrial area. While they reported th at they A. Epidemiologic considerations in interpreta were unable to show the striking increase in their tion of pathogenesis.--The epidemiologic data re test mice th at Campbell reported in a similar ex lating to lung cancer as reviewed here show periment, they did show an exaggerated produc marked contradictions. The inconsistencies appear tion of pulmonary adenomas in their test mice. to be wholly irreconcilable if but a single initiating Si Kotin and Falk (63) reported an exaggerated and promoting agent were to be regarded as being i incidence of pulmonary tumors in strain A mice responsible for the increasing incidence of lung exposed to an atmosphere of ozonized gasoline. cancer. Of the two suggested major etiologic fac They removed mice from the inhalation chamber tors, cigarette smoking alone appears the least i at 4-week intervals, beginning with the 24th week 1P. Kotin, unpublished data. lt,, S88 Cancer Research capable of adaptation to the panorama of lung cancer as it is currently manifest. The limitations of the tobacco concept of etiology are evident in studies showing differences in lung cancer rates on the basis of socio-economic status. Geographic studies singling out urbanization as the exclusive variable in groups with contrasting lung cancer rates cast further doubt on the validity of the major role assigned to tobacco in pulmonary can cer. The urban-rural difference in incidence could very possibly, in one or even several countries, be a manifestation of difference in smoking habits. I t is unlikely, however, that peoples with different cultures, economies, and mores should have iden tical smoking habits and patterns. More readily tenable is the hypothesis that urban-rural differ ences are a true reflection primarily of the differ ence in the cleanliness of the atmospheric environ ment. D ata already referred to in the text as a basis for this observation are available from epidemio logic studies made in Norway (68, 69), New Zea land (38), the United States (31, 32, 43, 44, 55), and Denmark (14-17). These data emphasize dif ferences in urban-rural cancer rates and the great er incidence in the low socio-economic groups. Clemmessen interprets differences in rural-urban incidences as evidence of difference in the onset of exposure to the carcinogenic agent. A more likely explanation would be th at the concentrations of carcinogens in the atmosphere in rural areas are lower as a result of fewer sources of atmospheric pollution and dilutions of the pollutants carried to the country by winds from the city. Accepting dif ferences in urban-rural rates as a function of a t mospheric pollution, one can readily reconcile the parallelisms noted between lung cancer rates and density of population and degree of urbanization. Not the least significant of the epidemiologic con siderations concerned with atmospheric pollution are the documented instances of occupational lung cancer. The parameters of the exposure and the response of the host are defined within measurable limits. Sex differences in lung cancer rates are not irreconcilable with the concept of atmospheric pollution as a major etiologic factor. The contrast ing social and economic roles of men and women result in a more prolonged exposure to a greater concentration of polluted atmosphere on the part of men. T he conviction th at air pollutants possess the ability to initiate and promote pulmonary cancer in no way precludes the role of other possible fac tors. Stocks and Campbell (98) in their recent study of the combined effect of smoking and air pollution on lung cancer death rates demonstrated an urban-rural ratio of 9:1 for nonsmokers residing in Liverpool as compared with those living in adjacent rural areas. At every level of smoking in tensity, Liverpool