ICANCER RESEARCH57. 4787-4794, November 1, 1997)
Associations
with
of Sedentary
the Risk
Lifestyle,
of Colorectal
Obesity,
Smoking,
Alcohol
Use, and Diabetes
Cancer'
LoIc Le Marchand,2 Lynne R. Wilkens, Laurence N. Kolonel, Jean H. Hankin, and Li-Ching Lyu
Etiology Program, Cancer Research Center of Hawaii, University of Hawaii, Honolulu, Hawaii 96813
ABSTRACT
Variation
in the etiology
in colorectal
cancer rates between countries
and within eth
mc groups upon migration and/or Westernization suggests a role for some
aspects of Western
lifestyle in the etiology of this disease. We conducted
a
population-based case-control study in the multiethnic population of
Hawaii to evaluate
associations
between colorectal
cancer and a number
of characteristics of the Western lifestyle (high caloric Intake, physical
inactivity, obesity, smoking, and drinking) and some of their associated
diseases. We interviewed in person 698 male and 494 female United
States-born or immigrant Japanese, Caucasian, Filipino, Hawaiian, and
Chinese patients diagnosed in 1987-1991 with colorectal cancer and 1192
population controls matched on age, sex, and ethnicity. Conditional logis
tic regression
was used to estimate
odds ratios adjusting
for dietary
and
nondietary risk factors. Place of birth and duration of residence in the
United States were unrelated to colorectal cancer risk. Energy intake
(independent of the calorie source) and body mass index were directly
associated
versely
with
risk,
associated
and
lifetime
with risL
recreational
physical
The associations
activity
was
with these factors
in
were
independent of each other, additive (on the logistic scale) and stronger in
men. When individuals were cross-categorized in relation to the medians
of these variables, those with the higher energy intake and body mass
index and lower physical activity were at the highest risk (for males, OR,
3.0; 95% confIdence interval, 1.8—5.0, and for females, OR, 1.7; 95%
confidence interval, 1.0—3.2). Smoking in the distant, as well as recent,
of colorectal
disease.
Physical
activity,
both at work
and during leisure time, has consistently been inversely associated
with this disease (6—23).A direct association between caloric intake
and colorectal cancer has also been found in some studies and has
appeared to be independent of the dietary energy source (15, 16,
24—31). Data on obesity and colorectal
especially in women (32). Few studies
cancer have been inconsistent,
have considered energy intake,
energy expenditure, and weight gain in combination, although it has
been suggested that these factors act in concert (33).
Western lifestyle is known to often result in a host of other condi
tions (e.g., diabetes, hypertension, dyslipidemia, coronary heart dis
ease, gallbladder disease, arthritis, and constipation) that would be
expected to aggregate in colorectal cancer patients. It is also associ
ated with a number of lifestyle habits (e.g., smoking and drinking) that
are suspected to be risk factors for the disease and, thus, may act as
confounders of the energy balance-colorectal cancer association.
We report here on the findings for energy intake and expenditure,
body size, smoking, drinking, and “Westerndiseases―from a large
population-based case-control study of colorectal cancer conducted in
Hawaii
among
ethnic
groups
with
different
risks
for the disease.
Specific dietary factors associated with colorectal cancer risk in this
study were reported elsewhere (3 1, 34).
past and alcohol use were directly associated with colorectal cancer in
both sexes. Individuals with a history of diabetes or frequent constipation
were at increased risk for this cancer, whereas past diagnosis of hyper
cholesterolemia
was inversely associated with risL The findings were
MATERIALS
consistent between sexes, among ethnic groups, and across stages at
diagnosis, making bias an unlikely explanation. These results confirm the
data from immigrant studies that suggest that the increase in colorectal
cancer risk experienced by Asian immigrants to the United States oc
curred in the first generation because we found no difference in risk
by the rapid-reporting system of the Hawaii Tumor Registry, a member of the
between
ancestry
the immigrants
themselves
and subsequent
generations.
Patients were identified in all main hospitals on the island of Oahu, Hawaii,
They
also agree with recent findings that suggest that high energy intake, large
body mass, and physical inactivity Independently increase risk of this
disease and that a nutritional imbalance,
diabetes, may lead to colorectal cancer.
similar
AND METHODS
to the one involved in
Surveillance,
Epidemiology,
and End Results
Program
Institute. Eligible cases were all Oahu residents
of the National
Cancer
who had been diagnosed
between 1987 and 1991 with histologically confirmed adenocarcinoma of the
large bowel. Only patients who had any percentage
or at least were 75% Japanese,
Caucasian,
of native Hawaiian
Filipino,
or Chinese
were
included. Patients with a history of previous coloreclal cancer or who were
over 84 years at diagnosis were excluded. An interview was completed for
66% (698 men and 494 women) of the eligible cases. The main reasons for
nonparticipation were death before contact (15%), refusal (I I %), severe illness
(5%), and inability to locate (3%).
Controls were randomly selected from a list of Oahu residents interviewed
by the Hawaii State Department of Health as part of a health survey based on
INTRODUCTION
Colorectal cancer is predominantly a disease of Westernized coun
tries, with about two-thirds of the world cases occurring in developed
nations (1). Rates in immigrants from low-risk areas to these high-risk
areas are known to rapidly increase to the level of the host population
(e.g., in Japanese immigrants to the United States; Refs. 2—4).Rates
in countries undergoing Westernization have also been increasing
markedly (e.g., in Japan; Ref. 5).
Thus, some aspects of Western lifestyle, primarily a high caloric
intake and little physical activity, resulting in a positive energy bal
ance, weight gain and, ultimately, obesity, are suspected to play a role
a 2% random sample of state households. One control was matched to each
case on sex, ethnicity,
and age within
±2.5 years,
using age at diagnosis
for
cases and age at interview for controls. Controls with a previous history of
colorectal cancer were excluded. The overall participation rate for the controls
was 71%. Reasons for nonparticipation
locate
(5%),
serious
illness
included refusal (18%), inability to
(3%), and death
(3%).
Interviews were conducted in person, at the subjects' homes, by trained
interviewers. Generally, members of a case-control pair were interviewed by
the same interviewer. The questionnaire included detailed demographic infor
mation,
including
country
of birth, number
of years spent in the United
States,
and ethnicity of each grandparent; a quantitative food frequency questionnaire;
a lifetime history of tobacco and alcohol use; a personal history of various
Received 5/14/97; accepted 9/9/97.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement
18 U.S.C. Section 1734 solely to indicate this fact.
@
in accordance with
This research was supported in part by NIH Grant P0l-CA-33619 and contract
NOl-CN-05223
2 To
whom
from the National Cancer Institute.
requests
for
reprints
should
be
addressed,
at Etiology
Program,
Cancer
Research Center of Hawaii, 1236 Lauhala Street, Suite 407, Honolulu, HI 96813.
relevant medical conditions; a family history of colorectal cancer; a history of
constipation
and laxative
use; information
on height
and weight
at different
ages; and a history of recreational sport activities since age 18, including the
type, frequency, and duration of the activities. Usual physical activity was also
assessed using an adaptation of the Five-City Project questionnaire (35). This
method was selected because, when compared to seven other physical activity
questionnaires, it yielded an average estimate of total caloric expenditure that
4787
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1997 American Association for Cancer Research.
WESTERNLIFESTYLEANDCOLORECTAL
CANCER
most closely approximated caloric intake (96%; Ref. 36). With this method, we
inquired
about the numbers
of hours and minutes
spent in light, moderate,
and
intense activities, as well as sifting and sleeping, during a typical weekday and
weekend day during the diet reference period (see below). Cue cards listing
examples of light, moderate, and intense activities were used to help respon
Quetelet index 5 years ago; total caloric intake, and dietary fiber, calcium and
egg intake.
We performed likelihood ratio tests comparing models with and without a
trend variable to determine linearity in the logit of colorectal cancer risk across
dietary intake levels. The trend variable was assigned the median value of each
dents to classify the intensity level of their physical activity at home, work, and
quantile of dietary intake. We also used likelihood ratio tests to determine
leisure. Several days prior to the interviewer's visit, subjects were mailed a
self-administered lifetime occupational history to be filled out in advance of
the interview, documenting the job title, the type of industry, the dates of
employment, and the numbers of hours per week and months per year worked
for each position held. The interviewer checked this form and obtained any
interaction among certain variables with respect to colorectal cancer risk.
