[CANCER
RESEARCH
35, 3388-3394, November 1975]
The Epidemiology of Large Bowel Cancer'
Ernst L. Wynder
Division of Epidemiology,
Naylor Dana Institute for Disease Prevention, American Health Foundation, New York, New York /00/9
Summary
to the rectum, it should be stressed that many vital statistics
are misleading in that they have included rectal sigmoid
Results from epidemiological
studies have provided clues
lesions with rectal cancer.
The etiology
of rectal cancer,
as to etiological factors involved in the developmentof large when defined in terms of the last 8 cm of the large bowel,
bowel cancer. Overnutrition, especially in terms of dietary appears to be different from cancer of the colon and cancer
fat consumed, appears to be a key etiological variable
affecting the rate of colon cancer. Epidemiologists
can
provide the leads for chemists and bacteriologists to pursue
in population groups and for experimentalists to test in
laboratory animals. Coordination ,of and cooperation be
tween many disciplines is necessary in order to contribute to
of the rectal sigmoid and, therefore, should be evaluated
separately.
2. A world-wide correlation exists between colon cancer
and fat consumption in various countries (Chart 2). Obvi
ously, correlation does not necessarily mean causation,
although it is apparent that, in the absence of correlation,
the prevention of this man-madedisease.
causation is unlikely. One can establish a similar correlation
with many other variables connected with increasing eco
nomic development,
although, to be taken seriously, a
Introduction
correlation needs to have some logical basis.
3. There exists a negative correlation between gastric
A primary thinking process of an epidemiologist when
cancer
and colon cancer (Chart 3). This finding suggests
trying to elucidate the etiology of a given disease involves
that
factors
enhancing the risk of colon cancer may at the
simple logic. When it comes to the large bowel, one would
obviously consider factors present in the lumen of the large same time protect against cancer of the stomach and vice
bowel, although one cannot exclude components that are versa.
carried to the wall ofthe large intestine by the blood. On the
basis of such reasoning, we began to suspect, in our
epidemiological
study reported in 1967 (20), that colonic
contents contained tumorigenic components. Additionally,
because of the wide variation in the incidence of large bowel
cancer throughout the world, we suspected that there had to
be factors, probably dietary in nature and specifically
related to dietary fat, that directly and indirectly affected
the make-up of the feces and consequently tumorigenic
components.
Demographic
Leads
Significant leads may be derived from the markedly
diverse distribution of large bowel cancer in various parts of
4. A positive correlation exists between myocardial in
farction and colon cancer (Chart 4). The main risk factors
for myocardial infarction (hypercholesteremia, hyperten
sion, and cigarette smoking) have been well established (7).
The question that needs to be answered is whether any of
these factors is related to colon cancer.
5. Incidence data from the United States, comparing
data for 1947 and 1969, suggest a gradual increase in the
incidence of cancer of the colon while showing a decrease in
cancer of the rectum (Chart 5). Could this increased
incidence in colon cancer, if real and not due to an increase
in the discovery and reporting of Grade I cancers, be
explained on the basis of environmental factors that have
also increased?
is near unity,
in
constast to cancer of the rectum, which is predominantly
6. The sex ratio
of cancer
of the colon
a
the world and population groups within a given country. male disease especially in later life (Chart 6). If one
The major clues that, when viewed together, should be able examines the sex ratio of colon cancer by age, one finds that
to provide us with an idea of the etiological factors involved before age 60 colon cancer tends to be slightly more
in the development of large bowel cancer, may be summa
rized as follows.
1. In general, the more economically developed a society
is, the greater is its incidence of cancer of the colon, al
though not necessarily of the rectum (Chart 1). In reference
I Presented
at the Conference
on Nutrition
in the Causation
of Cancer,
May 19 to 22, 1975, Key Biscayne, Fla. This study is in part supported by
National Cancer Institute Grant CA-I2376, American Cancer Society
Grant Ct-I ISA, and National Cancer Institute Grant CA-l6382 from the
National Large Bowel Cancer Project.
