Download Dietary Flavonoid Intake and Breast Cancer Risk among Women on

American Journal of Epidemiology Advance Access published December 11, 2006
American Journal of Epidemiology
Copyright ª 2006 by the Johns Hopkins Bloomberg School of Public Health
All rights reserved; printed in U.S.A.
DOI: 10.1093/aje/kwk033
Original Contribution
Dietary Flavonoid Intake and Breast Cancer Risk among Women on Long Island
Brian N. Fink1, Susan E. Steck2, Mary S. Wolff3, Julie A. Britton3, Geoffrey C. Kabat4, Mia M.
Gaudet1, Page E. Abrahamson1, Paula Bell1, Jane C. Schroeder1, Susan L. Teitelbaum3, Alfred
I. Neugut5,6, and Marilie D. Gammon1
1
Department
Department
3
Department
4
Department
5
Department
6
Department
2
of
of
of
of
of
of
Epidemiology, School of Public Health, University of North Carolina, Chapel Hill, NC.
Nutrition, School of Public Health, University of North Carolina, Chapel Hill, NC.
Community and Preventive Medicine, Mt. Sinai School of Medicine, New York, NY.
Epidemiology and Population Health, Albert Einstein College of Medicine, Bronx, NY.
Epidemiology, Joseph L. Mailman School of Public Health, Columbia University, New York, NY.
Medicine, College of Physicians and Surgeons, Columbia University, New York, NY.
Received for publication February 9, 2006; accepted for publication July 26, 2006.
Flavonoids are found in a variety of foods and have anticarcinogenic properties in experimental models. Few
epidemiologic studies have examined whether flavonoid intake is associated with breast cancer in humans. In this
study, the authors investigated whether dietary flavonoid intake was associated with reduced risk of breast cancer
in a population-based sample of US women. They conducted a case-control study among women who resided in
Nassau and Suffolk counties on Long Island, New York. Cases and controls were interviewed about known and
suspected risk factors and asked to complete a food frequency questionnaire regarding their average intake in the
prior 12 months. A total of 1,434 breast cancer cases and 1,440 controls provided adequate responses. A decrease
in breast cancer risk was associated with flavonoid intake; the decrease was most pronounced among postmenopausal women for flavonols (odds ratio (OR) ¼ 0.54, 95% confidence interval (CI): 0.40, 0.73), flavones (OR ¼
0.61, 95% CI: 0.45, 0.83), flavan-3-ols (OR ¼ 0.74, 95% CI: 0.55, 0.99), and lignans (OR ¼ 0.69, 95% CI: 0.51,
0.94). The authors conclude that intake of flavonols, flavones, flavan-3-ols, and lignans is associated with reduced
risk of incident postmenopausal breast cancer among Long Island women. These results suggest that US women
can consume sufficient levels of flavonoids to benefit from their potential chemopreventive effects.
breast neoplasms; diet; flavonoids
Abbreviations: CI, confidence interval; ER, estrogen receptor; LIBCSP, Long Island Breast Cancer Study Project; OR, odds
ratio; PR, progesterone receptor; USDA, US Department of Agriculture.
influence breast cancer development (1, 8, 9, 13–18). Furthermore, dietary intake of certain flavonoids has been
reported to potentially protect humans from developing
certain types of cancer (19–22), including breast cancer
(23, 24).
Until very recently, epidemiologic research regarding
flavonoids and breast cancer development in women was
limited, primarily because of the difficulty in estimating
Flavonoids are a group of more than 4,000 polyphenolic
compounds that occur naturally in fruits, vegetables, and
beverages of plant origin (1, 2). In numerous laboratory
studies, flavonoids have demonstrated the ability to inhibit
aromatase activity and thus lower estrogen biosynthesis and
circulating estrogen levels (3–7), inhibit tumor cell proliferation (8, 9), and inhibit the formation of reactive oxygen
species (10–12), all of which are mechanisms thought to
Correspondence to Brian N. Fink, CB# 7435, McGavran-Greenberg Hall, Department of Epidemiology, School of Public Health, University of
North Carolina, Chapel Hill, NC 27599-7435 (e-mail: fi[email protected]).
1
2 Fink et al.
flavonoid intake. Previous hospital-based, case-control studies in Greece (23) and Italy (24) have had their respective
dietary data linked with two flavonoid databases from the US
Department of Agriculture (USDA) (25, 26). Reduced risks
of breast cancer were observed for intake of two classes of
flavonoids: flavones (23, 24) and flavonols (24). Whether
similar risk reductions are detectable among US women,
for whom intake of flavonoid-rich foods is traditionally
lower than in Mediterranean women, is unknown (27–29).
In this analysis, we investigated whether breast cancer
risk in a population-based case-control study of women in
the Long Island Breast Cancer Study Project (LIBCSP) was
reduced in relation to flavonoid intake.
MATERIALS AND METHODS
Participants
The LIBCSP was conducted on Long Island, New York, in
Nassau and Suffolk counties (26). Cases were Englishspeaking women with in-situ or invasive breast cancer newly
diagnosed between August 1, 1996, and July 31, 1997. Cases
were identified using a rapid reporting system developed
specifically for the study. Controls were randomly selected
through random digit dialing for persons under age 65 years
and through Health Care Financing Administration lists for
persons aged 65 years or older. Controls were frequencymatched to cases in 5-year age groups (30). The institutional
review boards of all participating institutions approved the
study protocol, and the individual women all signed informed consent forms.
In-person interviews were completed for 1,508 breast cancer cases (81.2 percent of eligible cases) and 1,556 controls
(62.8 percent of eligible controls). Reasons for nonparticipation included subject refusal, illness, cognitive impairment, inability to locate the subject, moving out of the
area, and death (26). The average length of time between
the reference date (date of diagnosis for cases and date of
identification for controls) and the interview date was 96
days for cases and 167 days for controls (30).
Exposure assessment
Women were administered a standardized questionnaire
and asked to report on a variety of known and suspected
breast cancer risk factors. Cases who signed a medical record release form at the interview had their medical records
reviewed for clinical and pathologic characteristics related
to the breast cancer diagnosis and treatment, including tumor estrogen receptor (ER) and progesterone receptor (PR)
status.
