Full text of DOI:10.1207/S15327914nc391_7

Nutrition and Cancer
ISSN: 0163-5581 (Print) 1532-7914 (Online) Journal homepage: http://www.tandfonline.com/loi/hnuc20
Energy Intake, Body Mass Index, Physical Activity,
and the Colorectal Adenoma-Carcinoma Sequence
Marie-Christine Boutron-Ruault , Pierre Senesse , Sèverine Mèance , Claude
Belghiti & Jean Faivre
To cite this article: Marie-Christine Boutron-Ruault , Pierre Senesse , Sèverine Mèance ,
Claude Belghiti & Jean Faivre (2001) Energy Intake, Body Mass Index, Physical Activity, and the
Colorectal Adenoma-Carcinoma Sequence, Nutrition and Cancer, 39:1, 50-57, DOI: 10.1207/
S15327914nc391_7
To link to this article: http://dx.doi.org/10.1207/S15327914nc391_7
Published online: 18 Nov 2009.
Submit your article to this journal
Article views: 92
View related articles
Citing articles: 63 View citing articles
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=hnuc20
Download by: [Central Michigan University]
Date: 05 November 2015, At: 16:40

NUTRITION AND CANCER, 39(1), 50–57
Copyright © 2001, Lawrence Erlbaum Associates, Inc.
Energy Intake, Body Mass Index, Physical Activity, and the
Colorectal Adenoma-Carcinoma Sequence
Marie-Christine Boutron-Ruault, Pierre Senesse,
Séverine Méance, Claude Belghiti, and Jean Faivre
Abstract: Little is known about the precise relationship be-
tween energy intake, overweight, sedentary lifestyle, and
steps in the colorectal adenoma-carcinoma pathway. We
studied these parameters within a case-control study. Pa-
tients with adenomas <10 mm (n = 154) or >10 mm (n =
208) were compared with polyp-free controls (n = 426) for
determining factors associated with adenoma formation,
i.e., observed for small and large adenomas, or with
adenoma growth only. Colorectal cancer cases (n = 171)
were compared with population controls (n = 309) to deter-
mine factors specific to the final stage, cancer. Exercise re-
duced the risk of cancer [odds ratio (OR) = 0.3, 95%
confidence interval (CI) = 0.2–0.5 for high vs. low physical
activity] but had little influence on adenomas. High energy
intake increased the risk of cancer [OR for 5th vs. 1st
quintile (OR5) = 1.6, 95% CI = 0.9–2.9, p = 0.02], but not of
adenomas. High body mass index (BMI) significantly in-
creased the risk of large adenomas (OR5 = 2.1, 95% CI =
1.2–3.5, p = 0.02 and OR5 = 1.7, 95% CI = 1.0–3.1, p =
0.25) for large and small adenomas vs. polyp-free controls.
Neither height nor weight nor BMI influenced the risk of
cancer. Results were unmodified when controlling for di-
etary risk factors and family history. Energy intake, a seden-
tary 1ifestyle, and high BMI were independently associated
with a high risk of cancer itself or large adenomas, which in-
dicates an effect on promotion of colorectal tumors. These
findings suggest that preventive advice regarding these fac-
tors should be provided, even late in life, to decrease the risk
of colorectal cancer.
overweight could be explained by hyperinsulinism or insulin
resistance, which might promote colonic cancer cells (2,3).
It is admitted that, in Western countries, most CRCs arise
from the so-called adenoma-carcinoma pathway, i.e., small
adenoma formation, adenoma growth, and cancer. Thus far,
no case-control study and only two cohort studies (4,5) sepa-
rately investigated the influence of high energy intake, sed-
entary lifestyle, and overweight on the risk of small and
large adenomas separately. In view of preventive strategy
focused on the proper at-risk groups, we attempted to deter-
mine which step of the adenoma-carcinoma pathway was in-
fluenced by these risk factors by simultaneously studying
the effect of occupational as well as recreational physical ac-
tivity, BMI, and energy intake on small adenomas, large
adenomas, and cancers within a case-control study. Factors
associated with small and large adenomas were considered
to influence adenoma formation, factors associated with
only large adenomas were considered to influence adenoma
growth, and factors associated with cancer were considered
to influence malignant transformation.