3388
common among females, while later in life it becomes more
predominant among males. This finding is dissimilar to
what is known for myocardial infarction, causing one to
speculate whether female steroids that tend to lower serum
cholesterol levels might in turn induce a greater excretion of
cholesterol into the feces.
7. Among American blacks, there are different rates in
certain rural parts of the South compared to blacks living in
industrialized Northern cities (9). Could these differences be
due to variation in diet?
8. Significant differences in colon cancer have been
CANCER RESEARCH VOL. 35
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1975 American Association for Cancer Research.
@
@1@‘—@-@
Large Bowel Cancer
12
70
•@
@MALE: r=+@@@
[@@EMALE:r:+.1332J
@
10
60
-
,
MICIUM £P@G&AP4O
&WAILS
,/
8
@
6
/
/I
@
SWITZERLAND
/
—
—@------.@
a
.9
•sMoN•i@ti*,o
/
@40
•12
@8
AUSTRALIA ,,/
BELGIUM
0
O(NCIAND
&WALlS
. .15
AUSTRIA
[email protected]@-wii.
I@IAND
TAL@
• P(*ZtALANDØ
I@Lmo
0
•
18
a@@w@v•1,
@•A@@RIçA
e@&MRt.AND @LA$D
Fl@LA@@ swuTzI@A@I@
0 JAP@_@Q@@-
@
•@.IRELA@D
• .:RLA@D
/
/
(\ •
4
r@—0.763
•AUSTRIA
;/
—
@
@
1
50
-
//
@
.3
SCOTLAND•
•17
•@
823 .10
c)
•16
U
•19
S@t*@
OISRAf@L@Ø@AY
.22
_
@13
@@20
OS.AIRICA
•14
8
21
20
.6
2
•24 ¶5
10
@
5
10
15
AGE-ADJUSTED DEATH RATE OF COLON CANCER
PER 100000
Chart I . Correlation
n
0
between colon and rectum cancer.
I
2
4
I
6
8
10
Colonic Cancer Rate per 100.000
12
14
16
Chart 3. Correlation between gastric cancer and colonic cancer.
Age-adjusted mortality rates for men, 1964 to 1965. /, South Africa; 2,
Canada; 3, Chile; 4, United States white; 5, United States nonwhite; 6,
Israel; 7, Japan; 8, Germany; 9, Austria; 10, Belgium; 11, Denmark; /2,
Finland; /3, France; 14, Ireland; 15, Italy; 16, Norway; 17, Netherlands;
/8, Portugal; 19, England and Wales; 20, Scotland; 21, N. Ireland; 22,
Sweden; 23, Switzerland; 24, Australia; 25, New Zealand.
100
@
F
@•:‘@“@
I,r=+O.6@J
..@,m.*
.@,
@
5w@
FM@'
‘I—
Chart 2. Large bowel cancer mortality
(1966 to 1967) and dietary fat and oil
consumption (1964 to 1966).
50
.
G._.p,
“@@‘:
a s@
. Sw,1i*i@@ •4/S
**s'.
•
.c.,@
4s#@.
S
‘9,..,
. .@,—
.
. C.,,,.
•Jopos
a.
@
‘0
-
I
.
&
.
.
10
AGE-ADJUSTED
I
$5
MORTAIJTY RATE FROM COLONIC CANCER/IOO,000
shown between Puerto Rico and the United States (13)
rural areas, a finding consistent with differences in dietary
(Chart 7). The intake of fat and cholesterol is significantly
lower on the island than on the United States mainland, but
habits (5).
10. Among economically
what additional environmental differences, especially in
relation to diet between these 2 population groups, might
account for the differences in the cancer rates?
Denmark and Finland, the rates of both colon cancer and
rectal cancerare higher in the former country. What are the
9. In developing
countries
such as Colombia,
the rate of
polyps of the colon is considerably higher in cities than in
developed countries such as
dietary differences between these 2 populations that could
account for this?