Cases and controls were asked to recall their diet history
in the previous 12 months, including assessment of frequency and portion size, with a modified version of the
Block food frequency questionnaire (31). A total of 1,481
cases (98.2 percent) and 1,518 controls (97.6 percent) completed this self-administered questionnaire. To facilitate
comparison of our results with those of other studies, 18
cases and 18 controls with daily energy intakes above or below three standard deviations of the log-transformed mean
(in kcal/day) were excluded from the analysis (32). An additional 29 cases and 60 controls were excluded because
their menopausal status was unknown. This resulted in a total
of 1,434 cases and 1,440 controls.
Assessment of dietary flavonoid intake
The content of total flavonoids and seven classes of flavonoids (flavonols, flavones, flavan-3-ols, flavanones, anthocyanidins, and isoflavones, as well as lignans) in foods and
beverages was estimated with a database created for use in
the LIBCSP (33). The LIBCSP database included both the
USDA Database for the Flavonoid Content of Selected
Foods (26) and the USDA–Iowa State University Database
on the Isoflavone Content of Selected Foods (25). Additional
sources (34–37) were utilized to include isoflavone content
provided by fruits, vegetables, nuts, and grains, which are important dietary contributors among US women (38). These
sources also provided content information for lignans, phytochemicals not included in the USDA databases but for
which laboratory evidence has demonstrated potential anticarcinogenic properties (1, 39–43).
Using this database, 50 items listed on the modified Block
food frequency questionnaire were found to contain measurable amounts of at least one flavonoid class or lignans. Individual foods and beverages were listed under each class
they contained, from the richest source to the smallest source
(top to bottom) (33). The richest sources of total flavonoids
include tea, including herb tea (111.41 mg of flavan-3-ols
per 100 g), cherries (116.31 mg of anthocyanidins per 100 g),
and grapefruit (54.50 mg of flavanones per 100 g) (26).
Statistical analysis
Odds ratios and 95 percent confidence intervals were estimated using unconditional logistic regression (44), including terms for energy intake (kcal/day) and age (in 5-year age
groups). Data on total flavonoids, lignans, and each phytochemical class were categorized into quintiles and deciles
based on the distribution of intake among controls, but both
categorizations produced similar results; thus, only the results for quintiles are reported here. Tests for trend were
conducted using the continuous values in mg/day.
Confounding was assessed using backward elimination
with multivariable models. Potential confounders included
menopausal status (pre- or postmenopausal at the reference
date), age at menarche, lifetime alcohol intake (g/day), cigarette smoking (current, former, never), family history of breast
cancer in a mother or sister, benign breast disease, average
physical activity level from menarche to the reference date
(hours/day), body mass index (weight (kg)/height (m)2) at the
reference date (date of interview), household income, education, parity, mammography use, oral contraceptive use, and
consumption of fruits, vegetables, and antioxidants in the previous 12 months. None of the potential confounders altered
the estimates of effect by more than 10 percent.
Effect modification was first examined through use of
stratified analysis and then by comparing log-likelihood statistics for regression models that included a multiplicative
interaction term with those without such a term (45). From
Flavonoid Intake and Breast Cancer Risk 3
TABLE 1. Mean intakes (mg/day) of flavonoids and lignans among cases and controls in the Long Island
Breast Cancer Study Project, 1996–1997
Total flavonoids
Flavonols
Flavones
Premenopausal
cases (n ¼ 457)
Premenopausal
controls (n ¼ 487)
All cases
(n ¼ 1,434)
211.12
212.19
217.82
0.94
0.53
0.10
10.09
10.11
9.81
0.96
0.35
0.02
p value*
p valuey
p valuez
0.14
0.14
0.13
0.91
0.44
0.002
Flavanones
25.60
27.13
31.29
0.44
0.54
0.91
Flavan-3-ols
161.00
161.06
162.73
0.99
0.41
0.12
3.16
3.03
3.15
0.75
0.64
0.23
Anthocyanidins
Isoflavones
5.50
5.03
0.78
0.37
0.38
0.40
Lignans
5.97
5.92
6.00
0.85
0.50
0.04
Postmenopausal
cases (n ¼ 977)
Postmenopausal
controls (n ¼ 953)
All controls
(n ¼ 1,440)
220.74
242.66
230.43
Flavonols
9.68
10.70
10.44
0.002
Flavones
0.13
0.15
0.15
0.0002
Flavanones
34.12
34.17
31.43
Flavan-3-ols
163.29
182.68
Anthocyanidins
3.14
Isoflavones
4.58
Lignans
6.01
Total flavonoids
p value§
0.02
p value{
p value#
0.02
0.23
0.0003
0.02
<0.0001
0.003
0.97
0.99
0.98
173.82
0.03
0.009
0.16
3.66
3.51
0.17
0.02
0.14
4.86
0.70
0.46
0.69
0.85
6.62
6.36
0.005
0.002
0.03
* t test comparing mean values among premenopausal women.
y Wilcoxon rank-sum test comparing median values among premenopausal women.
z t test comparing mean values among all cases and all controls.
§ t test comparing mean values among postmenopausal women.
{ Wilcoxon rank-sum test comparing median values among postmenopausal women.
# Wilcoxon rank-sum test comparing median values among all cases and all controls.
the covariates listed above, only menopausal status was
found to modify the association between flavonoid or lignan
intake and breast cancer risk. For the analyses, menopausespecific quintiles were created on the basis of the respective
intakes of pre- and postmenopausal control women.
Differences in risk estimates by the hormone receptor
status of case tumors were examined in stratified analyses.
ER-positive, PR-positive cases were considered as one
group and were compared with all other hormone receptor
types combined (ER-positive, PR-negative; ER-negative,
PR-positive; and ER-negative, PR-negative).
RESULTS
The distribution of flavonoid and lignan intakes among
the breast cancer cases and controls is presented in table 1.
Overall, postmenopausal cases consumed a smaller amount
of total flavonoids per day (mean ¼ 220.74 mg/day; median,
141.78 mg/day) than did postmenopausal controls (mean ¼
242.66 mg/day; median, 165.00 mg/day). However, among
premenopausal women, few or no case-control differences
were observed, as cases consumed a mean of 211.12 mg/day
(median, 143.14 mg/day) and controls consumed a mean of
212.19 mg/day (median, 137.12 mg/day). Flavan-3-ols were
the largest contributor to total intake and were most dispa-
rate between postmenopausal cases (mean ¼ 163.29 mg/
day) and postmenopausal controls (mean ¼ 182.68 mg/day).