Subjects and Methods
Cases and Controls
A several-group case-control study was set up in Bur-
gundy, France, to investigate environmental and familial
risk factors for colorectal tumors, following the adenoma-
carcinoma pathway. Details of the studied populations have
been presented elsewhere (6). Briefly, adenoma cases were
inhabitants of the Côte d’Or area (Burgundy, France), aged
30–79 years, with a colorectal adenoma diagnosed for the
first time in their lives. They were studied in two groups ac-
cording to the size of the largest adenoma: the small (<10
mm) adenoma group (85 men and 69 women) and the large
adenoma group (129 men and 79 women). Patients were re-
cruited through all private and public endoscopy clinics in
Introduction
The beneficial effect of physical activity on the risk of
colorectal cancer (CRC) is now well established, whereas
high levels of body mass index (BMI) and energy intake are
often associated with a high risk of CRC (1). It has been sug-
gested that the risk associated with a sedentary lifestyle and
M.-C. Boutron-Rualt, P. Senesse, C. Belghiti, and J. Faivre are affiliated with the Registre Bourguignon des cancers digestifs, Faculté de Médecine, 21033-
Dijon Cedex, France. M.-C. Boutron-Ruault and S. Méance are affiliated with the Institut Scientifique et Technique de la Nutrition et de l’Alimentation,
Conservatoire National des Arts et Métiers, 75003 Paris, France. P. Senesse is affiliated with the Centre Régional de Lutte Contre le Cancer, Val d’Aurelle
34298 Montpellier cedex 05, France.

the area. Polyp-free controls (182 men and 245 women)
were recruited after colonoscopy through the same proce-
dure as the cases and had never had any adenoma or any
other type of polyp. For polyp cases and polyp-free controls,
colonoscopy must have reached at least the sigmoid-to-
descending colon junction. When incomplete, it was com-
pleted in most cases by a double-contrast barium enema. The
hepatic flexure was reached in 64.4%, 66.2%, and 60.8% of
subjects, respectively. Histopathological examinations were
performed in the University hospital laboratory or a private
laboratory, where pathologists had common training ses-
sions to agree on the pathological definition of the different
types of polyps. Patients with only hyperplastic polyps were
excluded from the study, inasmuch as they have been dem-
onstrated to share some risk factors with adenomatous pol-
yps. Mean age was similar in the two adenoma groups (61.6
10.2 and 59.5 11.2 yr for large and small adenomas, re-
spectively) but older than in the polyp-free group (54.1
14.0 yr, p < 0.01). Cancer cases were 109 male and 62 fe-
male residents of the Côte d’Or area, aged 30–79 yr, with a
histologically proven adenocarcinoma of the large bowel di-
agnosed for the first time in their lives. Population controls
were 159 male and 150 female Côte d’Or residents who had
been randomly selected from the census list. Mean age for
cancer cases was 65.6 9.4 and 61.7 11.3 years for men
and women, respectively, and 62.9 11.1 and 61.2 12.0
years for male and female population controls, respectively.
Refusal rates were higher among population controls
(46.5%) than in the other groups, where refusal rates were
20.1% for cancers, 14.0% for large adenomas, 20.2% for
small adenomas, and 22.0% for polyp-free controls.
methods, a questionnaire by meals and a questionnaire by
list of foods, for assessing current dietary intakes (7). This
pilot study resulted in the elaboration of a two-hour detailed
qualitative and quantitative questionnaire about the past
year’s diet that followed the pattern of meals throughout the
day. Nutritional data were transformed into nutrients using a
compilation food composition table worked out for the
needs of the study, which included 318 simple foods and 188
mixed dishes. We studied energy as total energy intake and
as alcohol-free energy because alcohol had been demon-
strated to be a separate risk factor (6). BMI was calculated as
weight/height * height. Information about weight and height
was obtained from the subjects as the weight and height at
the time of the interview; data were missing in four subjects
with large adenomas. Quintiles of energy intake or BMI or,
when studying subgroups, tertiles were computed separately
by gender on the basis of the distribution in each control
group. Physical activity was recorded as absent (totally sed-
entary lifestyle), high (regular physical exercise as sports or
as an occupation requiring strenuous physical activity), or
moderate, separately for occupational and recreational activ-
ity. General physical activity was coded as high if both occu-
pational and recreational activities were high or, for retired
people, if physical activity was high, as moderate if all activ-
ities were at least moderate, and as mild in the other in-
stances. Information about physical activity was lacking in
18 population controls (5.8%), one cancer case (0.4%), two
large adenoma patients (1%), four small adenoma patients
(2.6%), and four polyp-free controls (0.9%). Distribution of
the main variables for each group is presented in Table 1.