11. The Seventh-Day Adventists, who consume less meat
NOVEMBER 1975
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1975 American Association for Cancer Research.
3389
E. L. Wynder
.MALE:
rs+.645O@
COLONICCANCER
L°
FE@@LE:
RECTAL
CANCER
160@
Men
-----
Women
350.
@:8 15
@
@,.
,@
U.S.
CA@D&
I@0@
140-
g@_
120@
@.250-
100W
@200-
60
@.150.
60
100
40
50
20
I-Lu
Lu
c3
5
-@
0
Lu
@
C
50
100
150
200
250
AGE-ADJUSTEDDEATH RATE OF ARTERIOSCLEROTICHEART
DISEASEPER 100.000
Chart 4. Correlation
0
ilIllIllIjIl
40 50 60 70 80 90
AgeatDiagnosis
between colon cancer (153) and arteriosclerotic
heart disease(420).
0
Ageat Diagnosis
Chart 6. Age-specific incidence for cancer of the colon and rectum in
white Americans(l969)(from Cutler, 1975).
@
.
0
.
0
•
0
@
@
@:
CJLON
CANCER
RECTUM
CANCER
0
0
1947
1969 1947
1969 1947
1969 1947
1969
All Regions
North
South
West
Chart S. Regional incidence for cancer of the colon and rectum in white
Americans (1947 and 1969). Rates adjusted on age distribution of total
United States population in 1950 (Cutler, S. J. Colon and Rectum. In: D.
Schottenfeld (ed.), Cancer Epidemiology and Prevention, pp. 375-385.
Springfield, Ill.: Charles C Thomas, Publisher, 1975).—, colonic can
CONNECICUT
cer; - - -, rectal cancer.
than others, are reported to have a relatively low rate of
colon cancer.2
12. Of particular interest have been the marked differ
ences in a wide variety of cancers, including cancer of the
colon but not cancer of the rectum, between Japanese and
American populations (18). These differences are not only
major but, since Japanese medical facilities and vital
statistics records are equal to those prevalent in the U.S., we
regard the differences to be real. There appear to be no
genetic factors affecting this low rate of colon cancer in the
Japanese, because the incidence of colon cancer increases
among those Japanese moving to America in line with their
adopting American dietary habits ( 10) (Chart 8). Of equal
interest is the fact that colon cancer seems to be increasing
in Japan itself, a finding consistent with the increasing
Westernization of its diet (4).
13. A rather ususual etiological lead is furnished by the
2 R.
L.
Phillips,
J.
W.
Kuzma,
F.
R.
Lemon,
and
R.
T.
Walden.
Mortality
from Colon Rectal Cancer among California Seventh-Day
Adventists. Presented at the Annual Meeting of the American Public
Health Association, San Francisco, Calif., November 8, 1973.
3390
—
MALE
iI@
PUERTO
RICO
jAPAN@
MIYAGI
CONNECTICUT
PUERTO
RICO
.JA@N.
MIVAGI
@FEMALE
Chart 7. Colon cancer incidence and rectal cancer incidence in Con
necticut, Puerto Rico, and Japan (Doll, R., Muir, C., and Waterhouse, J.
CancerIncidencein FiveContinents.Vol. 2. Geneva:U.I.C.C., 1970).
observation that patients with ureterocolonic anastomosis
have an increased risk of developing cancer of the colon
(12). What is particularly striking about this finding is that
patients who, early in life, have such anastomosis invofring
the colon, develop cancer at this site at a relatively young
age. Since such a correlation has not been reported for other
conditions, it has been suggested that urine itself in some
way acts as a tumorigenic agent. This unusual finding
certainly should be of interest to experimentalists
working
in colon carcinogenesis.