As is shown in table 2, the odds ratios for breast cancer
were reduced in relation to intakes of flavones (for the highest quintile of intake versus the lowest, odds ratio (OR) ¼
0.73, 95 percent confidence interval (CI): 0.57, 0.93) and
flavonols (OR ¼ 0.75, 95 percent CI: 0.59, 0.95). When
the associations were stratified by menopausal status, breast
cancer risk was decreased among postmenopausal women
in relation to intake of all flavonoids except flavanones, anthocyanidins, and isoflavones. Odds ratios were reduced by
25 percent among postmenopausal women in the highest
fifth of intake of total flavonoids (OR ¼ 0.75, 95 percent
CI: 0.56, 1.01), by nearly 40 percent for flavones (OR ¼
0.61, 95 percent CI: 0.45, 0.83), by nearly 50 percent for
flavonols (OR ¼ 0.54, 95 percent CI: 0.40, 0.73), by 31
percent for lignans (OR ¼ 0.69, 95 percent CI: 0.51, 0.94),
and by 26 percent for flavan-3-ols (OR ¼ 0.74, 95 percent
CI: 0.55, 0.99). There was a significant decreasing trend
across quintiles for total flavonoids, flavonols, flavones,
and lignans. In contrast, among premenopausal women,
there was no evidence of a decreased risk of breast cancer
for any class of flavonoids or lignans.
When all potential confounders were included in the
model (age at menarche, lifetime alcohol intake, cigarette
smoking, family history of breast cancer in a mother or sister,
4 Fink et al.
TABLE 2. Age- and energy-adjusted odds ratios for the associations between menopausal status-specific flavonoid and lignan
intakes and breast cancer incidence in the Long Island Breast Cancer Study Project, 1996–1997
Premenopausal women (n ¼ 944)
Intake
(mg/day)
OR*,y
Postmenopausal women (n ¼ 1,930)
Intake
(mg/day)
95% CI*
0–34.5
34.6–84.5
84.6–199.5
199.6–343.0
343.1
p for trendz
1.00
1.20
1.29
1.46
1.12
0.95
0.79,
0.84,
0.96,
0.72,
1.84
1.97
2.22
1.74
0–51.8
51.8–119.1
119.2–253.3
253.4–377.2
377.3
0–3.7
3.8–6.0
6.1–10.2
10.3–15.1
15.2
p for trend
1.00
1.32
1.48
1.53
1.38
0.92
0.86,
0.97,
0.99,
0.88,
2.03
2.27
2.35
2.15
0–4.3
4.4–6.8
6.9–11.1
11.2–17.1
17.2
0–0.04
0.05–0.07
0.08–0.12
0.13–0.21
0.22
p for trend
1.00
0.94
1.29
1.07
1.07
0.94
0.62,
0.86,
0.70,
0.70,
1.43
1.84
1.63
1.65
0–0.04
0.05–0.08
0.09–0.14
0.15–0.21
0.22
0–3.1
3.2–10.8
10.9–24.5
24.6–40.3
40.4
p for trend
1.00
0.69
0.69
0.85
0.80
0.34
0.46,
0.46,
0.57,
0.53,
1.04
1.04
1.26
1.21
0–5.3
5.4–18.8
18.9–32.1
32.2–54.2
54.3
0–5.1
5.2–26.4
26.5–120.8
120.9–264.1
264.2
p for trend
1.00
1.22
1.32
1.52
1.21
0.87
0.80,
0.87,
1.00,
0.78,
1.87
2.01
2.30
1.86
0–7.6
7.7–54.0
54.1–192.0
192.1–277.9
278.0
0–0.04
0.05–0.56
0.57–1.60
1.61–4.19
4.20
p for trend
1.00
1.15
0.77
1.07
1.08
0.81
0.77,
0.50,
0.71,
0.71,
1.72
1.17
1.61
1.63
0–0.03
0.04–0.56
0.57–1.84
1.85–4.84
4.85
0–0.31
0.32–1.10
1.11–3.17
3.18–7.62
7.63
p for trend
1.00
1.03
0.98
0.88
1.14
0.56
0.68,
0.65,
0.58,
0.76,
1.56
1.47
1.33
1.72
0–0.27
0.28–0.62
0.63–1.94
1.95–7.63
7.64
0–2.0
2.1–4.0
4.1–5.4
5.5–9.3
9.4
p for trend
1.00
1.43
0.98
1.62
1.24
0.72
0.95,
0.63,
1.07,
0.81,
2.17
1.51
2.45
1.92
0–2.4
2.5–4.2
4.3–6.4
6.5–10.2
10.3
* OR, odds ratio; CI, confidence interval.
y Adjusted for age (years) and energy intake (kcal/day).
z p for trend for continuous variable.
ORy
Total flavonoids
1.00
0.94
0.79
0.80
0.75
0.05
Flavonols
1.00
0.56
0.62
0.63
0.54
<0.001
Flavones
1.00
0.90
0.95
0.70
0.61
<0.001
Flavanones
1.00
1.09
1.10
1.08
1.00
0.87
Flavan-3-ols
1.00
0.94
0.80
0.82
0.74
0.06
Anthocyanidins
1.00
1.09
0.97
0.82
0.85
0.23
Isoflavones
1.00
0.97
1.16
1.14
1.02
0.72
Lignans
1.00
1.07
0.82
0.79
0.69
0.01
All women (n ¼ 2,874)
95% CI
Intake
(mg/day)
ORy
95% CI
0.71,
0.60,
0.60,
0.56,
1.24
1.05
1.06
1.01
0–44.6
44.7–101.2
101.3–230.2
230.3–364.7
364.8
1.00
1.01
0.98
0.95
0.88
0.14
0.80,
0.78,
0.76,
0.69,
1.27
1.23
1.20
1.12
0.42,
0.47,
0.47,
0.40,
0.74
0.82
0.83
0.73
0–4.0
4.1–6.4
6.5–10.7
10.8–16.2
16.3
1.00
0.80
0.83
0.87
0.75
0.05
0.64,
0.66,
0.69,
0.59,
1.01
1.04
1.09
0.95
0.68,
0.72,
0.52,
0.45,
1.19
1.26
0.94
0.83
0–0.05
0.06–0.09
0.10–0.14
0.15–0.22
0.23
1.00
0.94
0.99
0.83
0.73
0.004
0.75,
0.79,
0.66,
0.57,
1.18
1.25
1.05
0.93
0.82,
0.83,
0.81,
0.75,
1.46
1.46
1.43
1.34
0–4.5
4.6–15.2
15.3–30.0
30.1–50.3
50.4
1.00
0.90
0.93
0.99
0.89
0.64
0.71,
0.74,
0.79,
0.70,
1.13
1.17
1.25
1.12
0.72,
0.60,
0.62,
0.55,
1.24
1.06
1.08
0.99
0–6.5
6.6–39.5
39.6–189.8
189.9–267.9
268.0
1.00
0.96
0.99
1.00
0.85
0.17
0.76,
0.79,
0.80,
0.67,
1.21
1.24
1.26
1.08
0.83,
0.73,
0.62,
0.64,
1.44
1.28
1.09
1.14
0–0.04
0.05–0.56
0.57–1.75
1.76–4.57
4.58
1.00
1.10
0.95
0.87
0.91
0.27
0.88,
0.76,
0.69,
0.72,
1.38
1.19
1.10
1.15
0.72,
0.87,
0.85,
0.76,
1.30
1.55
1.53
1.38
0–0.17
0.18–0.26
0.27–0.38
0.39–0.61
0.62
1.00
0.86
1.00
0.96
0.95
0.31
0.68,
0.79,
0.75,
0.74,
1.09
1.26
1.22
1.22
0.81,
0.61,
0.59,
0.51,
1.40
1.09
1.05
0.94
0–2.3
2.4–4.2
4.3–6.2
6.3–9.8
9.9
1.00
1.13
0.86
0.97
0.82
0.06
0.90,
0.68,
0.77,
0.64,
1.42
1.09
1.23
1.04
Flavonoid Intake and Breast Cancer Risk 5
benign breast disease, average physical activity level from
menarche to the reference date, body mass index at the reference date, household income, education, parity, mammography use, oral contraceptive use, and consumption of fruits,
vegetables, and antioxidants in the previous 12 months), similar odds ratios were observed among postmenopausal women
for the highest fifth of intake of total flavonoids (OR ¼ 0.69,
95 percent CI: 0.49, 0.98), flavonols (OR ¼ 0.48, 95 percent
CI 0.34, 0.68), flavones (OR ¼ 0.59, 95 percent CI: 0.41,
0.84), flavan-3-ols (OR ¼ 0.63, 95 percent CI: 0.45, 0.88),
and lignans (OR ¼ 0.62, 95 percent CI: 0.43, 0.87).
When results were stratified by ER/PR status, there was
