TABLE `I.-Mortality rates for lung cancer and cancer of the
   respiratory tract for white females in the United
   States per 100,000 population for selected years, 1940
        to 1976

Yew               Lung and Bronchus      Respiratory System

1940                          -                   3.6
1945                          -                   4.6
1950                           4.7                   5.4
1955                           5.1                 5.7
1960                          5.9                  6.4
1965                           8.0                 8.6
1970                        12.3                  13.1
1975                         17.8                  18.8
1976                        19.5                  20.5

SOURCE: National Center for Health Statistica (150)

TABLE 8.-Percent of adult population who were current cigarette
    smokers in selected years in the United States
                            Percent smokers
Year                        Females               Males

1964                          31.5                 529

1966                          33.1                 51.9

1970                          36.5                  422

1975

Percent reduction
since 1964

28.9                 39.3



2.6                 13.6

SOURCE: National Clearinghouse for Smoking and Health (155)

reduction in the number of adult females who smoke cigarettes,
whereas there has been a 13.6 percent reduction in the number of
adult males smoking. Trends in the percentage of teenagers who arc
regular cigarette smokers are presented in Table 9. Cigarette smoking
among girls has increased steadily, so that at the present time equal
numbers of boys and girls are smoking cigarettes and many of the
differences which existed in the past between male and female
smokers have disappeared.

Epidemiological Studies

Three of the large prospective epidemiological studies contain informa-
tion on lung cancer in women. Data from these studies are summarized
in Table 10. A number of retrospective studies have examined the

5-20


TABLE g--Percent of teenagers who were current cigarette
     smokers in selected years in the United States

Year

Girls

Percent smokers
Ages 12-18

1968                           8.4                  14.7


1970                          11.9                  18.5


1972                          13.3                 15.7


1974                          15.3                 15.8

SOURCE: National Clearinghouse for Smoking and Health (1.5.%x)

TABLE lo.-Lung cancer mortality ratios for women-prospective
       studies

Study         Population

Number
of
deaths

   Mortality ratio

Female        Female
nonsmokers       smokers

A.C.S. %-         562671
State Study(Q)      Females         183          1.00          22fJ

Swedish           27,732
study(SP)          Females          8          l.aO          4.56

Japanese
study(78)

142&57
Females

143          1.00          2.63

relationship of lung cancer to smoking habits in women (46 63,64,8%
122,128,13g,160,16~, 167,198,222,227,232,236,242,24247).

&se-Response Relationships

Dose-response relationships between lung cancer and cigarette smok-
ing have been described for females by the number of cigarettes
smoked per day, the degree of inhalation, and the duration of smoking.
These relationships from selected studies are presented in Tables 11
through 14. The mortality ratios are as high as 10.0 for females who
have smoked more than 20 cigarettes per day and for females who
have smoked for more than 30 years.

Path-728 of Cigarette Use

Although death rates from lung cancer are increasing at an accelerat-
ed rate in females, it may be that the peak will be somewhat less than
in males; this may be due to substantial differences in the way males

5-21


TABLE Il.-Lung cancer mortality ratios for females, by number
    of cigarettes smoked per day: A.C.S. 25state Study

Cigarettes
smoked
wr dav

Mortality
ratios

. .

Nonsmoker
l-19
20+

1.00
1.06
4.76

SOURCE: Hammond, E.C. (65)

TABLE 12.-Lung cancer mortality ratios for females, by number
     of cigarettes smoked per day: Haenszel and Taeuber

cigarettes
smoked
per day

Mortality
ratios

Nonsmoker                                     1.00
Owasional                                     1.33
1-19                                         249
Xl+                                         10.80

SOURCE: Heenscel W. (bl)

TABLE 13.-Lung cancer mortality ratios for females, by

duration of smoking: Swedish Study

Duration of
smoking in
YeaA

Mortality
ratios

Nonsmokem                                    1.0
l-29 years                                      1.6
30+ years                                     9.6

SOURCE: Cederlof. R. (W

TABLE Il.-Lung cancer mortality ratios for females, by degree
      of inhalation: A.C.S. 25-State Study

kP-=
of
inhalation

Mortality
ratios

Nonsmokers                                     1.00
None to slight                                  1.78
Moderate to deep                                 3.70

SOURCE: Hammond. E.C. (65)

5-22


and females smoke cigarettes. A recent survey (155) of cigarette
smoking behavior shows that women do not smoke as far down on the
cigarette where proportionally more nicotine and tar are inhaled. More
than 91 percent of females use filter cigarettes, compared with 80
percent of males. Females report that they do not inhale cigarette
smoke as deeply into their lungs as males do. Women also smoke fewer
cigarettes per day and select brands of cigarettes with lower tar and
nicotine yields, compared to men. In 19'75, 76.7 percent of current
female smokers smoked a pack or less per day, whereas this was true
for only 69.6 percent of males (155). In the past, women began smoking
later than men, but at the present time this is no longer true. The
available evidence suggests that women who smoke cigarettes in the
same amount and with equal depth of inhalation as men are likely to
experience death rates similar to those found in men.

Twins

The best way to control genetic factors as a potentially complicating
variable in studies of lung cancer and cigarette smoking is to conduct
the investigation in a population of twins who are discordant as to
smoking habits (one smokes, the other does not). Cederlof, et al. (33)
published new data on smoking and lung cancer from the Swedish
Twin Registries in 1977. Although the number of deaths from lung
cancer among the monozygotic twins is quite low, the trend is clear.
The authors state,  "The welldocumented evidence of a causal
association between smoking and lung cancer found in other studies
has been further supported."