In summary,
the numerous leads derived from the
demographic distribution of cancer of the colon need to be
examined by the epidemiologists, interpreted for each of the
populations, and considered in line with the incidence of the
disease by its anatomic location in the large bowel, available
retrospective and prospective epidemiological data, and
general environmental data including nutrition.
CANCER RESEARCH VOL. 35
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1975 American Association for Cancer Research.
@
@,
@u
Large Bowel Cancer
Correa and Haenszel (4) have documented this finding by
calculating a sigmoid/cecum ascending site ratio obtained
from different cancer registries (Table 1). A particularly
Anatomic Distribution
The differences in the anatomic distribution of large
striking
bowel cancer among high- and low-risk population groups
aspect of the lower rate of colon cancer among
are of considerable interest to the epidemiologists. In Seventh-Day Adventists is that, among the 25 reported
general, there is a relatively greater incidenceof right-sided cases of colon cancer, 16 occurred in the right side of the
colon cancer in the low-risk population groups, a fact that is colon (R. L. Phillips and E. L. Wynder. Cancer of the
particularly evident when comparing the distribution of this Breast and Colon among Seventh-Day Adventists, manu
disease in American and Japanese populations (Chart 9).
script in preparation).
Along this line, it is not accidental
that among Japanese immigrants to Hawaii it is in particu
lar their rate of sigmoid lesions that increases (20).
As we have stressed previously, the incidence of cancer of
the rectum seems to differ significantly less between high
Ags @p@cific
DeathR@ssforWssW@olConcv
@
20
and low-risk population groups, especially when including
only those lesions occurring in the last 8 cm of the large
1OC
bowel. As we have also indicated, rectal cancer is more
common than colon cancer in males; this sex ratio differ
ence further supports our contention that its etiology is at
least in part different from that ofthe colon. Could it be that
t
the usual absence of feces in the lower rectum could
TableI
ii
Colon cancer incidence, in ascending order, as reported by cancer registries
f or 7 area?
Males
S/RIndia
S/C-A8S/RFem S/C-Aales
1964-1967)0.200.040.330.04Colombia
(Bombay,
100.420.12Japan(Miyagi,
(Cali, 1962- 1968)0.200.
1959-1961)
0.28Finland,
Puerto
Rico, 1950-19680.41
0.710.10 0.160.48 0.890.14
1964-19650.600.160.860.27Norway,
Chart 8. Age-specific mortality rates for colon cancer among Japanese
immigrants to Hawaii, Issei (1959-1962), and Nisei (1959—1962);United
1965-1966
1.560.55 0.680.92
Connecticut, 1960-19621.17
States Whites (1959-1961); Japan (1960-1961) [from Haenszel and
a 5@
Kurihara (10)].
b S/C-A,
Ref.
1.040.76 0.83
4, p. 389.
sigmoid/cecum-ascending;
SIR,
sigmoid/rectum.
COLON CANCER MORTALITY BY SITE IN U.S.A. AND JAPAN
AGE35-84,
@
R
sA
$968
JAPAN
MALE
Chart 9. Colon cancer mortality by site in
United States and Japan, 1968, ages 35 to 84.
FEMALE
0
0
0
0
0
a.
4
VERSE
VERSE
U.S/JAPAN
RATIO
2.5
.9
4.2
5.0
2.4
2.4
4.6
NOVEMBER 1975
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1975 American Association for Cancer Research.
6.8
3391
E. L. Wynder
contribute
to this difference
in incidence?
In this regard, we should note the observations by
Finegold (6) that the ratio of anaerobic to aerobic bacteria
increases as feces pass from the ileum to the rectum, with
the highest ratio in the sigmoid area (Table 2). Reasoning
would suggest that the anaerobic bacteria contribute in
some way to large bowel carcinogenesis.