rates exceeded the rural rates. As smoking increased, the disparity progressively decreased to a level where the difference ap proached a factor of less than two. They concluded that a dual role now appears to be tenable as a working hypothesis for the guidance of future studies. Since not all residents in any area smoke and the very method of smo king varies from indi vidual to individual, attempts to correlate lung cancer incidences with tobacco consumption on a per capita basis in any given population group denote a type of epidemiologic gerrymandering. In a given community, however, a constancy in exposure to atmospheric pollution or lack of ex posure can be related to residence site, occupation, and duration of these two. The inconsistencies and the irreconcilable fac tors in the epidemiologic data make it amply ap parent to this reviewer that the contributions of laboratory data will be of ever increasing signifi cance in the ultimate revelation of the mechanisms and etiologies of lung cancer. B. Role of chemical and 'physical factors in patho genesis.--Of equal significance to the epidemio logic data are the chemical and physical data re lating atmospheric pollution to the pathogenesis of lung cancer. The demonstration of 3,4-benz pyrene in urban atmospheres has been successful wherever undertaken. On the basis of current knowledge, the concentration and the atmospheric survival of this carcinogen are sufficiently great to postulate a biological effect in humans. The recent demonstration by Kotin and his co-workers of the carcinogenicity of oxidation products of aliphatic hydrocarbons has-made the incrimination of the atmosphere more certain. The ubiquity of gaso line, the most common source of these compounds, combined with the temporal aspects of the intro duction of liquid fuels permits of epidemiologic integration as well. The combination of carcinogenic aromatic poly cyclic hydrocarbons, oxidation products of ali phatic hydrocarbons, and known occupational carcinogenic agents which pollute the atmosphere warrants suspicion of the atmosphere as a factor in the pathogenesis of human lung cancer. Com plete ignorance of any possible additive or syner gistic effects they may manifest and the absence of exactitude of dosage in man make their diligent study imperative. The presence of 3,4-benzpyrene in the combus tion products of tobacco has been reported. Cooper and Lindsey (21) recovered 4 yg. of this carcino genic agent from the smoke of 500 cigarettes. This K otin--Atmospheric Pollution and Pulmonary Cancer 389 concentration is less than th at noted in the atmos phere in terms of total amounts respired by those exposed. I t is clear that the alleged cause-andeffect relationship ascribed to smoking cannot cur rently be predicated on the presence of 3,4-benz pyrene in tobacco smoke, Assuming that the aver age total of inspired air measures up to 7,500 cubic meters per year as calculated by Stocks (98), the average resident of Liverpool, whether smoker or nonsmoker, would inhale 450 pg. of benzpyrene. This does not take into account the other presuma bly carcinogenic materials in the air. Physical factors play a significant role in the deposition of particulate m atter in the res piratory tract. The per cent retention of particu late m atter in the lung on the basis of particle size is shown in Charts 6 and 7. Dautrebande (23) cor related the location of particulate deposition with specific sites in the tracheobronchial tree. The par ticle size of the carcinogenic material in the Los Angeles atmosphere as measured by Kotin1 and others is wholly consistent with penetration and settling out of these particles distal to the trachea. The relative absence of primary tracheal carcino ma is not an indication of local tissue immunity but rather reflects the failure of particles to settle out. Those particles capable of settling out on the tracheal mucosa are identical in size with those trapped in the nose, epipharynx, and accessory nasal sinuses. Particles of smaller size down to approximately 0.25 p settle progressively distally in the arborization of the bronchi. Those less than 0.25 p remain suspended in tidal air until a suffi ciently miniscule size is reached so that Brownian movement can produce precipitation. C. Significance of experimental laboratory data in pathogenesis.