Content uploaded by Leo J Schouten
Author content
All content in this area was uploaded by Leo J Schouten
Content may be subject to copyright.
A Prospective Study of Dietary Acrylamide Intake and
the Risk of Endometrial, Ovarian, and Breast Cancer
Janneke G. Hogervorst,
1
Leo J. Schouten,
1
Erik J. Konings,
2
R. Alexandra Goldbohm,
3
and Piet A. van den Brandt
1
1
Department of Epidemiology, Maastricht University, Maastricht, the Netherlands;
2
Food and Consumer Product Safety Authority,
Region South, Department Research & Development, Eindhoven, the Netherlands; and
3
Department of Food and Chemical Risk
Analysis, TNO Quality of Life, Zeist, the Netherlands
Abstract
Background: Acrylamide, a probable human carcinogen,
was detected in various heat-treated carbohydrate-rich
foods in 2002. The few epidemiologic studies done thus
far have not shown a relationship with cancer. Our aim
was to investigate the association between acrylamide
intake and endometrial, ovarian, and breast cancer risk.
Methods: The Netherlands Cohort Study on diet and
cancer includes 62,573 women, aged 55-69 years. At
baseline (1986), a random subcohort of 2,589 women
was selected using a case cohort analysis approach for
analysis. The acrylamide intake of subcohort members
and cases was assessed with a food frequency ques-
tionnaire and was based on chemical analysis of all
relevant Dutch foods. Subgroup analyses were done for
never-smokers to eliminate the influence of smoking;
an important source of acrylamide.
Results: After 11.3 years of follow-up, 327, 300, and
1,835 cases of endometrial, ovarian, and breast cancer,
respectively, were documented. Compared with the
lowest quintile of acrylamide intake (mean intake,
8.9 Mg/day), multivariable-adjusted hazard rate ratios
(HR) for endometrial, ovarian, and breast cancer in
the highest quintile (mean intake, 40.2 Mg/day) were
1.29 [95% confidence interval (95% CI), 0.81-2.07;
P
trend
= 0.18], 1.78 (95% CI, 1.10-2.88; P
trend
= 0.02),
and 0.93 (95% CI, 0.73-1.19; P
trend
= 0.79), respectively.
For never-smokers, the corresponding HRs were 1.99
(95% CI, 1.12-3.52; P
trend
= 0.03), 2.22 (95% CI, 1.20-
4.08; P
trend
= 0.01), and 1.10 (95% CI, 0.80-1.52;
P
trend
= 0.55).
Conclusions: We observed increased risks of postmen-
opausal endometrial and ovarian cancer with increas-
ing dietary acrylamide intake, particularly among
never-smokers. Risk of breast cancer was not associated
with acrylamide intake. (Cancer Epidemiol Bio-
markers Prev 2007;16(11):2304–13)
Introduction
In 2002, Swedish researchers reported the presence of
acrylamide, classified by the IARC as a probable human
carcinogen, in several heat-treated carbohydrate-rich
foods (1). This message received a lot of media attention
and caused concern among the general population and
the scientific community, particularly because acrylam-
ide is present in foods at considerably higher levels than
other well-known food carcinogens, such as polycyclic
aromatic hydrocarbons and ethyl carbamate (2). Before
this date, acrylamide exposure was thought to occur
mainly through occupational exposure because it has
been widely used in industry since the 1950s, and further
through cigarette smoke and, to a minor extent, through
drinking water (3, 4). Since 2002, research in several
countries has also shown high acrylamide levels in foods
such as French fries, coffee, potato crisps, and cookies (4).
In the Netherlands, high levels of acrylamide were found
in the typically Dutch spiced cakes (5), which are often
consumed daily.
Acrylamide has been shown to be mainly formed in
food by Maillard browning reactions in which amino
acids, mainly asparagine, react with reducing sugars at
high temperatures (>120jC; refs. 6-8).
It is generally thought that acrylamide might cause
cancer through cytochrome P4502E1 (CYP2E1)–catalyzed
conversion to glycidamide, which has proven to be
genotoxic (4).
There is a lot of debate on whether acrylamide in food
justifies the concern that it provoked in 2002. Animal
studies have shown positive dose-response relationships
between acrylamide given in drinking water and cancer
in multiple organs in both mice and rats (9-12), among
which were several hormone-sensitive organs, such as
the mammary glands and the uterus. However, studies
on humans with occupational acrylamide exposure have
been negative thus far, apart from an increased risk of
pancreatic cancer, but this was based on a small number
of cases (13-18). Dietary acrylamide intake and its
relation with various types of cancer have been studied
in a few case-control studies (19-21) and only two pros-
pective cohort studies on breast and colorectal cancer
(22, 23). None of these have found indications for a posi-
tive relationship.
Because of the pitfalls of case-control studies and the
paucity of cohort studies, there is a need of more
Cancer Epidemiol Biomarkers Prev 2007;16(11). November 2007
Received 6/27/07; revised 8/14/07; accepted 9/5/07.
Grant support: Dutch Food and Consumer Product Safety Authority (VWA). The
NLCS was established with funding from the Dutch Cancer Society (KWF).
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 in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Janneke Hogervorst, Department of Epidemiology, Maastricht
University, P.O. Box 616, 6200 MD, Maastricht, the Netherlands. Phone: 31-433882391;
Fax: 31-433884128. E-mail: JGF.Hogervorst@EPID.Unimaas.nl
Copyright D 2007 American Association for Cancer Research.
doi:10.1158/1055-9965.EPI-07-0581
2304
prospective cohort studies. In the present study, we
investigated the relationship between dietary acrylamide
and cancer of the endometrium, ovaries, and breasts
within the framework of the Netherlands Cohort Study
on diet and cancer (NLCS). These cancer types were
chosen because of their association with acrylamide in
animal studies and because there are indications of a
positive association between fried food consumption and
both ovarian and breast cancer (24, 25). We did subgroup
analyses for women who reported to have never smoked
to eliminate the influence of smoking, which is a major
source of acrylamide exposure.
Materials and Methods
Study Cohort, Cases, and Follow-up. The Nether-
lands Cohort Study on diet and cancer started in
September 1986 with the enrollment of 62,573 women
aged 55 to 69 years. All of these women were presumed
to be postmenopausal. At baseline, data on dietary habits
and other risk factors were collected by means of a self-
administered questionnaire. The informed consent for
this study consisted of the completing and returning of
the questionnaire in 1986. Participants were informed
that by returning a completed questionnaire, they gave
consent to participate in a study about the etiology of
cancer and for record linkage to the Netherlands Cancer
Registry and the Netherlands Pathology Registry. This
procedure of informed consent was approved by the
Medical Ethics Committees of the University Hospital
Maastricht and TNO Nutrition in February 1985 and July
1986, respectively.
Data processing and analysis were based on the case-
cohort approach because the desired level of detail of the
questionnaire (e.g., including open-ended questions)
prohibited high-throughput processing. Therefore, the
questionnaires were not processed for the total cohort,
but only for a subcohort and the cases. Following the
case-cohort approach, cases were enumerated for the
entire cohort (providing the numerator information for
estimating incidence rates), whereas the accumulated
person-years for the entire cohort were estimated from a
subcohort of 2,589 women randomly sampled from the
entire cohort at baseline (providing the denominator
information for estimating incidence rates). Thus, all the
incident cases were used in the analyses. The size of the
subcohort was based on calculations that were described
in a previous publication (26). Since the start of the study,
the subcohort has been followed up regularly for vital
status information. Incident cases in the total cohort have
been detected by annual computerized record linkages to
the regional cancer registries and the Netherlands
Pathology Registry. The completeness of cancer follow-
up through linkage with the national cancer registry
was estimated to be at least 96% (27). The follow-up of
the subcohort was 100% complete at the end of the
follow-up period. Further details on the design of the
study and methods of follow-up are presented elsewhere
(26, 28-30).
