Tesis RomanosNanclares21

School of Medicine
DIETARY FACTORS AND BREAST CANCER

RISK IN TWO COHORT STUDIES
ANDREA ROMANOS NANCLARES
PAMPLONA, 2021
Facultad de Medicina
Departmento de Medicina
Preventiva y Salud Pública
Pamplona, a 29 de marzo de 2021
Dña. ESTEFANÍA A. TOLEDO ATUCHA, PROFESORA TITULAR DEL

DEPARTAMENTO DE MEDICINA PREVENTIVA Y SALUD PÚBLICA DE LA
UNIVERSIDAD DE NAVARRA,
D. ALFREDO GEA SÁNCHEZ, PROFESOR CONTRATADO DOCTOR DEL

DEPARTAMENTO DE MEDICINA PREVENTIVA Y SALUD PÚBLICA DE LA
UNIVERSIDAD DE NAVARRA,
CERTIFICAN
Que Dña. ANDREA ROMANOS NANCLARES ha realizado el estudio “dietary

factors and breast cancer risk in two cohort studies" que aquí se presenta para optar al
grado de Doctor en Medicina bajo nuestra dirección. El trabajo refleja la capacidad
técnica e interpretativa tan aventajada de su autor que le hacen merecedor del título de
Doctora en Medicina, siempre a juicio del Tribunal designado para tal efecto por la
Universidad de Navarra.
Dra. Estefanía Toledo Atucha Dr. Alfredo Gea Sánchez
Directora Codirector
TABLE OF CONTENTS
ACKNOWLEDGMENTS ...............................................................................................3
ABBREVIATIONS .........................................................................................................6
INTRODUCTION .........................................................................................................18
1. Anatomy of the breast ....................................................................................................... 19
2. Breast carcinogenesis ...................................................................................................... 21
3. Breast Cancer Subtypes ................................................................................................... 26
4. Epidemiology of breast cancer ....................................................................................... 32
5. Risk factors .......................................................................................................................... 34
HYPOTHESIS AND AIMS .......................................................................................... 60
1. Hypothesis ............................................................................................................................ 61
2. General aim .......................................................................................................................... 61
3. Specific aims ........................................................................................................................ 61
METHODS ................................................................................................................... 63
1. SUN Project .......................................................................................................................... 64
2. Nurses’ Health Studies ...................................................................................................... 81
RESULTS .................................................................................................................... 94
1. Sugar-sweetened beverage consumption and breast cancer risk in the SUN
project .............................................................................................................................................. 96
2. Sugar-sweetened, artificially sweetened beverages and breast cancer risk in
the Nurses’ Health Studies ........................................................................................................ 99
3. Phenolic acids subclasses, individual compounds and breast cancer risk in the
SUN project .................................................................................................................................. 109
4. Fat intake and breast cancer risk in the SUN project. ............................................ 114
5. Carbohydrate quality index and breast cancer risk in the SUN project ............ 119
6. Healthful and unhealthful provegetarian food patterns and breast cancer risk in
the SUN project ........................................................................................................................... 124
7. Healthful and unhealthful plant-based diets and risk of breast cancer in the
Nurses’ Health Studies ............................................................................................................. 131
DISCUSSION ............................................................................................................ 139
1. SUN Project ........................................................................................................................ 140
2. Nurses’ Health Studies .................................................................................................... 140
project and Nurses’ Health Studies ....................................................................................... 140
4. Phenolic acids subclasses, individual compounds and breast cancer risk in the
SUN Project. ................................................................................................................................. 143
5. Fat intake and breast cancer risk in the SUN project. ............................................ 145
6. Carbohydrate quality index and breast cancer risk in the SUN project. ........... 146
7. Healthful and unhealthful provegetarian food patterns and breast cancer risk in
the SUN project and Nurses’ Health Studies ...................................................................... 148
8. Limitations and strengths .............................................................................................. 151
9. Public health interest and future directions .............................................................. 159
BIBLIOGRAPHY ....................................................................................................... 168
APPENDICES ........................................................................................................... 204
2
ACKNOWLEDGMENTS
3
I would like to thank my parents for their wise counsel and sympathetic ear, for
encouring me every step of the way and being always there for me. I am incredibly
grateful for your love and support throughout this experience. I will always cherish the
bond we have.
I also want to acknowledge the invaluable insight and support I have received
from my brother and best friend, Alvaro, whose ambition and eagerness to pursuit his
Medical Residency is unquestionable, he will achieve everything he sets his mind to.
I extend my sincerest appreciation to my supervisor, Dr. Estefania Toledo, for her

dedication, guidance and support and for making this process an enriching and valuable
experience. Your insightful, careful and punctilious feedback, your extensive expertise
and valuable counselling throughout my PhD pushed me to sharpen my thinking and
brought my work to a higher level.
I am grateful to my supervisor, Dr. Alfredo Gea, whose patience, serenity and

empathy are particularly noteworthy. Thanks for providing guidance, being so caring and
a sounding board when required. Our meetings and discussions on statistical and
methodological issues were crucial in inspiring me to think outside the box, I learned a
lot from you.
I would like to thank Dr. Miguel Angel Martinez-Gonzalez for being always
mindful, caring and supportive and providing me additional advice, encouragement,
mentorship and direction.
The Department of Preventive Medicine has been my workplace since

September 2017, and I thank my colleagues for a creative and stimulating work
environment. I am grateful to have counted on Clara, Sol and Alice during this long
reaching and intense period, I thank them for their sincere and priceless friendship.
This thesis could never have been written without the more than 10,000 women
who answered the SUN Project questionnaires - their contribution is much appreciated.
I also thank Dr. Toledo, Dr. Gea and Dr. Martinez-Gonzalez for the opportunity
they gave me to move to Boston and expand my research within the Nurses’ Health
Studies, which has undoubtedly marked a before and after in my career.
4
I would like to single out and thank my supervisor at Harvard, Dr. Heather
Eliassen for being an outstanding mentor and researcher, committed and invested in my
growth. Her dynamism, vision, empathy, support and motivation have deeply inspired
me. The amount that I am learning under her guidance is immeasurable and I could not
be more grateful for all that she has walked me through and what we are planning for
the near future.
Many thanks to my folks from the Breast Cancer Working Group at Harvard for
welcoming me from the really beginning and counting on me for interest discussions,
project updates and brainstormings. I would also like to extend my gratitude to Dr. Walter
Willett for his precious feedback and good words and Dr. Meir Stampfer for being so
caring, humble and emphatetic.
Despite the extraordinary circumstances as a visiting grad student engulfed in

the COVID-19 shutdown in Boston, I am thankful for all the people that lifted me up while
being in isolation for quite a few months miles away from home. I thank my family, my
supervisors (Estefania, Alfredo, Miguel Ángel and Heather) and friends scattered around
the world. Thank you for your thoughts, support, well-wishes, constant care, invaluable
advices, prayers, video and phone calls, and being there when I needed it most.
Particularly, I’d like to express my deepest gratitude to Clara, Sol, Alice, Carol, Carmen,
Marina and Alessio, I will be eternally grateful for your untiring support, to keep me
upbeat and positive.
I would like to express my sincere gratitude to the Spanish Association Against

Cancer for granting me with one Predoctoral fellowship to conduct this project.
Last but not least, I would like to extend my heartfelt gratitude to my grandparents
-Nela, Pepe and Amparo- for their enduring support and love.
5
ABBREVIATIONS
6
AICR: American Institute for Cancer Research
AHEI: Alternative Healthy Eating Index
ASB: Artificially sweetened beverages
BMI: Body Mass Index
BRK: Breast Tumor Kinase
CI: Confidence Interval
CDK: Cyclin Dependent Kinases
CDKI: Cyclin Dependent Kinases Inhibitor
CQA: Caffeoylquinic Acids
CQI: Carbohydrate Quality Index
CSC: Cancer Stem Cells
DCIS: Ductal Carcinoma In Situ
DII: Delta-Like
EPIC: European Prospective Investigation into Cancer and Nutrition
ER: Estrogen Receptor
HEI: Healthy Eating Index
HER2: Human epidermal growth factor receptor 2
HR: Hazard Ratio
HRT: Hormone Replacement Therapy
hPDI: healthful plant-based diet index
hPVG: healthful provegetarian
IARC: International Agency for Research on Cancer
IBC: Inflammatory Breast Cancer
IDC: Invasive Ductal Carcinoma
ILC: Invasive Lobular Carcinoma
MAPK: Mitogen-Activated Protein Kinase
MET: Metabolic Equivalent of Task
NHS: Nurses’ Health Study
NHSII: Nurses’ Health Study II
NIH-AARP: National Institute of Health- American Association of Retired Persons
OR: Odds Ratio
PDI: Plant-based diet index
PI3K: Phosphatidylinositol 4,5-Isphosphate 3-Kinase
PR: Progesterone Receptor
PREDIMED: PREvención con DIeta MEDiterránea
PURE: Prospective Urban Rural Epidemiology
7
PVG: Provegetarian
RR: Relative Risk
RCT: Randomized Controlled Trial
SSB: Sugar-Sweetened Beverages
SUN: Seguimiento Universidad de Navarra
TMAs: Tissue Microarray
TNM: Tumor, Node, and Metastasis
uPDI: unhealthful plant-based diet index
uPVG: unhealthful provegetarian
VITAL: Vitamins and Lifestyle
WCRF: World Cancer Research Fund
8
LIST OF TABLES AND FIGURES
9
TABLES
Table 1. Anatomic stage groups of breast cancer.
Table 2. Scoring criteria for the provegetarian (PVG) food pattern.
Table 3. Criteria used to calculate carbohydrate quality in the SUN Project (1999-2018).
Table 4. Examples of food items constituting the 18 food groups (from the 1984
Nurses’Health Study Food Frequency Questionnaire).
Table 5. Descriptive characteristics of women with a confirmed diagnosis of breast
cancer and the total sample of women from the “Seguimiento Universidad de Navarra”
(SUN) Project.
Table 6. Baseline characteristics of 10,713 women in the SUN cohort according to the
categories of sugar-sweetened beverage (SSB) consumption.
Table 7. Hazard Ratio (HR) and 95% confidence intervals (CI) of confirmed breast
cancer cases according to the categories of baseline SSB consumption in 10,713 women
of the SUN Project.
Table 8. Hazard ratio (HR) and 95% confidence intervals (CI) of confirmed breast cancer
cases for each category of SSB consumption among premenopausal and
postmenopausal women of the SUN Project.
Table 9. Age and age-standardized baseline characteristics of women according to
consumption of sugar-sweetened beverages at baseline (NHS,1980; NHSII, 1991).
consumption of artificially sweetened beverages at baseline (NHS,1980; NHSII, 1991).
Table 11. Total breast cancer incidence according to cumulative average intake of SSBs
and ASBs in the NHS, NHSII and pooled cohorts.
Table 12. Total breast cancer incidence according to cumulative average intake of SSBs
and ASBs in the NHS, NHSII and pooled cohorts.
Table 13. Breast cancer incidence by estrogen receptor status and molecular subtypes
according to cumulative average intake of SSBs using pooled data from the Nurses'
Health Studies (NHS and NHSII).
Table 14. Breast cancer incidence by estrogen receptor status and molecular subtypes
according to cumulative average intake of ASBs using pooled data from the Nurses'
Health Studies (NHS and NHSII).
Table 15. Multivariable hazard ratios (HRs) and 95% confidence intervals (CIs) for the
association between categories of cumulatively updated SSB and ASB intake and
premenopausal or postmenopausal breast cancer risk using pooled data from the
Nurses’ Health Studies (NHS and NHSII).
Table 16. Multivariable hazard ratios (HRs) and 95% confidence intervals (CIs) for the
association between categories of cumulatively updated SSB intake and premenopausal
10
or postmenopausal breast cancer risk stratified by BMI using pooled data from NHS and
NHSII.
Table 17. Baseline characteristics of 10,812 women according to tertiles of phenolic acid
intakes in the SUN Project, 1999-2018.
Table 18. Main sources of variability (cumulative R2 and change in R2) and main sources
of quantity (%) of total phenolic, hydroxycinnamic, hydroxybenzoic and chlorogenic acid
intake according to food groups. The SUN Cohort, 1999-2018.
Table 19. Hazard Ratios (HRs) and confidence intervals (95 % CIs) for the association
between tertiles of individual phenolic acid intake and confirmed breast cancer risk in the
SUN Cohort.
Table 20. Hazard Ratios (HRs) and confidence intervals (95 % CIs) for the association
between tertiles of chlorogenic acids (5-CQI, 4-CQI and 3-CQI) intake and confirmed
breast cancer risk in the SUN Cohort.
Table 21. Baseline characteristic for the entire cohort at enrolment and by tertiles of total
fat intake as a percentage of energy among 10,812 women in the SUN cohort.
Table 22. Energy- and multivariable-adjusted hazard ratios (HRs) and 95% confidence
intervals (CIs) for breast cancer risk in association with total fat intake using the density
energy adjustment among women in the SUN Cohort.
Table 23. Energy- and multivariable-adjusted hazard ratios (HRs) and 95% confidence
intervals (CIs) for breast cancer risk in association with subtypes of fat intake using the
density energy adjustment among women in the SUN Cohort.
Table 24. Estimated HRs (95% CIs) of breast cancer risk associated with isocaloric
substitutions (5% energy increment) of one dietary component for another.
Table 25. Baseline characteristics of participants according to quartiles of the
carbohydrate quality index: the SUN Project, 1999-2018.
Table 26. Hazard ratios and 95% confidence intervals for confirmed and probable
a
incident breast cancer by quartiles of carbohydrate quality index in the SUN Project,
1999-2018.
Table 27. Hazard ratios and 95% confidence intervals for confirmed incident breast
cancer by tertiles of individual components of CQI and glycemic load in the SUN Project,
1999-2018.
Table 28. Baseline characteristics of participants according to tertiles of the
provegetarian food pattern, the SUN cohort, 1999-2018.
Table 29. Hazard Ratio (HR) and 95% confidence intervals (CI 95%) according to tertiles
of the provegetarian (PVG) food pattern and a healthful and unhealthful provegetarian
food patterns in the overall sample of women from the SUN cohort.
11
quintiles of an overall plant-based diet index (PDI) in the Nurses’ Health Study (NHS)
and NHSII.
Table 31. Multivariable-adjusted hazard ratios (95% confidence intervals) for the
association between quintiles (Q) of cumulatively updated PDI, hPDI and uPDI and
breast cancer tumor subtypes in the NHS and NHSII.
Table 32. Multivariable-adjusted hazard ratios (95% confidence intervals) for the
association between quintiles of cumulatively updated hPDI and breast cancer tumor
subtypes in the NHS and NHSII.
Table 33. Risk of ER-negative breast cancer according to PDI, hPDI, uPDI with different
latency periods using pooled data from NHS and NHSII.
FIGURES
Figure 1. Breast anatomy.
Figure 2. Quadrants of left breast and axillary tail of Spence.
Figure 3. Microenvironmental alterations during tumor progression. Schematic view of
normal breast, DCIS (ductal carcinoma in situ), IDC (invasive ductal carcinoma), and
metastatic breast carcinoma.
Figure 4. Two hypothetical theories of breast cancer initiation and progression. (A)
Cancer stem cell theory. (B) Stochastic theory.
Figure 5. Invasive Ductal Carcinoma subtypes.
Figure 6. Common metastatic sites in breast cancer.
Figure 7. Summary of breast cancer subtypes.
Figure 8. Map showing the most frequent tumors in Spain (Figure 8A) and in the US
(Figure 8B) in 2020 for all cancer sites.
Figure 9. Hereditary conditions or gene mutations that can increase a person’s risk of
breast cancer.
Figure 10. Known risk factors that vary by race and ethnicity.
Figure 11. Fraction (%) of all cancer cases (at all anatomical sites) among both sexes
(worldwide) in 2012 attributable to excess body mass index, by country.
Figure 12. Timeline of data collection in the Nurses’Health Studies.
Figure 13. Total breast cancer according to SSB intake (servings/day) and ASB intake
stratified by BMI, physical activity, and diet quality (AHEI score) based on pooled data
from both cohorts (NHS/NHSII).
Figure 14. Estrogen receptor positive and Luminal A breast cancer risk according to
SSB intake (servings/day) and ASB intake stratified by BMI based on pooled data from
both cohorts.
12
Figure 15. Main components of total polyphenols (%) according to their chemical
structure (A) and their main subclasses and individual compounds (B) in a sample of
10,812 women from the SUN cohort (1999-2018).
Figure 16. Hazard ratios and 95% confidence intervals for confirmed (a) and probable
(b) incident premenopausal and postmenopausal breast cancer by quartiles of CQI.
Figure 17. Sensitivity analyses. Hazard ratios (HRs) and 95% confidence intervals (CIs)
of incident breast cancer for the fourth quartile of CQI compared with the first quartile.
Figure 18. Restricted cubic splines for the hazard ratio (HR) and 95% confidence
intervals (CI) of adherence to the Provegetarian (PVG) food pattern and incidence of
breast cancer in the SUN cohort.
Figure 19. Hazard Ratios (HR) and 95% confidence intervals (CI) of breast cancer
incidence according to a a) healthful and an b) unhealthful provegetarian food pattern in
the SUN cohort.
Figure 20. Restricted cubic splines for the hazard ratio (HR) and 95% confidence
intervals (CI) of adherence to a provegetarian (PVG) food pattern and incidence of (a)
premenopausal and (b) postmenopausal breast cancer in the SUN cohort.
Figure 21. Age-adjusted and multivariable-adjusted hazard ratios (95% confidence
intervals) for total breast cancer according to quintiles of plant-based diet indices (PDI,
hPDI, uPDI) in the NHS and NHSII.
Figure 22. Multivariable spline analysis of the association between adherence to a
healthful (A) and unhealthful (B) plant-based dietary indeces and risk of incident breast
cancer in the NHS (1984-2016) and NHSII (1991-2017).
Figure 23. Risk of ER-negative breast cancer according to animal, healthy plant and
less healthy plant foods.
Figure 24. Pooled hazard ratios of estrogen receptor negative breast cancer per 10-
units increment in the three dietary indices (PDI, hPDI and uPDI) across subgroups
(physical activity, current BMI and menopausal status).
BOXES
Box 1. Less common types of breast cancer.
13
RESUMEN
14
Globalmente, el cáncer de mama es el tumor más diagnosticado en la mujer, la
primera causa de mortalidad por cáncer en ésta y un problema de salud pública
preponderante a nivel mundial. El 5-10% de todos los tumores se deben a
predisposiciones genéticas (factores no modificables), el 90-95% se atribuyen a factores
ambientales y de estilos de vida (factores modificables), como el consumo de alcohol,
patrones dietéticos insanos y un estado de obesidad derivado de conductas
desfavorables.
De todos ellos, los factores relacionados con la dieta se consideran de gran

trascendencia debido a la continua exposición a la que puede estar sometida la
población, y en este caso las mujeres, a lo largo de su vida. La evidencia epidemiológica
muestra resultados heterogéneos en la relación entre la dieta y el cáncer de mama,
aunque se cree que podría desempeñar un papel protector, tanto en la etiología como
incluso a la hora de modular el desarrollo del tumor. La asociación inconsistente entre
alimentos y nutrientes y el riesgo de cáncer de mama puede deberse, en parte, a que
las personas no consumen alimentos o nutrientes de forma aislada sino una
combinación de estos dentro de patrones dietéticos.
En este contexto, el objetivo global de la presente tesis doctoral fue valorar

prospectivamente la asociación de algunos factores modificables relacionados con la
dieta sobre la incidencia de cáncer de mama en el estudio de cohortes español Proyecto
SUN (Seguimiento Universidad de Navarra) y en las cohortes americanas Nurses’
Health Studies o Estudios de las Enfermeras.
El proyecto SUN es una cohorte prospectiva, dinámica y multipropósito iniciada

en 1999, en la que se han incluido y seguido más de 22,500 graduados universitarios
durante una media de 10,3 años, de las cuales aproximadamente 10,000 son mujeres.
En el cuestionario basal se recogió información sobre hábitos dietéticos con un
cuestionario de frecuencia de consumo de alimentos, además de otras variables
antropométricas y de estilo de vida. Mediante cuestionarios de seguimiento cada dos
años se identificaron casos incidentes de cáncer de mama, que fueron confirmados
posteriormente con informes médicos. La relación entre las distintas variables de
exposición y la incidencia de cáncer de mama se estudió con modelos de regresión de
Cox ajustados por múltiples factores de confusión.
15
Además, en la presente tesis doctoral, hemos utilizado datos del Nurses’ Health
Study I que incluye 121,700 enfermeras (de 30 a 55 años) que comenzaron a formar
parte del estudio en 1976, y del Nurses’ Health Study II que ha seguido a 116,429
enfermeras (de 25 a 42 años) desde 1989. Ambos son estudios de cohortes
prospectivos. Cada dos años, los participantes proporcionan información sobre factores
relacionados con la salud, nutrición, estilo de vida y antecedentes médicos. Se ajustaron
modelos de regresión de Cox con medidas repetidas para estudiar la asociación entre
la dieta y la incidencia de cáncer de mama.
En el Proyecto SUN, una adhesión moderada a un patrón provegetariano, donde

se promueve el consumo de alimentos de origen vegetal junto con reducciones
compensatorias en el consumo de alimentos de origen animal, se asoció a una menor
incidencia de cáncer de mama. Por otro lado, en los Nurses’ Health Studies, una versión
saludable de un patrón dietético basado en alimentos de origen vegetal se asoció con
una menor incidencia de cáncer de mama, sobre todo, con receptores de estrógenos
negativos.
Existen alimentos que son ampliamente consumidos dentro de patrones

dietéticos que merecen ser evaluados de forma individual. Un ejemplo de ello son las
bebidas azucaradas, fuente de carbohidratos simples (principalmente sacarosa), y
promotor de numerosas patologías cardiometabólicas. En el Proyecto SUN, el consumo
frecuente de bebidas azucaradas entre mujeres postmenopáusicas se asoció a una
mayor incidencia de cáncer de mama en comparación con las que no consumían estas
bebidas o lo hacían con una frecuencia inferior a una vez al mes.
En línea con lo anterior, se debería de poner más énfasis en la calidad de los

carbohidratos que se ingieren más que en la cantidad. En el Proyecto SUN, una mayor
calidad de los carbohidratos ingeridos, es decir, aquellos con mayor contenido en fibra
dietética, con menor índice glucémico, que dan preferencia a carbohidratos sólidos
frente a los líquidos y de grano entero o integral, se asoció a un menor riesgo de padecer
cáncer de mama, sobre todo, en mujeres premenopáusicas.
Los polifenoles son responsables de algunas de las propiedades de los

alimentos que los contienen, como sus cualidades sensoriales y nutricionales. A pesar
de no ser elementos esenciales para el bienestar a corto plazo (como sí lo son las
vitaminas), se ha observado que tienen efectos favorables sobre la prevención y
tratamiento de enfermedades crónicas como el cáncer. Por todo ello, una dieta rica en
16
determinados polifenoles podría determinar una menor incidencia de cáncer de mama.
En el Proyecto SUN, observamos una asociación inversa entre la ingesta de ácidos
hidroxicinámicos y, concretamente, clorogénicos, ampliamente presentes en café, frutas
y verduras, con el riesgo de cáncer de mama postmenopáusico.
Por último, numerosos estudios experimentales y epidemiológicos han puesto de

manifiesto la relación existente entre las grasas o lípidos de la dieta y el cáncer de
mama, sin embargo, esta posible asociación sigue siendo controvertida y sujeta a
cuestiones metodológicas. En el Proyecto SUN, no encontramos evidencia de una
asociación entre el total de grasa dietética o sus subtipos (monoinsaturadas,
poliinsaturadas o saturadas) con la incidencia de cáncer de mama.
En general, los resultados de esta tesis doctoral aportan nueva evidencia del
papel de la dieta sobre la incidencia de cáncer de mama en dos estudios de cohortes.
Finalmente, cabe insistir en la necesidad de llevar a cabo más estudios en grandes
cohortes que avalen los resultados de nuestro trabajo para la prevención primaria del
cáncer de mama.
17
INTRODUCTION
18
1. Anatomy of the breast
The female breast lies on the anterior thoracic wall with the base extending from
the second to the sixth rib1. The breast is made up of skin, subcutaneous tissue, and
breast tissue. There are two fascial layers: the superficial fascia lies deep to the dermis
and the deep fascia lies anterior to the pectoralis major muscle fascia 2. The breast tissue,
which includes epithelial parenchymal elements and the stroma, lies in the superficial
fascia just deep to the dermis. Breast tissue is attached to the fascia of the pectoralis
major muscle by the suspensory ligaments of Cooper. The epithelial parenchymal
component takes about 10% to 15% of the overall breast volume, and remainder of the
volume is consisted of the stromal elements.
The breast is composed of 15 to 20 lobes 3. The lobes of the breast are further
divided into lobules, which range from 20 to 40. The lobules are made up of branched
tubuloalveolar glands. Each lobe drains into a major lactiferous duct. The lactiferous
ducts dilate into a lactiferous sinus beneath the areola and then open through a
constricted orifice onto the nipple. The space between the lobes is filled by adipose tissue
(Figure 1).
Figure 1. Breast anatomy. (Taken from Harbeck et al.4).
19
The breast is divided into quadrants (Figure 2): upper inner, upper outer, lower
inner, and lower outer quadrants with the tail of Spence as an extension of the upper
lateral quadrant of the breast into the axila. The majority of the breast volume lies in the
upper outer quadrant and the most common breast tumors are located in this quadrant.
Figure 2. Quadrants of left breast and axillary tail of Spence. (Taken from Seidel HM et
al.5).
Nipple-areola complex
The nipple is located over the fourth intercostal space in a non-pendulous breast
and is surrounded by a round pigmented areola (Figure 1). The areola contains
sebaceous glands and apocrine sweat glands. Throughout puberty, the pigment
becomes darker, and the nipple elevates from the surface, becoming more prominent.
The sebaceous glands enlarge during pregnancy. Moreover, the areola contains in its
periphery some nodular elevations known as tubercles of Morgagni formed by the
openings of the Montgomery glands which can secrete milk and represent a stage
between sweat and mammary glands.
Arterial and Venous Supply

The breast receives its blood supply through 3 major arterial routes6:
1) Branches of the internal thoracic artery, or internal mammary artery, which
supply the medial and central portion of the breast accounting for 60% of the
blood supply to the breast7.
2) Branches of the lateral thoracic artery emerging from the axillary artery, and the
pectoral branches of the thoracoacromial artery, also branch of the axillary artery
supply the upper outer quadrant of the breast and are responsible for 30% of the
blood supply to the breast.
20
3) Branches from the posterior intercostal arteries supply the remainder of the
blood to the breast.
The blood supply to the breast skin depends on the subdermal plexus, which is
communicated with the deeper vessels supplying the breast parenchyma. The internal
thoracic artery is an essential and sustained contributor of blood supply to the nipple-
areola complex by means of its perforating branches and anterior intercostals branches 8.
The venous drainage of the breast and chest wall accompanies the course of the
arteries and drains towards the axilla. The veins form an anastomotic circle, called the
circulus venosus around the nipple8. In the end, the veins from this circle and the gland
drain into vessels joining the internal thoracic and axillary vein. The three main veins are
the perforating branches of the internal thoracic vein (largest venous plexus to provide
drainage of the mammary gland), the perforating branches of the posterior intercostal
veins and the tributaries of the axillary vein8.
Lymphatic Drainage
Lymphatic drainage from the breast plays an important role in the proliferation of
breast malignancies and is the main get away for tumor cells. In fact, more than 75% of
the lymphatic drainage of the breast is through the axillary lymph nodes. The interlobular
lymphatic vessels intersect to shape the subareolar plexus and from there drain to the
axillary, internal mammary, and supraclavicular lymph nodes. Lymphatic drainage of the
epithelial and mesenchymal components of the breast is the primary route of metastatic
spread of breast cancer6.
2. Breast carcinogenesis
Carcinogenesis is a multifaceted process that involves numerous etiologies

which work at different stages. It refers to the process by which normal cells become
cancerous. The precise mechanisms by which breast cancer is initiated is unknown;
however, much effort has been made to molecularly characterize breast cancer and
depict its development and progression4.
The most common types of breast cancer are malignant proliferations of epithelial
cells in the ducts (ductal carcinoma) or lobules (lobular carcinoma) 9. Breast tumor
develops through a multistep process 9, it commonly starts from the ductal
hyperproliferation, and consequently advances to benign tumors or metastatic
carcinomas after a continuous stimulation by various carcinogenic factors (Figure 3).
21
Tumor-associated macrophages appear to take a vital role in breast cancer initiation and
progression of breast cancer 10. In fact, in the early stages of carcinogenesis,
macrophages are capable of promoting angiogenesis, cell migration, and invasiveness
of tumor cells11.
At the molecular level, there is a parallelism between normal cell development

and tumor progression12. In fact, a series of epithelial-mesenchymal interactions occur
in both13. Under normal conditions, cell development is strongly controlled by complex
signaling pathways that permit the transmission of signals with each other and with the
environment in which they are settled. During the carcinogenesis process, many of these
signaling pathways are disrupted. Ultimately, cancer develops due to genetic and
epigenetic alterations which allow cells to escape the mechanisms responsible for
controlling their proliferation, survival and migration. There is a cellular microenvironment
(fibroblasts, lymphocytes and macrophages, myoepithelial cells and endothelial cells,
among others) which increases as tumors progress, although the degree of this is tumor
subtype specific. This cellular microenvironment interacts with extracellular matrix
molecules to modulate the specific function of normal breast tissue, as well as the
growth, survival and invasiveness of the mammary cells when they become malignant
(Figure 3)11.
Figure 3. Microenvironmental alterations during tumor progression. Schematic view of

normal breast, DCIS (ductal carcinoma in situ), IDC (invasive ductal carcinoma), and
metastatic breast carcinoma. (Taken from Polyak et al.11).
Recently, a new subclass of malignant cells within tumor called cancer stem
cells (CSC) are observed and associated with tumor initiation, escape and recurrence.
The CSCs could come from mammary stem cells. They show the ability to self-renew
and to resist to traditional therapies such as chemotherapy and radiotherapy14.
22
Mammary stem cells exist in a small proportion in the mammary gland as undifferentiated
cells with the ability for self-renewal and the capacity to generate cells with differentiated
functions15. In healthy tissue, stem cells fulfill functions of tissue maintenance, repair and
regeneration16.
There are two hypothetical theories for breast cancer initiation and progression, the
cancer stem cell theory and the stochastic theory 9 which are supported by plentiful data,
although neither can comprehensively prove the origin of human breast cancer.
• The cancer stem cell theory proposes that all tumor subtypes are derivative of the
same stem cells or progenitor cells. Acquired genetic and epigenetic mutations in
stem cells or progenitor cells will induce different tumor phenotypes (Figure 4A).
• The stochastic theory proposes that each tumor subtype is initiated from a single
cell type (stem cell, progenitor cell, or differentiated cell) (Figure 4B). Random
mutations can progressively accumulate in any breast cells, leading to their
transformation into tumor cells when enough mutations have accumulated.
Figure 4. Two hypothetical theories of breast cancer initiation and progression. (A)
Cancer stem cell theory. (B) Stochastic theory. (Taken from Sun et al.9)
23
Major signaling pathways in breast cancer development and progression
Typically, cancer is influenced by genetic and epigenetic alterations which allow

cells to evade the mechanisms that generally control their proliferation, survival and
migration17. Many of these changes map to signaling pathways which control cell
proliferation and division, cell death, cell differentiation and fate, and cell
motility17. Hence, turning on mutations of proto-oncogenes can lead to hyperactivation
of these signaling pathways, while inactivation of tumor suppressors eliminates crucial
negative regulators of signaling17.
The most predominant signaling pathways implicated in the normal development of

the mammary gland and in its neoplastic degeneration can be gathered into four
categories:
1. ER signaling and ER-Positive breast cancer
Estrogen receptors (ERs) comprise membrane ERs and nuclear ERs (ERα, ERβ)18,
which are transcriptional factors that either activate or suppress the expression of target
genes upon ligand binding. ERα (coded by ESR1) plays a pivotal role in the
pathogenesis of breast cancers. In fact, approximately 75% of breast cancers have
positive expression of this specific type of hormonal receptor19. One of the best
characterized mechanisms by which ERα promotes the growth of breast tumor cells is
through its interaction with cyclin D120, which is an important activator of cyclin-
dependent kinases (CDKs) 4 and 6. These kinases coordinate the transition of the cell
cycle from G1 to S phase in many cancer cells21.
Even though ERα is the one of the mainly frequently examined parameters in breast
cancer patients, ERβ has been gently drawing attention. ERβ has partial overlap with
ERα in the tissue expression profile. The level of ERβ drops along with the progression
of breast tumors22. In vitro and in vivo studies support its role as a breast cancer
suppressor23,24. Nonetheless, ERβ signaling in breast cancer and other circumstances
are not as clear as ERα25. However, since more evidence suggests relevant role of ERβ
in the mammary tumorigenesis, it is likely to end up being as crucial as ERα26.
2. HER2 signaling and HER2-Positive breast cancer
Human epidermal growth factor receptor (HER)s 1 to 4 comprise a family of tyrosine

kinase receptors expressed in normal tissues and in many types of cancer 27. HER2 is
amplified and/or overexpressed in 13-15% of breast cancers, and leads to an activation
of the HER2 pathway4. HER2 signaling amplification results in HER2 protein
24
overexpression which is related to tumor cell proliferation and cancer progression28.
HER2/NEU or c-ERBB2 consists of an extracellular ligand-binding domain, a
transmembrane domain, and an intracellular domain 29. HER2 has the capability to form
dimers with other molecules and can affect many cellular functions through various
pathways when it is activated12. Ligand binding and successive dimerization stimulate
phosphorylation of tyrosine residues in the intracellular domain of HER2, provoking the
activation of different downstream signaling pathways such as the mitogen-activated
protein kinase (MAPK) and the phosphatidylinositol 4,5-bisphosphate 3-kinase (PI3K)
pathways which have been robustly associated with breast tumorigenesis30.
While the interplay of HER2 and ERs has long been identified31, it was latterly found
that a new intermediary factor -Mediator Subunit 1- has significant impact on HER2-
driven tumorigenesis32. The precancerous effect of HER2 was also found to be related
to inflammation and the expansion of CSCs in breast cancer33.
3. Canonical Wnt/β-catenin signaling in breast cancer
Wnt proteins are a family of highly glycosylated, secreted proteins with pivotal roles
in various developmental actions including embryonic induction, cell polarity generation,
and cell fate specification, along with maintenance of adult tissue homeostasis34. Wnt
signaling is implicated primarily in the processes of breast cancer proliferation and
metastasis, as supported by genome-wide sequencing and gene expression profile
analyses. In recent decades, a growing number of studies have demonstrated that Wnt
signaling involves cell proliferation, metastasis, immune microenvironment regulation,
maintenance of breast cancer cell stemness, therapeutic resistance, and phenotype
shaping of breast cancer35.
4. Other signaling pathways
• Cyclin dependent kinases (CDKs): cyclins, CDKs, and CDK inhibitors (CDKIs)
regulate cell cycle progression36. All cancers activate the cell cycle to preserve their
survival. The overexpression of cyclin D1 and cyclin E along with the decline in the
expression of CDKI p27Kip1 were found in human breast cancer 37. Cyclin D1
amplification is detected in roughly 60% of breast cancers.
• Notch signaling: Notch signaling is implicated in the pathogenesis of breast cancer

and as such may serve as a novel therapeutic target. Notch signaling involves five
Notch ligands, namely Delta-like (Dll) 1, 3, 4, and Jagged 1, 2, which are single
25
transmembrane proteins38 and there are four Notch receptors. Notch 3 and Jagged1
have emerged as key regulators of cancer stem cell renewal and hypoxia survival in
primary breast cancer and breast cancer cell line-derived tumorspheres39. Moreover,
Notch signaling also interacts with HER2 signaling pathway, which is active in
approximately 20% of breast cancers and linked with a more aggressive form of the
disease.
• Sonic Hedgehog signaling plays an essential role in coordinating cell growth and
differentiation during embryonic tissue patterning 40. Interruption of its downstream
transcriptional targets, Patched homolog-1 or glioma-associated oncogene-2
produced severe defects in ductal morphogenesis 41 which have also occurred in
breast cancer42. This indicates a role for the Sonic Hedgehog signaling pathway in
breast tumorigenesis.
• Breast tumor kinase (BRK): BRK is a non-receptor tyrosine kinase overexpressed

in more than 60% of breast cancer cases, but not in normal mammary glands or
benign lesions43–45. BRK may play an important role in promoting cell proliferation
and migration in breast cancer cells.
• PI3K/AKT/mTOR pathway: oncogenic mutation of PI3K is generally known in

human breast cancer. It may induce the dedifferentiation of luminal or basal
mammary progenitor cells, and helps them to achieve multi-lineage potential46.
3. Breast Cancer Subtypes
Types of breast cancer based on pathology, invasiveness and prevalence

Breast cancers can occur in different areas of the breast, such as the ducts, the
lobules, or the tissue in between and is conditional to the exact cells that are altered.
Depending on which cell origin is implicated, breast cancers can be divided into two
broad classifications, carcinomas and sarcomas. Nevertheless, in some occasions, a
single breast tumor can be a mixture of different cell types47–49.
Sarcomas are an unusual type of breast cancer (<1% of primary breast cancer)
emerging from the stromal components of the breast, which comprise myofibroblasts
and blood vessel cells. Carcinomas, which represent the vast majority of breast cancers,
are cancers arising from the epithelial component of the breast, which embody cells lining
the terminal ductal lobular units that comprise the ductal system of the breast 12. There
are numerous subtypes of breast carcinomas. Given the different prognoses and
treatment options, a meticulous distinction between the various subtypes is essential.
26
Common breast cancers can be divided into three groups: non-invasive (or in situ),
invasive and metastatic breast cancers contingent to the criteria of pathological
features and invasiveness.
• Non-invasive (or in situ) breast cancer

DCIS is a non-invasive or pre-invasive breast cancer, which evolves inside of pre-
existing normal ducts and has the potential to become an invasive cancer even though
DCIS is not invasive itself.
• Invasive or infiltrating breast cancer

Invasive breast cancers have cancer cells that invade and spread outside of the normal
breast lobules and ducts, and grow into the surrounding breast stromal tissue47–49. About
two-thirds of women with an invasive form of breast cancer are 55 or older when they
are diagnosed. Invasive carcinomas have the potential to spread to other sites of the
body, such as the lymph nodes or other organs and to form metastases. When that is
the case, they enter the classification of metastatic breast cancers47–49. According to the
tissue and cell types implicated, invasive breast cancers are further divided into the
following two types:
1. Invasive Ductal Carcinoma (IDC)

It is the most common type of breast cancer with about 80% of all breast cancers being
constituted by invasive ductal carcinomas47–49. The IDC histopathological classification
includes several subtypes (Figure 5)47–49:
Mucinous Papillary
carcinoma carcinoma
of the of the
breast breast
Medullary Cribriform
carcinoma carcinoma
of the of the
breast breast
Tubular
IDC-Not
carcinoma IDC
otherwise
of the subtypes
specified
breast
Figure 5. Invasive Ductal Carcinoma subtypes47–49.
27
2. Invasive Lobular Carcinoma (ILC)
ILC is the second most common type of breast cancer and accounts for 10–15% of
all breast cancers47–49. ILC tends to occur later in life than IDC, i.e., in the early 60s as
opposed to the mid-to late 50s for IDC47–49.
IDC and ILC cancers each display distinctive pathologic features. Lobular
carcinomas and ductal carcinomas are distinguished since the first ones grow as single
cells arranged individually and have distinct molecular and genetic aberrations. They
may have distinct prognoses and treatment choices and it is thus important to precisely
differentiate one from another.
• Metastatic breast cancer

Metastatic breast cancers, also recognized as stage IV or advanced breast cancers,
are late stage breast cancers, which have spread to other organs in the body 47–49.
Metastases from breast cancers can be found in lymph nodes in the armpit, and/or in
distant places such as the brain, lung, bone or liver, among others (Figure 6). The
histopathological classification influences the profile in terms of timing and sites of
metastatic disease. Microscopic tumor cells or micro-metastases may remain in the body
(even after the removal of the primary tumor) and allow the cancer to turn back and
disseminate. The risk of breast cancer relapse and with the potential to metastasize is
not clearly understood or predictable.
This risk of metastasis varies among breast carcinomas, conditional to the unique
molecular biology of the tumor and the stage at the time of the first diagnosis.
Approximately 30% of the women diagnosed with early-stage breast cancer will develop
a metastatic form of the disease47–49.
28
Figure 6. Frequent metastatic sites in breast cancer. (Taken from from Harbeck, N et
al.4)
• Less common types of breast cancer: while IDC and ILC explain approximately
90–95% of all breast cancer cases, there are still some rare types of breast cancers
that can be identified in the clinical setting (Box 1)47–49:
Inflammatory breast cancer (IBC)

• Uncommon (1%-5% of all breast cancers)
• Symptoms include inflammation-like breast swelling, purple or red color of the skin, and
pitting or thickening of the sking of the breast.
• Not detectable on mammograms.
• More common in younger women, African American and overweight or obese women.
• More aggresive (in one-third of cases, IBCs are already metastasized at diagnosis).
Male breast carcinomas

• Uncommon (<1% of all breast cancer cases)
• Either in situ or invasive
• Most cases comprise invasive ductal carcinomas with ER expression.
• Most common breast lesion is gynecomastia (or breast enlargement).
Paget disease of the breast

• Makes up <3% of all breast cancer cases.
• Originates in the breast ducts, spreads to the skin of the nipple and expands to the areola.
• Paget's cells divide rapidly and look different from normal cells (50% of the cells are + for
ER and PR, and most are + for the HER2 protein).
29
Papillary carcinoma
• Makes up <3% of all breast cancer cases.
• Papillary carcinoma cancer cells are usually arranged in finger-like projections.
• Most are invasive and are treated in the same way as IDCs.
• Better prognosis than another invasive breast cancer.
• Papillary carcinomas may also be distinguished when they are still noninvasive.
• A noninvasive papillary carcinoma is commonly treated as a variety of DCIS.
Phyllodes tumor
• Rare breast tumors.
• Develop in the stromal cells of the breast.
• One-quarter are malignant.
• Most frequently found in women in their 40s and in women with Li-Fraumerni syndrome.
Angiosarcoma of the breast

• Makes up <1% of all breast cancers.
• Involves the breast tissue or the skin of the breast.
• Tend to grow and spread quickly.
Box 1. Less common types of breast cancer47–49.
Molecular or intrinsic subtypes of breast cancer

The molecular alterations that drive breast carcinogenesis are many, and several
classifications have been developed to group tumors in a suitable way.
Immunohistochemical staining of paraffin sections using antibody panels has been
reported to be an acceptable, albeit not perfect, proxy for molecular classification of
invasive breast cancers as classified by gene expression profiling studies 50–53. This
conceptual change has led to a new paradigm on how breast cancer patients are
stratified and treated, and provides more accuracy of disease prognosis and therapeutic
decision making54. The rationale behind such classification, which is at the same time
prognostic, predictive, and therapeutic, is that the differences underlying the gene
expression patterns among cancer subtypes reveals the fundamental differences of the
tumors at the molecular level55.
Classical immunohistochemistry (IHC) markers such as ER, PR and HER2, along

with traditional clinicopathological variables (e.g., tumor size, tumor grade and nodal
involvement) are commonly used for patient prognosis and management 56. Through the
application of gene expression profiling techniques, the list of intrinsic genes that
differentiate these subtypes is now made up of several clusters of genes in relation to
estrogen receptor (ER) expression (the luminal cluster), HER2 expression, proliferation,
and a unique cluster of genes called the basal cluster 47–49. The pioneer studies
conducted by Sørlie et al.57 reported a distinctive ‘molecular portrait’ of breast cancer
using 456 cDNA clones, according to which tumors were classified into five intrinsic
30
subtypes with distinct clinical outcomes, i.e., luminal A, luminal B, HER2 over-
expression, basal and normal-like tumors57,58. Therefore, breast cancers are usually
divided into five molecular subtypes based on the expression pattern of some genes.
Figure 7 shows a summary of the subtypes:
Intrinsic subtype Molecular Signatures
Luminal A ER+, PR±, HER2-, Low Ki67

Luminal B ER+, PR±, HER2±, High Ki67
HER2-enriched ER-, PR-, HER2+
Basal-like ER-, PR-, HER2-
Normal-like ER+, PR±, HER2-, Low Ki67
Figure 7. Summary of breast cancer subtypes57,58.
Clinical staging and survival rates of breast cancer

Once breast cancer is diagnosed, tests are executed to ascertain the stage of the
disease, which will influence the treatment patients receive. The clinical staging of breast
cancer is identical across breast cancer subtypes according to the American Joint
Committee on Cancer and the International Union for Cancer Control Tumor, Node, and
Metastasis (TNM) breast cancer staging system: Stage 0, Stage I, Stage II, Stage III and
Stage IV, as detailed in Table 1.
Table 1. Anatomic stage groups of breast cancer.