--As suggested above, the incon sistencies and contradictions in the epidemiologic data placed increasing responsibility and emphasis on the laboratory for the elucidation of tiologie C hart 6.--Per cent retention of particulate m atter in lung in relation to particle size. TPACNEA .0 7 B r ONCH/OL ES .3 9 A Y O l ) 2 .0 B Bronchi J2 P A S S A G E TIM E ( S e c .) Chart 7.---Anatomical pattern of deposition and time of passage of particulate m atter in relation to particle size (after Dautrebande). 390 Cancer Research agents and mechanisms for lung cancer develop In addition to the eluting role and carcinoee i ment. The chemical demonstration of carcinogenic ty demonstrated for these aliphatic materials t| " agents in the environment and their successful use have been shown to be potent irritants for for the production of tumors in experimental ani respiratory mucosa. In common with other * mals do not prove or even especially strongly sug mospheric irritants introduced secondary to inti gest a like relationship in the instance of man. trialization, these materials transiently affect t l ' When, however, a demonstrable parallelism exists respiratory mucosa by interfering with cilin** between epidemiologic data and laboratory find activity and normal mucous secretion. The int/^ ings, greater significance accrues to both. Medical ference with these normal resistance factors tu, history is replete with examples in which labora mits the accumulation of particulate matter I tory findings have been proved ultimately to have selected sites in the tracheobronchial tree their counterpart in the human experience. Excep shown by Kotin. The sites are chiefly at bifurcu, tions have been very few. tions and angulations in the respiratory arboriia, D. A theoretical mechanism for the pathogenesistions. This accumulation allows prolonged re*[, of lung cancer.--By means of integrating labora dence of particles on the respiratory epithelium F tory findings and experimental data, the author The elution of carcinogens is facilitated, as hn| F has arrived at a theoretical mechanism for the been demonstrated by analytical, procedures. Dif. ti< epnactheoogfesnoeostisinofthluenagtmcaonspcehre.reInfosrpmiteanoyf ctheentpurrieess fpuisraiotonroyfeipnittahcetlisuomo,t apsarhtaicslebseeonccduerms oinntsotrathteedreb*y to and its association with cutaneous cancer, it is only electron microscopy. A study of the respiratory 8>' within the last several decades th at lung cancer epithelium in lungs obtained at necropsy from un. C has become a problem of epidemic proportions. selected cases shows th at the most common site Skin cancers as occupational responses to soot of metaplasia are similar to those in which soot have long been known. I t has been noted by accumulates in experimental animals.2 Steiner (94) and Falk and Steiner (41) that car In essence, it has been demonstrated that car cinogenic hydrocarbons adsorbed on soot are bio cinogens have become significant in the patho. logically ineffective until they are separated by genesis of pulmonary cancer within the past half- elution from the soot particles. I t has been postu century in spite of their atmospheric presence for lated th at in the case of the skin, sebaceous secre several centuries. This