little or no heterogeneity in breast cancer risk in relation to
flavonoid intake for postmenopausal women. A consistent
trend towards a reduced risk was found for all hormone
receptor types in relation to flavonols, flavones, and total
flavonoids (table 3). The number of premenopausal women
in the study limited our ability to stratify the results by
hormone receptor status in these younger women.
DISCUSSION
In this analysis, inverse associations with breast cancer
risk were found for intake of total flavonoids and for most
flavonoid classes. The associations were most evident in
postmenopausal women. These results are consistent with
those of two previous hospital-based case-control studies
conducted in Greece (21) and Italy (24) that found a slightly
more modest reduction in risk with increasing flavones (23,
24) and flavonols (24). Both studies used the same two
USDA databases (25, 26) to measure flavonoid intake, and
each had a study population size similar to ours. Our enhancement of these US databases to more fully capture intake of flavonoid-rich foods may have improved our ability
to detect a stronger association between flavonoid intake
and breast cancer risk in our US population.
Among postmenopausal women with ER-positive, PRpositive tumors, we observed a reduced risk of breast cancer
for increasing intakes of flavones and flavonols. The ERpositive, PR-positive tumor receptor type is the most common type diagnosed among breast cancer patients in the
United States (46, 47). Thus, if replicated in other studies,
our findings may be of public health significance. However,
because our study population included a limited number of
women diagnosed with tumors of other receptor types, results
from our subgroup analyses should be interpreted with care.
In contrast, our data do not support an inverse association
between isoflavones and breast cancer risk. Previous studies
have also not observed an association (48–53), including
a study conducted in a multiethnic population in the San
Francisco Bay Area (53). The diet history instrument used
in the LIBCSP was limited in its coverage of soy products,
which may have resulted in a slight underestimation of intake of isoflavone-rich foods. This nondifferential misclassification may have resulted in masking of any potential
beneficial effects of these compounds.
While previous efforts have been made to estimate the
isoflavone content of soy-based products in the United
States (54, 55), the continued introduction of new soy products to the market, as well as nontraditional uses of soy such
as inclusion of soy flour in doughnuts and soy protein in
fast-food hamburgers, demonstrates the need for their inclusion in future dietary assessment tools. Furthermore,
the Block food frequency questionnaire did not include
questions on blueberries and raspberries, both rich sources
of anthocyanidins (26). Omission of these berries may have
contributed to the lower anthocyanidin intake reported in
our study as compared with the studies in Greece (23) and
Italy (24), although neither of those studies found a risk
reduction with anthocyanidins.
Additionally, flavonoid content in foods is variable, influenced in part by environmental conditions (56). In fruits and
vegetables, particularly, flavonoid content varies because
of differences in cultivars, cultural practices, climatic conditions, geographic location, degree of ripeness, storage
conditions, and industrial processing (57–61). Thus, it is possible that there were differences in flavonoid and lignan
content in the products consumed by the Long Island study
population as compared with the products from which estimates were taken and used to create the databases. However,
it is unknown how large or small these differences were
and where many of the products, especially fruits and vegetables, were grown or produced. However, this source of
variation is common to all studies that rely on nutrient databases to estimate dietary consumption (62).
Our study did, however, expand the coverage of phytochemicals in comparison with previous studies of flavonoids
and breast cancer risk (23, 24) by including lignans, which
are typically found in the woody portions of plants, the coats
of seeds, and the bran layer of grains (34, 63). Lignans are
thought to act through the same mechanisms as other flavonoids in preventing breast cancer (63). Our finding of a reduced breast cancer risk for increasing lignan intake among
women on Long Island supports other reported data on animals (64–67) and humans (1, 39, 43, 68–71).
Given the hypothesized anticarcinogenic effects of flavonoids and lignans (1, 8–18, 72), consumption would be
expected to benefit both premenopausal women and postmenopausal women. However, the biologic mechanism for
the effect modification by menopausal status observed in our
data is unclear. A previous study of fruit, vegetable, and
micronutrient intake in the LIBCSP (32) found a decreased
risk of breast cancer among postmenopausal women with
increasing levels of intake of vegetables and many micronutrients, including alpha- and beta-carotene. Our findings
suggest that the impact of flavonoids and lignans may also
be greater in postmenopausal women. It is possible that the
antiestrogenic properties of some flavonoid classes or lignans are only effective in the low-endogenous-estrogen
environment observed in postmenopausal women and are
ineffective in the high-endogenous-estrogen environment
of premenopausal women. Further research based on large
numbers of both premenopausal women and postmenopausal women is needed to help clarify this issue.