The analyses are based on 11.3 years of follow-up,
from September 1986 to December 1997. This is the most
recent data set available for analysis because cancer
registry procedures, linkage to the cancer registry, checks
of the reported linkages, and the processing of the food-
frequency questionnaire cause considerable delays. After
11.3 years of follow-up, there were 327, 300, and 1,835
incident, microscopically confirmed, primary carcinomas
of the endometrium ([ICD-O]-3:C54), ovaries ([ICD-O]-
3:C56.9), and breasts ([ICD-O]-3:C50), respectively. Cases
were excluded from analysis if they had been diagnosed
with non-epithelial tumors and borderline invasive
tumors, and if their dietary data were incomplete or
inconsistent.Subcohortmemberswereexcludedfrom
Figure 1. Flow diagram
of subcohort mem bers
and cases on whom the
analyses were based.
Cancer Epidemiology, Biomarkers & Prevention
Cancer Epidemiol Biomarkers Prev 2007;16(11). November 2007
2305
the analysis if they had been diagnosed with cancer other
than skin cancer at baseline, if information on dietary
habits was missing or inconsistent, and if they reported
at baseline to have had a hysterectomy (excluded from
the endometrial cancer analysis) or an oophorectomy
(excluded from the ovarian cancer analysis). Figure 1
shows the selection and exclusion steps that resulted in
the numbers of cases and subcohort members that were
available for analysis.
Acrylamide Intake Assessment. In the NLCS, a food
frequency questionnaire with questions on 150 food
items was used to estimate dietary habits (29). The
acrylamide intake was estimated from the mean acryl-
amide level of food items, and the frequency of
consumption and portion size of the food items.
For the acrylamide intake estimation in the NLCS, we
focused on foods that are likely to contain considerable
acrylamide concentrations according to the mechanism
Table 1. Acrylamide levels in foods, used for the NLCS acrylamide intake assessment
n samples Mean (Ag/kg) Minimum (Ag/kg) Maximum (Ag/kg)
Potato crisps 40 1,249 310 2,800
Dutch spiced cake 13 1,018 260 1,410
French fries 33 351 <LOQ* 1,220
Salty snacks 12 277 45 867
Crisp bread 12 229 15 914
Cookies 20 204 10 829
Corn flakes 12 121 <LOQ* 300
Peanut butter 2 113 107 118
Chocolate 6 60 22 116
Nuts 8 33 <LOQ
c
83
Rusk 3 25 16 33
Rye bread 13 24 <LOQ
b
60
Sweet pastry 19 18 <LOQ
c
111
Coffee 9 17 9 28
Chocolate milk 4 <LOQ
c
<LOQ
c
<LOQ
c
Bread 22 <LOQ
c
<LOQ
c
<LOQ
c
*Limit of quantitation in 2002: 60 Ag/kg.
c
Limit of quantitation in 2005: 15 Ag/kg.
b
Limit of quantitation in 2002: 30 Ag/kg.
Figure 2. Absolute and relative contri-
bution of foods to the mean daily dietary
acrylamide intake of the NLCS subco-
hort.
Acrylamide and Endometrial, Ovarian, and Breast Cancer
Cancer Epidemiol Biomarkers Prev 2007;16(11). November 2007
2306
of formation and reports from various European
countries (31). Thus, the following list of food products
was established: French fries, potato crisps, bread and
rye bread, cookies, pastry, rusk, crisp bread, chocolate
and chocolate milk, nuts, peanut butter, breakfast cereals,
and coffee. To obtain an intake estimate representative
for the Dutch situation, we used data on acrylamide
levels in food products on the Dutch market. In 2002, the
Dutch Food and Consumer Product Safety Authority
analyzed acrylamide levels in various foods, such as
bread, French fries, pastry and cake, Dutch spiced cake,
crisps, corn flakes, meat, fish, and several beverages,
following an elaborate sampling scheme (5). In 2005,
another series of foods was analyzed to specifically
accommodate the estimation of acrylamide intake of the
NLCS cohort, and this series consisted of bread, rusk,
specific types of biscuits, rye bread, crisp bread, pastry
and cake, chocolate and chocolate milk, nuts and salty
snacks, peanut butter, and coffee (31). Five types of bread
were analyzed again in 2005 because the quantitation
limit of the analytic method had decreased from 30 ppb
(even 60 ppb for some foods) in 2002 to 15 ppb in 2005,
which offered the opportunity to more accurately
estimate the acrylamide content of bread. Acrylamide
in coffee was analyzed in brewed filtered coffee that was
prepared according to general Dutch preparation instruc-
tions (7 g of coffee powder/125 mL of water).
Acrylamide measurements in cookies were done in
several types of cookies known to be eaten most
frequently by the population comparable to the NLCS,
according to information from the development phase of
the questionnaire. Thus, an acrylamide level for cookies
was based on the acrylamide level of the specific types
of cookies weighted by the frequency of consumption of
the NLCS-comparable population. The same was done
for other composite food items, such as pies and
chocolates.
The foods listed above were assigned the mean value
of the acrylamide values per food or a value half the
quantitation limit when levels were lower than the
Table 2. Characteristics of the subcohort members across quintiles of daily dietary acrylamide intake; the
Netherlands Cohort Study on diet and cancer (NLCS), 1986 to 1997
Variable Q1 Q2 Q3 Q4 Q5 P*
Number of subcohort members 471 482 455 466 470
Dietary variables
Acrylamide intake, Ag/d
c
9.5 (3.7) 14.0 (1.5) 17.9 (2.2) 24.3 (4.7) 36.8 (9.6) <0.001
Acrylamide intake, Ag/kg body weight per day
c
0.14 (0.06) 0.20 (0.04) 0.26 (0.05) 0.34 (0.08) 0.53 (0.19) <0.001
Coffee consumption, g/d
c
250 (125) 500 (125) 500 (250) 500 (375) 500 (375) <0.001
Dutch spiced cake consumption, g/d
c
0.0 (0.0) 0.0 (0.7) 0.0 (1.8) 1.8 (7.8) 20.4 (12.6) <0.001
Cookie consumption, g/d
c
3.7 (8.7) 9.6 (20.4) 16.7 (14.8) 16.7 (14.8) 16.7 (14.8) <0.001
Potato crisp consumption, g/d
c
0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.3) 0.0 (0.7) <0.001