Stages Definition
Stage 0 Ductal Carcinoma in Situ
Stage I IA Primary invasive tumor with a size of ≤20 mm. No nodal involvement
IB Nodal micro metastases (>0.2 mm, <2.0 mm) with or without ≤ 20 mm primary tumor
Stage II IIA Movable ipsilateral Level I, II lymph node metastases with ≤20 mm primary tumor; Or > 20 mm,
≤ 50 mm tumor with no nodal involvement
IIB Movable ipsilateral Level I, II lymph node metastases with >20 mm, ≤ 50 mm tumor; Or > 50 mm
tumor with no nodal involvement
Stage III IIIA Movable ipsilateral Level I, II lymph node metastases with >50 mm tumor; Or any size primary
tumor with fixed ipsilateral Level I, II or internal lymph node metastases
IIIB Primary tumor with chest wall and/or skin invasion
IIIC Any size primary tumor with supraclavicular or ipsilateral Level III lymph node metastases; Or
with ipsilateral Level I, II and internal lymph node metastases
Stage IV Any case with distant organ metastasis
(Taken from American Joint Committee on Cancer Cancer Staging Manual, 8 th Edition, The American College of Surgeons 59).
31
4. Epidemiology of breast cancer
Female breast cancer is the most commonly occurring cancer worldwide (11.7%
of the total new cases or 2,261,771 cases), followed by lung cancer (11.4%), colorectal
cancer (10.0%), prostate cancer (7.3%), and stomach cancer (5.6%). In women, breast
cancer is the first cause of cancer-associated death and the most commonly diagnosed
malignant tumor worldwide, followed by colorectal cancer and lung cancer for incidence
and lung cancer and colorectal cancer for mortality. Although the advances in diagnosis,
treatment and intensive research have undoubtedly increased the survival rates among
breast cancer patients in the past years, its incidence has increased worldwide,
particularly in developing countries.
Breast cancer epidemiology pattern differences across countries are further

compounded by cultural and lifestyle factors and national awareness campaigns 4. Breast
cancer is the most common cancer in terms of incidence among women in 157 of 185
countries (Figure 8A) and the most common cause of cancer death in women in more
than 108 countries (684,996 deaths in women) (Figure 8B). Currently, the highest
incidence of breast cancer occurs in Belgium, while it is lowest in Central Africa and
South-Central Asia.
32
B
Figure 8. Map showing the most frequent tumors in incidence (Figure 8A) and mortality
(Figure 8B) in 2020 for each country among women (Taken from from Globocan 202060).
Over 355,000 women in the European Union were estimated to be diagnosed

with breast cancer in 2020 (13.3% of all cancer diagnoses). The most frequent causes
of death are from cancers of the lung (20.4% of all cancer deaths), followed by colorectal
(12.4%), female breast (7.3%) and pancreatic cancer (7.1%). The highest incidence of
breast cancer within the European framework occurs in Belgium (age-standardized
incidence rate per 100,000 is 107.7), whereas the minimum is found in Ukraine (age-
standardized incidence rate per 100,000 is 42.6). In Spain, breast cancer continues to
be the most frequent cancer among women (28.7% of the total, with 34,088 new cases
estimated in 2020), ahead of colorectum (13.3%) and lung (6.5%) cancers.
In the US, female breast cancer was estimated to be the most commonly
diagnosed cancer in 2020, accounting for 253,465 new cases, followed by lung (10.6%)
and colorectal (7%) cancers. Based on 2015 to 2017 data, around 12.9% of women will
be diagnosed with female breast cancer during their lifetime 61.
The high incidence in developed countries has been associated with the high
prevalence of known risk factors such as hormonal factors (early menarche, late
menopause), reproductive factors (nulliparity, fewer children, increased age of first
pregnancy, shorter time of breastfeeding), hormonal therapies (oral contraceptives and
hormone replacement therapy), suboptimal nutrition (alcohol consumption),
anthropometric characteristics (increased weight, weight gain during adulthood, body fat
distribution); and lower levels of physical activity.
33
Moreover, prominent changes in lifestyle, socio-cultural contexts, and built
environments are having a major impact on the prevalence of breast cancer risk factors
in many less developed countries where an increase in incidence has been observed.
5. Risk factors
A risk factor is defined as anything that affects an individual’s chance of getting a

disease, in this case, breast cancer. Multiple factors modify the risk of developing breast
cancer and certain factors increase a woman's risk, and some have a bigger impact on
risk than others. We describe both non-modifiable and modifiable risk factors:
5.1. Non-modifiable risk factors
These are defined as those which cannot be modified (e.g., genetics) or that are
unlikely to be modified with the aim of reducing breast cancer risk. However, some of
these factors do have modifiable components since height, age at menarche, and age
at menopause are partially determined by diet and body size.
5.1.1. Sex. Although men can also get breast cancer, female sex is the main risk factor
for breast cancer. In fact, the incidence among women is about 100 higher than men,
who do not have the same complex breast growth and development as women 62.
5.1.2. Age. Age is a well-established risk factor for breast cancer. In fact, the aging
process is the biggest risk factor for breast cancer. The risk of breast cancer increases
with age and therefore, about 85% of the diagnoses are made after the age of 50 while
5% occur in women under the age of 40. According to the American Cancer Society,
about 1 out of 8 invasive breast cancers occur in women under 45 years old and
approximately 2 out of 3 invasive breast cancers are found in women aged 55 years or
older63. In Europe, an average of 20% of breast cancer cases are diagnosed in women
when they are younger than 50 years old; 36% occur at age 50–64 and the remaining
cases in women above this age60.
5.1.3. Birthweight. Women with a higher birthweight have an increased risk of breast
cancer (especially before menopause)64–66. In fact, birthweight may be a marker of
factors related to intrauterine growth (e.g., height) that affect breast cancer risk67. A
higher birthweight does not seem to be associated to higher estrogen levels in
adulthood68.
34
5.1.4. Age at menarche. Starting menstrual periods at a younger age is associated to a
moderately increase in breast cancer risk67,69–71. A pooled analysis of data from 117
studies found for every year younger a woman was when she began her periods, her
breast cancer risk increased by 5%70.
The increased risk of breast cancer associated to a younger age at first period is
possible due, partly, to the amount of estrogen a woman is exposed to throughout her
life. Indeed, a higher lifetime exposure to estrogen is linked to an increment in breast
cancer risk72. A woman cannot change the age at first period, but, importantly, there are
numerous healthy lifestyle factors that can influence the age a girl initiates her period
(e.g., girls who are lean and are more physically active tend to begin their periods at a
later age than other girls).
What is more, to some extent, menarche is only one late part of the complex female
pubertal transition, which also involves the thelarche (the onset of breast development)
which is typically the first sign of puberty. It commonly happens 2-4 years before
menarche73. A cohort study of 104,931 women (Breakthrough Generations Study)
assessed the association between the timing of pubertal stages and breast cancer risk.
In this study, an earlier recalled age at thelarche was associated with a 20% increased
breast cancer risk74. In the last 50 years, there has been a temporal drop in the age of
thelarche but not menarche, which reflects a prolonged pubertal tempo over time75.
Furthermore, independent of either age at thelarche or menarche, the pubertal tempo
(length of time between thelarche and menarche) was positively linked to an increased
risk of breast cancer74. Nonetheless, relating pubertal timing with a breast cancer
diagnosis that takes place more than 50 years later is a barrier for epidemiological
studies76. One way to undertake this challenge is to evaluate intermediate markers that
can be measured earlier in life (e.g., breast density). The latter is one of the strongest
predictors of premenopausal and postmenopausal breast cancer risk 77,78 and may be a
useful intermediate marker between early life factors (e.g., pubertal development) and
breast cancer risk.
5.1.5. Height. Numerous studies have shown that taller women have a higher risk of
breast cancer79–81.
The major determinant of height across diverse populations is economic

development and nutrition in lieu of genetics, and especially, the average population
35
height between countries also correlates with cancer rates 82. Thus, both genetic and
environmental factors contribute to height and, as a consequence, to cancer risk.
In the Million Women Study, a large UK prospective cohort, breast cancer incidence
increased with increasing adult height 83. Particularly, the relative risk for breast cancer
was 1.17 (95% CI 1.15, 1.19) for every 10 cm increase in height. A potential explanation
for this increase may be related to the growth spurt tall women can have in their youth,
which can be produced by higher levels of hormones such as IGF-1 or other growth
factors. These higher hormone levels together with rapid cell growth amid a growth spurt,
may affect breast cancer risk later in life.
As indicated by Green et al.83, the increase (by approximately 1 cm per decade

throughout the 20th century) in average adult height in European populations would
explain a 10-15% increase in cancer incidence exceeding that expected if the average
population height had been maintained over this time period.
5.1.6. Family history of breast cancer. Women with breast cancer in their first-degree
family have a twofold higher risk of developing the disease. This increased risk is larger
among women with a first-degree relative diagnosed before the age of 50 compared to
a diagnosis made over the age of 50 years84,85.
There are numerous inherited genes associated with an increased risk of breast
cancer. In fact, approximately 5–10% of all breast cancers are predicted to evolve due
to high-impact germline mutations in breast cancer susceptibility genes. Up to 30% result
of pathogenic mutations in BRCA1 and BRCA2 and a smaller proportion of women carry
mutations in other susceptibility genes (PTEN, TP53, CHEK2, PALB2 and STK11) 86.
BRCA1 or BRCA2 are linked to an average cumulative risk of acquiring breast cancer
by the age of 80 years of 72% and 69%, respectively 87. Mutations in these genes are
associated to a higher breast and ovarian cancer risk, as well as other types of cancer.
The risk of breast cancer among men, along with other cancers such as prostate cancer,
is also increased if he has a mutation in BRCA1 or BRCA2.
36
Nonetheless, there are other, although rare, gene mutations or hereditary conditions
which can increase a person’s risk of breast cancer. Some of these genes and
syndromes are (Figure 9):
Lynch syndrome,
linked with Cowden syndrome, Li-Fraumeni Peutz-Jeghers
the MLH1, MSH2, linked syndrome, linked syndrome, linked
MSH6, and PMS2 with the PTEN gene with the TP53 gene with the STK11gene
genes
Ataxia telangiec- Hereditary diffuse

tasia, linked with gastric cancer, linked PALB2 gene
the ATM gene with the CDH1 gene
Figure 9. Hereditary conditions or gene mutations that can increase a person’s risk
of breast cancer88.
5.1.7. Personal history of DCIS. DCIS is non-invasive, but without treatment, the
abnormal cells grow within the milk ducts of the breast and can progress to invasive
breast cancer over time. If they are not treated, about 40-50% of DCIS cases may
advance to invasive breast cancer89.
5.1.8. Personal history of certain cancers can augment the risk of breast cancer.
• Radiotherapy for Hodgkin lymphoma: women who received mantle field radition
for Hodgkin lymphoma in their childhood or early adulthood have an increased
risk of breast cancer90–95.
• Ovarian cancer: women with a previous history of ovarian cancer have an
increased risk of breast cancer 96, probably due to genetic factors. As described
above, women who have an inherited BRCA1 or BRCA2 gene mutation have an
increased risk of either breast cancer and ovarian cancer96.
• Other cancers: there are certain genetic syndromes resulting from rare gene
mutations that increase the risk of breast cancer. A personal history of colorectal,
thyroid or uterine cancer, that is related to a rare genetic syndrome can broaden
the risk of breast cancer69,97.
5.1.9. Race, ethnicity. Rates of breast cancer in the US diverge by race and ethnicity.
While Caucasian women and African American women have the highest incidence
overall62, American Indian and Alaska Native women have the lowest incidence. To
explain these differences, one possible argument is that the prevalence rates of certain
risk factors for breast cancer vary by race and ethnic group79.
37
Caucasian women are more prone to have children at a later age, to have fewer
children and to use menopausal hormone therapy, compared to Hispanic/Latina women,
African American women, and non-Hispanic black women79,98,100. Reproductive factors
such as the already mentioned increase breast cancer risk 79. In general, black women
have a moderately lower rate of breast cancer compared to Caucasians62. Still, there are
differences by age; black women have higher rates of breast cancer compared to white
women in those under 40 years. In contrast, between ages 60-84, Caucasians have
higher rates of breast cancer compared to black women. Although the reasons are still
under study, they may be due to differences in prevalence rates of some reproductive
and lifestyle factors related to breast cancer risk along with differences in tumor biology 98–
101
(Figure 10).
Age at first period
Menopausal
Age at first
hormone
childbirth
therapy
Age at Number of
menopause childbirths
Body weight Breastfeeding
Figure 10. Risk factors that differ by race and ethnicity98–100.
5.1.10. Age at menopause: going through menopause at a later age increases the risk
of breast cancer70,79,102, which is likely due, at least in part, to the amount of estrogen a
woman is exposed to in her life. A higher lifetime exposure to estrogen is related to an
increase in breast cancer risk72. A pooled analysis of data from more than 400,000
women found for every year older a woman was at menopause, her breast cancer risk
was increased by about 3%70. Studies have shown that women who go through
menopause after age 55 have about a 40 % higher risk of breast cancer than women
who do so at age 45 or younger103.
5.1.11. Radiation exposure: exposure to extensive amounts of radiation early in life

(e.g., radiation therapy to the chest area for childhood cancer) increases the risk of breast
cancer91,92,104.
38
Some women who had Hodgkin’s lymphoma at a young age and who had been
treated with radiation therapy have a 3-7 times higher risk of breast cancer than women
who had Hodgkin’s disease at a young age but were not treated with radiation therapy
to the chest area91. Nonetheless, although radiation therapy increases the risk of breast
cancer later in life, its benefits far outweigh this risk. Another concern for breast cancer
risk is the amount of radiation given and a person’s age at the time of the treatment for
Hodgkin’s disease or another cancer. Generally, the higher radiation a person is exposed
to and the younger the age at treatment, the greater the risk of breast cancer91,104.
Airline crew exposure to radiation: the restricted amount of radiation exposure to

airline crews is improbable to boost breast cancer risk. Female flight crews are likely to
have slightly higher rates of breast cancer than other women 105–107. Nonetheless, this
may be due to reproductive and lifestyle factors, as well as the job’s characteristics
including the night shift work. In fact, women who work night shifts for numerous years
have a small increased risk of breast cancer108–110.
5.2. Modifiable risk factors
Current research on modifiable risk factors, including dietary habits, estimated that
a large percentage of cancer is preventable by behavior modification.
5.2.1. Body weight and weight gain: studies have shown a link between body mass
index (BMI) and breast cancer risk; however, BMI affects risk diversely before and after
menopause.
• Before menopause, women who are overweight or obese have a 20-40% lower risk
of breast cancer than those who are lean 111–114. Although being overweight or obese
may decrease the risk of breast cancer before menopause, weight gain should be
avoided, and the health risks of obesity far outweigh any potential benefit on risk of
premenopausal breast cancer. Nonetheless, this remarkable association suggests
probable adiposity effects in this crucial early life window, which are under current
investigation.
• After menopause, being overweight or obese increases breast cancer risk111,112,114
(Figure 11). In fact, women who are overweight or obese after menopause have a
30-60% higher breast cancer risk than those who are lean111,112,114.
39
Figure 11. Fraction (%) of all cancer cases (at all anatomical sites) among both sexes
(worldwide) in 2012 attributable to excess body mass index, by country. (Taken from
Globocan 202060).
While before menopause most estrogens in the body are originated in the ovaries,
after menopause the ovaries no longer produce much estrogen and estrogens primarily
originate in the fat tissue. Heavier women have higher blood estrogen levels than leaner
women72 and, therefore, have an increased risk of breast cancer compared to women
with lower estrogen levels72. Moreover, women who are heavier also tend to have higher
levels of insulin compared to leaner women115 and numerous studies have shown a
higher breast cancer risk among postmenopausal women with high insulin levels,
including women with type 2 diabetes116–118. Among premenopausal women, a possible
link between insulin levels and breast cancer risk is less clear 119.
Weight gain: gaining weight in adulthood increases the risk of breast cancer before
and after menopause 120–122. In the Nurses’ Health Study, women who gained 10kg or
25kg or more after age 18 had a 18% and 45% higher breast cancer risk, respectively,
compared to women who maintained their weight120. This association was stronger
among women who had never taken hormone replacement therapy. Losing weight after
menopause may lower the risk of breast cancer120,123,124. An analysis from the Pooling
Project of Prospective Studies of Diet and Cancer comprising 180,885 postmenopausal
women suggested that sustained weight loss, even moderate amounts, was associated
with lower breast cancer risk for women aged 50 years and older compared to women
40
with stable weight124. The Women’s Health Initiative (WHI) study found a clear linear
dose–response association between BMI and duration of being overweight or obese with
an increased risk of post-menopausal breast cancer125,126. Nonetheless, not all the
research evidence suggests that weight loss after menopause lowers breast cancer
risk124. In fact, weight loss in adulthood and the risk of breast cancer before menopause
remains under study127.
Body shape: body shape may also affect breast cancer risk. Some findings showed
that women had a modest increased risk of breast cancer if they put on additional weight
around their middle sections (often known as “apple-shaped”), in contrast to their hips
and thighs (commonly known as “pear-shaped”)128–131. Other studies have shown that
after BMI is taken into consideration in the analyses, body shape is no longer associated
with breast cancer risk132. A recent study from the SUN Project showed that women who
were lean in early life and experienced a notable increase in their body shape level had
a significantly higher risk of developing breast cancer when compared to women who
remained stable in a medium body shape level. This association was observed for
postmenopausal breast cancer133 which is in line with previous evidence134,135. Further
research should be fostered with the aim of better understanding the relationship
between body shape trajectories and the specific impact on pre- and postmenopausal
breast cancer incidence as well as in breast cancer subtypes.
5.2.2. Diet
Diet is an essential backbone of any lifestyle and one of the most modifiable elements
of lifestyle to help prevent cancer in general, and breast cancer in particular. Regardless
of site-specific details in the tumor pathogenesis, more than 90% of cancers are
recognized to be attributable to modifiable risk factors such as tobacco smoking, excess
body weight, physical inactivity, alcohol consumption, infectious factors, environmental
pollution, and poor diet136,137. In fact, poor diet is responsible for about 5–10% of total
cancer cases136,138.
The field of nutrition has built upon its initial focus on specific nutrients to now cover
the examination of whole dietary patterns, with the motivation that behavior change may
be more easily reached by tackling the diet as a whole. As food items and nutrients are
consumed in combination, dietary patterns have been successfully implemented as a
tool to assess the additive or synergistic effect of food in nutritional epidemiology 139,140.
41
Dietary patterns are defined as the quantities, proportions, or combinations of
different foods and beverages in diets and the frequency with which they are
consumed139. Typically, dietary pattern analysis is divided into two major groups based
on whether the patterns are determined empirically or using investigator-defined criteria.
They are also recognized as a posteriori or a priori141 dietary patterns. Data-driven or a
posteriori approaches apply multivariate statistical methods such as principal component
analysis, factor analysis and cluster analysis directly to a participant’s food data collected
by food-frequency questionnaires (FFQs), dietary recall, or food record to derive the
dietary patterns. They are data-driven methods that aim to identify which components of
the diet occur together to explain variations in food or beverage intake across dietary
patterns, regardless of the study outcomes. In contrast, a priori approaches score an
individual’s actual intake as compared against an already-defined dietary index (e.g., the
Healthy Eating Index142), based on a series of dietary recommendations, reached by
scientific consensus or by investigators with an evidence-based approach. There is a
third, less commonly used, method based on qualitative self-reported behaviors rather
than detailed questionnaires (e.g., vegetarianism or veganism) which evaluates
individuals' intake preferences built on the foods and drinks that are included and
excluded in their diets.
The popularity of dietary patterns research is reflected in the inclusion of dietary

patterns in the last report of the World Cancer Research Fund / American Institute for
Cancer Research (WCRF/AICR)143 and the 2020-2025 Dietary Guidelines for
Americans144. The WCRF/AICR Third Expert Report incorporated statements consistent
with a whole-food plant-based diet, characterized by high amounts of whole grains,
vegetables, fruits, and beans as well as limited consumption of red and processed meats,
sugar, sugar-sweetened beverages (SSBs), and fast food and other processed foods.
While the cancer prevention recommendations widely apply to cancer prevention, the
factors addressed in these recommendations are directly relevant to the prevention of
breast cancer and may have a significant impact on survival after breast cancer
diagnosis145.
Previous studies showed that higher adherence to the 2007 WCRF/AICR nutrition-
related recommendations was associated with lower breast cancer in most—though not
all146,147—studies. Those studies that found a protective association include the
European Prospective Investigation into Cancer and Nutrition (EPIC) cohort 148, the Iowa
Women’s Health Study149, the Black Women’s Health Study150, the Swedish
Mammography Cohort151, the Canadian National Breast Screening Study152, and the
42
Vitamins and Lifestyle (VITAL) cohort153. Moreover, only two studies have evaluated the
association between the most updated version of the WCRF/AICR recommendations
and breast cancer risk154,155. In the first one, a higher adherence to the recommendations
was associated with reduced breast cancer risk, independently of menopausal status.
The study also included a dose-response meta-analysis which indicated that women who
highly adhere to the recommendations had an approximately 30% lower risk of breast
cancer compared to those with low adherence. In the Seguimiento Universidad de
Navarra (SUN) cohort155, higher compliance to the 2018 WCRF/AICR recommendations
was associated with lower risk of postmenopausal breast cancer.
Plant-based diets have been traditionally regarded as vegetarian diets, but other
definitions of plant-based diets can be considered by assigning negatively some or all
animal foods, and even accounting for the quality of plant-based foods in the diet. Plant-
based foods (fruits, vegetables, cereals, nuts and seeds, legumes, and vegetable oils)
are the main source of fiber and other bioactive compounds in the diet 156. Plant-based
nutrition and dietary recommendations are a large part of the WCRF/AICR
recommendations143. They are generally similar in several characteristics, such as
emphasis of fruits and vegetables and de-emphasis of red and processed meats 157. In
fact, there is growing evidence supporting breast cancer incidence can be reduced with
an overall healthy or prudent diet, which includes a high-quality diet consisting of fruits,
vegetables, whole grains and legumes. In 2019, a meta-analysis of 32 observational
studies including 43,285 breast cancer cases 158, suggested a reduced breast cancer risk
with a prudent dietary pattern (18% relative reduction), especially among premenopausal
women (23% relative reduction) and ER-positive and/or PR-positive and ER-negative
and/or PR-negative tumors.
A comprehensive systematic review and meta-analysis of 93 studies on a posteriori

dietary patterns (classified as “healthy” or “unhealthy”) and cancer risk reported that high
adherence to a “healthy” dietary pattern was associated with a reduced risk of breast
cancer in case-control (RR 0.66; 95% CI, 0.55, 0.80) and cohort studies (RR 0.90; 95%
CI, 0.84, 0.96)159.
One of the most well-represented plant-based dietary patterns in the literature is the
Mediterranean diet160, characterized by high amounts of fruits, vegetables, nuts,
legumes, fish, whole grains, and extra-virgin olive oil161. A large body of epidemiological
and clinical studies have observed the protective effect of the Mediterranean Diet on
cardiovascular disease, diabetes and obesity162. In 2020, an updated systematic review
43
and meta-analysis163 reported conflicting findings between the Mediterranean diet and
breast cancer risk. An inverse association of breast cancer with highest adherence to
the Mediterranean diet was found in observational studies (RR 0.94, 95% CI 0.90, 0.97,
I2 = 31%), however, considering separate designs there was a reduction in case-control
studies, but not in cohort studies. The authors stated that lack of association in cohort
studies might suggest that significant findings found in case-control studies could be
explained by bias linked to study design.
The only clinical trial included on this topic, the PREDIMED (PREvención con DIeta
MEDiterránea) trial, examined the Mediterranean diet supplemented with extra-virgin
olive oil, the Mediterranean diet with additional mixed nuts, and a low-fat control arm164.
Among more than 4,000 postmenopausal women, in 5 years of follow-up, there were 35
incident breast cancer cases. Breast cancer incidence was significantly lower in the
Mediterranean diet supplemented with extra-virgin olive oil arm when compared to the
control group who was advised to follow a low-fat diet. Olive oil has attracted
considerable attention during the last decades. According to a new systematic review
and dose-response meta-analysis of 10 studies, 8 of which are case-control studies165,
there may be a potential inverse association between olive oil intake and breast cancer;
however, since the estimates are non-significant and the certainty level is very low,
additional prospective studies with better assessment of olive oil intake are needed.
Traditionally, studies on plant-based diets and breast cancer have defined them as
“vegetarian” diets, categorizing study populations dichotomously into participants who
do or do not consume some or any animal foods. An important question from clinical and
public health standpoints, however, is whether gradually moving towards a plant-rich diet
by progressively decreasing animal food intake lowers breast cancer risk. If so, public
health recommendations could suggest progressive dietary changes. Existing studies of
vegetarian diets and breast cancer are also limited by a lack of differentiation among
plant foods with divergent effects on breast cancer (e.g., SSBs as opposed to fruits and
vegetables), and less nutrient-dense plant foods such as refined grains, potatoes, and
SSBs have been lately associated with higher breast cancer risk 166. Satija et al.167
conceptualized a graded dietary pattern that positively weighs plant foods and negatively
weighs animal foods, similar to the approach previously used by Martínez-González et
al.168. To overcome the limitations derived from the fact that all plant foods were treated
equally in the original provegetarian food pattern, Satija et al.167 proposed two new
scores (healthful and unhealthful plant-based dietary patterns). This is an appealing
approach since little consideration has been given to the quality and types of plant foods
44
consumed compared with the proportion and frequency of animal foods consumed within
the plant-based diet. The authors then examined the association of an overall plant-
based diet, and a priori, healthful and unhealthful versions of a plant-based diet, with
type 2 diabetes167 and coronary heart disease 169. Previous studies that examined
associations between plant-based diets based on the methods described by either
Martínez-González et al.,168 or Satija et al.,167 and breast cancer are not consistent in the
literature170. In the NutriNet Santé cohort study, a higher pro plant-based dietary score
was associated with decreased risk of overall cancer though no association was found
for breast cancer170 which may partly due to the restricted number of cases. In a recent
Iranian case-control study171, greater adherence to an overall plant-based diet index and,
particularly, its healthful version, was inversely associated with the risk of breast cancer,
while its unhealthful version was associated with an increased risk.
A commonality across plant-based dietary patterns is the high consumption of fruits

and vegetables. In a pooled analysis of 20 prospective cohort studies with 34,526 cases
of breast cancer (19,869 ER-positive and 4,821 ER-negative breast cancers)172,
consumption of total fruits and vegetables was not associated with overall risk of breast
cancer, but vegetable consumption was inversely associated with risk of ER‐
negative/PR‐negative tumors (HR Q5 vs. Q1 0.84; 95%CI 0.75, 0.93; P-trend=0.001). Total
vegetable consumption was associated with lower risk of overall breast cancer as well
as ER‐negative/PR‐negative tumors in a recent EPIC study (10,197 breast cancer
cases)173.
In a large analysis, pooling repeated measures over 30 years of follow-up from two
large prospective cohorts, total fruit and vegetable consumption were associated with
lower breast cancer incidence, particularly the more aggressive tumors including ER‐
negative, HER2‐enriched and basal‐like tumors174. This study was the first to evaluate
the risk of breast cancer by molecular subtype in relation to fruit and vegetable
consumption. Moreover, as most studies assess diet with a single questionnaire at
baseline, and in many of them, follow‐up is relatively short, researchers performed
latency analyses174. Results showed weak associations with 0–8-year lags, illustrating
that fruit and vegetable intake may be essential eight or more years before diagnosis 174.
This is consistent with an effect acting in the early stages of carcinogenesis. The
protective effect of fruits and vegetables seems to be linked to their high content of
beneficial substances such as vitamins, minerals, salicylates, phytosterols,
175
glucosinolates, polyphenols, sulfides and lectins .
45
Carotenoids have been hypothesized to reduce cancer risk through antioxidant or
antiproliferative activity176. Prior studies showed that both dietary intake and circulating
levels of carotenoids were inversely associated with breast cancer, especially ER‐
negative tumors177–179. A large pooled analysis179 of 8 cohort studies (comprising >80%
of the world's published prospective data on this topic) found that circulating carotenoid
concentration was inversely associated with breast cancer development, especially with
ER-negative disease. Recently, a nested case-control study within the NHS and NHSII
concluded that the inverse associations between circulating carotenoids and breast
cancer risk appeared to be more noticeable in high-risk women, as defined by germline
genetic makeup or mammographic density180.
Moreover, polyphenols are a wide and diverse family of more than 8,000 natural
substances present in plant foods (i.e., cereals, vegetables, fruits, nuts, legumes) and
beverages (i.e., tea, coffee, cocoa). Even though they are not essential for short-term
well-being as it is the case for vitamins, they have shown beneficial effects on the
incidence of chronic diseases, including cancer181. According to their chemical structure,
polyphenols are often grouped into four main classes: flavonoids, phenolic acids,
lignans, and stilbenes182. Among these, phenolic acids are further divided into
subclasses—hydroxybenzoic and hydroxycinnamic acids—, composed of individual
compounds —3-, 4-, 5- caffeoylquinic acids (CQA) —also known as chlorogenic acids.
Various reviews and meta-analyses of observational studies have summarized and

analyzed the available evidence on the relationship between polyphenol-rich foods and
cancer risk183–185. While the majority of studies regarding classes of polyphenols have
focused on flavonoid intake (one of the major contributors to total polyphenol intake) 184,
very few epidemiological studies have evaluated the associations of dietary phenolic
acids and their compounds on chronic disease. Indeed, there is only one case-control
study that found an inverse association between phenolic acids and a hormone-related
cancer, specifically, prostate cancer186. Recently, Gardeazabal et al.,187 reported an
inverse association between phenolic acids intake and breast cancer risk among
postmenopausal women in the SUN project. Phenolic acids contribute to a large
proportion of the total polyphenol intake consumption. Indeed, on average about 58% of
polyphenols are ingested as phenolic acids across Europe188. Current interest on the
relationship between these phenolic acids and breast cancer risk stems from their
promising protective role against cancer cell growth, oxidative stress, inflammation and
diabetes in experimental studies 189–192. However, evidence in humans remains still
inconclusive and controversial.
46
In contrast to cancer protective dietary patterns, patterns which tended to reflect a
more Westernized type of diet and also less healthy plant-based foods positively
weighted in unhealthful plant-based diets have been associated with higher risk of breast
cancer in other studies. Specifically, a meta-analysis of breast cancer reviewed 14 cohort
and 18 case-control studies158 and concluded that there is an increased risk with the
“Western” pattern (RR 1.14; 95% CI 1.02,1.28) thought the association was limited to
case-control studies and not cohort studies. These results suggest that recall bias might
at least partially explain the discrepant results in different study designs. In addition, the
Western dietary pattern was significantly associated with an increased risk of ER-positive
and/or PR-positive, but not ER-negative and/or PR-negative breast tumors. Moreover, a
new umbrella review indicated that the Western diet, which involves a high intake of
refined sugars, saturated fats and alcohol, is strongly associated with an increased risk
of breast cancer193,194.
Among all the individual food products included in a Western dietary pattern, SSBs
have been consistently associated with weight gain195, type 2 diabetes196,
hypertension197, coronary heart disease 198,199, stroke200 and mortality201 in
epidemiological studies. Independent from the obesity and adiposity pathways, SSBs
also lead to activation of the insulin signaling pathway as well as elevation of markers of
oxidative stress and inflammation, which jointly may raise cancer risk 202, especially those
related to breast cancer203,204, hepatocellular cancer205 and diabetes-related carcinomas
(including bladder, breast, colon-rectum, endometrium, liver and pancreas)206. Artificially
sweetened beverages (ASBs) are marketed as healthier alternatives to SSBs, but much
remains to be answered in regard to their long-term health effects. ASBs consumption,
the main source of artificial sweeteners in the diet, has dramatically increased in the last
40 years in children and adults 207. In a recent meta-analysis, ASB consumption was
linearly associated with risk of obesity and type 2 diabetes and nonlinearly associated
with risk of hypertension and all-cause mortality208, however, reverse causality cannot
be ruled out as driving factors in the observations.
In terms of breast cancer, only four prospective studies on SSBs and breast cancer
risk have been published showing contrasting results: one which suggested an increased
risk in premenopausal women (NutriNet-Santé prospective cohort209, 693 cases), one
which suggested an increased risk in postmenopausal women (Melbourne Collaborative
Cohort Study210, 946 cases), and two further studies which observed no association
(Framingham Offspring cohort211, 124 cases; Canadian Study of Diet, Lifestyle, and
Health212, 848 cases). Furthermore, only two studies209,210 have prospectively assessed
47
the relation between ASBs and breast cancer incidence, and results showed null
associations.
Apart from SSBs, associations between carbohydrate sources such as total

sugar, added sugar, fructose, sucrose, sugary foods 166,211, dietary fiber213, whole
grains214, glycemic index (GI) or glycemic load (GL)204 and breast cancer have been
previously reported. However, although dietary carbohydrates are the only food
constituents that directly impact blood glucose and insulin levels, carbohydrate quality
seems to have a more important role than the total amount of carbohydrate as a
percentage of dietary energy in chronic disease prevention 215. Several indices for
carbohydrate quality –based on GI, GL, fiber, whole grains, and other factors
dimensions– have been proposed215. Meta-analyses have found that high GI and GL
were weakly associated with increased risk of breast cancer 216. In a later study,
researchers from the EPIC-Italy study found that neither GI nor GL had any significant
association with breast cancer risk217. A multidimensional approach, considering at the
same time the GI, fiber content, solid or liquid carbohydrates and the degree of
processing, represents an ideal approach to assess the potential relationship between
the nutritional quality of carbohydrates and disease outcomes, even to a better extent
than just considering GI or GL. In this regard, several studies have previously used this
approach in relation to cardiovascular disease 218, obesity219 or micronutrient
adequacy220, otherwise evidence on breast cancer incidence remains unknown.
Fat quality may also exert an important role in breast carcinogenesis. Most
studies have shown no link between a high-fat diet in adulthood and an increased risk of
breast cancer221–226. However, the type of fat, rather than the total amount of fat, may be
related to breast cancer risk221,222,224,226. There is growing evidence for a plausible role
for dietary fat composition in breast cancer pathophysiology 227 which may differ
according to the menopausal status of the women, as is the case with adiposity in which
the direction of the association is reversed by menopause 228,229.
The influence of monounsaturated fatty acids (MUFAs) on breast cancer is still

controversial, although animal models have reported a potential protective effect on
breast cancer230, most epidemiological studies have not found clear associations231.
Discrepancies in results may be due to different contributing foods to total MUFA
intake232, i.e., MUFA intake from olive oil in Mediterranean populations vs. MUFA intake
from hydrogenated fat high in artificial trans-fatty acids (as in margarine), more common
in Western Europe countries. Moreover, high intake of saturated fat (but not total,
48
monounsaturated or polyunsaturated fat intake) was associated with increased risk of
breast cancer in a large European multicenter prospective study (519,978 participants) 233
and also in a French prospective study 234.
Regarding the influence of polyunsaturated fatty acids (PUFAs) intake on breast

cancer development, the most recent meta-analyses combining studies did not find an
association231. A recent meta-analysis, however, found an inverse relation between fish
ω-3 PUFAs consumption and breast cancer risk in Asian patients (OR 0.80, 95% CI 0.73,
0.87)235.
Comprehensive systematic reviews of observational studies suggested no

relationship between total polyunsaturated fat intake and risk of breast cancer 236 or ω-3
intake and breast cancer risk237. A meta-analysis of prospective cohort studies from 2016
did not find an association between ω-6 PUFAs intake and breast cancer risk231,
however, a recent analysis indicated that higher dietary intake ratio of ω-3/ω-6 is
associated with a lower risk of breast cancer in Asian countries rather than in Western
countries238.
Furthermore, evidence on the relationship between dietary protein intake and

breast cancer incidence remains inconsistent. In the WHI, vegetable protein intake was
associated with statistically significantly lower breast cancer incidence whereas higher
animal protein intake was associated with higher breast cancer risk 239. To date, two
meta-analyses of cohort studies240,241 have evaluated associations of protein source
intake and breast cancer incidence. In the first report of 8 cohort studies240, total red meat
was associated with higher breast cancer risk. No association was found for poultry, fish,
egg, nuts, total milk and whole-milk intake. The other report including 13 cohort, 3 nested
case-control studies and 2 clinical trials241 found that processed meat was associated
with 9% higher breast cancer risk when comparing extreme categories.
5.2.3. Alcohol consumption

The WCRF report concluded in 2018 that, based on the existing literature (16
prospective studies of premenopausal breast cancer and 34 of postmenopausal
disease), alcohol consumption is a “probable cause” for premenopausal breast cancer
and a “convincing cause” postmenopausal breast cancer. The meta-analysis showed
that for a 10-gram increase in alcohol consumed per day on average, risk increased 5%
among premenopausal women and 9% among postmenopausal women. A standard
drink contains approximately 14 grams of alcohol.
49
The WCRF-AICR meta-analysis also reported some differences by location5. For
premenopausal breast cancer, the summary meta-analysis was statistically significant
only for North America. Results were similar in magnitude but not statistically significant
for analyses of findings from Europe and Asia. For postmenopausal cancer, in the dose-
response meta-analysis, the association was statistically significant only for studies of
Europe and North America.
When examining the effects of alcohol consumption on health and disease, how
participants consumed the alcohol must be considered. Not only the absolute quantity
consumed, but also the intensity of consumption may have biological effects. For
example, the effects of an average consumption of seven drinks per week may differ for
consumption of one drink daily and for seven drinks on one day once per week. Just a
few studies have examined drinking intensity. In the Nurses’ Health Study (NHS), binge
drinking (defined as six or more drinks in one day) was associated with increased risk,
even after adjusting for total consumption 242. The frequency of alcohol consumption was
not associated with risk in that cohort after adjusting for total consumption. In the Sister
Study, a cohort of women with a family history of breast cancer, self-report of ever binge
drinking (defined as four or more drinks in one sitting) or ever having passed out while
drinking were associated with increased breast cancer risk 243, though these associations
were not adjusted for overall alcohol intake. Recently, in the SUN Project, a twofold
higher risk of premenopausal breast cancer was associated with binge drinking habit 244.
Even among people who drink lightly, evidence of increased risk has been
reported. In a systematic review of light drinking, which used the World Health
Organization definition of consumption of less than 21 grams of alcohol per day, Shield
et al.245 found consistent evidence of increased risk. In a meta-analysis, Choi et al.246
found statistically significant relative increases in risk of 4% for participants who drank
0.5 drinks per day, 9% for participants who drank less than less than or equal to 1 drink
per day, and 13% for individuals who drank 1 to 2 drinks per day; in this analysis, one
drink was defined as 12.5 grams of alcohol. There is no evidence of a lower threshold
for an effect of alcohol consumption on risk of breast cancer. Jointly, results from these
studies on intake indicate that drinking pattern may affect risk, as drinks per drinking day
are associated with increased risk even after adjusting for total intake.
50
5.2.4. Smoking
Smoking increases the risk of some types of cancer (i.e., lung, kidney and
pancreas cancers). Albeit findings on a possible link to breast cancer remain
heterogeneous, there is growing evidence smoking may slightly increase breast cancer
risk. While some studies have shown that women smoking before having their first child
may increase their breast cancer risk247–251, others have found no link252,253. More
research is required before solid conclusions can be made concerning the potential link
between smoking and breast cancer.
5.2.5. Exercise
In recent years, epidemiological studies have been conducted on the relationship
between physical activity and cancer outcomes, and results suggested a protective role
of physical activity in breast cancer 254,255. Women who exercise regularly showed a lower
risk of breast cancer than women who were not active55,256–259. Overall, periodic exercise
seems to lower breast cancer risk by 10-20 % in relative terms55,256–259, which appears
to be most clear among postmenopausal women55,256–259. Activity equal to walking 30
minutes a day may lower breast cancer risk by about 3%260.The American Cancer
Society recommends 150 minutes (2½ hours) of physical activity a week to lower overall
cancer risk261.
Physical activity during adulthood lowers breast cancer risk256–259,262.

Researchers are now evaluating whether staying active during childhood and the teen
years may also lower breast cancer risk. Findings to date are heterogeneous263,264. Some
studies suggest women who were more active as children and teens may have a lower
risk of breast cancer compared to women who were less active 264. Other findings show
physical activity during childhood and the teen years may marginally lower the risk of
breast cancer before menopause while no impact after menopause is suspected 263.
The protective effect of physical activity on breast cancer is explained by long-

term regulation of various circulatory risk factors such as sex hormones, metabolic
hormones, and inflammatory factors. Moreover, physical activity largely changes whole-
body homeostasis by inducing numerous other factors as well as by modifying the gut
microbiota profile. However, as breast cancer is commonly contemplated as a
heterogeneous disease, the influence of physical activity may diverge determined by the
clinicopathologic features (e.g., tumor stage and hormone receptor status) and body
composition (e.g., fat and skeletal muscle mass), generating different biological and
molecular mechanisms and the diverse and restricted outcomes of breast cancer265.
51
Moreover, even though the evidence for the benefits of physical activity in breast
cancer is expanding, most of the studies available have not considered all the
components of exercise prescription including frequency, intensity, time, and type265.
To conclude, the biology of physical activity and exercise is highly intricate and
shifty, and influences multiple organ systems through autocrine, paracrine, and
endocrine factors within a crosstalk network. Thus, an understanding of these integrative
mechanisms will enhance how physical activity can ultimately influence the risk and
prognosis of various cancers, including breast cancer.
5.2.6. Oral contraceptive use
The possibility that the use of hormonal contraceptives may increase the risk of
breast cancer has been raised for many years266. Current or recent use of birth control
pills or oral contraceptives increases the risk of breast cancer79,267–269. Some other
contraceptives contain -or release- hormones. A possible increase in risk has been found
among current, longer-term users compared to women who never used Depo Provera270.
Studies show while women are taking birth control pills (and shortly after) their breast
cancer risk is 20-30% higher in relative terms than women who have never used the
pill267–269. Once women stop taking the pill, their risk of breast cancer begins to
decrease267,269. Overall, today’s lower-dose pills increase breast cancer risk just like older
forms of the pill269. Two recent papers showed a statistically significant increase in breast
cancer with use of hormonal contraception, even with contemporary lose-dose
formulations269,271. Particularly, in the Danish Sex Hormone Register Study269, the current
or recent use of hormonal contraception was associated with a higher risk of breast
cancer compared to the risk among women who had never used hormonal
contraceptives.
Findings on hormone-releasing Intrauterine devices (IUDs) and breast cancer are

mixed. Some studies showed these IUDs have no impact on breast cancer risk272. Others
show hormone-releasing IUDs increase risk by about 20% (similar to birth control
pills)269. Other findings suggest past use of hormone-releasing IUDs may increase breast
cancer risk after menopause273.
The birth control patch and vaginal ring do not appear to increase breast cancer
risk269. However, data are limited.
52
5.2.7. Age at first birth and number of childbirths
The link between having children and the risk of breast cancer is complex. It is
related to age at first childbirth and the number of childbirths a woman has. A first
pregnancy has 2 effects on breast cancer risk, it increases short-term risk and then it
lowers long-term risk. The global impact of these risks depends on a woman’s age at
the time of her first pregnancy79,274,275.
In general, women who give birth to their first child at later ages have a higher risk of
breast cancer compared to women who have their first child at younger ages79,274,275. For
example, women who give birth for the first time after age 35 are about 40% more likely
to get breast cancer than women who have their first child before age 20 276. For these
women, the increased risk from a first pregnancy is never fully offset by its long-term
protective effect277. Breast cancer risk is increased for about 10 years after a first
birth277. Then, it drops below the risk of women who do not have children 277.
• Women who give birth to their first child at age 35 or younger tend to get a protective
benefit from pregnancy79,274,275.
• Women who are over age 35 when they give birth to their first child also have a small
increased lifetime risk of breast cancer compared to those who never give birth 278.
One possible reason for the different effects of age at first childbirth relates to breast
cells:
• During pregnancy, breast cells grow fast.
• If there is any genetic damage in the breast cells anterior to pregnancy, it is
copied as the cells grow.
• This elevated genetic damage in the cells can lead to breast cancer.
• The chance of getting such genetic damage increases with age.
Overall, the more childbirths a woman has, the lower her risk of breast cancer turns
out to be79,274,275. After the first child, each childbirth reduces the risk of breast cancer.
Some findings show that giving birth only lowers the risk of ER-positive breast cancer,
with no apparent impact on the risk of ER-negative (along with triple negative) breast
cancers274,279.
Women whose childbirths are distributed close together may grab more benefit than
those whose childbirths are spaced far apart79. Although the reasons are unclear, they
may be related to changes in breast cells that take place during pregnancy.
53
5.2.8. Breastfeeding
Breastfeeding lowers the risk of breast cancer, especially among premenopausal
women280,281. A pooled analysis of data from 47 studies found that compared to mothers
who never breastfed 281, women who breastfed for a lifetime total (combined duration of
breastfeeding for all children) of:
• 1 year were slightly less likely to get breast cancer.

• 2 years got about twice the benefit of those who breastfed for a total of 1 year.
• > 2 years got the most benefit
Although data are limited, breastfeeding for less than one year may also lower breast
cancer risk281. Regarding subtypes, breastfeeding may be able to lower the risk of ER-
negative, including triple negative breast cancers279,282,283.
5.2.9. Hormone replacement therapy (HRT)

HRT has been used mostly in western countries, with about 600 million woman-years
of use since 1970284,285. During the 1990s, its used increased rapidly and halved abruptly
in the early 2000s, and then stabilized during the 2010s with about 12 million current
users. Women tend to begin HRT at around the time of the menopause and can take
them many more years. At these ages, breast cancer is the most common malignancy
in western countries. While regulatory bodies in Europe and in the US recommend that
HRT should be used for the shortest time that it is needed, some clinical guidelines
recommend less restrictive prescribing286,287. There are 2 main types of HRT:
• Estrogen plus progestin HRT increases the risk of breast cancer incidence and
death102,285. When women take estrogen plus progestin within the first year of use,
their risk of having an abnormal mammogram raises288 while their risk of breast
cancer increases within the first 5 years of use 288,289. In fact, the risk of breast cancer
goes up slightly with each year a woman takes estrogen plus progestin 285. Some
large-scale observational studies have suggested women who use estrogen plus
progestin for 5 or more years (and are currently taking it) have a twofold increase on
breast cancer289,290. Nonetheless, as soon as a woman pauses taking estrogen plus
progestin, their risk of breast cancer starts to downturn. After approximately 5-10
years, the risk turn back to that of a woman who has never used HRT102,288,290,291 as
shown in the Millions Women Study and in the WHI.
54
• Estrogen alone increases the risk of uterine cancer. Therefore, it is only used by
women who no longer have a uterus because of a hysterectomy. If a woman still has
a uterus, they would use estrogen plus progestin.
Some studies have suggested estrogen alone increases the risk of breast cancer
by about 30%102,285,289,290. However, the WHI suggested a decreased risk of breast
cancer with estrogen alone compared to placebo after 7 years of use (on average)292.
In this study, most women began using estrogen many years after menopause, in
lieu of at the time of menopause. Another factor delineating the differences between
randomized trials and observational studies is the age at first use and the associated
concept of time from menopause (gap time) initiation293. Given the suggestion of the
gap time effect, it is uncertain whether women initiating estrogen alone use close to
menopause would experience a favorable effect on breast cancer incidence.
The results from 2 large cohort studies, where most hormone use started at
menopause, raise the question of the safety of longer-term use of estrogen alone:
• As in the WHI, the Nurses’ Health Study found no increased risk of breast cancer
among women who used estrogen alone for less than 10 years294. However, there
was an increased risk of breast cancer after 20 years of use.
• The NIH-AARP Diet and Health Study found the use of estrogen alone for 10 or
more years increased breast cancer risk, especially in women who were thin 295.
5.2.10. Socioeconomic status
High socioeconomic status (SES) is most often defined by high income and/or high
education level. High SES has been linked to an increased risk of breast cancer 296,297.
This increased risk is not due to the higher SES itself, but to differences in risk factors
found in women of different education and income levels. For example, compared to
women of lower SES, women of higher SES are more likely to drink more alcohol, have
fewer children, have their first child at a later age, use birth control pills and use
menopausal hormone therapy (postmenopausal hormones). Each of these factors
increases the risk of breast cancer296,297.
5.2.11. Night-shift work

Some studies show women who routinely work night shifts for many years have a
small increased risk of breast cancer 108–110. More research is needed to understand
55
which aspects of shift work may impact breast cancer risk. One possible reason for the
increased risk among these workers is their exposure to light at night.
Melatonin, often referred as the sleep hormone, is a central part of the body’s sleep-
wake cycle. With evening darkness, its production increases and exposure to light at
night disrupts this process. Findings on melatonin levels in the body and breast cancer
risk are mixed298,299. However, some findings suggest that women with lower levels of
melatonin may have a higher risk of breast cancer than women with higher levels 299.
Exposure to light at night may increase the risk of breast cancer by suppressing the
production of melatonin. Lowering melatonin levels also can increase the production of
estrogen, which in turn, can increase the risk of breast cancer72. These processes in the
body and their effect on breast cancer risk among night shift workers are still under study.
Project justification
Diet-related chronic diseases, such as cardiovascular disease, type 2 diabetes,
obesity, and some types of cancer pose a major public health problem. In fact, today,
more than half of adults have one or more diet-related chronic diseases144. Irrespective
of site-specific details in the pathogenesis of tumors, up to more than 90% of cancers
are considered to be attributable to modifiable risk factors such as tobacco smoking,
excessive body weight, physical inactivity, alcohol consumption, infectious agents,
environmental pollution, and suboptimal diet136. The latter is made responsible for about
5% to 10% of total cancer cases.
On a population level, the best way to substantially decrease incidence and

mortality for any disease is through prevention. On a large scale, we have done far less
advancement preventing breast cancer than preventing its derivatives sequels. The
treatments for breast cancer (both medical and surgery) currently available, are
numerous and extensive, but to date, advances in treatment alone will never be sufficient
to completely stem the burden of breast cancer. This is why redirecting research towards
the development of measures that prevent the onset of the disease is imperative.
Several hormonal, anthropometric, lifestyle, and genetic factors are known to be

associated with breast cancer, though these associations may differ by menopausal
status and disease subtypes. While there is growing evidence for a plausible role for
dietary factors in breast cancer pathophysiology, evidence in the literature is still
inconclusive and sparse. Hence, in this thesis dissertation, we aim to further explore the
associations through which several suspected dietary risk factors may influence breast
56
cancer risk. The data sources for this thesis dissertation included three large prospective
cohort studies, the “Seguimiento Universidad de Navarra” (SUN) Project, a
Mediterranean cohort study of university graduates in Spain and the original Nurses’
Health Study (NHS) and Nurses’ Health Study II (NHSII), which are among the worldwide
largest ongoing prospective cohort studies in women.
Traditionally, nutritional epidemiology has focused on single nutrients or foods in

isolation, especially when undernutrition and nutritional deficiencies were the rampant
diet-related diseases. However, the demographic and epidemiologic transitions
experimented lately, which can be attributed in part to changes in the food environment,
have switched disease burdens300. The dietary determinants of breast cancer differ
substantially from those of undernutrition and nutrient deficiencies that results from
insufficient intake or absorption of particular nutrients. It is, then, important to go beyond
the study of single-nutrient approach and evaluate the effect of the whole dietary pattern.
As food items and nutrients are consumed in combination, dietary patterns have
been successfully implemented as a tool to assess the additive or synergistic effect of
food in nutritional epidemiology139. In particular, provegetarian food pattern scores have
evolved to assess the level of adherence to a vegetarian-like diet with preference of
plant-based foods. Unlike the vegetarian dietary patterns, these scores consider
moderate intakes of animal foods (fish, poultry, and dairy) in the assessment of the score
as they may be easier to adhere to and in the event that a modest consumption of these
foods may confer some health benefits 301 and distinguish between healthful and
unhealthful versions. A priori defined dietary patterns quantitatively measure the quality
of dietary habits self-limited to the existent and current knowledge on diet and disease.
Even food groups used in most a priori scores are usually widely defined and do not take
into consideration subtle differences in the nutrient profile and other chemicals in
individual foods items, and also, they do not reflect the characteristics of a given cohort
group, it more promptly lends itself to application across populations.
The examination of existing dietary patterns and their association with breast
cancer risk could provide a wider and more realistic estimation of the relation between
food habits and health in a particular population. Although considering the eating pattern
as a whole has advantages, it cannot be as informative as other approaches when it
comes to elucidating the possible relationships between diet and breast cancer. So, in
order to obtain good-quality scientific evidence, we also need to address the various
components individually. In this sense, there are certain foods within dietary patterns that
57
are highly consumed. Taking this argument further, the concept of nutrient-poor or
discretionary foods exposes the significance of nutrients in the diet health debate. Foods
are considered to be discretionary when, overall, they provide so many calories and
virtually no other nutrients such as SSBs. In a large part of the world, sugary drink
consumption is rising dramatically due to widespread urbanization and beverage
marketing302. The impact of SSBs on cardiometabolic health is well studied, they are
associated with an increased risk of weight gain, obesity, type 2 diabetes (independently
of adiposity), hypertension, and cardiometabolic death 302. Nonetheless, discretionary
foods deserve separate and more detailed consideration in the development of dietary
guidelines and research studies, particularly for diseases less studied as breast cancer.
Moreover, nutrient-based research should be considered to explain the

mechanisms by which certain foods or dietary patterns exert their effects. In fact,
nutrient-focused research enhances the mechanistic understanding of food and diet
effects. In this regard, there are specific components present in foods that also might
have protective effects. Given the mounting evidence on the relevant role that a high
intake of polyphenols in the diet can play for breast cancer prevention, we plan to
complement our study with the assessment of the association between phenolic acids,
which we previously identified to be protective 187, and their subclasses and individual
compounds and the risk of developing breast cancer in the SUN Project.
Last but not least, macronutrient research in terms of breast cancer risk has
mainly focused on the quantity, but few studies have specifically addressed the potential
link between the quality of the carbohydrates consumed with the risk of breast cancer.
Moreover, dietary fat has been suggested as a risk factor for breast cancer. Nonetheless,
epidemiological studies have produced conflicting results, and data on humans are
sparse and inconsistent. Moreover, fatty acids are heterogeneous, with varied chemical
structures, functions, and effects on the human body, so investigating the impacts of
individual types of fat is important.
Because estrogens have long been hypothesized to play an essential role in

breast cancer development and the source and metabolic pathway of estrogens are
different between premenopausal and postmenopausal women 303, the etiology and risk
factors of breast cancer differs by menopausal status. Hence, it seems relevant to take
this into consideration and examine diet exposures by menopausal status. Evidence for
hormonal and molecular subtypes of breast cancer is additionally lacking, despite
subtypes having potentially distinct etiologies304–306. Hormone receptor status is an
58
important diagnostic and prognostic characteristic of breast tumor and, therefore, merits
consideration. Given the opportunity to work within the Nurses’ Health Studies, we could
expand our analysis to cancer subtypes in two of our hypotheses.
Primary prevention of breast cancer is a new frontier in healthcare, which

necessitates that we build a model of integrated methods to conduct research on risk at
the population level, the cell biology level, the technology level and the social science
level. Overall, although many epidemiological and biological observations suggest that
diet at different periods of life is related to breast cancer risk, current epidemiological
studies have not yet clearly identified which specific foods or nutrients play the key roles.
Diet and food are prime environmental factors that are modifiable, and thus, amenable,
more accessible for public health messaging following by understanding their
relationship to primary breast cancer prevention.
59
HYPOTHESIS AND AIMS
60
1. Hypothesis
The central hypothesis was that diet may play an important role in breast cancer
development and there may be certain dietary patterns, foods and nutrients that could
be driving these associations.
• Frequent consumption of sugar-sweetened beverages will be associated with a

higher risk of breast cancer, independently of known confounders.
• Phenolic acids subclasses and their primary individual compounds will be associated
with lower risk of breast cancer.
• A better quality of the carbohydrates measured by the carbohydrate quality index will
be associated with a lower breast cancer risk.
• Monounsaturated and ω-3 polyunsaturated fatty acids may be associated with lower
risk of breast cancer while saturated fat intake would increase its risk.
• Higher adherence to an overall plant-based diet, specially, its healthful version, will
be associated with a decreased breast cancer risk while adherence to an unhealthful
version will be associated with a higher breast cancer risk.
2. General aim
In this thesis dissertation, our overall aim was to provide insight into the role of
diet on breast cancer incidence.
3. Specific aims
1. To evaluate the association between sugar-sweetened beverages consumption and

breast cancer risk in the Seguimiento Universidad de Navarra cohort study.
2. To examine the association between sugar-sweetened beverages and artificially
sweetened beverage consumption and breast cancer risk in the Nurses’ Health
Studies (NHS and NHSII).
3. To assess the association between subclasses and individual compounds of
phenolic acids intake with risk of breast cancer in the Seguimiento Universidad de
Navarra cohort study.
4. To evaluate the association between an a priori defined carbohydrate quality index,
defined by four criteria: dietary fiber intake, glycemic index, whole grains/total grains
ratio and solid carbohydrates/total carbohydrates ratio, and the incidence of breast
cancer in the Seguimiento Universidad de Navarra cohort study.
5. To examine the association between total and subtypes of dietary fat and breast
cancer risk in the Seguimiento Universidad de Navarra cohort study.
61
6. To assess the relationship between an overall provegetarian food pattern and breast
cancer incidence in the Seguimiento Universidad de Navarra cohort study.
7. To examine the association between healthful and unhealthful provegetarian food
patterns and breast cancer incidence in the Seguimiento Universidad de Navarra
cohort study.
8. To evaluate the association between an overall plant-based dietary pattern and
breast cancer incidence in the Nurses’ Health Studies (NHS and NHSII).
9. To assess the association between healthful and unhealthful plant-based diets and
breast cancer incidence in the Nurses’ Health Studies (NHS and NHSII).
62
METHODS
63
1. SUN Project
This thesis dissertation is framed within the “Seguimiento Universidad de Navarra”

(SUN) Project. It is a dynamic, permanently open, prospective cohort, consisting of more
than 22,500 university graduates as of December 2020. Its overall aim is to evaluate the
impact of diet, with a particular focus on the Mediterranean Diet, and lifestyle on the
prevention of non-communicable diseases such as cardiovascular disease and cancer,
among others.
The SUN Project was designed in collaboration with the “Harvard T. H. Chan School
of Public Health”, using a methodology similar to that of the US “Nurses’ Health Studies”
or the “Health Professionals Follow-up Study”.
1.1. Recruitment
Since December 1999 university graduates from the University of Navarra and other
universities around Spain, are invited to participate. Participants are all university
graduates, and the recruitment is permanently open. The decision to include highly
educated participants yields on the approach known as restriction in epidemiology, as
described by Rothman et al.,307 which represents an excellent technique to prevent or at
least reduce confounding by known factors, and obtain high-quality data from
participants.
The recruitment of participants has been carried out in various groups:

• University of Navarra Alumni.
• Navarra’s Association of Nurses.
• University graduate members from the health insurer of the University Clinic
[University Clinic Assistance of Navarra (ACUNSA)].
• Parents (with a college degree) of current college students from the University
of Navarra.
• Members of other professional associations from various Spanish provinces.
All these members received an invitation letter, briefly explaining the study
objectives, what their participation would imply and what information would be asked
from them over the follow-up time. Along with the invitation letter, the baseline
questionnaire and a pre-stamped envelope were provided. Since 2004, potential
participants can also receive a personal code via email to answer the questionnaire at
the SUN website.
64
The tedious nature of the baseline questionnaire is meant to increase the retention
proportion, although it may reduce the participation proportion. Moreover, this initial
response includes the following: three addresses (home, work and relative, or friend),
telephone number, and email address (if possible). This information is essential to avoid
future attrition.
Upon completion of the first questionnaire, the participant becomes a member of the
cohort (see questionnaires in Appendices). The questions in the baseline questionnaire
are intended to gather participants' socio-demographic variables, lifestyle factors, as well
as information related to health and diet.
1.2. Follow-up
After the initial questionnaire, additional questionnaires (see questionnaires in
Appendices) are mailed every 2 years in order to follow-up the participant and track new
disease diagnoses and changes in food habits, and other lifestyle aspects.
Anthropometric measurements are self-reported and when measurements are
requested (after 6 and 8 years), a measuring tape and instructions are mailed with the
questionnaire to allow the participants to measure their waist and hip circumferences. In
certain follow-up questionnaires there are questions on diet, but it is not until the 10 th
year of follow-up when a new FFQ is distributed. Follow-up questionnaires are shorter
and less detailed than the initial questionnaire regarding exposures.
As previously described, for each participant, three postal addresses, an email and
phone numbers are requested. Up to five further mails are sent to non-respondents,
followed by a mailed brief questionnaire and eventually, they are contacted by email or
phone. If participants are lost to follow-up, deaths confirmations and their causes are
checked every year at the Spanish National Death Index and the National Statistics
Institute (please see308 for a comparison between the two official Spanish sources of
information on mortality). The positive predictive value for these sources of information
regarding fatal events is very high, expected to be around 100%. Families and postal
authorities also report deaths, which are confirmed by medical records review (with next
of kin permission).
In February 2018, the overall retention proportion in the cohort was 91% (any follow-
up), 88% at Q2 (n=18,932), 81% at Q4 (n=17,200), 82% at Q6 (n=15,327), 81% at Q8
(n=13,863), 81% at Q10 (n=12,234), 84% at Q12 (n=10,145) and 78% at Q14 (n=6,024).
As Alonso et al.,309,310 reported differences in follow-up in the SUN cohort did not seem
65
to cause a bias and did not necessarily affect the magnitude or direction of the
association estimates.
The baseline questionnaire and the follow-up questionnaires include information on

sociodemographics (sex, birth date, marital status), anthropometrics (weight, weight
change, birth weight, height, body image, waist circumference), quality of life (health self-
perception, health quality), diet, eating behaviors, lifestyle (smoking, drinking, driving,
sedentarism), clinical data and family/childhood history, disease diagnosis, preventive
screenings, preventive strategies (sunscreen, sea belt, airbag or helmet), medications
and supplements, personality traits, feelings and emotions, and physical activity
(validated questionnaire 311). Metabolic equivalents (METs) for each participant were
estimated to generate METs-h/week scores. Self-reported weight and height to yield
body mass index was previously validated in a subsample of the cohort 312.
1.3. Information processing

Participants have the possibility to fill in the questionnaire using a hard copy they get
by ordinary mail, and then returned it using a pre-stamped envelope or an online version
at the SUN website. With the online option, they can use the temporary identification
information and password, also included in the mailing, to access the SUN website,
where the participant is then directed to the appropriate questionnaire (depending on the
follow-up cycle).
In the time between follow-up questionnaires, participants are sent a newsletter

detailing the current scientific progress and a Christmas card to remind and thank them
for their dedication, commitment, and continued participation in the project. This card is
also meant to verify any changes in mailing addresses.
Project managers evaluate all completed questionnaires for consistency and quality.
After data have been cleared for quality, the information is uploaded on a secured web-
based database, where all participating researchers can download applicable
information in order to conduct proper analyses.
66
1.4. Ethical standards
The present studies were conducted according to the guidelines laid down in the
Declaration of Helsinki and all procedures involving participants were approved by the
Institutional Review Board of the University of Navarra. All potential participants were
duly informed of their right to refuse to participate in the SUN study or to withdraw their
consent to participate at any time without reprisal. Special attention was given to the
specific information needs of individual potential candidates as well as to the methods
used to deliver their information and the feedback that may receive in the future from the
research team. After ensuring that the candidate had understood the information,
researchers sought their potential freely given informed consent through their voluntary
completion of the baseline questionnaire. These methods were accepted by the
Institutional Review Board as to imply an appropriately obtained informed consent.
The study protocol was approved by the Clinical Research Committee of the
University of Navarra. The response to the questionnaire was considered to be
equivalent to the informed consent of the individuals to participate in the study.
1.5. Funding sources

The SUN Project has received funding from the Spanish Government-Instituto de
Salud Carlos III, the European Regional Development Fund (FEDER) (RD 06/0045,
CIBER-OBN, Grants PI10/02658, PI10/02293, PI13/00615, PI14/01668, PI14/01798,
PI14/01764, PI17/01795, and G03/140), the Navarra Regional Government (45/2011,
122/2014, 41/2016), and the University of Navarra. Other funding sources can be found
at: https://www.unav.edu/en/web/departamento-de-medicina-preventiva-y-salud-
publica/proyecto-sun/proyectos-y-financiacion
1.6. Exposure assessment

Diet is assessed using a 136-item semiquantitative FFQ. This questionnaire
assessed food habits in the previous year. The SUN cohort FFQ has been repeatedly
validated in Spanish participants313–315 and it has been also shown to correlate well with
biomarkers of food intake316. Each food item shows a typical portion size. Daily intake is
calculated multiplying the portion size by the frequency of consumption (never/seldom,
1–3 servings/month, 1 serving/week, 2–4 servings/week, 5–6 servings/week, 1
serving/day, 2–3 servings/day, 4–5 servings/day, and ≥ 6 serving/day). Nutrient intake is
estimated using Spanish Food Composition Tables317,318. A trained team of dietitians
using available information of food composition tables for Spain derived the nutrient
intakes.
67
1.6.1. Sugar-sweetened beverages consumption
The validated semiquantitative FFQ included a specific item on SSB consumption
(soda drinks) with 9 possible answers ranging from never or seldom to 6 or more times
per day. The baseline FFQ also collected information about habits of other beverage
consumption and included specific items for spirits, total wine, red wine, beer, diet drinks,
coffee, decaffeinated coffee, natural orange juice, other natural juices, canned juices,
whole-fat and low-fat milk, tap water and bottled water.
We did not consider artificially sweetened (non-caloric) beverages. For the main
analysis, participants were classified into three categories of consumption: never/seldom
drinkers, > 1 serving/month to ≤1 serving/week and > 1 servings/week. Due to the limited
consumption of SSBs and as a result of the small number of incident cases, we
categorized frequency of consumption as never/seldom (< 1 serving/month) or any
regular consumption (≥ 1 serving/month) when we stratified the analyses by menopausal
status.
1.6.2. Phenolic acids intake

The Phenol-Explorer database (www.phenol-explorer.eu) was used to collect data
on the polyphenolic content of the foods included in the FFQ. They were expressed in
mg/100 g (solid foods and oils) or in mg/100 mL (beverages). The agreement between
these food items and the Phenol-Explorer database was assessed following these steps:
1. Foods with merely traces or without polyphenols were deleted.