Lung Cancer and the Use of Other Forms of Tobacco

Pipe and cigar smokers in the United States have experienced lung
cancer mortality rates that are somewhat higher than those of
nonsmokers but substantially lower than those of cigarette smokers
(1). Most pipe and cigar smokers report that they do not inhale the
smoke, and as a consequence the total exposure is relatively low. There
is little evidence that lung cancer is associated with the use of chewing
tobacco or snuff. These relationships are explored in detail in the
Chapter on Other Forms of Tobacco Use (specifically in Tables 15, 16,
17 and 22 of that chapter).

Histology of Lung Cancer

There are several different histologic types of lung malignancies in
humans. These include squamous cell carcinoma, adenocarcinoma,
small cell carcinoma, large cell carcinoma, bronchiole-alveolar, and
mixed and undifferentiated carcinomas of the lung. The predominent
type of carcinoma in males is squamous cell carcinoma, whereas the
most common lung cancer in females is adenocarcinoma. Over the past

5-23


15 years there has been little change in the incidence of large-cell,
bronchiole-alveolar, and mixed and undifferentiated carcinomas. There
has been an increase in adenocarcinoma and a decrease in squamous
cell carcinomas.

In 1962, Kreyberg (111~~) categorized epidermoid, small-cell, and
large-cell carcinoma of the lung as Group I and adenocarcinoma and
bronchiole-alveolar carcinoma as Group II. He noted that the risk for
smokers was substantially greater for Group I than for Group II
tumors. This view has been supported by some investigators (40, 47,
,221). Other investigators have disputed this classification (9,14,15,100,
230,254).

Weiss, et al. (230) followed the experience of 6,136 men over a lO-
year period. They found that well-differentiated squamous cell
carcinoma, small-cell carcinoma, and adenocarcinoma displayed a dose-
response relationship to smoking, but poordifferentiated squamous
cell carcinoma did not.

More recentiy, Auerbach, et al. (10) examined histologic types of
lung cancer associated with smoking habits from autopsy data on 662
men who had had lung cancer. In this study all cell types seemed to be
related to smoking to about the same degree.

Most recently, Vincent, et al. (221) reviewed the histopathology of
lung cancer in patients seen over a 13-year period at the Roswell Park
Memorial Institute. Their data indicated that adenocarcinoma is
becoming progressively more prevalent, compared to other forms of
lung cancer. They were unable to disassociate smoking as a causative
factor in any of the presently defined pathological categories of lung
cancer.

Cessation of Smoking

There is a decrease in the risk of developing lung cancer after cessation
of smoking. This decrease in risk occurs over a period of several years.
After 10 to 15 years, the risk of dying of lung cancer for ex-smokers
has decreased to point where it is only slightly above the risk for
nonsmokers. All of the major studies show this reduction in risk. The
most recent data from the British Doctor's Study are presented here
for illustration (Table 15). The mortality ratios for ex-smokers were
higher in the first year after quitting than they were for continuing
smokers. The explanation for this is that both healthy and sick
individuals quit smoking. Higher mortality is experienced by those who
quit because of illness. Lower mortality is experienced by those who
quit while experiencing apparently good health. In the US. Veterans
Study, a differentiation is made between ex-smokers who stopped
smoking on the recommendation of a doctor and those who quit for
other reasons. About 10 percent of the smokers quit because of doctors'
orders and were presumably ill. This group had much higher death
rates from lung cancer than those who stopped for other reasons.

54%


TABLE 15.-Lung cancer mortality ratios in excigarette
      smokers, by number of years stopped smoking

Year8

smokioe

Mortality
ratio

Still Smoking                                    15.8
l-4                                         16.0
5-9                                         5.9
lo-14                                         5.3
15+                                         2.0
Nonsmokem                                     1.0

SOURCE: Doll. R. (UO)

The magnitude of the residual risk which ex-smokers experience is
determined by the cumulative exposure to cigarette smoke which the
individual experienced before he quit smoking. The risk at any point in
time would be determined by the maximum amount the individual
smoked, the years since stopping smoking, the age when smoking
began, degree of inhalation, and reasons for quitting smoking. The
lung cancer mortality experience of ex-smokers is graphically present-
ed in Figure 3. The risk of developing lung cancer increases with age,
for both smokers and nonsmokers. The incidence in cigarette smokers
is much higher than in nonsmokers. It can be seen that the lung cancer
mortality of ex-smokers is initially similar to that of smokers, but, with
the passage of time, the mortality risk moves progressively closer to
that of nonsmokers. It is interesting to note that, except for the first 2
years after stopping smoking, there is a continued increase in the risk
of developing lung cancer among ex-smokers, although it is less than
that of those who continue to smoke. The slope of this line is less than
that for nonsmokers, and so there is a convergence of these two curves.

Lung Cancer and Air Pollution

A number of studies have been conducted in which the relative
influence of cigarette smoking, urban residence, and air pollution in
the etiology of lung cancer is examined. Eight of the earlier studies
were reviewed in the 1971 Report of the Surgeon General (212). More
recent publications include: "Epidemiological review of lung cancer in
man" by Higginson and Jensen (75) and a report of a task group, "Air
pollution and Cancer," edited by Cederlof, et al. (31). There have also
been studies by Doll (,&?), Weiss (229), Carnow (30), and Kotin and Falk
(109).

Lung cancer is consistently more common in urban than in rural
areas. There is only a small urban-rural lung cancer gradient for
nonsmokers. There is a much larger urban-rural gradient for smokers.
Cigarette consumption is generally greater in urban areas, but it is

5-25


1000

lnctdencr as
peer cent of
rate at time
of sro~~olny
floyscalel

            I            I            ,           1

0            5           10           15          20

Years since stoppmg

FIGURE 3.-Lung cancer mortality in continuing cigarette smok-
ers and nonsmokers as a percentage of the rate among ex-cigarette
smokers at the time they stopped smoking
SOURCE: UICC Technical Reports (24.7)

difficult to estimate how much of the excess urban mortality can be

5-26


accounted for by cigarette smoking alone. It is possible that there is an
interaction between the carcinogens in cigarette smoke and other
compounds in the ambient atmosphere.