Retrospective
Studies
In 1967, we conducted a large-scale retrospective study
(20) on large bowel cancer based on interviews with 761
patients with this type of cancer and concluded that, except
for the established high risk for patients with ulcerative
colitis and familial polyposis, no environmental
factors
could be identified that differed significantly between the
control and study populations (Table 3). Some of these
nonassociated
factors, i.e., constipation,
use of laxatives,
and frequency of bowel movements, were of considerable
interest, however. The fact that constipation did not relate
to large bowel cancer and that women were significantly
more constipated
than men without a correspondingly
higher rate indicated that transit time might not be an
important variable in cancer ofthe colon. We suggested that
there was no basis for the contention that increased transit
time resulted in a decreased bacterial action on the contents
of the colon. We believe that, even with the fast transit time
reported among Africans (2), there is sufficient time for
bacterial conversion of substrates in the content of the
colon. On the basis of a review of total epidemiological data
available at that time, we concluded that dietary factors,
especially the high intake of fat, appear to be associated
with the etiology of large bowel cancer. There are differ
ences in the etiology of cancer of the colon and of the
rectum. Etiologically, cancer of the rectosigmoid regions
Table 2
Bacterial counts at various positions of the gastrointestinal
All counts log10/gsample.
effluentFecesAnaerobes
TransverseIleostomy
colostomy
effluent
8.289.97Aerobes
5.72
7.366.59Total
8.29
8.419.97Anaerobes:aerobes
11000:1a
8.32
1:1000
10:
tract°
appears to fit more closely with cancer of the colon than
with cancer of the rectum.
The recent suggestionmade by Breslow and Enstrom (1)
on the basis of positive correlations between beer drinking
and rectal cancer in different countries could not be
confirmed on the basis of retrospective studies. In a similar
retrospective study on large bowel cancer patients in Japan
in 1969 (18), the data suggested a correlation between the
Westernization of the Japanese diet and colon cancer. We
did not find a similar correlation for cancer of the lower
rectum. It is reasonable to assume that differences and
changes in dietary intake, especially when involving West
ernization of the diet, could be determined by retrospective
techniques in a study in Japan. Such data-gathering
tech
niques do not, however, appear applicable to United States
populations. In the United States the total intake of fats,
carbohydrates,
and proteins is quite similar throughout the
country and between various groups, although the sources
from which these constituents come may be quite diversi
fled. Therefore, we have discontinued taking dietary histo
ries on American populations for some time, believing that
such information is relatively useless, both quantitatively
and qualitatively. This is particularly true when we are
interested in what was eaten over the last 20 years, rather
than what was eaten yesterday or during the previous year.
The “shoe
leather―epidemiologist should be the first to
recognize that, in his inability to take meaningful dietary
histories, he is faced with limitations that are unlikely to be
overcome by an improvement in interview techniques alone.
We also reported data on serum cholesterol in our 1967
study conducted in the United States (20) and again in our
Japanese study in 1969 (18). In neither study did we find a
correlation between serum cholesterol and an increased risk
of developing colon cancer, data that have been confirmed
in a prospective study by Rose et al. (16).
Thus, the correlation noted for myocardial infarction and
colon cancer cannot be explained on the basis of serum
cholesterol levels as was also apparent from the finding that
no correlation between myocardial infarction and colon
cancer has been observed in individual patients. This does
not deny, however, the possibility that diet, including
cholesterol, could be a common denominator. It could be
theorized that in one case the diet might contribute to an
increase in serum cholesterol leading to arteriosclerosis,
while in the other it might lead to an increased excretion of
cholesterol into the feces and/or an increase in the forma
tion of bile acids that, upon excretion into the feces and
upon being metabolized, may contribute to an increased risk
Ref. 6, p. 376.Table
3Epidemiologicalfindings
patientsPositive
oflarge
bowel cancer
Noassociation
associationUlcerative
useFamilial colitis
smokingCrohn's
polyposis
disease(?)
Tobacco
(?)Cigar
useConstipationOther
Alcohol
diseases(?)
largebowel
Obesity
3392
for colon cancer.
Selikoff et a!. (17) have presented evidence that asbestos
workers have an increased risk of developing colon cancer.