Analysis
Data
Multiple logistic regression was performed by calculat-
ing odds ratios (ORs) and 95% confidence intervals (CIs),
controlling for age and gender, and taking as the reference
Before the case-control study was set up, a pilot study
was performed to compare and validate two diet history
Table 1. Distribution of the Study Population for Relevant Factorsa
Polyp-Free
Controls
(n = 427)
Small
Adenomas
(n = 154)
Large
Adenomas
(n = 208)
Population
Controls
(n = 309)
Cancer
(n = 171)
Age, yr
Gender
Male
54.1
14.1
59.5
11.2
61.6
10.2
62.1
11.6
64.2
10.3
42.6
55.2
62.0
51.5
63.7
Female
Tumor locationb
57.4
44.8
38.0
48.5
36.3
Proximal colon
Distal colon
Rectum
21.4
56.5
32.5
6.2 (10.1)
72.6 (77.4)
21.6 (25.5)
25.1
36.8
38.0
Physical activity
Low
Median
65.0
27.9
64.9
22.5
64.2
26.1
51.9
25.1
62.9
25.3
High
7.1
12.6
9.7
23.0
11.8
Total energy intake, kcal/day
Alcohol-free energy intake, kcal/day
BMI, kg/m2
2,370.8 862.1
2,247.6 786.0
2,330.2 770.8
2,186.8 693.0
2,510.2 847.9
2,288.0 724.8
2,251.9 770.7
2,072.7 638.36
2,531.3 848.6
2,305.3 736.9
24.1
4.1
25.0
4.6
25.3
4.3
24.7
3.7
25.2
3.9
a: Values are expressed as percentages, except for age, energy, and body mass index (BMI), which are means SD expressed in units as noted.
b: Location of large adenomas (location of any adenoma in parentheses).
Vol. 39, No. 1
51

the first quintile for energy intake and for BMI and the low-
est level for physical activity. The potential confounding
role of a family history of colorectal tumors and of dietary
risk or protective factors for colorectal tumors was tested in
the logistic model by adding these factors to the model. The
interaction between two variables was also tested in the lo-
gistic model by adding an interaction term in the model. The
statistical significance of each variable was tested by the
likelihood ratio test.
free energy). Total energy intake also had a stronger influ-
ence on left colon cancers [OR for 3rd vs. 1st tertile (OR3) =
2.1, 95% CI = 1.0–4.3, p = 0.04] than on right colon (OR3 =
1.5, 95% CI = 0.6–3.3, p = 0.37) or rectal cancers (OR3 =
1.7, 95% CI = 0.9–3.4, p = 0.10). Values were very similar to
those for alcohol-free energy. A first-degree family history
of CRC (FHCRC) had an influence on the relationship be-
tween cancer and total energy intake, and the interaction
term was statistically significant (p = 0.007). In the case of
an FHCRC (n = 27 cases), the ORs for energy intake were as
follows: OR2 = 2.3 (95% CI = 0.7–8.1) and OR3 = 4.4 (95%
CI = 1.4–13.9, p = 0.006). When there was no FHCRC (n =
144 cases), the corresponding ORs became 1.3 (95% CI =
0.8–2.1) and 1.5 (95% CI = 0.9–2.5, p = 0.09), respectively.
There was a significant interaction between the effects of
physical activity and total energy intake (p < 0.005) or alco-
hol-free energy intake (p < 0.01) on risk of CRC. Total
energy intake did not influence the risk of cancer when phys-
ical activity was high. OR2 and OR3 for energy intake were
0.8 (95% CI = 0.3–2.3) and 0.6 (95% CI = 0.2–2.0, p for
trend = 0.40), respectively, when physical activity was high.
For moderate physical activity, OR2 and OR3 became 1.1
(95% CI = 0.4–2.7) and 2.1 (95% CI = 0.9–4.6, p for trend =
0.05), respectively, and 1.8 (95% CI = 1.0–3.2) and 2.0
(95% CI = 1.2–3.7, p for trend = 0.01), respectively, when
physical activity was low. When considering alcohol-free
energy, the latter increased the risk of cancer only in subjects
with the lowest level of physical activity. OR2 and OR3 for
energy intake were 2.9 (95% CI = 1.4–5.9) and 3.4 (95% CI
= 1.7–6.9, p for trend = 0.0006), respectively, when physical
activity was low. For moderate physical activity, OR2 and
OR3 became 0.7 (95% CI = 0.3–2.0) and 1.1 (95% CI =
0.4–3.0, p for trend = 0.75), respectively. In the case of high
physical activity, OR2 and OR3 were 0.4 (95% CI = 0.1–1.3)
and 0.5 (95% CI = 0.1–1.8, p for trend = 0.86), respectively.
Results
Height, Weight, BMI, and Risk of Colorectal Tumors
The relationship between body size and risk of colorectal
tumors is presented in Table 2. Height was not associated
with the risk of small adenomas [OR for 5th vs. 1st quintile
(OR5) = 0.8, 95% CI = 0.4–1.5], large adenomas (OR5 = 1.1,
95% CI = 0.6–1.9), or cancer (OR5 = 0.9, 95% CI = 0.5–1.7).
Weight was mostly related to the risk of large adenomas
(OR5 = 2.3, 95% CI = 1.3–3.9, p = 0.01). High BMI was also
mostly associated with an increased risk of large adenomas
(OR5 = 2.1, 95% CI = 1.2–3.5, p = 0.02). This effect was
similar in men (OR5 = 2.1, 95% CI = 1.0–4.5, p = 0.08) and
women (OR5 = 1.9, 95% CI = 0.8–4.3, p = 0.16). It was
somewhat stronger for the left colon than for rectal
adenomas. When controlling for tobacco, alcohol, fiber, fat,
and folate, the effect of high BMI remained unchanged, with
the fifth quintile of BMI still being associated with a two-
fold-increased risk.