These tests compared a main effects, no interaction model with a fully
parameterized model containing all possible interaction terms for the variables
missing information at the time of the interview. The interviewer also meas
of interest.
The analysis was first conducted for each subsite of the large bowel (right
colon, proximal to the splenic flexure: 30% of the cases; left colon: 39% of the
ured the waist and hip circumferences of the subjects at the end of the
cases; rectosigmoid junction or rectum: 29% of the cases; the remaining 2% of
interview.
cases with unknown
The food frequency questionnaire used in this study has been described
previously (31, 34, 37) and validated in this population (38). Frequencies and
amounts consumed were obtained for more than 280 food items or categories.
The questionnaire was designed to include all foods commonly consumed in
our population and to capture the complete diet, including total caloric intake.
lance, Epidemiology,
Participants
portion
reported
their average
frequencies
of consumption
sizes for those items eaten at least 12 times a year during
and average
food items, showing three different portion sizes, as well as measuring cups
and spoons, were used during the interview to facilitate quantification of
intakes. The food composition data were primarily based on the nutrient
database of the United States Department of Agriculture (39) and supple
mented with data from other primary sources and other research and commer
(40—43).
For lifetime recreational physical activity, two indices were computed; one
was the sum of the time (in hours) spent in the reported activities since age 18;
the other summed the products of duration and intensity of each activity over
all reported activities. An intensity value (in METs;3 1 MET = 1 Kcal/kg/h)
was assigned to each activity using the extensive coding system in Ref. 44. An
index of usual physical activity during the diet reference period was computed
for a usual 24-h period by taking the weighted average of the products of the
duration (in minutes) and an intensity level (in METs) for each category of
activities over a weekday (multiplied by 5) and a weekend day (multiplied by
2). Sleeping was assigned an intensity level of 1 MET; sitting, 1.5 MET; light
activities,
4 METs;
moderate
activities,
6 METs;
and intense
activities,
10
METs. Occupational physical activity was assessed by assigning one of five
levels of physical activity (sedentary, light, medium, heavy, and very heavy) to
each job as rated by the United States Department of Labor. Occupations held
before 1960 were coded according to Department of Labor physical require
ments published in 1956 (45). Physical activity ratings published in 1966 (46)
were used to code the occupations held between 1960 and 1969. Those
published in 1981 (47) were used to rates jobs held after 1969. All physical
activity coding was performed without knowledge of the case-control status of
the subject.
Smokers and drinkers were considered as current users if they smoked or
drank up to the date of diagnosis for cases or the date of interview for controls.
Lifetime tobacco use was computed as pack-years (packs/day X years of
smoking);
analyses,
cigarettes,
pipes,
we used conditional
and
cigars
logistic
were
regression
treated
models
equivalently.
For
(48) to compute
all
ORs
(and 95% CIs) for several levels of an exposure variable, with adjustment for
relevant covariates. Each matched case-control pair made up a stratum for the
logistic
model.
Quartile
and tertile
cutpoints
were sex specific
and based
on
data from the controls. The sex-specific interquartile ranges for each contin
uous variable are presented in the table footnotes for comparison with other
populations. Exposure variables were adjusted by entry into the model of the
following variables found to be independently associated with risk (continuous
unless otherwise specified): age (to account for a differential between cases
and controls within the 5-year matching interval); family history of colorectal
cancer (using an indicator variable); number of alcoholic drinks per week;
pack-years of cigarette smoking; lifetime recreational activity (in hours);
3 The
abbreviations
used
are:
MET,
metabolic
equivalent;
CI,
confidence
were excluded
from these analyses).
The Surveil
and End Results summary staging classification
was
used in the stage-specific analyses: in situ tumors (9.4%) had remained
intraepithelial; localized tumors (44.0%) were confined to the colon or rectum;
regional tumors (38.8%) had either extended through the muscularis to adja
cent tissue or metastasized to regional mesenteric lymph nodes; and distant
tumors (7.0%) had metastasized to distant sites.
the reference
period, which was the 3-year period before onset of symptoms for cases and
the 3-year period before interview for controls. Color photographs of most
cial publications
subsite
interval;
OR,
odds ratio; BMI, body mass index.
RESULTS
Table 1 presents the demographic characteristics of cases and
controls by sex. Cases were less educated and were more likely to be
single, divorced, or separated than controls in both sexes. In males,
cases were more often Buddhist and less often non-Catholic Christian
than controls. Cases were more likely to have been regular smokers
(i.e., to have smoked daily for at least 6 months) and regular drinkers
(i.e. , to have drunk
alcoholic
beverages
once
a week
for at least 6
months) in the past than controls in both sexes.
The adjusted odds ratios for most of these variables and selected
past medical conditions are presented in Table 2. No association was
found with place of birth or length of residency in the United States.
Education
and religion
were
not significantly
associated
with
risk
after adjustment for confounders (data not shown). Men who were
single, divorced, or widowed were at a 3-fold increased risk for cancer
of the right colon. Marital status was not associated with risk for the
other subsites or in females. Former smokers were at somewhat
increased
risk of cancers
in the left colon and rectum
for males and the
right colon and rectum for females. No association was found with
years since quitting. For all three subsites combined, the odds ratios
for increasing quartiles of pack-years of tobacco smoking were 1.0,
1.2, 1. 1, and 1.6 (95% CI, 1. 1—2.3;P for trend = 0.04) for males and
1.0, 0.9, 1.0, and 1.6 (95% CI, 1.0—2.6;P for trend = 0.04) for
females. This association was strongest for the left colon in males and
the right colon and rectum in females. Odds ratio estimates were
elevated for past regular drinkers for all subsites but especially for the
right colon in both sexes. Current drinkers were also at increased risk
for cancer of the right colon in both sexes. For all subsites combined,
the odds ratios for increasing quartiles of drink-years (numbers of
drinks/day x duration of drinking in years) were 1.0, 1.5, 1.3, and 1.5
(95%CI, l.l-.2.2;Pfortrend
= 0.Il)formalesand
1.0, 1.2, l.6,and
1.6 (95% CI, 0.8—3.0;P for trend
0. 14) for females, respectively.
This association with lifetime use of alcohol was strongest for the
right colon in both sexes and was not limited to any specific type of
alcoholic beverage. A more detailed analysis of the association with
alcohol and its interaction with nutrient intake will be given in a
separate report.
An increased risk of cancer in the left colon was observed with a
history
of diabetes
in both sexes,
whereas
hypercholesterolemia
was
inversely associated with cancer risk at all subsites in both sexes. For
all subsites combined, the odds ratios for history of diabetes was 1.2
(95% CI, 0.8—1.7)in males and 1.8 (95% CI, 1.1—2.8)in females. This
association was stronger when diabetes had been diagnosed at least 9
4788
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1997 American Association for Cancer Research.
WESTERNLIFESTYLEANDCOLORECFALCANCER
Table 1 Selected characteristics
of colorectal cancer cases and controls bysex, Hawaii.
1987—1991@Males
(698 case-control pairs)Females
(%)
(%)Cases
(%)Age
pairs)Cases
(494 case-control
(%)Controls
Controls
(median in years)67
656565Ethnicity
Japanese
(50)
Caucasian
Filipino
347 (50)
180(26)
70 ( I0)
(49)
116(23)
50 ( I0)
180(26)
70 ( 10)
Hawaiian
52 (7)
52 (7)
42 (9)
Chinese347
42(9)Biiihplace
49(7)
49(7)244
42(9)244
UnitedStates
(49)
116(23)
50 ( I0)
42(9)
615(88)
Asia
Other614(88)
6(1)Education
75(11)
6(1)
75(11)
6(1)
78(16)
11(2)P=0.47402(82)
80(16)
P=l.Ob399(82)
<High school
Highschool
(36)
184(26)
197 (28)
185(26)
(38)
137(28)
Some college
College250
86(17)Marital
156 (22)
108(16)
150 (22)
166(24)
I 10(22)
61 (12)P=0.0l150(30)
586 (84)
112 (16)
(58)
210 (42)P=0.30300(61)
194
125(25)
133 (27)
P=0.00l186
status
Married
Single/divorced/widowed533
(39)Religion
(76)
165(24)
P=0.OOl284
None
(16)
106(15)
Catholic
156(22)
147(21)
123(25)
124(25)
Other Christian
Buddhist
209 (30)
213(31)
257 (37)
175(25)
183 (37)
149(30)
173 (35)
154(31)
Other111
3(l)Smoking
5(1)
9(1)
3(1)P=0.9738(8)
P=0.0435(7)
status
Never
Former
Current167(24)
64(13)Drinking
217(31)
364(52)
117(17)
434(62)
97(14)
135(27)
60(12)P=0.05328(66)
102 (21)
P=0.00l299(60)
status
Never
Former
Current191
(28)
226(33)
277 (40)
248 (36)
164(24)
284 (41)
(74)
(79)
62(13)
43(9)
67 (14)P=0.ll388
62 ( I3)
P=0.OOl363
a Numbers
@
do
a p@ were
not
always
computed
from
add
the
up
to
test
total
of
number
of
cases
and
controls
because
of
years before diagnosis for cases or interview for controls [OR, 1.4
(95% CI, 0.2—2.3)for males and 2. 1 (95% CI, 1.2—3.7)for females].