2. Recipes and processed foods, which comprised several food components, were
divided according to their ingredients.
3. Retention factors from Phenol-Explorer were used to measure the impact of food
processing and cooking on polyphenol content values.
4. Some products that may contain refined wheat flour were introduced in the list of
foods (i.e., pizza, pastries) and phenolic acid content was obtained from their wheat
flour contents.
Details of calculation of total and subtypes of polyphenols intake have been
previously described319. Total phenolic acid intake was calculated as the sum of all
subclasses of phenolic acid intakes (hydroxybenzoic acids, hydroxycinnamic acids,
hydroxyphenylacetic acids, and hydroxyphenylpropanoic acids) from all food sources.
68
We also checked 3-, 4-, and 5-CQA, often grouped as chlorogenic acids, as 98% of
the between-person variability in total hydroxycinnamic acids intake was explained by
these specific acids. Finally, intake of phenolic acids, their subclasses, and individual
compounds were adjusted for total energy intake using the residual method and
participants were then grouped into tertiles.
1.6.3. Total and subtypes of fat intake

Using the information from the FFQ, the percentage of energy from each type of
dietary fat was calculated by dividing energy intake from each fat by total energy intake.
The reproducibility and validity of intakes of fat and individual fat-contributing foods by
FFQ have been evaluated elsewhere 313. It showed good validity for assessing fat intake
(energy-adjusted intraclass correlation coefficients for different types of fats versus four
72-hour food records ranged from 0.49 to 0.75) 315.
Baseline intake of every dietary fat: monounsaturated fat, saturated fat and trans-
fat were analyzed as continuous variables (per 1% increase in total fat intake) and
categorized into tertiles. We used the multivariate nutrient density model where the
exposure is made up with the nutrient densities (% of energy) from diverse fat subtypes
and the covariate comprise the total energy320. This method supports a wieldy separation
of diet into its composition and total amount and seems to deserve more comprehensive
use. This approach is preferable when the main exposure is categorized and can
calculate relative risks, avoids misspecification of the model and decreases influences
of outliers321.
1.6.4. Healthful and unhealthful provegetarian dietary patterns

Using the dietary data available from the FFQ, we built an overall provegetarian
(PVG) food pattern previously defined by Martínez-Gonzalez et al.,322 and two fitted
versions of a healthful and unhealthful PVG food patterns, as proposed by Satija et al.,167.
We distinguished between healthy and less healthy plant foods based on existing
epidemiologic knowledge on the association between foods and type 2 diabetes and
cardiovascular disease 167,169,323, as well as intermediate conditions such as obesity,
hypertension, or inflammation. The food items included in the PVG food pattern are
shown in the following Table 2.
69
Table 2. Scoring criteria for the provegetarian (PVG) food pattern1.
Component
Vegetable food groups, by quintile2
1. Vegetables Carrot, Swiss chard, cauliflower, lettuce, tomatoes, green beans,
eggplant, peppers, asparagus, spinach, other fresh vegetables
2. Fruit Citrus, banana, apple, pear, strawberry, peach, cherry, fig, melon,
watermelon, grapes, kiwi, avocado, olives, dried fruit, canned fruit,
natural fruit juices
3. Legumes Lentils, chickpeas, beans, peas
4. Cereals White bread, whole-grain bread, cold breakfast cereal, rice, pasta
5. Potatoes Potato chips, French fries, boiled or baked potatoes
6. Nuts Almonds, peanuts, hazelnuts, pistachios, pine nuts, walnuts
7. Vegetable fat Olive oil, corn oil, sunflower oil and margarine
8. Pastries Sweets and desserts
9. Sugar-sweetened Sugar-sweetened beverages and bottled fruit juices
beverages
10. Coffee Regular coffee
Animal food groups, by reverse quintile3
11. Meat and meat products Beef, pork, lamb, rabbit, liver, chicken, turkey, cooked ham, Parma
ham, mortadella, salami, foie grass, spicy pork sausage, bacon, cured
meats, hamburger, hot dog
12. Animal fats for cooking or Butter, lard
as a spread
13. Eggs Eggs
14. Fish and other seafood White fish, dark-meat fish, salad or smoked fish, clams, mussels,
shrimp, squid
15. Dairy products Whole milk, skim or low-fat milk, condensed milk, cream, milk shake,
yogurt, custard, cheese, ice cream
1
The overall PVG pattern was built by summing both components with a potential range of 15–75;2 The consumption (g/d)
of each food group was transformed into energy-adjusted quintiles by using the residuals method (1 = first quintile, 2 =
second quintile, 3 = third quintile, 4 = fourth quintile, 5 = fifth quintile). The sum of quintile values across the 10 food groups
gave a potential range of 10–50;3 Consumption (g/d) was transformed into energy-adjusted quintiles (residuals), and the
quintile values were reversed (1 = fifth quintile, 2 = fourth quintile, 3 = third quintile, 4 = second quintile, 5 = first quintile).
The sum of reverse quintile values across the 5 food groups had a potential range of 5–25.
To build the score, we adjusted the consumption (g/d) of 10 plant-origin food

groups (vegetables, fruits, legumes, cereals, potatoes, nuts, vegetable fat, pastries,
SSB, and coffee) and 5 animal-origin food groups (meat and meat products, animal fats
for cooking or as a spread, eggs, fish and other seafood, and dairy products) for total
energy intake using the residual method324. The energy-adjusted estimates (residuals)
were ranked into quintiles and the values received reversed punctuations in the case of
animal products (assigning a value of 5 for the first quintile, 4 for the second quintile, and
consecutively until the value of 1 was assigned to the fifth quintile). To obtain the final
PVG food pattern, quintile values were summed, and the final score ranged theoretically
from 15 to 75. A higher score showed a higher contribution of plant foods to the diet.
Subsequently, we classified adherence to the provegetarian food pattern in tertiles for
the overall sample and separately for pre- and postmenopausal women.
70
Moreover, with the use of the approaches outlined by Satija et al.,167 we
considered 18 food groups built on nutrient and culinary similarities within the bigger
categories of healthy plant foods (vegetables, fruits, legumes, whole grains, boiled or
baked potatoes, nuts, vegetable oil, and coffee), less healthy plant foods (refined grains,
French fries, SSB, pastries, and other vegetable fat [margarine]), and animal foods (meat
and meat products, animal fats for cooking or as a spread, eggs, fish and other seafood,
and dairy products). These food groups were transformed into energy-adjusted quintiles
using the residual method. For the healthful provegetarian food pattern (hPVG), positive
scores were given to healthy plant foods, and reverse scores to both less healthy plant
foods and animal foods. Finally, for the unhealthful provegetarian food pattern (uPVG),
positive scores were given to less healthy plant foods, and reverse scores to both healthy
plant foods and animal foods. The 18 food groups scores for an individual were summed
to obtain the indices, with a theoretical range from 18 (lowest possible score) to 90
(highest possible score). The observed ranges at baseline were 22 to 66 for the PVG,
30 to 80 for hPVG and 30 to 82 for uPVG according to the data from the SUN cohort.
1.6.5. Carbohydrate quality index

We used baseline dietary intake data to determine the Carbohydrate Quality
Index (CQI) according to the approach previously proposed by Zazpe et al. 325 (Table 3).
We calculated liquid carbohydrate intake adding up SSBs and fruit juices consumption
(four items in the FFQ), while solid carbohydrate corresponded to the carbohydrate
content of the rest of foods with any carbohydrate content. Refined grains were
determined according to 13 items linked to the consumption of white rice, refined bread,
refined pasta and different bakery products made with refined grains, and then, whole
grain consumption was ascertained with the item “whole bread consumption” (serving
size 60 g). The CQI was defined by adding up the quintiles for the following four criteria:
dietary fiber intake (g/day, positively weighted), GI (negatively weighted), whole-grain
carbohydrate: total grain carbohydrate ratio (positively weighted), and, lastly, solid
carbohydrate: total carbohydrate ratio (solid + liquid carbohydrates) (positively
weighted)325. For each of these four components, we categorized participants into
quintiles and obtained a value (ranging from 1 to 5) according to each quintile (only for
GI, those in the fifth quintile received 1 point and those in the first quintile received 5
points). All criteria had the same weighting, and the total score potential range was 4 to
20, with higher values meaning better quality of carbohydrates.
71
Table 3. Criteria used to calculate carbohydrate quality in the SUN Project (1999-2018).
Index
Components of carbohydrate Criteria for minimum
range Criteria for maximum index
quality index index
(points)a
Dietary fiber intake (g/d) 1–5 Minimum dietary fiber Maximum dietary fiber intake
intake (first quintile) (fifth quintile)
Glycemic index 1–5 Maximum glycemic Minimum glycemic index (first
index (fifth quintile) quintile)
Ratio of whole-grain: total- 1–5 Minimum value of this Maximum value of this ratio
grain carbohydrates ratio (first quintile) (fifth quintile)
Ratio solid 1–5 Minimum value of this Maximum value of this ratio
carbohydrates:(solid ratio (first quintile) (fifth quintile)
carbohydrates + liquid
carbohydrates)
Total index (range) 4–20
a
Proportional dietary indices were computed for intakes ranging between the maximum and minimum criteria.
The CQI is described in detail elsewhere 218,325 and has been previously used in other
studies assessing nutritional adequacy, weight gain, incident obesity and cardiovascular
disease218,219,325–327. The GI for food and beverage items was calculated by using
average values from the 2002 International tables of GI and GL values and extended in
2008328 with glucose as the reference. Additionally, dietary GL was calculated
considering the quality and the amount of carbohydrate [GL= (GI x amount of available
carbohydrate)/100]329. Finally, both dietary GI and dietary GL were categorized into
tertiles.
1.7. Breast cancer ascertainment

The primary end point for the present analysis was the diagnosis of incident breast
cancer. We requested the medical record from participants or their families when a
diagnosis was self-reported in any of the biennial follow-up questionnaires. An expert
oncologist confirmed the cases, and those confirmed cases were considered for the main
analyses. When the medical record to certify the diagnosis of breast cancer was not yet
accessible possibly due to a delay in the delivery of the documentation, we treated them
as probable incident cases. As previously described in “1.2. Follow-up”, we consulted
the National Death Index to obtain the vital status and to identify the cause of death of
cohort members in case they were lost to follow-up. When the cause of the death was
breast cancer, we consider it to be a confirmed case.
1.8. Inclusion and exclusion criteria

For the analysis on SSBs and breast cancer risk, we used the December 2016
version of the database. Thus, by December 2016, in the SUN project, 13,843 women
had been recruited and had completed the baseline questionnaire. To warrant a
minimum follow-up time of 2 years, we included those women recruited before March
2014 (n=13,645). Additionally, women were excluded if they were lost to follow-up
72
(n=1,286), had a previous history of breast cancer at baseline (n=102), presented
implausible energy intakes using a predefined allowable range of 500–3500 kcal/day for
women324 (n=1,345), or reported menopause before 35 years of age (n=199). Therefore,
our final sample for the analysis of SSB consumption and incidence of breast cancer
included 10,713 women.
A new database became available in December 2018. Thus, we used data up to July
2018 for the rest of analyses. Accordingly, 13,999 women had been recruited and had
completed the baseline questionnaire. To warrant a minimum follow-up time of 2 years,
we included those women recruited before October 2015 (n=13,770). Women were also
excluded if they were lost to follow-up (n=1,295), had prevalent breast cancer (n=104),
showed implausible total energy intake according to the predefined allowable range
proposed by Willett330 (<500 or >3500 kcal/day) (n=1,353) or reported menopause before
the age of 35 years (n=206). Therefore, our final sample consisted of 10,812 women for
the analyses on the provegetarian food patterns, the CQI, the dietary fat intake, or the
phenolic acids intake evaluation.
1.9. Statistical analyses

In all the research papers using the SUN cohort database, we carried out stratified
analyses by menopausal status. Information on age at menopause was collected in the
baseline questionnaire and after 16 years of follow-up. If there was lacking information
regarding the age at menopause, we imputed postmenopausal status according to the
75th percentile of age at menopause in the study population (52 years of age) 331. When
assessing premenopausal breast cancer as the outcome, we excluded those women
who reported having had menopause before study inception and censored follow-up time
at the age of 52 years or at the age of menopause. For postmenopausal breast cancer,
we considered women at risk only those after having turned 52 years old or having had
their menopause. In the analysis with postmenopausal breast cancer as the outcome,
we additionally adjusted for time since recruitment until the beginning of the time at risk
and age at menopause (< 50 years, ≥ 50 years).
All analyses were performed using Stata, version 12.0. All P values were two-tailed
and a P value <0.05 was deemed as significant.
73
1.9.1. Sugar-sweetened beverages consumption and breast cancer risk
We described the baseline characteristics of the participants across the three
categories of SSB consumption. To outline baseline characteristics, we used
percentages for categorical variables and mean and standard deviations for quantitative
variables.
To assess the relationship between baseline SSB consumption (categorized as

never/seldom drinkers, ≤ 1 serving/week and > 1 servings/week) and the subsequent
risk of breast cancer, we used Cox regression models, and included confirmed incident
cases as the outcome. We calculated person-years of follow-up for each participant from
the date of fulfillment of the baseline questionnaire to the date of breast cancer diagnosis
for cases and to the last returned questionnaire or the date of death for non-cases.
Probable incident cases were considered as non-cases and censored at the date of self-
reported diagnosis. Hazard ratios (HR) and their 95% CI were estimated considering the
lowest category of SSB consumption as the reference category and with age as
underlying time-variable. Categories of age and period of recruitment were used as
stratification variables.
For all analyses, we fitted an age-adjusted model and two multivariable models
after adjusting for potential confounders. Model 1 was adjusted for height (continuous),
family history of breast cancer (yes/no), smoking status (never smoker, former smoker,
current smoker), physical activity (MET-h/week, continuous), alcohol intake (g/day,
continuous), BMI (tertiles), age of menarche (< 10 years, 10–11 years, 12–13 years, 14–
16 years, > 16 years), menopause (no menopause, < 50 years, ≥ 50 years), number of
pregnancies of more than 6 months (continuous), pregnancy before the age of 30 years
(yes/no), months of breastfeeding (continuous), use of hormone replacement therapy
(yes/no) and its duration (continuous) and years at university (continuous).
Model 2 included additional adjustment for diabetes (yes/no), GI (continuous,

excluding SSB for its calculation) total energy intake (kcal/day, continuous), ultra-
processed consumption (continuous), daily servings of coffee (continuous) and
decaffeinated coffee (continuous), and Mediterranean diet (continuous).
We rerun our analyses for confirmed breast cancer cases by excluding those
women with a probable breast cancer.
74
To reduce the number of covariates, we fitted multivariate model 3, adjusted for
tertiles of a propensity score estimating the probability of being exposed to SSB. The
probability of being exposed to SSB given covariates in model 2 was estimated using
binary or multinomial logistic regression.
Tests of linear trend were conducted assigning to each category of SSB

consumption its median and using the resulting variable as continuous in the above-
mentioned models.
As an additional analysis, we performed sensitivity analyses to explore the

robustness of our findings and to account for potential uncertainties in our assumptions
with respect to possible sources of bias or measurement errors. Therefore, we further
assessed the association between categories of baseline SSB consumption and the risk
of overall breast cancer and pre- and postmenopausal breast cancer including also
probable breast cancer cases and excluding those women with total energy intake
beneath the percentile 1 or above the percentile 99 (P1–P99). Then we conducted our
analyses over 11,787 subjects free of breast cancer at baseline within these latter energy
limits.
1.9.2. Phenolic acids intake and breast cancer risk

Baseline characteristics of the participants were described across tertiles of
phenolic acids intake. We used percentages for categorical variables and means and
standard deviations for quantitative variables.
Person-years of follow-up were calculated for each participant from the date of
completion of the baseline questionnaire to the date of breast cancer diagnosis, or the
date of death due to breast cancer cases, or the date of completion of the last follow-up
questionnaire or date of death for a cause other than breast cancer for non-cases.
To assess the relationship of baseline dietary intake of phenolic acid subclasses

or individual compounds of phenolic acids with the subsequent risk of breast cancer, Cox
regression models were used. We estimated hazards ratios (HRs) and 95% CIs of
incident breast cancer during the follow-up for the subclasses and individual compounds
of phenolic acids tertiles. The lowest tertile was used as the reference category.
Categories of age and period of recruitment were used as stratification variables.
75
To reduce the potential effect of a variation in diet during follow-up, Cox
proportional hazard models were fitted with repeated dietary measurements using the
updated data on food consumption after 10 years of follow-up.
Cox regression models were adjusted for several potential confounders defined
a priori. Potential confounders were identified based on available literature rather than
deferring to statistical criteria. Potential confounders included height (continuous), family
history of breast cancer (yes/no), smoking status (never smoker, former smoker, current
smoker), physical activity (metabolic equivalents h/week, continuous), alcohol intake
(g/day, continuous), BMI (tertiles), age at menarche (categories), menopause (no
menopause, <50 years, ≥50 years), number of pregnancies of more than 6 months
(continuous), pregnancy before the age of 30 years (yes/no), months of breastfeeding
(continuous), use of hormone replacement therapy (yes/no) and its duration
(continuous), and years at university (continuous). Additional adjustments for diabetes
(yes/no), total energy intake (kcal/day, continuous), and adherence to the Mediterranean
diet (continuous) were also included.
Tests of linear trend were conducted, assigning to each tertile its median and
using the resulting variable as continuous in the aforementioned models.
Proportional hazard assumption was checked with the Grambsche-Therneau test

of the scaled Schoenfeld residuals332 and introducing hydroxycinnamic and chlorogenic
acids intake as a time-varying covariate in the model.
Furthermore, the contribution of each food group to the between-person

variability in total phenolic acid intake was determined, as well as their main subclasses
—hydroxycinnamic and hydroxybenzoic acids—and individual compounds (chlorogenic
acid intakes). A series of nested linear regression models after stepwise-selection
algorithm was constructed. The additional contribution of a given food group (e.g., fruits,
nuts, vegetables, and other foods) was shown in the change of cumulative R 2.
Additional analyses after stratifying by smoking status and BMI were performed
in order to assess potential effect modification by these variables, and we formally tested
the interactions introducing product-terms in the multivariable Cox regression models.
76
We also performed sensitivity analyses to explore the robustness of our findings
by excluding breast cancer cases diagnosed in the first 2 years of follow-up or by adding
other potential confounders to the multivariable-adjusted model, such as fiber intake.
1.9.3. Total and subtypes of fat intake and breast cancer risk
Baseline characteristics are shown according to extreme tertiles of percentage of
total fat, MUFA, PUFA and SFA intake, as means (SD) for quantitative traits and n (%)
for categorical variables.
The risk of breast cancer was assessed by fitting Cox proportional hazards
regression models using age and time since recruitment as the primary time variable 333.
We calculated person-years of follow-up for each participant from the date of fulfilment
of the baseline questionnaire to the date of breast cancer diagnosis for cases and to the
last returned questionnaire for non-cases.
Our primary exposures of interest were intakes of total fat and three major types
of fat (MUFA, PUFA and SFA) in tertiles. We conducted all analyses using fat or subtypes
of fats as either continuous or categorical variables. Therefore, we used the multivariate
nutrient density model where the exposure is made up with the nutrient densities
(percentages of energy) from diverse fat subtypes and the covariate comprise the total
energy320. The HR for fat represents the association of breast cancer risk with the
substitution of fat for energy-yielding nutrient(s) excluded from the model. So, when non-
alcohol energy intake is held constant, an increment in fat intake is unavoidably followed
by a reduction in carbohydrate and protein intakes. Ultimately, the estimated HR in the
corresponding nutrient density model is useful in public health terms and is expressed in
the same units. We also perform the analyses stratified by menopausal status.
As long as alcohol consumption is an accepted risk factor for postmenopausal

breast cancer and is probably a cause of premenopausal breast cancer140, non-alcohol
sources of energy were employed as a continuous variable to adjust for energy intake in
all models and used alcohol intake as an independent confounder in the multivariable
models. For overall women, we fitted a model including potential confounders of the fat
intake and breast cancer association: height (continuous), family history of breast cancer
(yes/no), smoking status (never smoker, former smoker, current smoker), physical
activity (MET-h/week, continuous), alcohol intake (g/d, continuous), BMI (tertiles), age of
menarche (<10 years, 10-11 years, 12-13 years, 14-16 years, >16 years), menopause
(no menopause, <50 years, ≥50 years), number of pregnancies of more than 6 months
77
(continuous), years at university (continuous), use of hormone replacement therapy
(yes/no) and its duration (continuous), Mediterranean diet adherence (continuous) 161 and
non-alcohol energy intake (continuous). We did not adjust for use of hormone
replacement therapy (yes/no) and its duration (continuous) in premenopausal women.
Just for the analysis addressing postmenopausal breast cancer as the outcome, we
further adjusted for years since recruitment until the beginning of the time at risk
(continuous). In additional models, we further adjusted for protein intake to assess the
effect of substituting fat intake for carbohydrate intake and for carbohydrate intake to
assess the effect of substituting fat intake for protein intake.
Furthermore, we aimed to investigate the predicted effects of the isocaloric

replacement of 5% of energy from total fat or subtypes of fat with carbohydrate or protein
using multivariate nutrient density substitution models as described in detail
elsewhere321.
To test for linear trend across successive categories, the median value for each
tertile was used for each category of fat intake as a continuous variable.
1.9.4. Carbohydrate quality index and breast cancer risk

We compared the baseline characteristics of participants according to quartiles
of a CQI. We calculated means (SD) or percentages for each variable across quartiles
for descriptive purposes.
We used Cox regression models to assess the relationship between CQI and the
risk of breast cancer. Hazard ratios (HRs) with 95% confidence intervals (CIs) were
calculated for the three upper quartiles using the lowest quartile of CQI as the reference
category. We also used the CQI as a continuous variable and assessed changes in the
outcome associated with a two-point increase. Furthermore, we repeated Cox regression
models using individual components of CQI [dietary fiber intake, GI, ratio of whole-grain:
total-grain carbohydrates and ratio solid carbohydrates: (solid carbohydrates/liquid
carbohydrates)] and GL as the main exposure.
78
For all the exposures, we fitted two multivariable models after adjusting for the
main breast cancer risk factors and other potential confounders. Model 1 included height
(continuous), years at university (continuous), family history of breast cancer (yes/no),
smoking status (never smoker, former smoker, current smoker), physical activity (METs-
h/ week, tertiles), alcohol intake (continuous), BMI (continuous), age at menarche ( 11
years, 12-13 years, >14 years), number of pregnancies of more than 6 months
(continuous), use of hormonal therapy (yes/no) and its duration (continuous),
menopause (no menopause, < 50 years, ≥ 50 years). Model 2 was additionally adjusted
for energy intake (continuous), adherence to the Mediterranean diet (kcal/day,
continuous) and adherence to a special diet (yes/no).
Tests of linear trend across successive quartiles of CQI (or tertiles for individual
components of CQI and GL) were conducted assigning the median value to each
category and treating the resulting variables as continuous. We also presented the p-
value obtained for the independent variable introduced in the model as a continuous
variable.
For postmenopausal women, we additionally adjusted the multivariable models

for time since recruitment until the beginning of the time at risk and age at menopause
(<50 years, ≥50 years).
We also assessed the interaction between the amount of energy intake from
carbohydrates and CQI and assessed the relationship between the joint classification of
the tertiles of CQI and the percentage of energy intake from carbohydrates (≤50%, >50%
of total energy intake) and breast cancer risk.
1.9.5. Healthful and unhealthful provegetarian dietary patterns and breast

cancer risk
Baseline characteristics of the study participants according to tertiles of the
provegetarian food pattern, healthful and unhealthful provegetarian are presented as
means (SD) for quantitative traits and n (%) for categorical variables.
79
For the primary analysis, we fitted Cox proportional-hazards regression models
to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) to evaluate
separately the associations of tertiles of each index (PVG, uPVG, and hPVG) with breast
cancer incidence in the overall sample of women and by menopausal status. Follow-up
was calculated for each participant from the baseline questionnaire return date until
breast cancer diagnosis, death, or end of follow-up.
We used age as the time scale, with stratification by time period and age. For all
analyses, we adjusted for height (cm, continuous), family history of breast cancer
(yes/no), smoking status (never smoker, former smoker, current smoker), physical
activity (METs-h/ week, continuous), alcohol intake (g/day, continuous), BMI (tertiles),
age at menarche (11 years, 12-13 years, 13-14 years, >15 years), number of
pregnancies of >6 months (continuous), pregnancy before the age of 30 years (yes/no),
months of breastfeeding (continuous), years at university (continuous), total energy
intake (kcal/day, continuous), time since recruitment, menopause (no menopause, <50
years, ≥50 years), use of hormone replacement therapy (yes/no), and its duration
(continuous). The latter three covariates were not considered for the analysis among
premenopausal women. In the analysis with postmenopausal breast cancer as the
outcome we made an additional adjustment for time between recruitment until the
beginning of the time at risk of postmenopausal breast cancer and age at the time of
menopause (<50 years, ≥50 years).
We conducted tests for linear trends by assigning each tertile its median value
and treating the resulting variable as continuous. We fitted restricted cubic splines to the
fully adjusted model334. We used adjusted Kaplan-Meier curves to describe breast
cancer risk according to tertiles of adherence to a hPVG or an uPVG. We also compared
regression coefficients from the different models using either uPVG or hPVG as
exposure to determine whether the association for uPVG was significantly different from
the association for hPVG335,336.
Sensitivity analyses by menopausal status and different energy cutoff points were
performed to explore the robustness of our findings and account for potential
uncertainties in our assumptions with respect to possible sources of bias or
measurement errors.
80
2. Nurses’ Health Studies
2.1. Study population and recruitment
The Nurses’ Health Study I
In 1976, Speizer et al. mailed the initial questionnaire to suitable married female
registered nurses aged 30 to 55 years who lived in 1 of 11 US states with the biggest
number of registrants (New York, California, Pennsylvania, Ohio, Massachusetts, New
Jersey, Michigan, Texas, Florida, Connecticut, and Maryland). The 71.2% of those
invited nurses (i.e.,121,700 nurses) returned the questionnaire. The cohort was
composed of mainly white women (97%) which correctly describe the racial composition
of nurses at that time, and with a faintly higher socioeconomic status than the general
population. These characteristics, in turn, may have initially affected generalizability, but
this population selection enhances internal validity as a result of their health knowledge
and their commitment to research translates into high-quality and complete self-reported
health data and high follow-up rates.
The Nurses’ Health Study II
Registered nurses who were aged 30 to 55 years at the beginning of the original
NHS (1976) would have had few opportunities to have used oral contraceptives for long
periods before their first pregnancy. Therefore, in order to see the health effects of oral
contraceptive use and other risk factors during early reproductive life, Willett et al., in
1989, included 116,430 registered nurses aged 25 to 42 years. They resided in 1 of 14
US states (California, Connecticut, Indiana, Iowa, Kentucky, Massachusetts, Michigan,
Missouri, New York, North Carolina, Ohio, Pennsylvania, South Carolina, and Texas),
and started the Nurses’ Health Study II (NHSII). Given the development of computers
and software abilities, and to lessen costs, participants were given the option to answer
to Web-based questionnaires starting in 2001, and by 2011, around 70% of all NHSII
participants had switched to online questionnaires. Furthermore, the NHSII also
introduced a large biorepository starting with blood samples collected in 1996.
Questionnaires
Development
The biennial questionnaires have established a distinctive resource of lifestyle-
and health-related data collected continuously over the past 40 years. Beyond the
questions regarding diet and lifestyle behaviors, researchers have regularly assessed
anxiety, depression, optimism, and social networks using validated scales337.
Researchers from the NHS and NHSII have also developed measures of environmental
exposures (e.g., air pollution and ultraviolet radiation) and neighborhood aspects (e.g.,
81
socioeconomic status, walkability…) based on residential histories. Just in the NHSII,
participants reported also body size at aged 5, 10, and 20 years in 1989 and physical
activity at aged 12 to 17 years in 1997; a subset of 41% of the women completed a
supplemental FFQ asking about adolescent diet in 1998. In 2001/2002, 39,904 mothers
of NHS and NHSII participants provided detailed information about their pre-pregnancy,
pregnancy, and early life experiences338. The Nurses’ Health Studies provide unique
data to study associations of disease risk to a broad set of social, lifestyle, and
environmental exposures at various crucial time points across the life course.
Diet questionnaires
Starting in 1980, the NHS questionnaires included a validated food frequency

questionnaire every 2 to 4 years339. After a series of pilot studies, a 61-item
semiquantitative FFQ was mailed in 1980 and was returned by approximately 95,000
women in the NHS. An extended questionnaire consisting of 130 foods was developed
in 1984. With little modifications to justify changes in the food supply and new
hypotheses, this questionnaire has been delivered every 4 years in both the NHS since
1986 and the NHSII since 1991 (Figure 12 and Appendices).
FFQs evaluate usual frequencies of intake using a structured list of foods over a
specified period of time in the past; in the Nurses’ Health Studies, researchers have
focused on food intake over the previous year 324. For the NHS, they developed a
semiquantitative questionnaire that incorporates information about usual portion sizes 340
and was to some extent inexpensive, simple to administer by mail, and had low
participant burden.
Repeated diet assessments using FFQs also allowed researchers to understand

how associations observed with chronic diseases are influenced by different ways of
modeling dietary exposures339. As all measurements, FFQs are prone to random and
systematic error. The assessment of the reliability and validity of diet assessment
methods, particularly FFQs, have been notable NHS contributions340,341, in fact, the
original NHS has evolved into one of the standards for dietary validation and calibration
study design. The low cost, ease of administration, and demonstrated reproducibility and
validity of the FFQ have made it one of the most widely used diet assessment methods
in large-scale, long-term observational studies.
The associations researchers have reported between diet and various diseases
in prospective analyses provide further evidence that these FFQs are able to detect
82
important biological relationships339. Additional behavioral and lifestyle risk factors have
been repeatedly assessed over the years, an extensive biorepository has been
established (Figure 12) and the array of chronic diseases and other outcomes have
expanded steadily over time.
Figure 12. Timeline of Data Collection in the Nurses’ Health Studies. FFQ = Food
Frequency Questionnaire; Qx = questionnaire; SF-36 = Short Form-36. (Adapted from
Bao Y et al.,342).
The scope of lifestyle and health outcome data collected has notably expanded
over time, beginning with the inclusion in 1980 of questions about physical activity and a
semiquantitative FFQ, which was completed by the cohort under the direction of Walter
Willett. The biennial study questionnaires include information on:
• Employment status, shift work • Leisure time physical activity,

• Smoke exposure sedentary time
• Family history of cancer or heart or • Social networks
other disease • Sleep patterns
• Living arrangement • Optimism scale
• Reproductive history and • Alcohol use
menopause • Caregiving and caregiving stress
• Neighborhood characteristics • Weight, height, waist, and hip
• Prescription and over-the-counter measurements
medications • Quality of life
• Environmental exposures • Diet (including during adolescence)
• Cancer and other screening tests
• Mental health
83
2.2. Follow-up
2.2.1. Response proportion
Women continue to be followed via biennial questionnaires, making the NHS the
first cohort of its magnitude with repeated data collection. Certified mail has been
described as the most effective approach for contacting study participants who did not
respond to the first mail. For those participants who did not answer to three regular mails,
researchers resent the questionnaire via certified mail. Nonetheless, this method is
expensive, so over the past years, this approach was replaced with hand-addressed
mailing for the greater number of participants. Response proportion was 30% for the first
hand-addressed mails and around 15% for the second and final mails after 6 to 7 regular
mails without response.
As of 2012, with all the mailing, the overall response proportion for participants is
86.2%. This high retention proportion emerges from many factors, containing but not
limited to the selection of motivated health professionals, periodic contact with
participants, repeated mailings followed by hand-addressed letters to non-responders,
the use of a shorter questionnaire in the final mailings to non-responders, and an annual
participant newsletter.
2.2.2. Documentation of diseases and deaths

Participants report any newly diagnosed diseases on their biennially follow-up
questionnaires. For new disease reports, researchers ask permission to check medical
records and gather pathology specimens. Once permission from the participant is
obtained, the software produces repeated mailings to hospitals to obtain medical
records.
2.2.3. Biorepository
NHS and NHSII have a rich resource of biospecimens, including blood, urine,
toenails, buccal cells, stool, and saliva (Figure 12), and the collection of repeated
samples substantially increases statistical power for analyses of age and latency effects.
The Brigham and Women’s Hospital/Harvard Cohorts Biorepository manages a large
number of biological specimens, including plasma, white blood cells, red blood cells,
whole blood, urine, DNA, toenails, saliva, stool, hair, mammograms, and paraffin
embedded human tissue. These samples are particularly valuable because participants
already have been followed as much as 30 years since collection and they are followed
by an abundance of data on medical histories and health-related behaviors. Specifically,
in NHS, in 1989-90, blood samples were collected from 32,826 women, ages 43-69. In
2000-2002, researchers collected a second blood sample from 18,743 of these women.
In NHSII, in 1996-99, blood samples were collected from 29,611 women, ages 32-55
years, 18,521 of whom donated samples timed within the menstrual cycle. Second blood
and urine samples were collected using a similar protocol in 2008-11 from 16,512 of
these women. Blood was separated into plasma, buffy, and red blood cells and aliquoted
into labeled cryotubes. NHS women with two samples are similar to those with one and
both groups are similar to the overall cohort.
The Nurses’ Health Study has also one of the biggest tumor tissue repositories
nested within a cohort study. Jointly with the Dana-Farber and Harvard Cancer Center
Core, researchers and clinicians have constructed hundreds of tissue microarrays
(TMAs), a powerful approach to measuring biomarkers utilizing many samples on the
same block.
2.3. Information processing

The Statistics Group and the Channing Division of Network Medicine computer
facility are key infrastructure resources for the management of the database as it grows
with the continuous receipt of biennial questionnaires, new -omics data and new
additions to the nutrient composition database. The Statistics Group collaborates on
complex statistical questions, develops statistical software, maintains and updates the
growing databases and leads regular seminars and training sessions, including teaching
newer methods for integrative analyses of -omics data sets.
As analysis of cohort data can be complex, custom utilities and analytic macros have
been developed to facilitate and standardize analyses (the vast majority are SAS
macros). These include basic macros (e.g., reading in the data from multiple
questionnaires) as well as more sophisticated tools such as nonlinear, advanced
statistical analysis, or direct statistical testing of heterogeneity of risk factor associations
across disease subtypes using competing risks analysis (prospective studies) or
polytomous logistic regression (nested case–control studies).
2.4. Ethical standards

The study protocol was approved by the institutional review boards of the Brigham
and Women’s Hospital and Harvard T.H. Chan School of Public Health, and those of
participating state cancer registries as appropriate. The return of the completed self-
administered questionnaire was considered to imply informed consent.
85
2.5. Funding sources
These studies were supported by grants UM1 CA186107, U01 CA176726, P01
CA87969, and R01 CA50385 from the National Institutes of Health, and the Breast
Cancer Research Foundation. The funding sources did not participate in the design or
conduct of the study; collection, management, analysis or interpretation of the data; or
preparation, review, or approval of the manuscript.
2.6. Exposure assessment

2.6.1. Sugar-sweetened and artificially sweetened beverage consumption
Diet was assessed with a validated FFQ administered in the NHS in 1980, 1984,
1986, and every 4 years thereafter, and in the NHSII in 1991 and every 4 years
thereafter. The numbers of FFQ food items have evolved: in the NHS, there were 61
items in 1980, 116 items in 1984, and 1986, and ≥130 items afterwards; in the NHSII,
the FFQ from 1991 had ≥130 items. The FFQs included foods with a portion size, and
participants were asked how often on average during the previous year they had
consumed the foods specified. A standard portion size and nine possible responses for
the frequency of consumption, ranging from “never/almost never” to “6 or more times per
day”, were given for each food item (Available at:
https://www.nurseshealthstudy.org/participants/questionnaires).
Nutrient and energy intakes were calculated by multiplying the frequency of

consumption of each unit of food and beverage recorded with the FFQs by its nutrient
and energy contents and summing across all items, using the US Department of
Agriculture database and complemented with information from the manufacturers 343,344.
Total SSBs were defined as caffeinated and non-caffeinated colas with sugar, other
(non-cola) carbonated beverages with sugar, and noncarbonated sweetened beverages
(e.g., punch, lemonade, fruit drink, or sugared iced tea). We summed the consumption
of these beverages as total SSBs. In addition, ASBs were defined as caffeinated, non-
caffeinated, and noncarbonated low-calorie or diet beverages. Questions included the
frequency of consumption over the past year for a standard 355 mL (12 oz.) serving (1
glass/can/bottle) of each SSB or ASB.
The FFQ has been extensively validated in the Nurses’ cohorts by comparison
with more detailed methods341,344,345 and biomarkers of intake341. For example, in a
comparison of the 1986 FFQ with multiple dietary records obtained in 1986, the mean
correlation coefficients were 0.84 for SSBs and 0.36 for ASBs 345. Moreover, high and
86
low-energy beverages had low to moderate correlations with biomarkers (IL-6, CRP,
TNFaR2, adiponectin) in a recent study346.
2.6.2. Healthful and unhealthful plant-based dietary patterns

Diet was assessed with FFQs administered in the NHS in 1980, 1984, 1986 and
every 4 years thereafter, and in the NHSII in 1991 and every 4 years thereafter. FFQs
include foods with a portion size, and participants were asked to specify the food-specific
average consumption during the previous year (9 choices from “almost never” to “6 or
more times per day”). Participants’ nutrient intakes were calculated by multiplying the
nutrient content of a food serving (based on updated USDA databases and other
sources) and consumption frequency.
As described in a prior publication 167, we derived three versions of a plant-based

diet using the FFQ data, an overall plant-based diet index (PDI), a healthful plant-based
diet index (hPDI), and an unhealthful plant-based diet index (uPDI). Briefly, we included
18 food groups (Table 4) based on nutrient and culinary similarities within the larger
categories of healthy plant foods, less healthy plant foods, and animal foods. Given that
alcoholic beverages have different directions of association for various health outcomes,
and margarine’s fatty acid composition has changed over time from high trans to high
unsaturated fats, we did not include these foods in the indices; we adjusted for alcohol
in the main analysis. Healthy plant food groups included whole grains, fruits, vegetables,
nuts, legumes, vegetable oils, and tea/coffee; less healthy plant food groups included
fruit juices, SSBs, refined grains, potatoes, and sweets; and animal food groups included
animal fats, dairy, eggs, fish/seafood, meat, and miscellaneous animal-based foods.
Food groups were ranked into quintiles and given positive or reverse scores. With
positive scores, participants in the highest quintile of a food group received a score of 5,
following on through to participants in the lowest quintile who received a score of 1. With
reverse scores, this pattern of scoring was opposite, with a score of 5 for the low quintile.
For the overall PDI, all plant food groups were given positive scores, while all animal
food groups were given reverse scores. For the hPDI, positive scores were given to
healthy plant food groups, and reverse scores to less healthy plant food groups and
animal food groups. Finally, for uPDI the opposite pattern of scoring was applied. The
18 food group scores were summed to obtain the indices.
87
Table 4. Examples of food items constituting the 18 food groups (from the 1984 Nurses’ Health
Study food frequency questionnaire).
Plant Food Groups
Healthy Plant Foods
1. Whole grains Whole grain breakfast cereal, other cooked breakfast cereal, cooked oatmeal,
dark bread, brown rice, other grains, bran, wheat germ, popcorn
2. Fruits Raisins or grapes, prunes, bananas, cantaloupe, watermelon, fresh apples or
pears, oranges, grapefruit, strawberries, blueberries, peaches or apricots or
plums
3. Vegetables Tomatoes, tomato juice, tomato sauce, broccoli, cabbage, cauliflower, Brussels
sprouts, carrots, mixed vegetables, yellow or winter squash, eggplant or
zucchini, yams or sweet potatoes, spinach cooked, spinach raw, kale or
mustard or chard greens, iceberg or head lettuce, romaine or leaf lettuce,
celery, mushrooms, beets, alfalfa sprouts, garlic, corn
4. Nuts Nuts, peanut butter
5. Legumes String beans, tofu or soybeans, beans or lentils, peas or lima beans
6. Vegetable oils Oil-based salad dressing, vegetable oil used for cooking
7. Tea and coffee Tea, coffee, decaffeinated coffee
Unhealthy Plant Foods
8. Fruit juices Apple juice or cider, orange juice, grapefruit juice, other fruit juice
9. Refined grains Refined grain breakfast cereal, white bread, English muffins or bagels or rolls,
muffins or biscuits, white rice, pancakes or waffles, crackers, pasta
10. Potatoes French fries, baked or mashed potatoes, potato or corn chips
11. Sugar-sweetened Colas with caffeine & sugar, colas without caffeine but with sugar, other
beverages carbonated beverages with sugar, non-carbonated fruit drinks with sugar
12. Sweets and Chocolates, candy bars, candy without chocolate, cookies (home-baked &
Desserts ready-made), brownies, doughnuts, cake (home-baked & ready-made), sweet
roll (home-baked & ready-made), pie (home-baked & ready-made), jams or
jellies or preserves or syrup or honey
Animal Food Groups
13. Animal fat Butter added to food, butter or lard used for cooking
14. Dairy Skim low fat milk, whole milk, cream, sour cream, sherbet, ice cream, yogurt,
cottage or ricotta cheese, cream cheese, other cheese
15. Egg Eggs
16. Fish or Seafood Canned tuna, dark meat fish, other fish, shrimp or lobster or scallops
17. Meat Chicken or turkey with skin, chicken or turkey without skin, bacon, hot dogs,
processed meats, liver, hamburger, beef or pork or lamb mixed dish, beef or
pork or lamb main dish
18. Miscellaneous Pizza, chowder or cream soup, mayonnaise or other creamy salad dressing
animal-based foods
2.7. Breast cancer ascertainment

Participants report newly diagnosed breast cancer biennially on their follow-up
questionnaires. Then, researchers requested permission from women reporting breast
cancer to review hospital records and pathology reports for diagnosis confirmation and
ascertainment of invasive vs. in situ tumors, and ER, PR and HER2 status. Once
permission is secured, their software produces repeated mailings to hospitals to get
medical records. Medical practitioners, blinded to questionnaire exposure data, revise all
medical records to confirm self-reported breast cancer. Given the high confirmation
proportion of reported breast cancer cases in the NHS and NHSII (>99%), we included
both breast cancer cases confirmed via medical record review and self-reported cases
confirmed by the nurse but lacking medical record. For deceased cases, the next of kin
was contacted for this permission; deaths were reported by family members or by the
postal service in response to follow-up questionnaires, or they were identified through
88
the National Death Index, which is a highly effective method for monitoring deaths,
reducing the need to obtain death certificates.
2.7.1. Tissue Microarrays, Immunohistochemical Analysis and Subtype

Classification
We have already reported details of breast cancer tissue block collection and
TMA construction347. Briefly, we collected archived formalin-fixed paraffin-embedded
breast cancer blocks from participants with incident breast cancer diagnosed up to 2006.
For molecular subtype classification, immunohistochemical staining information was
available for the markers of ER, PR, HER2, cytokeratins 5/6 and epidermal growth factor
receptor. Further staining for the proliferative marker Ki-67 was achieved in NHS cases;
Ki-67 data were not available for NHSII cases. Cases with TMAs were very similar to all
suitable invasive cases in terms of demographics, breast cancer risk factors and tumor
characteristics.
Definitions that correlated with gene expression profile classifications were used
for tumor molecular subtyping for a subgroup of cases 50–52,348–350. If Ki-67 expression data
was missing (NHSII tumors), histological grade was used instead. Hence, luminal A
tumors were ER-positive and/or PR-positive, HER2-negative, and Ki-67-negative (or
histologic grade 1 or 2). Luminal B tumors were either (a) ER-positive and/or PR-positive
and HER2-positive or (b) ER-positive and/or PR-positive, HER2-negative, and Ki-67-
positive (or histologic grade 3). HER2-enriched tumors were ER-negative, PR-negative
and HER2-positive. Basal-like tumors were ER-negative, PR-negative, HER2-negative
and CK 5/6-positive and/or EGFR-positive. For evaluating ER-positive vs. ER-negative
tumors, ER status was determined primarily from TMA slides, and if not available,
secondarily from pathology reports.
2.8. Inclusion and exclusion criteria
2.8.1. Sugar-sweetened and artificially sweetened beverage consumption

Women were followed from baseline (NHS, 1980; NHSII, 1991), when dietary
information was first available, to 2016 in NHS and to 2017 in NHSII. We excluded
women who died prior to the baseline dietary assessment, who had prevalent cancer,
had missing data on SSB or ASB intake, reported implausible total energy intake (<600
or >3500 kcal/day); leaving 82,713 women from NHS, and 93,085 from NHSII.
89
2.8.2. Plant-based diets
For this analysis, women were followed from baseline (NHS, 1984; NHSII, 1991),
when dietary information was first available, to 2016 in NHS and to 2017 in NHSII. This
FFQ was expanded from 61 to 116 items in 1984; therefore, given the complexity and
broad array of foods included in our plant-based dietary patterns, we used the 1984 FFQ
as baseline questionnaire.
We excluded women who died prior to baseline dietary assessment, had prevalent
cancer, had missing dietary information, or reported implausible total energy intake
(<600 or >3500 kcal/day); leaving 76,690 women from NHS, and 93,295 from NHSII.
2.9. Statistical analyses
2.9.1. Sugar-sweetened beverages and breast cancer risk
Data from the NHS and NHSII were pooled to increase statistical power.
Participants contributed person-years from the date of return of the baseline FFQ (NHS,
1980; NHSII 1991) to the date of any cancer diagnosis except non-melanoma skin
cancer, death, or end of follow-up [2016 (NHS) or 2017 (NHSII) for the main analysis
and 2006 for the molecular subtype analysis], whichever happened first. The main
outcome of the analysis was incident breast cancer (occurring in 1980-2016 in the NHS
and in 1991-2017 in the NHSII) and secondary outcomes were the different breast
cancer tumor subtypes.
We assessed the associations between categories of consumption of SSBs and