Flavonoids and lignans are found in numerous products,
including fruits and vegetables; thus, flavonoid and lignan
consumption may reflect part of an overall healthy diet and
lifestyle (73, 74). Furthermore, many lifestyle factors that
may potentially confound the relation between flavonoids or
lignans and breast cancer are highly correlated with fruit and
6 Fink et al.
TABLE 3. Age- and energy-adjusted odds ratios for the associations between flavonoid and lignan
intakes and breast cancer incidence among postmenopausal women in the Long Island Breast Cancer
Study Project, by tumor hormone receptor status, 1996–1997
Hormone receptor status
Variable and
intake (mg/day)
No. of controls
(n ¼ 953)
ER*-positive, PR*-positive
No. of cases
(n ¼ 378)
OR*,z
95% CI*
All othersy
No. of cases
(n ¼ 274)
ORz
95% CI
Total flavonoids
0–51.7
190
89
1.00
72
1.00
51.8–119.0
192
78
0.90
0.62, 1.31
62
0.89
0.59, 1.32
119.1–253.2
190
72
0.83
0.57, 1.21
42
0.60
0.39, 0.93
253.3–377.1
191
77
0.86
0.59, 1.25
49
0.68
0.44, 1.04
377.2
190
62
0.75
0.50, 1.12
49
0.72
0.47, 1.11
p for trend§
0.35
0.09
Flavonols
0–4.2
191
113
1.00
93
1.00
4.3–6.7
190
66
0.59
0.41, 0.86
47
0.51
0.34, 0.78
6.8–11.0
190
67
0.60
0.41, 0.87
42
0.46
0.30, 0.71
11.1–17.0
191
74
0.66
0.46, 0.96
46
0.49
0.32, 0.75
17.1
191
58
0.55
0.37, 0.82
56
0.51
0.33, 0.79
p for trend
0.12
0.03
Flavones
0–0.04
191
101
1.00
69
1.00
0.05–0.08
190
74
0.76
0.52, 1.10
72
1.05
0.71, 1.56
0.09–0.14
191
85
0.89
0.62, 1.29
55
0.83
0.54, 1.26
0.15–0.21
191
61
0.64
0.43, 0.94
44
0.67
0.43, 1.04
0.22
190
57
0.59
0.40, 0.89
34
0.51
0.32, 0.82
p for trend
0.02
0.003
Flavanones
0–5.3
190
70
1.00
50
1.00
5.4–18.8
192
77
1.18
0.80, 1.74
52
1.05
0.68, 1.63
18.9–32.1
189
84
1.21
0.83, 1.78
61
1.20
0.78, 1.84
32.2–54.2
191
80
1.14
0.77, 1.67
58
1.16
0.75, 1.80
54.3
191
67
0.95
0.63, 1.42
53
1.06
0.68, 1.66
p for trend
0.77
0.49
Table continues
vegetable intake (32, 75, 76), making it difficult to firmly
establish their independent effects (77). The main LIBCSP
questionnaire extensively assessed exposures throughout
the life course, including recreational physical activity levels from menarche to age at diagnosis, lifetime active and
passive smoking, and lifetime alcohol consumption (32).
However, when we controlled for these factors mutually
or individually, our results were not substantially altered.
Our study relied on retrospective reporting of dietary intake, which is subject to error, particularly in the reporting
by breast cancer cases as compared with controls. The controls may have overreported consumption of flavonoid-rich
foods, such as fruits and vegetables, in an attempt to appear
socially correct. This would have overemphasized the benefits of flavonoids or lignans. However, if this had actually
occurred, it would be expected that overreporting would
occur among all control women, regardless of menopausal
status. The lack of an association in premenopausal women
argues against this possibility. An additional concern with
errors in recall is that case reports of food intake may be
affected by whether patients have initiated chemotherapy
by the time of the study interview (78). However, in the
LIBCSP, most of the case women were interviewed prior
to any chemotherapy, and among those who had started
chemotherapy, reported intake levels were not found to differ from those of women who had not started (32).
There is a possibility that these results are due to errors in
reporting among the postmenopausal controls only. Daily
total flavonoid intake was higher among postmenopausal
controls than among premenopausal controls or among cases
Flavonoid Intake and Breast Cancer Risk 7
TABLE 3. Continued
Hormone receptor status
Variable and
intake (mg/day)
No. of controls
(n ¼ 953)
ER*-positive,
ER-positive, PR*-positive
PR-positive
No. of cases
(n ¼ 378)
OR*,z
ORz
95% CI*
All othersy
No. of cases
(n ¼ 274)
ORz
95% CI
Flavan-3-ols
0–7.6
190
93
1.00
71
1.00
7.7–54.0
192
75
0.81
0.56, 1.18
62
0.88
54.1–192.0
189
65
0.71
0.49, 1.04
48
0.69
0.45, 1.05
192.1–277.9
192
81
0.86
0.60, 1.24
48
0.66
0.43, 1.02
278.0
190
64
0.75
0.51, 1.10
45
0.67
0.43, 1.03
p for trend
0.45
0.59, 1.31
0.13
Anthocyanidins
0–0.03
189
77
1.00
58
1.00
0.04–0.56
192
88
1.19
0.82, 1.73
70
1.22
0.81, 1.84
0.57–1.84
190
88
1.25
0.86, 1.81
44
0.80
0.51, 1.24
1.85–4.84
191
69
0.93
0.63, 1.37
50
0.90
0.59, 1.39
4.85
191
56
0.77
0.51, 1.16
52
0.95
0.62, 1.46
p for trend
0.005
0.63
Isoflavones
0–0.27
190
82
1.00
55
1.00
0.28–0.62
191
63
0.78
0.52, 1.17
64
1.19
0.78, 1.83
0.63–1.94
191
81
1.09
0.75, 1.60
50
0.98
0.62, 1.52
1.95–7.63
191
88
1.21
0.83, 1.77
58
1.16
0.75, 1.80
7.64
190
64
0.92
0.61, 1.37
47
1.00
0.63, 1.58
p for trend
0.91
0.48
Lignans
0–2.4
215
89
1.00
72
1.00
2.5–4.2
167
81
1.09
0.76, 1.58
68
1.07
0.72, 1.59
4.3–6.4
190
76
0.96
0.66, 1.40
48
0.74
0.48, 1.14
6.5–10.2
191
71
0.89
0.60, 1.30
43
0.64
0.41, 0.99
10.3
190
61
0.82
0.55, 1.24
43
0.67
0.43, 1.05
p for trend
0.35
0.02
* ER, estrogen receptor; PR, progesterone receptor; OR, odds ratio; CI, confidence interval.
y ER-positive, PR-negative; ER-negative, PR-positive; and ER-negative, PR-negative.
z Adjusted for age (years) and energy intake (kcal/day).