Men or women with a history of elevated serum cholesterol had half
the risk of colorectal cancer of those who did not [OR, 0.5 (95% CI,
0.3—0.6) for males and 0.5 (95% CI, 0.4—0.7) for females]. This
inverse
association
was diagnosed
constipation
was observed
even
when
hypercholesterolemia
at least 7 years earlier. Frequency of episodes of
during
missing
values.
association.
the past 5 years was directly
associated
with risk
of cancer of the left colon in both sexes and the right colon in females
in a dose-dependent manner. It was not associated with rectal cancer.
The OR for colorectal cancer (all sites combined) for subjects who
reported having experienced constipation once a year compared to
once a month during the past 5 years was 0.5 (95% CI, 0.4—07) for
males and 0.6 (95% CI, 0.4—0.9)for females. Use of laxatives during
the past 5 years was inversely associated with risk after adjusting for
terms for the time periods 1—5,6—10,11—20,2 1—30,3 1—40,41—50,
and
5 1 years prior to colorectal cancer diagnosis for cases and
interview for controls. Never smokers were assigned 0 for all pack
year variables. ORs are given for a arbitrarily chosen difference of 10
pack-years (i.e. , the values given are the incremental odds ratio for 1
pack-year raised to the 10th power). Although estimates of the odds
ratios for pack-years progressively increase with the remoteness of the
smoking
period,
risk was significantly
elevated
for the most
recent
periods as well, for both colon and rectal cancers. These data imply
that the effect of smoking in the distant past may be greater [as
recently suggested (49)] but that recent smoking also increases risk of
colorectal
cancer.
The adjusted odds ratios for caloric intake, physical activity, and
body size are included in Table 4 and by subsite of the large bowel in
Table 5. Energy intake was directly associated with colorectal cancer
in males only. This association was stronger for the left colon and
frequency of constipation
[OR, 0.5 (95% CI, 0.3—0.8) for males and
rectum than for the right colon. Hours spent in recreational physical
0.6 (95% CI, 0.3—1.1)for females]. There was no significant associ
activity during the subjects' lifetimes were inversely associated with
ation between colorectal cancer and history of coronary heart disease,
risk in both sexes, although a monotonic trend was seen in men only.
high blood pressure, stroke, arthritis, duodenal or gastric ulcers,
Odds ratios below 1.0 were observed for all subsites of the large
cholecystectomy, appendectomy, adenoidectomy, or asthma (data not
bowel. Very similar risk estimates were found when the intensity (in
shown).
METs) of these recreational activities was taken into consideration
Table 3 presents the colon and rectal cancer risks associated with
(data not shown), suggesting that intensity may not be as important as
pack-years
of smoking during various periods of time in the past. A
duration in decreasing risk. There was no indication in our data that
model was fit that simultaneously included continuous pack-year
4789
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1997 American Association for Cancer Research.
WESTERN LIFESTYLE AND COLORECFAL CANCER
colorectal cancerby selected demographic, snwking, drinking.and
Table 2 OR? (95% CIs)for 1987—1991MalesFemalesRight
129)Place
colon
197)bLeft
(n
colon
270)Rectum
(n
medical history vanables, Hawaii
colon
164)Left
(n
(n = 221)Right
colon
194)Rectum
(n
(n
BirthUnited
of
States1.01.01.01.01.01.0Asia1.8
(0.4—2.9)Years
(0.5—5.8)0.7
(0.3—2.0)0.9
(0.4—2.1)0.7
(0.3—1.8)1.1
(0.6—2.3)0.8
(0.3—1.7)2.0
(0.2—2.0)
(0.7—5.8)0.8
(0.5—6.7)
(0.6—2.1)1.1
States591.01.01.01.01.01.060—680.8
in the United
(0.4—1.7)1.4(0.5—4.1)@691.4
(0.3—1.9)1.1
(0.5—3.0)
P = 0.700.7
(0.5—4.1)
P = 0.78c1.2
0.90Marital
(0.2—3.3)
(0.3—2.4)
P =
P = 0.630.9
P = 0.210.9
P = 0.451.9
statusMarried1.01.01.01.01.01.0Divorced/single/widowed3.1
(0.7—2.3)Smoking
(0.5—1.4)1.2
(0.5—1.9)0.8
(0.6—1.7)1.0
(0.7—2.2)1.0
(1.5—6.5)1.3
statusNeverI
.01.0Former1.0(0.5—1.9)1.4(0.9—2.4)1.4(0.8—2.3)2.4(1.0—5.6)1.1
.01
.01
(0.6—2.0)1.6(0.7—3.4)Current0.7
(0.5—3.7)Pack-years―T11.01.01.01.01.01.0T,1.1
(0.4—1.9)0.8
(0.3—1.6)0.9
(0.5—4.4)T30.8
0.35Drinking
statusNever1
.01.0Past2.6
(0.6—4.1)Current1.8
(0.5—1.6)1.2(0.7—2.1)2.1
(0.7—2.5)
(1.1—3.5)
P = 0.411.6
P = 0.0061.3
.01
.0I
(0.8—3.3)1.4
(0.4—2.6)0.7
(0.3—1.5)1.3
(0.9—5.1)0.5
(0.7—3.6)
P = 0.471.3
(0.3—1.2)1.5
(0.7—2.5)
P = 0.411.5
.01
(1.0—9.4)1.3
.01
(0.8—2.4)3.1
(0.6—2.0)2.5
.1 (0.7—2.0)1.1
(1.2—4.0)1.2
(0.5—3.4)1.5
(0.3—2.0)1.1
(0.6—1.8)
P = 0.591.4
(0.7—2.1)
P = 0.641.0
(0.7—3.0)
P =
(0.9—7.0)1.0
(0.7—2.0)1.6(0.9—2.8)2.7(1.1—6.9)0.8
(1.1—3.6)
P = 0.141.3
0.40History
(0.4—1.8)1.1
(0.6—2.2)0.9
(0.4—1.6)
P = 0.412.0
(1.4—5.2)1.7
(0.5—2.3)1.0(0.3—3.0)Drink-years―T11.01.01.01.01.01.0T22.2
(1.0—3.4)1
(0.4—2.7)T32.0
.01
.01
(0.4—1.8)
P=
P = 0.681.6(0.6—4.5)
(0.5—3.9)
P = 0.680.8
diabetes―No1.01.01.01.01.01.0Yes0.9
of
(0.7—4.4)History
.7
(0.4—1.8)1.9
(1.1—3.5)0.7
(0.3—1
.6)1.3
(0.6—2.7)3.0
(1.2—7.1)1
(0.3—0.9)0.4
(0.3—0.7)0.4
(0.2—0.7)0.5
(0.2—0.9)0.4
(0.2—0.7)0.7
(0.5—2.3)0.7
(0.4—1.5)
(0.3—1.5)0.4
(0.1—0.5)
(0.2—1.0)0.5
(0.4—1.6)
(0.2—1.3)0.6
(0.2—0.7)
hypercholesterolemiINo1.01.01.01.01.01.0Yes0.5
of
(0.3—1.3)Frequency
yrI/month1.01.01.01.01.01.01/3—6
of constipation in past 5
(0.4—3.1)1/year0.8
mo.1.0
P = 0.00010.8
P = 0.340.3
(0.3—1.3)1.1
(0.3—1.0)
P = 0.061.0(0.5—1.9)
P = 0.91
P = 0.0050.5
P = 0.990.3
a Adjusted by conditionallogistic regressionfor age; family historyof colorectalcancer;alcoholicdrinks/week (except in modelsfor drinking statusand drink-years); pack-years
(except in models for smoking status and pack-years); lifetime recreational physical activity; BMI 5 years ago; and egg, dietary fiber, calcium, and caloric intakes.
b No. of case-control pairs. These numbers do not add up to the total number of cases and controls because of missing subsite information.