ASBs and total and subtype-specific breast cancer using multivariable-adjusted time-
varying Cox proportional hazards regression models, stratified by age, 2-year time-
period at risk, and cohort (NHS or NHSII), to estimate HRs and 95% CIs. In order to
reduce measurement error and within-person variation and to better represent long-term
diet during follow-up, we performed the analysis using cumulative averages of dietary
data that were created using repeated measures from the FFQs (calculated by taking
the mean intake from all FFQs up to the beginning of a 2-year follow-up interval). We
categorized SSB and ASB in groups based on the frequency of intake, 1/month
(reference), 1 to 4/month, 2 to 6/week and ≥1/day, and linear trends were evaluated
using the Wald test on a continuous variable representing median intakes of each
category.
90
To evaluate the latency between SSBs or ASBs intake and breast cancer, we
performed latency analyses based on dietary data collected at different time points 324,351.
Briefly, in the cumulative average model, mean SSB or ASB consumption from all FFQs
up to the beginning of a follow-up interval was calculated; in the simple update model,
SSB or ASB consumption reported on the most recent FFQ before each follow-up
interval was used; in the latency models, we used SSB or ASB consumption reported at
different latencies (i.e., 4-8, 8-12, 12-16 and 16-20 years) before breast cancer
diagnosis. For example, in the 8-12 -year lag analysis, we examined the association of
SSB consumption in 1980 with the risk of breast cancer between 1988 and 1992. Since
the FFQs were collected every four years, the simple update model could be considered
as a latency analysis of 0-4 years.
To test whether consumption of SSBs or ASBs and breast cancer risk differed by
BMI (<25 vs. ≥25 kg/m2), physical activity (<21 vs. ≥21 METs-h/week352) or diet quality
(below vs. above AHEI median), we added interaction terms and used the Wald test for
cross-product terms based on SSBs or ASBs intake (continuous variable) and the
stratification variables.
To examine differential associations of SSBs and ASBs with breast cancer risk
by hormone receptor and molecular subtypes, we used the Lunn-McNeil approach to
derive the P for heterogeneity353. All statistical tests were 2-sided with a P value of <0.05
and performed using SAS version 9.4 (SAS Institute Inc).
2.9.2. Healthful and unhealthful plant-based dietary patterns
Data from the NHS and NHSII were pooled. Person-time for each participant was
calculated from the date of return of the baseline questionnaire until the date of breast
cancer diagnosis, other cancers (excluding non-melanoma skin cancers), death, or the
end of follow-up (2016 (NHS) or 2017 (NHSII) for the main analysis and 2006 for
molecular subtype analysis), whichever occurred first.
Indices were cumulatively averaged over follow-up to better capture long-term

diet. We evaluated the association between PDI, uPDI and hPDI quintiles and incident
breast cancer using multivariable-adjusted time-varying Cox proportional hazards
regression models, stratified by age, 2-year time-period at risk and cohort (NHS or
NHSII), to estimate hazard ratios and 95% CIs. We tested for linear trends by evaluating
the quintile median values as a continuous variable. We also performed an analysis
using cumulative average indices in relation to fatal breast cancer, defined by diagnosis
91
of invasive breast cancer during follow-up that led to death due to breast cancer before
the end of follow-up.
To take advantage of repeated diet assessments in these cohorts and evaluate

the latency between these indices and breast cancer incidence, we conducted latency
analyses, whereby we created different regression models based on dietary data
collected at distinct time points. In the simple update model, index scores reported on
the most recent FFQ before each follow-up interval was used; in the latency models, we
used index scores reported at different latencies (i.e., 4–8, 8–12, 12–16, 16-20 y) before
breast cancer diagnosis 351. Furthermore, we added interaction terms and used the Wald
test to test whether the PDI, uPDI and hPDI and breast cancer association differed by
menopausal status, current BMI and physical activity. Mediation analyses 354 were
performed in order to assess the extent to which associations may be potentially
mediated by weight gain from age 18 years and estimated the mediation proportion 355,356
(the proportion of the observed association attributable to a mediator) 356. To evaluate
whether associations differed by molecular subtype or ER status, we used the Lunn-
McNeil approach to derive the p for heterogeneity353. Statistical tests were two-sided with
P-values<0.05 indicating statistical significance. All analyses were performed using SAS
version 9.4 (SAS Institute Inc).
Covariates
In the Nurses’ Health Studies (NHS/NHSII), information on lifestyle factors and
medical history was obtained from biennial questionnaires. For the analyses on SSBs
and ASBs, we included race (self-reported), height, BMI at age 18, age at menarche,
parity/age at first birth, breastfeeding, oral contraceptive use, family history of breast
cancer in a first-degree relative, history of benign breast disease, age at menopause,
postmenopausal hormone use, census-tract socioeconomic status, physical activity
(METs-h/week), cumulatively updated alcohol intake and cumulatively updated total
caloric intake. We additionally adjusted for a modified Alternate Healthy Eating Index
(AHEI) score357, with SSBs and alcohol removed. This score was calculated based on 9
foods and nutrients that are predictive of chronic disease risk, including fruit, vegetables,
nuts and legumes, red and processed meat, whole grains, sodium, trans fat, long chain
ω-3, and other polyunsaturated fats. A higher score in the AHEI denoted better diet
quality. All models were mutually adjusted for SSB and ASB in categories. Because
change in weight since age 18 may be intermediate between SSBs or ASBs and breast
cancer, we subsequently added this variable in a separate model.
92
For the plant-based diets analyses, covariates included race, socioeconomic
status, age at menarche, age at menopause, postmenopausal hormone use, oral
contraceptive use, parity, age at first birth, breastfeeding history, height, alcohol intake,
total caloric intake, physical activity, and BMI at age 18 years. As change in weight from
age 18, total carotenoids intake and dietary fiber may be intermediates between diet and
breast cancer, we additionally adjusted for them in separate models.
93
RESULTS
94
For the analyses on the SUN Project, we included information from 10,812 women.
The mean age was 34.7 years (SD: 10.7 years). Of the total number of women studied,
9,593 were premenopausal at the time of entry into the study (88.7% of the cohort), of
which 1,870 (19.5%) developed menopause during follow-up (Table 5). At the beginning
of the study, the remaining 1,219 women were already postmenopausal (11.3%).
With a median follow-up of 11.8 years and 115,368 women-years at risk, 190
probable cases of breast cancer were identified. When only confirmed cases were
included in the analysis, 101 cases were included among 115,802 person-years of
follow-up. The crude incidence rate was 16.4 per 10,000 woman-years for probable
cases and 8.7 per 10,000 woman-years for confirmed cases.
Table 5. Descriptive characteristics of women with a confirmed diagnosis of breast cancer and the
total sample of women from the “Seguimiento Universidad de Navarra” (SUN) Project.
Confirmed Breast Cancer Total cohort of women
Menopausal status at Menopausal status at
baseline baseline
All cases All women
Premenopa Postmenop Premenopa Postmenopau
usal ausal usal sal
N 101 81 20 10,812 9,593 1,219
Age at baseline 41.7 (8.5) 32.3 (8.3) 53.8 (6.4) 34.3 32.3 (8.3) 53.8 (6.4)
Age at diagnosis 45.9 (11.4) 43.2 (7.9) 52.6 (4.5) - - -
BMI (kg/m2) at baseline 22.6 (2.9) 22.0 (2.9) 23.9 (3.4) 22.2 (3.1) 22.0 (2.9) 23.9 (3.4)
Postmenopausal BC 34 17 20 - - -
Premenopausal BC 57 57 - - - -
Family history of BC, n 18 16 2 1,143 1,004 139
Mean time since FFQ 10.7 (4.3) 10.7 (4.3) 10.5 (4.1) - - -
and DX
SSB intake *, mean 1.3 (1.5) 1.4 (1.6) 0.6 (0.8) 1.4 (1.4) 1.4 (1.4) 0.7 (1.2)
(SD)
Hydroxycinnamic 414.7 (84.4) 410.8 (82.7) 450.5 (104.2) 434.6 (128.2) 431.0 (126.2) 458.8 (139.0)
acids, mean (SD)
Chlorogenic acids, 272.7 (62.9) 272.5 (64.4) 274.5 (57.3) 292.5 (98.3) 290.2 (96.4) 306.6 (108.7)
mean (SD)
CQI, median 11 11 12 11 11 12
PVG, median 45 44 47 45 45 46
BC: Breast Cancer; BMI: Body Mass Index; CQI: Carbohydrate Quality Index; PVG: Provegetarian food pattern; SD:
Standard Deviation; SSB: Sugar-sweetened beverages
*Previous SUN database used for this exposure (N=10,713 women)
95
project
The present paper assesses the association between SSB and risk of breast cancer
among women from the SUN project. As previously indicated, 10,713 women with a
median age of 33 years were included in this study. Out of the total of women, 9,507
(88.7%) were premenopausal at baseline, out of which 1,761 (18.5% of 9,507)
developed menopause during the follow-up time. The remaining 1,206 women were
postmenopausal at baseline (11.3%).
Baseline characteristics of 10,713 women are presented in Table 6. Participants in

the highest category of total SSB consumption were younger, were less physically active,
had a higher BMI, higher total energy intake and ultra-processed food consumption,
higher intake of alcohol, meat, dairy products and fat intake, lower consumption of fruit,
vegetables, fish, and olive oil, and lower adherence to the Mediterranean diet compared
to women who never or seldom consumed SSBs.
Table 6. Baseline characteristics of 10,713 women in the SUN cohort according to the
categories of sugar-sweetened beverage (SSB) consumption.
Sugar-sweetened beverage category
Variable Never/seldom 1/month-≤1/week >1/week
N 4063 4537 2113
Median of SSB consumption (servings/week) 0 1 3
Age (years) 38.8 (11.3) 32.9 (9.6) 31.0 (8.5)
Body mass index (kg/m2) 22.5 (3.1) 22.1 (3.0) 22.0 (3.2)
Physical activity (METs-h/week) 25.9 (21.8) 23.7 (19.4) 22.4 (19.0)
Total energy intake (kcal/d) 2193 (585) 2322 (556) 2440 (550)
Alcohol intake (g/d) 4.1 (6.7) 3.5 (4.9) 5.0 (6.0)
Years at university 4.9 (1.4) 4.8 (1.3) 4.8 (1.3)
Height (cm) 163 (6) 164 (6) 164 (6)
Number of pregnancies of more than 6 0.9 (1.3) 0.6 (1.1) 0.4 (1.0)
months
Breast-feeding (months) 3.2 (5.8) 2.1 (4.6) 1.4 (3.9)
Time of hormone replacement therapy 1.3 (2.4) 1.3 (2.4) 1.2 (2.4)
(years)1
Diabetes (%) 1.87 0.75 0.43
Pregnancy before 30 years (%) 25.8 16.8 12.1
Smoking (%)
Former smoker 22.9 25.3 35.5
Current smoker 27.1 18.5 15.4
Family history of breast cancer (%) 2 11.2 10.3 10.2
Age of menarche (%)
<10 years 1.3 1.2 0.9
10-11 years 20.2 18.2 18.7
12-13 years 53.8 56.0 54.0
14-16 years 22.3 22.2 23.7
>16 years 2.5 2.5 27
Menopausal status (%)
Premenopausal 81.1 92.5 95.4
Postmenopausal 18.9 7.7 4.6
Age at menopause1
Postmenopausal <50 years (%) 47.3 55.5 51.2
Postmenopausal ≥50 years (%) 52.7 44.3 49.4
Macronutrient distribution
Carbohydrate intake (% energy) 43.5 (8.1) 43 (6.8) 43.3 (7)
Protein intake (% energy) 18.9 (3.8) 18.3 (3) 17.6 (3)
96
Table 6 (continuation). Baseline characteristics of 10,713 women in the SUN cohort
according to the categories of sugar-sweetened beverage (SSB) consumption.
Variable Never/seldom 1/month-≤1/week >1/week
Fat intake (% energy) 36.2 (7.2) 37.5 (6.2) 37.7 (6.4)
Saturated fatty acids (% energy) 11.8 (3.3) 12.8 (3) 13 (3)
Monounsaturated fatty acids (% 15.9 (4.2) 16.2 (3.7) 16.1 (3.7)
energy)
Polyunsaturated fatty acids (% 5.0 (1.6) 5.2 (1.6) 5.3 (1.6)
energy)
Adherence to the Mediterranean diet 3 4.5 (1.6) 3.9 (1.7) 3.7 (1.7)
Ultra-processed consumption (g/d) 2.7 (1.6) 3.2 (1.4) 4.4 (1.7)
Coffee intake (servings/d) 1.2 (1.3) 1.2 (1.2) 1.2 (1.3)
Decaffeinated coffee (servings/d) 1.3 (2.0) 1.3 (1.9) 1.2 (1.8)
Fruit consumption (g/d) 403 (312) 358 (283) 315 (277)
Vegetable consumption (g/d) 614 (384) 546 (318) 515 (323)
Olive oil consumption (g/d) 18 (15) 17 (14) 16 (14)
Cereal consumption (g/d) 97 (69) 97 (64) 94 (63)
Legume consumption (g/d) 22 (19) 22 (15) 21 (16)
Meat/meat product consumption (g/d) 155 (78) 177 (74) 186 (76)
Fish and seafood consumption (g/d) 101 (65) 96 (56) 92 (59)
Dairy product consumption (g/d) 153 (178) 191 (186) 206 (196)
Values are expressed as mean (SD), unless otherwise stated. SFA, saturated fatty acid; MUFA, monounsaturated
fatty acid; PUFA, polyunsaturated fatty acid; MET, metabolic equivalents.1Only for postmenopausal women;
2
Information from mother, sisters, and both grandmothers was collected; 3Score proposed by Trichopoulou et al.,161.
Out of 168 probable incident cases of breast cancer, we confirmed 100 new-
onset cases of breast cancer using medical records. The incidence rate was 15.9 per
10,000 woman-years for probable cases and 9.4 per 10,000 woman-years for confirmed
cases. A frequency of SSB consumption of >1 serving per week was associated,
although not statistically significant, with a higher incidence of breast cancer (Table 7)
(HR 1.36; 95% CI 0.74, 2.50; P-trend=0.312) compared to never or seldom consumers.
Results remained practically unchanged after merging the two highest category
into one, after applying other exclusion energy criteria or considering probable cases as
the main outcome.
Table 7. Hazard Ratio (HR) and 95% confidence intervals (CI) of confirmed breast cancer
cases according to the categories of baseline SSB consumption in 10,713 women of the SUN
Project.
Never/seldom 1/month- 1/week >1 week P-trend
N 4063 4537 SSB 2113
Cases/person-years 40/38,699 43/45,921 17/21,568
Incidence rate /10.000 person- 10.34 9.36 SSB 7.88
years
Age-adjusted model 1.00 1.32 (0.85, 2.05) 1.33 (0.74, 2.40) 0.312
Multivariate model 1 1.00 1.30 (0.83, 2.04) 1.36 (0.74, 2.50) 0.312
Never/seldom: no consumption, barely intake throughout the year; Regular consumption: ≥1 serving/month;
Multivariable model 1 adjusted for height, family history of breast cancer (yes/no), smoking status (never smoker,
former smoker, current smoker), physical activity (METs-h/week, continuous), alcohol intake (g/d, continuous), BMI
(tertiles), age of menarche (<10 years, 10-11 years, 12-13 years, 14-16 years, >16 years), menopause (no menopause,
<50 years, ≥50 years), number of pregnancies of more than 6 months (continuous), pregnancy before the age of 30
years (yes/no), months of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its duration
(continuous), years at university (continuous), diabetes (yes/no), glycemic index (continuous), total energy intake
(kcal/d, continuous), decaffeinated coffee intake (continuous), coffee intake (continuous), ultra-processed food
consumption (continuous) and Mediterranean diet adherence (continuous).
97
When we divided women according to their menopausal status, among
premenopausal women, regular consumption of SSBs was not associated with breast
cancer (Table 8), (HR 1.13; 95% CI 0.64, 2.01; P-trend=0.669) when compared to
women who never or seldom consumed SSBs. Moreover, postmenopausal women with
regular SSB consumption (≥1 serving/month) showed a higher incidence of breast
cancer (HR 2.12; 95% CI 1.02, 4.41; P value=0.045) compared to postmenopausal
women with the lowest SSBs consumption. In multivariate model 2, we adjusted for
tertiles of the propensity score based on variables already included in multivariate model
1, and results remained unchanged.
Table 8. Hazard ratio (HR) and 95% confidence intervals (CI) of confirmed breast cancer cases
for each category of SSB consumption among premenopausal and postmenopausal women of
the SUN Project.
Never/seldom Regular consumption P value
Premenopausal breast cancer
N 3236 6180
Cases/person-years 20/36,405 37/59,035
Incidence rate /10.000 person-years 7.57 6.28
Age-adjusted model 1.00 1.18 (0.67, 2.06) 0.570
Multivariate model 1 1.00 1.16 (0.66, 2.07) 0.602
Multivariate model 2 1.00 1.13 (0.64, 2.01) 0.669
Postmenopausal breast cancer
N 1610 1235
Cases/person-years 14/10,644 20/7,054
Incidence rate /10.000 person-years 13.15 28.35
Age-adjusted model 1.00 2.08 (1.04, 4.13) 0.037
Multivariate model 1a 1.00 2.12 (1.01, 4.41) 0.045
a 1.00 2.05 (1.02, 4.13) 0.045
Multivariate model 2
Never/seldom: no consumption, barely intake throughout the year; Regular consumption: ≥1 serving/month;
Multivariable model 1 adjusted for height, family history of breast cancer (yes/no), smoking status (never smoker,
former smoker, current smoker), physical activity (METs-h/week, continuous), alcohol intake (g/d, continuous), BMI
(tertiles), age of menarche (<10 years, 10-11 years, 12-13 years, 14-16 years, >16 years), number of pregnancies of
more than 6 months (continuous), pregnancy before the age of 30 years (yes/no), months of breastfeeding
(continuous), years at university (continuous), diabetes (yes/no), glycemic index (continuous), total energy intake
(kcal/d, continuous), decaffeinated coffee intake (continuous), coffee intake (continuous), ultra-processed food
consumption (continuous) and Mediterranean diet adherence (continuous). Multivariable model 2 adjusted by tertiles
of propensity score based on variables in model 1; a additionally adjusted for time since recruitment, age at menopause
(<50 years, (<50 years, ≥50 years), use of hormone replacement therapy (yes/no) and its duration (continuous).
98
2. Sugar-sweetened, artificially sweetened beverages and breast cancer risk in
the Nurses’ Health Studies
The aim of this study was to investigate the association between SSBs and ASBs
and breast cancer incidence in the original Nurses’ Health Study (NHS) and the Nurses’
Health Study II (NHSII). During 3,990,008 person-years of follow-up, we identified 11,379
invasive breast cancer cases (NHS 7,495; NHSII 3,884).
Table 9. Age and age-standardized baseline characteristics of women according to consumption of

sugar-sweetened beverages at baseline (NHS,1980; NHSII, 1991).
NHS NHSII
<1/ 1-4/ 2-6/ ≥1/ <1/ 1-4/ 2-6/ ≥1/
month month week day month month week day
Participants, n 31,702 18,424 20,326 12,261 31,171 22,613 24,137 15,164
Age* 47.8 46.9 45.6 44.5 37.4 36.6 36.2 35.8
(7) (7.1) (7.1) (7.1) (4.5) (4.6) (4.7) (4.8)
SSB, serv/day 0 0.1 0.5 1.9 0 0.1 0.5 2
(0) (0) (0.2) (1.2) (0) (0) (0.2) (1.2)
ASB, serv/day 0.7 0.3 0.3 0.4 1.5 1.1 0.7 0.5
(1.1) (0.7) (0.6) (0.8) (1.6) (1.4) (1.1) (1)
Self-reported African heritage (%) 0.9 1.3 1.8 3.1 0.8 1.1 2.0 3.0
Age at menarche <12 years, % 24.3 22.4 21.0 21.1 27.1 25.1 22.7 21.1
BMI, kg/m2 24.4 24.2 24.5 25 24.8 24.5 24.1 24.5
(4.4) (4.3) (4.5) (5.1) (5.3) (5.2) (5.1) (5.8)
BMI at age 18 years, kg/m2 21.7 21.3 21.1 21.1 21.8 21.3 20.8 20.7
(3.1) (2.9) (2.8) (3.1) (3.5) (3.2) (3.1) (3.4)
Weight change since age 18 years, kg 7 7.9 9 10.5 8 8.7 8.9 10.1
(10.6) (10) (10.3) (11.6) (11.8) (11.1) (10.7) (12)
Ever oral contraceptive use, % 50.2 49.2 49.3 49.3 84.9 84.4 84.3 84.7
Age at menopause (natural) 49 49 49 48.8 37.1 37.7 39.1 37.9
(3.6) (3.8) (4.1) (4.7) (6.1) (5.2) (4.2) (4.8)
Parous, % 91.8 92.8 93.3 92.5 71.4 74.5 77.7 75.3
Parity, n 3.1 3.1 3.2 3.1 2.1 2.1 2.1 2.1
(1.5) (1.5) (1.5) (1.5) (0.9) (0.9) (0.9) (0.9)
Breastfeeding, ≤6 months 35.9 36.2 35.8 34.8 16.9 16.5 16.8 18.1
Family history breast cancer, % 6.1 6.2 6.1 6.1 6.1 5.9 5.8 6.1
History of benign breast disease, % 24.9 24.9 23.7 23.9 9.6 9.4 9.5 10.2
Height, inches, continuous 64.5 64.5 64.5 64.5 64.9 64.9 64.8 64.8
(2.4) (2.4) (2.4) (2.4) (2.6) (2.6) (2.6) (2.6)
Physical activity, METs-h/week 13.2 12 11.4 10.4 23.3 20.8 19.2 18.3
(14) (13.3) (12.6) (11.9) (29.7) (26.5) (25.1) (26)
Alcohol consumption, g/day 7.8 5.8 5.2 5.3 3.4 3.2 3 2.5
(11.6) (9.8) (9.2) (10.3) (6.5) (6) (5.8) (5.6)
Alternative Healthy Eating Index, 42.1 39.8 38.2 36.9 48.6 44.5 40.1 35.6
score (9.4) (8.8) (8.6) (8.6) (9.9) (9.1) (8.4) (8.3)
Total energy intake, kcal/day 1454 1525 1645 1834 1618 1726 1863 2118
(462) (472) (483) (534) (501) (508) (528) (562)
Fruit consumption, serv/day 2.2 2.1 2 2 1.3 1.2 1.2 1
(1.6) (1.4) (1.4) (1.5) (1) (0.9) (0.9) (0.9)
Vegetable consumption, serv/day 2.1 1.9 1.9 1.9 3.2 3 2.9 2.7
(1.3) (1.1) (1.1) (1.1) (2.1) (1.7) (1.7) (1.7)
Coffee consumption, serv/day 2.5 2.4 2.2 1.9 1.8 1.6 1.4 1.2
(2.1) (2) (2) (1.9) (1.8) (1.7) (1.6) (1.5)
ASB= Artificially sweetened beverages; BMI= Body Mass Index; METs= Metabolic equivalents; NHS= Nurses’ Health Study; NHSII=
Nurses’ Health Study II; SSB= Sugar-sweetened beverages. Values are means (standard deviations) for continuous variables and
percentages for categorical variables. All variables except age are standardized to the age distribution of the study population. Each
food intakes expressed as servings/day. * Value is not age adjusted.
99
Baseline characteristics of participants according to frequency of SSB and ASB
consumption are shown in Table 9. Women with higher SSBs consumption were more
likely to be younger, less physically active, and to have a lower diet quality as measured
by the Alternate Healthy Eating Index (AHEI) score.
SSB consumption was also associated with higher total energy intake, and lower
intake of fruits, vegetables, coffee and alcohol. Women with higher ASBs consumption
had a greater BMI, were more physically active, consumed more alcohol and had higher
AHEI scores. Frequent consumption of ASBs was also associated with a lower intake of
total energy, and a higher intake of fruits, vegetables and coffee (Table 10).
Table 10. Age and age-standardized baseline characteristics of women according to consumption of
artificially sweetened beverages at baseline (NHS, 1980; NHSII, 1991).
NHS NHSII
<1/ 1-4/ 2-6/ ≥1/ <1/ 1-4/ 2-6/ ≥1/
month month week day month month week day
Participants, n 41,895 13,142 13,239 14,437 27,886 9,649 21,456 34,094
Age* 46.8 46.8 46.7 45.7 36.5 36.7 36.8 36.6
(7.2) (7.1) (7.1) (7.1) (4.6) (4.7) (4.6) (4.7)
SSB, serv/day 0.5 0.3 0.3 0.4 0.9 0.5 0.3 0.2
(0.9) (0.6) (0.6) (0.8) (1.1) (0.8) (0.6) (0.5)
ASB, serv/day 0 0.1 0.5 1.8 0 0.1 0.6 2.4
(0) (0) (0.2) (1.2) (0) (0) (0.2) (1.5)
Self-reported African heritage (%) 1.6 1.5 1.4 1.5 2.4 1.8 1.5 0.8
Age at menarche <12 years, % 20.5 23.8 24.3 25.5 20.3 23.9 25.4 27.6
BMI, kg/m2 23.5 24.8 25.4 26 23.2 24 24.4 25.7
(4.1) (4.4) (4.6) (5) (4.8) (5.1) (4.9) (5.7)
BMI at age 18 years, kg/m2 20.9 21.6 21.9 22.1 20.4 21.1 21.3 22
(2.8) (3) (3) (3.3) (3) (3) (3.1) (3.6)
Weight change since age 18 years, kg 7 8.5 9.5 10.4 7.7 8 8.4 10
(9.7) (10.4) (11) (12.1) (10.2) (10.9) (10.9) (12.6)
Ever oral contraceptive use, % 48.4 50.6 51.0 51.3 82.2 82.9 85.4 86.6
Age at menopause (natural) 49 49 49.1 48.8 38.8 36.8 37.2 37.7
(3.7) (3.8) (3.9) (4.4) (4.2) (5.9) (5.8) (5.8)
Parous, % 92.5 92.7 92.8 92.1 76.4 74.8 75.7 71.9
Parity, n 3.2 3.1 3.1 3.1 2.1 2.1 2.1 2.1
(1.5) (1.5) (1.5) (1.4) (0.9) (0.9) (0.9) (0.9)
Breastfeeding, ≤6 months 34.9 36.6 37.3 36.0 15.5 15.6 16.7 19.0
Family history breast cancer, % 6.1 6.0 6.1 6.3 5.9 5.7 6.2 5.9
History of benign breast disease, % 25.0 23.7 24.0 23.8 10.2 9.3 9.6 9.2
Height, inches, continuous 64.5 64.5 64.6 64.6 64.8 64.9 64.9 65
(2.4) (2.4) (2.4) (2.4) (2.6) (2.6) (2.6) (2.6)
Physical activity, METs-h/week 11.5 12.5 12.8 12.4 19.5 20.6 21.6 21.5
(12.9) (13.5) (13.6) (13.5) (26.4) (26) (27.8) (27.8)
Alcohol consumption, g/day 6.4 5.9 6.5 6.8 2.7 2.9 3.3 3.4
(10.7) (10) (10) (11) (5.9) (5.5) (6) (6.3)
Alternative Healthy Eating Index, 39 40.7 40.9 40.7 41.2 44.5 44.5 43.9
score (9.4) (9) (8.8) (9) (10.8) (10.6) (9.8) (9.7)
Total energy intake, kcal/day 1607 1530 1533 1550 1861 1759 1742 1770
(506) (481) (480) (499) (561) (538) (528) (547)
Fruit consumption, serv/day 2 2.1 2.2 2.2 1.2 1.3 1.3 1.2
(1.5) (1.4) (1.4) (1.5) (1) (1) (0.9) (0.9)
Vegetable consumption, serv/day 1.9 2 2 2.1 2.8 3 3 3.1
(1.2) (1.2) (1.2) (1.3) (1.8) (1.8) (1.8) (1.9)
100
Table 10 (continuation). Age and age-standardized baseline characteristics of women according to
consumption of artificially sweetened beverages at baseline (NHS, 1980; NHSII, 1991).
Coffee consumption, serv/day 2.3 2.3 2.3 2.2 1.3 1.6 1.7 1.6
(2.0) (2.0) (2.0) (2.0) (1.6) (1.7) (1.7) (1.7)
ASB= Artificially sweetened beverages; BMI= Body Mass Index; METs= Metabolic equivalents; NHS= Nurses’ Health Study;
NHSII= Nurses’ Health Study II; SSB= Sugar-sweetened beverages. Values are means (standard deviations) for continuous
variables and percentages for categorical variables. All variables except age are standardized to the age distribution of the study
population. Each food intakes expressed as servings/day. * Value is not age adjusted.
Sugar-sweetened beverage consumption and overall breast cancer risk

In the NHS, a suggestive positive, though not statistically significant, association
was observed in multivariable-adjusted model 1 (HR,1.11;95% 1,00, 1.22; P-trend=0.06)
(Table 11). Multivariable model 1 and model 2 (where change in weight since age 18
was further adjusted for) showed similar results.
Table 11. Total Breast Cancer Incidence According to cumulative average intake of SSBs in the
NHS, NHSII and pooled cohorts.
<1/month 1-4/month 2-6/week ≥1/day P-trend

Number of cases 2,141 1,947 2,864 543
Person-years 691,530 608,056 901,755 195,256
NHS
Age-adjusted model 1.00 1.01 (0.95, 1.07) 1.00 (0.94, 1.06) 1.04 (0.95, 1.15) 0.53
Multivariate model 1 1.00 1.01 (0.94, 1.07) 1.01 (0.95, 1.07) 1.11 (1.00, 1.22) 0.06
Number of cases 928 984 1,507 465
Person-years 557,608 540,283 848,158 312,506
NHSII
Number of cases 3,069 2,931 4,371 1,008
Person-years 1,249,138 1,148,340 1,749,913 507,762
Pooled*
Multivariable model 1: stratified by age in months and calendar year, adjusted for ASBs intake, race (Non-Hispanic
Caucasian, African, Asian, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause
(premenopausal, <45, 45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current user– estrogen only
for <5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen +
progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at first birth
(nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age),
breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no),
history of benign breast disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m), cumulatively updated
alcohol intake (0, <5, 5-9, 10-14, 15+ g/day), cumulatively updated total caloric intake (kcal/day, quintiles), physical activity
(linear MET-h/week), body mass index at age 18 years (<20.0, 20.0-21.9, 22.0-23.9, 24.0-26.9, ≥27.0), a modified Alternate
Healthy Eating Index (AHEI) score (with SSBs and alcohol removed) and socioeconomic status (continuous). Multivariable
model 2: Multivariable model 1 + change in weight since age 18 (lost ≥ 2, lost 0-1 or gained 0-2, gained 3-5, gained 6-10,
gained 11-20, gained 21-25, gained > 25 kg). *Pooled model also stratified by cohort.
101
Artificially sweetened beverage consumption and overall breast cancer risk
Compared with women who consumed ASBs less than once per month, women
who consumed ≥1 serving of ASBs per day had a 7% lower risk of breast cancer
(Multivariable Model 1, HR 0.93; 95% CI 0.84, 1.01; P-trend=0.04). Additional adjustment
for weight change since age 18 (Multivariable Model 2), slightly reinforced the
association (HR 0.91, 95% CI 0.83, 1.00; P-trend=0.02) (Table 12).
Table 12. Total Breast Cancer Incidence According to cumulative average intake of ASBs in
the NHS, NHSII and pooled cohorts.
Artificially sweetened beverage category
Number of cases 2,242 1,123 2,870 1,260
Person-years 758,074 353,549 866,359 418,616
NHS
Multivariate model 2 1.00 1.00 (0.93, 1.08) 0.97 (0.92,1.03) 0.99 (0.92, 1.07) 0.87
Number of cases 935 408 1,327 1,214
Person-years 548,370 225,896 703,317 780,973
NHSII
Age-adjusted model 1.00 1.02 (0.90, 1.14) 1.00 (0.92, 1.09) 0.88 (0.81, 0.97) <0.01
Number of cases 3,177 1,531 4,197 2,474
Person-years 1,306,444 579,445 1,569,676 1,199,588
Pooled*
Multivariable model 1: stratified by age in months and calendar year, adjusted for SSBs intake, race (Non-Hispanic
(premenopausal, <45, 45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current user–
estrogen only for <5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years,
current estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever),
parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25,
2+ children ≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family
history of breast cancer (yes or no), history of benign breast disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69,
1.70-1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5-9, 10-14, 15+ g/day), cumulatively updated total
caloric intake (kcal/day, quintiles), physical activity (linear MET-h/week), body mass index at age 18 years (<20.0,
20.0-21.9, 22.0-23.9, 24.0-26.9, ≥27.0), a modified Alternate Healthy Eating Index (AHEI) score (with SSBs and alcohol
removed) and socioeconomic status (continuous). Multivariable model 2 = Multivariable model 1 + change in weight
since age 18 (lost ≥ 2, lost 0-1 or gained 0-2, gained 3-5, gained 6-10, gained 11-20, gained 21-25, gained > 25 kg).
*Pooled model also stratified by cohort.
Furthermore, we conducted stratified analyses by BMI, physical activity and diet

quality (using the AHEI score) for total invasive breast cancer as the main endpoint
(Figure 13). Although we did not observe a significant interaction by BMI (P-
interaction=0.08), a modest higher risk of breast cancer with each serving per day
increment of SSBs was found in normal weight women (BMI<25 kg/m2) (HR 1.06, 95%
CI 1.01, 1.11), but not among overweight or obese women (BMI ≥ 25 kg/m 2) (HR 1.00,
0.95, 1.06). Although interaction by physical activity (<21 METs-h/week vs. ≥21 METs-
h/week) was not significant (P-interaction=0.23), each serving per day of ASBs in women
102
with high activity level was associated with 4% lower risk of breast cancer (HR 0.96, 95%
CI, 0.91, 1.00).
Figure 13. Total breast cancer according to SSB intake (servings/day) and ASB intake stratified
by BMI, physical activity, and diet quality (AHEI score) based on pooled data from both cohorts
(NHS/NHSII).
Stratified by age in months, cohort and calendar year and adjusted for SSBs or ASBs intake; race (Non-Hispanic
(premenopausal, <45, 45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current user– estrogen
only for <5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current
estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity
and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children
≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of
breast cancer (yes or no), history of benign breast disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74,
1.75 + m), body mass index at age 18 years (<20.0, 20.0-21.9, 22.0-23.9, 24.0-26.9, ≥27.0), physical activity (linear
MET-h/week), diet quality using a modified Alternate Healthy Eating Index (AHEI) score (with SSBs and alcohol removed),
cumulatively updated alcohol intake (0, <5, 5-9, 10-14, 15+ g/day), cumulatively updated total caloric intake (kcal/day,
quintiles), change in weight since age 18 (lost ≥ 2, lost 0-1 or gained 0-2, gained 3-5, gained 6-10, gained 11-20, gained
21-25, gained > 25 kg) and socioeconomic status (continuous).
103
Sugar-sweetened beverage and artificially sweetened beverage consumption and
breast cancer risk by subtypes
Table 13 shows breast cancer incidence by ER status and molecular phenotype
according to cumulative average intake categories of SSBs and ASBs using pooled data
from the Nurses’ Health Studies. SSB intake was not associated with an increased risk
of any breast cancer subtype, and we did not observe heterogeneity by ER status or
molecular phenotype.
Table 13. Breast Cancer Incidence by estrogen receptor status and molecular subtypes according
to cumulative average intake of SSBs using pooled data from the Nurses' Health Studies (NHS
and NHSII).
By ER status (NHS with follow-up of 1980-2016 and NHSII with follow-up of 1991-2017)
Estrogen receptor positive (n=7,302 cases)
Number of cases 2,019 1,852 2,832 599
Person-years 1,250,140 1,149,365 1,751,379 508,153
Multivariate model 1.00 0.99 (0.92, 1.05) 0.99 (0.93, 1.06) 1.02 (0.92, 1.13) 0.68
Estrogen receptor negative (n=1,693 cases)
Number of cases 453 428 635 177
Person-years 1,251,580 1,150,699 1,753,462 508,552
P- heterogeneity‡ by ER status= 0.99
By molecular subtype* (NHS with follow-up of 1980-2006 and NHSII with follow-up of 1991-2005)
Luminal A (n= 2,673 cases)
Person-years 942,382 818,719 1,192,135 384,338
Luminal B (n= 1,147 cases)
Person-years 942,773 819,060 1,192,571 384,477
HER-2 (n= 250 cases)
Person-years 942,994 819,265 1,192,915 384,538
Basal-like (n=303 cases)
Person-years 942,980 819,259 1,192,889 384,529
P- heterogeneity‡ by molecular subtype= 0.65
Multivariate model stratified by age in months, cohort, and calendar year, adjusted for ASBs intake, race (Non-Hispanic
only for <5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen
+ progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at
first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years
of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer
(yes or no), history of benign breast disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m),
quintiles), physical activity (linear MET-h/week), body mass index at age 18 years (<20.0, 20.0-21.9, 22.0-23.9, 24.0-26.9,
≥27.0), a modified Alternate Healthy Eating Index (AHEI) score (with SSBs and alcohol removed), socioeconomic status
and change in weight since age 18 (lost ≥ 2, lost 0-1 or gained 0-2, gained 3-5, gained 6-10, gained 11-20, gained 21-25,
gained > 25 kg).
*Due to smaller sample sizes in analyses, to ensure that models would run, covariate categorizations were simplified.
‡ For testing heterogeneity by subtype, we used the Lunn-McNeil approach.
104
Nonetheless, ASB intake was associated with a lower risk of luminal A breast
tumors, HR (95% CI) across categories (<1/month, 1-4/month, -6/week, ≥1/day) were
1.00 (reference), 1.05 (0.93, 1.19), 0.95 (0.86, 1.05), 0.90 (0.80, 1.01), P-trend=0.02.
(Table 14). Each serving per day increment in ASB was associated with a 6% lower risk
of this subtype (HR, 0.94; 95% CI, 0.90, 0.98).
Table 14. Breast Cancer Incidence by estrogen receptor status and molecular subtypes according
to cumulative average intake of ASBs using pooled data from the Nurses' Health Studies (NHS
and NHSII).
By ER status (NHS with follow-up of 1980-2016 and NHSII with follow-up of 1991-2017)
Estrogen receptor positive (n=7,302 cases)
Number of cases 2,014 1,011 2,715 1,562
Person-years 1,307,507 579,966 1,571,087 1,200,478
Estrogen receptor negative (n=1,693 cases)
Person-years 1,308,930 580,719 1,573,107 1,201,537
P- heterogeneity‡ by ER status=0.62
By molecular subtype* (NHS with follow-up of 1980-2006 and NHSII with follow-up of 1991-2005)
Luminal A (n=2,673 cases)
Person-years 1,006,598 422,837 1,055,686 852,453
Luminal B (n=1,147 cases)
Person-years 1,007,001 423,043 1,056,133 852,706
HER-2 (n=250 cases)
Person-years 1,007,221 423,157 1,056,440 852,895
Basal-like (n=303 cases)
Person-years 1,007,204 423,158 1,056,422 852,875
P- heterogeneity‡ by molecular subtype=0.20
Multivariate model stratified by age in months, cohort, and calendar year, adjusted for SSBs intake, race (Non-Hispanic
only for <5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen
+ progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at
first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years
of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer
≥27.0), a modified Alternate Healthy Eating Index (AHEI) score (with SSBs and alcohol removed), socioeconomic status
and change in weight since age 18 (lost ≥ 2, lost 0-1 or gained 0-2, gained 3-5, gained 6-10, gained 11-20, gained 21-25,
gained > 25 kg).
*Due to smaller sample sizes in analyses, to ensure that models would run, covariate categorizations were simplified.
‡ For testing heterogeneity by subtype, we used the Lunn-McNeil approach.
105
Furthermore, we also tested whether ASB intake and ER-positive or luminal A
breast tumors differed by BMI (in kg/m 2) in Figure 14. Each serving per day increment
in ASBs was associated with an 8% and 6% lower relative risk of luminal A and ER-
positive breast cancer, respectively, among overweight/obese women [(HR 0.92; 95%
CI 0.88, 0.96); (HR 0.94; 95% CI 0.91, 0.98)] but not among normal weight women [(HR
0.97; 95% CI 0.92, 1.02); (HR 0.99; 95% CI 0.95, 1.03)] and P-interaction was 0.10 and
<0.01.
ER-positive
Luminal A
ER-positive
Luminal A
Figure 14. Estrogen receptor positive and Luminal A breast cancer risk according to SSB intake
(servings/day) and ASB intake stratified by BMI based on pooled data from both cohorts.
Stratified by age in months, cohort and calendar year and adjusted for SSBs or ASBs intake; race (Non-Hispanic
only for <5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current
estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and
age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25
years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast
cancer (yes or no), history of benign breast disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m),
physical activity (linear MET-h/week), diet quality using a modified Alternate Healthy Eating Index (AHEI) score (with SSBs
and alcohol removed), cumulatively updated alcohol intake (0, <5, 5-9, 10-14, 15+ g/day), cumulatively updated total
caloric intake (kcal/day, quintiles), change in weight since age 18 (lost ≥ 2, lost 0-1 or gained 0-2, gained 3-5, gained 6-
10, gained 11-20, gained 21-25, gained > 25 kg) and socioeconomic status.
No interaction by menopausal status was observed for SSB or ASB, nonetheless,

a suggestive, non-significant, higher risk of postmenopausal breast cancer was observed
among women who consumed ≥1 serving of SSBs per day compared to women in the
lowest category (HR 1.07; 95% CI 0.97, 1.19; P-trend=0.40) (Table 15). Furthermore,
women who consumed ≥1serving/day of ASBs compared to those with a low
consumption (<1 serving/month), had a 10% relative lower risk of premenopausal breast
106
cancer (HR 0.90; 95% CI 0.80, 1.01) although test for a linear trend was not significant
(P-trend=0.20).
Table 15. Multivariable hazard ratios (HRs) and 95% confidence intervals (CIs) for the association
between categories of cumulatively updated SSB and ASB intake and premenopausal or
postmenopausal breast cancer risk using pooled data from the Nurses’ Health Studies (NHS and
NHSII).
P-trend Per-1 serving/day
<1/month 1-4/month 2-6/week ≥1/day
Premenopausal breast cancer (n= 2,503 cases)
Cases/person-years 671/421,460 596/389,514 906/578,796 330/236,214
Multivariate model 1.00 0.94 (0.84, 1.06) 0.90 (0.81, 1.01) 0.96 (0.82, 1.12) 0.95 1.02 (0.95, 1.08)
Postmenopausal breast cancer (n=8,140 cases)
Cases/person-years 2195/717,260 2160/662,476 3203/1,018,196 582/208,073
P-interaction by menopausal status = 0.20
P-trend Per-1 serving/day
Premenopausal breast cancer (n= 2,503 cases)
Postmenopausal breast cancer (n= 8,140 cases)
P-interaction by menopausal status = 0.71
Multivariate model stratified by age in months, cohort and calendar year, adjusted for SSBs or ASBs, race (Non-Hispanic
Caucasian, African-American, Asian-American, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at
menopause (premenopausal, <45, 45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current user–
estrogen only for <5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current
estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and
age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25
years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast
cancer (yes or no), history of benign breast disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m),
≥27.0), neighborhood-based socioeconomic status indicator (continuous) and change in weight since age 18 (lost ≥ 2, lost
0-1 or gained 0-2, gained 3-5, gained 6-10, gained 11-20, gained 21-25, gained > 25 kg).
We also tested whether SSBs and breast cancer differed by BMI in

premenopausal or postmenopausal women (Table 16), but no further associations were
found.
107
Table 16. Multivariable hazard ratios (HRs) and 95% confidence intervals (CIs) for the association
between categories of cumulatively updated SSB intake and premenopausal or postmenopausal breast
cancer risk stratified by BMI using pooled data from NHS and NHSII.
P-trend
BMI <25 kg/m2
Multivariable model 1.00 0.92 (0.80, 1.05) 0.90 (0.79, 1.03) 0.97 (0.81, 1.16) 0.78
BMI ≥25 kg/m2
P for interaction by BMI=0.14
BMI <25 kg/m2
BMI ≥25 kg/m2
P for interaction by BMI=0.96
Model stratified by age in months, cohort and calendar year, adjusted for SSBs or ASBs, race (Non-Hispanic Caucasian, African-
American, Asian-American, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause
(premenopausal, <45, 45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current user– estrogen only for
<5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin
user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at first birth (nulliparous,
1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding h istory
(never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast
disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5-9, 10-
14, 15+ g/day), cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass
index at age 18 years (<20.0, 20.0-21.9, 22.0-23.9, 24.0-26.9, ≥27.0) and neighborhood-based socioeconomic status indicator
(continuous).
108
3. Phenolic acids subclasses, individual compounds and breast cancer risk in
the SUN project
Based on a previous analysis regarding total and subtypes of polyphenol intake and
breast cancer risk in the SUN Project187, we aimed to further evaluate the association
between subclasses and individual compounds of phenolic acids, and breast cancer risk
in the SUN cohort study. Phenolic acid intake was derived from the foods included in the
FFQ using Phenol Explorer358. In the SUN project, reported median phenolic acid intake
was 260 mg/day (interquartile range, 170 to 376 mg/day) accounting for 40% of total
polyphenol intake. Moreover, hydroxycinnamic acids were the main source (37%) of total
polyphenol intake in this sample of 10,812 women, as represented in Figure 15.
6,8 % 0,1 %
Flavonoids
0,1 %
A Phenolic acids
Other polyphenols
40% 53%
Lignans
Stilbenes
2% Flavanols
6% 7% Hydroxycinnamic acids
4% 34% Anthocyanins
3% Hydroxybenzoic acids
B 7%
Flavanones
Flavonols
37% Other polyphenols

Flavones
Figure 15. Main components of total polyphenols (%) according to their chemical structure
(A) and their main subclasses and individual compounds (B) in a sample of 10,812 women from
the SUN cohort (1999-2018).
Baseline characteristics of participants according to tertiles of phenolic acids

intake in the SUN cohort are shown in Table 17. Women in the highest tertile of phenolic
acid intake were older, were more likely to be physically active, had lower total energy
intake, had higher alcohol intake, were more likely to be former smokers, and had higher
adherence to the Mediterranean diet.
109
Table 17. Baseline characteristics of 10,812 women according to tertiles of phenolic acid
intakes in the SUN Project, 1999-2018.
Phenolic acid intake
Variable Tertile 1 Tertile 2 Tertile 3
N 3,604 3,604 3,604
Median phenolic acid intake (mg/d) 135 259 416
Age (years) 31.7 (9.9) 35.4 (10.7) 37.0 (10.5)
2
Body mass index (kg/m ) 21.9 (3) 22.2 (3) 22.5 (3.1)
Physical activity (METsa-h/week) 19.9 (19) 20.2 (18.8) 20.5 (18.6)
Total energy intake (Kcal/d) 2335 (575) 2272 (573) 2283 (568)
Alcohol intake (g/d) 3.2 (5) 4.2 (6) 4.6 (6.5)
Years at university 4.7 (1.3) 4.9 (1.3) 4.9 (1.4)
Height (cm) 164 (6) 163 (6) 163 (6)
Number of pregnancies of more than 6 months 0.2 (0.6) 0.2 (0.7) 0.3 (0.7)
Pregnancy before 30 years, n (%) 202 (5.6) 266 (7.4) 300 (8.3)
Breast-feeding (months) 1.9 (4.6) 2.5 (5.1) 2.7 (5.3)
b
Hormone replacement therapy (yes), n (%) 109 (40.4) 148 (32.7) 192 (38.6)
Time of hormone replacement therapy (years) b 1.2 (2.3) 1.3 (2.6) 1.3 (2.2)
Diabetes, n (%) 23 (0.6) 47 (1.3) 48 (1.3)
Smoking, n (%)
Never smoker 2,201 (61.1) 1,858 (51.6) 1,466 (40.7)
Former smoker 684 (19) 770 (26.4) 1,007 (30.6)
Current smoker 686 (19) 950 (21.4) 1,103 (27.9)
c
Family history of breast cancer, n (%)
No 3,215 (89.2) 3,248 (90.1) 3,206 (89.0)
Yes 389 (10.8) 356 (9.9) 398 (11.1)
Age of menarche, n (%)
< 10 years 7 (2.6) 4 (0.9) 2 (0.4)
10-11 years 51 (18.9) 93 (20.6) 108 (21.7)
12-13 years 158 (58.5) 245 (54.2) 268 (53.9)
14-16 years 47 (17.4) 105 (23.2) 105 (21.1)
≥16 years 7 (2.6) 5 (1.1) 14 (2.8)
Menopausal status, n (%)
Premenopausal 2,892 (80.2) 2,496 (69.3) 2,188 (60.7)
Postmenopausal 712 (19.8) 1,108 (30.7) 1,416 (39.3)
b
Age at menopause, n (%)
Postmenopausal < 50 years 174 (64.4) 293 (64.8) 333 (67)
Postmenopausal ≥ 50 years 96 (35.6) 159 (35.2) 164 (33)
Mediterranean diet adherence 3.6 (1.7) 4.3 (1.8) 4.5 (1.8)
a
METs: metabolic equivalents; b Only for postmenopausal women. Values are expressed as mean (SD), unless
otherwise stated.c Information from mother, sisters, and both grandmothers was collected.
Individual and cumulative contributions of different foods to the variability in total

phenolic, hydroxycinnamic, hydroxybenzoic, and chlorogenic acid intake according to
food groups are presented in Table 18. Food groups that contributed most to total
phenolic acid intake were coffee (48%), fruits (25%), and vegetables (23%). In fact, 99%
of between-person variability in total phenolic acid intake was explained by these food
groups (coffee, fruits, and vegetables).
110
Table 18. Main sources of variability (cumulative R2 and change in R 2) and main sources of
quantity (%) of total phenolic, hydroxycinnamic, hydroxybenzoic and chlorogenic acid intake
according to food groups. The SUN Cohort, 1999-2018.
Food group Cumulative R2 a Change in R2 Sources of quantity (%)
Total phenolic acid intake
Coffee b 0.78 48
c
Fruits 0.94 0.17 25
Vegetables d 0.99 0.05 23
Hydroxycinnamic acid intake
Coffee 0.83 51
Vegetables 0.92 0.09 26
Fruits 0.99 0.08 18
Hydroxybenzoic acid intake
Fruits 0.75 51
Nuts e 0.99 0.24 29
f
Alcohol 0.99 0.01 9
Chlorogenic acid intake
Coffee 0.79 52
Fruits 0.91 0.13 22
a 2
Cumulative R values were determined from nested regression analyses after stepwise selection.
b
Regular and decaffeinated coffee
c
Cherry, strawberry, apple, peach, grape, kiwi, mango, orange, banana, avocado, olive, canned fruit and dried fruit.
d
Swiss chard, spinach, artichoke, thistle, leek, tomato, carrot, pumpkin, lettuce, endive, eggplant, zucchini, beans,
pepper, cabbage, broccoli, cauliflower, gazpacho, asparagus, potato (roasted and fried).
e
Walnuts, peanuts, hazelnut and almonds.
f
Wine, Beer, other wines and whisky
Subclasses of phenolic acid intake and risk of breast cancer