transformation may be tions provide the necessary polar substances for attributed to (a) the atmospheric presence of car- the elution of the carcinogenic materials. Experi cinogens in a size range consistent with their res mentally, Kotin and his co-workers (62) have piration and retention within the lung; (6) the in demonstrated poor elution powers for the mucous troduction into the atmosphere of polar substance secretions in the respiratory tract. However, the capable of eluting adsorbed carcinogens from soot introduction of aliphatic polar compounds into the particles following their deposition in the respira atmosphere makes available an environmental tory mucosa; (c) interference with the normal de source of an eluting agent capable of entry into the fensive mechanisms of the bronchial epithelium by respiratory tract. abnormally affecting ciliary motion and mucous The simultaneous respiration of the carcinogen- secretion; (d) the introduction of cancerigenic laden soot and the eluting agent provides a mech nonaromatic polycyclic hydrocarbon agents into anism for biological activity of the carcinogen. the air. Included in this group are aliphatic hydro This mechanism is unlikely to occur when soot carbons and their oxidation products, metals, in alone is breathed. organic dusts, and probably macromolecular sub Aliphatic hydrocarbons and their oxidation stances. products in the air were further incriminated in the Any conceivable role of tobacco smoking in the pathogenesis of cancer when it was shown th at pathogenesis of lung cancer appears to this re they could induce both skin cancers and alveolo- viewer to be at the level of a nonspecific irritant or genic carcinomas in mice. Cutaneous cancers re eluting agent for previously deposited carcinogenic sulted from the painting of mice, and lung tumors agents. There is a t present no convincing evidence were produced by inhalation. The dual capacity th at tobacco possesses the necessary qualifications shown for these compounds makes them suspect for the initiation and promotion of lung cancer. on a laboratory basis. When the laboratory data CONCLUSION are considered along with epidemiologic data, it is apparent th at these aliphatic materials were intro Pulmonary cancer in common with all neo i duced into the atmospheric environment at a time plasms may be properly regarded as having several consistent with their having tiologie significance. 2D . T a tte r, E . M. B u tt, a n d P. K otin, unpublished data. fr***"" arcinogenicitterials, they ants for the :h other at ari- to indusIy affect the with ciliary i. The interfactors pere matter at aial tree as v at bifurca>rv arborizailonged resiepithelium. ated, as has dures. Difinto the resonstrated by * respiratory psy from unominon sites 1 which soot 2 ted that cari the pathohe past half presence for ion may be sence of carith their resg; (6) the iniar substance ns from soot i the respirae normal depithelium by and mucous cancerigenic agents into >hatic hydro- metals, inolecular sub- oking in the s to this re lic irritant or carcinogenic :ing evidence lualiii cations mg cancer. ith all neoaving several published data. K otin--Atmospheric Pollution and Pulmonary Cancer 391 factors concerned with its initiation and promo tion. A review of the factors relating atmospheric pol lution to lung cancer both on the epidemiologic diagnosi e indirizzi terapeutici. Lotta C. Tuberc., 25:519, 1955. 13. Chambers, L. A.; Foter, M. J.; and Cholak, J. A. A Comparison of Particulate Loadings in the Atmosphere of Certain American Cities. Proc. 3d Nat;. Air Pollution and experimental levels warrants its incrimination as one of the dominant agents etiologically asso ciated with the increase in mortality from lung cancer now being reported in various regions of the world. Symposium, pp. 25-31, 1955. 