BMI had no significant influence on the risk of cancer in
men (OR5 = 0.9, 95% CI = 0.4–2.1, p for trend = 0.9) or
women (OR5 = 1.8, 95% CI = 0.6–5.3, p for trend = 0.8). The
stage at diagnosis did not influence this lack of association
between BMI and cancer.
Physical Activity and Risk of Colorectal Tumors
Energy Intake and Risk of Colorectal Tumors
The relationship between physical activity and colorectal
tumors is presented in Table 4. Physical activity had no sig-
nificant influence on the risk of small adenomas. OR3 values
for high physical activity were 0.8 (95% CI = 0.5–1.3, p =
0.23) for occupational, 1.8 (95% CI = 1.0–2.3, p = 0.19) for
recreational, and 1.3 (95% CI = 0.7–2.5, p = 0.95) for overall
physical activity. Regarding large adenomas, only occupa-
tional physical activity was significantly associated with a
decreased risk: OR3 = 0.6 (95% CI = 0.4–1.0, p = 0.04) for
occupational, 0.9 (95% CI = 0.5–1.6, p = 0.62) for recre-
ational, and 0.8 (95% CI = 0.4–1.5, p = 0.35) for overall
physical activity. This effect was greater in women (OR3 =
0.3, 95% CI = 0.09–0.8, p < 0.05) than in men (OR3 = 0.8,
95% CI = 0.4–1.4, p > 0.10).
The relationship between colorectal tumors and energy
intake is presented in Table 3. Energy intake had no signifi-
cant influence on the risk of colorectal adenomas whatever
their size: OR5 = 0.9 (95% CI = 0.5–1.7) for total energy and
OR5 = 0.7 (95% CI = 0.4–1.3) for alcohol-free energy for
small adenomas and OR5 = 1.2 (95% CI = 0.7–2.1) and OR5
= 1.0 (95% CI = 0.6–1.7) for total and alcohol-free energy,
respectively, for large adenomas.
High energy intake was significantly associated with the
risk of colorectal cancer: OR5 = 2.1 (95% CI = 1.1–4.0, p for
linear trend = 0.007) for total energy and OR5 = 1.6, (95% CI
= 0.9–2.9, p for linear trend = 0.02) for alcohol-free energy.
The effect of energy intake was stronger in women (OR5 =
3.1, 95% CI = 1.1–9.0, p = 0.01 for total energy and OR5 =
2.2, 95% CI = 0.8–6.0, p = 0.03 for alcohol-free energy) than
in men (OR5 = 1.8, 95% CI = 0.8–4.1, p = 0.12 for total en-
ergy and OR5 = 1.3, 95% CI = 0.6–2.9, p = 0.24 for alcohol-
Risk of cancer was inversely related to the level of physi-
cal activity: OR3 = 0.5 (95% CI = 0.3–0.9, p = 0.005) for
occupational, 0.3 (95% CI = 0.2–0.5, p < 0.0001) for recre-
ational, and 0.4 (95% CI = 0.2–0.6, p = 0.0003) for overall
52
Nutrition and Cancer 2001

Table 2. Relative Risk of Colorectal Adenomas and Cancer According to Body Size: Logistic Model Controlling With
Age, Categories Established by Gendera,b
Quintiles
P Value
1
2
3
4
5
for Trend
Height
Adenomas
Cutoff (males/females) cm
167/155
171/160
173/163
178/165
Small adenomas vs. polyp-free controls
n
OR
35/90
1.0
43/118
1.0
28/66
1.1
26/76
1.0
22/77
0.8
0.55
0.41
95% CI
0.6–1.7
0.6–2.0
0.5–1.8
0.4–1.5
Large adenomas vs. polyp-free controls
n
OR
38/90
1.0
53/118
1.2
41/66
1.5
46/76
1.6
30/77
1.1
95% CI
0.7–2.0
0.9–2.7
0.9–2.8
0.6–1.9
Cancer vs. population controls
Cutoff (males/females) cm
n
OR
166/155
29/61
0.7
169/159
34/66
0.8
172/162
34/67
0.8