During a median follow-up time of 11.5 years and 115,368 person-years at risk,
190 probable incident cases of breast cancer were identified, of which 101 were
confirmed by our team of oncologists after reviewing the medical records. An inverse
association between hydroxycinnamic acids intake and postmenopausal breast cancer
risk was found for both confirmed (HR T3 vs T1 0.37, 95% CI 0.16, 0.85; P-trend=0.029)
and probable (HR T3 vs T1 0.43, 95% CI 0.23, 0.80; P-trend=0.011) breast cancer cases.
Cox proportional hazard models were also fitted with repeated measurements using the
updated data on food consumption after 10 years of follow-up (Table 19). Results from
these analyses remained basically unchanged, and the HR comparing extreme tertiles
of hydroxycinnamic acids intake was 0.40 (95% CI 0.17, 0.92) among postmenopausal
women. These compounds were also inversely associated with breast cancer risk when
we also excluded women diagnosed with breast cancer during the first 2 years of follow-
up or after adjusting the main analysis for other potential confounders, such as fiber
intake.
111
Table 19. Hazard Ratios (HRs) and confidence intervals (95 % CIs) for the association between
tertiles of individual phenolic acid intake and confirmed breast cancer risk in the SUN Cohort.
Tertiles of hydroxycinnamic acids intake
Overall cases Tertile 1 Tertile 2 Tertile 3 P-trend
Incidence rate/10,000 PY 6.9 8.9 10.3
Age adjusted HR 1.00 1.02 (0.61, 1.70) 1.01 (0.61, 1.66) 0.98
b 1.00
Multivariable adjusted 1.00 (0.60, 1.68) 0.98 (0.60, 1.63) 0.94
c 1.00
Repeated measures 1.00 1.03 (0.61, 1.73) 1.08 (0.65, 1.80) 0.75
Premenopausal women
Age adjusted HR 1.00 1.55 (0.74, 3.22) 1.51 (0.74 3.09) 0.35
Multivariable adjusted d 1.00 1.54 (0.74, 3.22) 1.45 (0.70, 3.00) 0.43
c 1.00
Postmenopausal women
Age adjusted HR 1.00 0.46 (0.21, 1.04) 0.38 (0.17, 0.86) 0.027
e 1.00
c 1.00
Repeated measures f 1.00 0.51 (0.22, 1.17) 0.40 (0.17, 0.92) 0.046
b
PY: person-years; Adjusted for age (underlying time variable), height (continuous), family history of breast cancer
(yes/no) smoking status (never smoker, former smoker, current smoker), physical activity (METs-h/ week, continuous),
alcohol intake (g/day, continuous), BMI (tertiles), age of menarche (categories), menopause (no menopause, <50
years, ≥50 years), number of pregnancies of more than 6 months (continuous), pregnancy before the age of 30 years
(yes/no), months of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its duration
(continuous) and years at university (continuous); c Additionally adjusted for diabetes (yes/no), total energy intake
(kcal/day, continuous) and Mediterranean diet adherence (continuous); d Adjusted for age (underlying time variable),
height (continuous), family history of breast cancer (yes/no) smoking status (never smoker, former smoker, current
smoker), physical activity (METs-h/ week, continuous), alcohol intake (g/day, continuous), BMI (tertiles), age of
menarche (categories), number of pregnancies of more than 6 months (continuous), pregnancy before the age of 30
years (yes/no), months of breastfeeding (continuous), and years at university (continuous); e Adjusted for age
(underlying time variable), height (continuous), family history of breast cancer (yes/no) smoking status (never smoker,
former smoker, current smoker), physical activity (METs-h/ week, continuous), alcohol intake (g/day, continuous), BMI
(tertiles), age of menarche (categories), menopause (<50 years, ≥50 years), number of pregnancies of more than 6
months (continuous), pregnancy before the age of 30 years (yes/no), months of breastfeeding (continuous), use of
hormone replacement therapy (yes/no) and its duration (continuous) , years at university (continuous) and time since
recruitment; f Multivariable adjusted model with repeated measures (updated data at 10 years of follow-up).
Individual compounds of phenolic acid and risk of breast cancer

Furthermore, we also analyzed individual compounds that constitute the
hydroxycinnamic subclass. The main contributors to the between-person variability
(98%) and the main sources of hydroxycinnamic acids were, in decreasing order, 4-CQA,
5-CQA, and 3-CQA, which are classified as chlorogenic acids. Chlorogenic acids were
the main individual compounds of hydroxycinnamic acid, with a median intake of 162
mg/day (interquartile range: 93 to 226 mg/day). The main food sources of chlorogenic
acids were coffee, fruits, and vegetables (Table 18). A significant inverse association
112
between chlorogenic acid intake and breast cancer was observed only among
postmenopausal women when both confirmed cases (HR T3 vs T1 0.33, 95% CI 0.14, 0.78;
P-trend=0.012) (Table 20) and probable cases (HR T3 vs T1 0.42, 95% CI 0.23, 0.79; P-
trend=0.008) were evaluated. Moreover, results remained practically unchanged after
further adjusting for fiber intake, performing repeated measures after 10 years of follow-
up or excluding breast cancer cases diagnosed during the first 2 years of follow-up.
Table 20. Hazard Ratios (HRs) and confidence intervals (95 % CIs) for the association between
tertiles of chlorogenic acids intake and confirmed breast cancer risk in the SUN Cohort.
Tertiles of total chlorogenic acid intake
Tertile 1 P for
Tertile 2 Tertile 3
trend
Overall confirmed cases
Incidence rate/10,000 person-years 6.7 9.1 10.3
Age adjusted HR 1.00 1.09 (0.65, 1.82) 1.08 (0.65, 1.78) 0.80
b
Multivariable adjusted 1.00 1.06 (0.63, 1.79) 1.06 (0.64, 1.76) 0.83
c
f
Premenopausal women
Age adjusted HR 1.00 1.51 (0.72, 3.17) 1.62 (0.79, 3.30) 0.22
Multivariable adjusted d 1.00 1.49 (0.71, 3.14) 1.57 (0.76, 3.23) 0.26
Multivariable adjusted c 1.00 1.49 (0.70, 3.16) 1.62 (0.78, 3.38) 0.23
f
Age adjusted HR 1.00 0.52 (0.23, 1.14) 0.34 (0.15, 0.80) 0.013
e
c
f
b
Adjusted for age (underlying time variable), height (continuous), family history of breast cancer (yes/no) smoking
status (never smoker, former smoker, current smoker), physical activity (METs-h/ week, continuous), alcohol intake
(g/day, continuous), BMI (tertiles), age of menarche (categories), menopause (no menopause, <50 years, ≥50 years),
number of pregnancies of more than 6 months (continuous), pregnancy before the age of 30 years (yes/no), months
of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its duration (continuous) and years
at university (continuous); c Additionally adjusted for diabetes (yes/no), total energy intake (kcal/day, continuous) and
Mediterranean diet adherence (continuous); d Adjusted for age (underlying time variable), height (continuous), family
history of breast cancer (yes/no) smoking status (never smoker, former smoker, current smoker), physical activity
(METs-h/ week, continuous), alcohol intake (g/day, continuous), BMI (tertiles), age of menarche (categories), number
of pregnancies of more than 6 months (continuous), pregnancy before the age of 30 years (yes/no), months of
breastfeeding (continuous), and years at university (continuous); e Adjusted for age (underlying time variable), height
(continuous), family history of breast cancer (yes/no) smoking status (never smoker, former smoker, current smoker),
physical activity (METs-h/ week, continuous), alcohol intake (g/day, continuous), BMI (tertiles), age of menarche
(categories), menopause (<50 years, ≥50 years), number of pregnancies of more than 6 months (continuous),
pregnancy before the age of 30 years (yes/no), months of breastfeeding (continuous), use of hormone replacement
therapy (yes/no) and its duration (continuous) , years at university (continuous) and time since recruitment; f
Multivariable adjusted model with repeated measures (updated data at 10 years of follow-up).
113
4. Fat intake and breast cancer risk in the SUN project
Total fat intake contributed to 37.1% of total energy intake. MUFA was the major
contributor to total energy intake (16.0%) and olive oil was its main contributor. PUFA
and SFA contributed to 5.2% and 12.5%, respectively. Baseline characteristics of the
study population according to extreme tertiles of total fat, MUFA, PUFA and SFA intake
are shown in Table 21. Women in the highest tertile (T3) of all fat components were
more likely to be younger, less active, current smokers and spend less time with hormone
replacement therapy compared to those in the lowest tertile (T1). Daily energy intake
and mean intake (% of energy) of total fat, MUFA, PUFA and SFA were higher among
women in the highest tertile (T3) compared to women in the lowest tertile (T1).
Carbohydrate intake (% of energy) was also found to be lower among women in the
highest intake of total fat, MUFA, PUFA and SFA. Moreover, physical activity, expressed
in METs-h/week, was higher among women in the first tertile of total fat, MUFA, PUFA
and SFA.
Table 21. Baseline characteristic for the entire cohort at enrolment and by tertiles of total fat
intake as a percentage of energy among 10,812 women in the SUN cohort.
Total fat MUFA PUFA SFA
Variable
T1 T3 T1 T3 T1 T3 T1 T3
Mean intake (% of energy) 30.0 44.2 11.9 20.6 3.5 6.7 8.5 15.2
Age (years) 36.4 33.4 48.3 47.1 48.3 47.2 49.3 46.3
(11.3) (9.9) (7.4) (6.8) (7.4) (6.9) (7.5) (6.8)
Body mass index (kg/m2) 22.3 22.1 23.3 23.3 23.4 23.2 23.3 23.2
(3) (3.2) (3.2) (3.3) (3.1) (3.4) (3.2) (3.2)
Physical activity (METs- 22.5 18.8 20.5 18.9 20.0 19.5 17.0 13.1
h/week) (20.6) (18) (17.7) (16.6) (17.3) (17.4) (19.9) (17.8)
Height (cm) 164 164 162 162 162 162 162 162
(6) (6) (6) (6) (6) (6) (6) (6)
Years of education 4.9 4.8 5.0 5.0 5 5 5.0 5.0
(1.4) (1.3) (1.6) (1.5) (1.55) (1.52) (1.5) (1.6)
Diabetes (%) 1.1 1.0 2.0 1.7 2.0 2.1 2.2 1.5
Smoking (%)
Never smoker 53.4 48.0 43.5 35.0 40.1 39.7 41.0 39.8
Current smoker 19.4 26.0 18.8 21.8 19.7 21.1 18.6 22.5
Former smoker 26.5 24.9 37.0 42.7 39.6 38.7 39.6 37.2
Energy intake (kcal/day) 2203 2305 2206 2250 2157 2311 2203 2280
(576) (579) (591) (594) (581) (592) (604) (581)
Protein intake (% of energy) 18.6 18.2 18.9 18.1 19.2 17.7 18.2 19.1
(3.4) (3.5) (3.6) (3.8) (3.6) (3.6) (3.7) (3.8)
Carbohydrate intake (% of 50.0 36.5 50.1 37.6 48 40.4 49.5 38.6
energy) (5.4) (5.2) (6.4) (6.3) (7.4) (7.3) (7) (6.7)
Family history of breast
cancer (%)1
No 90.3 89.1 89.3 88.3 89.3 88.7 89.9 88.9
Yes 9.7 10.9 10.7 11.7 10.7 11.3 10.1 11.1
Age of menarche (%)
<10 years 1.3 1.0 1.2 1.8 1.5 1.6 1.5 1.4
10-11 years 20.1 18.3 17.6 19.3 19.4 17.6 19.2 19.0
12-13 years 54.4 55.2 56.0 55.1 56.2 54.7 55.6 55.0
14-16 years 21.5 23.1 22.0 22.0 21.0 23.5 21.3 22.4
≥ 16 years 2.8 2.5 3.2 1.9 1.9 2.6 2.4 2.2
114
Table 21 (continuation). Baseline characteristic for the entire cohort at enrolment and by
tertiles of total fat intake as a percentage of energy among 10,812 women in the SUN cohort.
Pregnancy before 30 years 19.5 19.0 37.3 40.7 37.4 41.7 36.2 41.8
(%)
Number of pregnancies of 0.7 0.7 1.4 1.4 1.4 1.4 1.3 1.5
more than 6 months (1.2) (1.1) (1.5) (1.4) (1.4) (1.5) (1.4) (1.5)
Breast-feeding (months) 2.4 2.2 4.5 4.7 4.3 4.7 4.1 5.0
(5.0) (4.9) (6.8) (6.6) (6.5) (7.1) (6.2) (7.4)
Premenopausal (%) 85.7 92.2 60.5 65.8 60.6 66.3 55.3 70.6
Postmenopausal (%) 14.3 7.8 39.5 34.2 39.4 34.7 44.7 29.4
Time of hormone 1.4 1.1 1.4 1.1 1.6 1.1 1.3 1.3
replacement therapy (years)2 (2.5) (2.2) (2.5) (2.2) (2.6) (2.2) (2.4) (2.3)
Adherence to the 4.1 2.8 4.2 3.4 3.9 3.5 4.6 2.9
Mediterranean diet (1.4) (1.4) (1.4) (1.5) (1.5) (1.5) (1.3) (1.3)
Values are expressed as mean (SD), unless otherwise stated, MET, metabolic equivalents, T, tertile.
1
Information from mother, sisters, and both grandmothers was collected. 2Only for postmenopausal women.
Total fat intake and breast cancer incidence

Table 22 shows the multivariable-adjusted HR and the 95% CI for breast cancer
risk in association with total fat intake using the density energy-adjustment among all
women and stratified by menopausal status. When total fat intake was considered as a
categorical variable, the multivariate-adjusted model revealed a non-significant
protective association between total fat intake and breast cancer risk in the overall
sample (HR T3vs.T1 0.67; 95% CI 0.30, 1.46) and after stratifying by menopausal status,
premenopausal women (HR T3 VS. T1 0.52; 95% CI 0.18, 1.52) and postmenopausal
women (HR T3vs.T1 0.55; 95% CI 0.14, 2.16). When we considered fat intake as a
continuous variable, the HR for increasing fat intake were not statistically significant and
were close to the null value for the overall sample of women and for premenopausal
women. Nonetheless, in the model adjusted for carbohydrate intake, we found a
marginal significant protective association (HR T3vs.T1 0.89; 95% CI 0.79, 1.00)
representing the theoretical substitution for 1% of energy from total fat intake for protein
intake among postmenopausal women.
Subtypes of fat intake and breast cancer incidence

In this population, the major fat subtype was MUFA intake for overall women,
premenopausal and postmenopausal women. In general, the association between
subtypes of fat and breast cancer risk remained non-significant (Table 23). Even though
our point estimates also suggest an inverse association between MUFA and breast
cancer in the overall sample (HR T3vs.T1 0.87; 95% CI 0.51, 1.47), premenopausal (HR
T3vs.T1 0.66; 95% CI 0.33, 1.30) and postmenopausal women (HR T3vs.T1 0.58; 95% CI
0.21, 1.55), we could not confirm these results probably due to the small number of
breast cancer cases reported in our young cohort and also as a result of the overall
healthy eating habits of our cohort’s participants. Contrarily, PUFA seemed to exert a
115
negative effect on breast cancer, particularly, in the overall sample (HR T3 VS. T1 1.28; 95%
CI 0.79, 2.10) and in postmenopausal women (HR T3 VS. T1 1.53; 95% CI 0.64, 3.63),
although results remained non-significant.
Table 22. Energy- and multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals
(CIs) for breast cancer risk in association with total fat intake using the density energy adjustment
among women in the SUN Cohort.
Total fat intake
Tertiles of total fat intake per 1% increase
as a continuous
Overall women 1 2 3 P-trend variable
Median value (% energy) 30.8 37.1 43.2
Median caloric intake 2208 2313 2328
(kcal/day)
Cases/person-years 40/ 37744 29/ 38870 32/39188
Energy-adjusted HR (95% CI) 1.00 0.77 (0.48, 1.26) 0.90 (0.56, 1.44) 0.629 0.99 (0.96, 1.02)
Multivariable model 1 1.00 0.69 (0.42, 1.14) 0.79 (0.47, 1.31) 0.346 0.99 (0.95, 1.02)
Premenopausal women §
(kcal/day)
Cases/person-years 22/ 27912 18/31454 17/ 32470
Postmenopausal women †
(kcal/day)
Cases/person-years 16/8586 9/6157 9/5565
Multivariable model 1: energy-adjusted bivariate Cox regression models adjusted for non-alcohol energy (continuous), using
age and time since recruitment as the underlying time metric. Multivariable Cox regression models adjusted for alcohol and
non-alcohol energy intakes (both continuous), smoking (never smoker, former smoker, current smoker),height (continuous),
family history of breast cancer (yes/no), physical activity (METs-h/week, continuous), BMI (three categories, 25 to <30, ≥30,
versus <25 kg/m2), age of menarche (4 categories), age of menopause (3 categories), number of pregnancies of more than 6
months (continuous), use of hormone replacement therapy (dichotomous) and its duration (continuous), years at university
(continuous), adherence to the Mediterranean diet (Trichopoulou score)for overall women.
§For premenopausal women, we excluded from the adjustment use of hormone replacement therapy (dichotomous) and its
duration (continuous). † For postmenopausal women, years since recruitment until the beginning of the time at risk and use
of hormone replacement therapy (dichotomous) and its duration (continuous). Multivariable model 2: further adjusted for
protein intake (continuous) *Multivariable model 3: further adjusted for carbohydrate intake (continuous).
116
Table 23. Energy- and multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals
(CIs) for breast cancer risk in association with subtypes of fat intake using the density energy
adjustment among women in the SUN Cohort.
Overall women
Tertiles of fat subtype Tertile 1 Tertile 2 Tertile 3 P-trend
Monounsaturated fat
Energy-adjusted HR (95% CI) 1.00 1.35 (0.84, 2.16) 0.92 (0.55, 1.53) 0.693
Multivariable-adjusted HR (95% CI) 1.00 1.30 (0.81, 2.10) 0.87 (0.51, 1.47) 0.517
Saturated fat
Cases/person-years 38/ 37459 33/ 38581 30/ 39762
Polyunsaturated fat
Premenopausal women
p trend
Tertile 1 Tertile 2 Tertile 3
Monounsaturated fat
Saturated fat
Cases/person-years 19/ 28919 21/ 30641 17/ 32276
Polyunsaturated fat
p trend
Tertile 1 Tertile 2 Tertile 3
Monounsaturated fat
Saturated fat
Cases/person-years 12/ 7277 11/ 6653 11/6377
117
Table 23 (continuation). Energy- and multivariable-adjusted hazard ratios (HRs) and 95%
confidence intervals (CIs) for breast cancer risk in association with subtypes of fat intake
using the density energy adjustment among women in the SUN Cohort.
Polyunsaturated fat
The results are shown for the density model after fat and energy were log-transformed. Energy-adjusted bivariate Cox
regression models adjusted for nonalcohol energy (continuous), using age and time since recruitment as the underlying
time metric. Multivariable Cox regression models adjusted for alcohol and nonalcohol energy intakes (both continuous),
smoking (never smoker, former smoker, current smoker),height (continuous), family history of breast cancer (yes/no),
physical activity (METs-h/week, continuous), BMI (three categories, 25 to <30, ≥30, versus <25 kg/m 2), age of
menarche (4 categories), age of menopause (3 categories), number of pregnancies of more than 6 months
(continuous), use of hormone replacement therapy (dichotomous) and its duration (continuous), years at university
(continuous), adherence to the Mediterranean diet (Trichopoulou score) for overall women.
*For premenopausal women, we excluded from the adjustment use of hormone replacement therapy (dichotomous)
and its duration (continuous). ** For postmenopausal women, years since recruitment until the beginning of the time
at risk and use of hormone replacement therapy (dichotomous) and its duration (continuous).
Isocaloric substitutions
The associations of various isocaloric dietary substitutions of one dietary component
for another (5% of energy) on the risk of breast cancer showed a suggested increased
risk when increasing the amount of proteins while decreasing the amount of total fat (HR
1.26; 95% CI 0.92, 1.73). Replacing carbohydrates for either total fat or MUFA also
seemed to increase the risk of breast cancer in the overall sample of women and by
menopausal status. Furthermore, the only apparent protective association was found
when replacing polyunsaturated fats in favor of either carbohydrates or proteins (Table
24).
Table 24. Estimated HRs (95% CIs) of breast cancer risk associated with isocaloric substitutions (5%
energy increment) of one dietary component for another.
Premenopausal Postmenopausal
Overall women
women* women**
Increase Decrease
Proteins Fat 1.26 (0.92, 1.73) 1.33 (0.86, 2.06) 1.57 (0.93, 2.62)
Carbohydrates Fat 1.08 (0.92, 1.28) 1.11 (0.89, 1.39) 1.19 (0.90, 1.57)
Carbohydrates MUFA 1.12 (0.80, 1.56) 1.23 (0.76, 1.99) 1.36 (0.73, 2.50)
Proteins MUFA 1.36 (0.91, 2.03) 1.45 (0.81, 2.60) 1.88 (0.95, 3.77)
Carbohydrates PUFA 0.71 (0.36, 1.39) 0.95 (0.35, 2.53) 0.49 (0.16, 1.43)
Proteins PUFA 0.87 (0.44, 1.71) 1.12 (0.41, 3.05) 0.67 (0.23, 1.98)
Carbohydrates SFA 1.20 (0.76, 1.90) 1.00 (0.55, 1.85) 1.35 (0.63, 2.91)
Proteins SFA 1.46 (0.79, 2.70) 1.19 (0.53, 2.66) 1.88 (0.67, 5.28)
Energy-adjusted bivariate Cox regression models adjusted for non-alcohol energy (continuous), using age and time since recruitment
as the underlying time metric. Multivariable Cox regression models adjusted for alcohol and nonalcohol energy intakes (both
continuous), smoking (never smoker, former smoker, current smoker),height (continuous), physical activity (METs-h/week,
continuous), BMI (three categories, 25 to <30, ≥30, versus <25 kg/m2), age of menarche (4 categories), age of menopause (3
categories), number of pregnancies of more than 6 months (continuous), use of hormone replacement therapy (dichotomous) and its
duration (continuous), years at university (continuous), adherence to the Mediterranean diet (Trichopoulou score)for overall women.
*For premenopausal women, we excluded from the adjustment use of hormone replacement therapy (dichotomous) and its duration
(continuous). ** For postmenopausal women, years since recruitment until the beginning of the time at risk and use of hormone
replacement therapy (dichotomous) and its duration (continuous).
118
5. Carbohydrate quality index and breast cancer risk in the SUN project
Baseline characteristics of the study participants according to quartiles of a CQI are

presented in Table 25. Higher CQI was observed in older, postmenopausal, and more
physically active women, who have breastfed their offspring for a longer period. They
were also more likely to be former smokers and have a higher prevalence of diabetes.
Moreover, these women were more prone to follow a special diet and had better
adherence to the Mediterranean diet and higher energy intake. As expected, we
observed a higher total energy intake from carbohydrates, solid carbohydrates, fiber, and
carbohydrates from whole grains among women with better CQI. Contrarily, CQI was
inversely associated with GI and consumption of liquid carbohydrates or carbohydrates
from refined grains.
Table 25. Baseline characteristics of participants according to quartiles of the carbohydrate

quality index: the SUN Project, 1999-2018.
Variable Quartile 1 Quartile 2 Quartile 3 Quartile 4
N 3,296 2,371 3,089 2,056
Index range (points) 4-9 10-11 12-14 15-20
Age (years) 32 (10) 34 (10) 35 (11) 37 (11)
Body mass index (kg/m2) 22.0 (3.0) 22.2 (3.0) 22.4 (3.1) 22.4 (3.0)
Physical activity (METs-h/week) 17.5 (16.4) 19.0 (16.4) 21.1 (19.0) 24.6 (23.2)
Alcohol intake (g/d) 4.0 (5.8) 4.1 (6.1) 4.0 (6.0) 4.0 (5.7)
Years at university 4.8 (1.3) 4.8 (1.3) 4.9 (1.4) 4.9 (1.4)
Height (cm) 164 (6) 164 (6) 164 (6) 164 (6)
Pregnancy before 30 years (%) 20.5 23.5 24.4 25.5
Number of pregnancies of > 6 months 0.6 (1.1) 0.7 (1.2) 0.71 (1.1) 0.7 (1.2)
Breast-feeding (months) 2.3 (4.9) 2.8 (5.3) 2.8 (5.1) 3.1 (5.6)
Hormone replacement therapy 1 39.4 42.7 38.5 34.4
Time of hormone replacement therapy 1.4 (2.5) 1.1 (2.2) 1.4 (2.5) 1.2 (2.2)
(years) 1
Diabetes (%) 0.7 1.0 1.1 1.9
Smoking (%)
Never smoker 49.4 51.2 52.1 52.1
Current smoker 27.3 24.2 20.8 16.9
Former smoker 22.7 23.8 26.2 30.1
Family history of breast cancer (%) 2 10.7 9.9 10.6 11.2
Age at menarche (%)
≤ 11 years 18.8 20.4 20.5 21.8
12-13 years 55.2 54.2 55.2 54.1
14 years 26.0 25.4 24.3 24.1
Premenopausal (%) 92.4 89.9 87.2 83.8
Postmenopausal (%) 7.6 10.1 12.8 16.2
Age at menopause 1
Postmenopausal < 50 years (%) 57.0 54.0 48.9 53.6
Postmenopausal 50 years (%) 43.0 46.0 51.1 46.4
Adherence to the Mediterranean diet 2.5 (1.4) 3.3 (1.5) 3.8 (1.4) 4.6 (1.4)
Special diet 5.5 6.5 9.4 13.3
Total energy intake (kcal/day) 2152 (564) 2326 (573) 2352 (567) 2422 (548)
Glycemic index 53.0 (4.6) 52.0 (4.4) 51.1 (4.4) 50.0 (4.3)
Carbohydrate intake (% of total energy) 42.2 43.0 43.7 44.8
Solid carbohydrate intake (g/d) 192.0 (66.7) 219.0 (71.4) 227.4 (74.0) 245.1 (71.2)
Liquid carbohydrate intake (g/d) 36.1 (19.6) 32.2 (20.3) 30.9 (20.1) 25.7 (16.6)
Fiber intake (g/d) 19.4 (6.1) 26.3 (7.7) 31.8 (10.5) 41.6 (12.8)
Carbohydrates from whole grains (g/d) 0.4 (3.0) 2.6 (7.8) 8.9 (15.1) 21.8 (23.6)
Carbohydrates from refined grains i (g/d) 86.4 (46.5) 92.1 (51.2) 82.6 (48.7) 67.4 (39.2)
Values are expressed as mean (SD) unless otherwise stated. 1 Only for postmenopausal women. 2 Information from
mother, sisters, and both grandmothers was collected.
119
During a median follow-up of 11.8 years and 115,802 person-years, we observed
101 confirmed cases of breast cancer. In the fully adjusted model, participants in the
highest quartile of the CQI showed a significant lower risk of confirmed breast cancer
during follow-up than those in the lowest quartile (HR Q4vs.Q1 0.39, 95% CI 0.17, 0.87)
(Table 26). As a sensitivity analysis, we included probable incident cases (self-reported
at baseline). After 11.5 years of follow-up and 115,368 person-years, 190 probable
incident cases were reported. In the multivariable adjusted model, the highest quartile of
carbohydrate quality compared to the lowest quartile was associated with a relative risk
reduction of 48% (HR Q4vs.Q1 0.52, 95% CI 0.31, 0.89).
Table 26. Hazard ratios and 95% confidence intervals for confirmed and probable incident breast
cancer by quartiles of carbohydrate quality index a in the SUN Project, 1999-2018.
Quartiles of CQI
P-trend
Quartile 1 Quartile 2 Quartile 3 Quartile 4
Confirmed cases
N 3296 2371 3089 2056
CQI range 4-9 10-11 12-14 15-20
Incident cases 29 31 32 9
Person-years 36,674 25,770 32,618 20,740
Age-adjusted 1.00 1.37 (0.82, 2.27) 1.07 (0.64, 1.77) 0.43 (0.20, 0.92) 0.068
Model 1 1.00 1.36 (0.82, 2.26) 1.09 (0.65, 1.81) 0.43 (0.20, 0.92) 0.075
Model 2 1.00 1.30 (0.77, 2.19) 1.02 (0.59, 1.75) 0.39 (0.17, 0.87) 0.049
Confirmed + probable cases
N 3296 2371 3089 2056
CQI range 4-9 10-11 12-14 15-20
Person-years 36,546 25,679 32,463 20,679
Age-adjusted 1.00 1.02 (0.69, 1.50) 0.99 (0.68, 1.42) 0.58 (0.36, 0.93) 0.057
Model 1 1.00 1.02 (0.69, 1.49) 0.99 (0.69, 1.43) 0.58 (0.36, 0.94) 0.062
Model 2 1.00 0.97 (0.65, 1.44) 0.93 (0.63, 1.37) 0.52 (0.31, 0.89) 0.034
Model 1 additionally adjusted for height (continuous), years at university (continuous), family history of breast cancer
(yes/no), smoking status (never smoker, former smoker, current smoker), physical activity (METs-h/ week, categories),
alcohol intake (continuous), BMI (continuous), age at menarche (≤ 11 years , 12–13 years, ≥14 years), number of
pregnancies of more than 6 months (continuous), pregnancy before the age of 30 years (yes/no), months of breastfeeding
(continuous), use of hormonal therapy (yes/no) and its duration (continuous), menopause (no menopause, < 50 years,≥
50 years). Model 2 additionally adjusted for energy intake (kcal/day, continuous), adherence to the Mediterranean diet
(continuous) and special diet (yes/no).
Moreover, when we assessed individual CQI components, we observed that the

highest vs. the lowest ratio of whole-grain: total-grain carbohydrates were associated
with a relative risk reduction of 44% for total breast cancer incidence (Table 27).
120
Table 27. Hazard ratios and 95% confidence intervals for confirmed incident breast cancer by
tertiles of individual components of CQI and glycemic load in the SUN Project, 1999-2018.
Tertile 1 Tertile 2 Tertile 3 P-trend
Tertiles of glycemic load
N 3604 3604 3604
Incident cases 30 33 38
Person-years 37,774 38,698 39,330
Age-adjusted 1.00 1.11 (0.67, 1.82) 1.22 (0.76, 1.97) 0.416
Model 1 1.00 1.12 (0.68, 1.84) 1.24 (0.77, 2.01) 0.381
Model 2 1.00 1.23 (0.69, 2.17) 1.45 (0.70, 3.01) 0.327
Tertiles of glycemic index
N 3604 3604 3604
Person-years 37,672 38,807 39,322
Age-adjusted 1.00 1.16 (0.70, 1.93) 1.46 (0.91, 2.36) 0.115
Model 1 1.00 1.16 (0.70, 1.93) 1.49 (0.92, 2.41) 0.103
Model 2 1.00 1.19 (0.71, 2.01) 1.53 (0.92, 2.54) 0.097
Tertiles of solid carbohydrates:(solid carbohydrates + liquid carbohydrates) ratio
N 3604 3604 3604
Person-years 39,107 38,700 37,995
Age-adjusted 1.00 1.08 (0.65, 1.79) 1.34 (0.83, 2.15) 0.239
Model 1 1.00 1.07 (0.64, 1.78) 1.33 (0.82, 2.15) 0.259
Model 2 1.00 1.07 (0.64, 1.79) 1.33 (0.82, 2.17) 0.253
Tertiles of whole-grain: total-grain carbohydrates ratio
N 6322 886 3604
Person-years 69,336 9,480 36,985
Age-adjusted 1.00 0.66 (0.29, 1.52) 0.57 (0.36, 0.92) 0.026
Model 1 1.00 0.68 (0.29, 1.57) 0.57 (0.35, 0.91) 0.024
Model 2 1.00 0.67 (0.29, 1.55) 0.56 (0.34, 0.90) 0.022
Tertiles of dietary fiber intake (g/d)
N 3604 3604 3604
Person-years 38,941 39,243 37,618
Age-adjusted 1.00 0.80 (0.48, 1.32) 0.98 (0.60, 1.57) 0.942
Model 1 1.00 0.81 (0.49, 1.35) 1.00 (0.62, 1.62) 0.869
Model 2 1.00 0.81 (0.46, 1.42) 0.98 (0.50, 1.93) 0.872
Model 1 additionally adjusted for height (continuous), years at university (continuous), family history of breast cancer
(yes/no), smoking status (never smoker, former smoker, current smoker), physical activity (METs-h/ week, categories),
alcohol intake (continuous), BMI (continuous), age at menarche (≤ 11 years , 12–13 years, ≥14 years), number of
pregnancies of more than 6 months (continuous), pregnancy before the age of 30 years (yes/no), months of
breastfeeding (continuous), use of hormonal therapy (yes/no) and its duration (continuous), menopause(no
menopause, < 50 years, ≥ 50 years). Model 2 additionally adjusted for energy intake (kcal/day, continuous), adherence
to the Mediterranean diet (continuous) and special diet (yes/no).
Moreover, we stratified our analyses by menopausal status (Figure 16). When

we compared the highest quartile vs. the lowest quartile of CQI, we found a statistically
significant inverse association between CQI and breast cancer risk among
premenopausal women both in the analyses for confirmed breast cancer (HR T3vs T1 0.32,
95% CI 0.10, 1.00) and when we also added the probable cases of breast cancer
121
(HR T3vs.T1 0.46, 95% CI 0.22, 0.96). We also performed sensitivity analyses in order to
assess the robustness of our results, but they did not substantially change in any of these
scenarios. Particularly, we excluded participants with baseline diabetes, changed
implausible energy limits to percentiles 1 and 99 and included either confirmed or
probable cases (Figure 17).
Figure 16. Hazard ratios and 95% confidence intervals for confirmed (a) and probable
(b) incident premenopausal and postmenopausal breast cancer by quartiles of CQI.
Adjusted for height (continuous), years at university (continuous), family history of breast cancer (yes/no),
smoking status (never smoker, former smoker, current smoker), physical activity (METs-h/ week, categories),
alcohol intake (continuous), BMI (continuous), age at menarche (≤ 11 years , 12–13 years, ≥14 years), number
of pregnancies of more than 6 months (continuous), pregnancy before the age of 30 years (yes/no), months of
breastfeeding (continuous), energy intake (kcal/day, continuous), adherence to the Mediterranean diet
(continuous) and special diet (yes/no). For postmenopausal women we further adjusted for use of hormonal
therapy (yes/no), its duration (continuous), menopause (no menopause, <50 years, ≥ 50 years) and time since
recruitment.
122
Figure 17. Sensitivity analyses. Hazard ratios (HRs) and 95% confidence intervals (CIs) of
incident breast cancer for the fourth quartile of CQI compared with the first quartile.
Adjusted for height, years at university, family history of BC, smoking status, physical activity, alcohol intake, BMI, total
energy intake, age of menarche, menopause, number of pregnancies of more than 6 months, pregnancy before the age
of 30 years, months of breastfeeding, use of hormone replacement therapy and its duration, adherence to the
Mediterranean diet and special diet.
123
6. Healthful and unhealthful provegetarian food patterns and breast cancer risk
in the SUN project
Between 1999 and 2017 (median follow-up time: 11.5 years), we identified 190
probable breast cancer cases, of which 101 were confirmed by review of the medical
records. The remaining cases were self-reported by highly educated women, and
confirmation is pending of the retrieval of the medical records. Only the 101 confirmed
cases were used for the main analyses. At baseline, the PVG ranged from 22 to 43 in
the lowest tertile and 48 to 66 in the highest tertile.
The baseline characteristics of the participants according to tertiles of a PVG, a

healthful PVG, and an unhealthful PVG food patterns are described in Table 28.
Participants with a higher PVG and hPVG were more likely to be older, more physically
active, have a lower energy intake, and lower fat and protein (% of energy) intake.
Furthermore, these participants also had a higher consumption (g/d) of vegetables, fruits,
legumes, cereals, nuts, vegetable fats, and coffee, and have a lower consumption of
meat and meat products, animal fats, eggs, fish and other seafood, and dairy products.
Nevertheless, participants with the highest adherence to a PVG were more likely to eat
more pastries and SSBs compared with those in the lowest adherence group. On the
other hand, participants with the highest adherence to an uPVG were more likely to be
younger, less physically active, and have a higher energy intake.
Table 28. Baseline characteristics of participants according to tertiles of the provegetarian

food pattern: the SUN cohort, 1999-2018.
Provegetarian food Healthful Unhealthful
pattern provegetarian provegetarian
Variable T1 T3 T1 T3 T1 T3
N 4162 3441 4041 3178 4089 3359
Median (range) 22-43 48-66 30-51 58-80 30-51 58-82
Age (years) 33.7 35.9 32.1 37.9 37.3 32.0
(10.2) (11) (9.2) (11.5) (11.2) (9.3)
Body mass index (kg/m2) 22.2 22.2 22.0 22.4 22.6 21.8
(3) (3.1) (3.0) (3.0) (3.1) (3.0)
Physical activity (METs- 19.9 20.6 17.7 23.4 22.8 17.5
h/week) (19) (19.2) (16.6) (21.3) (20.4) (16.1)
Alcohol intake (g/d) 3.97 4.06 3.6 4.6 4.1 4.0
(6.22) (5.55) (5.3) (6.6) (5.8) (6.0)
Years at university 4.79 4.9 4.8 4.9 4.8 1.8
(1.28) (1.38) (1.3) (1.4) (1.3) (1.3)
Height (cm) 164 163 164.0 163.3 163.3 164.0
(6) (6) (6.0) (6.1) (6.0) (6.0)
Pregnancy before 30 years 0.4 0.4 0.3 0.5 0.5 0.3
(%) (0.8) (0.8) (0.8) (0.9) (0.9) (0.7)
Number of pregnancies of 0.7 0.7 0.6 0.8 0.8 0.5
more than 6 months (1.1) (1.1) (1.1) (1.2) (1.2) (1.1)
124
Table 28 (continuation). Baseline characteristics of participants according to tertiles of the
provegetarian food pattern: the SUN cohort, 1999-2018.
Breast-feeding (months) 2.7 2.7 2.3 3.2 3.3 2.2
(5.2) (5.1) (5.1) (5.4) (5.5) (5.0)
Hormone replacement
therapy 1
Yes 40.3 34.8 40.3 33.7 37.8 33.9
Time of hormone 1.4 1.2 1.4 1.2 1.3 1.3
replacement therapy (2.5) (2.4) (2.5) (2.3) (2.3) (2.5)
(years) 1
Diabetes (%) 0.94 1.05 0.7 1.5 1.7 0.7
Smoking (%)
Never smoker 53.9 48.8 54.5 48.5 49.0 55.2
Current smoker 22.2 24.2 23.9 20.8 21.1 24.2
Former smoker 23.9 26.9 21.5 30.7 30.0 20.6
cancer (%) 2
Yes 10.8 10.8 10.4 11.2 10.8 10.8
Age at menarche (%)
≤ 11 years 20.4 20.4 19.1 21.7 21.9 17.9
12-13 years 53.7 55.6 54.5 55.2 54.3 55.8
13-14 years 17.2 16.5 17.6 15.4 16.5 17.8
16 years 8.7 7.4 8.9 7.7 7.3 8.6
Premenopausal (%) 90.6 86.0 94.0 82.0 83.8 93.5
Postmenopausal (%) 9.4 14.0 6.0 18.0 16.2 6.5
Age at menopause 1
Postmenopausal < 50 52.6 52.9 52.7 51.7 51.5 51.4
years (%)
Postmenopausal 50 47.5 47.1 47.3 48.3 48.5 48.6
years (%)
Adherence to the 3.5 4.9 3.1 5.4 3.1 5.4
Mediterranean diet (1.6) (1.7) (1.5) (1.5) (1.5) (1.5)
Energy intake (kcal/d) 2396 2214 2299 2389 2193 2481
(539) (611) (571) (554) (541) (570)
Carbohydrate (% of 40.7 46.1 42.1 45.0 42.2 44.8
energy) (7.3) (6.7) (6.9) (7.7) (7.3) (7.1)
Protein (% of energy) 19.6 17.0 18.6 18.0 20.0 16.5
(3.4) (3.0) (3.3) (3.2) (3.2) (2.9)
Total fat (% of energy) 38.6 35.6 38.3 35.6 36.5 37.6
(6.6) (6.6) (5.9) (7.1) (6.7) (6.6)
Consumption of:
Vegetables (g/d) 503 637 417 756 718 401
(326) (327) (231) (416) (345) (271)
Fruit (g/d) 392 532 316 644 569 335
(301) (331) (219) (383) (312) (300)
Legumes (g/d) 19 25 18 26 25 18
(17) (17) (12) (21) (17) (15)
Cereals (g/d) 78 102 88 91 82 98
(60) (53) (55) (60) (49) (67)
Potatoes (g/d) 44 57 52 47 46 56
(42) (39) (39) (41) (36) (47)
Nuts (g/d) 5 9 4 11 9 5
(10) (12) (6) (16) (13) (9)
Vegetable fats (g/d) 19 25 19 27 24 19
(15) (15) (13) (17) (14) (16)
Pastries (g/d) 45 52 58 36 33 67
(42) (36) (39) (35) (24) (49)
Sugar-sweetened 50 66 81 31 32 90
beverages (g/d) (95) (94) (109) (73) (60) (12)
Coffee (g/d) 48 74 49 71 71 47
(58) (64) (57) (64) (63) (57)
125
Table 28 (continuation). Baseline characteristics of participants according to tertiles of the
provegetarian food pattern: the SUN cohort, 1999-2018.
Meats /meat products (g/d) 198 140 189 146 481 393
(74) (57) (66) (70) (248) (238)
Animal fats for cooking or 1 1 1 0.6 1 1
as a spread (g/d) (3) (2) (3) (2.1) (2) (3)
Eggs (g/d) 26 18 25 19 24 20
(15) (11) (15) (13) (13) (13)
Fish and other seafood 107 86 92 103 119 72
(g/d) (64) (54) (54) (69) (60) (47)
Dairy products (g/d) 521 345 463 407 174 164
(268) (194) (233) (255) (68) (72)
Values are expressed as mean (SD) unless otherwise stated. T (tertile).
1 Only for postmenopausal women. 2 Information from mother, sisters, and both grandmothers was
collected.
With more than 115,802 person-years of follow-up, 101 women developed a

confirmed case of breast cancer. In the age-adjusted model, a higher overall PVG was
not associated with the risk of breast cancer (HR T3vs.T1 1.03; 95% CI, 0.67, 1.59; P-
trend=0.857) (Table 29). The results remained unchanged after further adjustment for
potential confounders (HR T3vs.T11.03; 95% CI, 0.67, 1.60; P-trend=0.875). Nonetheless,
a moderate adherence (range, 44–47) to a PVG compared with the lowest adherence
was associated with a 45% relative lower risk of breast cancer (HRT2 vs T1 0.55; 95% CI,
0.32, 0.95).
Table 29. Hazard Ratio (HR) and 95 confidence intervals 95% (CI 95%) according to tertiles
of the provegetarian (PVG) food pattern and a healthful and unhealthful provegetarian food
patterns in the overall sample of women from the SUN cohort.
Overall women Tertile 1 Tertile 2 Tertile 3 P-trend
Provegetarian food pattern
n 4162 3209 3441
Median (range) 22-43 44-47 48-66
Age-adjusted 1.00 0.55 (0.32, 0.95) 1.03 (0.67, 1.59) 0.857
Healthful provegetarian food pattern
n 4041 3593 3178
Median (range) 30-51 52-57 58-80
Age-adjusted 1.00 0.93 (0.58, 1.48) 0.80 (0.49, 1.32) 0.382
Unhealthful provegetarian food pattern
n 4089 3365 3359
Median (range) 30-51 52-57 58-82
126
Table 29 (continuation). Hazard Ratio (HR) and 95 confidence intervals 95% (CI 95%)
according to tertiles of the provegetarian (PVG) food pattern and a healthful and unhealthful
provegetarian food patterns in the overall sample of women from the SUN cohort.
Age-adjusted 1.00 1.44 (0.89, 2.30) 1.41 (0.86, 2.31) 0.161
Multivariable adjusted for height, family history of breast cancer (yes or no), smoking status (never smoker, former
smoker, current smoker), physical activity (METs-h/ week, continuous), alcohol intake (g/day, continuous), BMI
(tertiles), age at menarche (≤11 years, 12-13 years, 13-14 years, ≥15 years), menopause (no menopause, < 50
years, ≥ 50 years), number of pregnancies of more than 6 months (continuous), pregnancy before the age of 30
years (yes/no), months of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its duration
(continuous), years at university (continuous), total energy intake (kcal/day, continuous).
The p value when we assigned the median value to each quartile and entered this as a continuous variable in the
model.
Furthermore, to account for a nonlinear association, we performed a restricted

cubic spline analysis adjusting for the same potential confounding factors included in the
Cox regression analyses. We found a significant nonlinear relationship between
adherence to a PVG and the incidence of breast cancer (P for nonlinearity<0.001).
Particularly, we observed a U-shaped association with increased breast cancer risk in
the lower and upper ranges of the PVG. We also repeated the multivariable analyses
using the original PVG, and the results were similar (HR T3vs.T1 1.02; 95% CI 0.65,
1.60; P-trend=0.897 and HR T2vs.T1 0.66; 95% CI 0.40, 1.11) (Figure 18).
Figure 18. Restricted cubic splines for the hazard ratio (HR) and 95% confidence intervals (CI)
of adherence to the Provegetarian (PVG) food pattern and incidence of breast cancer in the
SUN cohort.
Black line represents the HR, and the pointed lines represent the 95% confidence intervals. A PVG value of 40 was
the reference value (median of tertile 1). Adjusted for height, family history of breast cancer (yes or no), smoking status
(never smoker, former smoker, current smoker), physical activity (METs-h/ week, continuous), alcohol intake (g/day,
continuous), BMI (tertiles), age at menarche (≤11 years, 12-13 years, 13-14 years, ≥15 years), menopause (no
menopause, <50 years, ≥50 years), number of pregnancies of more than 6 months (continuous), pregnancy before the
age of 30 years (yes/no), months of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its
duration (continuous), years at university (continuous), total energy intake (kcal/day, continuous). P for
nonlinearity=0.0003
When we analyzed hPVG and uPVG separately (Figure 19), we found an

inverse, non-significant suggestive association between hPVG and breast cancer
incidence in the age-adjusted model (HR T3vs.T1 0.80; 95% CI 0.49, 1.32; P-trend=0.382)
127
and the multivariable adjusted model (HR T3vs.T1 0.83; 95% CI 0.50, 1.37; P-trend=0.466).
A non-significant suggestive higher risk of breast cancer was also found with a higher
adherence to an uPVG food pattern (HR T3vs.T1 1.31; 95% CI 0.79, 2.19; P-trend=0.282)
in the multivariable adjusted model.
Figure 19. Hazard Ratios (HR) and 95% confidence intervals (CI) of breast cancer incidence
according to a a) healthful and an b) unhealthful provegetarian food pattern in the SUN cohort
Adjusted for height, family history of breast cancer (yes or no), smoking status (never smoker, former smoker, current
smoker), physical activity (METs-h/ week, continuous), alcohol intake (g/day, continuous), BMI (tertiles), age at
menarche (≤11 years, 12-13 years, 13-14 years, ≥15 years), menopause (no menopause, < 50 years, ≥ 50 years),
number of pregnancies of more than 6 months (continuous), pregnancy before the age of 30 years (yes/no), months
of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its duration (continuous), years at
university (continuous), total energy intake (kcal/day, continuous).
When we compared regression coefficients for the comparison of the second

versus first and third versus first tertiles across the models of uPVG and hPVG, we
found P values of 0.279 and 0.225, respectively. When we stratified our results by
128
menopausal status, the results were non-significant after adjusting for multiple
confounders, both for premenopausal and postmenopausal women. Nonetheless, a
moderate adherence to a PVG compared with the lowest adherence seemed to be
inversely associated with breast cancer in premenopausal women (HR T2vs.T1 0.40; 95%
CI 0.19, 0.86).
Among postmenopausal women, a moderate adherence to an uPVG was

associated with a significant increased risk of breast cancer compared with the reference
category (HR T2vs.T1 2.57; 95% CI 1.03, 6.43). When adherence to an hPVG was
evaluated, inverse, although non-significant, associations with breast cancer risk were
observed among premenopausal (HR T3vs.T1 0.79; 95% CI 0.40, 1.57; P-trend=0.503) and
postmenopausal (HR T3vs.T1 0.71; 95% CI 0.28, 1.84; P-trend=0.525) women.
Moreover, we found evidence of deviation from linearity for a PVG food pattern
in both premenopausal (Figure 20a) and postmenopausal (Figure 20b) breast cancer
incidence with a P value of 0.0167 and 0.0443, respectively.
129
b
Figure 20. Restricted cubic splines for the hazard ratio (HR) and 95% confidence intervals (CI)
of adherence to a provegetarian (PVG) food pattern and incidence of (a) premenopausal and
(b) postmenopausal breast cancer in the SUN cohort.
Black line represents the HR, and the pointed lines represent the 95% confidence intervals. A PVG value of 40 was
the reference value (median of tertile 1). Adjusted for height, family history of breast cancer (yes or no), smoking status
(never smoker, former smoker, current smoker), physical activity (METs-h/ week, continuous), alcohol intake (g/day,
continuous), BMI (tertiles), age at menarche(≤11 years, 12-13 years, 13-14 years, ≥15 years), menopause (no
menopause, < 50 years, ≥ 50 years), number of pregnancies of more than 6 months (continuous), pregnancy before
the age of 30 years (yes/no), months of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and
its duration (continuous), years at university (continuous), total energy intake (kcal/day, continuous). P value for
nonlinearity for premenopausal women= 0.0167 and for postmenopausal women=0.0443.
130
7. Healthful and unhealthful plant-based diets and risk of breast cancer in
the Nurses’ Health Studies
During 4,841,083 person-years of follow-up, 12,482 participants developed invasive