Epidemiologic investigations  thus far indicate that the most
important cause of lung cancer is cigarette smoking and that urban
factors such as air pollution have very little independent effect on the
development of lung cancer. In the absence of cigarette smoking, the
combined effects of all atmospheric agents do not increase the death
rates for lung cancer more than a very few cases per 100,000 persons
per year.

Lung Cancer and Occupational Factors

There are several occupations (described in Chapter 7) which are
associated with the development of lung cancer and cancer at other
sites (84). Estimates of the fraction of cancer deaths in the United
States that can be attributed to occupational exposure have been made
by several investigators. These estimates have been as low as 1 to 5
percent (45, 73, ?`4, 153, 241). Cole (37) has placed these estimates as
high as 10 to 15 percent.
There are difficulties in estimating the proportion of cancers
attributable to certain occupational exposures, tobacco, alcohol, or diet.
Most of these estimates are based on the assumption that specific
cancers are caused by specific agents. It is more likely that cancer is a
disease of interactions. The precipitating cause and subsequent
development of cancer is likely to be a process with multiple phases
and multiple agents. Both internal and external factors interact at
each of several stages before cancer becomes clinically apparent. The
development of cancer, then, is influenced by two or more different
external factors acting simultaneously or sequentially. This principle is
illustrated by the synergistic effects of tobacco and alcohol. Cigarette
smoking by itself is an important cause of oral cancer, whereas alcohol
by itself is a relatively minor cause of oral cancer. Combined exposure
to cigarette smoking and alcohol results in an increased risk of
developing cancer of the oral cavity which is considerably higher than
the risk experienced by cigarette smokers alone or drinkers alone.
The synergistic relationship between cigarette smoking and occupa-
tional exposure as it relates to the development of cancer is
complicated. Most hazardous occupational exposures are to single
agents or to a few at most. Cigarette smoking results in exposure to
more than 2,000 chemical compounds, among which are carcinogens,
tumor initiators, and promoters (see Chapter 14). It might be expected
that cigarette smoking would have an adverse interaction with several
Wupational exposures, which it is important to try to identify. Insofar
as possible, workers should be provided with a safe working environ-
ment, free from potentially harmful agents. It is equally true that
workers can substantially reduce or eliminate the potential for

5-27


harmful occupational interactions by eliminating cigarette smoking
from their lifestyle. This would probably eliminate the vast majority of
the lung cancers which are occupationally related.

Short of giving up smoking entirely, it might be impossible for the
worker to avoid many of the risks of developing cancer which may be
related to his employment. Smoking at home but not on the job will not
avoid this interaction, because the tars which are trapped in the
airways will still be there when. the individual goes to work.

Asbestos

In 1935, Lynch and Smith (127) in the United States and Gloyne (61) in
in the United Kingdom reported an association between asbestos and
lung cancer. In 1968, Selikoff, et al. (188, 189) first took into account
the interaction between cigarette smoking and asbestos exposure in
the development of lung cancer. They estimated that asbestos workers
who smoked cigarettes had eight times the lung cancer risk of smokers
without this occupational exposure. This was estimated to be 92 times
the risk of nonsmokers who did not work with asbestos. This study has
been continued and is supported by other investigations which
consistently show a potent synergism between the carcinogens of
tobacco smoke and asbestos (29,69). There is evidence that exposure to
asbestos carries some real risk to nonsmokers; however, this is of a low
order of magnitude compared to the risks experienced by cigarette
smokers (135,15?`).

Uranium Mining

Lung cancer is an occupational risk associated with uranium mining.
The causative agents in the atmosphere of mines are alpha particles
resulting from the decay of short-lived radon daughters (12, 48).
Several investigators (7, 126, 173, 224, 225, 226) have extensively
studied underground uranium miners in the United States. The
combined effect of tobacco smoke and radon daughter exposure results
in high death rates from lung cancer among uranium miners. The risk
for cigarette-smoking uranium miners is at least four times greater
than for cigarette smokers who do not work in the mines.

Nickel

Epidemiological studies by Morgan (146) and Doll (44) and experimen-
tal studies by Hueper (89) and Sunderman, et al. (200,201,202) suggest
that exposure to nickel or nickel carbonyl is a potent carcinogen for the
respiratory tract in humans and animals. The interaction of cigarette
smoking on the risk of respiratory cancer in nickel workers will
probably never be adequately studied, since the Mond process for
refining nickel is rarely used and conditions in nickel refining factories
have improved.

5-28


Chkwome th yl Ethers
Epidemiological and experimental studies (59, 114) have identified
chloromethyl ethers as potent carcinogens for the human and animal
respiratory tract. Investigations are in progress to more fully
characterize these relationships, but the closing of the plants producing
these substances makes it unlikely that the relative contribution of
cigarette smoking to this type of occupational lung cancer will ever be
known.

Animal Studies

Experimental animal models have been developed in which to study
tobacco-induced carcinogenesis. Over the past 30 years, this field has
acquired considerable sophistication and has enhanced our understand-
ing of carcinogenesis in humans.
Experimental carcinogenesis has advanced to the point where it is
now possible to reproduce in animals the major categories of
respiratory tumors observed in humans and to link the induction of
certain types of respiratory tumors to definite categories of exposure
(176). By intratracheal administration of polynuclear hydrocarbons in
rats and hamsters, bronchogenic squamous cell carcinoma is induced.
Certain systemic carcinogens, particularly diethylnitrosamine in
hamsters, give rise to adenomatous tumors of bronchial and bronchiolar-
alveolar origin, as well as to papillary tumors in the trachea. Of the
main types of respiratory tumors seen in human pathology, only one,
the oat cell carcinoma, has not yet been found to be reproducible in
experimental animals (176).