§ p for trend for continuous variable.
(regardless of cases’ menopausal status). The LIBCSP was
a federally mandated study (30), and thus it garnered considerable media attention. Therefore, it is possible that some
of the control women in this study may have recently chosen
to adopt healthier dietary habits as a result of the publicity
surrounding the study. However, if dietary changes were
made in response to media attention, one would expect that
median intake levels would be higher among all controls, not
just postmenopausal controls. Furthermore, the LIBCSP was
primarily undertaken to examine whether breast cancer was
influenced by environmental factors, such as pesticides (30).
Media attention during the data collection phase was not
focused on potential dietary causes of breast cancer. Thus,
whether there was systematic misclassification of exposure
among only the postmenopausal control women is unclear,
although there is no strong evidence to suggest that this
differential error occurred. Nevertheless, our findings should
be interpreted cautiously until confirmation is provided from
studies conducted in other areas of the country.
As with all epidemiologic studies of cancer that rely on
food frequency questionnaires to estimate diet history, it is
unknown whether reported recent diet accurately reflected
the dietary patterns prevailing during the time periods most
relevant to cancer development. Because researchers rarely
have accurate information on subjects’ lifetime dietary intake patterns, it is unclear whether our findings are overestimates or underestimates of the true association between
flavonoids and lignans and breast cancer development.
8 Fink et al.
To our knowledge, very few studies have addressed the
impact of dietary flavonoid intake on the risk of breast cancer, particularly in a US population. Only the studies conducted in Greece (23) and Italy (24) have used the USDA
Database for the Flavonoid Content of Selected Foods (26)
and the USDA–Iowa State University Database on the Isoflavone Content of Selected Foods (25) together. We combined these two databases with additional literature (34–38)
to construct a more comprehensive instrument for assessing flavonoid intake, including intake of another phytochemical, lignans. Overall, the mean difference in total flavonoid
intake between postmenopausal cases and controls was
roughly equivalent to half a glass of tea or one apple per
day. However, these absolute differences must be interpreted
with care, given that the intake estimates were based on data
collected as part of a food frequency questionnaire; the food
frequency questionnaire is designed to rank relative intakes
rather than identify the absolute amounts of food that must
be consumed in order to influence disease outcomes.
This study had the advantage of a large sample size with
a population-based design, reducing selection bias and allowing for greater generalizability of results as compared
with hospital-based studies. The LIBCSP population consumed a wide variety and significant amounts of flavonoidand lignan-containing products, such as fruits, vegetables,
and tea, enabling us to address our specific aims. Furthermore, the Block food frequency questionnaire utilized in
this study is a validated, reliable dietary assessment tool
for estimating usual food-group intake and for ranking individuals in categories of intake (78–80).
In summary, this case-control study provides evidence
that increased intake of lignans and flavonoids is associated
with reduced risk of breast cancer in a population-based
sample of US women. When the results were stratified by
menopausal status, the inverse associations were restricted
to postmenopausal women and were particularly strong for
flavones, flavonols, flavan-3-ols, and lignans. These findings
support the similar but more modest reductions observed
in Greece and Italy (23, 24). Most research to date has not
utilized the two recently available USDA databases together,
with additional sources, to study flavonoids and breast cancer incidence. Further research using these instruments needs
to be conducted, particularly in US populations.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
ACKNOWLEDGMENTS
This work was supported in part by grants from the
National Cancer Institute and the National Institute of
Environmental Health Sciences (grants UO1CA/ES66572,
UO1CA66572, CA52283, P30ES10126, and 5T32CA
009330-25) and from the Lance Armstrong Foundation.
Conflict of interest: none declared.
20.
21.
22.
23.
REFERENCES
1. Dai Q, Franke AA, Jin F, et al. Urinary excretion of phytoestrogens and risk of breast cancer among Chinese women in
24.
Shanghai. Cancer Epidemiol Biomarkers Prev 2002;11:
815–21.
Ren W, Qiao Z, Wang H, et al. Flavonoids: promising anticancer agents. Med Res Rev 2003;23:519–34.
Le Bail JC, Laroche T, Marre-Fournier F, et al. Aromatase
and 17beta-hydroxysteroid dehydrogenase inhibition by flavonoids. Cancer Lett 1998;133:101–6.
Adlercrutz H, Bannwart C, Wahala K, et al. Inhibition of
human aromatase by mammalian lignans and isoflavonoid
phytoestrogens. J Steroid Biochem 1993;44:147–53.
Kellis JT Jr, Vickery LE. Inhibition of human estrogen synthetase (aromatase) by flavones. Science 1984;225:1032–4.
Kao YC, Zhou C, Sherman M, et al. Molecular basis of the
inhibition of human aromatase (estrogen synthetase) by flavone and isoflavone phytoestrogens: a site-directed mutagenesis study. Environ Health Perspect 1998;106:85–92.
Ibrahim AR, Abul-Hajj YJ. Aromatase inhibition by flavonoids. J Steroid Biochem Mol Biol 1990;37:257–60.
Chan WS, Wen PC, Chiang HC. Structure-activity relationship of caffeic acid analogues on xanthine oxidase inhibition.
Anticancer Res 1995;15:703–7.
Chang WS, Lee YJ, Lu FJ, et al. Inhibitory effects of flavonoids on xanthine oxidase. Anticancer Res 1993;13:2165–70.
Le Marchand L, Murphy SP, Hankin JH, et al. Intake of flavonoids and lung cancer. J Natl Cancer Inst 2000;92:154–60.
Lahiri-Chatterjee M, Katiyar SK, Mohan RR, et al. A flavonoid antioxidant, silymarin, affords exceptionally high protection against tumor promotion in the SENCAR mouse skin
tumorigenesis model. Cancer Res 1999;59:622–32.