French fries consumption, g/d
c
0.0 (0.0) 0.0 (0.0) 0.0 (4.8) 0.0 (12.3) 0.0 (9.7) <0.001
Total energy intake, kcal
c
1467 (449) 1580 (452) 1671 (465) 1730 (550) 1821 (548) <0.001
Carbohydrate intake, g/d
c
153 (55) 162 (50) 176 (53) 183 (61) 200 (69) <0.001
Saturated fat intake, g/d
c
25.5 (12.0) 28.0 (12.6) 28.9 (11.6) 29.8 (13.2) 30.8 (12.3) <0.001
Trans-unsaturated fatty acid intake, g/d
c
1.8 (1.2) 2.1 (1.3) 2.3 (1.4) 2.6 (1.5) 2.7 (1.6) <0.001
Fiber intake, g/d
c
22.9 (9.5) 23.9 (8.5) 24.5 (9.1) 25.7 (9.6) 26.6 (8.6) <0.001
Vegetable intake, g/d
c
184 (107) 182 (102) 179 (83) 190 (104) 184 (96) 0.34
Fruit intake, g/d
c
184 (151) 179 (149) 175 (140) 174 (138) 180 (146) 0.42
Alcohol intake, g/d
c
0.9 (8.4) 1.6 (7.8) 2.0 (8.6) 1.9 (7.0) 1.4 (7.1) 0.09
Nondietary variables
Age, y
c
62 (7) 61 (7) 60 (7) 61 (8) 61 (7) <0.001
BMI, kg/m
2
c
24.6 (4.3) 24.8 (4.4) 24.8 (4.3) 24.6 (4.3) 24.3 (4.2) 0.07
Height, cm
c
165 (8) 165 (9) 165 (8) 165 (8) 165 (10) 0.18
Parity, n children
c
2 (3) 3 (3) 3 (2) 3 (2) 2 (3) 0.003
Age at menarche, y
c
13 (2) 13 (2) 14 (3) 14 (2) 14 (3) 0.12
Age at menopause, y
c
50 (6) 50 (6) 49 (7) 50 (6) 50 (6) 0.38
Age at first childbirth, y
c
27 (5) 26 (5) 27 (5) 26 (5) 26 (5) 0.92
Current cigarette smoking, % yes
b
17.0 18.9 25.9 21.5 19.2 0.02
n cigarettes per day among ever-smokers
c
10 (15) 10 (15) 10 (15) 10 (10) 10 (15) 0.79
n smoking years among ever-smokers
c
31 (20) 27 (20) 30 (20) 28 (21) 29 (19) 0.03
Non-occupational physical activity, minutes/d
c
51.4 (57.9) 51.4 (54.6) 53.6 (54.3) 53.6 (53.6) 51.4 (60.0) 0.91
Oral contraceptive use, % ever
b
22.6 25.1 26.1 26.5 23.6 0.61
Oral contraceptive use duration among ever-users, y
c
6 (9) 6 (9) 6 (8) 7 (10) 8 (8) 0.22
Postmenop. hrt
x
use, % ever
b
13.6 14.9 12.6 11.4 11.0 0.76
Postmenop. hrt
x
use duration among ever-users, y
c
2 (5) 1 (4) 1 (2) 2 (3) 2 (4) 0.15
Education, %
b
<0.001
Primary school 29.9 30.9 34.5 35.2 35.1
Lower vocational school 18.3 20.5 25.7 26.2 24.0
Intermediate vocational/high school 37.2 39.6 33.6 29.0 32.3
Higher vocational school/University 13.6 8.5 5.5 9.0 8.1
Family history of breast cancer, % yes
b
8.5 8.9 7.9 7.5 11.1 0.35
Family history of endometrial cancer, % yes
b
2.3 2.3 3.5 3.4 2.3 0.60
Family history of ovarian cancer, % yes
b
0.0 0.0 0.0 0.2 0.0 0.52
*P value for Kruskal-Wallis test or m
2
test (proportions).
c
Data represent medians (interquartile range).
b
Percentages.
x
Postmenop.hrt = postmenopausal hormone treatment.
Cancer Epidemiology, Biomarkers & Prevention
Cancer Epidemiol Biomarkers Prev 2007;16(11). November 2007
2307
quantitation limit because they probably do contain some
acrylamide. Other foods (meats, cooked and raw
vegetables, dairy products) were assigned the value
zero. Table 1 shows the acrylamide levels of foods that
were used in the acrylamide intake assessment.
Statistical Analysis. Acrylamide was included in the
statistical models as a continuous variable and as quintiles.
Covariables were included in the models if they
changed age-adjusted regression coefficients of acryl-
amide by more than 10%. A priori, the following
variables were considered as potential confounders:
age at menarche, duration of oral contraceptive use,
age at first childbirth, parity, age at menopause,
duration of postmenopausal hormone use, non-occupa-
tional physical activity, body mass index (BMI), height,
Table 3. Age and multivariable-adjusted association between acrylamide intake and risk of endometrial, ovarian
and breast cancer; the Netherlands Cohort Study on diet and cancer (NLCS), 1986 to 1997
Increment per 10 Ag
acrylamide/d
Q1 Q2 Q3 Q4 Q5 P
trend
Endometrial cancer
All
Cases/person-years* 221/15,836 44/3,202 40/3,078 39/3,095 46/3,175 52/3,285
HR (95% CI)
c
1.01 (0.90-1.14) 1.00
b
0.96 (0.61-1.51) 0.94 (0.59-1.49) 1.09 (0.70-1.70) 1.17 (0.76-1.79) 0.38
HR (95% CI)
x
1.04 (0.91-1.19) 1.00
b
0.95 (0.59-1.54) 0.94 (0.56-1.56) 1.21 (0.74-1.98) 1.29 (0.81-2.07) 0.18
Endometrial cancer
Never-smokers
Cases/person-years*,
never-smokers
150/9,422 27/2,113 25/1,786 29/1,709 29/1,883 40/1,931
HR (95% CI)
c
1.07 (0.93-1.24) 1.00
b
1.10 (0.62-1.97) 1.36 (0.77-2.40) 1.23 (0.70-2.16) 1.63 (0.97-2.74) 0.07
HR (95% CI)
x
1.12 (0.95-1.33) 1.00
b
1.16 (0.63-2.15) 1.35 (0.73-2.51) 1.30 (0.69-2.46) 1.99 (1.12-3.52) 0.03
Ovarian cancer
All
Cases/person-years* 195/19,037 31/3,803 39/3,939 32/3,595 39/3,823 54/3,877
HR (95% CI)
c
1.11 (0.99-1.24) 1.00
b
1.26 (0.77-2.06) 1.16 (0.69-1.94) 1.33 (0.82-2.18) 1.77 (1.11-2.82) 0.02
HR (95% CI)
k
1.11 (0.99-1.25) 1.00
b
1.22 (0.73-2.01) 1.12 (0.65-1.92) 1.28 (0.77-2.13) 1.78 (1.10-2.88) 0.02
Ovarian cancer
Never-smokers
Cases/person-years*,
never-smokers
129/11,446 18/2,540 27/2,373 22/1,992 29/2,250 33/2,292
HR(95% CI)
c
1.16 (1.00-1.34) 1.00
b
1.62 (0.87-3.02) 1.64 (0.85-3.16) 1.90 (1.03-3.50) 2.06 (1.13-3.77) 0.02
HR(95% CI)
k
1.17 (1.01-1.36) 1.00
b
1.60 (0.85-3.02) 1.64 (0.84-3.19) 1.86 (1.00-3.48) 2.22 (1.20-4.08) 0.01
Breast cancer
All
Cases/person-years* 1,350/19,036 296/3,732 267/3,941 257/3,616 254/3,877 276/3,871
HR (95% CI)
c
0.98 (0.92-1.04) 1.00
b
0.86 (0.69-1.07) 0.91 (0.72-1.14) 0.84 (0.67-1.05) 0.90 (0.73-1.13) 0.37
HR(95% CI)
{
0.99 (0.92-1.06) 1.00
b
0.80 (0.64-1.02) 0.92 (0.72-1.17) 0.86 (0.67-1.10) 0.93 (0.73-1.19) 0.79
Breast cancer
Never-smokers
Cases/person-years*,
never-smokers
767/11,540 159/2,496 159/2,379 147/2,054 144/2,317 158/2,294
HR(95% CI)
c
1.01 (0.93-1.09) 1.00
b
1.06 (0.80-1.41) 1.18 (0.88-1.59) 1.01 (0.75-1.36) 1.10 (0.82-1.47) 0.66
HR(95% CI)
{
1.01 (0.93-1.11) 1.00
b
0.97 (0.72-1.32) 1.17 (0.85-1.61) 1.00 (0.73-1.38) 1.10 (0.80-1.52) 0.55
NOTE: The numbers of cases and person-years are the numbers that resulted after list-wise deletion of observations with missing values on the selected
confounders.