14. Ci.emmessen, J . Bronchial Carcinoma--a Pandemic. Danish Med. Bull. 1:37-46, 1954. 15. ------- . Bronchial Carcinoma--a Pendeimc. II. Incidence and Tobacco Consumption in Various Countries. Ibid., pp. 194-99. Other agents, including the frequently accused 16. Clemmessen, J., and N ielsen, A. Social Distribution of excessive use of tobacco, appear to be capable of playing only a secondary role in the increase in lung cancer rates. They may conceivably act as promoting agents so that in the presence of a pre- Cancer in Copenhagen, 1943 to 1947. Brit. J. Cancer, 5:159-71, 1951. 17. Clemmessen, J.; N ielsen, A.; and J ensen, E. The In crease in Incidence of Carcinoma of the Lung in Denmark, 1931 to 1950. Brit. J. Cancer, 7:1-9, 1953. i pared or initiated soil they can act either synergis- 18. Clkmo, G. R., and M iller, E. W. The Carcinogenic tically, as additives, or as cocarcinogenic agents. Refinements in both epidemiologic and labora tory data are indicated as are the new methodolo gies for the study of the phenomenon of lung cancer. Action of City Smoke. Chem. & Ind., p. 38, 1955. 19. Cohart, E. M. Socioeconomic Distribution of Cancer of the Lung in New Haven. Cancer, 8:1126-29, 1955. 20. Cooper, R. L., and L indsey, A. J. Atmospheric Pollu tion by Polycyclic Hydrocarbons. Chem. & Ind., pp. 1177-78, 1953. ACKNOWLEDGMENTS 21. -- -----. 3 :4-Benzpyrene and Polycyclic Hydrocarbons in Cigarette Smoke. Brit. J. Cancer, 49:304-9, 1955. The assistance of Dr. Hans L. Falk is gratefully acknowl 22. Curwen, M. P.; K ennaway, E. L.; and K ennaway, edged in relation to the chemical phases of this review. N. M. The Incidence of Cancer of the Lung and Larynx in Urban and Rural Districts. Brit. J. Cancer, 8:181-98, REFERENCES 1954. 23. Dautrebande, L. Physiological and Pharmacological 1. Baetjer, A. M. Pulmonary Carcinoma in the Chromate Characteristics of Liquid Aerosols. Physiol. Rev., 32: Workers. I. A Heview of the Literature and Report of the 214-51, 1952. ' Cases. A.M.A. Arch. Indust. Hyg. & Occup. Med. 24. D awson, K. B. Radioactive Material in the Atmosphere. 2:487-504, 1950. Brit. J. Cancer, 6:22-31, 1952. 9. -------- . Presentation a t Conference on Environmental 25. D enoix, P. F., and Gelle, X . Estimation de l'impor- Carcinogens. National Cancer Inst., Sept. 12-18, 1955. tance compares du cancer broncho-pulmonaire en France 8. B erk so n , J. The Statistical Study of Association between et dans d 'autres pays. Bull, du Cancer, 42:247-78, 1955. Smoking and Lung Cancer. Proc. Staff Meetings Mayo 26. Doll, R. The Causes of D eath among Gas-Workers with Clinic, 30:319-48, 1955. Special Reference to Cancer of the Lung. Brit. J . Indust. 4. Bidstrvp, P . L. Carcinoma in Lung of Chromate Work Med., 9:180-85, 1952. ers. Brit. J. Indust. Med., 8:302-5,1951. 27. -------- . Mortality from Lung Cancer among Non- 5. B lacklock, J. W. S.; K ennaway, E. L.; Lewis, G. M .; Smokers. Brit. J. Cancer, 7:303-12, 1953. 1 and L rqchart, M. E. The Carbon Content of the Hu 28. -------- . Bronchial Carcinoma: Incidence and Aetiology. man Lungs and Bronchial Glands. Brit. J. Cancer, 8:40- Brit. M. J., 4835:521-37, 4836:585-90, 1953. oo, 1954. 6. Bonsek, G. M .; F aulds, J. S.; and Stewart, M . J. 29. -------- . Mortality from Lung Cancer in Asbestos Work Occupational Cancer. Am. J. Clin. Path., 25:126-34, ers. Brit. J. Indust. Med., 12:81-86, 1955. 1955. SO. Doll, R., and Hill, A. B. A Study of the Aetiology of 7. Breslow, L.; H oaglin, L.; R asmussen, G.; and Abrams, Carcinoma of the Lung. Brit. Med. J., 2:1271-86, 1952. H. K. Occupations and Cigarette Smoking as Factors in 31. D orn, H. F. I s Lung Cancer on the Increase--Evalua Lung Cancer. Am. J . Pub. Health, 44:171-81, 1954. tion of Present Day Evidence. Cancer of the Lung--an 8. Campbell, J. A. Cancer of the Skin and Increase in Evaluation of the Problem; Proc. Scientific Session Ann. Incidence of Primary Tumours of Lung in Mice Exposed Meeting Am. Cancer Society, Nov. 3--4, 1953; New to D ust Obtained from Tarred Roads. Brit. J. Exper. York: Am. Cancer Society, Inc., pp. 5-15, 1954. Path., 16:287-94, 1934. 