‘P for
trend.
T tertile.
The interquartile
range (25th—75th percentiles)
for pack-years
was 0—39 for males and 0—28 for females, and for drink-years,
0—83 for males and 0—29 for females.
Drink-years = no. of alcoholic drinks/day X duration of drinking in years.
e No.
of
subjects
with
positive
history:
males,
122
cases
and
87
controls;
females,
82
cases
and
51
controls.
I No. of subjects with positive history: males, 111 cases and 208 controls; females, 92 cases and 163 controls.
the effect of recreational physical activity was greater for any life
period in the past. An inverse association was also present for usual
daily physical activity, but it was present in females only. This
association in females was almost limited to the right colon. There
was also the suggestion of a direct association between the number of
years spent in jobs involving only sedentary or light work and cancer
of the right colon in both sexes.
No association was found with height, whereas an inverse associ
Table 3 Colon and rectal cancer risks―
associated
1987—1991Colon
with 10 pack-years of tobacco smoking in various time periods, Hawaii,
RectumMale
Female
Male
Female
CI1—5
ORb
95% Cl
OR
95% Cl
OR
95% CI
OR
years ago
6—10years ago
1.09
1.09
(1.04—1.14)
(1.04—1.15)
1.16
1.18
(1.06—1.28)
(1.07—1.31)
1.08
1.09
(1.01—1.16)
(1.01—1.16)
1.08
1.08
(0.94—1.24)
(0.94—1.18)
I 1—20
years
21—30
years
31—40
years
41—50
years
1.10
1.13
1.17
1.25
(1.05—1.16)
(1.06—1.20)
(1.07—1.29)
(1.07—1.46)
1.21
1.30
1.54
1.94
(1.08—1.35)
(1.11—1.51)
(1.19—1.99)
(1.14—3.28)
1.09
1.12
1.12
1.20
(1.01—1.18)
(0.98—1.28)
(0.98—1.28)
(0.97—1.50)
1.08
1.11
1.11
1.13
(0.92—1.28)
(0.87—1.35)
(0.78—1.57)
(0.58—2.20)
ago
ago
ago
ago
(0.20—35.17)a
5l years ago
The
ORs
represent
1.40
the
risk
associated
(1.01—1.94)
with
a difference
2.77
of
10
pack-years
(0.69—11.0)
(i.e.,
they
are
the
incremental
1.46
OR
for
(0.93—2.30)
1 pack-year
b ORs adjusted by conditional logistic regression for age; family history of colorectal cancer; alcoholic drinks/week;
raised
to
95%
2.66
the
10th
power).
lifetime recreational physical activity; BMI 5 years ago; egg,
dietary fiber, calcium, and caloric intakes; and the other variables in the table.
4790
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1997 American Association for Cancer Research.
WESTERNLIFESTYLEANDCOLORECFAL
CANCER
Table 4 ORs―(95% CIs)for
colorectal 1987—1991Males
cancer by quartile of energy intake, physical activity, and body size, Hawaii,
pairs)Qb
(698 case-control pairs)
(high);(low)
Females (494 case-control
Q4 (high);
trendCaloric
Q2
Q3
QI
95%CI
P for trend
Q4
(low)
Q2
Q3
95%CI
P for
0.61Lifetime
intake
0.32Dailyrecreational physical activity (h)
1.0
1.0
1.1
0.8
1.2
0.7
2.0(1.3—3.0)
0.6 (0.4—0.8)
0.001
0.007
1.0
1.0
0.9
0.7
1.0
0.7
1.1 (0.7—1.8)
0.7 (0.5—1.1)
0.05Years
physical activity (METs)
0.10Height
in sedentary or light work
0.70Current
I .0
I .0
1.0
1.2
1.5
0.7
1.5
1. 1
0.6
1.0 (0.7—1
.5)
1.3 (0.9—1
.9)
0.9 (0.6—1.4)
0.94
0.61
0.15
1.0
1.0
1.0
0.9
0.8
1.0
0.8
0.9
1.0
0.6 (0.4—1
.0)
1.5 (0.9—2.3)
1.1 (0.6—2.0)
0.04Weight
weight
0.28BMI5 years ago
0.21BMI
5 years ago
0.03BMI
at age 35
1.0
1.0
1.0
1.0
0.8
1.3
1.2
1.0
0.9
1.8
1.5
1.2
0.9 (0.6—1.3)
2.1 (1.4—3.3)
2.2 (1.5—3.2)
1.5 (1.0—2.2)
0.64
0.0003
0.0001
0.02
1.0
1.0
1.0
1.0
0.8
1.3
0.8
0.9
0.7
1.3
0.4
1.0
0.6 (0.4—1.0)
1.3 (0.8—2.1)
1.2 (0.8—1.9)
1.6 (1.0—2.5)
0.06BMI
at age 25
1.0
1.2
1.2
1.5(1.0—2.2)
0.05
1.0
1.0
1.3
1.5(0.9—2.3)
0.06Adult
at age 15
0.56Current
weight changec
0.71a waist-to-hip ratio
1.0
1.0
1.0
0.8
1.0
1.2
1.0
1.6
1.3
1.0(0.7—1.5)
1.6 (1.0—2.4)
1.2 (0.8—1.8)
0.62
0.003
0.26
1.0
1.0
1.0
1.2
0.7
0.9
1.1
0.9
1.1
1.6 (1.0—2.6)
0.8 (0.5—1.3)
1.1 (0.7—1.7)
(exceptin
Adjusted by conditional logistic regression for age; family history of colorectal cancer; alcoholic drinks/week; pack-years; egg, dietary fiber, and calcium intakes; calories
models for caloric intake); lifetime recreational physical activity (except in models for physical activity variables); and BMI 5 years ago (except in models for body size variables).
@
physicalactivity
b
quartile.
The interquartile
range (25th—75th percentiles)
for the variables
were as follows:
calories
(kcal),
males,
1671—2761, females,
1281—2096; lifetime
recreational
(h), males, 408—10,080,females, 864—3,300;daily physical activity (METs), males, 47—62,females, 49—61;years in sedentary or light work, males, 0—35,females, 6—30;
height (cm), males, 165—175,females, 152—163;current weight (kg), males, 63—78,females, 50—64;weight 5 years ago (kg), males, 63—78,females, 50—64;BMI 5 years ago (kg/m2).
18—21,females,
males, 22—26,females, 21—26;BMI at age 35 (kg/rn2), males, 21—25,females, 20—23;BMI at age 25 (kg/m2), males, 20—23,females, 19—22;BMI at age 15 (kg/m2), males,
0.78—0.87.C
17—21;
adult weight change (kg), males, 2—14,
females, 2—I
1; waist:hip ratio, males, 0.88—0.96,females,
Weight change between age 25 and 5 years ago (i.e., 5 years prior to diagnosis for cases or interview for controls).
ation was found with current weight (i.e., at diagnosis for cases and at
interview for controls) in females. In contrast, weight S years prior to
that time was directly associated with risk, especially in men. The
association with BMI was strongest for BMI 5 years ago in men and
for BMI at age 35 in females. The association with BMI S years ago
in men extended to all segments of the large bowel, whereas the
association with BMI at age 35 in women was present for the colon
only. The association with weight change between age 25 and S years
prior to diagnosis or interview in men was weaker than that with
1987—1991Right
cancer by tertile ofenergy
Table 5 OR?for cobrectal
weight in the latter period. No association was found with the waist
to-hip ratio in men or women.
Table 6 presents the odds ratios for lifetime recreational physical
activity for each ethnic group. Decreased odds ratios were found for
8 of the 10 ethnic-sex groups, although monotonic trends were not
always observed, and some of the risk estimates were unstable due to
small sample sizes. Similarly, an increased colorectal cancer risk was
observed in 8 of the 10 ethnic-sex groups for BMI (data not shown).