If this increased risk should prove to be real, it will be of
interest to consider the mechanismwhereby asbestosfibers
increase the chances ofdeveloping cancer ofthe large bowel.
Could such an increased rate relate to the carcinogenicity of
such asbestos fibers or could it be that asbestos fibers
localize within the mucosa of the large bowel and act as a
point around which various carcinogens in the contents of
lumen could crystalize?This lead should be further explored
both epidemiologically
and experimentally.
CANCER RESEARCH VOL. 35
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1975 American Association for Cancer Research.
Large Bowel Cancer
@
Nutrition
We
Cal.
have
proposed
that
nutritional
intake
is the
key
Fetal
Energy:
Proleln
0 carwi@ytiraie
etiological variable affecting the rate of colon cancer (19,
20). Burkitt et a!. (3), reaching a similar conclusion, have
emphasized that the fiber content of the diet may act as a
protective factor. They have stressed changes in the dietary
fiber, such as consumption time and consistency of stool
bulk. There is, however, no correlation between the fiber
content of foods and the incidence of colon cancer when
1t%X1
judged on a worldwide basis. Also arguing against refined
sugar as a factor is the fact that in Argentina, where the
incidence ofcolon cancer is high, the consumption of refined
VIt.A
U.
carbohydrates is low (14). Furthermore, Haenszelet a!. (8)
1*51
are high in fiber content.
a significant effect on bacterial flora, especially on the
concentration of anaerobic bacteria. The large-scale meta
bolic epidemiological studies by Hill et a!. (I 1) and similar
studies by Reddy and Wynder (15) have reached similar
conclusions. The fat content of the diet also increases the
substrates, in terms of bile acids and cholesterol, leading to
a greater production of bile acid metabolites and choles
terol metabolites among which, we suggest, there are com
ponents that possess tumorigenic activity. The questions
that need to be answered are: whether carcinogens are pres
ent in the colonic lumen or whether they are synthesized by
the colonic mucosa, and how the bacterial metabolites act
and/or interact.
The epidemiologist
in this regard can provide only the
leads. It has been up to the metabolic epidemiologist,
working together with chemists and bacteriologists,
as
exemplified by Williams and Hill in England and Reddy and
Mastromarino of our group, to: (a) establish the differences
in the makeup of the fecal constituents between high- and
low-risk population groups and between patients with colon
Vit.82
Vit.C
8L9
tag
2.0
2.0
1
4i@
1i@@
‘:E
have shown in the Hawaiian studies a positive correlation
between colon cancer and the consumption of legumes that
On the other hand, we have shown a correlation between
the intake of dietary fat and colon cancer. The studies by
Haenszel et a!. (8) in Hawaii have also shown a correlation
between beef and colon cancer. We are suggesting that there
may be nothing “carcinogenic―
about beef itself, but rather,
since meat contributes about 42% of the total fat calories in
our population (Chart 10), it is the fat and/or cholesterol
content of our diet that is principally related to cancer of the
large bowel. We are suggesting that the fat content of the
diet, which normally parallels cholesterol consumption, has
VIt.81
mg
U.S.J@tan
U.S.J@tan
U.S.Jipan
U.S.J@Lan
Chart 10. Intake of nutrients in the United States and Japan in 1962
(per person per day). Figures in bars, percentages of total calories. Sources
of data: statistical abstract of the United States, 1964; United States
Department
of Commerce;
Nutrition
in Japan
1963, Ministry
of Health
and Welfare,Japan.
Preventive Measures
A major epidemiological implication is that colon cancer
could be prevented by dietary modification. While we are
undertaking various programs in metabolic epidemiology
and experimental studies, the question that should be
answered is what dietary recommendation should be pre
sented to the public today. The history of preventive
medicine has repeatedly shown that waiting for all the facts
to be studied before making appropriate recommendations
frequently causes valuable years to be lost, years during
which preventive measures could have produced beneficial
effect. History has amply shown that preventive measures
have been successful long before the pathogenesis of a
disease was fully understood. In making recommendations
or suggestions, the risk/benefit issue must be considered.