176/165
26/46
0.9
48/69
1.0
0.77
95% CI
0.4–1.2
0.4–1.4
0.4–1.3
0.5–1.7
Weight
Adenomas
Cutoff (males/females) kg
63/52
71/56
76/60
83/70
Small adenomas vs. polyp-free controls
n
OR
27/94
1.0
35/89
1.4
24/73
1.2
38/97
1.4
30/74
1.4
0.32
0.01
95% CI
0.8–2.5
0.6–2.2
0.8–2.5
0.8–2.6
Large adenomas vs. polyp-free controls
n
OR
30/94
1.0
48/89
1.6
28/73
1.2
46/97
1.5
56/74
2.3*
95% CI
0.9–2.8
0.6–2.2
0.9–2.6
1.3–3.9
Cancer vs. population controls
Cutoff (males/females) kg
n
OR
65/53
36/60
1.1
71/57
26/63
0.8
76/60
38/62
1.2
84/68
35/58
1.1
36/66
1.0
0.64
95% CI
0.6–2.0
0.4–1.5
0.6–2.0
0.6–2.0
BMI
Adenomas
Cutoff (males/females) kg/m2
21.7/20.1
24.0/22.0
25.6/23.4
27.3/25.7
Small adenomas vs. polyp-free controls
n
OR
24/85
1.0
33/87
1.3
31/84
1.2
22/85
0.9
44/86
1.7
0.25
0.02
95% CI
0.7–2.5
0.7–2.3
0.5–1.7
1.0–3.1
Large adenomas vs. polyp-free controls
n
OR
30/85
1.0
40/87
1.3
34/84
1.1
34/85
1.0
66/86
2.1*
95% CI
0.7–2.2
0.6–2.0
0.6–1.8
1.2–3.5
Cancer vs. population controls
Cutoff (males/females) kg/m2
n
OR
22.9/20.3
45/61
24.4/22.6
23/62
25.9/23.9
40/61
28.7/26.1
34/62
29/63
1.0
1.7
0.8
1.4
1.1
0.92
95% CI
0.9–3.0
0.4–1.6
0.8–2.6
0.6–2.1
a: Abbreviations are as follows: n, cases/controls; OR, odds ratio; 95% CI, 95% confidence interval; BMI, body mass index.
b: Statistical significance is as follows: *, p < 0.01. Data were missing from 4 patients.
Vol. 39, No. 1
53

Table 3. Relative Risk of Colorectal Adenomas and Cancer According to Caloric Intake: Logistic Model Controlling
With Age, Categories Established by Gendera
Consumption Levels
P Value
for Trendd
1
2
3
4
5
Total calories
Adenomas
Cutoff (males/females), kcal/day
2,141/1,522
2,626/1,806
3,037/2,054
3,520/2,299
Small adenomas vs. polyp-free controls
n
OR
31/85
1.0
47/86
1.5
31/85
1.1
20/85
0.7
25/86
0.9
0.17
0.44
95% CI
0.9–2.7
0.6–2.0
0.4–1.4
0.5–1.7
Large adenomas vs. polyp-free controls
n
OR
42/85
1.0
42/86
1.1
46/85
1.3
40/85
1.3
38/86
1.2
95% CI
0.6–1.8
0.8–2.2
0.7–2.2
0.7–2.1
Cancer vs. population controls
Cutoff (males/females), kcal/d
n
ORb
2,053/1,437
31/62
2,393/1,712
30/62
2,747/1,900
42/61
3,158/2,299
44/62
24/62
1.0
1.3
1.4
2.0
2.1
0.007
0.005
95% CI
ORc
0.7–2.5
1.4
0.7–2.6
1.4
1.1–3.7
2.0
1.1–4.0
2.2
1.0
95% CI
0.7–2.6
0.7–2.7
1.1–3.8
1.2–4.2
Alcohol-free calories
Adenomas
Cutoff (males/females), kcal/day
1,933/1,494
2,389/1,764
2,753/2,010
3,289/2,255
Small adenomas vs. polyp-free controls
n
OR
36/85
1.0
45/86
1.3
26/85
0.8
25/86
0.8
22/85
0.7
0.08
0.69
95% CI
0.7–2.1
0.4–1.4
0.4–1.4
0.4–1.3
Large adenomas vs. polyp-free controls
n
OR
44/85
1.0
55/86
1.3
39/85
1.0
37/86
1.1
33/85
1.0
95% CI
0.8–2.2
0.6–1.8
0.6–1.8
0.6–1.7
Cancer vs. population controls
Cutoff (males/females), kcal/day
n
ORb
1,817/1,395
22/62
2,066/1,661
36/62
2,442/1,845
44/62
2,898/2,268
40/62
29/61
1.0
0.8
1.4
1.6
1.6
0.02
0.01
95% CI
ORc
0.4–1.5
0.7
0.7–2.5
1.4
0.9–3.0
1.5
0.9–2.9
1.7
1.0
95% CI
0.3–1.3
0.7–2.6
0.8–2.9
0.9–3.2
a: See Table 2 footnote for definition of abbreviations.
b: Control with age.
c: Control with age and physical activity.
d: Test for linear trend.