The association with history of hypercholesterolemia was also ob
mtake, physical activity, and body size. Hawaii.
(270/194)Rectum(221/129)T1T2
colon (197/1f,4)bLeft
T3PCaloriesMale1.00.6
colon
T3PT1T2T3PT1T2
1.80.051.01.72.70.001Female1.01.7
1.30.251.01.5
1.50.231.00.80.80.67Lifetime
activityMale1.01.4
recreational physical
0.70.341.00.50.50.07Female1.01.9
0.60.221.00.40.80.97Daily
(METs)Male1.01.2
physical activity
1.00.921.02.01.20.51Female1.00.3
0.70.101.00.8
0.60.081.00.7
1.10.861.01.00.60.19Years
0.20.0031.01.8
1.10.961.01.3
0.90.601.01.2
workMale1.01.6
in sedentary or light
1.80.061.01.6
1.30.261.00.90.80.49Female1.00.8
1.00.881.01.71.10.87BMI
2.10.051.00.7
agoMale1.01.4
5 years
1.80.031.01.72.90.001Female1.01.6
1.90.041.01.2
1.10.841.01.00.80.45BMI
2.00.091.01.3
35Male1
at age
.00.92.00.02Female1.01.6
.01
.0
.01
131
1.50.
2.60.021.01.9
.2
0.80.451
1.80.051.00.71.40.24
a Adjusted by conditional logistic regression for the same variables as in Table 4.
b No. of male case-control pairs/no. of female case-control pairs. These numbers do not add up to the total number of cases and controls because of missing subsite information.
991MalesFemalesT1T2T3PT1
Table 6 ORs―
for colorectal cancer by tertile of lifetime recreational physicaI activity by ethnicity, Hawaii.1987—1
T3P―Japanese
T2
.0
.0
(337/236)c
I.0
0.8
0.8
Caucasian (174/1 12)
1.0
0.5
0.6
0.40
1.0
0.5
0.3
0.03
Filipino(69/48)
1.0
0.6
0.4
0.19
1.0
0.2
0.9
0.87
1.0
0.7
0.2
0.37
0.071
1.0
1.00.5
6.9
0.81
0.9
0.30.43
0.33
Hawaiian (49/41)
0.17aChinese (44/38)
1.0
0.8
0.20.30
Adjusted by conditional logistic regression for the same variables as in Table 4.
b p for linear trend.
C No.
of
male
case-control
pairs/no.
of
female
case-control
pairs.
4791
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1997 American Association for Cancer Research.
WESTERNLIFESTYLEANDCOLORECFALCANCER
Table 7ORs―for
colorectal
cancerby
the diree.way interactionbetweencaloric
intake.
lifetime recreational activity, and BMI 5 years ago, Hawaii,
1987—199!Lifetime
activityMalesFemalesMedian>Median
hours in recreational
MedianCaloriesBMI
>Median
5 yr agoOR95%
Median
@Median
>Median
>Median
@Median
>Median
@Median
>Median
CIOR95%
1.0
2.4
1.8
3.0
(1.4—3.9)
(1.0—3.0)
(1.8—5.0)
CI
1.2
1.5
1.5
1.8
OR
(0.7—2.1)
(0.9—2.5)
(0.9—2.4)
(1.1—3.0)
1.0
1.5
1.3
1.7
95% CI
OR
95% Cl
(0.4—1.5)
(0.8—2.6)
(0.7—2.4)
(1.0—3.2)
0.8
1.5
0.9
1.2
(0.8—2.7)
(0.4—1.7)
(0.6—2.2)
a Adjusted by conditional logistic regressionfor age; family history of colorectalcancer;alcoholicdrinks/week;pack-years;and egg, dietary fiber, and calcium intakes.
served in 8 and that with history of diabetes in 6 of the 10 groups.
Stratified analyses by stage showed that the associations with calories,
obesity, smoking, alcohol, and high serum cholesterol were observed
for early, as well as advanced, stages, and associations with lifetime
recreational physical activity were observed for the regional and
distant stages.
Table 7 explores the three-way interaction between caloric intake,
BMI 5 years ago, and lifetime hours in recreational physical activity.
Each of these three risk factors appears to be independent and to have
an additive effect (on the coefficients in the logistic model), especially
in men. In both sexes, the most elevated ORs were observed for the
highest categories of caloric intake and BMI and the lowest category
of physical activity. Similarly, there was no interaction between
smoking and drinking on the risk of colorectal cancer.
DISCUSSION
In this study, because we matched on ethnicity, we were unable to
compare the colorectal cancer risks among ethnic groups. However,
cancer registry data have clearly shown that rates for this cancer are
high for Asians in the United States (1, 5). This study confirms the
observation made in ecological studies, namely, that Asians who
migrated as young adults to the United States do not develop cob
rectal cancer less frequently than their local-born counterparts and,
thus, in contrast to other cancers, such as breast cancer, that this
increase in risk occurred as early as the first generation. We also found
that duration of residence in the United States was not associated with
risk, in contrast to the case-control study of Whittemore et al. (15)
among American Chinese, in which an association was found with
number of years spent in North America. This discrepancy probably
reflects the fact that the majority of Chinese Americans in the latter
study had immigrated 5—15years previously to North America, an
interval that may have been too short to modify their risk, whereas
Japanese, Chinese, and Filipino immigrants in our study had lived in
Hawaii for many more years (42 years for males and 29 years for
females, on the average). What characteristics of the Western lifestyle
adopted by Asian immigrants to Hawaii may have contributed to their
increased risk of coborectal cancer?
Intake of energy beyond what is needed for energy expenditure,
resulting in obesity, is observed commonly in Westernized countries
and rarely in Asia. In this study, we found that caloric intake [inde
pendently of the source of calories (31)], physical activity, and obesity
were all independently associated with coborectal cancer risk and that
overweight individuals with high caloric intake and little physical
activity had the highest risk. The data suggested that these associa
tions were quite consistent
across
ethnic
groups
but were somewhat
stronger in men, especially for caloric intake and obesity. There was
no clear specificity of these associations for any subsite of the large
bowel.
In many case-control studies, cases have reported a higher total
caloric
intake than controls,
usually
in both sexes (15, 23—30). Among
the four studies that partitioned energy intake by dietary sources of
calories, only one study found that the caloric effect was due to a
single source of calories (15); the others found associations with
several sources (27, 28, 30). Thus, in agreement with our results, data
from case-control studies support an independent effect of energy on
colorectal cancer risk. This is in contrast to the findings from cohort
studies,
which
generally
did not show
an association
with
caloric
intake (50—53).The most frequently proposed explanation for this
discrepancy has been recall bias in case-control studies. Although this
is possible,
differential
measurement
error
does
not appear
likely
because findings from recent studies using both types of design have
been very consistent for a multitude of dietary and nondietary risk
factors (e.g., red meat, dietary fiber, calcium, fruits and vegetables,
alcohol, and physical activity). Other possibilities are selection bias
and inadequate caloric intake assessment in cohort studies because
they are rarely population-based and, by necessity, use a shortened
dietary questionnaire.
Physical activity has consistently been associated with a decreased
risk of coborectal cancer in past studies. Although most studies have
focused on occupational physical activity (6—8, 12—17,19), studies
examining leisure time and total physical activity (9, 11, 13, 15, 16,
18, 20) or participation in college athletics (20) have also shown a
reduced risk for the most active. Also consistent with our results, this
protective effect has been observed across a gradient of activity levels
and is not limited to strenuous exercise (7, 9, 10, 13, 18, 19, 20, 28).
A subsite specificity has not been shown for physical activity, al
though rectal cancer studies have been more inconsistent. Occupa
tional studies have usually examined cumulative lifetime physical
activity, whereas most other studies have examined activity levels a
few years before diagnosis. Few studies have attempted to determine
at which period in life physical activity has its greatest impact. Marcus
et a!. (21) found
that strenuous
activity
during
ages
14—22 was not
associated with colon cancer in women. Lee et a!. (20) reported that
adult physical activity assessed on two occasions 11—15years apart
was associated with risk, whereas either measurement alone was not.
In our data, although there was the suggestion of an association with
recent overall physical activity and lifetime occupational physical
activity, the strongest association was found with cumulative recre
ational physical activity during adulthood.