breast cancer (8,220 cases in NHS and 4,262 cases in NHSII). Women with higher PDI
were more likely to have lower BMI, lower weight change from age 18 years, lower
prevalence of diabetes, less alcohol intake, higher physical activity and energy intake,
lower protein intake and higher carbohydrate intake (% of energy) (Table 30).
Table 30. Age and age-standardized baseline characteristics of women according to quintiles of
an overall plant-based diet index (PDI) in the Nurses’ Health Study (NHS) and NHSII.
NHS (1984) NHSII (1991)
Quintile 1 Quintile 3 Quintile 5 Quintile 1 Quintile 3 Quintile 5
(n=14,629) (n=13,832) (n=15,857) (n=20,408) (n=16,541) (n=17,944)
Median PDI 46 54 62 47 55 63
Age, years1 50.2 (7) 50.7 (7.2) 51.7 (7.3) 36.5 (4.8) 36.6 (4.7) 36.9 (4.5)
Body mass index, kg/m2 25.7 (5.1) 25.1 (4.7) 24.5 (4.4) 21.6 (3.6) 21.3 (3.3) 20.9 (3.0)
Body mass index at age 18 years, 21.7 (3.2) 21.4 (3.0) 21.1 (2.8) 25.2 (5.7) 24.5 (5.2) 23.7 ((4.8)
kg/m2
Weight change from age 18 years, kg 10.7 (12.3) 9.8 (11.1) 9.0 (10.5) 9.8 (12.4) 8.7 (11.2) 7.7 (10.4)
Height, inches 64.5 (2.4) 64.5 (2.4) 64.5 (2.4) 64.9 (2.6) 64.9 (2.6) 64.9 (2.6)
Self-reported African heritage (%) 1.5 1.4 1.1 2.0 1.4 1.2
Self-reported history of diabetes (%) 3.8 3.3 2.3 1.2 1.0 0.7
Family history of breast cancer (%) 7.8 8.1 8.1 5.8 6.1 6.3
Personal history of benign breast 29.5 30.2 31.0 9.1 9.5 9.6
disease (%)
Age at menarche <12 years (%) 23.5 21.6 22.5 24.8 24.0 24.8
Oral contraceptives, ever (%) 50.5 49.3 48.3 85.6 84.7 83.4
Parous (%) 92.0 92.9 93.3 69.3 75.9 77.5
Parity, n2 3.1 (1.5) 3.2 (1.5) 3.2 (1.5) 2.1 (0.9) 2.1 (0.9) 2.2. (0.9)
Breastfeeding, ≤6 months (%)2 35.8 36.6 36.1 19.7 16.9 14.1
Postmenopausal (%) 48.4 48.6 48.8 3.1 3.2 3.4
Postmenopausal hormone use, never 52.9 52.7 52.8 6.1 6.9 7.8
(%)3
Physical activity, METs-h/week 11.4 (13.2) 11.9 (13.1) 12.9 (13.4) 17.8 (24.4) 20.6 (26.8) 24.9 (31.5)
Alcohol intake (g/day) 8.6 (13.6) 6.7 (11) 5.8 (9.3) 3.3 (7.1) 3 (5.8) 3.2 (5.6)
Total energy intake (kcal/day) 1458 (451) 1723 (488) 2056 (519) 1478 (464) 1792 (504) 2134 (527)
Saturated fat (% of energy) 13.9 (2.9) 12.5 (2.3) 11.2 (2.1) 12.6 (2.5) 11.1 (2.2) 9.8 (2.1)
Monounsaturated fat (% of energy) 13.4 (2.6) 12.7 (2.3) 12 (2.2) 12.7 (2.6) 11.9 (2.4) 11.2 (2.3)
Polyunsaturated fat (% of energy) 6.5 (1.9) 6.6 (1.7) 6.8 (1.6) 5.7 (1.5) 5.6 (1.4) 5.6 (1.3)
Trans fat (% of energy) 1.8 (0.6) 1.9 (0.6) 1.9 (0.6) 1.7 (0.7) 1.6 (0.6) 1.5 (0.5)
Protein intake (% of energy) 19.7 (3.9) 17.7 (3.1) 16.2 (2.5) 21 (3.8) 19.2 (3.2) 17.6 (2.9)
Carbohydrate intake (% of energy) 40.5 (8) 46.7 (6.8) 51.2 (6.5) 44.8 (7.4) 50.1 (6.5) 54.7 (6.6)
Healthy plant foods (servings/day)
Whole grains 0.7 (0.9) 1.1 (1.1) 1.6 (1.2) 0.9 (0.9) 1.4 (1.1) 2.1 (1.3)
Fruits 0.9 (0.9) 1.4 (1) 1.9 (1.1) 0.7 (0.7) 1.2 (0.9) 1.8 (1.1)
Vegetables 2.3 (1.4) 2.9 (1.5) 3.7 (1.8) 2.1 (1.4) 2.9 (1.8) 4.1 (2.1)
Nuts 0.1 (0.2) 0.2 (0.3) 0.3 (0.4) 0.1 (0.1) 0.2 (0.2) 0.3 (0.3)
Legumes 0.3 (0.2) 0.4 (0.3) 0.5 (0.3) 0.2 (0.2) 0.4 (0.3) 0.6 (0.4)
Vegetable oil 0.2 (0.3) 0.3 (0.4) 0.4 (0.4) 0.2 (0.3) 0.3 (0.4) 0.5 (0.5)
Tea and coffee 2.6 (1.9) 3.1 (1.9) 3.6 (2) 1.8 (1.8) 2.2 (1.9) 2.7 (2)
Less healthy plant foods
(servings/day)
Fruit juices 0.4 (0.6) 0.7 (0.7) 1 (0.8) 0.4 (0.6) 0.6 (0.7) 1.1 (1)
Refined grains 1.2 (1.1) 1.6 (1.3) 2.2 (1.5) 1.2 (0.9) 1.6 (1) 2 (1.1)
Potatoes 0.4 (0.3) 0.5 (0.4) 0.7 (0.4) 0.4 (0.3) 0.5 (0.4) 0.7 (0.4)
Sugar-sweetened beverages 0.2 (0.6) 0.3 (0.6) 0.4 (0.6) 0.4 (0.8) 0.5 (0.9) 0.6 (0.9)
Sweets and desserts 0.8 (0.9) 1.2 (1.2) 1.8 (1.4) 0.8 (0.9) 1.2 (1.1) 1.6 (1.2)
131
Table 30 (continuation). Age and age-standardized baseline characteristics of women according
to quintiles of an overall plant-based diet index (PDI) in the Nurses’ Health Study (NHS) and NHSII.
NHS (1984) NHSII (1991)
Quintile 1 Quintile 3 Quintile 5 Quintile 1 Quintile 3 Quintile 5
(n=14,629) (n=13,832) (n=15,857) (n=20,408) (n=16,541) (n=17,944)
Animal foods (servings/day)
Animal fat 0.5 (0.9) 0.4 (0.8) 0.3 (0.6) 0.2 (0.5) 0.2 (0.4) 0.1 (0.4)
Dairy 2 (1.4) 2 (1.4) 2 (1.3) 2.3 (1.5) 2.3 (1.5) 2.3 (1.4)
Eggs 0.4 (0.4) 0.3 (0.3) 0.3 (0.3) 0.2 (0.2) 0.2 (0.2) 0.2 (0.2)
Fish and seafood 0.3 (0.3) 0.3 (0.3) 0.3 (0.2) 0.3 (0.2) 0.3 (0.2) 0.3 (0.3)
Meat 1.7 (0.8) 1.7 (0.8) 1.7 (0.8) 1.6 (0.8) 1.7 (0.8) 1.6 (0.9)
Miscellaneous animal-based foods 0.5 (0.4) 0.4 (0.4) 0.4 (0.4) 0.4 (0.4) 0.4 (0.3) 0.4 (0.3)
Food groups (servings/day)
Healthy plant foods 7.1 (3) 9.3 (3.3) 12.1 (3.7) 6 (2.9) 8.6 (3.5) 12 (4.1)
Less healthy plant foods 3 (1.9) 4.4 (2.3) 6.1 (2.7) 3.1 (1.8) 4.4 (2.2) 5.9 (2.4)
Animal foods 5.4 (2.3) 5.1 (2.1) 4.9 (1.9) 5 (2.1) 5 (2.1) 4.8 (2)
NHS= Nurses’ Health Study; NHSII= Nurses’ Health Study II. Values are means (standard deviations) for continuous variables
and percentages for categorical variables. All variables except age are standardized to the age distribution of the study
population. Each food intakes expressed as servings/day. * Value is not age adjusted.
In multivariable-adjusted models (Figure 21), a higher adherence to a PDI was

significantly inversely associated with breast cancer (HR Q5vs.Q1 0.89; 95% CI 0.84, 0.95;
P-trend<0.01). When we analyzed hPDI and uPDI separately, we found a modest
inverse association between a hPDI and breast cancer incidence (HR Q5vs.Q1 0.89; 95%
CI 0.75, 0.92; P-trend<0.01). A 10-unit increment in PDI was associated with an 8% (HR
0.94; 95% CI 0.91, 0.97) and 6% (HR 0.92; 95% CI 0.89, 0.96) relative lower risk of
breast cancer.
Figure 21. Age-adjusted and multivariable-adjusted hazard ratios (95% confidence intervals) for
total breast cancer according to quintiles of plant-based diet indices (PDI, hPDI, uPDI) in the NHS
and NHSII.
Multivariable model stratified by age in months and calendar year, adjusted for race (Non-Hispanic Caucasian, African-American,
Asian-American, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45-49,
50-52, 53+), postmenopausal hormone use (never user, past user, current user– estrogen only for <5 years, current user – estrogen
only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin user for ≥5 years, current user of
other types), oral contraceptive use history (never, ever), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at
≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months,
breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast disease (yes or no), height (<1.60,
1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5-9, 10-14, 15+ g/day), cumulatively updated
total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass index at age 18 years (<20.0, 20.0-
21.9, 22.0-23.9, 24.0-26.9, ≥27.0) and neighborhood-based socioeconomic status indicator (continuous).
132
We observed a significant heterogeneity by ER status in the hPDI (P-
heterogeneity=0.01) and the uPDI (P-heterogeneity=0.01) (Table 31). We detected an
inverse association between a higher hPDI and ER-negative breast cancer (HR Q5vs.Q1
0.77; 95%CI, 0.65, 0.90; P-trend<001) and a positive association between uPDI and ER-
negative breast cancer (HRQ5vsQ11.28;95%CI, 1.08, 1.51; P-trend<0.01). This association
was further observed when we performed multivariable spline analysis (Figure 22).
Table 31. Multivariable-adjusted hazard ratios (95% confidence intervals) for the association between
quintiles (Q) of cumulatively updated PDI, hPDI and uPDI and breast cancer tumor subtypes in the NHS
and NHSII.
Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5 P Trend
ESTROGEN RECEPTOR STATUS (NHS with follow-up of 1984-2016 and NHSII with follow-up of 1991-2017)
Plant-based dietary index
Cases/ 1609/ 1599/ 1659/ 1636/ 1632/
ER+ Person-years 968,461 955,406 980,656 974,166 966,567
MV 1.00 0.97 (0.91, 1.04) 0.97 (0.90, 1.04) 0.94 (0.88, 1.02) 0.93 (0.86, 1.00) 0.05
Cases/ 368/ 358/ 334/ 345/ 345/
ER- Person-years 969,664 956,537 981,914 975,387 967,764
MV 1.00 0.96 (0.82, 1.11) 0.86 (0.74, 1.01) 0.87 (0.74, 1.02) 0.85 (0.72, 1.00) 0.03
Healthful Plant-based dietary index
Cases/ 1551/ 1577/ 1671/ 1656/ 1680/
ER+
Person-years 972,520 960,328 970,065 977,670 964,673
MV 1.00 0.96 (0.89, 1.03) 0.98 (0.91, 1.05) 0.94 (0.87, 1.01) 0.91 (0.85, 0.99) 0.02
Cases/ 377/ 390/ 325/ 336/ 322/
ER-
Person-years 973,605 961,476 971,282 978,887 966,017
MV 1.00 0.99 (0.86, 1.15) 0.81 (0.70, 0.95) 0.82 (0.70, 0.96) 0.77 (0.65, 0.90) <0.01
Unhealthful Plant-based dietary index
Cases/ 1750/ 1690/ 1577/ 1594/ 1524/
ER+
Person-years 972,733 967,320 972,116 963,230 969,856
MV 1.00 0.99 (0.93, 1.06) 0.95 (0.88, 1.02) 0.99 (0.92, 1.06) 1.00 (0.93, 1.08) 0.99
Cases/ 305/ 342/ 377/ 367/ 359/
ER-
Person-years 974,085 968,560 973,277 964,382 970,963
MV 1.00 1.17 (1.00, 1.36) 1.29 (1.10, 1.50) 1.26 (1.07, 1.48) 1.28 (1.08, 1.51) <0.01
ER= estrogen receptor. Multivariable model (MV) stratified by cohort, age in months and calendar year, adjusted for race (Non-
Hispanic Caucasian, African, Asian, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause
(premenopausal, <45, 45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current user– estrogen only for
<5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin
user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at first birth (nulliparous,
1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding h istory
(never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast
disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5-9, 10-
14, 15+ g/day), cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass
index at age 18 years (<20.0, 20.0-21.9, 22.0-23.9, 24.0-26.9, ≥27.0) and socioeconomic status (continuous). 1 For testing
heterogeneity by subtype, we used the Lunn-McNeil approach, for Multivariable model (MV), p for heterogeneity PDI=0.25,
hPDI=0.01; uPDI=0.01.
133
A B
Hazard Ratio for ER- breast cancer
Healthful Plant-based diet index Unhealthful Plant-based diet index
Figure 22. Multivariable spline analysis of the association between adherence to a healthful (A)
and unhealthful (B) plant-based dietary indeces and risk of incident breast cancer in the NHS
(1984-2016) and NHSII (1991-2017) (pooled).
Although we observed no significant heterogeneity by molecular subtypes (Table

32), each 10-unit increase in the hPDI was associated with lower risk of HER2-enriched
(HR 0.80; 95% CI, 0.65, 0.99) and basal-like tumors (HR 0.79; 95% CI, 0.65, 0.96).
Table 32. Multivariable-adjusted hazard ratios (95% confidence intervals) for the association between quintiles of
cumulatively updated hPDI and breast cancer tumor subtypes in the NHS and NHSII.
P- Per 10-unit
Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5
trend increase
MOLECULAR SUBTYPES (NHS with follow-up of 1984-2006 and NHSII with follow-up of 1991-2005)
561/ 523/ 626/ 604/ 605/
Luminal Cases/ PY
664,554 647,388 667,302 668,016 658,333
A2
MV 1.00 0.88 (0.78, 1.00) 0.99 (0.88, 1.11) 0.93 (0.82, 1.05) 0.88 (0.77, 1.00) 0.12 0.95 (0.89, 1.01)
254/ 247/ 265/ 265/ 286/
Luminal Cases/ PY
2
664,818 647,610 667,608 668,332 658649
B
MV 1.00 0.94 (0.79, 1.13) 0.96 (0.80, 1.15) 0.96 (0.80, 1.15) 1.00 (0.83, 1.20) 0.99 1.01 (0.92, 1.11)
134
Table 32 (continuation). Multivariable-adjusted hazard ratios (95% confidence intervals) for the association between
quintiles of cumulatively updated hPDI and breast cancer tumor subtypes in the NHS and NHSII.
P- Per 10-unit
Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5
trend increase
56/ 66/ 62/ 41/ 45/
Cases/ PY
HER-22 664,998 647,780 667,794 668,527 658,868
MV 1.00 1.10 (0.76, 1.58) 1.05 (0.72, 1.53) 0.69 (0.45, 1.06) 0.72 (0.47, 1.11) 0.04 0.80 (0.65, 0.99)
69/ 56/ 60/ 62/ 52/
Basal- Cases/ PY
664,993 647,785 667,792 668,505 658861
like2
MV 1.00 0.82 (0.57, 1.17) 0.85 (0.59, 1.21) 0.86 (0.60, 1.24) 0.70 (0.47, 1.04) 0.13 0.79 (0.65, 0.96)
PY= Person-Years; Multivariable model (MV) stratified by cohort, age in months and calendar year, adjusted for race (Non-Hispanic Caucasian,
African, Asian, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45-49, 50-52, 53+),
postmenopausal hormone use (never user, past user, current user– estrogen only for <5 years, current user – estrogen only for ≥5 years, current
estrogen + progestin user for < 5 years, current estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use history
(never, ever), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children be fore age 25, 2+ children ≥25
years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of
benign breast disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5-9, 10-14,
15+ g/day), cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass index at age 18 years
(<20.0, 20.0-21.9, 22.0-23.9, 24.0-26.9, ≥27.0) and socioeconomic status (continuous). 1 For testing heterogeneity by subtype, we used the Lunn-
McNeil approach, for Multivariable model (MV), P for heterogeneity hPDI=0.19
In ancillary analyses (Figure 23), we entered variables for the three food
categories together into the fully adjusted model in place of the indices. We found an
inverse association between extreme quintiles of healthy plant-based foods and ER-
negative breast cancer (HR 0.74; 95% CI 0.61, 0.88; P-trend<0.01).
Figure 23. Risk of ER-negative breast cancer according to animal, healthy plant and less healthy
plant foods.
Model stratified by age in months, cohort and calendar year, adjusted for race (Non-Hispanic Caucasian, African, Asian,
Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45-49, 50-
52, 53+), postmenopausal hormone use (never user, past user, current user– estrogen only for <5 years, current user –
estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin user for ≥5 years,
current user of other types), oral contraceptive use history (never, ever), parity and age at first birth (nulliparous, 1 child
before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding history
(never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign
breast disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m), cumulatively updated alcohol intake
(0, <5, 5-9, 10-14, 15+ g/day), cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-
hours/week), body mass index at age 18 years (<20.0, 20.0-21.9, 22.0-23.9, 24.0-26.9, ≥27.0) and socioeconomic status
(continuous). The 3 food categories (healthy and less healthy plant foods, and animal foods) were simultaneously included
in the same model.
135
We assessed effect modification by physical activity, current BMI and
menopausal status (Figure 24). Although we did not find any significant interaction,
among women with higher physical activity (≥21MET-h/week) and lower BMI (<25
kg/m2), higher adherence to the PDI was inversely associated with ER-negative breast
cancer [(HR per 10-unit increase 0.78; 95%CI 0.65, 0.95); (HR 0.81; 95%CI 0.70, 0.93)].
Moreover, we observed a significant inverse association between each 10-unit increase
in the hPDI and ER-negative breast cancer among normal weight women (HR 0.80; 95%
CI 0.71, 0.90), and a suggestive association among those with higher BMI (HR 0.90;
95% CI 0.80, 1.02). Associations were similar by menopausal status; however, the
association between the hPDI and ER-negative tumors was stronger for
postmenopausal women (HR 0.86; 95%CI 0.78, 0.95). Furthermore, each 10-unit
increase in the uPDI was positively associated with a higher risk of breast cancer among
those with low physical activity (HR 1.16; 95%CI 1.06, 1.27) and normal weight women
(HR 1.19; 95%CI 1.06, 1.33). Among postmenopausal women, there was a suggestion
of stronger associations among never and past users of postmenopausal hormones
(HRQ5vsQ1 0.69;95%CI 0.54,0.90; P-trend<0.01) than current users. Nonetheless, P for
interaction was non-significant.
136
Figure 24. Pooled hazard ratios of estrogen receptor negative breast cancer per 10-units increment in
the three dietary indices (PDI, hPDI and uPDI) across subgroups (physical activity, current BMI and
menopausal status).
The hazard ratios (HRs) and p values for women were obtained after combining all 2 cohorts (NHS; NHSII). Stratified by age in
months, cohort and calendar year, adjusted for race (Non-Hispanic Caucasian, African, Asian, Hispanic Caucasian), age at
menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45-49, 50-52, 53+), postmenopausal hormone
use (never user, past user, current user– estrogen only for <5 years, current user – estrogen only for ≥5 years, current estrogen
+ progestin user for < 5 years, current estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use
history (never, ever), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before
age 25, 2+ children ≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family
history of breast cancer (yes or no), history of benign breast disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74,
1.75 + m), cumulatively updated alcohol intake (0, <5, 5-9, 10-14, 15+ g/day), cumulatively updated total caloric intake (kcal/day,
quintiles), physical activity (linear MET-hours/week), body mass index at age 18 years (<20.0, 20.0-21.9, 22.0-23.9, 24.0-26.9,
≥27.0) and socioeconomic status (continuous).For the analysis of BMI, we further adjusted for current body mass index (linear,
kg/m2).To test whether the PDI, hPDI, uPDI and breast cancer association differed by current BMI, physical activity, or menopausal
status we added interaction terms and used the Wald test. BMI= body mass index; CI= confidence interval; MET= metabolic
equivalent task.
137
We evaluated the extent to which the inverse association with higher PDI and
hPDI may be mediated by less weight gain from age 18. The calculated mediation
proportion was 11.0% (95%CI=3.4% - 30.2%; P=<0.01), indicating that less weight gain
could statistically explain 11.0% of the inverse association with PDI. Moreover, the
proportion of hPDI effect mediated by weight change since age 18 was 7.3%
(95%CI=3.0% - 16.5%; P<0.01). Results from latency analysis (Table 33) indicated the
highest versus the lowest adherence to a hPDI 0-4 and 4-8 years before diagnosis was
associated with lower risk of ER-negative breast cancer [(HR Q5vs.Q1 0.85; 95% CI, 0.72,
1.01; P-trend=0.01); (HR Q5vs.Q1 0.80; 95% CI, 0.66, 0.96; P-trend=0.01)].
Table 33. Risk of ER-negative breast cancer according to PDI, hPDI, uPDI with different latency periods
using pooled data from NHS and NHSII.
Simple update
4-8 y lag 8-12 y lag 12-16 y lag 16-20 y lag
(0-4 y lag)
Cases/
person- 1,542/3,955,108 1,296/3,314,002 1,051/2,728,313 777/2,142,682 484/1,560,331
years
Plant-based diet index
Q1 1.00 1.00 1.00 1.00 1.00
Q2 1.10 (0.94, 1.29) 0.93 (0.78, 1.11) 0.85 (0.70, 1.03) 0.89 (0.71, 1.11) 0.95 (0.72, 1.25)
Q3 0.93 (0.79, 1.10) 0.97 (0.81, 1.15) 0.87 (0.72, 1.06) 1.03 (0.82, 1.28) 0.88 (0.66, 1.18)
Q4 0.99 (0.83, 1.16) 0.91 (0.76, 1.09) 0.85 (0.69, 1.03) 0.93 (0.73, 1.17) 0.84 (0.62, 1.14)
Q5 0.82 (0.69, 0.99) 0.92 (0.76, 1.11) 0.88 (0.72, 1.09) 0.90 (0.70, 1.15) 0.97 (0.71, 1.31)
P-trend 0.02 0.35 0.28 0.52 0.63
Healthful plant-based diet index
Q1 1.00 1.00 1.00 1.00 1.00
Q2 1.07 (0.91, 1.25) 1.06 (0.89, 1.26) 1.08 (0.89, 1.31) 0.90 (0.72, 1.12) 0.94 (0.70, 1.26)
Q3 0.98 (0.84, 1.15) 0.95 (0.80, 1.13) 1.04 (0.86, 1.26) 0.77 (0.61, 0.96) 0.94 (0.70, 1.25)
Q4 0.89 (0.76, 1.05) 0.94 (0.79, 1.12) 0.94 (0.77, 1.15) 0.80 (0.63, 1.01) 1.00 (0.74, 1.35)
Q5 0.85 (0.72, 1.01) 0.80 (0.66, 0.96) 0.91 (0.74, 1.12) 0.88 (0.70, 1.11) 0.98 (0.72, 1.32)
P-trend 0.01 0.01 0.20 0.18 0.99
Unhealthful plant-based diet index
Q1 1.00 1.00 1.00 1.00 1.00
Q2 0.99 (0.84, 1.17) 1.04 (0.87, 1.25) 1.09 (0.89, 1.33) 1.11 (0.88, 1.40) 1.17 (0.87, 1.58)
Q3 1.12 (0.95, 1.31) 1.19 (0.99, 1.42) 1.29 (1.06, 1.57) 1.04 (0.82, 1.31) 1.30 (0.97, 1.73)
Q4 1.03 (0.87, 1.22) 1.19 (0.99, 1.43) 1.08 (0.88, 1.33) 1.06 (0.84, 1.35) 0.95 (0.69, 1.30)
Q5 1.01 (0.85, 1.21) 1.22 (1.01, 1.48) 1.20 (0.97, 1.49) 1.29 (1.01, 1.64) 1.17 (0.85, 1.60)
P-trend 0.80 0.02 0.13 0.07 0.68
Q=quintiles; All values are hazard ratios with 95% CIs in parentheses, stratified by age in months, cohort and calendar year, adjusted
for race (Non-Hispanic Caucasian, African, Asian, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at
menopause (premenopausal, <45, 45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current user– estrogen
only for <5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin
user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at first birth (nulliparous, 1
child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding history (never,
breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast disease (yes
or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5-9, 10-14, 15+ g/day),
cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass index at age 18
years (<20.0, 20.0-21.9, 22.0-23.9, 24.0-26.9, ≥27.0) and socioeconomic status (continuous).
138
DISCUSSION
139
1. SUN Project
The SUN study began in 1999 at the University of Navarra’s Department of
Preventive Medicine and Public Health following the models established in large cohort
studies from the Harvard T.H. Chan School of Public Health (“Nurses' Health Studies”
and “Health Professionals Follow-up Study"). Since the early beginning, the SUN study
has integrated other universities (Palmas de Gran Canarias, Jaen, Cantabria, Zaragoza,
Malaga, Santiago de Compostela and the International University of Catalonia) to pursuit
their objectives. Currently, the SUN Project has more than 22,500 participants, and given
its dynamic nature, the number of participants enlarges every year.
Even though the initial rationale of the SUN project was to shed light on the benefits
of the Mediterranean diet on chronic diseases, the study allows the analysis of a wide
range of risk factors and outcomes. Therefore, in this thesis dissertation, we were able
to study the association between various dietary factors and the risk of breast cancer.
Below, we explain the results previously described and their implication given the
currently available evidence.
2. Nurses’ Health Studies

The original Nurses’ Health Study (NHS) started in 1976 with the central goal of
evaluating the lifelong consequences of oral contraceptive use and breast cancer risk.
Over the past 40 years, the Nurses’ Health Studies have advanced to a great extent our
understanding of numerous health- and disease-related outcomes in women through
questionnaires and biomarkers, including diet, hormones, and trace elements359. In
terms of breast cancer, research in the Nurses’ Health Studies has contributed
considerably to the understanding of risk factors for this disease and continues to assess
whether other factors such as hormone levels, mammographic density, and genetic
polymorphisms, enhance the capability to classify women at highest risk for breast
cancer. In summary, these studies serve as model prospective cohort studies with
repeated measurements that are having a huge impact on public health policy and
practice both in the US and globally.

project and Nurses’ Health Studies
We used three different cohort studies to assess the relationship between SSB
consumption and breast cancer risk. On the one hand, in the SUN Project, a frequent
consumption (>1serving/month) of SSBs was associated with a higher incidence of
postmenopausal breast cancer. We did not further evaluate ASBs and breast cancer risk
140
given its low variability among participants. On the other hand, in the Nurses’ Health
Studies, no significant associations of SSB or ASB with breast cancer risk were observed
overall. However, we observed a direct association between intake of SSBs and risk of
breast cancer among normal weight women, which was independent of weight change
and other established dietary and nondietary breast cancer risk factors. These results
are similar to the ones obtained in the SUN Project, since the average women’s BMI in
the SUN Project was approximately 22 kg/m2.
We also found a modest inverse association between ASB intake and risk of luminal
A breast tumors, particularly among obese women. Because this is the first evaluation
of ASBs and subtype-specific breast tumors, confirmation in other cohort studies is highly
needed. Given our study’s observational design, we cannot completely exclude the
possibility of residual confounding. Nonetheless, the observational findings persisted
after adjusting for known and suspected predictors of breast cancer, and sensitivity
analyses evaluated the scenarios under which unknown confounders might explain
some or all of the observed association between the dietary exposure and breast cancer.
Although these findings require replication, there are several potential mechanisms
which may contribute to these observed results. Adiposity plays a key role in
postmenopausal breast cancer, with higher concentration of adipocytokines and
hyperinsulinemia and insulin resistance leading to breast cancer cell growth 360.
Nonetheless, a mechanism independent of adiposity in the Nurses’ Health Studies is
likely given the direct association among normal weight women which persisted after
adjustment for weight change. Indeed, whereas among overweight/obese women, we
observed no association with SSB consumption, among normal weight women, greater
consumption of SSB was positively associated with risk, even after adjusting for weight
change. Aside from obesity, mechanisms underlying a link between SSBs and breast
cancer may stem from the high amounts of rapidly absorbable carbohydrates such as
any form of sugar or high-fructose corn syrup, the primary sweeteners used in SSBs
which bring rapid and dramatic increases in blood glucose and insulin concentrations.
Increased insulin concentrations may influence breast cancer risk either directly, by
stimulating insulin receptors in breast tissue, or indirectly, by augmenting the bioactivity
of IGF‐I, which stimulates cell proliferation and inhibits apoptosis 361. Moreover, there is
evidence that both insulin and IGF‐I stimulate the synthesis of sex steroids, particularly
androgens, and decrease the concentration of sex‐hormone‐binding globulin which in
turn, produces an increased tissue concentration of estrogens, formed by local
conversion of the androgens. Further, the caramel coloring used in SSBs (4-
141
methylimidazole)362 is high in advanced glycation end products, which may additionally
363
increase insulin resistance and inflammation .
While SSB and cardiometabolic health has been broadly studied302, little is known
about whether intake of these beverages impacts risk of total and subtype-specific breast
tumors. Inconsistent findings between consumption of SSBs and breast cancer in the
epidemiologic literature may be due to differences in exposure definitions, which have
included intake of sugars, sweets and desserts, individual foods or an overall dietary
pattern that included sweet foods and/or beverages364. Consumption of dessert foods,
sweet beverages, and added sugars was positively associated with breast cancer risk in
a population-based case-control study in Long Island, New York364. In the latter study,
and analogous to our results from the Nurses’ Heath Studies, women with lower BMI
exhibited a stronger association between dessert consumption and breast cancer
compared to those with higher BMI. A recent study from the Nurses’ Health Studies365
showed a higher adherence to a diabetes risk reduction score (where SSBs scored
negative) was inversely associated with breast cancer particularly among normal weight
women and independent of weight change. Hence, encouraging lifestyle modifications
to lessen the risk of developing insulin resistance and hyperinsulinemia may be a
promising breast cancer primary prevention strategy. In fact, a study evaluating the
association between adherence to the WCRF/AICR cancer prevention
recommendations and breast cancer incidence in African American women found that
adherence to the recommendation to limit intake of sugary beverages was protective
against breast cancer risk, particularly ER-positive/PR-positive tumors150.
Similar to our results from both the SUN Project and the Nurses’ Health Studies, for
total consumption of sugary beverages, null findings were observed for breast cancer in
multivariable-adjusted models in the Framingham Offspring cohort 211 and in the
Canadian Study of Diet, Lifestyle, and Health 212. A direct association between sugary
drinks consumption and the risk of breast cancer was found in the NutriNet-Santé cohort
study, specifically among premenopausal women 209. Moreover, in our results from the
Nurses’ Health Studies, we showed a suggestive, non-significant, higher risk of
postmenopausal breast cancer for women consuming 1 or more servings of SSBs per
day compared to those with infrequent consumption (<1/month). On the other hand, in
the SUN cohort study, we observed a statistically significant increased risk of
postmenopausal breast cancer which was in line with previous results from the
Melbourne Collaborative Cohort Study 210 and a case-control study with African American
and European American women209. Furthermore, in a previous study of the EPIC-France
142
cohort366, rapidly absorbed carbohydrates were associated with risk of postmenopausal
breast cancer in women who were overweight and women with large waist
circumference. Further analyses with longer follow-up and larger number of breast
cancer cases are needed to advance the knowledge.
In contrast to SSBs, which represents the largest source of added sugar in the
American diet142, ASBs contain few to no calories, which makes them a likely attractive
substitute for SSBs. However, observational studies and intervention trials have shown
inconclusive results regarding their health effects 367,368. In the Nurses’ Health Studies,
null results were observed for ASBs intake and overall breast cancer risk, consistent with
results from the NutriNet-Santé cohort study209. While emerging research shows mixed
results369, additional high quality, long-term research is needed to understand the
mechanisms of action of SSBs and ASBs, and further explore breast cancer subtypes.
4. Phenolic acids subclasses, individual compounds and breast cancer risk in

the SUN Project.
The results from this prospective study of Mediterranean university graduates
support the hypothesis that high consumption of hydroxycinnamic acids (>435 mg/day)
and chlorogenic acids (>290 mg/day) was associated with lower risk of breast cancer
among postmenopausal women.
Phytochemicals are now gaining considerable attention in cancer prevention as a

result of all their antioxidant, antiproliferative, antiangiogenic and proapoptotic properties,
among others370. In the SUN Project, phenolic acids were the main contributors to total
polyphenol intake, along with flavonoids, and coffee was undoubtedly the major food
source (48%) and the main source of between-person variability (change in R2=0.78) in
total phenolic acids intake. These results are consistent with previous studies from
Mediterranean countries371,372.
Although a number of in vitro and in vivo studies suggest that polyphenol intake
373
and phenolic acids intake may influence carcinogenesis and tumor development , no
previous prospective observational study has been conducted to test potential
associations between subclasses and individual phenolic acids intake and breast cancer
risk.
Phenolic acids are of substantial interest because foods rich in these molecules,
such as coffee, have demonstrated a decreased risk of breast cancer in animal and cell
143
models374. Epidemiological studies have relied on the estimation of dietary intakes and
food sources accrued from phenolic acids in different cohorts,188,375,376 but the study of
phenolic acid intake in prospective cohort studies is scarce in cancer research. In fact,
only one case-control study in Southern Italy has particularly evaluated phenolic acid
intake in relation to a hormone-related tumor, specifically prostate cancer186. In that case-
control study, caffeic and ferulic acids were both associated with a reduced risk of
prostate cancer, and caffeic and hydroxybenzoic acids were linked to a lower risk of
advanced prostate cancer.
Phenolic acids and, particularly, hydroxycinnamic acids, are important sources of

antioxidants due to their ubiquity in the plant kingdom. They are mostly present in coffee
beans, tea, berries, tomatoes, citrus fruits, grapes, green vegetables, potatoes, beetroots
and cereals370. Thus, they could have a potential role in the prevention of oxidative
stress-related diseases, such as cancer374. These compounds are found both covalently
adhered to the plant cell wall polysaccharides and in cytoplasm as soluble forms. In the
PREDIMED study371, hydroxycinnamic acids were the main source of polyphenol intake
(33%), a figure close to our current study (37%). In the EPIC study, Zamora-Ros et al.188
observed that the hydroxycinnamic acid subclass was the major contributor to total
phenolic acids intake among 36,037 subjects aged 35 to 74 years from 10 European
countries. More concretely, in the latter study, hydroxycinnamic acids explained between
84.6% and 95.3% of total polyphenol intake, depending on the location. Similar results
were observed by Pérez-Jiménez et al.377 in a French study.
Hydroxycinnamic acids are known to display anticancer, antiatherogenic,

antimalarial, antifungal, antimicrobial, and antioxidant activities 370. In terms of
postmenopausal breast cancer risk, there are several mechanisms that could explain the
protective association found for hydroxycinnamic acids intake. Hydroxycinnamic acids
have been reported to inhibit cancer invasion and metastasis,378 as well as to modulate
apoptosis and ERα expression374. Inhibition of ERα expression decreases the peripheral
conversion of androgens to estrogens, a key mechanism for postmenopausal breast
cancer379–381.
According to Johnston et al.,382 chlorogenic acids may lower food cravings, reduce
daily calorie intake, induce body fat loss by thermogenesis, and enhance glucose
tolerance and insulin sensitivity by acting as a peroxisome proliferator activated receptor
antagonist. This mechanism could partially explain its potential protective effect in
postmenopausal women 383 as this group is particularly susceptible to insulin resistance
144
and accumulation of fatty acids. In fact, chlorogenic acids, through their potential
antioxidant effects, may also reduce the proliferation of new fat cells 384. Nakatani et al.385
suggested that some chlorogenic acids isomers (3-, 4-, and 5-CQA) isolated from prunes
may have antioxidant activity, such as scavenging activity on superoxide anion radicals,
and inhibition of methyl linoleate oxidation. This antioxidant activity of chlorogenic acids
has been observed either in in vitro and in vivo studies, and in various disease models.
They also have anti-inflammatory activity, which explains their ability to downregulate
proinflammatory cytokines through modulation of key transcription factors 386.
5. Fat intake and breast cancer risk in the SUN project.
Dietary fat intake has been widely investigated as a possible risk factor for breast
cancer because it is thought to increase endogenous estrogen levels 387, however, there
is overall limited evidence for an association140 and results from prospective studies are
conflicting388.
In the SUN Project, we did not find an association of total fat or its subtypes with
breast cancer risk. In line with these results, most prospective studies have not observed
an association, even at very low or very high levels of total fat intake, as in the Pooling
Project of eight prospective cohort studies with over 350,000 women, the largest study
of dietary fat and breast cancer to date 221. Furthermore, two randomized trials of total fat
reduction showed no significant effect on breast cancer incidence223,389. Yet dietary fats
are heterogeneous, and a few prospective studies suggested a direct association
between intake of saturated/animal fat, particularly from red meat and high-fat dairy, and
breast cancer incidence221,228,390 as well as an inverse association between
monounsaturated/vegetable fat and breast cancer risk 221.
In the Nurses’ Health Studies, higher pre-diagnosis fat intake was not associated
with greater risk of lethal breast cancer in these large prospective cohort studies,
consistent with the weight of the evidence against a causal role for fat intake and breast
cancer incidence391.
Moreover, the concept of breast cancer heterogeneity has become a driving force for
breast cancer epidemiology. Recent research suggests that breast cancer tumor
subtypes such as luminal A, luminal B, triple-negative, and others, which are differentially
related to disease prognosis392,393, have distinct etiologies394. Therefore, a potentially
important new research direction is to identify risk factors for subtypes of breast cancers.
145
Whether dietary fat intake influences development of certain subtypes of breast cancer
is unknown and requires a careful examination.
A previous analysis from the EPIC study found a weak association between
saturated fat intake and breast cancer risk395. Particularly, women in the highest quintile
of saturated fat intake had a statistically significant greater risk of breast cancer than
those in the lowest quintile. Increases in total and monounsaturated fat intake
(continuous variables) were also associated with greater breast cancer risk. The
association between fat intake and breast cancer did not vary by menopausal status at
baseline or at diagnosis.
Our findings are consistent with the weight of the evidence, which argues against a
causal role for high fat intake and incidence of breast cancer. Nonetheless, mounting
basic science evidence shows that some fatty acids may influence breast cancer risk
through a variety of immune-inflammatory mechanisms396. For example, while
monounsaturated and polyunsaturated fats, especially long-chain ω-3 polyunsaturated
fatty acids, appear to be beneficial for long-term health397, trans fatty acids may increase
breast cancer risk through proinflammatory mechanisms 398,399.
Healthy fats are an important part of a balanced diet and do not appear to increase
breast cancer risk or mortality. However, a limitation of observational studies is that
dietary questionnaires are limited in assessing eating out behaviors, and high fat
processed foods consumed out of home might not be fully captured.
From a public health perspective, the oversimplified advice to reduce all fat or follow
a low-fat diet has not stood the test of science; robust evidence supports the need to
consider fat quality within a healthy dietary pattern.
6. Carbohydrate quality index and breast cancer risk in the SUN project.
In the SUN Project, we observed that those women with best CQI, especially
premenopausal women, had a significantly lower risk of breast cancer. The CQI was
defined by considering jointly the dietary fiber, the GI, the solid or liquid form of
carbohydrates and the degree of processing of carbohydrate-rich foods (whole or refined
grains). The total quantity of carbohydrate intake or the proportion of energy explained
by carbohydrates, their quality and their dietary sources comprise an emerging concern
for breast cancer prevention. Nonetheless, GI and GL have been widely used as markers
of carbohydrate quality, although there is no agreement on the best standard400. Recent
146
discussion around GI indicate that fiber intake and whole grain consumption could also
be appropriate markers of carbohydrate quality, even to a better extent than GI or GL 401.
Several studies have previously used the CQI, which has been associated to a lower
risk of cardiovascular disease 220, overweight and obesity219, and better micronutrient
intake adequacy both in elderly326 and among middle-age participants220. Additionally,
evidence on the association between several individual components that made up the
multidimensional CQI and breast cancer has been previously described. A recent meta-
analysis showed a 12% relative decrease in breast cancer risk with dietary fiber intake
by extracting the risk estimate of the highest and lowest reported categories of intake
from each of the twenty-four selected studies213. Furthermore, an increase of 10 g/day
of dietary fiber intake was associated with a 4% relative risk reduction of breast cancer.
Dietary fiber can promote the discharge and decomposition of harmful and carcinogenic
substances in the gut, promote the growth of probiotics and inhibit the growth of
pathogenic bacteria, thereby inhibiting production of carcinogens and promoting their
decomposition in the intestine 213. It also improves the phagocytosis of macrophages,
blocks nitrosamine synthesis, and reduces estrogen levels 213.
In the last decades, the role of GI in developing chronic conditions related to diet has
received major attention, albeit it is still an issue of debate. In a recent meta-analysis, for
every additional 10 units/day of GI the risk of breast cancer was relatively increased by
6% in postmenopausal women, whereas no increased risk was observed among
premenopausal women. In an umbrella review published in 2020, a higher consumption
of foods with a high GI seem to be associated with a higher risk of breast cancer175.
Studies have indicated that diets high in GI or GL might be linked to

hyperinsulinemia402, higher concentrations of IGF-1402, type 2 diabetes403, and
inflammatory biomarkers402, all of which play a role in breast carcinogenesis204.
Chronically, raised blood insulin –as a consequence of chronically high blood glucose–
is thought to be the underlying mechanism for GI/GL-related cancers. Insulin acts as
growth factor and also increases the bioactivity of IGFs, such as IGF-1, which can
promote tumor development by inhibiting apoptosis, stimulating cell proliferation and
sex-steroid synthesis404 and promoting angiogenesis405. Additional conditions linked to
chronically high blood glucose, such insulin resistance, obesity, and type 2 diabetes may
also impact cancer risk406.
147
Furthermore, the physical form of carbohydrates, namely solid vs. liquid sources, is
an important concern. In this sense, prospective investigations in several cohorts have
confirmed that consumption of liquid carbohydrates may be positively associated with
breast cancer risk209,210 whereas the consumption of unrefined or unprocessed solid
carbohydrates, such as whole grains may be linked to decreased breast cancer risk407.
Whole grains may lessen the post-prandial glucose and insulin responses leading to
better glycemic control408 which in turn may reduce breast cancer risk214.
These dietary components have also been found to be associated with reduced
levels of inflammatory markers and liver enzymes, and therefore, a decreased risk of
cancer. Whole grains are rich in antioxidants, including vitamins vitamin C and E and β-
carotene and minerals selenium, zinc, copper, and manganese which are elements of
enzymes with antioxidant activities and have been linked to be inversely associated with
breast cancer risk409. Finally, whole grains are an important source of some non-
nutrients, such as phytoestrogens and polyphenols (phenolic acids and lignans), that
have antioxidant properties and potential to inhibit cell proliferation and angiogenesis,
and to induce cell apoptosis407.
7. Healthful and unhealthful provegetarian food patterns and breast cancer risk
in the SUN project and Nurses’ Health Studies
In two large prospective cohort studies in the US, we found that a higher hPDI score,
a measure of adherence to a high-quality plant-based diet, was associated with lower
risk of total breast cancer, independent of weight change and total carotenoid intake.
The association was most evident in relation to ER-negative tumors for which non-
hormonal exposures may be most important 84,410,411. Less weight gain could statistically
explain 7.3% of the inverse association with hPDI.
Only few prospective studies have assessed the association between healthful and
unhealthful plant-based diets, defined by Satija et al.167,169,412–416. Because of the
impracticality of assessment of a pure vegetarian diet in these cohorts and the fact that
it may not easily be embraced by many individuals, consuming preferentially plant-
derived foods would be a more comprehensible message. In this context, these scores
are designed to understand common dietary patterns that incorporate a range of
progressively increasing proportions of plant foods and accompanying reductions in
animal-foods.
148
Previous studies that examined associations between these plant-based diets and
breast cancer are not consistent in the literature170. In the SUN project, a moderate, but
not greater, adherence to a provegetarian dietary pattern, was associated with a
decreased risk of breast cancer, otherwise no particular associations were found when
assessing healthful and unhealthful provegetarian dietary patterns separately.
Furthermore, a higher pro plant-based dietary score was associated with decreased risks
of overall cancer though no association was found for breast cancer170 which may partly
due to the restricted number of cases (n=487 breast cancer cases, only 13.6%
representing ER‐negative/PR‐negative and ER-negative/PR-positive).
In 2010, a systematic review and meta-analysis showed that a Prudent dietary

pattern characterized by high intakes of fruit, vegetables, whole grains, low-fat dairy
products, fish and poultry, was linked to a relative 11% breast cancer risk reduction193.
Subsequently, in 2017, the World Cancer Research Fund supported that there was
insufficient evidence to make a judgment about the association between dietary patterns
and the risk of breast cancer140. In a recent systematic review and meta-analysis of 32
observational studies158, a Western and a Prudent diet patterns were associated with a
relative 14% increased and a relative 18% reduced risk of breast cancer, respectively.
While no association with neither premenopausal nor postmenopausal breast cancer

was found in the SUN Project, in the NHS/NHSII cohorts, the healthful version of the
plant-based dietary pattern was associated with a relative 14% reduction in breast
cancer risk among postmenopausal but not among premenopausal women. The different
associations between dietary patterns and breast cancer risk by menopausal status may
be explained by the diet-estrogen pathway. After menopause, when ovarian production
of estrogen ceases, the serum levels of estrogen come from aromatization of
androstenedione to estrone in the stroma of fat cells followed by conversion to estradiol;
therefore, adipose tissue is the major source of estrogen among postmenopausal
women417.
Hormone receptor status in an important diagnostic and prognostic characteristic of

breast tumor and, therefore, deserves consideration. In previous analyses within the
NHS, the Prudent dietary pattern, the Dietary Approaches to Stop Hypertension 418,419,
the Mediterranean Diet420 and a diabetes risk reduction diet365 have been significantly
associated with a decreased risk of ER-negative breast tumors. Thus, results are in close
agreement with previous findings, which could be expected, as most healthy plant foods
positively weighted in the hPDI (e.g., vegetables, fruits 172,174,421, whole grains422, olive
149
oil164 or coffee365,423) have been associated with lower risk of breast cancer in prospective
cohort studies, including our own. Moreover, higher intakes of dietary fiber 424,425 and
carotenoids, particularly α-carotene, β-carotene, and lutein/zeaxanthin as shown in a
pooled analysis of 18 prospective cohort studies, were inversely associated with risk of
ER-negative breast cancer178. These compounds have been hypothesized to reduce
cancer risk through antioxidant, anti-inflammatory or antiproliferative activity 176 and by
decreasing concentrations of circulating estrogens and androgens, and reducing insulin-
resistance426. However, adjustment for carotenoid or dietary fiber intake did not
substantially change the observed associations with regard to hPDI adherence and ER-
negative breast tumors, indicating that other constituents in plant-based foods likely
account for the observed findings. In this sense, other phytochemicals present in fruits,
vegetables, and legumes, such as flavonoids or other phenolic compounds, or an
interaction among several phytochemicals, could be responsible for the observed
association427. Future studies should explore the associations between additional
phytochemical components and breast cancer risk. On the other hand, less healthy
plant-based foods positively weighted in the uPDI (refined grains, pastries 428, sugary
drinks209,210 or processed foods429) have been associated with higher risk of breast
cancer in other studies.
In line with our results, the Nurses’ Health Study reported null results for total breast
cancer and all molecular subtypes except for inverse associations between the DASH
diet and ER-negative and HER-2 type breast cancers418,419. Many prospective studies
have examined relations between the Mediterranean diet and breast cancer risk, with
most164,420,430–433 but not all419,434 studies observing inverse associations, especially
among those diagnosed with ER-negative or ER-negative/PR-negative cancers.
Previous studies with the strongest inverse associations for breast cancer risk were
conducted in Spain164, Greece433, and Europe as a whole431.
In our results from the Nurses’ Health Studies, we also showed a higher adherence
to an uPDI was associated with a higher risk of breast cancer, particularly, ER-negative
breast cancer and HER-2 positive breast cancer. Interestingly, the unhealthful version of
the plant-based diet was associated with a 10% relative higher risk of postmenopausal
breast cancer which is in line with a previous result from a systematic review and meta-
analysis of observational studies158.
In our analyses of the Nurses’ Health Studies, the consistent opposite associations
of hPDI and uPDI also strengthened the value of considering plant-based dietary quality
150
and increasing intake of healthy plant foods while lessening consumption of less healthy
plant foods. A healthful version of a PDI represents many advantages such as the
representation of a healthy plant-based diet without complete exclusion of animal foods.
In a previous analysis from the NHS and the Health Professionals Follow-Up Study416
where changes in plant-based diet quality and mortality were evaluated, a 10-point
increase in hPDI could be reached by increasing healthy plant foods (i.e., fruits,
vegetables and whole grains) by about 3 servings/day and decreasing less healthy plant
foods (i.e., refined grains and sugary beverages) and some animal foods (i.e., processed
meat) by approximately 2 servings/day. Such an approach is preferable because it is
flexible and allows individuals to make gentle changes to their diets. Besides, excluding
all animal foods might not be convenient for all populations as there is limited-suggestive
evidence that moderate intakes of some animal foods such as fish and dairy may be
associated with lower breast cancer risk140.
As diet is advocated by the WCRF as a potentially modifiable means to reduce

cancer risk, a healthful plant-based dietary pattern should be adopted to protect against
the development of breast cancer. Healthy diet plays an important role in cancer
prevention as shown in the literature. Hormone receptor-positive breast tumors are
contingent to hormonal risk factors for ongoing proliferation, while hormone receptor-
negative breast cancer relies more on non-hormonal risk factors like diet435.
Given the poorer prognosis and challenges couple with the treatment, prevention of
ER-negative breast cancer is of significant public health importance. While
nonmodifiable factors are strong predictors for ER-positive breast cancer, they have
weaker associations with ER-negative cancer436. Therefore, lifestyle risk factors may
have a comparatively greater influence in the ER-negative subtype and, therefore, a
potential opportunity for ER-negative breast cancer prevention.
8. Limitations and strengths
SUN Project
Common limitations
We acknowledge that the results of the present thesis dissertation should be

interpreted in light of some limitations.
In the first place, we are aware that the number of incident cases of breast cancer
that we have to date is small (both confirmed and probable), which is a determining factor
151
in obtaining sufficient statistical power to find significant differences, if any. The SUN
cohort is made up of a large number of young women, and the majority is, so far,
premenopausal. As previously explained, breast cancer diagnoses increase with age,
which contributes to reducing the number of incident cases of breast cancer for the time
being evaluated. Furthermore, the small number of observed postmenopausal breast
cancer cases, may have also led to low statistical power in some analyses and thus,
excessively wide confidence intervals. However, after finding similar point estimates
between pre- and postmenopausal women, despite not being significant, one could
expect the effect to be similar.
Second, and also due to the low number of incident cases of breast cancer, it has
not been possible to perform a sub-analysis according to different subtypes of breast
cancer. Nonetheless, we hope to carry out these analyses in the foreseeable future.
Third, the information on the diagnosis of breast cancer was self-reported. With
the intention of confirming this information, we asked participants to send as a copy of
the medical report with the disease diagnosis. This is why we carried out our analyses
separately, considering, on the one hand, those cases that we were able to confirm with
a medical report and, on the other hand, all cases (both confirmed and those for which
we did not have a medical report at the time of analysis). Despite this, we may have lost
some diagnosed breast cancer cases, which may have decreased the sensitivity to
detect incident breast cancer cases. However, the close monitoring of our participants
and the periodic consultation of the National Death Index may have reduced this number
of breast cancer cases losses. Nonetheless, the identified age-adjusted breast cancer
incidence in the SUN Project was consistent with the reported age-adjusted incidence of
breast cancer in the Spanish population 437.
Fourth, the use of a self-reported FFQ for dietary assessment could have led to
non-differential misclassification and measurement errors and may have probably
underestimated true associations. However, the FFQ is the most efficient and feasible
tool to evaluate food habits in large epidemiological studies438. Furthermore, reliability
and validity of the FFQ used in our cohort have been evaluated and showed good
correlation with nutrient intake313,315,439.
Fifth, we cannot rule out residual confounding by unknown factors. Nonetheless,

we attempted to adjust for a wide variety of established or hypothesized medical,
lifestyle, and dietary risk factors for breast cancer. Confounding is often considered a
152
special type of bias that results in “a confusion of effects” 440. Although we have adjusted
for potential confounding factors, given the observational nature of the study, we cannot
completely rule out the possible existence of residual confounding. Residual confounding
may arise when confounders are inadequately measured (e.g., self-reported history of
benign breast disease) or are not included (e.g., mammogram screening).
Sixth, our sample was composed of highly educated participants who are not
representative of the general Spanish population which could have limited the external
validity of our results. However, restriction to highly educated participants help to prevent
or at least to lessen confounding by known factors, such as educational level and socio-
economic status, and could increase our internal validity due to the high-quality
information given by the SUN participants. Nonetheless, our results need to be
generalized on the basis of biological mechanisms and not on the representativeness of
our sample.
Specific limitations for each research paper:

Sugar-sweetened beverage consumption and incidence of breast cancer
The use of self-reported information on SSB consumption may imply some
degree of misclassification which in turn, may bias our results towards the null.
Nevertheless, dietary information was collected with a previously validated FFQ.
Particularly, the intra-class correlation coefficient between the FFQ and multiple dietary
records for SSBs was 0.67 and in the reproducibility analyses the correlation coefficient
was 0.69313,315.
As mentioned above, our sample was not representative of the general

population. Nevertheless, it is unlikely that underlying mechanisms for the association
between sugary drinks consumption and breast cancer risk are different for highly
educated women than for women with a lower educational level.
Finally, the consumption of sugary drinks might merely represent a marker of an

overall unhealthful diet, however, we adjusted for the overall dietary pattern and the
results remained basically unchanged.
Phenolic acids intake and breast cancer risk

In the SUN Project, we did not have blood samples from the participants, so we
were unable to validate the estimated consumption of polyphenols. In relation to phenolic
acids intake, it should be highlighted that the content of polyphenols may not be uniform
153
for foods of all brands and over time. For example, the polyphenol content can vary
between different types of chocolate and coffee over time. This could have led to a non-
differential misclassification bias of the exposure that could have biased our results
towards the null.
Moreover, the lack of information on the consumption of spices and/or aromatic

herbs in the FFQ may have led to a potential underestimation of polyphenol intake.
Although they are consumed in small amounts, they provide a large number of
polyphenols. On the other hand, the way food is prepared can also condition the
bioavailability of polyphenols. Since processing is not captured by the food frequency
questionnaire, this could also lead to some degree of non-differential misclassification.
Dietary fat consumption and breast cancer risk

First, considering the method used for dietary assessment was an FFQ, some
measurement errors in total fat and subtypes of fat intake were unavoidable and remain
one of the main problems in nutritional epidemiology. Nonetheless the FFQ used was
validated313,314.
We also estimated the hypothetical effect of isocaloric substitutions of dietary fat

for another dietary component under causality of assumption.
Carbohydrate quality index and breast cancer risk