*Number of cases that accumulated in the total cohort/person-years at risk of the subcohort.
c
Age-adjusted (same data set as in multivariable-adjusted analyses).
b
Reference category.
x
Adjusted for: age (y), age at menarche (y), age at menopause (y), age at first childbirth (dummy: nulliparous, 17-19 y, 20-24 y, 25-29 y, z 30 y), parity (n
children), duration of oral contraceptives use (y), duration of postmenopausal hormone use (y), BMI (kg/m
2
), height (cm), current smoking (yes/no),
quantity of smoking (cigarettes/d), duration of smoking (n smoking years), non-occupational physical activity (min/d), energy intake (kcal/d), trans-
unsaturated fatty acid intake (energy-adjusted; g/d), carbohydrate intake (energy-adjusted; g/d), alcohol consumption (g/d).
k
Adjusted for age (y), age at menarche (y), age at menopause (y), parity (n children), duration of oral contraceptives use (y), duration of postmenopausal
hormone use (y), BMI (kg/m
2
), height (cm), current smoking (yes/no), quantity of smoking (cigarettes/d), duration of smoking (n smoking years),
saturated fat intake (energy-adjusted; g/d), trans-unsaturated fatty acid intake (energy-adjusted; g/d).
{
Adjusted for age (y), age at menarche (y), age at menopause (y), age at first childbirth (dummy: nulliparous, 17-19 y, 20-24 y, 25-29 y, z30 y), duration of
oral contraceptives use (y), duration of postmenopausal hormone use (y), BMI (kg/m
2
), height (cm), current smoking (yes/no), quantity of smoking
(cigarettes/d), duration of smoking (n smoking years), socioeconomic status (four levels of education), energy intake (kcal/d), saturated fat intake (energy-
adjusted; g/d), carbohydrate intake (energy-adjusted; g/d), family history of breast cancer (yes/no), history of benign breast disease (yes/no).
Acrylamide and Endometrial, Ovarian, and Breast Cancer
Cancer Epidemiol Biomarkers Prev 2007;16(11). November 2007
2308
education level, family history of endometrial cancer
(in the endometrial cancer analyses), ovarian or breast
cancer (in the ovarian cancer analyses), breast cancer
(in the breast cancer analyses), and energy intake (32).
We also checked the following broad categories of foods
or nutrients for confounding potential: alcohol con-
sumption, intake of vegetables and fruits, and energy-
adjusted intake of saturated fat, trans-unsaturated fatty
acids, carbohydrates, and dietary fiber. Smoking status
(current versus not current), the duration of smoking,
and the number of cigarettes per day were always
included in the model because cigarette smoke is a very
Table 4. Association between acrylamide intake and risk of endometrial and ovarian cancer, adjusted for
acrylamide-containing foods; the Netherlands Cohort Study on diet and cancer (NLCS), 1986 to 1997
Additionally
adjusted for
Increment
per 10 Ag
acrylamide/d
Q1 Q2 Q3 Q4 Q5 P
trend
Endometrial cancer
All
Cases/person-years* 221/15,836 44/3,202 40/3,078 39/3,095 46/3,175 52/3,285
HR (95% CI)
c
na 1.04 (0.91-1.19) 1.00
b
0.95 (0.59-1.54) 0.94 (0.56-1.56) 1.21 (0.74-1.98) 1.29 (0.81-2.07) 0.18
HR (95% CI)
c
Coffee 1.06 (0.93-1.22) 1.00
b
1.02 (0.61-1.71) 1.04 (0.60-1.87) 1.35 (0.79-2.32) 1.46 (0.86-2.46) 0.09
HR (95% CI)
c
Dutch spiced cake 1.07 (0.79-1.43) 1.00
b
0.96 (0.59-1.56) 0.96 (0.58-1.61) 1.36 (0.78-2.35) 1.85 (0.86-3.97) 0.17
HR (95% CI)
c
Cookies 1.05 (0.92-1.20) 1.00
x
0.98 (0.61-1.59) 0.98 (0.58-1.65) 1.27 (0.77-2.09) 1.37 (0.84-2.21) 0.13
HR (95% CI)
c
French fries 1.01 (0.88-1.17) 1.00
b
0.95 (0.59-1.54) 0.93 (0.56-1.55) 1.16 (0.71-1.92) 1.24 (0.76-2.00) 0.29
HR (95% CI)
c
Potato crisps 1.03 (0.90-1.19) 1.00
b
0.95 (0.59-1.54) 0.94 (0.56-1.56) 1.20 (0.73-1.97) 1.28 (0.80-2.06) 0.20
Endometrial cancer
Never-smokers
Cases/person-years*,
never-smokers
150/9,422 27/2,113 25/1,786 29/1,709 29/1,883 40/1,931
HR (95% CI)
c
na 1.12 (0.95-1.33) 1.00
b
1.16 (0.63-2.15) 1.35 (0.73-2.51) 1.30 (0.69-2.46) 1.99 (1.12-3.52) 0.03
HR (95% CI)
c
Coffee 1.13 (0.95-1.35) 1.00
b
1.25 (0.65-2.38) 1.50 (0.77-2.93) 1.44 (0.74-2.79) 2.16 (1.16-4.05) 0.02
HR (95% CI)
c
Dutch spiced cake 1.05 (0.71-1.55) 1.00
b
1.17 (0.64-2.17) 1.38 (0.75-2.57) 1.45 (0.73-2.88) 2.67 (1.08-6.61) 0.08
HR (95% CI)
c
Cookies 1.14 (0.96-1.35) 1.00
b
1.23 (0.66-2.27) 1.45 (0.77-2.75) 1.39 (0.74-2.63) 2.13 (1.18-3.85) 0.02
HR (95% CI)
c
French fries 1.11 (0.93-1.32) 1.00
b
1.16 (0.63-2.15) 1.34 (0.73-2.49) 1.26 (0.65-2.47) 1.92 (1.07-3.45) 0.05
HR (95% CI)
c
Potato crisps 1.14 (0.95-1.35) 1.00
b
1.16 (0.63-2.14) 1.35 (0.73-2.51) 1.32 (0.70-2.48) 2.01 (1.14-3.57) 0.02
All
Ovarian cancer
Cases/person-years* 195/19,037 31/3,803 39/3,939 32/3,595 39/3,823 54/3,877
HR (95% CI) na 1.11 (0.99-1.25) 1.00
b
1.22 (0.73-2.01) 1.12 (0.65-1.92) 1.28 (0.77-2.13) 1.78 (1.10-2.88) 0.02
HR (95% CI)
x
Coffee 1.08 (0.95-1.22) 1.00
b
1.11 (0.66-1.86) 0.97 (0.56-1.70) 1.09 (0.63-1.90) 1.53 (0.93-2.53) 0.10
HR (95% CI)
x
Dutch spiced cake 1.33 (1.08-1.65) 1.00
b
1.25 (0.75-2.06) 1.18 (0.69-2.02) 1.53 (0.90-2.60) 3.24 (1.57-6.71) 0.01
HR (95% CI)
x
Cookies 1.10 (0.98-1.25) 1.00
b
1.20 (0.73-1.99) 1.10 (0.64-1.91) 1.26 (0.75-2.12) 1.74 (1.05-2.89) 0.04
HR (95% CI)
x
French fries 1.09 (0.96-1.24) 1.00
b
1.21 (0.73-2.01) 1.11 (0.65-1.90) 1.24 (0.74-2.09) 1.72 (1.04-2.82) 0.05
HR (95% CI)
x
Potato crisps 1.11 (0.99-1.25) 1.00
b
1.22 (0.73-2.01) 1.12 (0.65-1.92) 1.28 (0.77-2.13) 1.78 (1.10-2.88) 0.03
Ovarian cancer
Never-smokers
Cases/person-years*,
never-smokers
129/11,446 18/2,540 27/2,373 22/1,992 29/2,250 33/2,292
HR (95% CI)
x
na 1.17 (1.01-1.36) 1.00
b
1.60 (0.85-3.02) 1.64 (0.84-3.19) 1.86 (1.00-3.48) 2.22 (1.20-4.08) 0.01
HR (95% CI)
x
Coffee 1.12 (0.95-1.32) 1.00
b
1.43 (0.75-2.75) 1.37 (0.68-2.76) 1.54 (0.77-3.05) 1.86 (0.98-3.51) 0.08
HR (95% CI)
x
Dutch spiced cake 1.44 (1.07-1.94) 1.00
b
1.62 (0.86-3.06) 1.67 (0.85-3.26) 2.01 (1.04-3.88) 2.91 (1.10-7.67) 0.02
HR (95% CI)
x
Cookies 1.18 (1.00-1.38) 1.00
b
1.63 (0.87-3.06) 1.68 (0.85-3.32) 1.92 (1.01-3.65) 2.29 (1.21-4.34) 0.02
HR (95% CI)
x
French fries 1.17 (1.01-1.37) 1.00
b
1.61 (0.85-3.05) 1.65 (0.84-3.22) 1.91 (1.00-3.64) 2.26 (1.21-4.24) 0.01
HR (95% CI)
x
Potato crisps 1.17 (1.00-1.36) 1.00
b
1.60 (0.85-3.02) 1.63 (0.84-3.18) 1.85 (0.99-3.47) 2.19 (1.19-4.04) 0.01