Tsyrlov IB, Mikhailenko VM, Gelboin HV. Isozyme- and
species-specific susceptibility of cDNA-expressed CYP1A
P-450s to different flavonoids. Biochim Biophys Acta 1994;
1205:325–35.
Frei B, Higdon JV. Antioxidant activity of tea polyphenols in
vivo: evidence from animal studies. J Nutr 2003;133:3275S–84S.
Jeong HJ, Shin YG, Kim IH, et al. Inhibition of aromatase
activity by flavonoids. Arch Pharm Res 1999;22:309–12.
Lampe JW. Isoflavonoid and lignan phytoestrogens as dietary
biomarkers. J Nutr 2003;133(suppl 3):956S–64S.
Pouget C, Fagnere C, Basly JP, et al. Synthesis and aromatase
inhibitory activity of flavanones. Pharm Res 2002;19:286–91.
Rice-Evans CA, Miller NJ, Paganga G. Structure-antioxidant
activity relationships of flavonoids and phenolic acids. Free
Radic Biol Med 1996;20:933–56.
Wang C, Makela T, Hase T, et al. Lignans and flavonoids inhibit aromatase enzyme in human preadipocytes. J Steroid
Biochem Mol Biol 1994;50:205–12.
Lagiou P, Samoli E, Lagiou A, et al. Flavonoid intake in relation to lung cancer risk: case-control study among women in
Greece. Nutr Cancer 2004;49:139–43.
Lagiou P, Samoli E, Lagiou A, et al. Flavonoids, vitamin C and
adenocarcinoma of the stomach. Cancer Causes Control 2004;
15:67–72.
Hertog MG, Feskens EJ, Hollman PC, et al. Dietary flavonoids
and cancer risk in the Zutphen Elderly Study. Nutr Cancer
1994;22:175–84.
Hollman PC, Hertog MG, Katan MB. Role of dietary flavonoids in protection against cancer and coronary heart disease.
Biochem Soc Trans 1996;24:785–9.
Peterson J, Lagiou P, Samoli E, et al. Flavonoid intake and
breast cancer risk: a case-control study in Greece. Br J Cancer
2003;89:1255–9.
Bosetti C, Spertini L, Parpinel M, et al. Flavonoids and breast
cancer risk in Italy. Cancer Epidemiol Biomarkers Prev 2005;
14:805–8.
Flavonoid Intake and Breast Cancer Risk 9
25. Nutrient Data Laboratory, Agricultural Research Service,
US Department of Agriculture. USDA–Iowa State University
Database on the Isoflavone Content of Foods, release 1.3—
2002. Beltsville, MD: US Department of Agriculture, 2002.
(http://www.nal.usda.gov/fnic/foodcomp/Data/isoflav/isoflav.
html).
26. Nutrient Data Laboratory, Agricultural Research Service, US
Department of Agriculture. USDA Database for the Flavonoid
Content of Selected Foods. Beltsville, MD: US Department of
Agriculture, 2003. (http://www.nal.usda.gov/fnic/foodcomp/
Data/Flav/flav.pdf).
27. Blot WJ, Chow WH, McLaughlin JK. Tea and cancer: a review
of the epidemiological evidence. Eur J Cancer Prev 1996;5:
425–38.
28. Arts IC, Jacobs DR Jr, Folsom AR. Dietary catechins and
cancer incidence: The Iowa Women’s Health Study. IARC Sci
Publ 2002;156:353–5.
29. Peeters PH, Keinan-Boker L, van der Schouw YT, et al. Phytoestrogens and breast cancer risk. Review of the epidemiological evidence. Breast Cancer Res Treat 2003;77:171–83.
30. Gammon MD, Neugut AI, Santella RM, et al. The Long
Island Breast Cancer Study Project: description of a multiinstitutional collaboration to identify environmental risk
factors for breast cancer. Breast Cancer Res Treat 2002;74:
233–54.
31. Potischman N, Swanson CA, Coates RJ, et al. Intake of food
groups and associated micronutrients in relation to risk of
early-stage breast cancer. Int J Cancer 1999;82:315–21.
32. Gaudet MM, Britton JA, Kabat GC, et al. Fruits, vegetables,
and micronutrients in relation to breast cancer modified by
menopause and hormone receptor status. Cancer Epidemiol
Biomarkers Prev 2004;13:1485–94.
33. Fink BN, Steck SE, Wolff MS, et al. Construction of a flavonoid database for assessing intake in a population-based
sample of women on Long Island, New York. Nutr Cancer
2006 (in press).
34. de Kleijn MJ, van der Schouw YT, Wilson PW, et al. Intake
of dietary phytoestrogens is low in postmenopausal women
in the United States: The Framingham Study. J Nutr 2001;131:
1826–32.
35. Liggins J, Bluck LJ, Runswick S, et al. Daidzein and genistein
content of fruits and nuts. J Nutr Biochem 2000;11:326–31.
36. Liggins J, Bluck LJ, Runswick S, et al. Daidzein and genistein
contents of vegetables. Br J Nutr 2000;84:717–25.
37. Liggins J, Mulligan A, Runswick S, et al. Daidzein and genistein content of cereals. Eur J Clin Nutr 2002;56:961–6.
38. Smiciklas-Wright H, Mitchell DC, Mickle SJ, et al. Foods
commonly eaten in the United States: quantities consumed per
eating occasion and in a day, 1994–96. Beltsville, MD: US
Department of Agriculture, 2002.
39. Ingram D, Sanders K, Kolybaba M, et al. Case-control study of
phyto-oestrogens and breast cancer. Lancet 1997;350:990–4.
40. Horn-Ross PL, John EM, Canchola AJ, et al. Phytoestrogen
intake and endometrial cancer risk. J Natl Cancer Inst 2003;
95:1158–64.
41. Horn-Ross PL, Hoggatt KJ, Lee MM. Phytoestrogens and
thyroid cancer risk: the San Francisco Bay Area thyroid cancer
study. Cancer Epidemiol Biomarkers Prev 2002;11:43–9.
42. McCann SE, Freudenheim JL, Marshall JR, et al. Risk of
human ovarian cancer is related to dietary intake of selected
nutrients, phytochemicals and food groups. J Nutr 2003;133:
1937–42.
43. Pietinen P, Stumpf K, Mannisto S, et al. Serum enterolactone
and risk of breast cancer: a case-control study in eastern
Finland. Cancer Epidemiol Biomarkers Prev 2001;10:339–44.
44. Lemeshow S, Hosmer D. Applied logistic regression analysis.
New York, NY: John Wiley & Sons, Inc, 1989.