The controversy in defining whole grains, carbohydrate quality, or restricted
carbohydrate diet could make the comparison with previous results reasonably difficult.
Exposure was defined using self-reported dietary intake at baseline, and it is

possible that participants have changed their dietary habits during follow-up. Therefore,
we have lost some information on the defined CQI.
Moreover, our definition of the CQI cannot fully capture all elements of
carbohydrate intake in the context of the overall dietary pattern (e.g., pasta and rice are
refined grains, but if eaten within a Mediterranean diet may have different effects on
health outcomes than refined carbohydrates consumed in a framework of an unhealthy
dietary pattern).
Healthful and unhealthful provegetarian food patterns and breast cancer risk
The distinction between healthy and less healthy plant foods is based on existing
knowledge of the association of foods with type 2 diabetes and cardiovascular
154
disease167,169,323, as well as intermediate conditions such as obesity, hypertension, or
inflammation and not specifically with breast cancer. Nonetheless, the present methods
have been previously explored in remarkable and solid cohort studies with longer follow-
up periods and larger numbers of participants and disease cases.
Our cohort study recently showed a better conformity with a healthy

provegetarian diet associated with a reduced long-term risk of being overweight or
obese413.
Moreover, in the SUN Project, we evaluated these food patterns at a particular

timepoint, making fully capturing the relationship of a time-varying exposure, such as
diet, on the development of breast cancer difficult given the reasonably long induction
period.
Finally, the provegetarian, unhealthful provegetarian and healthful provegetarian

included dairy products, fish, seafood, and poultries, assuming they have negative
effects on breast cancer risk, even though there was no convincing evidence for these
effects140.Therefore, we acknowledge that the truthfulness of an a priori index approach
is limited by the current level of dietary knowledge concerning the diet-breast cancer
relationship, as well as uncertainties accompanying the index development. The aim
behind the provegetarian, unhealthful provegetarian and healthful provegetarian
construction was not to get the best predictive scores for breast cancer but to correctly
test their association with breast cancer risk.
Strengths
The SUN Project has also solid strengths. The main one is its prospective design,
which allowed as to assess the exposure before the disease appearance, thus avoiding
reverse causality bias commonly present in cross-sectional or in case-control studies.
The large number of participants included in the cohort (>22,000), which currently
includes 13,843 women, and its high retention (approximately 90%), which limits the risk
of incurring selection bias due to losses throughout of the study, need to be highlighted.
Recruitment of only highly educated participants follows the approach known as

restriction in epidemiology. The purpose was to control for confounding by
socioeconomic status and educational level. By doing so, restriction to university
graduates could also enhance the validity of our study, as the high educational level may
increase the accuracy of self-reports.
155
Another strong aspect of our study is the prior validation of our variables such as
weight and BMI, physical activity and the FFQ, repetitiously validated in Spain 314,315,441.
It should be noted that we have exhaustive information on the participants' diet, collected
through repeatedly validated FFQs315 which allows as to trust the validity of the nutritional
information collected.
The use of confirmed cases of breast cancer as a result variable in addition to

the information self-reported by the participants allows increasing the specificity of the
dependent variable, avoiding false positives.
Furthermore, we have collected a wide range of potential confounders which we

included in our multivariate analyses. This actually reduces the possibility of residual
confounding in our results. In addition, several of the possible confounders considered
have also been the subject of validation sub-studies.
In relation to the content of polyphenols in food, we used the Phenol-explorer

database (www.phenol-explorer.eu)358,442 to determine the intake of phenolic acids as
precisely as possible. The Phenol-Explorer database was used because it is the most
comprehensive food composition platform on polyphenols available 358.
Nurses’ Health Studies

Limitations
Several limitations should be widely recognized. In the first place, generalizability
may be limited because participants in these studies were all registered nurses and were
predominately Caucasians. However, the high educational status of these participants
should be treated as a strength because it permitted researchers to gather meticulous
and precise information on diet, lifestyle, and other health variables and minimize
confounding by socioeconomic status. Moreover, biological relations among the women
in the NHS and NHSII are unlikely to differ substantially from women in general. Further,
the distributions of most risk factors and rates of breast cancer in the NHS and NHSII
population are remarkably similar to the US population. While the prevalence of risk
factors often differs across population subgroups, many breast cancer risk factors have
been documented to operate across racial/ethnic groups, as would be expected from a
common underlying biology.
156
Second, whilst we controlled for a large variety of lifestyle factors and excluded
participants with cancer, or implausible energy intakes, the possibility of residual
confounding cannot be rule out owing to the observational nature of the study. Dietary
information was self-reported using validated FFQs, which unavoidably generate
measurement errors; yet these would presumably be non-differential in relation to risk of
breast cancer and therefore, have led to an underestimation of the associations. While
inaccuracy in self-reported diet cannot be fully avoided, the FFQs used in the current
studies have been extensively validated against diet records and biomarkers, and results
from validation studies showed good correlations341,344,345. In addition, the use of
repeated measures of diet and lifestyle in our analyses could further reduce random
measurement error and represent long-term habits340,345.
Assessment of SSBs/ASBs may be less prone to measurement error because

these beverages are relatively easy to measure. Higher SSB intake could serve as a
marker of a globally unhealthy diet and incomplete adjustment for various factors could
lead to an overestimation of the association between SSBs and breast cancer. Reverse
causation cannot be totally ruled out; however, latency analyses showed similar
associations as in the main analyses.
Third, one challenge in examining ER-negative breast cancer in epidemiologic

studies is that it only accounts for 15% to 20% of breast cancer443. In all studies, ER-
negative breast cancer cases were identified or confirmed through cancer registries, or
clinical or pathology records. Moreover, further analyses exploring molecular subtypes
beyond ER status are warranted given our limited power.
Strengths
The strengths of this study include our use of pooled data from two, large, well-
established cohorts with long-term follow-up, many repeated diet assessments, and a
large number of cases, allowed us to detect reductions and increases in risk, examine
specific a priori dietary patterns and beverages, and the availability of evaluate different
breast cancer subtypes. In NHS and NHSII, detailed data have been collected at
baseline and throughout follow-up on all major known and most hypothesized breast
cancer risk factors, including age at menarche, menopausal status, age at menopause,
parity and age at first birth, family history of breast cancer, BMI, physical activity,
smoking, alcohol consumption, and dietary intake, among others.
157
In this context, the approximate 90% response proportion of the NHS women at
each of the 20 follow-up cycles is simply unique. Overall, because participants may skip
a cycle but still remain in the study, the NHS estimates that it has retained more than
94% of its original population.
These studies have made key contributions in nutritional epidemiology research

with the creation and validation of the Harvard semiquantitative FFQ345. This improved
efficiency has enabled the NHS to create an exceptional, large database of
comprehensive, long-term, multidimensional information. Most nutritional variables are
measured by self-administered FFQs every four years that has been developed, tested
and refined by the group over the past 40 years. The validity of the FFQ has been
documented extensively by comparisons with multiple weighed diet records and
biomarkers of diet including urinary nitrogen and doubly labeled water. Development and
validation of the semiquantitative FFQ. The FFQ represents a standard tool in the
epidemiologist toolkit worlwide.
Such long follow-up allows the study and detailed analysis of the links between
long-term or life-long exposures and diseases with very long lag times. Prospective and
repeated assessment of SSB and ASB intake via validated FFQs allowed us to capture
long-term intake, reflecting true changes and minimizing the extent of measurement error
and recall bias.
The success proportion for collecting breast cancer-related medical records is

96%. Medical records are reviewed by physicians using established criteria; >99% of
reported breast cancers are confirmed upon review. Researchers code the histology,
size, grade, node status, and ER/PR/HER2 status. While routinely collected from
pathology reports, they have assessed ER status through tumor TMAs when available.
Concordance is excellent between these sources in NHS and NHSII 444. Therefore, we
were able to expand our analyses by breast cancer subtypes, which allows further
exploration of risk.
Detailed collection of updated dietary, lifestyle and medical information, namely

tissue information for the determination of molecular subtypes, over several decades
permitted the evaluation of quality of plant-based diets and the adjustment for widely
recognized confounders of these associations.
158
While residual confounding cannot be excluded, incorporation of detailed,
updated data on lifestyle factors and other potential confounders had minimal effects on
associations. Although the majority of participants were white, educated females, the
underlying biologic mechanisms were not likely to differ substantially by race or
education. Type I error is possible given we made multiple comparisons. However, the
central finding of an inverse association with healthful plant-based diets, particularly for
ER-negative cases, was our primary hypothesis, and additional analyses supported
these results.
9. Public health interest and future directions
Even Europe and the US are two major forces when it comes to research,
technology and innovation, these worldwide powers are facing many healthcare
challenges including the rising and potentially unsustainable health and care costs,
largely due to the growing prevalence of chronic diseases.
The increasing prevalence of obesity and diet-related chronic diseases is one of

the major societal challenges in Europe and in the US. As countries take on Western
food cultures, technologies, and advertising driven by a powerful economic force, chronic
diseases including breast cancer become a growing public health issue. Cancer is
currently the second leading cause of death globally, and is expected to be the first cause
of death in the 21st century445, but also the single largest obstacle to boosting life
expectancy in scores of nations. There is an emerging epidemic of obesity-related
cancers in many parts of the developed and developing world 446. Among different
obesity-related tumors, breast cancer is the most commonly diagnosed in Europe and in
the US and the worldwide leading cause of cancer deaths among women. Therefore,
prevention should be integral, not optional and solving the front end of the problem would
save more lives. Nonetheless, in government, prevention budgets are always the first
items to be cut and the last to be restored. In short, cancer must be outlined not just as
a curable disease but equitably as a preventable one.
As dietary patterns are modifiable, these results may provide viable strategies for
breast cancer primary prevention through changes in diet. Particularly, according to the
results of this thesis dissertation, it would be convenient to reinforce the message of the
importance of diet and lifestyles in the prevention of breast cancer, especially through a
healthy plant-based diet rich in polyphenols, taking into account the quality of the
159
carbohydrates consumed and lowering the consumption of sugary drinks. These
measures would result from a simple implementation, with benefits that could go beyond
breast cancer and have a positive impact on health as a global concept.
Nonetheless, in public health and in other areas, no single study can bring a
definitive answer and no study design is without its limitations. In this regard, although
the results from this thesis dissertation are promising, translating the findings into smart,
simple, scalable and sustainable recommendations for the public would require further
confirmation and more-detailed specification of dietary data. Over recent decades, the
combination of a multitude of research designs, approaches, including carefully
conducted nutritional epidemiologic studies, and measurement techniques with
complementary sets of strengths and weaknesses, has provided much new knowledge
about the important influence of diet on human health.
Observational studies and self-report diet assessment methods do have

limitations, and it is essential that we acknowledge them while moving forward to
enhance study and diet assessment methodologies, reduce measurement errors, and
leverage new technologies (e.g., omics approaches). Nevertheless, these limitations are
not an argument for reject an entire body of research as inherently and indomitably
flawed.
It is also worth emphasizing the need to identify other risk factors that allow us to
better understand the etiopathogenesis of the disease and the importance to improve
diet assessment methods and to distinguish between pre‐ and postmenopausal breast
cancer, as well as hormone receptor and molecular subtypes, because of differences in
etiology. One approach to improve exposure measurement and better understand the
mechanisms underlying diet-breast cancer relationships is via objective dietary
biomarkers. However, this area is relatively recent and few biomarkers have been
described and most remain to be validated 447–449. Therefore, these technologies are likely
to complement rather than replace self-reported dietary assessment tools450.
Combining basic science and cohort studies would also facilitate the discovery
and validation of new biomarkers. Basic science holds the key to determining how that
identified biomarker links back to the disease state and whether it is contributing to the
disease’s onset –perhaps as a risk factor– or is a consequence of the disease. Moreover,
as reproductive and lifestyle exposures evolve over time, it is crucial to include younger
160
generations of women in cohort studies, particularly, as early-life exposures may be most
relevant for breast cancer risk.
Nutritional science has often been strongly criticized for relying mainly on
observational studies instead of randomized controlled trials (RCTs), considered the gold
standard of scientific evidence in medicine. The idea that for progress to arise in nutrition
science, large, simple trials should replace most observational research is unfeasible,
and is improbable to move forward the field or improve related policies. RCTs and
objective diet assessment methods are not exempt of limitations, and if these are
minimized as being of no serious concern, it can drive to misleading conclusions. Hence,
despite the flaws, observational research is much better fitted to answer nutritional
questions relative to long-term health, particularly, when the outcome has a long etiologic
period contingent to dietary exposures and the nutritional comparison of interest would
be practically or ethically infeasible in an interventional setting (i.e., SSBs), observational
studies can be an inestimable resource.
To further progress in nutrition research, we need multiple lines of evidence. More

funding for well-designed RCTs will indeed bolster the field, but RCTs are not the
panacea. In fact, it should be necessary to have strong observational evidence, and
highlight the etiologically relevant periods during which detailed ranges of exposure
contrasts may influence a disease onset. If correctly designed, they can have a huge
potential to inform prevention strategies.
Overall, modifications in public policies in the face of mounting evidence from

well-conducted observational studies such as the SUN Project or the NHS/NHSII will
play a great role in framing future policies for the prevention of breast cancer. We are
facing a scenario where multiple lines of evidence, including carefully conducted
prospective cohort studies with repeated measures are necessary to further confirm
these findings. Translating current nutritional knowledge into policies and public health
practice continues to be a top research priority.
A cancer prevented is even better than a cancer cured, thus, when cancer
becomes our leading cause of death, cancer prevention will become our leading cause
of improved life-expectancy. In order to make a dent in a public health sense, we must
prevent cancer.
161
CONCLUSIONS
162
1. In the SUN Project, SSB consumption was not associated with overall breast cancer
incidence. Nonetheless, we observed a higher risk of postmenopausal breast cancer
among women with SSB consumption over 1 serving/month compared to those who
never or seldom consumed SSBs.
2. In the Nurses’ Health Studies, SSB consumption was not associated with overall
breast cancer incidence. Nevertheless, we observed a suggestive interaction by BMI,
where a modestly higher incidence of breast cancer with each serving per day
increment of SSBs was found in normal weight women. Our findings also suggest no
substantial increase in risk of breast cancer with consumption of ASBs.
3. In the SUN Project, high consumption of hydroxycinnamic acids and chlorogenic

acids, present in coffee, fruits, and vegetables, were associated with lower breast
cancer incidence among postmenopausal women.
4. In the same Mediterranean cohort study, a better quality of dietary carbohydrate

intake, considering jointly the dietary fiber intake, the glycemic index, the whole grain:
total grain carbohydrates ratio and the solid carbohydrate: total carbohydrate ratio,
showed a significant inverse association with the incidence of breast cancer,
especially among premenopausal women.
5. In the SUN cohort study, total fat, monounsaturated, polyunsaturated or saturated fat
intake was not associated with overall breast cancer incidence.
6. A moderate adherence to a provegetarian food pattern was associated with a relative

reduction in the risk of breast cancer of 45% in the SUN Project, which may represent
not an entirely vegetarian diet but rather a pattern emphasizing plant-based foods
and some animal foods.
7. Furthermore, we did not find significant associations between healthful or unhealthful

provegetarian food patterns and breast cancer risk in the SUN Project.
8. In the Nurses’ Health Studies, adherence to a plant-based dietary pattern, which

assigns positive scores to plant foods and reverse scores to animal foods, may
reduce the risk of breast cancer.
163
9. A greater adherence to a healthful plant-based dietary pattern is associated with
lower risk of ER-negative breast cancer. Contrarily, a low-quality plant-based diet
may be associated with an increased risk of ER-negative breast cancer in the Nurses’
Health Studies.
164
CONCLUSIONES
165
1. En el Proyecto SUN, un mayor consumo de bebidas azucaradas no se asoció
de forma estadísticamente significativa con una mayor incidencia de cáncer de
mama. Sin embargo, observamos un aumento relativo en el riesgo de cáncer de
mama postmenopáusico con un consumo de bebidas azucaradas de al menos
una vez al mes en comparación con aquellas mujeres que nunca o casi nunca
consumían estas bebidas.
2. En los Nurses’ Health Studies, un mayor consumo de bebidas azucaradas no se

asoció de forma estadísticamente significativa con una mayor incidencia de
cáncer de mama. No obstante, observamos una posible interacción con el índice
de masa corporal, donde se observó una incidencia moderadamente más alta
de cáncer de mama asociada a cada incremento en una porción de bebidas
azucaradas al día en mujeres con peso normal. Por otro lado, no encontramos
un aumento significativo en el riesgo de cáncer de mama con el consumo de
bebidas artificialmente edulcoradas.
3. En el Proyecto SUN, un mayor consumo de ácidos hidroxicinámicos y ácidos

clorogénicos, presentes en café, frutas y verduras, se asoció con una menor
incidencia de cáncer de mama entre las mujeres postmenopáusicas.
4. En el Proyecto SUN, una mayor calidad de los carbohidratos consumidos en la

dieta, considerando conjuntamente la ingesta de fibra dietética, el índice
glucémico, el ratio de carbohidratos de grano entero: total de granos y el ratio de
carbohidratos sólidos: carbohidratos totales, se asoció de forma significativa con
una menor incidencia de cáncer de mama, especialmente entre mujeres
premenopáusicas.
5. En el Proyecto SUN, no se observó una asociación estadísticamente significativa

entre una mayor ingesta total de grasas o de sus subtipos (monoinsaturadas,
poliinsaturadas o saturadas) y la incidencia de cáncer de mama.
6. Una adhesión moderada a un patrón de alimentación provegetariano, que no

representa una dieta completamente vegetariana sino más bien un patrón que
enfatiza los alimentos de origen vegetal y algunos alimentos de origen animal,
se asoció con una reducción relativa en el riesgo de cáncer de mama del 45%
en el Proyecto SUN.
166
7. No se encontró ninguna asociación estadísticamente significativa entre el grado
de adhesión a patrones provegetarianos saludables o no saludables y el riesgo
de cáncer de mama en el Proyecto SUN.
8. En los Nurses’ Health Studies, una mayor adhesión a un patrón dietético basado
en alimentos de origen vegetal se relacionó con una reducción relativa del riesgo
de cáncer de mama del 11%.
9. Una mayor adhesión a un patrón dietético basado en alimentos de origen vegetal