Skin Painting and Subcutumous Injections

The earliest animal models for studying tobacco carcinogenesis
involved the single or repeated painting of shaved or unshaved animal
skin with solutions containing whole tobacco tar, various tobacco
condensate subfractions, or single chemical compounds known to be
Present in tobacco smoke (161). Subcutaneous injections of various
substances or fractions found in tobacco were also used as experimen-
tal models. Considerable criticism was directed towards these early
studies, but they effectively demonstrated that a variety of carcino-
genic compounds were found in tobacco smoke and that tobacco tar
wag a potent carcinogenic substance. Early experiments of these types
have been reviewed by Wynder and Hoffmann (245).

T?ddronchiul Implantation and Instillation
More complex experiments have been performed using direct implan-
tation, instillation, or fixation of suspected materials in the tracheo-
bronchial tree of animals. Several authors have reviewed these studies
(115,143, 175, 176, 245).

5-B


Lung tumors which closely resemble lesions found in human
cigarette smokers can be induced in hamsters by intratracheal
instillation of benzo(a)pyrene (BaP). BaP induces a low incidence of
bronchogenic tumors in hamsters when administered in saline; but
when it is adsorbed into <l ~1 ferric oxide carrier particles, its
carcinogenicity is increased. When administered in the absence of BuP,
ferric oxide particles alone do not induce tumors (176). The rate of
elimination of BaP from the lung influences its tumorigenicity (71, 72).
When BaP is administered alone or in simple mixtures with particles,
95 percent is eliminated within 24 hours. However, BaP adsorbed to
particles is retained within the lung for several days (71, 72). Thus, the
duration of the exposure to the carcinogen may be important to tumor
induction by polycyclic aromatic hydrocarbons (PAH). These studies
suggest that the particulate carrier increases the retention of PAH in
the lung with a consequent increase in the exposure of respiratory
tissue to the carcinogen.

In the hamster system, intratracheally-instilled BaP ferric oxide
particles and subcutaneously-administered diethylnitrosamine (1.&Z,
1&`) were synergistic. Inhaled ferric oxide particles have also been
found to enhance carcinogenicity of subcutaneously administered
diethylnitrosamine (158) in the peripheral lung.

Inhalation Carcinogenesis

Various species, including mice, rats, hamsters, and dogs, have been
exposed to cigarette smoke or to aerosols of its chemical constituents.
Most of these substances have been administered to the experimental
animal in a passive fashion. Active inhalation experiments more closely
simulating human smoking behavior have been conducted by Hockey
and Speer (169) and Auerbach, et al. (11, 66). In these experiments,
animals were trained to inhale voluntarily through openings in the
trachea.

Nitrosa mines

A number of nitrosamines present in tobacco products or smoke have
been found to produce respiratory tract tumors in animals. Various N-
nitroso compounds of a nicotine metabolite, which are present in cured
tobacco and chewing tobacco, can induce respiratory tract tumors in
mice and hamsters (70, 77). Diethylnitrosamine, a volatile component
of cigarette smoke, is a potent inducer of lung tumors in hamsters
(141). Other nitrosamines present in tobacco products or smoke which
have been shown to produce lung or tracheal tumors in animals include
nitrosopiperidine (99) and N-nitrosodiethanolamine (81). This last
compound is thought to be derived during curing from the maleic
hydrazide triethanolamine salt which is sprayed on growing tobacco
plants to reduce sucker formation.

5-30


Phgocytosis

Another factor which may be important is phagocytosis by macro-
phages. Some macrophages with engulfed particles remain in the lung
for an extended period of time. A recent study by Palmer, et al. (162)
showed that macrophages metabolized the potent carcinogen 7.12
dimethylbenz(a)anthracene (DMBA) and released the majority of the
resultant derivatives into the surrounding medium. Unlike macro-
phages, cells from lung and tracheal tissues tended to retain the
DMBA metabolites that they produced. This and related work by
Harris, et al. (69a) showed that the human pulmonary macrophages
under some conditions in ~+t,-o may permit the accumulation of
metabolic products of carcinogens.

Conclusions

1. Cigarette smoking is the major cause of lung cancer in both men
and women. This fact has been supported by prospective and
retrospective epidemiological studies, clinical studies, autopsy studies,
and experimental studies in animals. This conclusion is based on a
weight of evidence which exceeds by several times the evidence
available when this same conclusion was first reached in 1964.

2. The past 15 years have brought little significant progress in the
earlier diagnosis or treatment of lung cancer. Taken as a whole, 30
percent of lung cancer patients live 1 year, and only 10 percent live 5
years after diagnosis. Fortunately, lung cancer is largely a preventable
disease. Significant reductions in the number of deaths from lung
cancer can be achieved if a significant portion of the smoking
population can be persuaded to stop smoking and if a reduction can be
brought about in the number of young people who take up smoking.
3. Lung cancer mortality is increasing in women and is increasing
more rapidly than any other cause of death. If present trends continue,
lung cancer will be the leading cause of cancer death among women in
the next decade.