NOTE: Data represent HRs for acrylamide intake. The numbers of cases and person-years are the numbers that resulted after list-wise deletion of
observations with missing values on the selected confounders. Number (percentage) of cases who consumed coffee: endometrial cancer cases: 212 (95.9);
ovarian cancer cases: 192 (98.5). Number (percentage) of cases who consumed Dutch spiced cake: endometrial cancer cases: 110 (49.8); ovarian cancer
cases: 104 (53.3). Number (percentage) of cases who consumed cookies: endometrial cancer cases: 194 (87.8); ovarian cancer cases: 179 (91.8). Number
(percentage) of cases who consumed French fries: endometrial cancer cases: 82 (37.1); ovarian cancer cases: 56 (28.7). Number (percentage) of cases who
consumed potato crisps: endometrial cancer cases: 51 (23.1); ovarian cancer cases: 42 (21.5).
Abbreviation: na, not applicable. Hrs are from Table 3.
*Number of cases that accumulated in the total cohort/person-years at risk of the subcohort.
c
Adjusted for age (y), age at menarche (y), age at menopause (y), age at first childbirth (y), parity (n children), duration of oral contraceptives use (y),
duration of postmenopausal hormone use (y), BMI (kg/m
2
), height (cm), current smoking (yes/no), quantity of smoking (cigarettes/d), duration of
smoking (n smoking years), non-occupational physical activity (min/d), energy intake (kcal/d), trans-unsaturated fatty acid intake (energy-adjusted; g/d),
carbohydrate intake (energy-adjusted; g/d), alcohol consumption (g/d).
b
Reference category.
x
Adjusted for age (y), age at menarche (y), age at menopause (y), parity (n children), duration of oral contraceptives use (y), duration of postmenopausal
hormone use (y), BMI (kg/m
2
), height (cm), current smoking (yes/no), quantity of smoking (cigarettes/d), duration of smoking (n smoking years),
saturated fat intake (energy-adjusted; g/d), trans-unsaturated fatty acid intake (energy-adjusted; g/d).
Cancer Epidemiology, Biomarkers & Prevention
Cancer Epidemiol Biomarkers Prev 2007;16(11). November 2007
2309
important source of acrylamide. Smokers have been
shown to have on average four times higher levels of
acrylamide-hemoglobin adducts, which is a marker of
internal dose of acrylamide, than nonsmokers (33, 34).
Also, for this reason, subgroup analyses were done for
never-smokers.
The proportional hazards assumption was tested using
scaled Schoenfeld residuals. Hazard rate ratios (HR) and
confidence intervals were obtained through Cox propor-
tional hazards regression with STATA software (package
9.2). SEs were estimated using the robust Huber-White
sandwich estimator to account for additional variance
introduced by sampling from the cohort. Tests for dose-
response trends were done by fitting ordinal exposure
variables as continuous variables.
To investigate if the observed associations could be
attributed to acrylamide specifically or whether the
association was due to other characteristics of acrylam-
ide-containing foods, models were run with acrylamide
and the foods (one at a time) that explained most
variance in acrylamide intake.
To check for the influence of protopathic bias, the
analyses were also done with exclusion of cases occur-
ring in the first 2 years of follow-up.
Effect modification of the association between acryl-
amide intake and cancer by other variables was tested
using Wald m
2
tests. The variables that were tested
for effect modification were selected based on their
ability to modify the activity of CYP2E1, the enzyme
that converts acrylamide to glycidamide. These variables
are diabetes (ever/never diagnosis of diabetes), obesity
(BMI > 30 kg/m
2
), smoking (both duration of smoking
in years and number of cigarettes smoked per day),
alcohol consumption (0, >0-5, >5 g/day), and physical
activity (<30, 30 to <60, 60 to <90, z90 minutes/day;
refs. 35-38).
Two-sided P values are reported throughout this paper.
Results
On average, the subcohort members had a daily
acrylamide intake of 21.0 F 11.9 Ag, which corresponds
to 0.32 F 0.19 Ag acrylamide per kg body weight per day.
Although coffee was overall the most important contrib-
utor to acrylamide intake, Fig. 2 shows that not coffee,
but Dutch spiced cake was chiefly responsible for the
variationinacrylamideintakeinthispopulation,
followed by coffee, French fries, potato crisps, and
cookies.
Table 2 shows the values of the covariables that were
assessed for confounding potential across acrylamide
intake quintiles. From this table, it can be concluded that
some variables differed considerably between the acryl-
amide intake categories. There were linear relationships
between acrylamide intake and total energy, carbohy-
drate, saturated fat, trans-unsaturated fatty acids, and
fiber intake. There were some significant differences in
the nondietary variables between the quintiles of
acrylamide intake, but these variables did not show a
linear correlation with acrylamide intake.
In Table 3, the age-adjusted and multivariable-
adjusted associations between acrylamide intake and
endometrial, ovarian, and breast cancer are shown.
There was no statistically significant association between
acrylamide as a continuous variable and endometrial
cancer, and in the total group of women, there were also
no statistically significant associations between quintiles
of acrylamide intake and endometrial cancer. However,
in the group of never-smoking women, the multivari-
able-adjusted HR in the fifth quintile [HR, 1.99; 95%
confidence interval (95% CI), 1.12-3.52] was statistically
significantly higher than that in the lowest quintile of
acrylamide intake (P
trend
= 0.03).
In the total group, acrylamide intake as a continuous
variable was borderline significantly associated with
ovarian cancer (HR, 1.11; 95% CI, 0.99-1.25 for an
increment of acrylamide intake with 10 Ag/day). In the
never-smoker subgroup, the association was statistically
significant (HR, 1.17; 95% CI, 1.01-1.36 for an increment
of acrylamide intake with 10 Ag/day). Also, the HR in the
highest quintile of acrylamide intake was statistically
significantly higher [total group: HR, 1.78 (1.10-2.88);
never-smokers: HR, 2.22 (1.20-4.08)] than that of the
lowest quintile, and the P
trend
was significant (P = 0.02
and 0.01, respectively).