Most past studies have also shown a direct association between
body mass and coborectal cancer (9, 15, 16, 26, 32, 54—57),with men
in the highest exposure category having approximately a 2-fold in
creased risk. Data for women have been more inconsistent. with a
majority of studies showing no association (9, 15, 16, 26, 56—58).Our
results suggest that this discrepancy may be due to either a somewhat
weaker effect in women, an age-specificity for the association, or
weight loss due to the disease process, because weight was usually
assessed
examined
only a few years prior to diagnosis.
body fat distribution
The only other study that
also found no association
between
4792
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1997 American Association for Cancer Research.
the
WESTERNLIFESTYLEANDCOLORECTAL
CANCER
waist-to-hip ratio and colorectal cancer risk (51). Height has usually
not been associated with risk (32, 55, 58—60).
Several biological mechanisms have been pmposed to explain the
associations
of coborectal cancer with energy balance,
physical
activity,
and obesity. Animal studies have well documented that moderate energy
restriction reduces tumor development (61). Physical activity stimulates
colon peristalsis, decreasing the time that dietary carcinogens reside in the
colon. Exercise is also known to enhance the immune defense by affect
ing T cells, B cells, natural killer cells, and interleukin 1 levels (62). In
recent years, two related unifying hypotheses have been proposed based
on similarities of risk factors for coborectal cancer and metabolic profiles
that may promote cobonic tumors. McKeown-Eyssen (33) noted that
several aspects of Western diets, including alcohol use and physical
inactivity,
increase
serum
glucose
and triglycerides,
which
themselves
However,
our data are not completely
consistent
with
the
hyperinsulinemia hypothesis because the association with diabetes was
limited to the left colon (in both sexes), whereas those with the risk
factors common to coborectal cancer and hyperinsulinemia did not show
such site specificity.
Few past studies have reported on diabetes and colorectal cancer
risk. Most have found no association (64—67). However, our results
agree with those of two other retrospective studies (68—69) and a
prospective study (70) in which an increased risk was found among
subjects with a history of diabetes, as well as with an animal study in
which injections of insulin promoted colon tumors in rats (71).
Cases in our study were less likely than controls to have been diag
nosed with hypercholesterolemia in the past. This is consistent with the
inverse association frequently observed in past studies between serum
cholesterol and coborectal cancer (72). Detailed analyses of some of these
studies have shown that this association appears to result from a progres
sive decline in serum cholesterol starting about 8 years before diagnosis,
presumably as a result of a precinical effect of the disease (73, 74).
Although tobacco smoking has consistently been found to increase
the risk of adenomatous polyps, which are considered precursor
lesions for coborectal cancer, it was not associated with this cancer
before the l980s (75). However, studies conducted several decades
after the onset of the smoking epidemic in the United States have been
more consistent in showing an association (75). It has been proposed
that smoking increases the risk of large bowel cancer only after a
4-decade induction period, as suggested in two cohort studies in which
smoking
in the distant
past was associated
with
coborectal
associations
for this association,
we genotyped
of in situ
coborectal
cancer
with
the generalizability
of our results
to the entire
patient
popu
ter, Straub Clinic and Hospital, St. Francis Medical Center, Tripler Medical
Center, and Wahiawa General Hospital for their support of this study. The
authors also thank Aleli Vinoya for assistance with computer programming.
REFERENCES
cancer,
a germ
in this study may reflect
We thank the Hawaii Tumor Registry, Castle Medical Center, Kaiser
Permanente Medical Center, Kuakini Medical Center, Queen's Medical Cen
a subset of our subjects
in individuals
observed
ACKNOWLEDGMENTS
for whom we had collected a blood sample. We found an increased
risk
activity
lation. However, the associations found in this study were equally
strong for each stage, except the associations with physical activity,
which was more marked for the most advanced stages. Thus, the risk
estimates were probably not overestimated but could have been un
derestimated for exercise because patients who were missed due to
early death were more likely to have an advanced form of the disease.
In summary, these findings confirm the data from immigrant stud
ies suggesting that coborectal cancer risk in Asian immigrants to the
United States increased rapidly, as early as in the first generation.
They also agree with recent findings suggesting that high energy
intake, large body mass, and physical inactivity independently in
crease risk of this disease and with the proposal that a metabolic
imbalance, leading to hyperinsulinemia and diabetes, may be involved
in the etiology of coborectal cancer.
whereas more recent smoking was not (76). Our data also showed an
association with smoking in the distant past but did not find evidence
that recent smoking would not affect risk as well. Exploring a possible
mechanism
with physical
an over-representation of health conscious individuals in the control
group. However, this does not appear likely because our participation
rate was high (71%) and because the associations were found with
surprising consistency among ethnic groups that vary in their cultural
beliefs and socioeconomic status. Similarly, we were unable to inter
view the 15% of eligible cases who died before contact, potentially
limiting
correlate with increased levels of putative colon tumor promoters (e.g.,
fecal bile acids) or growth factors (e.g., insulin) or may provide energy to
neoplastic cells. Giovannucci (63) focused on hyperinsulinemia as a
tumor promoter because the main determinants of insulin resistance
(obesity, particularly central obesity, physical inactivity, and possibly a
low ratio of dietary polyunsaturated fat to saturated fat) and hyperinsu
linemia (a diet high in refined carbohydrates and low in water-soluble
fiber) may also be related to coborectal cancer. Our study brings some
support to Giovannucci's hypothesis because we found inverse associa
tions with physical activity, vegetable fiber (34), and ratio of dietary
polyunsaturated fat to saturated fat (31) and direct associations with
energy intake, obesity (although not central obesity), and a history of
diabetes.
risk of colon and rectal cancers (78). Alcohol may promote carcino
genesis by inducing metabolic activation of procarcinogens by cyto
chrome P450 enzymes by delaying DNA repair or by disrupting DNA
methylation.
Potential limitations of our study deserve consideration. A propor
tion of eligible subjects was not interviewed. It is possible that the
line
mutation in the 3' end of the CYPJAJ gene (77). This genetic
polymorphism has been related to an increased metabolic activation of
polycyclic aromatic hydrocarbons present in tobacco smoke.
Alcohol use has been found in a majority of past studies to increase
I . Parkin, D. M., Pisani, P., and Ferlay, J. Estimates of the worldwide incidence of
eighteen major cancers in 1985. Int. J. Cancer. 54: 594—606. 1993.
2. Haenszel, W. Cancer mortality among the foreign born in the United States. J. NatI.
Cancer Inst., 26: 37—132,1961.
3. La Marchand, L.. and Kolonel, L. N. Cancer among Japanese migrants to Hawaii:
gene-environment interactions. Rev. Epidemiol. Sante Publique. 40: 425—430,1992.
4. McMichael, A. J., and Giles, G. G. Cancer in migrants to Australia: extending the
descriptive
epidemiological
data. Cancer
Res., 48: 751—756. 1988.
5. Broder, S. Perspectives on cancer in Japan and the United States. Jpn. J. Cancer Res..
84: 821—830, 1993.
6. Garabrant, D. H., Peters, J. M., Mack, T. M., and Bernstein, L. Job activity and colon
cancer risk. Am. J. Epidemiol., 119: 1005—1014,
1984.
7. Vena, J. E., Graham, S., Zielezny. M., Swanson, M. K., Barnes, R. E., and Nolan. J.
Lifetime occupational exercise and colon cancer. Am. J. Epidemiol., 122: 357—365,1985.
8. Gerhardsson, M., Norell, S. E., Kiviranta, H., Pedersen, N. L., and Ahlbom, A.
Sedentary jobs and colon cancer. Am. J. Epidemiol., 123: 775—780,1986.
9. Wu, A. H., Paganini-Hill, A., Ross, R. K., and Henderson, B. E. Alcohol. physical
activity, and other risk factors for colorectal cancer: a prospective study. Br. J.
Cancer, 55: 687—694, 1987.
10. Gerhardsson de Verdier, M., Broderus, B., and Norell, S. E. Physical activity and
colon cancer risk. mt. J. Epidemiol., 17: 743—746,1988.
11. Slattery, M. L., Schumacher, M. C., Smith, K. R., West, D. W., and Abd-Elghani. N.
Physical activity, diet, and risk of colon cancer in Utah. Am. J. Epidemiol., 128:
989—999,1988.
12. Fredriksson, M., Bengisson, N. 0., Hardell, L., and Axelson, 0. Colon cancer,
physical activity, and occupational exposures. Cancer (Phila.). 63: 1838—1842, 1989.
13. Severson, R. K., Nomura, A. M. Y.. Grove, J. S., and Stemmermann, G. N. A prospective
analysis of physical activity and cancer. Am. J. Epidemiol., 130: 522—529,1989.
14. Slattery, M. L., Abd-Elghany, N., Kerber, R., and Schumacher, M. C. Physical
activity and colon cancer: a comparison of various indicators of physical activity to
evaluate the association. Epidemiology. 1: 481—485,1990.
15. Whittemore, A. S., Wu-Williams, A. H., Lee, M., Shu, Z., Gallagher.R. P.. Deng-ao.
J., Lun, Z., Xianghui, W., Kun, C.. Jung, D., The, C-Z., Chengde, L., Yao, X. J.,
Paffenbarger, R. S., Jr., and Henderson, B. E. Diet. physical activity, and colorectal
4793
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1997 American Association for Cancer Research.
WESTERNLIFESTYLEANDCOLORECTAL
CANCER
cancer among Chinese in North American and China. J. Nail. Cancer Inst., 82:
915—926,1990.
46. United States Department of Labor. Selected CHaracteristics
16. Gerhardsson de Verdier, M., Steineck, G., Hagman, U., Rieger, A., and Norell, S. E.
Physical activity and colon cancer: a case-referent study in Stockholm. mt. J. Cancer,
46: 985—989,1990.
17. Kato, I., Tominaga, S., and Ikari, A. A case-control study of male colorectal cancer
in Aichi Prefecture, Japan: with special reference to occupational activity level,
drinking habits and family history. Jpn. J. Cancer Res., 81: 115—121,1990.
18. Ballard-Barbash, R., Schatzkin, A., Albanes, D., Schiffman, M. H., Kreger, B. E.,
Kannel, W. B., Anderson, K. M., and Helsel, W. E. Physical activity and risk of large
bowel cancer in the Framingham Study. Cancer Res., 50: 3610—3613, 1990.
19. Brownson, R. C., Chang, J. C., Davis, J. R., and Smith, C. A. Physical activity on the
job and cancer in Missouri. Am. J. Public Health, 81: 639—642, 1991.
20. Lee, I. M., Paffenbarger, R. S., Jr., and Hsieh, C-c. Physical activity and risk of developing
colorectal cancer among college alumni. J. Nail. Cancer Inst., 83: 1324—1329,1991.
21. Marcus, P. M., Newcomb, P. A., and Storer, B. E. Early adulthood physical activity
and colon cancer risk among Wisconsin women. Cancer Epidemiol. Biomarkers
Prey., 3: 641—644,1994.
of Occupations (Phys
ical Demands, Working Conditions, Training Time). A Supplement to the Dictionary
of Occupational Titles, Ed. 3. Washington, DC: United States Government Printing
Office, 1966.
47. United States Department of Labor. Selected Characteristics of Occupations Defined
in the Dictionary of Occupational Titles. Washington, DC: United States Government
Printing Office, 1981.
48. Breslow, N. E., and Day, N. E. Statistical Methods for Cancer Epidemiology.
The
Analysis of Case-Control Studies, Vol. I. IARC Sci. Pubi. No. 32. Lyon, France:
IARC, 1980.
49. Giovannucci, E., Rimm, E. B., Stampfer, M. J., Colditz, G. A., Ascherio, A., Kearney,
J., and Willett, W. C. A prospective study of cigarette smoking and risk of colorectal
adenoma and colorectal cancer in U. S. men. J. Nail. Cancer Inst., 86: 183—191,
1994.
50. Willett, W. C., Stampfer, M. J., Colditz, G. A., Rosner, B. A., and Speizer, F. E.
Relation of meat fat and fiber intake to the risk of colon cancer in a prospective study
among women. N. Engl. J. Med., 323: 1664—1672,1990.
51. Bostick, R. M., Potter, J. D., Kushi, L. H., Sellers, T. A., Steinmetz, K. A., McKenzie,
D. R., Gapstur, S. M., and Folsom, A. R. Sugar, meat, and fat intake, and non-dietary
risk factors for colon cancer incidence in Iowa women (United States). Cancer Causes
22. Slattery, M. L., Potter, J., Cam, B., Edwards, S., Coates, A., Ma, K-N., and Berry, D.
Control, 5: 38—52,1994.
Energy balance and colon cancer: beyond physical activity. Cancer Res., 57: 75—80,
52. Goldbohm, R. A., van den Brandt, P. A., van't Veer, P., Brains, H. A. M., Dorant, E.,
1997.
23. Potter, J. D., and McMichael, A. J. Diet and cancer of the colon and rectum: a
case-control study. J. NatI. Cancer Inst., 76: 557—569,1986.
24. Lyon, J. L., Mahoney, A. W., West, D. \V., Gardner, J. W., Smith, K. R., Sorenson,
A. W., and Stanish, W. Energy intake: its relationship to colon cancer risk. J. Nati.
Cancer Inst., 78: 853—861, 1987.
25. Graham, S., Marshall, J., Haughey, B., Miuelman, A., Swanson, M., Zielezny, M.,
Sturmans, F., and Hermus, R. J. J. A prospective cohort study on the relation between
meat consumption and the risk of colon cancer. Cancer Res., 54: 718—723,1994.
53. Giovannucci, E., Rimm, E. B., Stampfer, M. J., Colditz, G. A., Ascherio, A., and
Willett, W. C. Intake of fat, meat, and fiber in relation to risk of colon cancer in men.
Cancer Res., 54: 2990—2997, 1994.
54. Nomura, A. M. Y., Heilbrun, L. K., and Stemmermann, G. N. Body mass index as a
Byers, T., Wilkinson, G., and West, D. Dietary epidemiology of cancer of the colon
in western New York. Am. J. Epidemiol., 128: 490—503,1988.
26. Gerhardsson de Verdier, M., Hagman, U., Steineck, G., Rieger, A., and Norell, S. E.
Diet, body mass and colorectal cancer: a case-referent study. list. J. Cancer, 46:
832—838.
1990.
27. Benito, E., Stiggelbout, A., Bosch, F. X., Obrador, A., Kaldor, J., Mulct, M., and
Munoz, N. Nutritional factors in colorectal cancer risk: a case-control study in
Majorca. tat. J. Cancer, 49: 161—167,
1991.
28. Peters, R. K., Pike, M. C., Garabrant, D., and Mack, T. M. Diet and colon cancer in
Los Angeles County, California. Cancer Causes Control, 3: 457—473,1992.
predictor of cancer in men. J. Natl. Cancer Inst., 74: 319—323, 1985.
55. Le Marchand, L., Wilkens, L. R., and Mi, M-P. Obesity in youth and middle age and
risk of colorectal cancer in men. Cancer Causes Control, 3: 249—254,1992.
56. Lew, E. A., and Garfinkel, L. Variations in mortality by weight among 750,000 men
and women. J. Chronic Dis., 32: 563—576,1979.
57. Phillips, R. L., and Snowdon, D. A. Dietary relationships with fatal colorectal cancer
among Seventh-day Adventists. J. Natl. Cancer Inst., 74: 307—317,1985.
58. Chute, C. G., Willett, W. C., Colditz, G. Q., Stampfer, M. J., Baron, J. A., Rosner, B.,
and Speizer, F. E. A prospective study of body mass, height, and smoking on the risk
of colorectal cancer in women. Cancer Causes Control, 2: 117—124,
1991.
29. Freudenheim, J. L., Graham, S., Marshall, J. R., Haughey, B. P., and Wilkinson, G.
59. Benito, E., Obrador, A., Stiggelbout, A., Bosch, F. X., Mulct, M., Numoz, N., and
A case-control study of diet and rectal cancer in western New York. Am. J. Epide
Kaldor, J. A population-based case-control study of colorectal cancer in Majorca. I.
Dietary factors. mt. J. Cancer, 45: 69—76,1990.
60. La Vecchia, C., Negri, E., Parazzini, F., Boyle, P., D'Avanzo, B., Levi, F., Gentile,
A., and Franceschi, S. Height and cancer risk in a network of case-control studies in
miol.,
131:612—624,
1990.