4. There are dose-response relationships for developing lung cancer
with the number of cigarettes smoked per day, the duration of
smoking, the age of starting to smoke, degree of inhalation, tar and
nicotine content of cigarettes, and several other measures of dosage.
5. The long-term use (10 years or more) of filter cigarettes is
associated with lower death rates from lung cancer than those
experienced by persons who smoke an equal number of nonfilter
cigarettes.
6. Ex-cigarette smokers experience decreasing lung cancer mortality
mtes, relative to continuing cigarette smokers. The risk of developing
lung cancer for ex-smokers depends on the type of smoker he or she
used to be. The risk is proportional to the number of cigarettes
previously smoked per day, degree of inhalation. the age when smoking

5-31


was started, and duration of smoking. Whether the risk based on the
previous smoking profile is high or low, there is a fairly rapid initial
decline in risk following cessation of smoking which occurs over a 2- to
3- year period. It takes from 10 to 15 years, however, until the risk of
developing lung cancer approaches the risk of nonsmokers.

7. Pipe and cigar smokers have lung cancer mortality rates which are
higher than those of nonsmokers but which are considerable lower
than those of cigarette smokers (see conclusions in the Chapter on
Other Forms of Tobacco Use for further refinements and qualifica-
tions concerning pipe and cigar smoking).

8. Air pollution may be associated with the development of lung
cancer; however, detailed epidemiological surveys indicate that the
influence of air pollution on the development of lung cancer is small
compared to the overriding effect of cigarette smoking. It is probable
that there is a synergistic effect between cigarette smoking and air
pollution in causing lung cancer. Air pollution does not appreciably
influence lung cancer mortality rates in nonsmokers.

9. Certain occupational exposures, particularly uranium mining and
working with asbestos, act synergistically with cigarette smoking,
resulting in lung cancer mortality rates which exceed by several times
the lung cancer mortality rates of unexposed cigarette smokers. Lung
cancer mortality in these situations can be attributed to both cigarette
smoking and the occupational exposure.
10. In the past few years, progress has been made in the
development of animal models in which to study lung cancer. At the
present time it is possible to reproduce in animals the major categories
of respiratory tumors observed in man, using tobacco smoke, subfrac-
tions of tobacco tar, or specific compounds found in c'garette smoke.

Cancer of the Larynx

Approximately 1 percent of all deaths from cancer are from cancer of
the larynx. It is estimated that in 197'8 there were 3,350 deaths from
cancer of the larynx, with 2,900 occurring in males and 450 occurring in
females. The National Center for Health Statistics reported 3,351
deaths from cancer of the larynx in 1976. There were 2,308 deaths in
males and 543 deaths in females (150). The most common histological
lesion is squamous cell carcinoma. Approximately 70 percent are
located in the glottis and 25 percent in the supraglottic region (132).
Laryngeal cancer is predominantly a disease of males, although the
incidence for females has increased somewhat over the past 20 years
(181, 238). A typical patient with cancer of the larynx would be a 6@
year-old male who was a heavy cigarette smoker and also a moderate-
to-heavy alcohol drinker (132). The 5-year survival rate is improving
and is presently at approximately 60 percent for all stages in both
males and females.

5-32


TABLE 16.--Mortality ratios for cancer of the larynx-
      prospectiive studies

NUmtR?r    Mortalitv ratio

Study           Population size dezths Nonsmokers Smokers      Comments

A.C.S. B-
State Study(68)

                          All larynx
lE!a,ooo males     24     -                cancer deaths
                                orrurmd in
                                smokers

British
doctors(47a)

U.S. vetemns(so)

A.C.S. 25-
State Study@%)

Miiomia males
in 9 occupations
hw

Moo0 males     38     1.00

13.00

cancer of
larynx and
other upper
respiratory
sites.

239,cca males

440,ooO males

54     1.00       9.95


57     1.00     6.09-males,   W-45-61
          &BB-males,   ages=-

68,000 males

11

-.

-     All larynx
       cancer deaths
       occurred in
       smokem

Jrqmese
atudy(77q80)        l2z,mmalea     38     1.00       11.83
              142800 females    6     1.00       9.00

Epidemiological Studies

Many epidemiological studies have investigated the relationship
between smoking habits and cancer of the larynx. The major
Prospective studies are outlined in Table 16. In these studies, cigarette
smokers had a mortality ratio which was 6 to 13 times greater than
that of nonsmokers. In three of the prospective studies, mortality
ratios could not be calculated because all of the deaths from cancer of
the larynx occurred in cigarette smokers.

Recent retrospective studies confirm prior evidence of a strong
Positive association between cancer of the larynx and cigarette
smoking (56, 238, 252, 253). Wynder, et al. (238) found that the large
sex difference has diminished somewhat over the past 20 years. This is
most likely due to the increase in female cigarette smokers in age
groups for which laryngeal cancer rates are high. The relative risk for
developing laryngeal cancer for male cigarette smokers was 15.8; for
female cigarette smokers it was 9.0. There was also a strong dose-
response relationship in the relative risk of laryngeal cancer with both

5-33


the number of cigarettes smoked per day and the duration of smoking.
A distinct synergism with combined alcohol and tobacco use was also
described, with a relative risk of 22.1 for the smoker of more than 35
cigarettes a day who was also a heavy drinker. This study also
examined the relative risks experienced by long-term filter cigarette
smokers. At every level of consumption, both males and females who
smoked filter cigarettes had a lower risk than did nonfilter smokers.
Among men, the reduction in risk ranged from 25 to 49 percent for
cancer of the larynx, and a substantial lowering of risk was also found
for women. For ex-smokers, the risk of developing laryngeal cancer
diminished gradually with time in a curve that paralleled that for
cancer of the lung. The most rapid reduction in risk occurred during
the first 5 years after cessation of smoking. After approximately 10
years, the risk approached that of nonsmokers. Several of these
relationships are demonstrated in Figures 4 through `7.