45. Allison P. Survival analysis using SAS: a practical guide. Cary,
NC: SAS Institute, Inc, 1995.
46. Yasui Y, Potter JD. The shape of age-incidence curves of female breast cancer by hormone-receptor status. Cancer Causes
Control 1999;10:431–7.
47. Potter JD, Cerhan JR, Sellers TA, et al. Progesterone and estrogen receptors and mammary neoplasia in the Iowa Women’s
Health Study: how many kinds of breast cancer are there?
Cancer Epidemiol Biomarkers Prev 1995;4:319–26.
48. Lee HP, Gourley L, Duffy SW, et al. Dietary effects on breastcancer risk in Singapore. Lancet 1991;337:1197–200.
49. Wu AH, Ziegler RG, Horn-Ross PL, et al. Tofu and risk of
breast cancer in Asian-Americans. Cancer Epidemiol Biomarkers Prev 1996;5:901–6.
50. Witte JS, Ursin G, Siemiatycki J, et al. Diet and premenopausal bilateral breast cancer: a case-control study. Breast
Cancer Res Treat 1997;42:243–51.
51. Yuan JM, Wang QS, Ross RK, et al. Diet and breast cancer in
Shanghai and Tianjin, China. Br J Cancer 1995;71:1353–8.
52. Hirose K, Tajima K, Hamajima N, et al. A large-scale, hospitalbased case-control study of risk factors of breast cancer
according to menopausal status. Jpn J Cancer Res 1995;86:
146–54.
53. Horn-Ross PL, John EM, Lee M, et al. Phytoestrogen consumption and breast cancer risk in a multiethnic population:
The Bay Area Breast Cancer Study. Am J Epidemiol 2001;
154:434–41.
54. Pillow PC, Duphorne CM, Chang S, et al. Development of
a database for assessing dietary phytoestrogen intake. Nutr
Cancer 1999;33:3–19.
55. Horn-Ross PL, Barnes S, Lee M, et al. Assessing phytoestrogen exposure in epidemiologic studies: development of a
database (United States). Cancer Causes Control 2000;11:
289–98.
56. Bravo L. Polyphenols: chemistry, dietary sources, metabolism,
and nutritional significance. Nutr Rev 1998;56:317–33.
57. Kuhnau J. The flavonoids. A class of semi-essential food
components: their role in human nutrition. World Rev Nutr
Diet 1976;24:117–91.
58. Mazza G. Anthocyanins in grapes and grape products. Crit
Rev Food Sci Nutr 1995;35:341–71.
59. Imeh U, Khokhar S. Distribution of conjugated and free phenols in fruits: antioxidant activity and cultivar variations.
J Agric Food Chem 2002;50:6301–6.
60. Re R, Bramley PM, Rice-Evans C. Effects of food processing
on flavonoids and lycopene status in a Mediterranean tomato
variety. Free Radic Res 2002;36:803–10.
61. Kim DO, Chun OK, Kim YJ, et al. Quantification of polyphenolics and their antioxidant capacity in fresh plums. J Agric
Food Chem 2003;51:6509–15.
62. Willett W. Nutritional epidemiology. New York, NY: Oxford
University Press, 1998.
63. Webb AL, McCullough ML. Dietary lignans: potential role in
cancer prevention. Nutr Cancer 2005;51:117–31.
64. Ward WE, Jiang FO, Thompson LU. Exposure to flaxseed or
purified lignan during lactation influences rat mammary gland
structures. Nutr Cancer 2000;37:187–92.
65. Dabrosin C, Chen J, Wang L, et al. Flaxseed inhibits metastasis and decreases extracellular vascular endothelial growth
factor in human breast cancer xenografts. Cancer Lett 2002;
185:31–7.
66. Thompson LU. Experimental studies on lignans and cancer.
Baillieres Clin Endocrinol Metab 1998;12:691–705.
10
Fink et al.
67. Tou JC, Thompson LU. Exposure to flaxseed or its lignan component during different developmental stages influences rat
mammary gland structures. Carcinogenesis 1999;20:1831–5.
68. Linseisen J, Piller R, Hermann S, et al. Dietary phytoestrogen
intake and premenopausal breast cancer risk in a German casecontrol study. Int J Cancer 2004;110:284–90.
69. McCann SE, Muti P, Vito D, et al. Dietary lignan intakes and
risk of pre- and postmenopausal breast cancer. Int J Cancer
2004;111:440–3.
70. Keinan-Boker L, van Der Schouw YT, Grobbee DE, et al.
Dietary phytoestrogens and breast cancer risk. Am J Clin Nutr
2004;79:282–8.
71. McCann SE, Moysich KB, Freudenheim JL, et al. The risk of
breast cancer associated with dietary lignans differs by CYP17
genotype in women. J Nutr 2002;132:3036–41.
72. Bruggemeier R. Aromatase, aromatase inhibitors, and breast
cancer. Am J Ther 2001;8:333–44.
73. Jain M, Miller AB, To T. Premorbid diet and the prognosis of
women with breast cancer. J Natl Cancer Inst 1994;86:1390–7.
74. Holmes MD, Stampfer MJ, Colditz GA, et al. Dietary factors
and the survival of women with breast carcinoma. Cancer
1999;86:826–35.
75. Trudeau E, Kristal AR, Li S, et al. Demographic and psychosocial predictors of fruit and vegetable intakes differ: implications for dietary interventions. J Am Diet Assoc 1998;98:
1412–17.
76. Thompson B, Demark-Wahnefried W, Taylor G, et al. Baseline
fruit and vegetable intake among adults in seven 5 A Day study
centers located in diverse geographic areas. J Am Diet Assoc
1999;99:1241–8.
77. Subar AF, Heimendinger J, Patterson BH, et al. Fruit and
vegetable intake in the United States: the baseline survey of
the Five A Day for Better Health Program. Am J Health Promot 1995;9:352–60.
78. Potischman N, Swanson CA, Coates RJ, et al. Dietary relationships with early onset (under age 45) breast cancer in
a case-control study: influence of chemotherapy treatment.
Cancer Causes Control 1997;8:713–21.
79. Block G, Hartman AM, Dresser CM, et al. A data-based approach to diet questionnaire design and testing. Am J Epidemiol 1986;124:453–69.
80. Block G, Woods M, Potosky A, et al. Validation of a selfadministered diet history questionnaire using multiple diet
records. J Clin Epidemiol 1990;43:1327–35.