3 2 . ------- . Morbidity and M ortality from Cancer of the 9. ---------. The Effects of Exhaust Gases from Internal Lung in the United States. Acta Union internat. contre Combustion Engines and of Tobacco Smoke upon le cancer, 9:126-85, 1953. ' Mice with Special Reference to the Incidence of Tumors 33. D uchen, L. W. Bronchogenic Carcinoma. Incidence of the Lung. Ibid., 17:146-58,1936. and Pathology as Seen on the Witwatersrand. S. Afr. J. 10. -------- . Carcinogenic Agents Present in the Atmosphere Med. Sc., 19:65-74, 1954. and Incidence of Primary Lung Tumours in Mice. 34. Dungal, N . Carcinoma of the Lung in Iceland. Lancet, Ibid.i 20:122-32, 1939. 2:245-47, 1950. 11. - -. Lung Tumore in Mice, Incidence as Affected by 35. -. Cancer in Iceland. Ann. Boy. Coll. Surg. Eng- Inhalation of Certain Carcinogenic Agents and Some land, 16:211-26, 1955. Dusts. Brit. M. J., 1:217-21, 1942. 86. D unn, H. L. Lung Cancer in the Twentieth Century. 12. Ceino, O., and B ombo, V. Incidenza della mortalit J. Internat. Coll. Surg., 23:826-42, 1955. per tumori maligni dell'apparato respiratorio in Italia: 37. D utra, F. R .; Laroent, E. J.; a mi R oth, 3. L. Osteo- 392 Cancer Research genic Sarcoma after Inhalation of Beryllium Oxide. Arch. Path., 51:474-79, 1951. 38. Eastcott, D . F. The Epidemiology of Lung Cancer in New Zealand. Lancet, 1:37-39, 1956. 39. F alk, H. L., and K otin, P. The Fate of Polycyclic Aromatic Hydrocarbons in Soot Following Their Emis sion into the Atmosphere and Deposition in the Lungs. Proc. Am. Assoc. Cancer Research, 2 (1): 15, 1955. 40. Falk, II. L.; M ahkul, I.; and Kotin, P. Aromatic Hydrocarbons. IV. Their Fate Following Emission into the Atmosphere and Experimental Exposure to Washed Air and Synthetic Smog. A.M.A. Arch. Indust. Health, 13:13-17, 1956. 4L F alk, H. L., and Steiner, P. E. The Identification of Aromatic Polycylic Hydrocarbons in Carbon Blacks. Cancer Research, 12:30-39,1952. 42. F owler, R. Some Observations on the Epidemiology of Lung Cancer. Med. J. Australia, 1:485-94, 1955. 43. Gilliam, A. G. Opportunities for Application of Epidemi ologic Method to Study of Cancer. Am. J. Publ. Health, 43:1247-57, 1953. 44. -------- . Mortality Attributed to Lung Cancer in the Large Cities of the United States in 1948 and 1949. J. N at. Cancer Inst., 15:1307-12, 1955. 45. Goulden, F.; K ennaway, E. L.; and Urqubart, M. E. Arsenic in the Suspended M atter of Town Air. Brit. J. Cancer, 6:1-7, 1952. 46. Goulden, F., and T ipler, M. M. Experiments on the Identification of 3 :4-Benzpvrene in Domestic Soot by Means of the Fluorescence Spectrum. Brit. J. Cancer, 3:157-60, 1949. 47. Graham, E. A. Remarks on the Aetiology of Broncho genic Carcinoma. Lancet, pp. 1305-8, 1954. 48. IIaagen-Smit, A. J. The Air Pollution Problem in Los Angeles. Engin. & Sc. (Calif. Inst. Tech.), 14:7-13, 1950. 49. -------- . Chemistry and Physiology of Los Angeles Smog. Indust. & Eng. Chem., 44:1342-46, 1952. 50. Hammond, E. C., and Horn, D. The Relationship be tween Human Smoking Habits and Death Rates. J.A.M.A., 155:1316-28, 1954. 51. Harting, F. H., and H esse, W. Der Lungenkrebs, die Bergkrankheit in den Schneeberger Gruben. Vierteljahrsschr. Gerichtl. Med. u. Offentl. Sanittsw. (N.S.), 30:102-32, 296-309, 1879. 52. Henschen, F. Ueber Vernderungen im Krankheits panorama Schwedens whrend der letzten 50 Jahre. Schweiz. Med. Wchnschr., 77:968,1947. 53. H offman, E. F., and Gilllam, A. G. Lung Cancer Mor tality. Geographie Distribution in the United States for 1948-1949. Publ. Health Rep., 69:1033-42, 1954. 