physical activity. Physical activity reduced the risk of cancer
in men (OR3 = 0.3, 95% CI = 0.1–0.5, p = 0.0001) and
women (OR3 = 0.2, 95% CI = 0.05–1.1, p = 0.03) and for all
cancer sites [OR3 = 0.09 (95% CI = 0.01–0.7) for the right
colon, OR3= 0.5 (95% CI = 0.2–1.0) for the left colon, and
OR3 = 0.4 (95% CI = 0.2–0.8) for rectal cancer, p = 0.02 for
each site]. This protective effect of physical activity was
identical when controlling for FHCRC and for intakes of al-
cohol, tobacco, fat, fiber, and folate (OR3 = 0.4, 95% CI =
0.2–0.7, p < 0.001 in all models). When controlling for en-
ergy intake, the effect of physical activity became slightly
stronger: OR2 = 0.7 (95% CI = 0.4–1.2), and OR3 = 0.3 (95%
CI = 0.1–0.5, p = 0.0001). Stratified analysis suggested
some interaction between physical activity and FHCRC: OR
= 0.8 (95% CI = 0.5–1.3) and 0.3 (95% CI = 0.2–0.5, p =
0.0001) in the absence of an FHCRC and 0.2 (95% CI =
0.07–0.8) and 0.2 (95% CI = 0.06–0.7, p = 0.002) in case of
an FHCRC, but the interaction term was not significant (p =
0.22).
Discussion
The study design enabled us to study patients known to
be at different steps of the adenoma-carcinoma sequence
with the same methodology and within the same geographic
area. By comparing patients with incident adenomas with
polyp-free controls selected in the same endoscopy units, we
54
Nutrition and Cancer 2001

Table 4. Relative Risk of Colorectal Adenomas and
Cancer According to Physical Activity: Logistic Model
Controlling With Age and Gendera,b
within the controls. Our study suggested that high BMI
increased the risk of large adenomas, whereas high energy
intake and sedentary lifestyle mostly influenced the risk of
cancer.
Physical Activity
A positive association between energy intake and risk of
CRC has already been largely documented in most case-
control studies, and a meta-analysis observed a significant
trend with increasing energy intake in men and women and
for all sites (9). Surprisingly, some cohort studies observed
no relationship (10) or even a protective one (11). Such dis-
crepancies between cohort and case-control studies have
been described for other risk factors related to CRC (6,12),
in particular, when the effect of the studied factor was differ-
ent on adenomas and on cancer, as may be the case here. Our
study further supports an association between high energy
intake and an increased risk of CRC and suggests that it spe-
cifically influences the last step of carcinogenesis, i.e.,
adenoma transformation to cancer.
P Value
Low
Medium
High
for Trend
Small adenomas vs. polyp-free controls
Occupational
n
OR
68/175
1.0
50/171
0.7
32/77
0.8
0.23
0.19
0.95
95% CI
Leisure
n
0.5–1.1
0.5–1.3
64/192
1.0
56/186
0.9
0.6–1.3
31/40
1.8*
1.0–3.2
OR
95% CI
Global
n
98/275
1.0
34/118
0.8
19/30
1.3
OR
95% CI
0.5–1.2
0.7–2.5
In a recent review of the literature, physical activity was
consistently associated with a reduced risk of CRC (1). It
was reported in 7 of 9 cohort studies and in 10 of 11 case-
control studies. In our study, physical activity decreased the
risk of CRC, the effect being maximal for right colon can-
cer. In a cohort study, the protective effect of physical activ-
ity was restricted to proximal colon cancer in men and
women (13). Our study also supports some protective effect
of leisure physical activity on ³10-mm adenomas, but not on
small adenomas, as was also observed in two large prospec-
tive studies (4,5). In our study, the effect of physical activity
was not confounded by the effect of energy intake, but there
was an interaction between the two factors. High energy in-
take had no significant impact on the risk of CRC when the
level of physical activity was high, and its major impact was
on subjects who had no physical activity, especially when
the exposure factor was alcohol-free energy. A large case-
control study also suggested such an interaction; the effect
of high energy intake on CRC was maximal for subjects with
low physical activity and excess weight (14). Therefore, the
most deleterious association is low physical activity associ-
ated with high energy intake, a combination that is likely to
induce hyperinsulinism, the hypothesized link between
colorectal tumors, obesity, sedentary lifestyle, and high en-
ergy intake (2,3). Caloric intake should be adapted to the
level of physical activity, but preventive advice should en-
courage physical activity rather than caloric restriction, inas-
much as reducing caloric intake will lead to reduce intake of
potentially protective micronutrients.
Large adenomas vs. polyp-free controls
Occupational
n
OR
95% CI
Leisure
n
91/175
1.0
75/171
0.8
40/77
0.6*
0.04
0.62
0.35
0.5–1.2
0.4–1.0
86/192
1.0
92/186
0.9
26/40
0.9
OR
95% CI
Global
n
OR
95% CI
0.6–1.4
0.5–1.6
133/275
1.0
54/118
0.8
20/30
0.8
0.6–1.3
0.4–1.5
Cancers vs. population controls
Occupational
n
OR
95% CI
Leisure
n
63/104
1.0
66/80
1.3
40/93
0.5*
0.005
<0.0001
0.0003
0.8–2.0
0.3–0.9
74/83
1.0
69/109
0.7
27/85
0.3‡
OR
95% CI
Global
n
OR
95% CI
0.4–1.1
0.2–0.5
107/151
1.0
43/73
0.8
20/67
0.3†
0.5–1.2
0.2–0.6
a: See Table 2 footnote for definition of abbreviations.
b: Statistical significance is as follows: *, p < 0.05; †, p < 0.01; ‡, p <
0.001.
studied factors related to the first step, i.e., adenoma forma-
tion. Factors observed to be associated with large, but not
small, adenomas were considered to be related to adenoma
growth. Because of the selection biases of subjects having
undergone endoscopy, the comparison of adenoma cases to
cancer cases did not appear suitable. As previously dis-
cussed by us and others (6,8), the comparison of cancer
cases with population controls analyzes mainly the final step
of carcinogenesis, i.e., transformation to cancer, because of
the presence of an unknown proportion of adenoma patients
Regarding body characteristics, height, weight, and BMI,
we did not find any association with CRC. Similarly, a num-
ber of case-control studies failed to find such an association,
and Potter (15), in a review of the literature, was unable to
establish a consistent relationship between body size and
cancer. This could be due to recent weight loss associated
with the cancer. However, CRC is rarely a wasting disease,
and we did not find any influence of BMI, even in patients
with Dukes A or B tumors. Cohort studies found an associa-
tion with CRC, but also with large adenomas (4,5). We sug-
Vol. 39, No. 1
55

gest that the effect of BMI on CRC is mainly exerted
through its effect on large adenomas, as also suggested by a
German case-control study (16).
weight, and exercising regularly may efficiently reduce the
risk of CRC, even in late-middle-aged people, inasmuch as all
these factors have an influence on the final steps of the ade-
noma-carcinoma sequence, either cancer itself or adenoma
growth, the step that occurs immediately before cancer.
Our study also suggested an interaction between energy
intake or physical activity and an FHCRC, where subjects
who had first-degree relatives with CRC were more sensi-
tive to high energy intake and low physical activity than
those without such a history. Because these data relied on
small numbers, the interaction term was not significant for
physical activity, but the smallest group, i.e., those with a
family history, still conveyed sufficient power for the ORs to
be significantly different from 1. Another case-control study
observed such an interaction (17) with energy intake. These
possible interactions with family history need to be further
investigated in terms of metabolic polymorphism.
Acknowledgments and Notes
The authors thank Brigitte Lieubray and Claude Grillet for performing
the interviews and Drs. Bataillon, Bedenne, Carli, Gambert, Garaudet,
Hillon, Jacquot, Klepping, Massart, Meny, Riot, Roy and Villand for ad-
vising their patients to participate. This study was supported by Institut
National de la Santé et de la Recherche Médicale Grant CRE 87-8011, the
Europe Against Cancer Program, and the Regional Council of Burgundy.
Address correspondence to Dr. Marie-Christine Boutron-Ruault, ISTNA,
CNAM, 2 rue Conté, 75003 Paris, France. Phone: 33.1.40.27.29.19. FAX:
33.1.40.27.28.26. E-mail: boutron@cnam.fr.
Some limitations of our data set have been discussed pre-
viously (6,18). One relates to the proportion of colonosco-
pies limited to the left colon. This could possibly reduce the
power of our study, but mainly for the comparison between
small adenomas and controls, inasmuch as some controls
could bear small proximal adenomas. However, analysis re-
stricted to cases and controls with a complete colonoscopy
conveyed the same negative findings for small adenomas.
Another limitation, which is common to all case-control
studies, is the retrospective assessment of risk factors. It can
be suggested, for example, that reduced physical activity in
subjects with cancer may be due to the onset of the disease.
However, this would more likely affect leisure physical ac-
tivity than occupational activity, and both were found to be
protective. The simplified evaluation of physical activity in
our questionnaire is another limitation due to a lengthy dietary
questionnaire. It did not enable us to determine which precise
level of energy expenditure was associated with a reduced
risk. Some misclassification of the level of physical activity
may also have occurred, but this can only reduce the true ef-
fect, which was already strong. Regarding energy intake in
cancer cases, one can argue that cancer patients may be over-
zealous in their reporting of exposure. However, this is less
likely to be the case with adenoma patients, and mean energy
intake was only slightly lower in the latter. Another limitation
of our study is the relatively small sample size. Therefore, CIs
are large when subgroups are studied, and the corresponding
results must be taken with care. However, misclassification
of BMI data is limited, thus limiting the potential problem of
power. Regarding energy intake, we took special care in
elaborating a questionnaire that appeared to be well adapted
to this aged French population by following the pattern of
meals throughout the day and including a high number of
food items. The problem of the high refusal rate among pop-
ulation controls is difficult to overcome and is common in
case-control studies using population controls. It could be
suggested that those who refused to participate were a less
healthy group, with a high energy intake and a sedentary
lifestyle. However, our findings are consistent with those of
the literature, which tends to rule out a major selection bias.