There were no indications of an association between
acrylamide intake and breast cancer.
Table 4 shows the multivariable-adjusted associations
between acrylamide intake and endometrial and ovarian
cancer additionally adjusted for coffee, Dutch spiced
cake, cookies, French fries, and potato crisps consump-
tion, respectively. For both endometrial and ovarian
cancer, adjustment for Dutch spiced cake increased the
HRs in the highest quintiles, and for ovarian cancer, it
also increased the HR of the continuous acrylamide
variable. Adjustment for coffee intake slightly decreased
the HRs of the highest quintiles of acrylamide intake and
the continuous acrylamide variable for ovarian cancer
only. Adjustment for the other foods did not materially
change the HRs.
When the cases that occurred during the first 2 years of
follow-up were excluded, the HRs for endometrial and
ovarian cancer were slightly higher, whereas the results
for breast cancer remained unchanged.
None of the interaction terms between acrylamide
intake and covariables was statistically significant.
Discussion
This prospective cohort study provides the first epide-
miologic indications that dietary acrylamide intake may
increase the risk of cancer in humans. For both
endometrial and ovarian cancer, a positive association
with dietary acrylamide was observed, particularly in
the subgroup of never-smoking women. Whether the
differences between the total group and the never-
smokers are due to effect modification by smoking or
due to masking of the dietary acrylamide effect by
acrylamide or other compounds in tobacco smoke
remains to be established. Smoking may induce epoxide
hydrolase, which results in enhanced metabolism of
glycidamide (33), after which there is less glycidamide
left to damage DNA.
There was no association between acrylamide intake
and postmenopausal breast cancer, which is in line with
the only other prospective cohort study on this topic (23),
although this latter study investigated mainly premeno-
pausal breast cancer. According to the formula of Cai et
Acrylamide and Endometrial, Ovarian, and Breast Cancer
Cancer Epidemiol Biomarkers Prev 2007;16(11). November 2007
2310
al. (39) for calculating power for case-cohort studies, we
had a power of 80% to detect a HR of breast cancer 1.38
or higher for the total group and 1.52 or higher for the
never-smokers for comparing the highest quintile of
acrylamide intake to the lowest.
The relationship between dietary acrylamide intake
and endometrial cancer has not been studied before in
humans, but the positive association is in line with an
increased occurrence of uterus tumors in rats after
acrylamide exposure (9). The only epidemiologic study
thus far on acrylamide and ovarian cancer, a case-control
study, showed no association (19). This may be due to
preclinical disease affecting food consumption in the case-
control study. Generally speaking, case-control studies
may suffer from recall bias, but whether this played a role
in this particular case-control study is not clear. Recall bias
relating to acrylamide itself is unlikely, since the dietary
data were gathered before 2002, when acrylamide in food
was first reported. However, recall bias may have
occurred for acrylamide-containing foods, such as coffee
or French fries. Protopathic bias is not present in our
prospective cohort study, in which exclusion of cases in
the first 2 years of follow-up did not alter the conclusions.
Another reason for the absence of an association in the
case-control study might be the non-differential misclas-
sification of the acrylamide intake, which biases the effect
estimate toward null. Pelucchi et al. based their acrylam-
ide intake assessment on acrylamide levels of foods that
were not all specifically from the country of the study
population, and they did not obtain information from the
study participants on individual portion sizes, except for
fried potatoes.
The acrylamide intake assessment is an important
asset of the present study, together with its large study
size and prospective nature, which precludes recall bias
and makes it less vulnerable to selection. We used
acrylamide levels of foods from the Dutch market only
and specifically sampled and analyzed foods that were
relevant for the NLCS population. Furthermore, the
acrylamide measurements were of good quality, which
was shown by participation in several international
comparison studies of acrylamide analyses (5, 40).
Although we have no direct data for acrylamide itself,
the reliability and validity of the Netherlands Cohort
Study on diet and Cancer (NLCS) Ford Frequency
Questionaire (FFQ) for acrylamide can be derived from
nutrients that are correlated to acrylamide, namely,
carbohydrates and dietary fiber. For carbohydrates, the
test-retest correlation was 0.67 for women. For fiber, the
corresponding number was 0.69. Furthermore, the
decline of the correlation between the baseline question-
naire and the questionnaire administered after 5 years of
follow-up was 0.07 on average among the nutrients that
were tested. This indicates that although the question-
naire was administered only once, it characterizes long-
term food intake for over a period of at least 5 years (30).
As for validity, the correlation coefficients between the
FFQ and a dietary record method were 0.77 for
carbohydrates and 0.74 for fiber. For the food groups
potatoes, bread, and cakes and cookies, Spearman
correlation coefficients were 0.74, 0.80 and 0.65, respec-
tively (30).
Acrylamide intake could be difficult to assess due to
the large range of acrylamide levels within foods.
Thus, there is potential for non-differential misclassifi-
cation of acrylamide intake, which would bias risk
estimates toward null. We did a validation study to
investigate whether using mean acrylamide levels in
foods can still result in a sound estimate of total
acrylamide intake.
4
We estimated the acrylamide
content of 39 Dutch duplicate 24-h meals that were
collected by the Dutch National Institute for Public
Health and the Environment in 2004. From the diaries
of the participants of the duplicate meal study on the
amounts of specific foods that were eaten during the
24 h, the acrylamide content of the 24-h meals was
estimated (= amount of specific food
mean acrylam-
ide level of a specific food), using the mean acrylamide
levels of the specific foods that were also used for
the NLCS acrylamide intake assessment. Furthermore,
the acrylamide levels of the duplicate meals were
chemically analyzed (average F SD acrylamide content,
74 F 66 Ag) and correlated to the estimated acrylamide
content (average F SD estimated acrylamide content:
107 F 74 Ag), which rendered a Spearman correlation
coefficient of 0.78. This indicates that it is feasible to
make a sound rank ordering of the acrylamide intake
via a 24-h meal using these mean acrylamide levels for
individual foods. For estimating long-term, usual
acrylamide intake in the NLCS study, the mean
acrylamide levels for foods are expected to be even
more suitable because people do not always stick to
one particular brand for a certain food, but will switch
between brands, at least to some extent. Furthermore,
there is also variation in acrylamide levels within
batches of the same brand. Over time, differences in
acrylamide intake of an individual due to different
brands and within-brand fluctuations are, to some
extent, averaged out.
This study has a number of limitations. The foods that
were sampled and analyzed in 2002 and 2005 may not be
completely representative of the foods that were on the
market in 1986. Since its discovery in food in 2002,
industry and science are working together to reduce
acrylamide levels in foods. Nevertheless, it proves to be
quite hard to substantially lower the levels (41). Before
2002, some changes in acrylamide levels in foods may
have occurred due to changes in production procedures,
but never with the intent of lowering acrylamide levels.
The changes that may have occurred will have resulted
in non-differential misclassification of the intake of our
cohort, which will then have led to some underestima-
tion of the true associations between acrylamide and the
cancer types we studied.
We did not gather information on whether food was
bought or prepared at home and have therefore not taken
this into account in the analyses. Of the important
acrylamide-containing foods, French fries were most
likely to be prepared at home in the NLCS population.
However, French fries contribute relatively little to the
acrylamide intake and to the variance in acrylamide
intake in this cohort. Dutch spiced cake, which is an
important acrylamide source in this cohort, was not
prepared at home. Besides this, the misclassification
that comes from the consumption of home-cooked or
4
Janneke G. Hogervorst, Leo J. Schouten, Erik J. Konings, R. Alexandra
Goldbohm, Piet A. van den Brandt, unpublished results.
Cancer Epidemiology, Biomarkers & Prevention
Cancer Epidemiol Biomarkers Prev 2007;16(11). November 2007
2311
store-bought foods is probably non-differential and
would have biased the risk estimates toward null.
Furthermore, it has to be born in mind that the
variation in acrylamide intake in our study was largely
due to Dutch spiced cake. Dutch spiced cake resembles
gingerbread of the cake-like, non-crusty, type. It does not
always contain ginger, but it contains spices and
sometimes fruits. Levels of acrylamide in Dutch spiced
cake (mean, 1,018 ppb) are much higher than in plain
cake (mean, <30 ppb). Due to the addition of glucose-
and fructose-containing syrup, it contains a higher
amount of reducing sugars than other cakes. In the
presence of ammonium hydrogen carbonate, which is
often used as a baking agent in spiced cake, the
formation of acrylamide is further enhanced (42). Our
analyses show that it is not just Dutch spiced cake that is
responsible for the observed associations because the
associations between acrylamide intake and endometrial
and ovarian cancer were still present after adjustment for
spiced cake, and also in the subgroup of nonconsumers
of spiced cake (data for the latter analysis not shown).
The acrylamide-associated HRs that we have observed
in this study are considerably higher than the relative
risks for dietary acrylamide intake that were calculated
based on linear extrapolation of the cancer incidence
observed in animal studies (43) and that led some
epidemiologists to conclude that epidemiologic studies
probably do not have the statistical power to detect an
association between dietary acrylamide and cancer.
However, the results of extrapolating the high acrylam-
ide dosages that were given to animals to the low
dosages that humans are exposed to through food are
uncertain. The activity of enzymes in the metabolism of
acrylamide may be either higher or lower at low doses,
and humans may differ from animals in their ability to
handle acrylamide or glycidamide. Furthermore, the
genotoxic action of glycidamide (36, 44, 45) is currently
adopted as the mechanism on which acrylamide cancer
risk assessments are based. However, the fact that the
present study has found associations between acrylam-
ide intake and both endometrial and ovarian cancer
might indicate that disturbance of hormonal balances
may also be at the basis of acrylamide carcinogenesis
(46-48). Acrylamide has been shown to bind to many
proteins (49), among which may be proteins related to
maintenance of hormonal balances. Park et al. have
found a synergistic effect of acrylamide and estrogen on
the morphologic transformation of Syrian Hamster Cells
(SHE), an end point that is widely used in studies on
chemical carcinogenesis (49). This indicates that acryl-
amide may also be carcinogenic through hormonal
pathways, for which animals may not be a good model
or for which the linear extrapolation of the animal results
to the human situation may not be suitable.
The acrylamide intake of the NLCS population is in
the lower part of the WHO estimate range of 0.3-0.8 Ag/
kg BW/day for developed countries (50). It corresponds
well to the intake level that was obtained from a
duplicate meal study in Switzerland, which was 0.28
Ag/kg BW/day (51), but the authors commented that the
participants had consumed far less French fries than the
average Swiss population. Current populations, especial-
ly young people, consume larger amounts of potato
crisps and French fries than the NLCS population did.
Because of the high acrylamide levels of these foods,
current dietary intake of acrylamide is assumed to be
considerably higher (4, 5, 52) than the intake in this
study. On the other hand, industry and science are
investigating ways to reduce acrylamide levels in food
and have succeeded to some extent for certain foods (53).
If the associations we have observed in this study were
found to be causal, acrylamide in food today could prove
to be a relevant health hazard. Therefore, we encourage
other researchers to prospectively investigate the associ-
ation between dietary acrylamide intake and cancer in
hormone-sensitive organs.
Acknowledgments
We are indebted to the participants of this study and further
wish to thank the cancer registries (IKA, IKL, IKMN, IKN, IKO,
IKR, IKST, IKW, IKZ, and VIKC), and the Netherlands
nationwide registry of pathology (PALGA). We thank Arnold
Kester, Ph.D., from Maastricht University for statistical advice;
Sacha van de Crommert, Jolanda Nelissen, Conny de Zwart, and
Annemie Pisters from Maastricht University, and Henny Brants,
Willy van Dijk, and Martine Jansen from TNO Quality of Life for
assistance; Linda van den Bosch from TNO Quality of Life, and
Jack Berben and Harry van Montfort from Maastricht University
for programming assistance. We are grateful to the Laboratory
of Food and Residue Analysis of the Dutch National Institute for
Public Health and the Environment for providing the duplicate
meals for the validation study.
References
1. Tareke E, Rydberg P, Karlsson P, Eriksson S, Tornqvist M. Analysis
of acrylamide, a carcinogen formed in heated foodstuffs. J Agric
Food Chem 2002;50:4998 – 5006.
2. Summary and conclusions of the sixty-fourth meeting of the Joint
FAO/WHO Expert Committee on Food Additives (JECFA): JECFA,
2005.
3. IARC. Monographs on the evaluation of carcinogen risk to humans:
some industrial chemicals. Lyon: IARC; 1994.
4. Dybing E, Farmer PB, Andersen M, et al. Human exposure and
internal dose assessments of acrylamide in food. Food Chem Toxicol
2005;43:365 – 410.
5. Konings EJ, Baars AJ, van Klaveren JD, et al. Acrylamide exposure
from foods of the Dutch population and an assessment of the
consequent risks. Food Chem Toxicol 2003;41:1569 – 79.
6. Mottram DS, Wedzicha BL, Dodson AT. Acrylamide is formed in the
Maillard reaction. Nature 2002;419:448 – 9.
7. Stadler RH, Blank I, Varga N, et al. Acrylamide from Maillard
reaction products. Nature 2002;419:449 – 50.
8. Taeymans D, Wood J, Ashby P, et al. A review of acrylamide: an
industry perspective on research, analysis, formation, and control.
Crit Rev Food Sci Nutr 2004;44:323 – 47.
9. Johnson KA, Gorzinski SJ, Bodner KM, et al. Chronic toxicity and
oncogenicity study on acrylamide incorporated in the drinking water
of Fischer 344 rats. Toxicol Appl Pharmacol 1986;85:154 – 68.
10. Bull RJ, Robinson M, Laurie RD, et al. Carcinogenic effects of
acrylamide in Sencar and A/J mice. Cancer Res 1984;44:107 – 11.
11. Bull RJ, Robinson M, Stober JA. Carcinogenic activity of
acrylamide in the skin and lung of Swiss-ICR mice. Cancer Lett
1984;24:209 – 12.
12. Friedman MA, Dulak LH, Stedham MA. A lifetime oncogenicity
study in rats with acrylamide. Fundam Appl Toxicol 1995;27:95 – 105.
13. Marsh GM, Lucas LJ, Youk AO, Schall LC. Mortality patterns among
workers exposed to acrylamide: 1994 follow up. Occup Environ Med
1999;56:181 – 90.
14. Schulz MR, Hertz-Picciotto I, van Wijngaarden E, Hernandez JC, Ball
LM. Dose-response relation between acrylamide and pancreatic
cancer. Occup Environ Med 2001;58:609.
15. Sobel W, Bond GG, Parsons TW, Brenner FE. Acrylamide cohort
mortality study. Br J Ind Med 1986;43:785 – 8.
16. Collins JJ, Swaen GM, Marsh GM, Utidjian HM, Caporossi JC, Lucas
LJ. Mortality patterns among workers exposed to acrylamide.
J Occup Med 1989;31:614 – 7.
Acrylamide and Endometrial, Ovarian, and Breast Cancer
Cancer Epidemiol Biomarkers Prev 2007;16(11). November 2007
2312
17. Swaen GM, Haidar S, Burns CJ, et al. Mortality study update of
acrylamide workers. Occup Environ Med 2007;64:396 – 401.