30. Slattery, M. L., Cam, B. J., Potter, J. D., Berry, T. D., Coates, A., Duncan, D., and
Edwards, S. L. Dietary energy sources and colon cancer risk. Am. J. Epidemiol., 3:
199—210,1997.
31. Le Marchand, L., Wilkens, L. R., Hankin, J. H., Kolonel, L. N., and Lyu, L-C. A
case-control study of diet and colorectal cancer in a multiethnic population in Hawaii
(United States): lipids and foods of animal origin. Cancer Causes Control, 8: 637—
Northern Italy. lot. J. Cancer, 45: 275—279,1990.
61. Albanes, D. Total calories, body weight and tumor incidence in mice. Cancer Res.,
47: 1987—1992,
1987.
62. Simon, H. B. The immunology of exercise: a brief review. J. Am. Med. Assoc., 252:
2735—2738,1984.
648, 1997.
32. Le Marchand, L. Anthropometry, body composition and cancer. in: M. Micozzi, and
63. Giovannucci, E. Insulin and colon cancer. Cancer Causes Control, 6: 164—179,1995.
T. Moon (eds.), Macronutrients: Investigating Their Role in Cancer, pp. 321—342. 64. O'Mara, B., A., Byers, T., and Schottenfeld, E. Diabetes mellitus and cancer risk: a
New York: Marcel Dekker, Inc., 1992.
multisite case-control study. J. Chronic Dis., 38: 435—441, 1985.
33. McKeown-Eyssen, G. Epidemiology of colorectal cancer revisited: are serum ni
glycerides and/or plasma glucose associated with risk? Cancer Epidemiol. Biomark
ers Prey., 3: 687—695, 1994.
65.
34. La Marchand,L., Hankin,J. H., wilkens, L. R., Kolonel,L N., Englyst,H. N., and Lyu,
L-C. The association of dietary fiber and colorectal cancer risk: an evaluation of fiber
components. micronutrientsand foods of plant origin. Epidemiology,in press, 1997.
35. Sallis, J. F., Haskell, W. L., Wood, P. D., Fortmann, S. P., Rogers, T., Blair, S. N.,
and Paffenbarger, R. S., Jr. Physical activity assessment methodology in the Five City
66. Ragozzino, M., Melton, L. J., Chu, C. P., and Palumbo, P. J. Subsequent cancer risk
in the incidence cohort of Rochester, Minnesota, residents with diabetes mellitus. J.
Project. Am. J. Epidemiol., 121: 91—106,1985.
36. Albanes, D., Conway, J. M., Taylor, P. R., Moe, P. W., and Judd, J. Validation and
comparison of eight physical activity questionnaires. Epidemiology, I: 65—71,1990.
68. Williams, J. C., Walsh, D. A., and Jackson, J. F. Colon carcinoma and diabetes
mellitus. Cancer (Phila.), 54: 3070—3071, 1984.
37. Hankin, J. H. A diet history method for research,clinical and community use. Twenty
third Lenna FrancesCooper MemorialLecture.J. Am. Diet. Assoc.,86: 868-875, 1986.
diseases and the risk of colorectal cancer. Eur. J. Cancer, 27: 582—586,1991.
70. Steenland, K., Nowlin, S., and Palu, S. Cancer incidence in the National Health and
Nutrition Survey I Follow-up data: diabetes, cholesterol, pulse, and physical activity.
Cancer Epidemiol. Biomarkers Prey., 4: 807—811, 1995.
Adamui,
H. 0.,
McLaughlin,
J., Ekbom,
A.,
Beme,
C.,
Silverman,
D.,
Hacker,
D.,
and
Persson, I. Cancer risk in patients with diabetes mellitus. Cancer Causes Control, 2:
307—314,1991.
Chronic Dis., 35: 13—19,1982.
67. Kessler, I. I. Mortality experience among diabetics. J. NaIL Cancer Inst., 44: 673—
686, 1970.
69. La Vecchia, C., D'Avanzo, B., Negri, E., and Franceschi, S. History of selected
38. Hankin, J. H., Wilkens, L. R., Kolonel, L. N., and Yoshizawa, C. N. Validation of a
quantitative diet history method in Hawaii. Am. J. Epidemiol., 133: 616—628, 1991.
39. United States Department of Agriculture. Nutrient Database for Standard Reference.
Release 10. Bethesda, MD: Human Nutrition Information Service, United States
71. Tran, T. T., Medline, A., and Bruce, W. R. Insulin promotion of colon tumors in rats.
Cancer Epidemiol. Biomarkers Prey., 5: 1013—1015,
1996.
Department of Agriculture, 1992.
40. Pennington, J. A. Bowes and Church's Food Values of Portions Commonly Used, Ed.
72. Jacobs, D., Blackburn, A., Higgins, M., Reed, D., Iso, H., McMillan, G., Neaton, J.,
Nelson, J., Potter, J., Rifkind, B., Roussouw, J., Shekelle, R., and Yusuf, S. Report of
16. Philadelphia: J. B. Lippincott, 1994.
41. Holland, B., Welch, A. A., Unwin, I. D., Buss, D. M., Paul, A. A., and Southgate,
the conference on low blood cholesterol mortality associations. Circulation, 86:
1046—1060, 1992.
D. A. T. McCance and Widdowson's the Composition of Foods, Ed. 5. Cambridge,
England: Royal Society of Chemistry,
1991.
42. Kagawa, R. Standard Tables of Food Composition in Japan. Tokyo, Japan: Women's
University, 1995.
43. Abdon, I. C., Rosario, I. F., Aguinaldo, A. R., and Lontoc, A. V. Food, and Nutrition
73. Winawer, S. J., Flehinger, B. J., Buchlater, J., Herbert, E., and Shike, M. Declining
serum cholesterol levels prior to diagnosis of colon cancer. A time-trend, case-control
study. J. Am. Med. Assoc., 263: 2083—2085, 1990.
74. Law, M. R., and Thompson, S. G. Low serum cholesterol and the risk of cancer: an
analysis of the published prospective studies. Cancer Causes Control, 2: 253—261,1991.
Research Institute, Department of Science and Technology. Food Composition Tables
Recommended for Use in the Philippines. 6th Revision, Handbook I, Vol. I. Manila,
75. Giovannucci, E., and Martinez, M. E. Tobacco, colorectal cancer and adenomas: a
review of the evidence. J. Nail. Cancer Inst., 88: 1717—1730,
1996.
Philippines: Food and Nutrition Research Institute, 1990.
76. Giovannucci, E., Rimm, E. B., Stampfer, M. J., Colditz, G. A., Ascherio, A., Kearney,
44. Ainswonh, B. E., Haskell, W. L, Leon, A. S., Jacobs, D. R., Jr., Montoye, H. J., Sallis,
J. F., and Paffenbarger,R. S., Jr. Compendium of physical activities: classificationof
energy costs of human physical activities. Med. Sci. Sports Exercise, 25: 71—80,1993.
45. United States Department of Labor, Employment Service, Bureau of Employment
Security. Estimates of Worker Trait Requirements for 4000 Jobs as Defined in the
Dictionary of Occupational Titles (An Alphabetical Index). Washington, DC: United
States Government Printing Office, 1956.
J., and Willett, W. C. A prospective study of cigarette smoking and risk of colorectal
adenoma and colorectal cancer in US men. J. Nail. Cancer Inst., 86: 183—191,
1994.
77. Sivaraman, L., Leatham, M. P., Yee, J., Wilkens, L. R., Lau, A. F., and Le Marchand,
L. CYPIAI genetic polymorphism and in situ colorectal cancer. Cancer Res., 54:
3692—3695, 1994.
78. Potter, J. D. Nutrition and colorectal cancer. Cancer Causes Control, 7: 127—146,
1996.
4794
Downloaded from cancerres.aacrjournals.org on July 31, 2017. © 1997 American Association for Cancer Research.
Associations of Sedentary Lifestyle, Obesity, Smoking, Alcohol
Use, and Diabetes with the Risk of Colorectal Cancer
Loïc Le Marchand, Lynne R. Wilkens, Laurence N. Kolonel, et al.
Cancer Res 1997;57:4787-4794.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/57/21/4787
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 July 31, 2017. © 1997 American Association for Cancer Research.