54. H ueper, W. C. Recent Developments in Environmental Cancer. A.M.A. Arch. Path., 58:360-523, 645-82, 1954. 55. -----;-- . A Quest into Environmetal Causes of Cancer of the Lung. Pub. Health Mono. 36, U.S.P.H.S., 1956. 56. J ames, G.; P atton, R . E.; and H eslin, A. S. Accuracy of Cause-of-Death Statements on D eath Certificates. Pub. Health Rep., 70:39-51, 1955. 57. Kennaway, E. L. The D ata Relating to Cancer in the Publications of the General Register Office. Brit. J. Cancer, 4:158-72, 1950. 58. K ennaway, N . M., and K ennaway, E. L. Further Study of Incidence of Cancer of Lung and Larynx. Brit. J. Cancer, 1:260-98, 1947. 59. Kihkun, J . W.; M cD onald, J. R .; Clagett, O. T.; M oeksch, H. J.; and Gage, R . P. Bronchogenic Carci noma: Cell Type and Other Factors Relating to Prog nosis. Gynec. & Obst., 100:429-38, 1955. 60. Korteweg, A. M ortality from Lung Cancer in tli Netherlands during and after the Last War. Brit. Cancer, 8:34-39, 1954. 61. Kotin, P., and F alk, H. L. Production of Tumors i C57BL Mice with Atmosphere-extracted Aliphati Hydrocarbons. Proc. Am. Assoc. Cancer Researcl 2(1) :30, 1955. 6 2 . ------- . Local and Systemic Responses to Experiment;. Exposure to Atmospheric Pollutants. Proc. Third. Nat Air Pollution Symposium, April 18-20, 1955, Pasadena Calif. 6 3 . --------. The Experimental Induction of Pulmonar Tumors in Strain A Mice Following Their Exposure t an Atmosphere of Ozonized Gasoline. Cancer (in press). 64. Kotin, P.; F alk, II. L.; M ader, P.; and T homas M. Aromatic Hydrocarbons. I. Presence in the Lo Angeles Atmosphere and the Carcinogenicity of Atmos pheric Extracts. A.M.A. Arch. Indust. Hyg. & Occup Med., 9:153-63, 1954. 65. Kotin, P.; F alk, H. L.; and T homas, M. Aromati. Hydrocarbons. II. Presence in the Particulate Phase o Gasoline-Engine Exhausts and the Carcinogenicity o Exhaust Extracts. A.M..A. Arch. Indust. Hyg. & Occup Med., 9:164-77, 1954. 6 6 . ------- . Aromatic Hydrocarbons. III. Presence in thi Particulate Phase of Diesel-Engine Exhausts and tin Carcinogenicity of Exhaust Extracts. A.M.A. Arch Indust. Health, 11:118-20, 1955. 6 7 . ------- . Production of Skin Tumors in Mice with Oxidation Products of Aliphatic Hydrocarbons. Cancer (ii press). 68. Kreyberg, L. The Occurrence of Lung Cancer in Nor way. Brit. J. Cancer, 8:209-14, 1954. 6 9 . ------- . The Geographical Distribution of Histologic;; Subgroups of Primary Epithelial Lung Tumours in Nor way. Brit. J. Cancer, 8:599-604,1954. 70. Leiter, J., and Shear, M. J. Production of Tumors ii. Mice with Tars from City Air Dusts. J. N at. Cancer Inst., 3:167-74, 1942. 71. Leiter, J.; Shimkin, M . B.; and Shear, M. J. Produc tion of Subcutaneous Sarcomas in Mice with Tar; Extracted from Atmospheric Dusts. J. N at. Cancer Inst., 3:155-65, 1942. 72. Levin, M. L. Etiology of Lung Cancer: Present Status. N . Y. State J. Med., 64:769-77, 1954. 73. Lew, E. A. Cancer of the Respiratory Tract. J . In te rn a l Coll. Surg., 24:12-27, 1955. 74. Lilienfeld, A. M . Possible Existence of Predisposing Factors in the Etiology of Selected Cancers of Nonsexual Sites in Females. Cancer, 9:111-22, 1956. 75. Lisco, H., and F inkel, M. P. Observations on Lung Pathology Following Inhalation of Radioactive Cerium. Fed. Proc., 8:360-61, 1949. 76. Lynch, K., and Smith, W. A. Carcinoma of the Lung in Asbestos-Silicosis. Am. J. Cancer, 24:56-63, 1935. 77. M cDonald, S., and Woodhouse, D. L. On the Nature of Mouse Lung Adenomata with Special Reference to Effects of Atmospheric D ust on Incidence of These Tumours. J. Path. & Bact., 64:1-12,1942. 78. M achle, W., and Gregorius, F. Cancer of the Respira tory -System in the United States Chromate-Producing Industry. Pub. Health Rep., 63:1114-27,1948. 