No risks appear to be involved in suggesting that the present
American diet be modified to be consonant with the Prudent
Diet (Chart I 1). The benefits may already be considerable in
terms of the effect this diet may have on cardiovascular
diseases. Such a diet has been recommended
by many
experts concerned with cardiovascular diseaseprevention.
cancer, ulcerative colitis, familial polyposis, and polyps and What we are suggesting is that we join our colleagues in the
patients with no known large bowel disorders; (b) determine
how bacterial flora taken from patients with large bowel
disease and from high- and low-risk populations differ in
their ability to degrade bile acids and cholesterol; and (c)
give suggestions to the laboratory investigators as to certain
fecal components such as bile acids or cholesterol metabo
cardiovascular disease area by recommending the establish
ment of the Prudent Diet (both in terms of individual
practice and the modification of food at its source) so that
we can in fact have a national American Prudent Diet, a diet
especially essential for a largely sedentary population.
lites that should be tested in experimental animals. Further
data resulting from such an interdisciplinary effort will be Epilog
presented later at this conference. Another approach could
deal with a reduction of anaerobic bacteria by other than
dietary
means,
such as reducing
the pH of the stool, an
approach that deservesgreater attention.
It is apparent from this discussion and other presenta
tions at this conference that cancer etiology, perhaps more
than any other aspect of cancer research, requires an
NOVEMBER 1975
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1975 American Association for Cancer Research.
3393
E. L. Wynder
mg/day
cal/day
Rat.o
References
1. Breslow, N., and Enstrom, J. Geographic Correlates between Cancer
Mortality Rates and Alcohol-Tobacco Consumption in the United
States. J. Natl. Cancer Inst., 53: 631-639, 1974.
2. Burkitt, D. P. Epidemiology of Cancer of the Colon and Rectum.
Cancer, 28: 3-13, 1971.
3. Burkitt, D. P., Walker, A. R., and Painter, N. S. Effect of Dietary
Fiber of Stools and Transit Times and Its Role in the Causation of
Disease. Lancet, 2: 1408- 1412, 1972.
4. Correa,P., and Haenszel,W. ComparativeInternationalIncidence
TOTAL
CALORIC
INTAKE
CALORIES
FROM FAT
CROLES1EROL
INTAKE
P/S RATIO
Chart I 1. Prudent Diet (0) and present American diet (i).
and Mortality.
Prevention, pp.
lisher, 1975.
5. Correa, P., and
Cali, Colombia.
In: D. Schottenfeld (ed.), Cancer Epidemiology and
386-403. Springfield, Ill.: Charles C Thomas, Pub
Llanos, G. Morbidity and Mortality from Cancer in
J. NatI. Cancer Inst., 36: 717-745, 1966.
6. Finegold,S. M. The Normal Flora of Ileostomyand Transverse
interdisciplinary
approach. The epidemiologist can provide
basic leads; the chemist and bacteriologist can explore these
leads by using human material; and the experimentalist,
working in a laboratory situation with laboratory animals,
can test a variety of suspected large bowel carcinogens in a
number of combinations. The public health oriented scien
tist can make appropriate preventive suggestions. The
epidemiologist
will finally have to establish whether the
preventive measures have been successful in reducing the
incidence of disease. It needs to be recognized that each of
these specialists has limitations, but fortunately only his
own limitations. The animal experimentalist needs to be
aware of the fact that small laboratory animals cannot
develop a bacterial flora similar to humans, and thus
produce many of the cholesterol and bile acid metabolites
that humans are able to produce. What has taken evolution
millions of years to accomplish cannot be modified within a
few generations. There are parts of the human anatomy and
metabolism that cannot be duplicated in smaller animals.
On the other hand, it appears reasonable to assume that the
colonic wall, as a bioassay system, may be quite adequate to
test potential large bowel carcinogens and that the results
may then be applicable to the human condition.
The solution to the etiology of colon cancer requires a
coordinated
effort, which we are attempting
to carry out in
our institute and which is now being well organized within
the Division of Cancer Cause and Prevention of the
National Cancer Institute and the Task Force on Large
Bowel Cancer at M. D. Anderson Tumor Institute. As we
have reported, epidemiological evidence indicates that large
bowel cancer is largely man-made and thus is also largely
preventable. The interdisciplinary approach to this problem
can contribute to making such prevention a realistic goal.
3394
Colostomy Effluents. J. Infect. Diseases, /22: 376—38
1, 1970.
7. Gordon, T., and Kannel, W. B. The Framingham Study. J. Am. Med.
Assoc.,22!:661-680, 1972.
8. Haenszel, W., Berg, J. W., Segi, M., Kurihara, M., and Locke, F. B.
Large Bowel Cancer in Hawaiian Japanese. J. NatI. Cancer Inst., 5/:
1765-1779,
1973.
9. Haenszel, W., and Dawson, E. A. A Note on Mortality from Cancer of
the Colon and Rectum in the United States. Cancer, /8: 265—272,
1965.
10. Haenszel,W. M., and Kurihara, M. Studiesof JapaneseMigrants: I.
Mortality from Cancer and Other Diseasesamong Japanesein the
U.S. J. Nail. CancerInst.,40:43-51, 1968.
11. Hill, M. J., Drasar, B. S., Aries, V., Crowther, J. S., Hawksworth, G.,
and Williams, R. E. 0. Bacteria and Aetiology of Cancer of Large
Bowel. Lancet, 2: 95-100, 1971.
12. Kozak, J. A., Watkins, W. E., and Jewell, W. R. Neoplastic Stomal
Obstruction: A Complication of Ureterosigmoidostomy. J. Urol., 96:
691-696, 1966.
13. Martinez, I., Torres, R., and Frias, Z. Cancer Incidence in the United
States and Puerto Rico. Cancer Res., 35: 3265—3271,1975.
14. Puffer, R. R., and Griffith, G. W. Patterns of Urban Mortality.
Washington, D. C.: Pan American Health Organization, 1967.
15. Reddy, B. S., and Wynder, E. L. Large Bowel Carcinogenesis: Fecal
Constituents of Populations with Disease Incidence Rates of Colon
Cancer. J. NatI. Cancer Inst., 50: 1437-1442, 1973.
16. Rose, G., Blackburn, H., Keys, A., Taylor, H. 1., Kannel, W. B.,
Paul, 0., Reid, D. D., and Stamler, J. Colon Cancer and Blood
Cholesterol. Lancet, 1: 181- 183, 1974.
17. Selikoff, I. J., Churg, J., and Hammond, E. C. Asbestos Exposure and
Neoplasia. J. Am. Med. Assoc., /88: 22-26, 1964.
18. Wynder, E. L., Kajitani, T., Ishikawa, S., Dodo, H., and Takano, A.
Environmental Factors of Cancer of the Colon and Rectum: II.
Japanese Epidemiological
Data. Cancer, 23: 1210-1220,
1969.
19. Wynder, E. L., and Reddy, B. S. Studies of Large Bowel Cancer:
Human Leadsto ExperimentalApplication. J. NatI. Cancer Inst., 50:
1099-1106,
1973.
20. Wynder, E. L., and Shigematsu, T. Environmental Factors of Cancer
of the Colon and Rectum. Cancer, 20: 1520-1561, 1967.
CANCER RESEARCH VOL. 35
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1975 American Association for Cancer Research.
The Epidemiology of Large Bowel Cancer
Ernst L. Wynder
Cancer Res 1975;35:3388-3394.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/35/11_Part_2/3388
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1975 American Association for Cancer Research.