Williams and Horm (2.33), using data from the Third National
Cancer Survey, reported a strong dose-response relationship for the
number of cigarettes smoked per day and the risk of developing cancer
of the larynx. The relative risks for males, controlling for age and race,
were 2.9 for level-one smokers, 3.3 for level-two smokers, and 17.7 for
level-three smokers (the levels for cigarette-smoke exposure were
established by using both the amount and the duration of cigarette
use). Considering tobacco use at each level of alcohol consumption, the
risk of developing cancer of the larynx increased as tobacco exposure
increased. There was a positive association for the intake of alcoholic
beverages and the development of cancer of the larynx. In previous
reports of the U.S. Public Health Service (212, 217), most of the older
retrospective epidemiological studies have been reviewed (22, 56, 172,
174, 184, 185, 193, 196,iW,205,218,2.?7,246, 250).

Asbestos

Several authors have found an association between asbestos exposure
and cigarette smoking with development of laryngeal carcinoma (28,
121,148,190,197).

Animal Studies

The Syrian golden hamster has been found to be a suitable species for
the investigation of cancer of the larynx. The distribution of malignant
lesions in the upper airway of the hamster is not due to an unusual
susceptibility of the larynx for tumor induction but rather reflects the
distribution of smoke aerosol precipitation within the upper respira-
tory tract. The most recent experimental studies are those of Bernfeld,
et al. (18), Dontenwill, et al. (43, 50), Hornburger (86), and Karbe and
Koster (93). Cigarette smoke inhalation has not been found to induce
laryngeal tumors in other rodents. Such tumors have been induced,

5-34


40




35




30




25




20




15




10




5

   CASES
N:
CONTROLS

      1
-115
9

NON F        F NF     F NF     F NF    F NF
SMOKER l-10      11-20     21-30    31-40     41+

25       16
126    z

32       13

FIGURE &-Relative risk of developing larynx cancer for males, by
number of &arettes smoked per day -and use of filter (F) and
nonfilter (NF) cigarettes
SOURCE: Wynder, EL (253)

however, by the direct application of carcinogens known to be present
in cigarette smoke. This is accomplished by the intratracheal instilla-
tion of benzo(a)pyrene in combination with particulate dusts into
hamster lungs. In this animal model, laryngeal tumors, as well as
tumors in other parts of the respiratory tract, are induced (I.44 176,
173.

5-35


25

15








5

   CASES
N:
CONTROLS

8
ii?

9
54
7   61

NON         F NF         F NF
SMOKER        l-20          21+

FIGURE L-Relative risk of developing larynx cancer for females,
by number of cigarettes smoked per day and use of filter (F) and
nonfilter (NF) cigarettes
SOURCE: Wynder. E.L. (453)

Conclusions

1. Epidemiological, experimental, and autopsy studies indicate that
cigarette smoking is a significant causative factor in the development
of cancer of the larynx.

2. The risk of developing cancer of the larynx in pipe and cigar
smokers is similar to that for cigarette smokers.

5-36


16

8

4

PRESENT    l-3
SMOKER

   CASES
N: CONTROLS

5
443
Ifi

11+

  -
i
NON
SMOKER

FIGURE 6.-Relative risk of developing larynx cancer for male ex-
smokers, by years of smoking cessation
fjcmtcE: wynder, EL. ($53)

3. There are positive dose-response relationships for the development
of laryngeal cancer with the number of cigarettes smoked per day and
the duration of cigarette smoking.
4. There is a synergistic effect with the use of cigarettes and alcohol.
The risk of developing cancer of the larynx is much greater for heavy
smokers who also drink heavily, compared with individuals who only
have exposure to either substance.

5-37


18

6

2

           CASES
       N:
3        CONTROLS
72

1-3      4-6      7+      NON

SMOKER                           SMOKER

FIGURE 7.-Relative risk of developing larynx cancer for female
ex-smokers, by years of smoking cessation
SOURCE: Wynder. EL. (258)
5. There is a substantial decrease in the risk of developing cancer of
the larynx with the long-term use of filter cigarettes (10 years or
more), compared to the use of nonfilter cigarettes.
6. There is a gradual reduction in the risk of developing laryngeal
cancer after cessation of smoking. After approximately 10 years, the
risk of developing cancer of the larynx is similar to that of nonsmoken.
7. It has been reported that exposure to both asbestos and cigarette
smoking synergistically increases the likelihood of an individual
developing cancer of the larnyx.

5-38


8. Animal models have been found in which inhalation of cigarette
smoke induces cancer of the larynx.

Oral Cancer

Cancers included in the oral cancer category are those malignant
tumors of the lip, tongue, floor of the mouth, hard and soft palate, the
gums, buccal mucosa, and oropharynx. The National Center for Health
Statistics reported that in 1976 there were 8,114 deaths from cancer of
the oral cavity, buccal surfaces, and pharynx. There were 5,731 deaths
among males and 2,383 deaths among females (150). It is estimated
that, in 1978, 24,400 new cases were diagnosed with a total of 8,409
deaths (4). The incidence in males is three times that in females. For
the floor of the mouth, tongue, and pharynx, 5year survival rates vary
from 25 to 45 percent. A variety of histological types of malignant
neoplasms can affect these tissues, but squamous cell carcinoma is the
most common type, accounting for 90 percent of cancer of the oral
cavity.

Epidemiological Studies

The use of tobacco in various forms has been associated with the
development of cancer of the oral cavity and pharynx. Studies of
cancer of the oral cavity are international. Many investigations have
heen carried out in Asian nations, as well as in the West. Data from the
major prospective epidemiological studies show increased mortality
ratios for these cancers among cigarette smokers, as well as among
pipe and cigar smokers, compared to nonsmokers. There is some
variation in mortality ratios, ranging from about 3.0 to 10.0. The
results of these investigations are presented in Table 17.
There are a large number of retrospective studies which have
examined the relationship of cigarette smoking to the development of
cancer of the oral cavity (26, 57,94,95,116,117,119,13'3,134,138,139,
Q4,1.&163,170,174,178,193,220,223,239,246). These studies almost
uniformly show a significant relationship between the various forms of
tobacco use and cancer of the oral cavity and pharynx. One large
survey recently conducted in India was reported by Bhargava, Smith,
Malaowalla, and associates (21,130,192). The prevalence of oral cancer
WaS determined in 57,518 industrial workers in Gujarat, India. A Z-year
follow-up survey was conducted, and the incidence of oral cancer was
determined. There was a strong association with tobacco use in various
forms. In the Third National Cancer Survey (233), Williams and Horm
reported a significant correlation between cancer of the gum and
mouth and the use of pipes, cigars, cigarettes, and unsmoked tobacco.
In many of the studies dose-reponse relationships were examined.
Increasing relative risks with increasing tobacco use were noted.

5-39


TABLE 17.-Mortality ratios for cancer of the oral cavity-
      prospective studies

Study

         Number      cigerette
Population size of  Nonsmokers          Comments
          deaths         smokers

A.C.S. 3-
State Studyl68)

British
doetors(47u)

U.S. veterans@O)

A.C.S. 25-
State Study(65)

239,006 nudes

440,600 males

California malea
in 9 occup&iona
@w

68,006 make

Japan&%$         1222ood=
study(Y7a,80)       142@0 fen&a

Swedish
study@?)

55,000 Swedish
males and females

13E,ooO males

34,cm males

55

38

61

1.00






1.60

1.00


1.00





1.06



1.00
1.00

      only 3
18.00       deaths
         8mon8
        nousmokem

13.00

4.03


9.90     ABes
        4M4

276

288 males
1.22 females


       5 deaths

Mortality ratios not    in non-
  published     smoking
             males.
           10 deaths in
          smoking
              m&?LL

Other Forms of Tobacco

All forms of tobacco use expose the oral cavity to compounds found in
raw tobacco or tobacco smoke. In most of the prospective and
retrospective studies where other forms of tobacco use were accounted
for, significant correlations were found between the use of tobacco and
the development of oral cancer. These relationships are of the same
general magnitude or slightly greater than those found with cigarette
smoking. These relationships are examined in detail in the Chapter on
Other Forms of Tobacco Use.

Other Risk Factors

Other than tobacco use, alcohol consumption and possibly poor
dentition appear to be risk factors for the development of oral and
pharyngeal cancers. The most recent investigations of the interaction

5-40


between alcohol and tobacco in the development of oral cancer are the
studies of Rothman and Keller (ITO), Feldman, et al. (58), Graham, et
al. (62), Browne, et al. (28), and the Third National Cancer Survey
(233). In the latter survey, cancer of the oral cavity was associated
significantly with both cigarettes and alcohol. The relative strength of
each exposure after controlling for the other was evaluated by
multiple regression analysis. For cancer of the pharynx, the standard-
ized regression slope (based on standard deviation units) in males, after
controlling for age, race, education, and cigarettes or alcohol, was 0.104
for alcohol and 0.084 for cigarettes. For cancer of the oral cavity and
gums, the values were: alcohol 0.081 and cigarettes 0.018. For cancer of
the lip and tongue, the values were: alcohol 0.057 and cigarettes 0.043.
Hence, in this survey, oral cancer in males was somewhat more related
to drinking than to smoking.
Rothman and Keller (17'0) also reported a strong synergy between
the two exposures. They attributed 76 percent of oral cancer in males
to the interaction of tobacco and alcohol. Feldman, et al. (58) found
that nonsmoking alcohol users had only a slightly increased risk for
head and neck cancer, whereas smokers who did not use alcohol still
had two to four times the risk of abstainers from alcohol and tobacco.
The risk for the heavy drinker who smokes, however, was from 6 to 15
times greater than for the individual who did not use tobacco or
alcohol. In the study of Graham, et al. (62), the relative risk for heavy
smoking alone was only 1.54; for heavy drinking alone it was 1.70.
Heavy smoking and heavy drinking resulted in a relative risk of 2.49.
When this was combined with inadequate dentition, the risk rose to
7.66. Browne, et al. (28) reported that alcohol and tobacco use was
particularly prevalent among patients with oral squamous cell
carcinoma.

LRukoplakia

hukoplakia of the oral mucosa represents an abnormal thickening and
keratinization of the oral mucosa. Leukoplakia is generally recognized
as a precursor of malignancy in the oral cavity and is associated with
tobacco use in various forms. The largest survey of leukoplakia in a
Western population has been conducted by Banoczy and associates (13,
168, 199). Leukoplakia is quite common in India where tobacco and
betel-nut chewing occurs and where bidis are smoked. The prevalence
and incidence of leukoplakia has been reviewed in several large studies
(24 130, 137,192).

himal Studies

An ideal animal model in which to study oral carcinogenesis has not
been found. Cigarette smoke and cigarette-smoke condensates general-
1~ fail to produce malignancies when applied to the oral cavity of mice,
rabbits, or hamsters. Mechanical factors, such as secretion of saliva,

541


interfere with the retention of carcinogenic agents. The only positive
results with carcinogens have been obtained with benza(u)pyrene, Xl-
methyl-cholanthrene, and 9,1Odimetyl-12 benzanthracene applied to
the cheek pouch of hamsters. The cheek pouch, however, lacks the
salivary gland, and its structure and function differ from those of the
oral mucosa. These studies have been reviewed in previous reports of
the U.S. Public Health Service (212,217).

Conclusions

1. Epidemiological studies indicate that smoking is a significant
causal factor in the development of cancer of the oral cavity. Dose-
response relationships with the number of cigarettes smoked per day
have been described.

2. The use of pipes, cigars, and chewing tobacco is associated with
the development of cancer of the oral cavity. The risk of using these
forms is of the same general magnitude as that of using cigarettes.
3. There is a synergism between cigarette smoking and alcohol use
and the development of cancer of the oral cavity. The use of alcohol
and tobacco results in a higher risk of developing cancer than that
resulting from the use of either substance alone.

Cancer ot the Esophagus

The National Center for Health Statistics reported that there were
7,224 deaths from cancer of the esophagus in 1976. There were 5,343
deaths in males and 1,881 deaths in females (150). It has been
estimated that these figures rose to 7,100 deaths from cancer of the
esophagus in 1978 (4). In addition, esophageal cancer incidence and
mortality in the United States are substantially higher for blacks than
for whites (39). Epidermoid carcinoma is the most common cancer of
the esophagus (3). The prognosis is extremely poor with a 5year
survival rate of only 3 percent; the median survival time is less than 6
months after diagnosis (152).

Epidemiological Studies

Data from the major prospective epidemiological studies demonstrate
a significant relationship between smoking and esophageal cancer. The
mortality ratios for male cigarette smokers range from 1.82 to 8.75.
These relationships are shown in Table 18. In several of these studies a
positive dose-response relationship for the number of cigarettes
smoked per day is shown. Available evidence indicates a similar
relationship for men and women.

A number of retrospective studies have been published concerning
smoking and esophageal cancer. Risk ratios for smokers in these
studies range from 1.3 to 11.1, compared to nonsmoking controls (24,
105, 133, 174, 178, 186, 194, 204, 235, 246).

5-42


TABLE IS.-Mortality ratios for cancer of the esophagus-
      prospective studies

Study         Population size

Number of         Cigarette
deaths    Nonsmokers   smokers    Comments

A.C.S. S-
State Study(68)

British
dabs (471~)

U.S. veterans(s0)

A.C.S. 2s
State Study(65)

California males
in 9 occupations
@a)

Japr=
Studyf 77a)

Swedish

lES.Mx) males

34,OMl males

293,ocO



440,ooo males



@wo

lZ.200 males    215         1.00       2.35

  55,090 Swedish    1 nonsmoker
males and females    12 smokers

1.00       -

1 nonsmoker
33 smokers

6.5

111

46

32

1.00

   Qvhagus
     and other
-    respiratory
     sites

l.l-Hl

8.75

EWwm
and other
respiratory
sites

1.00       6.17

1.00       4.17

1.00       1.32

Other Forms of Tobacco Use

In most of the prospective and retrospective epidemiological investiga-
tions, the association of esophagus cancer with the use of tobacco in
other forms was examined. These relationships are discussed in some
detail in the Chapter on Other Forms of Tobacco Use. The mortality
ratios for cancer of the esophagus are approximately equal in users of
cigars, pipes, and cigarettes.

Other Risk Factors

Numerous investigations have been made into the synergistic relation-
ships between the use of tobacco in various forms, alcohol consumption,
and the development of cancer of the esophagus (78, 92, 105,182, 183,
204, 208, 233, 235, 249). Some investigators report that tobacco is a
more important carcinogen than alcohol in the development of cancer
of the esophagus, but others report that the reverse is true. Most of
these studies support a synergism with the combined use of tobacco
and alcohol, resulting in higher rates of cancer of the esophagus
compared to those resulting from the use of either substance alone.
The mechanism of the association is not known. Alcohol may act as a

5-43


solvent for carcinogenic hydrocarbons in tobacco smoke or alter
microsomal enzymes in the mucosal cells of the esophagus (23.4). This
hypothesis has received support from experimental observations by
Kuratsune, et al. (113). The picture is complicated by the fact that
alcoholism may be accompanied by severe nutritional deficiencies
which may also predispose an individual to certain diseases.

Autopsy Studies

Histologic changes in the esophagus in relationship to smoking of
tobacco in various forms were investigated by Auerbach, et al. (11). A
total of 12,598 sections were made from esophageal tissue obtained
from 1,269 subjects. It was found that tobacco smoking in any form
resulted in the formation of atypical nuclei, disintegrating nuclei,
hyperplasia, and hyperactive esophageal glands. Each of these
parameters was significantly more abnormal in smokers than in
nonsmokers; however, these changes were more frequently seen and
more severe in cigarette smokers (11).

Animal Studies

There is experimental evidence that benzo(a)pyrene is able to
penetrate the cell membranes of the esophageal epithelium, producing
papillomas and squamous cell carcinomas. This process can be
accelerated and better penetration achieved if the carcinogen is
dissolved in an aqueous ethanol solution. This effect was reported by
Kuratsune, Horie, and Kohchi (88, 113). Nitrosamine-induced esopha-
geal cancer in experimental animals has also been reported by a
number of investigators (34, 52, 53, 54, 179). These observations are
significant because a variety of nitrosamine compounds have been
identified in cigarette smoke.

Schmaehl (179) administered methyl-phenyl-nitrosamine orally or
subcutaneously to Sprague-Dawley rata. Carcinomas of the esophagus
were found in 46 to 87 percent of the animals. Simultaneous
application of 25 percent ethyl alcohol did not affect the tumor
incidence.

Mirvish (140) has reported that 3H-thymidine incorporation in rat
esophageal epithelium can be inhibited in the presence of nitrosamine
in wivo and in vitro, lending further support to the role of these
compounds in esophageal carcinogenic mechanisms.

Conclusions

1. Epidemiological studies demonstrate that cigarette smoking is a
significant causal factor in the development of cancer of the
esophagus. The risk of developing esophageal cancer increases with the
amount smoked.

5-44