In terms of prevention strategies, our findings suggest
that advice aimed at reducing energy intake, controlling
Submitted 11 September 2000; accepted in final form 17 November 2000.
References
1. World Cancer Research Fund: Food, Nutrition and the Prevention of
Cancer: A Global Perspective. Menasha: Banta, 1997, pp 216–251.
2. McKeown-Eyssen, G: Epidemiology of colorectal cancer revised: are
serum triglycerides and/or plasma glucose associated with risk? Can-
cer Epidemiol Biomarkers Prev 3, 687–695, 1994.
3. Giovannucci, E: Insulin and colon cancer. Cancer Causes Control 6,
164–179, 1995.
4. Giovannucci, E, Ascherio, A, Rimm, EB, Colditz, G, Stampfer, MJ, et
al.: Physical activity, obesity, and risk for colon cancer and adenoma in
men. Ann Intern Med 122, 327–334, 1995.
5. Giovannucci, E, Colditz, GA, Stampfer, MJ, and Willett, WC: Physical
activity, obesity, and risk of colorectal adenoma in women (United
States). Cancer Causes Control 7, 253–263, 1996.
6. Boutron, MC, Faivre, J, Dop, MC, Quipourt, V, and Senesse, P: To-
bacco, alcohol and colorectal tumor: a multistep process. Am J Epi-
demiol 141, 1038–1046, 1995.
7. Boutron, MC, Faivre, J, Milan, C, Lorcerie, B, and Esteve, J: A com-
parison of two diet history questionnaires that measure usual food in-
take. Nutr Cancer 12, 83–91, 1989.
8. Terry, MB, and Neugut, AI: Cigarette smoking and the colorectal
adenoma-carcinoma sequence: a hypothesis to explain the paradox. Am
J Epidemiol 147, 903–910, 1998.
9. Howe, GR, Aronson, KJ, Benito, E, Castelleto, R, Cornee, J, et al.: The
relationship between dietary fat intake and risk of colorectal cancer:
evidence from the combined analysis of 13 case-control studies. Can-
cer Causes Control 8, 215–228, 1998.
10. Giovannucci, E, Rimm, EB, Stampfer, MJ, Colditz, GA, Ascherio, A,
et al.: Intake of fat, meat and fiber in relation to risk of colon cancer in
men. Cancer Res 54, 2390–2397, 1994.
11. Bostick, RM, Potter, JD, Kushi, LH, Sellers, TA, Steinmetz, KA, et al.:
Sugar, meat, and fat intake, and non-dietary risk factors for colon can-
cer incidence in Iowa women (United States). Cancer Causes Control
5, 38–52, 1994.
12. Giovannucci, E, Rimm, EB, Stampfer, MJ, Colditz, GA, Ascherio, A,
et al.: A prospective study of cigarette smoking and risk of colorectal
adenoma and colorectal cancer in US men. JNCI 86, 192–199, 1994.
13. Thune, I, and Lund, E: Physical activity and risk of colon cancer in
men and women. Br J Cancer 73, 1134–1140, 1996.
14. Slattery, ML, Potter, J, Caan, B, Edwards, S, Coates, A, et al.: Energy
balance and colon cancer—beyond physical activity. Cancer Res 57,
75–80, 1997.
56
Nutrition and Cancer 2001

15. Potter, J: Colon cancer—do the nutritional epidemiology, the gut phys-
iology and the molecular biology tell the same story? J Nutr 123,
418–422, 1993.
17. Slattery, ML, Caan, BJ, Potter, JD, Berry, TD, Coates, A, et al.: Di-
etary energy sources and colon cancer risk. Am J Epidemiol 145,
199–210, 1997.
16. Bayerdörffer, E, Mannes, GA, Ochsenkühn, T, Köpcke, W, Wiebecke,
B, et al.: Increased risk of “high risk” colorectal adenomas in over-
weight men. Gastroenterology 104, 137–144, 1993.
18. Boutron, MC, Faivre, J, Quipourt, V, Senesse, P, and Michiels, C:
Family history of colorectal tumors and implications for the adenoma-
carcinoma sequence: a case-control study. Gut 37, 830–834, 1995.
Vol. 39, No. 1
57