79. Mancuso, T . F .; M acFarlane, E. M .; and P orter field, J . D . The Distribution of Cancer M ortality in Ohio. Am. J. Pub. Health, 46:58-70, 1955. 80. M erewether, E. R. Annual Report of Chief Inspector of Factories for the Year 1947, pp. 79-81; 1948, pp. 92-93. m y - * K otin--Atmospheric Pollution and Pulmonary Cancer 393 J 81. M llek, F. H. Tabakmissbrauch und Lungencarcinom. 94. Steiner, P. The Conditional Biological Activity of the I, Ztscbr. Krebsforsch., 49:57-85, 1939. Carcinogens in Carbon Blacks and Its Elimination. Can 82. National Office of Vital Statistics Cancer D eath Rates cer Research, 14:103-10, 1944. for Each State in the United States by Site, 1948. 95. Steiner, P. E.; Butt, E. M.; and E dmondson, H. A. 83. Pascua, M. Increased M ortality from Cancer of the Pulmonary Carcinoma Revealed at Necropsy with Refer Respiratory System. Bull. World Health Org., 12:687- ence to Increase in Incidence in Los Angeles County 704, 1955. Hospital. J. N at. Cancer Inst., 11:497-510, 1950. 84. Passey, R. D . Experimental Soot Cancer. Brit. Med. J., 96. Stocks, P. Endemiology of Cancer of the Lung in Eng 2:1112-13, 1922. land and Wales. Brit. J. Cancer, 6:99-111, 1952. , 85. P erry, K.; Bowler, R . G.; B uckell, H. M.; D ruett, 97. -------- . Lung Cancer, Tobacco and Atmospheric Pol H. A.; and Schilling, R. S. F. Studies in the Incidence lution. Brit. Emp. Cancer Camp., 31st Ann. Rep., pp. of Cancer in a Factory Handling Inorganic Compounds 410-12, 1953. of Arsenic. Brit. J. Indust. Med., 5:1-15, 1948. 98. Stocks, P., and Campbell, J. M. Lung Cancer D eath 86. Phillips, A. J. M ortality from Cancer of the Lung in Rates among Nonsmokers and Pipe and Cigarette Smok Canada, 1931-1952. Can. Med. Assoc. J., 71:242-44, ers. Brit. M. J,, 2:923-29, 1955. 1954. 99. Vorwald, A. J. Pulmonary Cancer in Experimental j 87. Pott, P. Chirurgical Observations Relative to the Cata- Exposures to Beryllium. Proc. Seventh Saranac Sym f ract, the Polypus of the Nose, the Cancer of the Scrotum, posium, Saranac Laboratory, Saranac Lake, N.Y., 1952. the Different Kinds of Ruptures, and the Mortification 100. Waller, R. E. The Benzpyrene Content of Town Air. of the Toes and Feet. Printed by: T. J. Carnegie, 1775; Brit. J. Cancer, 6:8-21, 1952. Hawes, Clarke, and Collins, London. 101. Walter, J. B., and Bryce, D . M. The Histology of Lung 88. Rigdon, R. H.; Kirchoff, II.; and M artin, N . Where Cancer. Thorax, 10:107-16, 1955. Cancer of the Lung Occurs in Texas. Tex. Rep. Biol. & 102. Warwick, O. II., and P hillips, A. J. Cancer among Med., 13:102-69, 1955. Canadian Indians. Brit. J. Cancer, 8:223-30, 1954. 89. Schrek, R.; Baker, L. A.; Ballard, G. P.; and Kol- 103. Weaver, E. R ., and Gunther, S. Condensable Im puri ) goff, S. Tobacco Smoking as an Etiologic Factor in ties in the Air of Los Angeles and Vicinity. Proc. 3d Disease. I. Cancer. Cancer Research, 10:49-57, 1950. N at. Air Pollution Symposium, pp. 86-96, 1955. 90. Sellig, M. G., and Benignus, E. L. Coal Smoke, Soot 104. Weil, C. S.; Smyth, II. F.; and N a le, T. W. Quest for a and Tumors of the Lung in Mice. Am. J. Cancer, 28: Suspected Industrial Carcinogen. A.M.A. Arch. Indust. 1 96-111, 1936. Hyg. & Occup. Med., 5:535-47, 1952. 91. ------. The Production of Experimental Cancer of the 105. Wynder, E. L. Studies on Lung Cancer in Relation to Lung in Mice. Ibid., 33:549-54, 1938. Smoking. A.M.A. Arch. Indust. Hyg. & Occup. Med., 6:218-27, 1952. 92. Sibton, J., and H adida, A. Role du tabae dans l'evo1lu06- . Wynder, E. L., and Graham, E. A. Tobacco Smoking as ' tion des cancers larynges et bronchopulmonaires. Bull. Possible Etiologic Factor in Bronchiogenic Carcinoma: Alger. Carcinol., 6:136-47, 1953. A Study of 684 Proved Cases. J.A.M.A., 143:329-86, 93. Stanford Research Institute: The Smog Problem in Los 1950. Angeles County, pp. 106-8, 1954.