saludable se asoció con una menor incidencia de cáncer de mama con
receptores de estrógenos negativos. Por el contrario, una dieta a base de
alimentos de origen vegetal de baja calidad se asoció con una mayor incidencia
de cáncer de mama con receptores de estrógenos negativos en los Nurses’
Health Studies.
167
BIBLIOGRAPHY
168
1. Rivard AB, Galarza-Paez L, Peterson DC. Anatomy, Thorax, Breast. In: StatPearls.
Treasure Island (FL): StatPearls Publishing; 2021. Available from:
https://www.ncbi.nlm.nih.gov/books/NBK519575/
2. Gatt A, Agarwal S, Zito PM. Anatomy, Fascia Layers. Treasure Island (FL): StatPearls
Publishing; 2021. Available from: https://www.ncbi.nlm.nih.gov/books/NBK526038/
3. Khan YS, Sajjad H. Anatomy, Thorax, Mammary Gland. In: StatPearls Treasure Island
(FL): StatPearls Publishing; 2021. Available from:
https://www.ncbi.nlm.nih.gov/books/NBK547666/
4. Harbeck N, Penault-Llorca F, Cortes J, et al. Breast cancer. Nat Rev Dis Primers. 2019;
5:66. doi:10.1038/s41572-019-0111-2
5. Seidel, H.M., Ball, J.W., Davis, J.E., et al. Mosby's Guide to Physical Examination, 7th
edn. New York: Mosby Elsevier, 2011.
6. Pandya S, Moore RG. Breast development and anatomy. Clin Obstet Gynecol. 2011;
54:91-95. doi:10.1097/GRF.0b013e318207ffe9
7. Bland, KI, Copeland, EM, Klimberg, VS & Gradishar, WJ 2017, The Breast:
Comprehensive Management of Benign and Malignant Diseases. Elsevier Inc.
doi.org/10.1016/C2014-0-01946-6
8. Van Deventer PV. The blood supply to the nipple-areola complex of the human
mammary gland. Aesthetic Plast Surg. 2004; 28:393-398. doi:10.1007/s00266-003-
7113-9
9. Sun YS, Zhao Z, Yang ZN, et al. Risk Factors and Preventions of Breast Cancer. Int J
Biol Sci. 2017; 13:1387-1397. doi:10.7150/ijbs.21635
10. Mittal S, Brown NJ, Holen I. The breast tumor microenvironment: role in cancer
development, progression and response to therapy. Expert Rev Mol Diagn. 2018;
18:227-243. doi:10.1080/14737159.2018.1439382
11. Polyak K, Kalluri R. The role of the microenvironment in mammary gland development
and cancer. Cold Spring Harb Perspect Biol. 2010; 2:a003244.
doi:10.1101/cshperspect.a003244
12. Feng Y, Spezia M, Huang S, et al. Breast cancer development and progression: Risk
factors, cancer stem cells, signaling pathways, genomics, and molecular pathogenesis.
Genes Dis. 2018; 5:77-106. doi:10.1016/j.gendis.2018.05.001
13. Slepicka PF, Somasundara AVH, Dos Santos CO. The molecular basis of mammary
gland development and epithelial differentiation. Seminars in Cell & Developmental
Biology. 2020. doi: 10.1016/j.semcdb.2020.09.014.
14. Phi LTH, Sari IN, Yang YG, et al. Cancer Stem Cells (CSCs) in Drug Resistance and
their Therapeutic Implications in Cancer Treatment. Stem Cells Int. 2018:5416923.
doi:10.1155/2018/5416923
15. Lloyd-Lewis B, Harris OB, Watson CJ, Davis FM. Mammary Stem Cells: Premise,
Properties, and Perspectives. Trends Cell Biol. 2017;8,556-67.
doi:10.1016/j.tcb.2017.04.001
169
16. Soteriou D, Fuchs Y. A matter of life and death: stem cell survival in tissue regeneration
and tumour formation. Nat Rev Cancer. 2018; 18:187-201. doi:10.1038/nrc.2017.122
17. Sever R, Brugge JS. Signal transduction in cancer. Cold Spring Harb Perspect Med.
2015; 5:a006098. doi:10.1101/cshperspect.a006098
18. Kumar V, Green S, Stack G, Berry M, Jin JR, Chambon P. Functional domains of the
human estrogen receptor. Cell. 1987; 51:941–51. doi:10.1016/0092-8674(87)90581-2
19. Nadji M, Gomez-Fernandez C, Ganjei-Azar P, Morales AR. Immunohistochemistry of
estrogen and progesterone receptors reconsidered: experience with 5,993 breast
cancers. Am J Clin Pathol. 2005; 123:21-27. doi:10.1309/4wv79n2ghj3x1841
20. Zwijsen RLM, Wientjens E, Klompmaker R, van der Sman J, Bernards R, Michalides
RJAM. Cdk-independent activation of estrogen receptor by cyclin
D1. Cell. 1997;88:405–415
21. Lundberg AS, Weinberg RA. Control of the cell cycle and apoptosis. Eur J Cancer.
1999; 35:1886-1894. doi:10.1016/s0959-8049(99)00292-0
22. Huang B, Omoto Y, Iwase H, et al. Differential expression of estrogen receptor α, β1,
and β2 in lobular and ductal breast cancer. Proc Natl Acad Sci U S A. 2014; 111:1933-
1938. doi:10.1073/pnas.1323719111
23. Paruthiyil S, Parmar H, Kerekatte V, Cunha GR, Firestone GL, Leitman DC. Estrogen
receptor beta inhibits human breast cancer cell proliferation and tumor formation by
causing a G2 cell cycle arrest. Cancer Res. 2004; 64:423-428. doi:10.1158/0008-
5472.can-03-2446
24. Lin CY, Ström A, Li Kong S, et al. Inhibitory effects of estrogen receptor beta on specific
hormone-responsive gene expression and association with disease outcome in primary
breast cancer. Breast Cancer Res. 2007; 9:R25. doi:10.1186/bcr1667
25. Bado, I., Nikolos, F., Rajapaksa, G. et al. Somatic loss of estrogen receptor beta and
p53 synergize to induce breast tumorigenesis. Breast Cancer Res 2017; 19,79.
doi.org/10.1186/s13058-017-0872-z
26. Saha Roy S, Vadlamudi RK. Role of estrogen receptor signaling in breast cancer
metastasis. Int J Breast Cancer. 2012; 2012:654698. doi:10.1155/2012/654698
27. Wee P, Wang Z. Epidermal Growth Factor Receptor Cell Proliferation Signaling
Pathways. Cancers (Basel). 2017; 9:52. doi: 10.3390/cancers9050052
28. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL. Human breast
cancer: correlation of relapse and survival with amplification of the HER-2/neu
oncogene. Science. 1987; 235:177-182. doi:10.1126/science.3798106
29. Jacobi N, Seeboeck R, Hofmann E, Eger A. ErbB Family Signalling: A Paradigm for
Oncogene Addiction and Personalized Oncology. Cancers (Basel). 2017; 9:33.
doi:10.3390/cancers9040033
30. Kruger DT, Opdam M, Sanders J, van der Noort V, Boven E, Linn SC. Hierarchical
clustering of PI3K and MAPK pathway proteins in breast cancer intrinsic
subtypes. APMIS. 2020; 128:298-307. doi:10.1111/apm.13026
170
31. Oda K, Matsuoka Y, Funahashi A, Kitano H. A comprehensive pathway map of
epidermal growth factor receptor signaling. Mol Syst Biol. 2005; 1:2005.0010.
doi:10.1038/msb4100014
32. Yang Y, Leonard M, Zhang Y, et al. HER2-Driven Breast Tumorigenesis Relies upon
Interactions of the Estrogen Receptor with Coactivator MED1. Cancer Res. 2018;
78:422-435. doi:10.1158/0008-5472.CAN-17-1533
33. Liu S, Lee JS, Jie C, et al. HER2 overexpression triggers an IL1a proinflammatory circuit
to drive tumorigenesis and promote chemotherapy resistance. Cancer Res
2018;8,2040-51. doi:10.1158/0008-5472.CAN-17-2761
34. Clevers H, Nusse R. Wnt/β-catenin signaling and disease. Cell. 2012;6,1192-1205
doi:10.1016/j.cell.2012.05.012
35. Xu X, Zhang M, Xu F, Jiang S. Wnt signaling in breast cancer: biological mechanisms,
challenges and opportunities. Mol Cancer. 2020;19:165. doi:10.1186/s12943-020-
01276-5
36. Nurse, P.M. Cyclin Dependent Kinases and Cell Cycle Control. Biosci Rep. 2002; 22,
487–499. doi.org/10.1023/A:1022017701871
37. Sutherland RL, Musgrove EA. Cyclins and breast cancer. J Mammary Gland Biol
Neoplasia. 2004; 9:95–104. doi: 10.1023/B:JOMG.0000023591.45568.77
38. Radtke F, Raj K. The role of Notch in tumorigenesis: oncogene or tumour
suppressor?. Nat Rev Cancer. 2003; 3:756-767. doi:10.1038/nrc1186
39. Sansone P, Storci G, Giovannini C, et al. p66Shc/Notch-3 interplay controls self-renewal
and hypoxia survival in human stem/progenitor cells of the mammary gland expanded
in vitro as mammospheres. Stem Cells. 2007; 25:807-815. doi:10.1634/stemcells.2006-
0442
40. Jiang J, Hui C chung. Hedgehog Signaling in Development and Cancer. Dev Cell. 2008;
6, 801-812. doi:10.1016/j.devcel.2008.11.010
41. García-Zaragoza E, Pérez-Tavarez R, Ballester A, et al. Intraepithelial paracrine
Hedgehog signaling induces the expansion of ciliated cells that express diverse
progenitor cell markers in the basal epithelium of the mouse mammary gland. Dev Biol.
2012; 372:28-44. doi:10.1016/j.ydbio.2012.09.005
42. Jeng KS, Sheen IS, Jeng WJ, Yu MC, Hsiau HI, Chang FY. High expression of Sonic
Hedgehog signaling pathway genes indicates a risk of recurrence of breast
carcinoma. Onco Targets Ther. 2013; 7:79-86. doi:10.2147/OTT.S54702
43. Mitchell PJ, Barker KT, Martindale JE, et al. Cloning and characterisation of cDNAs
encoding a novel non-receptor tyrosine kinase, brk, expressed in human breast
tumours. Oncogene. 1994; 9:2383-2390.
44. Barker KT, Jackson LE, Crompton MR. BRK tyrosine kinase expression in a high
proportion of human breast carcinomas. Oncogene. 1997; 15:799-805.
doi:10.1038/sj.onc.1201241
171
45. Aubele M, Auer G, Walch AK, et al. PTK (protein tyrosine kinase)-6 and HER2 and 4,
but not HER1 and 3 predict long-term survival in breast carcinomas. Br J Cancer. 2007;
96:801-807. doi:10.1038/sj.bjc.6603613
46. Koren S, Reavie L, Couto JP, et al. PIK3CA (H1047R) induces multipotency and multi-
lineage mammary tumours. Nature. 2015; 525:114-118. doi:10.1038/nature14669
47. Allison KH. Molecular pathology of breast cancer: what a pathologist needs to know. Am
J Clin Pathol. 2012; 138:770-780. doi:10.1309/AJCPIV9IQ1MRQMOO
48. Polyak K. Breast cancer: Origins and evolution. J Clin Invest. 2007.117:3155-3163.
doi:10.1172/JCI33295
49. Colditz GA, Kaphingst KA, Hankinson SE, Rosner B. Family history and risk of breast
cancer: nurses' health study. Breast Cancer Res Treat. 2012; 133:1097-1104.
doi:10.1007/s10549-012-1985-9
50. Brenton JD, Carey LA, Ahmed A, Caldas C. Molecular classification and molecular
forecasting of breast cancer: Ready for clinical application? J Clin Oncol. 2005: 29,
7350-60. doi:10.1200/JCO.2005.03.3845
51. Nielsen TO, Hsu FD, Jensen K, et al. Immunohistochemical and clinical characterization
of the basal-like subtype of invasive breast carcinoma. Clin Cancer Res. 2004;16, 5367-
5374. doi:10.1158/1078-0432.CCR-04-0220
52. Abd El-Rehim DM, Pinder SE, Paish CE, et al. Expression of luminal and basal
cytokeratins in human breast carcinoma. J Pathol. 2004; 203:661-671.
doi:10.1002/path.1559
53. Abd El-Rehim DM, Ball G, Pinder SE, et al. High-throughput protein expression analysis
using tissue microarray technology of a large well-characterised series identifies
biologically distinct classes of breast cancer confirming recent cDNA expression
analyses. Int J Cancer. 2005; 116:340-350. doi:10.1002/ijc.21004
54. Pusztai L, Broglio K, Andre F, Fraser Symmans W, Hess KR, Hortobagyi GN. Effect of
molecular disease subsets on disease-free survival in randomized adjuvant
chemotherapy trials for estrogen receptor-positive breast cancer. J Clin Oncol.2008; 26:
4679-83. doi:10.1200/JCO.2008.17.2544
55. Sørlie T, Tibshirani R, Parker J, et al. Repeated observation of breast tumor subtypes in
independent gene expression data sets. Proc Natl Acad Sci U S A. 2003; 100: 8418-23.
doi:10.1073/pnas.0932692100
56. Dai X, Li T, Bai Z, et al. Breast cancer intrinsic subtype classification, clinical use and
future trends. Am J Cancer Res. 2015; 5:2929-2943.
57. Sørlie T, Perou CM, Tibshirani R, et al. Gene expression patterns of breast carcinomas
distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A.
2001;98:10869-74. doi:10.1073/pnas.191367098
58. Perou CM, Sørile T, Eisen MB, Van De Rijn M, Jeffrey SS, Ress CA, et al. Molecular
portraits of human breast tumours. Nature 2000;406:747-752. doi:10.1038/35021093
172
59. Giuliano AE, Edge SB, Hortobagyi GN. Eighth Edition of the AJCC Cancer Staging
Manual: Breast Cancer. Ann Surg Oncol. 2018; 25:1783-1785. doi:10.1245/s10434-
018-6486-6
60. Ferlay J, Laversanne M, Ervik M, Lam F, Colombet M, Mery L, Piñeros M, Znaor A,
Soerjomataram I, Bray F (2020). Global Cancer Observatory: Cancer Tomorrow. Lyon,
France: International Agency for Research on Cancer. Available from:
https://gco.iarc.fr/tomorrow Accessed February 25, 2021.
61. Cancer of the Breast (Female) - Cancer Stat Facts. SEER.
https://seer.cancer.gov/statfacts/html/breast.html Accessed February 25, 2021.
62. Howlader N, Noone AM, Krapcho M, et al. SEER Cancer Statistics Review, 1975-2017.
Table 4.18. Cancer of the Female Breast (Invasive): Age-Adjusted Rates and Trends by
Race/Ethnicity, 2013-2017. National Cancer Institute.; 2019.
http://seer.cancer.gov/csr/1975_2017/ Accessed February 25, 2021.
63. American Cancer Society. Breast Cancer Facts & Figures 2019-2020. Atlanta: American
Cancer Society, Inc. 2019
64. Xue F, Michels KB. Intrauterine factors and risk of breast cancer: a systematic review
and meta-analysis of current evidence Lancet Oncol. 2007; 8:1088-1100.
doi:10.1016/S1470-2045(07)70377-7
65. Park, S.K., Kang, D., McGlynn, K.A. et al. Intrauterine environments and breast cancer
risk: meta-analysis and systematic review. Breast Cancer Res 2008; 10, R8.
doi.org/10.1186/bcr1850
66. Fei Xue, Bernard Rosner, Heather Eliassen, Karin B Michels, Body fatness throughout
the life course and the incidence of premenopausal breast cancer, International Journal
of Epidemiology.2016; 4:1103–12. doi:10.1093/ije/dyw149
67. Yang TO, Reeves GK, Green J, Beral V, Cairns BJ; Million Women Study Collaborators.
Birth weight and adult cancer incidence: large prospective study and meta-
analysis. Ann Oncol. 2014; 25:1836-1843. doi:10.1093/annonc/mdu214
68. Tworoger SS, Eliassen AH, Missmer SA, et al. Birthweight and body size throughout life
in relation to sex hormones and prolactin concentrations in premenopausal
women. Cancer Epidemiol Biomarkers Prev. 2006; 15:2494-2501. doi: 10.1158/1055-
9965.EPI-06-0671
69. Harris JR, Lippman ME, Morrow M, Kent Osborne C. Diseases of the breast: Fifth
edition. Wolters Kluwer Health Adis (ESP), 2014. 1224.
70. Hamajima N, Hirose K, Tajima K, Rohan T, Friedenreich CM, Calle EE et al. Menarche,
menopause, and breast cancer risk: Individual participant meta-analysis, including 118
964 women with breast cancer from 117 epidemiological studies. Lancet Oncol. 2012;
13:1141-51. doi.org/10.1016/S1470-2045(12)70425-4
71. Ritte R, Tikk K, Lukanova A, et al. Reproductive factors and risk of hormone receptor
positive and negative breast cancer: a cohort study. BMC Cancer. 2013; 13:584.
doi:10.1186/1471-2407-13-584
173
72. Endogenous Hormones and Breast Cancer Collaborative Group, Key TJ, Appleby PN,
et al. Circulating sex hormones and breast cancer risk factors in postmenopausal
women: reanalysis of 13 studies. Br J Cancer. 2011; 105:709-722.
doi:10.1038/bjc.2011.254
73. Biro FM, Lucky AW, Simbartl LA, et al. Pubertal maturation in girls and the relationship
to anthropometric changes: pathways through puberty. J Pediatr. 2003; 142:643-646.
doi:10.1067/mpd.2003.244
74. Bodicoat DH, Schoemaker MJ, Jones ME, et al. Timing of pubertal stages and breast
cancer risk: the Breakthrough Generations Study. Breast Cancer Res. 2014; 16: R18.
doi:10.1186/bcr3613
75. Kaplowitz P. Pubertal development in girls: Secular trends. Curr Opin Obstet Gynecol.
2006; 5: 487-91. doi: 10.1097/01.gco.0000242949.02373.09
76. Mahabir S, Aagaard K, Anderson LM, et al. Challenges and opportunities in research on
early-life events/exposures and cancer development later in life. Cancer Causes
Control. 2012; 23:983-990. doi:10.1007/s10552-012-9962-5
77. McCormack VA, dos Santos Silva I. Breast density and parenchymal patterns as
markers of breast cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev.
2006;15:1159-1169. doi:10.1158/1055-9965.EPI-06-0034
78. Engmann NJ, Golmakani MK, Miglioretti DL, Sprague BL, Kerlikowske K, for the Breast
Cancer Surveillance Consortium. Population-Attributable Risk Proportion of Clinical
Risk Factors for Breast Cancer. JAMA Oncol. 2017; 3:1228–1236.
doi:10.1001/jamaoncol.2016.6326
79. Willett WC, Tamimi R, Hankinson SE, Hazra A, Eliassen AH CG. Chapter 18:
Nongenetic Factors in the Causation of Breast Cancer, in Harris JR, Lippman ME,
Morrow M, Osborne CK. Diseases of the Breast, 5th Edition, Lippincott Williams &
Wilkins, 2014.; 2014.
80. Zhang B, Shu XO, Delahanty RJ, et al. Height and breast cancer risk: Evidence from
prospective studies and mendelian randomization. J Natl Cancer Inst.
2015;107:djv219. doi: 10.1093/jnci/djv219. Print 2015 Nov.
81. Guo W, Key TJ, Reeves GK. Adiposity and breast cancer risk in postmenopausal
women: Results from the UK Biobank prospective cohort. Int J Cancer. 2018;143:1037-
1046. doi:10.1002/ijc.31394
82. Albanes D. Height, early energy intake, and cancer. Evidence mounts for the relation of
energy intake to adult malignancies. BMJ. 1998; 317:1331–1332. doi:
10.1136/bmj.317.7169.1331.
83. Green J, Cairns BJ, Casabonne D, et al. Height and cancer incidence in the Million
Women Study: prospective cohort, and meta-analysis of prospective studies of height
and total cancer risk. Lancet Oncol. 2011; 12:785-794. doi:10.1016/S1470-
174
2045(11)70154-1
84. Barnard M.E. Boeke C.E. Tamimi R.M. Established breast cancer risk factors and risk of
intrinsic tumor subtypes. Biochim Biophys Acta. 2015; 1856: 73-85.
doi.org/10.1016/j.bbcan.2015.06.002
85. Collaborative Group on Hormonal Factors in Breast Cancer CG on HF in B, Cancer CG
on HF in B, Cancer CG on HF in B, et al. Breast cancer and breastfeeding:
collaborative reanalysis of individual data from 47 epidemiological studies in 30
countries, including 50302 women with breast cancer and 96973 women without the
disease. Lancet (London, England). 2002;360:187-195. doi:10.1016/S0140-
6736(02)09454-0
86. Economopoulou P, Dimitriadis G, Psyrri A. Beyond BRCA: New hereditary breast
cancer susceptibility genes. Cancer Treat Rev. 2015;41:1–8. doi:
10.1016/j.ctrv.2014.10.008
87. Kuchenbaecker KB, Hopper JL, Barnes DR, et al. Risks of breast, ovarian, and
contralateral breast cancer for BRCA1 and BRCA2 mutation carriers. JAMA.
2017;317:2402-2416. doi:10.1001/jama.2017.7112
88. Breast Cancer - Risk Factors and Prevention. Cancer.Net.
https://www.cancer.net/cancer-types/breast-cancer/risk-factors-and-prevention
Accessed February 25, 2021.
89. Van Zee KJ, White J, Morrow M HJ. Chapter 23: Ductal carcinoma in situ and
microinvasive carcinoma, in Harris JR, Lippman ME, Morrow M, Osborne CK. In:
Diseases of the Breast: Fifth Edition. Lippincott. ; 2014.
90. Hodgson DC, Gilbert ES, Dores GM, et al. Long-term solid cancer risk among 5-year
survivors of Hodgkin's lymphoma. J Clin Oncol. 2007;25:1489–1497.
doi:10.1200/JCO.2006.09.0936
91. Ibrahim EM, Abouelkhair KM, Kazkaz GA, Elmasri OA, Al-Foheidi M. Risk of second
breast cancer in female Hodgkin's lymphoma survivors: a meta-analysis. BMC
Cancer. 2012; 12:197. doi:1471-2407-12-197
92. Moskowitz CS, Chou JF, Wolden SL, et al. Breast cancer after chest radiation therapy
for childhood cancer. J Clin Oncol. 2014; 32:2217-2223.
doi:10.1200/JCO.2013.54.4601
93. Schaapveld M, Aleman BM, van Eggermond AM, et al. Second Cancer Risk Up to 40
Years after Treatment for Hodgkin's Lymphoma. N Engl J Med. 2015; 373:2499-2511.
doi:10.1056/NEJMoa1505949
94. Sud A, Thomsen H, Sundquist K, Houlston RS, Hemminki K. Risk of second cancer in
Hodgkin lymphoma survivors and influence of family history. J Clin Oncol. 2017;14:
1584-1590.doi:10.1200/JCO.2016.70.9709
95. Bright CJ, Reulen RC, Winter DL, et al. Risk of subsequent primary neoplasms in
survivors of adolescent and young adult cancer (Teenage and Young Adult Cancer
175
Survivor Study): a population-based, cohort study. Lancet Oncol. 2019; 4:531-545.
doi:10.1016/S1470-2045(18)30903-3
96. PDQ Cancer Genetics Editorial Board. Genetics of Breast and Gynecologic Cancers
(PDQ®): Health Professional Version. 2021 Feb 12. In: PDQ Cancer Information
Summaries. Bethesda (MD): National Cancer Institute (US); 2002–. PMID: 26389210.
97. Koivisto-Korander R, Scélo G, Ferro G, et al. Second primary malignancies among
women with uterine sarcoma. Gynecol Oncol. 2012; 126:30-35. doi:
10.1016/j.ygyno.2012.04.002
98. Chlebowski RT, Chen Z, Anderson GL, et al. Ethnicity and breast cancer: factors
influencing differences in incidence and outcome. J Natl Cancer Inst. 2005; 97:439-448.
doi:10.1093/jnci/dji064
99. Krieger N, Chen JT, Waterman PD. Decline in US breast cancer rates after the
Women's Health Initiative: socioeconomic and racial/ethnic differentials. Am J Public
Health. 2010;100: S132-S139. doi:10.2105/AJPH.2009.181628
100. Warner ET, Tamimi RM, Boggs DA, et al. Estrogen receptor positive tumors: do
reproductive factors explain differences in incidence between black and white
women?. Cancer Causes Control. 2013; 24:731-739. doi:10.1007/s10552-013-0153-9
101. Xiang AH, Chow T, Mora-Marquez J, et al. Breastfeeding Persistence at 6 Months:
Trends and Disparities from 2008 to 2015. J Pediatr. 2019; 208:169-175.e2. doi:
10.1016/j.jpeds.2018.12.055
102. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from
51 epidemiological studies of 52,705 women with breast cancer and 108,411 women
without breast cancer. Collaborative Group on Hormonal Factors in Breast Cancer.
Lancet. 1997; 350:1047-1059.
103. Li, K., Anderson, G., Viallon, V. et al. Risk prediction for estrogen receptor-specific
breast cancers in two large prospective cohorts. Breast Cancer Res 2018; 20: 147.
doi.org/10.1186/s13058-018-1073-0
104. Veiga LH, Curtis RE, Morton LM, et al. Association of Breast Cancer Risk After
Childhood Cancer with Radiation Dose to the Breast and Anthracycline Use: A Report
from the Childhood Cancer Survivor Study. JAMA Pediatr. 2019; 173:1171–1179.
doi:10.1001/jamapediatrics.2019.3807
105. Zeeb H, Blettner M, Langner I, et al. Mortality from cancer and other causes among
airline cabin attendants in Europe: a collaborative cohort study in eight countries.
American Journal of Epidemiology. 2003; 158:35-46. doi: 10.1093/aje/kwg107.
106. Rushton L, Bagga S, Bevan R, et al. Occupation and cancer in Britain. Br J Cancer.
2010; 102:1428-1437. doi:10.1038/sj.bjc.6605637
107. Pukkala E, Helminen M, Haldorsen T, et al. Cancer incidence among Nordic airline
cabin crew. Int J Cancer. 2012; 131:2886-2897. doi:10.1002/ijc.27551
176
108. Jia Y, Lu Y, Wu K, et al. Does night work increase the risk of breast cancer? A
systematic review and meta-analysis of epidemiological studies. Cancer Epidemiol.
2013; 37:197-206. doi: 10.1016/j.canep.2013.01.005
109. Wegrzyn LR, Tamimi RM, Rosner BA, et al. Rotating Night-Shift Work and the Risk of
Breast Cancer in the Nurses' Health Studies. Am J Epidemiol. 2017; 186:532-540.
doi:10.1093/aje/kwx140
110. Yuan X, Zhu C, Wang M, Mo F, Du W, Ma X. Night Shift Work Increases the Risks of
Multiple Primary Cancers in Women: A Systematic Review and Meta-analysis of 61
Articles. Cancer Epidemiol Biomarkers Prev. 2018; 27:25-40. doi: 10.1158/1055-
9965.EPI-17-0221
111. Van Den Brandt PA, Spiegelman D, Yaun SS, et al. Pooled analysis of prospective
cohort studies on height, weight, and breast cancer risk. Am J Epidemiol.
2000;152:514–27. doi:10.1093/aje/152.6.514
112. Reeves GK, Pirie K, Beral V, et al. Cancer incidence and mortality in relation to body
mass index in the Million Women Study: cohort study. BMJ. 2007; 335:1134.
doi:10.1136/bmj.39367.495995.AE
113. The Premenopausal Breast Cancer Collaborative Group. Association of Body Mass
Index and Age With Subsequent Breast Cancer Risk in Premenopausal Women. JAMA
Oncol. 2018; 4: e181771. doi:10.1001/jamaoncol.2018.1771
114. Huang Z, Hankinson SE, Colditz GA, et al. Dual effects of weight and weight gain on
breast cancer risk. JAMA. 1997; 278:1407-1411.
doi:10.1001/jama.1997.03550170037029
115. Hvidtfeldt UA, Gunter MJ, Lange T, et al. Quantifying mediating effects of endogenous
estrogen and insulin in the relation between obesity, alcohol consumption, and breast
cancer. Cancer Epidemiology, Biomarkers & Prevention : a Publication of the American
Association for Cancer Research, Cosponsored by the American Society of Preventive
Oncology. 2012;21:1203-1212. doi: 10.1158/1055-9965.epi-12-0310.
116. Michels KB, Solomon CG, Hu FB, et al. Type 2 diabetes and subsequent incidence of
breast cancer in the nurses’ health study. Diabetes Care. 2003. 26: 1752-1758.
doi:10.2337/diacare.26.6.1752
117. Boyle P, Boniol M, Koechlin A, et al. Diabetes and breast cancer risk: a meta-analysis.
Br J Cancer. 2012;107:1608-1617. doi:10.1038/bjc.2012.414
118. De Bruijn KM, Arends LR, Hansen BE, Leeflang S, Ruiter R, van Eijck CH. Systematic
review and meta-analysis of the association between diabetes mellitus and incidence
and mortality in breast and colorectal cancer. Br J Surg. 2013;100:1421-1429.
doi:10.1002/bjs.9229
119. Eliassen AH, Tworoger SS, Mantzoros CS, Pollak MN, Hankinson SE. Circulating
insulin and c-peptide levels and risk of breast cancer among predominately
premenopausal women. Cancer Epidemiol Biomarkers Prev. 2007;16:161-164.
doi:10.1158/1055-9965.EPI-06-0693
177
120. Eliassen AH, Colditz GA, Rosner B, Willett WC, Hankinson SE. Adult weight change
and risk of postmenopausal breast cancer. J Am Med Assoc. 2006; 296(2):193–201.
doi:10.1001/jama.296.2.193
121. Vrieling A, Buck K, Kaaks R, Chang-Claude J. Adult weight gain in relation to breast
cancer risk by estrogen and progesterone receptor status: a meta-analysis. Breast
Cancer Res Treat. 2010; 123:641-649. doi:10.1007/s10549-010-1116-4
122. Emaus MJ, van Gils CH, Bakker MF, et al. Weight change in middle adulthood and
breast cancer risk in the EPIC-PANACEA study. Int J Cancer. 2014; 135:2887-2899.
doi:10.1002/ijc.28926
123. Chlebowski RT, Luo J, Anderson GL, et al. Weight loss and breast cancer incidence in
postmenopausal women. Cancer. 2019; 125:205-212. doi:10.1002/cncr.31687
124. Teras LR, Patel AV, Wang M, et al. Sustained Weight Loss and Risk of Breast Cancer
in Women 50 Years and Older: A Pooled Analysis of Prospective Data. J Natl Cancer
Inst. 2020; 112:929-937. doi:10.1093/jnci/djz226
125. Arnold M, Jiang L, Stefanick ML, et al. Duration of Adulthood Overweight, Obesity, and
Cancer Risk in the Women's Health Initiative: A Longitudinal Study from the United
States. PLoS Med. 2016; 13:e1002081. doi:10.1371/journal.pmed.1002081
126. Neuhouser ML, Aragaki AK, Prentice RL, et al. Overweight, Obesity, and
Postmenopausal Invasive Breast Cancer Risk: A Secondary Analysis of the Women's
Health Initiative Randomized Clinical Trials. JAMA Oncol. 2015; 1:611-621.
doi:10.1001/jamaoncol.2015.1546
127. Michels KB, Terry KL, Eliassen AH, Hankinson SE, Willett WC. Adult weight change
and incidence of premenopausal breast cancer. Int J Cancer. 2012; 130:902-909.
doi:10.1002/ijc.26069
128. Harris HR, Willett WC, Terry KL, Michels KB. Body fat distribution and risk of
premenopausal breast cancer in the Nurses' Health Study II. J Natl Cancer Inst. 2011;
103:273-278. doi:10.1093/jnci/djq500
129. Kabat GC, Xue X, Kamensky V, et al. Risk of breast, endometrial, colorectal, and renal
cancers in postmenopausal women in association with a body shape index and other
anthropometric measures Cancer Causes Control. 2015; 26:219-229.
doi:10.1007/s10552-014-0501-4
130. Harding JL, Shaw JE, Anstey KJ, et al. Comparison of anthropometric measures as
predictors of cancer incidence: A pooled collaborative analysis of 11 Australian
cohorts. Int J Cancer. 2015; 137:1699-1708. doi:10.1002/ijc.29529
131. White AJ, Nichols HB, Bradshaw PT, Sandler DP. Overall and central adiposity and
breast cancer risk in the Sister Study. Cancer. 2015; 121:3700-3708.
doi:10.1002/cncr.29552
132. Gaudet MM, Carter BD, Patel AV, Teras LR, Jacobs EJ, Gapstur SM. Waist
circumference, body mass index, and postmenopausal breast cancer incidence in the
178
Cancer Prevention Study-II Nutrition Cohort. Cancer Causes Control. 2014; 25:737-
745. doi:10.1007/s10552-014-0376-4
133. Sanchez-Bayona R, Sayon-Orea C, Gardeazabal I, et al. Body shape trajectories and
risk of breast cancer: results from the SUN ('Seguimiento Universidad De Navarra')
Project. Public Health Nutr. 2021; 24:467-475. doi:10.1017/S1368980020004322
134. Song M, Willett WC, Hu FB, et al. Trajectory of body shape across the lifespan and
cancer risk. Int J Cancer. 2016; 138:2383-2395. doi:10.1002/ijc.29981
135. Macciò A, Madeddu C. Obesity, inflammation, and postmenopausal breast cancer:
therapeutic implications. ScientificWorldJournal. 2011; 11:2020-2036.
doi:10.1100/2011/806787
136. Islami F, Goding Sauer A, Miller KD, et al. Proportion and number of cancer cases and
deaths attributable to potentially modifiable risk factors in the United States. CA Cancer
J Clin. 2018; 68:31-54. doi:10.3322/caac.21440
137. Brown, K.F., Rumgay, H., Dunlop, C. et al. The fraction of cancer attributable to
modifiable risk factors in England, Wales, Scotland, Northern Ireland, and the United
Kingdom in 2015. Br J Cancer 2018;118: 1130–1141 doi.org/10.1038/s41416-018-
0029-6
138. Wiseman M. The second World Cancer Research Fund/American Institute for Cancer
Research expert report. Food, nutrition, physical activity, and the prevention of cancer:
a global perspective. Proc Nutr Soc. 2008; 67:253-256.
doi:10.1017/S002966510800712X
139. Tapsell LC, Neale EP, Satija A, Hu FB. Foods, Nutrients, and Dietary Patterns:
Interconnections and Implications for Dietary Guidelines. Adv Nutr. 2016; 7:445-454.
doi:10.3945/an.115.011718
140. World Cancer Research Fund International/American Institute for Cancer Research.
Continuous Update Project Report: Diet, Nutrition, Physical Activity and Breast Cancer.
2017. Available at: wcrf.org/breast-cancer-2017 Accessed February 25, 2021.
141. Edefonti V, Randi G, La Vecchia C, Ferraroni M, Decarli A. Dietary patterns and breast
cancer: a review with focus on methodological issues. Nutr Rev. 2009; 67: 297-314.
doi:10.1111/j.1753-4887.2009. 00203.x
142. U.S. Department of Agriculture. Scientific report of the 2015 Dietary Guidelines
Advisory Committee to the Secretary of Agriculture and the Secretary of Health and
Human Services. http://health.gov/dietaryguidelines/2015-scientificreport Accessed
February 25, 2021.
143. World Cancer Research Fund International. Diet, Nutrition, Physical Activity and
Cancer: a Global Perspective - The Third Expert Report.London, UK: World Cancer
Research Fund International; 2018 Available from: https://www.wcrf.org/dietandcancer.
Accessed February 25, 2021
144. U.S. Department of Agriculture and U.S. Department of Health and Human Services.
Dietary Guidelines for Americans, 2020-2025. 9th Edition. December 2020. Available at
179
DietaryGuidelines.gov Acessed February 25, 2021.
145. Orman A, Johnson DL, Comander A, Brockton N. Breast Cancer: A Lifestyle Medicine
Approach. American Journal of Lifestyle Medicine. 2020; 14:483-494.
doi:10.1177/1559827620913263
146. Lohse T, Faeh D, Bopp M, Rohrmann S; Swiss National Cohort Study Group.
Adherence to the cancer prevention recommendations of the World Cancer Research
Fund/American Institute for Cancer Research and mortality: a census-linked cohort. Am
J Clin Nutr. 2016; 104:678-685. doi:10.3945/ajcn.116.135020
147. Fanidi A, Ferrari P, Biessy C, et al. Adherence to the World Cancer Research
Fund/American Institute for Cancer Research cancer prevention recommendations and
breast cancer risk in the Cancer de Màma (CAMA) study. Public Health Nutr. 2015;
18:3337-3348. doi:10.1017/S1368980015000634
148. Romaguera D, Vergnaud AC, Peeters PH, et al. Is concordance with World Cancer
Research Fund/American Institute for Cancer Research guidelines for cancer
prevention related to subsequent risk of cancer? Results from the EPIC study. Am J
Clin Nutr. 2012;96:150-163. doi:10.3945/ajcn.111.031674
149. Nomura SJ, Inoue-Choi M, Lazovich D, Robien K. WCRF/AICR recommendation
adherence and breast cancer incidence among postmenopausal women with and
without non-modifiable risk factors. Int J Cancer. 2016; 138:2602-2615.
doi:10.1002/ijc.29994
150. Nomura SJ, Dash C, Rosenberg L, Yu J, Palmer JR, Adams-Campbell LL. Adherence
to diet, physical activity and body weight recommendations and breast cancer incidence
in the Black Women's Health Study. Int J Cancer. 2016; 139:2738-2752.
doi:10.1002/ijc.30410
151. Harris HR, Bergkvist L, Wolk A. Adherence to the World Cancer Research
Fund/American Institute for Cancer Research recommendations and breast cancer
risk. Int J Cancer. 2016; 138:2657-2664. doi:10.1002/ijc.30015
152. Catsburg C, Miller AB, Rohan TE. Adherence to cancer prevention guidelines and risk
of breast cancer. Int J Cancer. 2014; 135:2444-2452. doi:10.1002/ijc.28887
153. Hastert TA, Beresford SA, Patterson RE, Kristal AR, White E. Adherence to
WCRF/AICR cancer prevention recommendations and risk of postmenopausal breast
cancer. Cancer Epidemiol Biomarkers Prev. 2013; 22:1498-1508. doi:10.1158/1055-
9965.EPI-13-0210
154. Turati F, Dalmartello M, Bravi F, et al. Adherence to the World Cancer Research
Fund/American Institute for Cancer Research Recommendations and the Risk of Breast
Cancer. Nutrients. 2020; 12:607. doi:10.3390/nu12030607
155. Barrios-Rodríguez R, Toledo E, Martinez-Gonzalez MA, Aguilera-Buenosvinos I,
Romanos-Nanclares A, Jiménez-Moleón JJ. Adherence to the 2018 World Cancer
Research Fund/American Institute for Cancer Research Recommendations and Breast
Cancer in the SUN Project. Nutrients. 2020; 12:2076. doi:10.3390/nu12072076
180
156. Subramaniam S, Selvaduray KR, Radhakrishnan AK. Bioactive Compounds: Natural
Defense Against Cancer?. Biomolecules. 2019; 9:758. doi:10.3390/biom9120758
157. Du M, Liu SH, Mitchell C, Fung TT. Associations between Diet Quality Scores and Risk
of Postmenopausal Estrogen Receptor-Negative Breast Cancer: A Systematic
Review. J Nutr. 2018; 148:100-108. doi:10.1093/jn/nxx015
158. Xiao Y, Xia J, Li L, et al. Associations between dietary patterns and the risk of breast
cancer: a systematic review and meta-analysis of observational studies. Breast Cancer
Res. 2019; 21:16. doi:10.1186/s13058-019-1096-1
159. Grosso G, Bella F, Godos J, et al. Possible role of diet in cancer: systematic review
and multiple meta-analyses of dietary patterns, lifestyle factors, and cancer risk. Nutr
Rev. 2017; 75:405-419. doi:10.1093/nutrit/nux012
160. Menotti A, Puddu PE. How the seven countries study contributed to the definition and
development of the Mediterranean diet concept: a 50-year journey. Nutr Metab
Cardiovasc Dis 2015; 25:245–52. 10.1016/j.numecd.2014.12.001
161. Trichopoulou A, Costacou T, Bamia C, Trichopoulos D. Adherence to a Mediterranean
diet and survival in a Greek population. N Engl J Med. 2003;348:2599-2608.
doi:10.1056/NEJMoa025039
162. Galbete C, Schwingshackl L, Schwedhelm C, Boeing H, Schulze MB. Evaluating
Mediterranean diet and risk of chronic disease in cohort studies: an umbrella review of
meta-analyses. Eur J Epidemiol. 2018; 33:909-931. doi:10.1007/s10654-018-0427-3
163. Morze J, Danielewicz A, Przybyłowicz K, Zeng H, Hoffmann G, Schwingshackl L. An
updated systematic review and meta-analysis on adherence to mediterranean diet and
risk of cáncer. Eur J Nutr. 2020; 10.1007/s00394-020-02346-6. doi:10.1007/s00394-
020-02346-6
164. Toledo E, Salas-Salvado J, Donat-Vargas C, et al. Mediterranean Diet and Invasive
Breast Cancer Risk Among Women at High Cardiovascular Risk in the PREDIMED
Trial: A Randomized Clinical Trial. JAMA Intern Med. 2015;175:1752-1760.
doi:10.1001/jamainternmed.2015.4838
165. Sealy N, Hankinson SE, Houghton SC. Olive oil and risk of breast cancer: a systematic
review and dose-response meta-analysis of observational studies. Br J Nutr. 2020; 1-9.
doi:10.1017/S0007114520003499
166. Debras C, Chazelas E, Srour B, et al. Total and added sugar intakes, sugar types, and
cancer risk: results from the prospective NutriNet-Santé cohort. Am J Clin Nutr. 2020;
112:1267-1279. doi:10.1093/ajcn/nqaa246
167. Satija A, Bhupathiraju SN, Rimm EB, et al. Plant-Based Dietary Patterns and Incidence
of Type 2 Diabetes in US Men and Women: Results from Three Prospective Cohort
Studies. PLoS Med. 2016; 13:e1002039. doi:10.1371/journal.pmed.1002039
168. Martínez-González MA, Sánchez-Tainta A, Corella D, et al. A provegetarian food
pattern and reduction in total mortality in the Prevención con Dieta Mediterránea
(PREDIMED) study. Am J Clin Nutr. 2014; 100:320S-8S. doi:10.3945/ajcn.113.071431
181
169. Satija A, Bhupathiraju SN, Spiegelman D, et al. Healthful and Unhealthful Plant-Based
Diets and the Risk of Coronary Heart Disease in U.S. Adults. J Am Coll Cardiol. 2017;
70:411-422. doi:10.1016/j.jacc.2017.05.047
170. Kane-Diallo A, Srour B, Sellem L, et al. Association between a pro plant-based dietary
score and cancer risk in the prospective NutriNet-santé cohort. Int J Cancer. 2018;
143:2168-2176. doi:10.1002/ijc.31593
171. Rigi S, Mousavi SM, Benisi-Kohansal S, Azadbakht L, Esmaillzadeh A. The
association between plant-based dietary patterns and risk of breast cancer: a case–
control study. Sci Rep. 2021;11:3391. doi:10.1038/s41598-021-82659-6
172. Jung S, Spiegelman D, Baglietto L, et al. Fruit and vegetable intake and risk of breast
cancer by hormone receptor status. J Natl Cancer Inst. 2013; 105:219-236.
doi:10.1093/jnci/djs635
173. Emaus MJ, Peeters PH, Bakker MF, et al. Vegetable and fruit consumption and the
risk of hormone receptor-defined breast cancer in the EPIC cohort. Am J Clin Nutr.
2016; 103:168-177. doi:10.3945/ajcn.114.101436
174. Farvid MS, Chen WY, Rosner BA, Tamimi RM, Willett WC, Eliassen AH. Fruit and
vegetable consumption and breast cancer incidence: Repeated measures over 30
years of follow-up. Int J Cancer. 2019; 144:1496-1510. doi:10.1002/ijc.31653
175. Buja A, Pierbon M, Lago L, Grotto G, Baldo V. Breast Cancer Primary Prevention and
Diet: An Umbrella Review. Int J Environ Res Public Health. 2020; 17: 4731.
doi:10.3390/ijerph17134731
176. Bertram JS. Dietary carotenoids, connexins and cancer: what is the
connection? Biochem Soc Trans. 2004; 32:985-989. doi:10.1042/BST0320985
177. Aune D, Chan DS, Vieira AR, et al. Dietary compared with blood concentrations of
carotenoids and breast cancer risk: a systematic review and meta-analysis of
prospective studies. Am J Clin Nutr. 2012; 96:356-373. doi:10.3945/ajcn.112.034165
178. Zhang X, Spiegelman D, Baglietto L, et al. Carotenoid intakes and risk of breast cancer
defined by estrogen receptor and progesterone receptor status: a pooled analysis of 18
prospective cohort studies. Am J Clin Nutr. 2012; 95:713-725.
doi:10.3945/ajcn.111.014415
179. Eliassen AH, Hendrickson SJ, Brinton LA, et al. Circulating carotenoids and risk of
breast cancer: pooled analysis of eight prospective studies. J Natl Cancer Inst. 2012;
104:1905-1916. doi:10.1093/jnci/djs461
180. Peng C, Gao C, Lu D, et al. Circulating carotenoids and breast cancer among high-risk
individuals. Am J Clin Nutr. 2020; nqaa316. doi:10.1093/ajcn/nqaa316
181. Koch W. Dietary Polyphenols-Important Non-Nutrients in the Prevention of Chronic
Noncommunicable Diseases. A Systematic Review. Nutrients. 2019;11:1039.
doi:10.3390/nu11051039
182. Zamora-Ros R, Touillaud M, Rothwell JA, Romieu I, Scalbert A. Measuring exposure
to the polyphenol metabolome in observational epidemiologic studies: Current tools and
182
applications and their limits. Am J Clin Nutr. 2014;100:11-26.
doi:10.3945/ajcn.113.077743
183. Rothwell JA, Knaze V, Zamora-Ros R. Polyphenols: Dietary assessment and role in
the prevention of cancers. Curr Opin Clin Nutr Metab Care. 2017;20:512-521.
doi:10.1097/MCO.0000000000000424
184. Grosso G, Godos J, Lamuela-Raventos R, et al. A comprehensive meta-analysis on
dietary flavonoid and lignan intake and cancer risk: Level of evidence and limitations.
Mol Nutr Food Res. 2017;61. doi:10.1002/mnfr.201600930
185. Lei L, Yang Y, He H, et al. Flavan-3-ols consumption and cancer risk: A meta-analysis
of epidemiologic studies. Oncotarget. 2016; 7:73573-73592.
doi:10.18632/oncotarget.12017
186. Russo GI, Campisi D, Di Mauro M, et al. Dietary Consumption of Phenolic Acids and
Prostate Cancer: A Case-Control Study in Sicily, Southern Italy. Molecules. 2017;
22:2159. doi:10.3390/molecules22122159
187. Gardeazabal I, Romanos-Nanclares A, Martínez-González MA, et al. Total polyphenol
intake and breast cancer risk in the SUN cohort. Br J Nutr. 2018:1-23.
doi:10.1017/S0007114518003811
188. Zamora-Ros R, Rothwell JA, Scalbert A, et al. Dietary intakes and food sources of
phenolic acids in the European Prospective Investigation into Cancer and Nutrition
(EPIC) study. Br J Nutr. 2013; 110:1500-1511. doi:10.1017/S0007114513000688
189. Nagasaka R., Chotimarkorn C., Shafiqul I.M., Hori M., Ozaki H., Ushio H. Anti-
inflammatory effects of hydroxycinnamic acid derivatives. Biochem. Biophys. Res.
Commun. 2007; 358:615–619. doi: 10.1016/j.bbrc.2007.04.178
190. De P., Baltas M., Bedos-Belval F. Cinnamic acid derivatives as anticancer agents-a
review. Curr. Med. Chem. 2011; 18:1672–1703. doi: 10.2174/092986711795471347.
191. Kylli P, Nousiainen P, Biely P, Sipilä J, Tenkanen M, Heinonen M. Antioxidant potential
of hydroxycinnamic acid glycoside esters. J Agric Food Chem. 2008; 56:4797-4805.
doi:10.1021/jf800317v
192. Thom E. The effect of chlorogenic acid enriched coffee on glucose absorption in
healthy volunteers and its effect on body mass when used long-term in overweight and
obese people. J Int Med Res. 2007; 35:900-908. doi:10.1177/147323000703500620
193. Brennan SF, Cantwell MM, Cardwell CR, Velentzis LS, Woodside JV. Dietary patterns
and breast cancer risk: a systematic review and meta-analysis. Am J Clin Nutr. 2010;
91:1294-1302. doi:10.3945/ajcn.2009.28796
194. Albuquerque RC, Baltar VT, Marchioni DM. Breast cancer and dietary patterns: a
systematic review. Nutr Rev. 2014; 72:1-17. doi:10.1111/nure.12083
195. Malik VS, Pan A, Willett WC, Hu FB. Sugar-sweetened beverages and weight gain in
children and adults: a systematic review and meta-analysis. Am J Clin Nutr. 2013;
98:1084-1102. doi:10.3945/ajcn.113.058362
183
196. Imamura F, O'Connor L, Ye Z, et al. Consumption of sugar sweetened beverages,
artificially sweetened beverages, and fruit juice and incidence of type 2 diabetes:
systematic review, meta-analysis, and estimation of population attributable
fraction. BMJ. 2015; 351:h3576. doi:10.1136/bmj.h3576
197. Schwingshackl L, Schwedhelm C, Hoffmann G, et al. Food Groups and Risk of
Hypertension: A Systematic Review and Dose-Response Meta-Analysis of Prospective
Studies. Adv Nutr. 2017; 8:793-803. doi:10.3945/an.117.017178
198. Fung TT, Malik V, Rexrode KM, Manson JE, Willett WC, Hu FB. Sweetened beverage
consumption and risk of coronary heart disease in women. Am J Clin Nutr. 2009;
89:1037-1042. doi:10.3945/ajcn.2008.27140
199. De Koning L, Malik VS, Kellogg MD, Rimm EB, Willett WC, Hu FB. Sweetened
beverage consumption, incident coronary heart disease, and biomarkers of risk in
men. Circulation. 2012; 125:1735-S1. doi:10.1161/CIRCULATIONAHA.111.067017
200. Bernstein AM, de Koning L, Flint AJ, Rexrode KM, Willett WC. Soda consumption and
the risk of stroke in men and women. Am J Clin Nutr. 2012; 95:1190-1199.
doi:10.3945/ajcn.111.030205
201. Malik VS, Li Y, Pan A, et al. Long-Term Consumption of Sugar-Sweetened and
Artificially Sweetened Beverages and Risk of Mortality in US Adults. Circulation. 2019;
139:2113-2125. doi:10.1161/CIRCULATIONAHA.118.037401
202. Klement, R.J., Kämmerer, U. Is there a role for carbohydrate restriction in the
treatment and prevention of cancer?. Nutr Metab 2011; 8, 75. doi.org/10.1186/1743-
7075-8-75
203. Dong J-Y, Qin L-Q. Dietary glycemic index, glycemic load, and risk of breast cancer:
meta-analysis of prospective cohort studies. Breast Cancer Res Treat. 2011;126:287-
294. doi:10.1007/s10549-011-1343-3
204. Schlesinger S, Chan DSM, Vingeliene S, et al. Carbohydrates, glycemic index,
glycemic load, and breast cancer risk: a systematic review and dose-response meta-
analysis of prospective studies. Nutr Rev. 2017; 75:420-441. doi:10.1093/nutrit/nux010
205. Laguna JC, Alegret M, Roglans N. Simple sugar intake and hepatocellular carcinoma:
epidemiological and mechanistic insight. Nutrients. 2014; 6:5933-5954.
doi:10.3390/nu6125933
206. Choi Y, Giovannucci E, Lee JE. Glycaemic index and glycaemic load in relation to risk
of diabetes-related cancers: a meta-analysis. Br J Nutr. 2012; 108:1934-1947.
doi:10.1017/S0007114512003984
207. Sylvetsky AC, Rother KI. Trends in the consumption of low-calorie sweeteners. Physiol
Behav. 2016; 164:446-450. doi:10.1016/j.physbeh.2016.03.030
208. Qin P, Li Q, Zhao Y, et al. Sugar and artificially sweetened beverages and risk of
obesity, type 2 diabetes mellitus, hypertension, and all-cause mortality: a dose-
response meta-analysis of prospective cohort studies. Eur J Epidemiol. 2020; 35:655-
671. doi:10.1007/s10654-020-00655-y
184
209. Chazelas E, Srour B, Desmetz E, et al. Sugary drink consumption and risk of cancer:
results from NutriNet-Santé prospective cohort. BMJ. 2019; 366:l2408.
doi:10.1136/bmj.l2408
210. Hodge AM, Bassett JK, Milne RL, English DR, Giles GG. Consumption of sugar-
sweetened and artificially sweetened soft drinks and risk of obesity-related cancers.
Public Health Nutr. 2018:1-9.
211. Makarem N, Bandera EV, Nicholson JM, Parekh N. Consumption of Sugars, Sugary
Foods, and Sugary Beverages in Relation to Cancer Risk: A Systematic Review of
Longitudinal Studies. Annu Rev Nutr. 2018; 38:17-39. doi:10.1146/annurev-nutr-
082117-051805
212. Arthur RS, Kirsh VA, Mossavar-Rahmani Y, Xue X, Rohan TE. Sugar-containing
beverages and their association with risk of breast, endometrial, ovarian and colorectal
cancers among Canadian women. Cancer Epidemiol. 2021;70:101855.
doi:https://doi.org/10.1016/j.canep.2020.101855
213. Chen S, Chen Y, Ma S, et al. Dietary fibre intake and risk of breast cancer: A
systematic review and meta-analysis of epidemiological studies. Oncotarget. 2016;7:
80980-80989. doi:10.18632/oncotarget.13140
214. Xiao Y, Ke Y, Wu S, et al. Association between whole grain intake and breast cancer
risk: a systematic review and meta-analysis of observational studies. Nutr J.
2018;17:87. doi:10.1186/s12937-018-0394-2
215. Ludwig DS, Hu FB, Tappy L, Brand-Miller J. Dietary carbohydrates: role of quality and
quantity in chronic disease. BMJ. 2018; 361:k2340. doi:10.1136/bmj.k2340
216. Mullie P, Koechlin A, Boniol M, Autier P, Boyle P. Relation between Breast Cancer and
High Glycemic Index or Glycemic Load: A Meta-analysis of Prospective Cohort Studies.
Crit Rev Food Sci Nutr. 2016;56:152-159. doi:10.1080/10408398.2012.718723
217. Sieri S, Agnoli C, Pala V, et al. Dietary glycemic index, glycemic load, and cancer risk:
Results from the EPIC-Italy study. Sci Rep. 2017;7. doi:10.1038/s41598-017-09498-2
218. Zazpe I, Santiago S, Gea A, et al. Association between a dietary carbohydrate index
and cardiovascular disease in the SUN (Seguimiento Universidad de Navarra) Project.
Nutr Metab Cardiovasc Dis. 2016; 26: 1048-1056. doi:10.1016/j.numecd.2016.07.002
219. Santiago S, Zazpe I, Bes-Rastrollo M, et al. Carbohydrate quality, weight change and
incident obesity in a Mediterranean cohort: the SUN Project. Eur J Clin Nutr. 2015; 69:
297-302. doi:10.1038/ejcn.2014.187
220. Zazpe I, Sánchez-Taínta A, Santiago S, et al. Association between dietary
carbohydrate intake quality and micronutrient intake adequacy in a Mediterranean
cohort: the SUN (Seguimiento Universidad de Navarra) Project. Br J Nutr.
2014;111:2000-2009. doi:10.1017/S0007114513004364
221. Smith-Warner SA, Spiegelman D, Adami HO, et al. Types of dietary fat and breast
cancer: a pooled analysis of cohort studies. Int J cancer. 2001;92:767-774.
doi:10.1002/1097-0215(20010601)92:5<767::AID-IJC1247>3.0.CO;2-0
185
222. Boyd NF, Stone J, Vogt KN, Connelly BS, Martin LJ, Minkin S. Dietary fat and breast
cancer risk revisited: A meta-analysis of the published literature. Br J Cancer.
2003;89:1672-1685. doi:10.1038/sj.bjc.6601314
223. Prentice RL, Caan B, Chlebowski RT, et al. Low-fat dietary pattern and risk of invasive
breast cancer: the Women's Health Initiative Randomized Controlled Dietary
Modification Trial. JAMA. 2006; 295:629-642. doi:10.1001/jama.295.6.629
224. Sieri S, Krogh V, Ferrari P, et al. Dietary fat and breast cancer risk in the European
Prospective Investigation into Cancer and Nutrition. Am J Clin Nutr. 2008;88:1304-
1312. doi:10.3945/AJCN.2008.26090
225. Key TJ, Appleby PN, Cairns BJ, et al. Dietary fat and breast cancer: comparison of
results from food diaries and food-frequency questionnaires in the UK Dietary Cohort
Consortium. Am J Clin Nutr. 2011; 94:1043-1052. doi:10.3945/ajcn.111.015735
226. Farvid MS, Cho E, Chen WY, Eliassen AH, Willett WC. Premenopausal dietary fat in
relation to pre- and post-menopausal breast cancer. Breast Cancer Res Treat.
2014;145:255-265. doi:10.1007/s10549-014-2895-9
227. Rossi RE, Pericleous M, Mandair D, Whyand T, Caplin ME. The role of dietary factors
in prevention and progression of breast cancer. Anticancer Res. 2014; 34:6861-6875.
228. Cho E, Spiegelman D, Hunter DJ, et al. Premenopausal fat intake and risk of breast
cancer. J Natl Cancer Inst. 2003;95:1079-1085. doi:10.1093/jnci/95.14.1079
229. Löf M, Sandin S, Lagiou P, et al. Dietary fat and breast cancer risk in the Swedish
women’s lifestyle and health cohort. Br J Cancer. 2007;97:1570-1576.
doi:10.1038/sj.bjc.6604033
230. Escrich E, Solanas M, Moral R, Escrich R. Modulatory effects and molecular
mechanisms of olive oil and other dietary lipids in breast cancer. Curr Pharm Des.
2011; 17:813-830. doi:10.2174/138161211795428902
231. Cao Y, Hou L, Wang W. Dietary total fat and fatty acids intake, serum fatty acids and
risk of breast cancer: A meta-analysis of prospective cohort studies. Int J Cancer.
2016;138:1894-1904. doi:10.1002/ijc.29938
232. De Lorgeril M, Salen P. New insights into the health effects of dietary saturated and
omega-6 and omega-3 polyunsaturated fatty acids. BMC Med. 2012; 10:50.
doi:10.1186/1741-7015-10-50
233. Gonzalez CA, Riboli E. Diet and cancer prevention: Contributions from the European
Prospective Investigation into Cancer and Nutrition (EPIC) study. Eur J Cancer. 2010;
46:2555-2562. doi:10.1016/j.ejca.2010.07.025
234. Sellem L, Srour B, Guéraud F, et al. Saturated, mono- and polyunsaturated fatty acid
intake and cancer risk: results from the French prospective cohort NutriNet-Santé. Eur J
Nutr. 2019; 58:1515-1527. doi:10.1007/s00394-018-1682-5
235. Nindrea RD, Aryandono T, Lazuardi L, Dwiprahasto I. Protective Effect of Omega-3
Fatty Acids in Fish Consumption Against Breast Cancer in Asian Patients: A Meta-
Analysis. Asian Pac J Cancer Prev. 2019; 20:327-332.
186
doi:10.31557/APJCP.2019.20.2.327
236. Chan DSM, Abar L, Cariolou M, et al. World Cancer Research Fund International:
Continuous Update Project—systematic literature review and meta-analysis of
observational cohort studies on physical activity, sedentary behavior, adiposity, and
weight change and breast cancer risk. Cancer Causes Control. 2019;30:1183-1200.
doi:10.1007/s10552-019-01223-w
237. MacLean CH, Newberry SJ, Mojica WA, et al. Effects of omega-3 fatty acids on cancer
risk: a systematic review. JAMA. 2006; 295:403-415. doi:10.1001/jama.295.4.403
238. Nindrea RD, Aryandono T, Lazuardi L, Dwiprahasto I. Association of Dietary Intake
Ratio of n-3/n-6 Polyunsaturated Fatty Acids with Breast Cancer Risk in Western and
Asian Countries: A Meta-Analysis. Asian Pac J Cancer Prev. 2019; 20:1321-1327.
doi:10.31557/APJCP.2019.20.5.1321
239. Pan K, Larson JC, Chlebowski RT, et al. Protein intake and breast cancer incidence
and mortality. J Clin Oncol. 2020;38:1569. doi:10.1200/JCO.2020.38.15_suppl.1569
240. Wu J, Zeng R, Huang J, et al. Dietary Protein Sources and Incidence of Breast Cancer:
A Dose-Response Meta-Analysis of Prospective Studies. Nutrients. 2016; 8:730.
doi:10.3390/nu8110730
241. Farvid MS, Stern MC, Norat T, et al. Consumption of red and processed meat and
breast cancer incidence: A systematic review and meta-analysis of prospective
studies. Int J Cancer. 2018; 143:2787-2799. doi:10.1002/ijc.31848
242. Chen WY, Rosner B, Hankinson SE, Colditz GA, Willett WC. Moderate alcohol
consumption during adult life, drinking patterns, and breast cancer risk. JAMA.
2011;306:1884-1890. doi:10.1001/jama.2011.1590
243. White AJ, DeRoo LA, Weinberg CR, Sandler DP. Lifetime Alcohol Intake, Binge
Drinking Behaviors, and Breast Cancer Risk. Am J Epidemiol. 2017;186:541-549.
244. Sánchez-bayona R, Gea A, Gardeazabal I, et al. Binge drinking and risk of breast
cancer: Results from the SUN (‘Seguimiento universidad de Navarra’) project.
Nutrients. 2020;12. doi:10.3390/nu12030731
245. Shield KD, Soerjomataram I, Rehm J. Alcohol Use and Breast Cancer: A Critical
Review. Alcohol Clin Exp Res. 2016; 40:1166-1181. doi:10.1111/acer.13071
246. Choi YJ, Myung SK, Lee JH. Light Alcohol Drinking and Risk of Cancer: A Meta-
Analysis of Cohort Studies. Cancer Res Treat. 2018; 50:474-487.
doi:10.4143/crt.2017.094
247. DeRoo LA, Cummings P, Mueller BA. Smoking before the first pregnancy and the risk
of breast cancer: a meta-analysis. Am J Epidemiol. 2011; 174:390-402.
doi:10.1093/aje/kwr090
248. Gaudet MM, Gapstur SM, Sun J, Diver WR, Hannan LM, Thun MJ. Active smoking and
breast cancer risk: original cohort data and meta-analysis. J Natl Cancer Inst. 2013;
105: 515-525. doi:10.1093/jnci/djt023
187
249. Bjerkaas E, Parajuli R, Weiderpass E, et al. Smoking duration before first childbirth: an
emerging risk factor for breast cancer? Results from 302,865 Norwegian
women. Cancer Causes Control. 2013; 24:1347-1356. doi:10.1007/s10552-013-0213-1
250. Dossus L, Boutron-Ruault MC, Kaaks R, et al. Active and passive cigarette smoking
and breast cancer risk: results from the EPIC cohort. Int J Cancer. 2014; 134:1871-
1888. doi:10.1002/ijc.28508
251. Gram IT, Park SY, Maskarinec G, Wilkens LR, Haiman CA, Le Marchand L. Smoking
and breast cancer risk by race/ethnicity and oestrogen and progesterone receptor
status: the Multiethnic Cohort (MEC) study. Int J Epidemiol. 2019; 48:501-511.
doi:10.1093/ije/dyy290
252. Lawlor DA, Ebrahim S, Smith GD. Smoking before the birth of a first child is not
associated with increased risk of breast cancer: findings from the British Women's Heart
and Health Cohort Study and a meta-analysis. Br J Cancer. 2004; 91:512-518.
doi:10.1038/sj.bjc.6601916
253. Jones ME, Schoemaker MJ, Wright LB, Ashworth A, Swerdlow AJ. Smoking and risk of
breast cancer in the Generations Study cohort. Breast Cancer Res. 2017; 19:118.
doi:10.1186/s13058-017-0908-4
254. Moore SC, Lee IM, Weiderpass E, et al. Association of Leisure-Time Physical Activity
With Risk of 26 Types of Cancer in 1.44 Million Adults. JAMA Intern Med. 2016;
176:816-825. doi:10.1001/jamainternmed.2016.1548
255. Guo, W., Fensom, G.K., Reeves, G.K. et al. Physical activity and breast cancer risk:
results from the UK Biobank prospective cohort. Br J Cancer 2020;122, 726–732
.doi.org/10.1038/s41416-019-0700-6
256. Wu Y, Zhang D, Kang S. Physical activity and risk of breast cancer: A meta-analysis of
prospective studies. Breast Cancer Res Treat. 2013;137:869-882. doi:10.1007/s10549-
012-2396-7
257. Gong Z, Hong CC, Bandera EV, et al. Vigorous physical activity and risk of breast
cancer in the African American breast cancer epidemiology and risk consortium. Breast
Cancer Res Treat. 2016; 159:347-356. doi:10.1007/s10549-016-3936-3
258. Neilson HK, Farris MS, Stone CR, Vaska MM, Brenner DR, Friedenreich CM.
Moderate-vigorous recreational physical activity and breast cancer risk, stratified by
menopause status: a systematic review and meta-analysis. Menopause. 2017; 24:322-
344. doi:10.1097/GME.0000000000000745
259. Cifu G., Arem H. Adherence to lifestyle-related cancer prevention guidelines and
breast cancer incidence and mortality. Ann. Epidemiol. 2018; 28:767–773. doi:
10.1016/j.annepidem.2018.09.002
260. Eliassen AH, Hankinson SE, Rosner B, Holmes MD, Willett WC. Physical activity and
risk of breast cancer among postmenopausal women. Arch Intern Med. 2010;
170:1758-1764. doi:10.1001/archinternmed.2010.363
188
261. Kushi LH, Doyle C, McCullough M, et al. American Cancer Society Guidelines on
nutrition and physical activity for cancer prevention: reducing the risk of cancer with
healthy food choices and physical activity. CA Cancer J Clin. 2012; 62:30-67.
doi:10.3322/caac.20140
262. Pizot C, Boniol M, Mullie P, et al. Physical activity, hormone replacement therapy and
breast cancer risk: A meta-analysis of prospective studies. Eur J Cancer. 2016;52:138-
54. doi:10.1016/j.ejca.2015.10.063 pmid:26687833
263. Boeke CE, Eliassen AH, Oh H, Spiegelman D, Willett WC, Tamimi RM. Adolescent
physical activity in relation to breast cancer risk. Breast Cancer Res Treat. 2014;
145:715-724. doi:10.1007/s10549-014-2919-5
264. Niehoff NM, White AJ, Sandler DP. Childhood and teenage physical activity and breast
cancer risk. Breast Cancer Res Treat. 2017; 164:697-705. doi:10.1007/s10549-017-
4276-7
265. Hong BS, Lee KP. A systematic review of the biological mechanisms linking physical
activity and breast cancer. Phys Act Nutr. 2020; 24:25-31. doi:10.20463/pan.2020.0018
266. Del Pup L, Codacci-Pisanelli G, Peccatori F. Breast cancer risk of hormonal
contraception: Counselling considering new evidence. Crit Rev Oncol Hematol. 2019;
137:123-130. doi:10.1016/j.critrevonc.2019.03.001
267. Calle EE, Heath CW, Miracle-McMahill HL, et al.Breast cancer and hormonal
contraceptives: Collaborative reanalysis of individual data on 53 297 women with breast
cancer and 100 239 women without breast cancer from 54 epidemiological
studies. Lancet.1996; 347, 1713-1727. doi:10.1016/S0140-6736(96)90806-5
268. Gierisch JM, Coeytaux RR, Urrutia RP, et al. Oral contraceptive use and risk of breast,
cervical, colorectal, and endometrial cancers: a systematic review. Cancer Epidemiol
Biomarkers Prev. 2013; 22:1931-1943. doi:10.1158/1055-9965.EPI-13-0298
269. Mørch LS, Skovlund CW, Hannaford PC, Iversen L, Fielding S, Lidegaard Ø.
Contemporary Hormonal Contraception and the Risk of Breast Cancer. N Engl J Med.
2017; 377:2228-2239. doi:10.1056/NEJMoa1700732
270. Li CI, Beaber EF, Tang MT, Porter PL, Daling JR, Malone KE. Effect of depo-
medroxyprogesterone acetate on breast cancer risk among women 20 to 44 years of
age. Cancer Res. 2012; 72:2028-2035. doi:10.1158/0008-5472.CAN-11-4064
271. Busund M, Bugge NS, Braaten T, Waaseth M, Rylander C, Lund E. Progestin-only and
combined oral contraceptives and receptor-defined premenopausal breast cancer risk:
The Norwegian Women and Cancer Study. Int J Cancer. 2018; 142:2293-2302.
doi:10.1002/ijc.31266
272. Dinger J, Bardenheuer K, Minh TD. Levonorgestrel-releasing and copper intrauterine
devices and the risk of breast cancer. Contraception. 2011; 83:211-217.
doi:10.1016/j.contraception.2010.11.009
273. Heikkinen S, Koskenvuo M, Malila N, Sarkeala T, Pukkala E, Pitkäniemi J. Use of
exogenous hormones and the risk of breast cancer: results from self-reported survey
189
data with validity assessment. Cancer Causes Control. 2016; 27:249-258.
doi:10.1007/s10552-015-0702-5
274. Nichols HB, Schoemaker MJ, Cai J, et al. Breast Cancer Risk After Recent Childbirth:
A Pooled Analysis of 15 Prospective Studies. Ann Intern Med. 2019; 170:22-30.
doi:10.7326/M18-1323
275. Mullooly M, Khodr ZG, Dallal CM, et al. Epidemiologic Risk Factors for In Situ and
Invasive Breast Cancers Among Postmenopausal Women in the National Institutes of
Health-AARP Diet and Health Study. Am J Epidemiol. 2017; 186:1329-1340.
276. Ewertz M, Duffy SW, Adami HO, et al. Age at first birth, parity and risk of breast
cancer: a meta-analysis of 8 studies from the Nordic countries. Int J Cancer. 1990;
46:597-603. doi:10.1002/ijc.2910460408
277. Rosner B, Colditz GA, Willett WC. Reproductive risk factors in a prospective study of
breast cancer: the Nurses' Health Study. Am J Epidemiol. 1994; 139:819-835.
doi:10.1093/oxfordjournals.aje.a117079
278. Colditz GA, Rosner B. Cumulative risk of breast cancer to age 70 years according to
risk factor status: data from the Nurses' Health Study. Am J Epidemiol. 2000; 152:950-
964. doi:10.1093/aje/152.10.950
279. Fortner RT, Sisti J, Chai B, et al. Parity, breastfeeding, and breast cancer risk by
hormone receptor status and molecular phenotype: results from the Nurses' Health
Studies. Breast Cancer Res. 2019; 21:40. doi:10.1186/s13058-019-1119-y
280. Bernier MO, Plu-Bureau G, Bossard N, Ayzac L, Thalabard JC. Breastfeeding and risk
of breast cancer: a metaanalysis of published studies. Hum Reprod Update. 2000;
6:374-386. doi:10.1093/humupd/6.4.374
281. Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and
breastfeeding: collaborative reanalysis of individual data from 47 epidemiological
studies in 30 countries, including 50302 women with breast cancer and 96973 women
without the disease. Lancet. 2002; 360:187-195. doi:10.1016/S0140-6736(02)09454-0
282. Lambertini M, Santoro L, Del Mastro L, et al. Reproductive behaviors and risk of
developing breast cancer according to tumor subtype: A systematic review and meta-
analysis of epidemiological studies. Cancer Treat Rev. 2016; 49:65-76.
doi:10.1016/j.ctrv.2016.07.006
283. Ma H, Ursin G, Xu X, et al. Reproductive factors and the risk of triple-negative breast
cancer in white women and African-American women: a pooled analysis. Breast Cancer
Res. 2017; 19:6. doi:10.1186/s13058-016-0799-9
284. Collaborative Group On Epidemiological Studies Of Ovarian Cancer, Beral V, Gaitskell
K, et al. Menopausal hormone use and ovarian cancer risk: individual participant meta-
analysis of 52 epidemiological studies. Lancet .2015; 385:1835.
285. Collaborative Group on Hormonal Factors in Breast Cancer. Type and Timing of
menopausal therapy and breast cancer risk individual participant meta-analysis of the
190
worldwide epidemiological evidence. Lancet 2019; 394 1159-68. :10.1016/S0140-
6736(19)31474332
286. Clinical investigation of medicinal products for hormone replacement therapy
oestrogen deficiency symptoms in postmenopausal women - European Medicines
Agency. European Medicines Agency. https://www.ema.europa.eu/en/clinical-
investigation-medicinal-products-hormone-replacement-therapy-oestrogen-deficiency-
symptoms Accessed February 25, 2021.
287. Menopause. U.S. Food and Drug Administration.
https://www.fda.gov/consumers/womens-health-topics/menopause. Accessed February
25, 2021.
288. Chlebowski RT, Hendrix SL, Langer RD, et al. Influence of estrogen plus progestin on
breast cancer and mammography in healthy postmenopausal women: the Women's
Health Initiative Randomized Trial. JAMA. 2003; 289:3243-3253.
doi:10.1001/jama.289.24.3243
289. Bakken K, Fournier A, Lund E, et al. Menopausal hormone therapy and breast cancer
risk: impact of different treatments. The European Prospective Investigation into Cancer
and Nutrition. Int J Cancer. 2011; 128:144-156. doi:10.1002/ijc.25314
290. Beral V; Million Women Study Collaborators. Breast cancer and hormone-replacement
therapy in the Million Women Study. Lancet. 2003; 362:419-427. doi:10.1016/s0140-
6736(03)14065-2
291. Rossouw JE, Anderson GL, Prentice RL, et al. Risks and benefits of estrogen plus
progestin in healthy postmenopausal women: principal results From the Women's
Health Initiative randomized controlled trial. JAMA. 2002; 288:321-333.
doi:10.1001/jama.288.3.321
292. Manson JE, Chlebowski RT, Stefanick ML, et al. Menopausal hormone therapy and
health outcomes during the intervention and extended poststopping phases of the
Women's Health Initiative randomized trials. JAMA. 2013; 310:1353-1368.
doi:10.1001/jama.2013.278040
293. Chlebowski RT, Anderson GL. Menopausal hormone therapy and breast cancer
mortality: clinical implications. Ther Adv drug Saf. 2015;6:45-56.
doi:10.1177/2042098614568300
294. Chen WY, Manson JE, Hankinson SE, et al. Unopposed estrogen therapy and the risk
of invasive breast cancer. Arch Intern Med. 2006; 166:1027-1032.
doi:10.1001/archinte.166.9.1027
295. Brinton LA, Richesson D, Leitzmann MF, et al. Menopausal hormone therapy and
breast cancer risk in the NIH-AARP Diet and Health Study Cohort. Cancer Epidemiol
Biomarkers Prev. 2008; 17:3150-3160. doi:10.1158/1055-9965.EPI-08-0435
296. Palmer JR, Boggs DA, Wise LA, Adams-Campbell LL, Rosenberg L. Individual and
neighborhood socioeconomic status in relation to breast cancer incidence in African-
American women. Am J Epidemiol. 2012; 176:1141-1146. doi:10.1093/aje/kws211
191
297. Lundqvist A, Andersson E, Ahlberg I, Nilbert M, Gerdtham U. Socioeconomic
inequalities in breast cancer incidence and mortality in Europe-a systematic review and
meta-analysis. Eur J Public Health. 2016; 26:804-813. doi:10.1093/eurpub/ckw070
298. Brown SB, Hankinson SE, Eliassen AH, et al. Urinary melatonin concentration and the
risk of breast cancer in Nurses' Health Study II. Am J Epidemiol. 2015; 181:155-162.
doi:10.1093/aje/kwu261
299. Devore EE, Warner ET, Eliassen AH, et al. Urinary Melatonin in Relation to
Postmenopausal Breast Cancer Risk According to Melatonin 1 Receptor Status. Cancer
Epidemiol Biomarkers Prev. 2017; 26:413-419. doi:10.1158/1055-9965.EPI-16-0630
300. Vos T, Lim SS, Abbafati C, et al. Global burden of 369 diseases and injuries in 204
countries and territories, 1990–2019: a systematic analysis for the Global Burden of
Disease Study 2019. The Lancet. 2020. 396:1204-1222. doi: 10.1016/S0140-
6736(20)30925-9
301. Molina-Montes E, Salamanca-Fernández E, Garcia-Villanova B, Sánchez MJ. The
Impact of Plant-Based Dietary Patterns on Cancer-Related Outcomes: A Rapid Review
and Meta-Analysis. Nutrients. 2020; 12:2010. doi:10.3390/nu12072010
302. Malik VS, Hu FB. Sugar-Sweetened Beverages and Cardiometabolic Health: An
Update of the Evidence. Nutrients. 2019; 11:1840. doi:10.3390/nu11081840
303. Muti P, Bradlow HL, Micheli A, et al. Estrogen metabolism and risk of breast cancer: a
prospective study of the 2:16alpha-hydroxyestrone ratio in premenopausal and
postmenopausal women. Epidemiology. 2000; 11:635-640. doi:10.1097/00001648-
200011000-00004
304. Russnes HG, Lingjærde OC, Børresen-Dale AL, Caldas C. Breast Cancer Molecular
Stratification: From Intrinsic Subtypes to Integrative Clusters. Am J Pathol. 2017;
187:2152-2162. doi:10.1016/j.ajpath.2017.04.022
305. Anderson KN, Schwab RB, Martinez ME. Reproductive risk factors and breast cancer
subtypes: a review of the literature. Breast Cancer Res Treat. 2014; 144:1-10.
doi:10.1007/s10549-014-2852-7
306. Althuis MD, Fergenbaum JH, Garcia-Closas M, Brinton LA, Madigan MP, Sherman
ME. Etiology of hormone receptor-defined breast cancer: a systematic review of the
literature. Cancer Epidemiol Biomarkers Prev. 2004; 13:1558-1568.
307. Rothman, Kenneth J, and Sander Greenland. Design strategies to improve study
accuracy. In: Modern Epidemiology. Philadelphia, PA: Lippincott-Raven, 2011.
308. Molist G, Barrio G, Santos S, et al. Quality deficits in the Spanish National Death
Index: lessons learned from studying the mortality of two cohorts of people admitted to
drug abuse treatment. Gac Sanit. 2012; 26:261-266. doi:10.1016/j.gaceta.2011.08.008
309. Alonso A, de Irala J, Martínez-González MA. Representativeness, losses to follow-up
and validity in cohort studies. Eur J Epidemiol. 2007; 22:481-482. doi:10.1007/s10654-
007-9124-3
192
310. Alonso A, Seguí-Gómez M, de Irala J, Sánchez-Villegas A, Beunza JJ, Martínez-
Gonzalez MA. Predictors of follow-up and assessment of selection bias from dropouts
using inverse probability weighting in a cohort of university graduates. Eur J Epidemiol.
2006; 21:351-358. doi:10.1007/s10654-006-9008-y
311. Martinez-Gonzalez MA, Lopez-Fontana C, Varo JJ, Sanchez-Villegas A, Martinez JA.
Validation of the Spanish version of the physical activity questionnaire used in the
Nurses’ Health Study and the Health Professionals’ Follow-up Study. Public Heathl
Nutr. 2005;8:920-927. doi:S1368980005001230
312. Bes-Rastrollo M, Pérez Valdivieso JR, Sanchez-Villegas A, Alonso A, Martínez-
González M. Validation of self-reported weight and body mass index of the participants
of a cohort of university graduates. Rev Esp Obes. 2005;3:352-358.
313. De la Fuente-Arrillaga C, Vázquez Ruiz Z, Bes-Rastrollo M, Sampson L, Martinez-
González MA. Reproducibility of an FFQ validated in Spain. Public Health Nutr.
2010;13:1364-1372. doi:10.1017/S1368980009993065
314. Martin-Moreno JM, Boyle P, Gorgojo L, et al. Development and validation of a food
frequency questionnaire in Spain. Int J Epidemiol. 1993;22:512-519.
doi:10.1093/ije/22.3.512
315. Fernández-Ballart JD, Piñol JL, Zazpe I, et al. Relative validity of a semi-quantitative
food-frequency questionnaire in an elderly Mediterranean population of Spain. Br J
Nutr. 2010;103:1808-1816. doi:10.1017/S0007114509993837
316. Medina-Remón A, Barrionuevo-González A, Zamora-Ros R, et al. Rapid Folin–
Ciocalteu method using microtiter 96-well plate cartridges for solid phase extraction to
assess urinary total phenolic compounds, as a biomarker of total polyphenols intake.
Anal Chim Acta. 2009;634:54-60. doi:10.1016/j.aca.2008.12.012
317. Mataix J. Tablas de Composición de Alimentos (Spanish Food Composition Tables).
Vol 5th ed.; 2009.
318. Moreiras O. Tablas de Composición de Alimentos : Guía de Prácticas. Pirámide; 2015.
319. Gardeazabal I, Romanos-Nanclares A, Martínez-González MÁ, et al. Total polyphenol
intake and breast cancer risk in the Seguimiento Universidad de Navarra (SUN) cohort.
Br J Nutr. 2019;122. doi:10.1017/S0007114518003811
320. Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses.
AmJEpidemiol. 1986;124:17-27. doi:10.1093/aje/kwh134
321. Willet W, Howe G, Kushi L. Adjustment for total energy intake in epidemiologic studies.
Am J Clin Nutr. 1997;65:1220S-8S. doi:10.1007/s13398-014-0173-7.2
322. Martínez-González MA, Sánchez-Tainta A, Corella D, et al. A provegetarian food
pattern and reduction in total mortality in the Prevención con Dieta Mediterránea
(PREDIMED) study. Am J Clin Nutr. 2014;100:320S-328S.
doi:10.3945/ajcn.113.071431
323. Satija A, Hu FB. Plant-based diets and cardiovascular health. Trends Cardiovasc Med.
2018; 28:437-441. doi:10.1016/j.tcm.2018.02.004
193
324. Willett, Walter. Nutritional Epidemiology. Oxford: Oxford University Press, 2013.
325. Zazpe I, Sánchez-Taínta A, Santiago S, et al. Association between dietary
carbohydrate intake quality and micronutrient intake adequacy in a Mediterranean
cohort: the SUN (Seguimiento Universidad de Navarra) Project. Br J Nutr. 2014;
111:2000-2009. doi:10.1017/S0007114513004364
326. Sánchez-Tainta A, Zazpe I, Bes-Rastrollo M, et al. Nutritional adequacy according to
carbohydrates and fat quality. Eur J Nutr. 2016; 55:93-106. doi:10.1007/s00394-014-
0828-3
327. Fernandez-Lazaro CI, Zazpe I, Santiago S, Toledo E, Barbería-Latasa M, Martínez-
González MÁ. Association of carbohydrate quality and all-cause mortality in the SUN
Project: A prospective cohort study. Clin Nutr. 2020; S0261-5614(20)30572-0.
doi:10.1016/j.clnu.2020.10.029
328. Atkinson FS, Foster-Powell K, Brand-Miller JC. International tables of glycemic index
and glycemic load values: 2008. Diabetes Care. 2008; 31:2281-2283.
doi:10.2337/dc08-1239
329. Liu S, Willett WC, Stampfer MJ, et al. A prospective study of dietary glycemic load,
carbohydrate intake, and risk of coronary heart disease in US women. Am J Clin Nutr.
2000; 71:1455-1461. doi:10.1093/ajcn/71.6.1455
330. Willett W. Issues in Analysis and Presentation of Dietary Data. In: Nutritional
Epidemiology. Oxford: Oxford University Press, 2013.
331. Shivappa N, Sandin S, Löf M, Hébert JR, Adami HO, Weiderpass E. Prospective study
of dietary inflammatory index and risk of breast cancer in Swedish women. Br J Cancer.
2015; 113:1099-1103. doi:10.1038/bjc.2015.304
332. Royston, P. , and Lambert, P. C. 2011. Flexible Parametric Survival Analysis Using
Stata: Beyond the Cox Model. College Station, TX: Stata Press..
333. Thiébaut ACM, Bénichou J. Choice of time-scale in Cox’s model analysis of
epidemiologic cohort data: A simulation study. Stat Med. 2004;23:3803-3820.
doi:10.1002/sim.2098
334. Greenland S. Dose-response and trend analysis in epidemiology: alternatives to
categorical analysis. Epidemiology. 1995; 6:356-365. doi:10.1097/00001648-
199507000-00005
335. Clogg CC, Petkova E, Haritou A. Statistical Methods for Comparing Regression
Coefficients Between Models. Am J Sociol. 1995;100:1261-1293. doi:10.1086/230638
336. Weesie J. Seemingly unrelated estimation and the cluster-adjusted sandwich
estimator. Stata Tech Bulitin. 1999.
337. Brazier JE, Harper R, Jones NM, et al. Validating the SF-36 health survey
questionnaire: new outcome measure for primary care. BMJ. 1992; 305:160-164.
doi:10.1136/bmj.305.6846.160
194
338. Michels KB, Willett WC, Graubard BI, et al. A longitudinal study of infant feeding and
obesity throughout life course. Int J Obes (Lond). 2007; 31:1078-1085.
doi:10.1038/sj.ijo.0803622
339. Hu FB, Satija A, Rimm EB, et al. Diet Assessment Methods in the Nurses' Health
Studies and Contribution to Evidence-Based Nutritional Policies and Guidelines. Am J
Public Health. 2016; 106:1567-1572. doi:10.2105/AJPH.2016.303348
340. Willett WC, Sampson L, Stampfer MJ, et al. Reproducibility and validity of a
semiquantitative food frequency questionnaire. Am J Epidemiol. 1985; 122:51-65.
341. Willett W. Reproducibility and Validity of Food Frequency Questionnaires. In:
Nutritional Epidemiology. Oxford: Oxford University Press, 2013
342. Bao Y, Bertoia ML, Lenart EB, et al. Origin, Methods, and Evolution of the Three
Nurses' Health Studies. Am J Public Health. 2016; 106:1573-1581.
doi:10.2105/AJPH.2016.303338
343. Willett WC, Sampson L, