IJC

International Journal of Cancer

Dietary total antioxidant capacity and gastric cancer risk

in the European prospective investigation into cancer and
nutrition study
Mauro Serafini1, Paula Jakszyn2*, Leila Luján-Barroso2, Antonio Agudo2, H. Bas Bueno-de-Mesquita3,4,
Fränzel J.B. van Duijnhoven3,4, Mazda Jenab5, Carmen Navarro6,7, Domenico Palli8, Heiner Boeing9, Peter Wallström10,
Sara Regne r10, Mattijs E. Numans11, Fatima Carneiro12, Marie-Christine Boutron-Ruault13, Françoise Clavel-Chapelon13,
Sophie Morois13, Sara Grioni14, Salvatore Panico15, Rosario Tumino16, Carlotta Sacerdote17, Jose  Ramon Quirós18,
Esther Molina-Montes , Jose M. Huerta Castaño , Aurelio Barricarte , Pilar Amiano , Kay-Tee Khaw22,
7,19 6,7 7,20 7,21

Nicholas Wareham23, Naomi E. Allen23, Timothy J. Key23, Suzanne M. Jeurnink24, Petra H.M. Peeters11,
Christina Bamia25, Elisabeth Valanou25,26, Antonia Trichopoulou25,26, Rudolf Kaaks27, Annekatrin Lukanova27,
Manuela M. Bergmann9, Björn Lindkvist28, Roger Stenling29, Ingegerd Johansson30, Christina C. Dahm31,32,
Kim Overvad32, Majken Jensen32, Anja Olsen33, Anne Tjonneland33, Eiliv Lund34, Sabina Rinaldi5, Dominique Michaud35,
Traci Mouw35, Elio Riboli36 and Carlos A. González2
1
Antioxidant Research Laboratory, Istituto Nazionale di Ricerca per gli Alimenti e la Nutrizione (INRAN), Rome, Italy
2
Unit of Nutrition, Environment and Cancer, Cancer Epidemiology Research Program, Catalan Institute of Oncology, (ICO-IDIBELL) Barcelona, Spain
3
National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands
4
Department of Gastroenterology and Hepatology, University Medical Centre Utrecht (UMCU)
5
International Agency for Research on Cancer (IARC-WHO), Lyon, France
6
Department of Epidemiology, Murcia Health Council, Murcia, Spain
7
CIBER Epidemiologı́a y Salud Pública (CIBERESP), Spain
8
Molecular and Nutritional Epidemiology Unit, Cancer Research and Prevention Institute (ISPO)
9
Department of Epidemiology, German Institute of Human Nutrition Potsdam-Rehbrücke
10
Department of Surgery, Skåne University Hospital Malmö, Lund University, Malmö, Sweden
11
Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht
12
Institute of Molecular Pathology and Immunology of the University of Porto (IPATIMUP) and Medical Faculty/HS João, Porto, Portugal
13
INSERM, Centre for Research in Epidemiology and Population Health, Institut Gustave Roussy, Villejuif, France, Paris South University, Villejuif, France
14
Department of Preventive & Predictive Medicine, Nutritional Epidemiology Unit, Fondazione IRCCS Istituto Nazionale dei Tumori – Milan, Milan – Italy
15
Department Of Clinical And Experimental Medicine, Federico Ii University, Naples, Italy
16
Cancer Registry and Histopathology Unit, ‘‘Civile M.P. Arezzo’’ Hospital, Ragusa, Italy
17
Center for Cancer Prevention (CPO Piedmont), and Human Genetic Foundation (Hugef), Turin, Italy, Florence, Italy
18
Public Health and Participation Directorate, Health and Health Care Services Council, Asturias, Spain
19
Andalusian School of Public Health, Granada, Spain
20
Navarre Public Health Institute. Pamplona. Spain
Epidemiology

21
Public Health Division of Gipuzkoa, Basque Regional Health Department, Spain
22
Dept Public Health and Primary Care, University of Cambridge, Cambridge, UK
23
Cancer Epidemiology Unit, Nuffield Department of Clinical Medicine, University of Oxford, Oxford, UK
24
Department of Gastroenterology and Hepatology, University Medical Centre Utrecht (UMCU), Utrecht, The Netherlands
25
WHO Collaborating Center for Food and Nutrition Policies, Department of Hygiene, Epidemiology and Medical Statistics, University of Athens Medical
School, Athens, Greece
26
Hellenic Health Foundation, Athens, Greece

This article was published online on 9 November 2011. An error was subsequently identified. This notice is included in the online and
print versions to indicate that both have been corrected XX XXXX 2012.
Key words: stomach cancer, diet, antioxidant capacity, longitudinal studies
Grant sponsor: of the Spanish Ministry of Health (Health Research Fund, FIS); Grant numbers: RCESP-C03/09; RTICCC-C03/10, R06/0020;
Grant sponsor: Fundación La Caixa; Grant number: BM06-130-0; Grant sponsors: Dutch Ministry of Public Health, Welfare and Sports
(VWS), Netherlands Cancer Registry (NKR), LK Research Funds, Dutch Prevention Funds, Dutch ZON (Zorg Onderzoek Nederland), World
Cancer Research Fund (WCRF) (The Netherlands), Statistics Netherlands
DOI: 10.1002/ijc.27347
History: Received 24 Jan 2011; Accepted 22 Jul 2011; Online 9 Nov 2011
Correspondence to: Paula Jakszyn, PhD, Unit of Nutrition, Environment and Cancer, Cancer Epidemiology Research Programme, Catalan
Institute of Oncology-IDIBELL, AV Gran via 193 (08907) L’Hospitalet de Llobregat, Spain, Tel: þ34-93-260-74-01, Fax: þ34-93-260-77-87,
E-mail: paujak@iconcologia.net

Int. J. Cancer: 131, E544–E554 (2012) V

C 2011 UICC
Serafini et al. E545

27
Department of Cancer Epidemiology, German Cancer Research Center, Heidelberg, Germany
28
Department of Internal Medicine, Division of Gastroenterology and Hepatology, Sahlgrenska University Hospital, Gothenburg, Sweden
29
Department of Medical Biosciences, Pathology, Umeå University, Umeå, Sweden
30
Department of Public Health and Clinical Medicine, Nutritional Research, Umeå University, Umeå, Sweden
31
Department of Clinical Epidemiology, Aarhus University Hospital, Aalborg, Denmark
32
Department of Epidemiology, School of Public Health, Aarhus University, Aarhus, Denmark
33
Danish Cancer Society, Institute of Cancer Epidemiology, Diet Cancer and Health, Copenhagen, Denmark
34
Department of Community Medicine, University of Tromsø
35
Department of Epidemiology and Biostatistics, School of Public Health, Imperial College London, UK
36
School of Public Health, Imperial College London, St Mary’s Campus, Imperial College, London, UK

A high intake of dietary antioxidant compounds has been hypothesized to be an appropriate strategy to reduce gastric
cancer (GC) development. We investigated the effect of dietary total antioxidant capacity (TAC) in relation to GC in the
European Prospective Investigation into Cancer (EPIC) study including 23 centers in 10 European countries. A total of
521,457 subjects (153,447 men) aged mostly 35–70 years old, were recruited largely between 1992 and 1998. Ferric
reducing antioxidant potential (FRAP) and total radical-trapping antioxidant parameter (TRAP), measuring reducing and
chain-breaking antioxidant capacity were used to measure dietary TAC from plant foods. Dietary antioxidant intake is
associated with a reduction in the risk of GC for both FRAP (adjusted HR 0.66; 95%CI (0.46–0.95) and TRAP (adjusted HR
0.61; 95%CI (0.43–0.87) (highest vs. lowest quintile). The association was observed for both cardia and noncardia cancers. A
clear effect was observed in smokers with a significant reduction in GC risk for the fifth quintile of intake for both assays
(highest vs. lowest quintile: adjusted HR 0.41; 95%CI (0.22–0.76) p for trend <0.001 for FRAP; adjusted HR 0.52; 95%CI
(0.28–0.97) p for trend <0.001 for TRAP) but not in nonsmokers. In former smokers, the association with FRAP intake was
statistically significant (highest vs. lowest quintile: adjusted HR 0.4; 95%CI (0.21–0.75) p < 0.05); no association was
observed for TRAP. Dietary antioxidant capacity intake from different sources of plant foods is associated with a reduction in
the risk of GC.

Gastric cancer (GC) remains the second leading cause of can-
cer-related deaths overall, and accounts for nearly 9% of total
cancer incidence, representing the fourth most common type
of cancer worldwide.1 Although the exact mechanisms of the
carcinogenic process remain largely unknown, oxidative and
inflammatory stress induced by Helicobacter pylori (Hp)
infection and other risk factors, such as smoking, salt intake,
meat and smoked foods, are thought to represent a crucial
causing ‘‘disturbances’’ in the endogenous redox network.
Moreover, the contribution of Vitamin C, Vitamin E and
b-carotene to the antioxidant capacity of fruit and vegetables
is lower respect to the contributions of the hundreds of dif-
ferent antioxidants present in the food matrix. Antioxidant
molecules do not act in isolation and synergistic interactions
are part of the ordinary mechanism of protection played by
the redox network against oxidative stress. About a relevant

mechanism in the chain of events leading to neoplastic cell
transformation.2 Based on this, a high intake of antioxidant
molecules from diet has been hypothesized to be an appro-
priate strategy to reduce the damage induced by oxidative
and inflammatory stress.
Epidemiological studies suggest an inverse association
between intake of foods rich in bioactive redox substances,
such as fruit and vegetables, and risk of GC, particularly for
raw vegetables and citrus fruits rich in vitamin C and poly-
phenols.3–5 However, results from clinical trials have pro-
duced contrasting results6: a review of 14 randomized trials
(170,525 participants) found no evidence that antioxidant
supplementation with b-carotene, vitamin A, vitamin C, vita-
min E and selenium prevented GC. On the contrary, an
effect of antioxidant supplementation on increasing overall
mortality was described, raising strong concerns about the
use of antioxidants supplementation for GC prevention.
It must be pointed out that clinical trials have utilized
synthetic antioxidants at doses far higher than nutritional
recommendations and for long periods of time, potentially
Epidemiology
epidemiological finding, the importance of synergistic redox
interactions came from a case-control study by Ekstrom et al.
showed that the combined intake of dietary antioxidants such
as vitamin C, a-tocopherol and b-carotene was associated
with a 70% lower risk reduction of developing GC.7
In this view and to properly assess the impact of dietary
antioxidants on GC risk, information on the overall antioxi-
dant intake from diet, needs to be taken in account. Total
antioxidant capacity (TAC) represents a direct measurement
of the nonenzymatic antioxidant network considering single
antioxidant activity as well as the synergistic interactions of
the redox molecules present in the tested matrix such as food
extracts, biological fluids or tissues.8 Different studies in
human subjects have shown that diet is able to modulate
plasma TAC following consumption of foods rich in antioxi-
dants, such as tea, wine, fruit juices, onions, lettuce, chocolate
and vegetables.8 In larger trials, Pitsavos et al.9 showed that
plasma TAC was significantly associated with the Mediterra-
nean diet score and with consumption of fruits, vegetables
and olive oil in the ATTICA study in 3,024 subjects. A

Int. J. Cancer: 131, E544–E554 (2012) V

C 2011 UICC
 E546
significant correlation between dietary TAC intake estimated
by a food frequency questionnaire and plasma TAC was
described in Sweden.10 More recently, plasma TAC was
shown to increase with a Mediterranean-style diet rich in
virgin olive oil after 3 years of intervention in 187 high
cardiovascular risk patients of the Predimed trial.11 Moreover,
Serafini et al.12 showed that intake of TAC from fruit and
vegetables was associated with a lower risk of GC in a popu-
lation-based case-control study in Sweden. Agudo et al.13
showed that dietary TAC from fruit and vegetables was
inversely related to overall mortality rates in the Spanish
cohort of European Prospective Investigation into Cancer
(EPIC), however except for this study, no evidence is avail-
able from cohort studies.
Supplementary sources of antioxidants other than fruit
and vegetables are present in the diet at very high concentra-
tion (such as chocolate, tea, wine and spices).14 Foods, such
as grain, cereals and juices, despite being endowed with a
lower amount of antioxidants, may also contribute to overall
antioxidant dietary intake due to their high frequency of
daily consumption. To properly assess dietary antioxidant
intake, it is crucial to merge the information on dietary
intake from all plant foods in an overall TAC intake, which
better resembles a daily exposure to antioxidants. We
hypothesized that TAC intake from plant foods is associated
with a lower risk of GC occurrence with an improved
effectiveness in subjects exposed to high free radical
production (i.e., smoking). To test our hypothesis, we investi-
gated the effect of dietary TAC (measured through radical-
trapping antioxidant parameter (TRAP) and ferric reducing
antioxidant potential (FRAP)) in relation to GC in the EPIC
study.
Epidemiology
Material and Methods
Subjects
The methodological details and rationale behind the EPIC
study has been described previously.15 In summary, EPIC is a
prospective study involving 23 centers from 10 European
countries (Denmark, France, Germany, Greece, Italy, The
Netherlands, Norway, Spain, Sweden and United Kingdom).
A total of 521,457 subjects (153,447 men) aged mostly 35–
70 years old, were recruited largely between 1992 and 1998.
The majority of the participants were from the general popu-
lation, selected from a defined geographical area, region
or town, with exceptions for France (health insurance mem-
bers), Utrecht, the Netherlands) and Florence, Italy (partici-
pants of breast cancer screening programs), Oxford, United
Kingdom (mostly vegetarian volunteers) and parts of the
Spain and Italy cohorts (mostly blood donors). The ethical
review boards from the International Agency for Research on
Cancer (IARC) and all local participating centers approved
the study.
Dietary antioxidant capacity and gastric cancer
Data collection
Dietary intake was assessed by a number of different instru-
ments that had been developed and validated previously in a
series of studies within the various source populations partic-
ipating in EPIC.16 Extensive self-administered quantitative
dietary questionnaires systematically estimating individual
average portions were used in Italy, The Netherlands,
Germany, Greece and south of Italy by face to face interview,
Spain (Diet History administered by interviewers) and France.
Semiquantitative food-frequency questionnaires were used in
Denmark, Norway, Naples (Italy) and Umea (Sweden).
Combined dietary methods were used in the UK and Malmo
(Sweden). The UK used a semiquantitative food frequency
questionnaire and a 7-day record, and Malmo used a short,
semiquantitative food frequency questionnaire with a 14-day
record on hot meals. A lifestyle questionnaire collected infor-
mation about sociodemographic characteristics, lifestyles
(especially those related to cancer etiology such as lifetime
history of alcohol and smoking) and medical history. Anthro-
pometric measures were taken at recruitment as well as blood
samples (from
74% of the subjects).15
Follow-up and identification of cancer cases
Vital status was obtained through periodic linkage to regional
and national mortality registries. Information on cancer status
(including diagnosis of GC) was obtained by linkage with
population cancer registries, except for France, Germany,
Greece and Naples where a combination of different active
follow-up methods were used. The date of last complete
follow-up (recorded by central database at IARC) ranged from
2003 to 2006, depending on the center.
GC included cancers coded as C16 from the 10th Revision
of the International Statistical Classification of Diseases. They
were classified according to both anatomic location (cardia
and noncardia) and Lauren histological type (intestinal and
diffuse). As previously described, the majority of cancers
were validated and confirmed by a panel of pathologists who
reviewed specimen material and pathology reports from each
centre, detailed previously.17
Subjects lost to follow-up or with a prevalent cancer at
recruitment were excluded at baseline (n ¼ 27,090). During
the follow-up, 576 incident cases of GC were identified. As
the analysis included primary gastric adenocarcinomas as
cases, we censored at date of diagnosis (i) 24 subjects with
GC (adenocarcinoma) who had another type of cancer dur-
ing follow-up and previous to the GC and (ii) 89 nonadeno-
carcinoma GC (gastric lymphomas (n ¼ 33), gastric stump
cancers (n ¼ 8), other (n ¼ 23) and unspecified (n ¼ 26).
In addition, we excluded from the analysis subjects either
without dietary information available (13 cases and 6,147
noncases) or considered to have implausible dietary data,
defined as more than three standard deviations from the sex-
specific mean of the log transformed energy intake per day
(one case and 3,162 noncases). Therefore, the final analysis
Int. J. Cancer: 131, E544–E554 (2012) V
C 2011 UICC
Serafini et al. E547

included 449 primary incident gastric adenocarcinoma cases, Statistical analysis

of which 47 were gastroesophageal junction tumors. Descriptive statistics were presented for the whole cohort
according to TRAP and FRAP intake. To estimate adjusted
median intakes, we applied the t-test (adjusted by sex, age,
Dietary FRAP and TRAP
energy intake, BMI, tobacco smoking). Cox proportional haz-
Data on TAC from plant foods was gathered from published
ards regression models were used to assess the association
databases that provided the antioxidant capacity measured in
between the TRAP or FRAP and GC. Age was the primary
foods by total radical-trapping antioxidant parameter (TRAP)
time variable; entry time was defined as age at recruitment
and ferric reducing antioxidant power (FRAP), measuring,
and exit time defined as age of first GC diagnosis (for cases),
respectively, the chain-breaking antioxidant capacity and the
and diagnosis of a cancer other than GC, death or last com-
reducing power.14,18 Briefly, three food samples for each food
plete follow-up (for at risk subjects), depending on which
items were purchased, selecting the three cultivars and/or
occurred first. Sixteen subjects were excluded from the
brands with the highest sales in the market. Samples were
regression models due to lack of diagnosis date for a primary
then prepared, mixed in equal proportions and analyzed,
tumor other than a GC.
after appropriate extraction and dilution, in duplicate for
All models were stratified by country to control for differen-
TRAP and FRAP. The variation in TRAP and FRAP values
ces in follow-up time and questionnaire design between
for replicates was always between 3 and 10% relative stand-
countries, and by age at EPIC study entry (1-year intervals).
ard deviation, when it was higher than 10%, the analyses
Models were adjusted for sex, BMI (<25 kg/m2, 25–30 kg/m2,
were repeated. Food composition database from different
>30 kg/m2) education level (no formal education, primary
food groups (fruits, vegetables, wine, cereals, potatoes, choco-
school, secondary school, technical or professional training,
late, juice, tea, soups, legumes, condiments and soft drinks)
university and not specified), smoking status and intensity
was used, including information on 207 and 210 food items
(never, former from <10 years, former from 10 years, current
for TRAP and FRAP, respectively. The value of dietary intake
<20 cigarettes/day, current 20 cigarettes/day and not speci-
has been calculated based on the TRAP and FRAP values of
fied) and energy intake (Kcal/day).
the single food items multiplied for the frequency of
Hazard ratios (HR) were calculated for categorical FRAP
consumption.
and TRAP sex-specific quintiles, using the first quintile as the
For food items for which TAC data were not available,
reference category. Furthermore, log2 transformed FRAP and
the value of the nearest comparable food was assigned. When
TRAP values were also analyzed as continuous variables.
several matches for a food item were found, the weighted
Trend tests were calculated based on quintile-based scores
mean of all the suitable values were assigned according to the
1–5 used as continuous variables. The natural logarithm is
information from the 24 HR. Maillard products, produced
the most common transformation used to normalize right
during the process of coffee roasting between reducing sugars
skewed data; however, we used log2 transformation because it
and amino acids, are the main contributor to the in vitro
produces the same normalizing effect, but the HR is better
antioxidant capacity of coffee.19 However, due to their high
interpretable as it corresponds to the increase of risk of GC
molecular weight, it is still unclear if they are efficiently
for a doubling of intake. Additional models were created to
absorbed displaying an antioxidant effect in vivo.20 Because

of this discrepancy, coffee might be a strong confounder of
TAC intake, for this reason dietary information on coffee
was not taken in account.
Hp infection
Antibodies against Hp were determined in cases and controls
selected for the nested case–control study within the EPIC
cohort. To define Hp-positive infection, we included data
from a previous analysis (103 noncardia GC cases and 519
controls21 and data from a second analysis performed in a
different laboratory (INSERM, Bordeaux laboratory) which
included 75 new incident noncardia GC cases and 294
new controls. In both datasets, we considered subjects with
positives results in one or both ELISA Hp antibodies and
CagA antibodies as Hp positive. For each incident, GC case
with available blood sample, four control subjects were ran-
domly selected from the cohort, matched by sex, age group
(62.5 years), center and date of blood sample collection
(645 days).
Epidemiology
assess risk of GC by cardia and noncardia location and
diffuse and intestinal types, while censoring cases with
unclassified and mixed locations or types. The Wald statistic
was used to assess homogeneity of risk by location and histo-
logical type.22 When the data were analyzed according to
H. pylori infection status, unconditional logistic regression
modeling was used to estimate ORs. Interaction with meat
and smoking was estimated by likelihood ratio test. To fur-
ther evaluate whether the association between FRAP and
TRAP intakes and GC risk were linear, we created restricted
cubic splines (at 5th, 50th, 75th percentiles).23
Results
The range and median intake of TRAP and FRAP for males
and females by quintiles are described in Table 1. The me-
dian intake was 10,100.9 mmol Trolox equivalents for FRAP
and 3,442 mmol Fe2þ equivalents for TRAP.
Table 2 describes the contribution from different food
groups to overall TAC dietary intake, in all subjects and by
sex in the EPIC cohort. Tea, with a percentage of 28.3%

Int. J. Cancer: 131, E544–E554 (2012) V

C 2011 UICC
 Epidemiology

E548

Table 1. Mean and median intake of TRAP and FRAP by quintiles

TRAP (mmol TE1) FRAP (mmol FE2)
Quintiles Male Female Total Male Female Total
1 N 28,744 67,915 96,659 28,744 67,915 96,659
Range (118.22–2,106.0) (64.49–2,028.8) (64.49–2,050.9) (501.03–6,750.2) (433.85–6,399.7) (433.85–6,490.6)
Median 1,602.9 1,587.4 1,591.2 5,344 5,163.3 5,213.1
2 N 28,744 67,916 96,660 28,744 67,916 96,660
Range (2,106.0–3,038.1) (2,028.8–2,883.6) (2,050.9–2,927.8) (6,750.3–9,219.4) (6,399.8–8,617.4) (6,490.6–8,788.3)
Median 2,561.7 2,443.7 2,477.4 7,966.7 7,499.5 7,627.6
3 N 28,744 67,916 96,660 28,745 67,916 96,661
Range (3,038.1–4,329.6) (2,883.6–4,018.1) (2,927.8–4,107.1) (9,219.5–12,291.0) (8,617.4–11,389.0) (8,788.3–11,656.4)
Median 3,597.1 3,382.6 3,442 10,657 9,883.9 10,100.9
4 N 28,745 67,916 96,661 28,744 67,916 96,660
Range (4,329.7–6,444.5) (4,018.1–5,912.3) (4,107.1–6,063.6) (12,291.0–16,998.0) (11,389.0–15,795.0) (11,656.4–16,161.3)
Median 5,200.2 4,847 4,953.2 14,280 13,262 13,565.7
5 N 28,744 67,916 96,660 28,744 67,916 96,660
Range (6,444.5–14,991.0) (5,912.4–15,000.0) (6,063.6–15,000.0) (16,998.0–37,995.0) (15,795.0–37,874.0) (16,161.3–37,995.0)
Median 8,332.7 7,748.2 7,931.2 21,038 19,851 20,234.8
Total N 143,721 339,579 483,300 143,721 339,579 483,300
Median 3,597.1 3,382.7 3,442 10,657 9,883.9 10,100.9
1
TRAP is expressed as mmol Trolox equivalents. 2FRAP is expressed as mmol Fe2þ equivalents.

Int. J. Cancer: 131, E544–E554 (2012) V

C 2011 UICC
Dietary antioxidant capacity and gastric cancer
Serafini et al. E549

Table 2. Contribution of food groups to overall TRAP and FRAP intake (%) in all subjects stratified by sex
Potatoes Sugar and Soft
Sex Wine Cereals1 Condiment Fruits Legumes and derivate Soup confectionary2 Vegetables Juice drinks Tea
Male TRAP 19.0 4.0 0.8 14.1 0.4 4.9 2.0 9.6 11.6 3.3 4.5 26.0
FRAP 14.2 10.7 0.7 14.7 0.6 7.6 2.1 8.0 11.8 4.3 4.3 21.1
Female TRAP 9.7 3.2 0.7 19.1 0.4 3.6 1.5 10.5 15.3 3.5 3.3 29.3
FRAP 7.3 8.8 0.6 20.4 0.6 5.6 1.7 8.8 14.9 4.6 3.0 23.8
Total TRAP 12.6 3.4 0.7 17.5 0.4 4.0 1.7 10.2 14.2 3.4 3.7 28.3
FRAP 9.4 9.4 0.7 18.6 0.6 6.2 1.8 8.5 14.0 4.5 3.4 23.0
1 2
Cereals: flours, pasta, breakfast cereals, rice, spelt and corn. Sugar and confectionary: honey, jam and chocolate.

Table 3. Intake of dietary equivalents of FRAP and TRAP in the EPIC cohorts according to demographic, anthropometric and lifestyle variables
N FRAP1,2 (CI 95%) TRAP1,2 (CI 95%)
All 483,300 9,617 (9,590–9,644) 3,223 (3,212–3,234)
Sex
Male 143,721 9,976 (9,943–10,010) 3,321 (3,307–3,334)
Female 339,579 9,270 (9,243–9,297) 3,127 (3,116–3,138)
Age (years)
34 28,167 9,285 (9,232–9,339) 3,064 (3,043–3,085)
35–44 91,976 9,332 (9,298–9,367) 3,089 (3,075–3,102)
45–54 191,367 9,670 (9,641–9,699) 3,249 (3,238–3,261)
55–64 139,614 9,868 (9,837–9,899) 3,339 (3,326–3,351)
>64 32,176 9,947 (9,897–9,998) 3,384 (3,364–3,405)
BMI
Normal (18.5–25 kg/m2) 250,158 9,796 (9,766–9,826) 3,291 (3,279–3,303)
Overweight (>25 kg/m2) 167,910 9,647 (9,616–9,678) 3,239 (3,226–3,251)
Obese (>30 kg/m2) 65,232 9,411 (9,373–9,449) 3,140 (3,125–3,155)
Energy Intake3 (Kcal/day)
Q1 96,659 7,001 (6,976–7,026) 2,364 (2,354–2,374)
Q2 96,660 8,639 (8,608–8,671) 2,904 (2,892–2,917)
Q3 96,661 9,673 (9,638-9,708) 3,243 (3,229–3,258)

Epidemiology
Q4 96,660 10,805 (10,766–10,845) 3,612 (3,596–3,628)
Q5 96,660 13,012 (12,964–13,060) 4,320 (4,301–4,339)
Educational level
None 19,372 8,815 (8,754–8,876) 2,894 (2,870–2,918)
Primary school completed 112,234 9,127 (9,096–9,158) 3,024 (3,012–3,037)
Technical/professional school 108,572 9,733 (9,699–9,767) 3,279 (3,265–3,293)
Secondary school 110,347 10,028 (9,993–10,064) 3,398 (3,384–3,412)
Longer education (including University degree) 113,885 10,406 (10,370–10,442) 3,550 (3,535–3,565)
Not specified 18,890 9,680 (9,617–9,744) 3,234 (3,209–3,260)
Smoking status
Never 235,995 9,659 (9,635–9,684) 3,227 (3,218–3,237)
Former 128,635 9,951 (9,922–9,980) 3,356 (3,344–3,367)
Smoker 108,693 9,107 (9,079–9,135) 3,034 (3,022–3,045)
Unknown 9,977 9,771 (9,691–9,852) 3,282 (3,250–3,315)
1
Adjusted by sex, age, BMI, energy intake, educational level, smoking status and country. 2All differences between categories were statistically
significant (p < 0.001). 3Quintiles of energy intake (Kcal/day): Male 1: (949.25–1,843.39), 2: (1,843.39–2,188.26), 3: (2,188.26–2,518.38), 4:
(2,518.38–2,955.99), 5: (2,955.99–5,716.66). Female 1: (745.14–1,475.13), 2: (1,475.13–1,752.59), 3: (1,752.59–2,019.92), 4: (2,019.92–
2,378.96), 5: (2,378.96–4,667.18).

Int. J. Cancer: 131, E544–E554 (2012) V

C 2011 UICC
 Epidemiology

E550

Table 4. Risk of gastric cancer in association to dietary consumption of FRAP and TRAP in EPIC cohorts: hazard ratios (HR) and 95% confidence intervals (CI)1
HR and 95% CI by quintiles of TAC intake
Variables Categories Cases Q2 Q3 Q4 Q5 p Trend HR (log2)2
FRAP Total 444 0.70 (0.53–0.93) 0.71 (0.52–0.95) 0.60 (0.44–0.84) 0.61 (0.43–0.87) <0.0001 0.81 (0.69–0.94)
Sex Male 255 0.66 (0.45–0.96) 0.70 (0.47–1.04) 0.56 (0.36–0.87) 0.68 (0.43–1.08) 0.0672 0.85 (0.69–1.04)
Female 189 0.75 (0.49–1.15) 0.70 (0.45–1.10) 0.67 (0.41–1.10) 0.52 (0.28–0.95) 0.0379 0.78 (0.60–1.00)
Localization site3 Cardia 130 0.69 (0.41–1.15) 0.48 (0.26–0.87) 0.57 (0.31–1.02) 0.45 (0.24–0.86) 0.0003 0.76 (0.58–1.01)
Noncardia 203 0.68 (0.45–1.02) 0.77 (0.51–1.17) 0.55 (0.33–0.91) 0.73 (0.43–1.23) 0.0464 0.86 (0.68–1.08)
Histological type4 Intestinal 154 0.55 (0.34–0.89) 0.48 (0.28–0.82) 0.51 (0.29–0.89) 0.62 (0.35–1.11) 0.065 0.78 (0.60–1.01)
Diffuse 157 0.87 (0.55–1.40) 1.03 (0.64–1.65) 0.64 (0.36–1.15) 0.84 (0.45–1.57) 0.360 0.92 (0.70–1.22)
TRAP Total 444 0.93 (0.70–1.22) 0.69 (0.51–0.95) 0.78 (0.56–1.06) 0.66 (0.46–0.95) 0.0010 0.86 (0.75–0.98)
Sex Male 255 0.74 (0.50–1.09) 0.87 (0.59–1.29) 0.79 (0.52–1.21) 0.72 (0.46–1.14) 0.2560 0.90 (0.76–1.06)
Female 189 1.16 (0.78–1.73) 0.48 (0.29–0.81) 0.77 (0.47–1.26) 0.62 (0.34–1.12) 0.0266 0.83 (0.63–1.02)
3
Localization site Cardia 130 0.78 (0.46–1.30) 0.50 (0.27–0.93) 0.62 (0.35–1.11) 0.51 (0.27–0.95) 0.0012 0.81 (0.64–1.02)
Noncardia 203 1.03 (0.70–1.52) 0.69 (0.44–1.09) 0.70 (0.43–1.14) 0.86 (0.51–1.46) 0.1034 0.91 (0.75–1.10)
4
Histological type Intestinal 154 0.80 (0.51–1.27) 0.53 (0.31–0.92) 0.60 (0.34–1.03) 0.70 (0.39–1.25) 0.097 0.83 (0.67–1.03)
Diffuse 157 1.02 (0.65–1.62) 0.87 (0.52–1.44) 0.92 (0.54–1.59) 0.88 (0.54–1.59) 0.588 0.97 (0.77–1.21)
1
All models adjusted by sex, educational level, smoking status, BMI, red meat and energy intake. 2HR (log2): hazard ratio after log2 transformation of the variable. 3Localization site unknown: 111.
4
Histological type unknown: 133.
FRAP: p Value for sex interaction: 0.6593/cardia and noncardia (X2wald ¼ 0.40, p value ¼ 0.527)/intestinal and diffuse (X2wald ¼ 1.21, p value ¼ 0.271).
TRAP: p Value for sex interaction: 0.4339/cardia and noncardia (X2wald ¼ 0.57, p value ¼ 0.449)/intestinal and diffuse (X2wald ¼ 0.1.30, p value ¼ 0.254).

Int. J. Cancer: 131, E544–E554 (2012) V

C 2011 UICC
Dietary antioxidant capacity and gastric cancer
Serafini et al.
Figure 1. Hazard ratio (HR) for gastric cancer and TAC intakes. HR has been modeled using a third-knot cubic regression spline. HRs are
indicated by the solid line and 95% confidence intervals by dashed blue line, with knots placed at the first, fifth and 90th percentiles of
the distribution of FRAP (a) and TRAP (b). The HRs was adjusted for sex, educational level, smoking status, BMI, red meat intake and total
energy intake.
(TRAP) and 23.0% (FRAP), represents the main contributor
to dietary antioxidant intake, with a large proportion
accounted by UK and The Netherlands (data not shown).
Fruit and vegetables combined account for 31.7% (TRAP)
and 32.6% (FRAP) of the antioxidant intake. Other relevant
sources of antioxidant were wine (12.6% TRAP and 9.4%
FRAP) and sugar and confectionery (10.2% TRAP and 8.5%
FRAP) with chocolate as main contributor (data not shown).
Minor source of antioxidants were cereals, potatoes, juices
and soft drinks. Soup and condiment were negligible contrib-
utors to TAC intake. Tea was the first contributor of dietary
antioxidants in men and women for both assays. In males,
wine replaced fruit as second source of dietary antioxidants
for TRAP (19.0% vs. 14.1%) with a similar contribution
(14.2% vs. 14.7%) for FRAP, while in women, fruit was
the second source of antioxidants followed by vegetables,
sugar/confectionary and wine.
Dietary intake of FRAP and TRAP, according to demo-
graphic, anthropometric and lifestyle characteristics are out-
lined in Table 3. Results show that TRAP and FRAP intake
is slightly lower in women compared to men, increase with
age and energy intake and decline from normal to obese sub-
jects. A positive trend was shown for educational level, with
the group with the highest level of education displaying a
higher TRAP (þ22.6%) and FRAP (þ18%) intake compared
to people with lowest level of education. In the nonsmokers
group, intake of TRAP and FRAP was slightly higher (þ6%
for both markers) compared to smokers.
The association between dietary consumption of TRAP
and FRAP and risk of GC is shown in Table 4. Statistically
Int. J. Cancer: 131, E544–E554 (2012) V
C 2011 UICC
E551
significant trends of risk reduction is observed for both
TRAP (p < 0.001) and FRAP (p < 0.0001) for increasing of
quintiles of TAC consumption. The association seems to
reach a threshold of effect at the third and fourth quintile for
TRAP and FRAP, respectively. This inverse association is
also significant in the continuous model (log2 HR 0.86; 95%
CI: 0.75–0.98, p ¼ 0.0233 and log2 HR 0.81; 95% CI: 0.69–
0.94, p ¼ 0.0079 and for doubling the intake), for TRAP and
FRAP, respectively.
When subtypes of GC were considered, a clear and signifi-
Epidemiology
cant inverse association was observed for both cardia cancer
(TRAP and FRAP) and for noncardia cancer (FRAP only).
An inverse but statistically nonsignificant association was
observed for diffuse and intestinal types, although the magni-
tude of the negative association seemed stronger for the
intestinal type.
Figures 1a and 1b show the HRs for GC according to
FRAP and TRAP intake: GC risk decreased steadily before
flattening at 3,500 mmol Trolox/day and 13,000 mmol Fe/day
for TRAP and FRAP, respectively. These values correspond
to the third and fourth quintiles of TRAP and FRAP intake
(Table 1), where a threshold of effect was observed (Table 4).
Other specific risk factors such as Hp infection, meat
intake and smoking status were considered for their associa-
tion with dietary antioxidants. There was no difference in the
effect of dietary TAC on GC risk, measured with both assays
on subjects infected and not infected by Hp (p for interac-
tions: p ¼ 0.645 for FRAP and p ¼ 0.588 for TRAP) (data
not shown). Interaction between FRAP and TRAP and
tobacco smoking was not significant (p ¼ 0.22 and p ¼ 0.28,
 E552
Figure 2. Hazard ratio (HR) for gastric cancer in relation to FRAP
(a) and TRAP (b) intake by smoking status categories (Q1 was
considered as reference category of FRAP and TRAP). The HRs were
adjusted for sex, educational level, BMI, red meat intake and total
energy intake§. ***p value for trend < 0.001; **p value for trend
<0.01; *p value for trend <0.05. §p interaction between
categories: 0.71 TRAP and 0.60 FRAP.
respectively) (data not shown). No association was observed
between meat and TAC intake (p for interactions: p ¼ 0.84
Epidemiology
for FRAP and p ¼ 0.74 for TRAP).
On the contrary, a clear effect was observed in smokers,
with a significant risk reduction of about 60% is the fifth quin-
tile of intake for both assays (p for trend <0.001 for FRAP;
p for trend <0.001 for TRAP), as displayed in Figures 2a and
2b. In former smokers, the significant protective association of
FRAP intake was reduced but still present (p < 0.05) (Fig. 2a)
and disappeared for TRAP (Fig. 2b). When never smokers
were considered, no significant association was observed for
both TRAP or FRAP and risk of GC.
Discussion
We showed that dietary TAC intake, assessed through TRAP
and FRAP assays, is associated with a reduction in the risk of
GC in the EPIC study. This cohort study, with a large number
of subjects from different European countries, is the first to
evaluate TAC intake from the majority of plant foods. The
novel aspect presented in this article, which give an additional
value to the simple analyses of fruit and vegetable intake, is the
Dietary antioxidant capacity and gastric cancer
assessment of the ‘‘overall’’ antioxidant intake, obtained with a
marker of antioxidant function and not with single levels of
antioxidants of the majority of plant foods such as spices, alco-
holic beverages, cereals, nuts, legumes and beverages giving a
more realistic estimation of the antioxidant intake. The choice
of utilizing a marker such as TAC, representative of the antiox-
idant activity of diet, give us the possibility to investigate the
importance of consuming antioxidant from a wider source of
food respect to the antioxidants from fruit and vegetables
only.12 The assessment of dietary intake from fruit and/or veg-
etables do not give any indication about the mechanism of
action, differently from the TAC analyses, providing informa-
tion about the involvement of dietary antioxidant. In this
sense, our findings support the importance of achieving the
wide array of antioxidants, not only from fruit and vegetables
but also from other dietary sources of plant origin. For the
future, the measurements of markers of TAC in biological flu-
ids, will allow to understand if the protective effect displayed
by dietary antioxidants, reflect an improvement in endogenous
antioxidant defenses and a reduction of GC risk.
We carried out a sensitivity analysis by excluding tea, the
highest dietary TAC contributor, and a similar association
between TRAP and FRAP and GC was observed (p for trend
excluding tea 0.001 for FRAP and 0.0016 for TRAP). Despite
this, it is still unclear if the association between milk and tea
impairs the biological functions of tea in vivo.24,25 The com-
mon practice of adding milk to tea in the UK and in the
Netherlands (countries where tea intake was highest), might
reduce the antioxidant effect of the beverage. However, sensi-
tivity analysis excluding these countries does not change the
significance of our findings.
The inverse association with dietary antioxidant had a
threshold, located in the fourth and on the third quintile of
intake for FRAP and TRAP. The existence of a potential
threshold effect for the protective effect of dietary antioxi-
dants suggests the importance of achieving the ‘‘right’’ and
not the ‘‘highest’’ intake of antioxidants. When subjects were
divided according to smoking status, the inverse association
between GC and TAC was related to a reduction of the risk
of GC mainly among current smokers. In former smokers,
the association is still present but less pronounced, whereas
in nonsmokers, the association is lacking, highlighting the
importance of an adequate antioxidant intake in subjects
whose free radical production is high. Homeostatic control
mechanisms might be taking place, which could explain our
results: in recent work by Block et al.,26 the efficiency of vita-
min C supplementation in reducing a marker of lipid oxida-
tion, isoprostane, is closely linked to the starting levels of the
subjects. In this view, the lack of specific risk factors related
to oxidative stress might allow the endogenous defenses of
the body to better cope with free radical formation without
the need to synergize with dietary antioxidants. On the con-
trary, when oxidative stress is ongoing, endogenous battery of
redox defenses require the contribution of dietary bioactive
components to maximize strategies of stress reduction.
Int. J. Cancer: 131, E544–E554 (2012) V
C 2011 UICC
Serafini et al. E553

Dietary antioxidants, such as polyphenols, have been
shown to have an important contribution to the in vitro
TAC of plant foods, where they are present in millimolar
concentration. However, the low bioavailability of dietary
polyphenols in vivo (1–5%),27 leading to plasma concentra-
tions not exceeding 1 lM, raises doubts about their antioxi-
dant efficacy in biological fluids.28 An alternative mechanism
of action, which might justify the high presence of polyphe-
nols in food but not in body fluids, is linked with a postpran-
dial antioxidant action within the stomach. A dietary regimen
high in lipids and energy induces a postprandial oxidative
and inflammatory stress, mediated by proinflammatory
cytokines such as tumor necrosis factor-a (TNF-a) and inter-
leukin (IL)-629 and oxidized lipids.30 The presence of antioxi-
dants-rich foods during a high-fat meal31 might provide a
battery of exogenous antioxidants, able to quench radical
species produced at the gastric level, synergizing with endoge-
nous antioxidants and providing a more efficient protection
against oxidative stress.
Hp infection affects early gastric carcinogenesis by induc-
ing chronic gastritis with an inflammatory and oxidative
response, impairing gastric secretion of antioxidants.32
Tissues from subjects infected with Hp have been shown to
contain more radicals than normal tissues.33 However, in
agreement with our earlier findings,12 we failed to show any
interaction between Hp infection and TAC intake, but this
could also be due to small numbers in our study.
There are some limitations in this study: both FRAP and
TRAP assays are water soluble techniques and do not take in
account the contributions of antioxidants from oils and lipo-
philics, potentially leading to an underestimation of the anti-
oxidant effect. The TAC database was developed in Italy and
reflects antioxidant values of Italian food items and might
have produced different values compared to a TAC database
developed in other countries. It was an explicit choice to
idants, that are present in the diet such as from fish or meat
has not been measured in our database. Nevertheless, all per-
formed analyses were adjusted by meat intake without any
significant change. Strengths of the study are the large sample
size of the EPIC study and by the validated and detailed die-
tary questionnaires allowing the use of specific information
from the TAC database, including about 150 food items. In
addition, the study is mostly based on confirmed adenocarci-
noma cases validated by a panel of pathologists.
In conclusion, we showed that a high dietary intake of
antioxidant capacity from different plant food sources is asso-
ciated with a reduced GC risk in European countries from
the multicenter EPIC study. The existence of a threshold
effect at higher levels of TAC intake suggests the necessity to
identify the optimal antioxidant intake to avoid unnecessary
overloading. The effect of dietary antioxidants is more evi-
dent in subjects where specific risk factors linked to oxidative
stress (smoking) are present. Our results put new emphasis
on the role of dietary antioxidants in GC prevention. How-
ever, further research is warranted, including assessment of
biomarkers of TAC and oxidative stress in biological fluids,
to support nutritional strategies based on antioxidant equiva-
lents for GC prevention.
Acknowledgements
The authors thank the following pathologists for their valuable work on the
EURGAST pathology panel and/or for the collection of pathology material:
Johan Offerhaus, Amsterdam, Netherlands; Vicki Save and Laszlo Igali,
Cambridge, United Kingdom; Julio Torrado, San Sebastian, Spain; Gabriella
Nesi, Firenze, Italy; U Mahlke, Potsdam, Germany; Hendrik Bläker, Heildel-
berg; Germany; Claus Fenger, Denmark, Sonja Steigen, Tromso, Norway;
Dimitrious Roukos, Ioannina, Greece. The authors also thank the following
collaborators for their help with the collection of new pathology material:
Anna Zawadzka, Oxford, United Kingdom; Jutta Kneisel, Heidelberg,
Germany; Wolfgang Fleischhauer, Potsdam, Germany; Tine Plato, Hansen,
Denmark; and Åsa Ågren, Sweden. They also thank Catia Moutinho and
Bárbara Gomes (Porto, Portugal) for their technical work in the preparation

Epidemiology
include foods of plant origin only, as other sources of antiox- of pathologic material for the pathology panel.

References
1. World Cancer Research Fund/American Institute
for Cancer Research. Food, nutrition, physical
activity, and the prevention of cancer: a global
perspective. Washington, DC: AICR, 2007.
2. Correa P, Piazuelo MB, Camargo MC. The future
of gastric cancer prevention. Gastric Cancer 2004;
7:9–16.
3. IARC handbooks of cancer prevention, Vol. 8.
Fruits and vegetables. Lyon, France: IARC Press,
2003.
4. Nouraie M, Pietinen P, Kamangar F, Dawsey SM,
Abnet CC, Albanes D, Virtamo J, Taylor PR.
Fruits, vegetables, and antioxidants and risk of
gastric cancer among male smokers. Cancer
Epidemiol Biomarkers Prev 2005;14:2087–92.
5. González CA, Pera G, Agudo A, Bueno-de-
Mesquita HB, Ceroti M, Boeing H, Schulz M, Del
Giudice G, Plebani M, Carneiro F, Berrino F,
Sacerdote C, et al. Fruit and vegetable intake and
the risk of stomach and oesophagus
adenocarcinoma in the European Prospective
Investigation into Cancer and Nutrition (EPIC-
EURGAST). Int J Cancer 2006;118:2559–66.
6. Bjelakovic G, Nikolova D, Simonetti RG, Gluud
C. Antioxidant supplements for prevention of
gastrointestinal cancers: a systematic review and
meta-analysis. Lancet 2004;364:1219–28.
7. Ekström AM, Serafini M, Nyren O, Hansson LE,
Ye W, Wolk A. Dietary antioxidant intake and
the risk of cardia cancer and non-cardia cancer
of the intestinal and diffuse types—a population-
based case-control study in Sweden. Int J Cancer
2000;87:133–40.
8. Serafini M, Villano D, Spera G, Pellegrini N.
Redox molecules and cancer prevention: the
importance of understanding the role of the
antioxidant network. Nutr Cancer 2006;56:
232–40.
9. Pitsavos C, Panagiotakos DB, Tzima N,
Chrysohoou C, Economou M, Zampelas A,
Stefanadis C. Adherence to the Mediterranean
diet is associated with total antioxidant capacity
in healthy adults: the ATTICA study. Am J Clin
Nutr 2005;82:694–9.
10. Rautiainen S, Serafini M, Morgenstern R, Prior
RL, Wolk A. The validity and reproducibility of
food-frequency questionnaire-based total
antioxidant capacity estimates in
Swedish women. Am J Clin Nutr 2008;87:
1247–53.
11. Razquin C, Martinez JA, Martinez-Gonzalez MA,
Mitjavila MT, Estruch R, Marti A. A 3 years
follow-up of a Mediterranean diet rich in virgin
olive oil is associated with high plasma
antioxidant capacity and reduced body weight
gain. Eur J Clin Nutr 2009;63:1387–93.

Int. J. Cancer: 131, E544–E554 (2012) V

C 2011 UICC
 E554
12. Serafini M, Bellocco R, Wolk A, Ekström AM.
Total antioxidant potential of fruit and vegetables
and risk of gastric cancer. Gastroenterology 2002;
123:985–91.
13. Agudo A, Cabrera L, Amiano P, Ardanaz E,
Barricarte A, Berenguer T, Chirlaque MD,
Dorronsoro M, Jakszyn P, Larrañaga N, Martı́nez
C, Navarro C, et al. Fruit and vegetable intakes,
dietary antioxidant nutrients, and total mortality
in Spanish adults: findings from the Spanish
cohort of the European Prospective Investigation
into Cancer and Nutrition (EPIC-Spain). Am J
Clin Nutr 2007;85:1634–42.
14. Pellegrini N, Serafini M, Colombi B, Del Rio D,
Salvatore S, Bianchi M, Brighenti F. Total
antioxidant capacity of plant foods, beverages and
oils consumed in Italy assessed by three different
in vitro assays. J Nutr 2003;133:2812–9.
15. Riboli E, Hunt KJ, Slimani N, Ferrari P, Norat T,
Fahey M, Charrondière UR, Hemon B,
Casagrande C, Vignat J, Overvad K, Tjønneland
A, et al. European Prospective Investigation into
Cancer and Nutrition (EPIC): study populations
and data collection. Public Health Nutr 2002;5:
1113–24.
16. Margetts BM, Pietinen P. European Prospective
Investigation into Cancer and Nutrition: validity
studies on dietary assessment methods. Int J
Epidemiol 1997;26(suppl 1):S1–S5.
17. Carneiro F, Moutinho C, Pera G, Caldas C,
Fenger C, Offerhaus J, Save V, Stenling R, Nesi
G, Mahlke U, Bläker H, Torrado J, et al.
Pathology findings and validation of gastric and
esophageal cancer cases in a European cohort
(EPIC/EUR-GAST). Scand J Gastroenterol 2007;
42:618–27.
Epidemiology
Dietary antioxidant capacity and gastric cancer
18. Pellegrini N, Serafini M, Salvatore S, Del Rio D,
Bianchi M, Brighenti F. Total antioxidant
capacity of spices, dried fruits, nuts, pulses,
cereals and sweets consumed in Italy assessed by
three different in vitro assays. Mol Nutr Food Res
2006;50:1030–8.
19. Delgado-Andrade C, Morales FJ. Unraveling the
contribution of melanoidins to the antioxidant
activity of coffee brews. J Agric Food Chem 2005;
53:1403–7.
20. Morales FJ, Somoza V, Fogliano V. Physiological
relevance of dietary melanoidins. Amino Acids
2010. DOI: 10.1007/s00726-010-0774-1.
21. Palli D, Masala G, Del Giudice G, Plebani M,
Basso D, Berti D, Numans ME, Ceroti M, Peeters
PH, Bueno de Mesquita HB, Buchner FL, Clavel-
Chapelon F, et al. CagAþ Helicobacter pylori
infection and gastric cancer risk in the EPIC-
EURGAST study. Int J Cancer 2007;120:859–67.
22. Armitage P, Berry G, Matthews JNS. Statistical
methods in medical research, 4th edn. Oxford:
Blackwell Publishers, 2001.
23. Harrell FE. Regression modelling strategies: with
applications to linear models, logistic regression
and survival analysis. New York: Springer, 2001.
24. Serafini M, Ghiselli A, Ferro-Luzzi A. In vivo
antioxidant effect of green and black tea in man.
Eur J Clin Nutr 1996;50:28–32.
25. Lorenz M, Jochmann N, von Krosigk A, Martus P,
Baumann G, Stangl K, Stangl V. Addition of milk
prevents vascular protective effects of tea. Eur Heart
J 2007;28:219–23.
26. Block G, Jensen CD, Morrow JD, Holland N,
Norkus EP, Milne GL, Hudes M, Dalvi TB,
Crawford PB, Fung EB, Schumacher L, Harmatz
P. The effect of vitamins C and E on biomarkers
Int. J. Cancer: 131, E544–E554 (2012) V
of oxidative stress depends on baseline levels.
Free Radic Biol Med 2008;45:377–84.
27. Manach C, Williamson G, Morand C, Scalbert A,
Remesy C. Bioavailability and bioefficacy of
polyphenols in humans. I. Review of 97
bioavailability studies. Am J Clin Nutr 2005;81:
230S–42S.
28. Serafini M, Testa MF, Villaño D, Pecorari M, van
Wieren K, Azzini E, Brambilla A, Maiani G.
Antioxidant activity of blueberry fruit is impaired
by association with milk. Free Radic Biol Med
2009;46:769–74.
29. Adibhatla RM, Dempsy R, Hatcher JF.
Integration of cytokine biology and lipid
metabolism in stroke. Front Biosci 2008;13:
1250–70.
30. Lamb RE, Goldstein BJ. Modulating an oxidative-
inflammatory cascade: potential new treatment
strategy for improving glucose metabolism,
insulin resistance, and vascular function. Int J
Clin Pract 2008;62:1087–95.
31. Gorelik S, Ligumsky M, Kohen R, Kanner J. A
novel function of red wine polyphenols in
humans: prevention of absorption of cytotoxic
lipid peroxidation products. FASEB J 2008;22:
41–6.
32. Correa P, Malcom G, Schmidt B, Fontham E,
Ruiz B, Bravo JC, Bravo LE, Zarama G, Realpe
JL. Review article: antioxidant micronutrients and
gastric cancer. Aliment Pharmacol Ther 1998;
12(suppl 1):73–82.
33. Davies GR, Banatvala N, Collins CE, Sheaff MT,
Abdi Y, Clements L, Rampton DS. Relationship
between infective load of Helicobacter pylori and
reactive oxygen metabolite production in antral
mucosa. Scand J Gastroenterol 1994;29:419–24.
C 2011 UICC