18. Marsh GM, Youk AO, Buchanich JM, Kant IJ, Swaen G. Mortality
patterns among workers exposed to acrylamide: updated follow up.
J Occup Environ Med 2007;49:82 – 95.
19. Pelucchi C, Galeone C, Levi F, et al. Dietary acrylamide and human
cancer. Int J Cancer 2006;118:467 – 71.
20. Mucci LA, Dickman PW, Steineck G, Adami HO, Augustsson K.
Dietary acrylamide and cancer of the large bowel, kidney, and
bladder: absence of an association in a population-based study in
Sweden. Br J Cancer 2003;88:84 – 9.
21. Mucci LA, Lindblad P, Steineck G, Adami HO. Dietary acrylamide
and risk of renal cell cancer. Int J Cancer 2004;109:774 – 6.
22. Mucci LA, Adami HO, Wolk A. Prospective study of dietary
acrylamide and risk of colorectal cancer among women. Int J Cancer
2006;118:169 – 73.
23. Mucci LA, Sandin S, Balter K, Adami HO, Magnusson C, Weiderpass
E. Acrylamide intake and breast cancer risk in Swedish women. Jama
2005;293:1326 – 7.
24. Zhang M, Yang ZY, Binns CW, Lee AH. Diet and ovarian cancer risk:
a case-control study in China. Br J Cancer 2002;86:712 – 7.
25. Michels KB, Rosner BA, Chumlea WC, Colditz GA, Willett WC.
Preschool diet and adult risk of breast cancer. Int J Cancer 2006;118:
749 – 54.
26. van den Brandt PA, Goldbohm RA, van ’t Veer P, Volovics A,
Hermus RJ, Sturmans F. A large-scale prospective cohort study
on diet and cancer in the Netherlands. J Clin Epidemiol 1990;43:
285 – 95.
27. Goldbohm RA, van den Brandt PA, Dorant E. Estimation of the
coverage of Dutch municipalities by cancer registries and PALGA
based on hospital discharge data. Tijdschr Soc Gezondheidsz 1994;72:
80 – 84.
28. van den Brandt PA, Schouten LJ, Goldbohm RA, Dorant E, Hunen
PM. Development of a record linkage protocol for use in the Dutch
Cancer Registry for Epidemiological Research. Int J Epidemiol 1990;
19:553 – 8.
29. Goldbohm RA, van den Brandt PA, Brants HA, et al. Validation of a
dietary questionnaire used in a large-scale prospective cohort study
on diet and cancer. Eur J Clin Nutr 1994;48:253 – 65.
30. Goldbohm RA, van ’t Veer P, van den Brandt PA, et al. Rep-
roducibility of a food frequency questionnaire and stability of dietary
habits determined from five annually repeated measurements. Eur
J Clin Nutr 1995;49:420 – 9.
31. Database of the European Commission Institute for Reference
Materials and Measurements (IRMM; accessed August 3, 2006, at
http://www.irmm.jrc.be/html/activities/acrylamide/database.htm).
32. Schottenfeld D, Fraumeni JF, Jr. Cancer epidemiology and preven-
tion. New York City: Oxford University Press; 2006.
33. Schettgen T, Rossbach B, Kutting B, Letzel S, Drexler H, Angerer J.
Determination of haemoglobin adducts of acrylamide and glycida-
mide in smoking and non-smoking persons of the general popula-
tion. Int J Hyg Environ Health 2004;207:531 – 9.
34. Bergmark E. Hemoglobin adducts of acrylamide and acrylonitrile in
laboratory workers, smokers and nonsmokers. Chem Res Toxicol
1997;10:78 – 84.
35. Ardies CM, Smith TJ, Kim S, Yang CS. Induction of 4-(methylni-
trosamino)-1-(3-pyridyl)-1-butanone (NNK) activation in rat lung
microsomes by chronic ethanol consumption and repeated running
exercise. Cancer Lett 1996;103:209 – 18.
36. Ghanayem BI, Witt KL, Kissling GE, Tice RR, Recio L. Absence
of acrylamide-induced genotoxicity in CYP2E1-null mice: evidence
consistent with a glycidamide-mediated effect. Mutat Res 2005;578:
284 – 97.
37. Howard LA, Micu AL, Sellers EM, Tyndale RF. Low doses of nicotine
and ethanol induce CYP2E1 and chlorzoxazone metabolism in rat
liver. J Pharmacol Exp Ther 2001;299:542 – 50.
38. Wang Z, Hall SD, Maya JF, Li L, Asghar A, Gorski JC. Diabetes
mellitus increases the in vivo activity of cytochrome P450 2E1 in
humans. Br J Clin Pharmacol 2003;55:77 – 85.
39. Cai J, Zeng D. Sample size/power calculation for case-cohort studies.
Biometrics 2004;60:1015 – 24.
40. Wenzl T, Karasek L, Rosen J, et al. Collaborative trial validation
study of two methods, one based on high performance liquid
chromatography-tandem mass spectrometry and on gas chromato-
graphy-mass spectrometry for the determination of acrylamide in
bakery and potato products. J Chromatogr A 2006;1132:211 – 8.
41. Stadler RH. Acrylamide formation in different foods and potential
strategies for reduction. Adv Exp Med Biol 2005;561:157 – 69.
42. Amrein TM, Andres L, Manzardo GG, Amado R. Investigations
on the promoting effect of ammonium hydrogen carbonate on the
formation of acrylamide in model systems. J Agric Food Chem 2006;
54:10253 – 61.
43. Dybing E, Sanner T. Risk assessment of acrylamide in foods. Toxicol
Sci 2003;75:7 – 15.
44. Gamboa da Costa G, Churchwell MI, Hamilton LP, et al. DNA adduct
formation from acrylamide via conversion to glycidamide in adult
and neonatal mice. Chem Res Toxicol 2003;16:1328 – 37.
45. Doerge DR, da Costa GG, McDaniel LP, Churchwell MI, Twaddle
NC, Beland FA. DNA adducts derived from administration of
acrylamide and glycidamide to mice and rats. Mutat Res 2005;580:
131 – 41.
46. Akhmedkhanov A, Zeleniuch-Jacquotte A, Toniolo P. Role of
exogenous and endogenous hormones in endometrial cancer: review
of the evidence and research perspectives. Ann N Y Acad Sci 2001;
943:296 – 315.
47. Cramer DW, Welch WR. Determinants of ovarian cancer risk. II.
Inferences regarding pathogenesis. J Natl Cancer Inst 1983;71:
717 – 21.
48. Exon JH. A review of the toxicology of acrylamide. J Toxicol Environ
Health B Crit Rev 2006;9:397 – 412.
49. Park J, Kamendulis LM, Friedman MA, Klaunig JE. Acrylamide-
induced cellular transformation. Toxicol Sci 2002;65:177 – 83.
50. Health implications of acrylamide in food. Report of a Joint FAO/
WHO Consultation.: FAO/WHO; 2002.
51. Swiss Federal Office of Public Health. Assessment of acrylamide
intake by duplicate diet study. 2002. (accessed August 3, 2006, at
http://www.bag.admin.ch).
52. Matthys C, Bilau M, Govaert Y, Moons E, De Henauw S, Willems JL.
Risk assessment of dietary acrylamide intake in Flemish adolescents.
Food Chem Toxicol 2005;43:271 – 8.
53. Acrylamide Status Report December 2004: a summary of the efforts
and progress achieved to date by the European food and drink
industry (CIAA) in lowering levels of acrylamide in food. Bruxelles:
CIAA; 2004.
Cancer Epidemiology, Biomarkers & Prevention
Cancer Epidemiol Biomarkers Prev 2007;16(11). November 2007
2313
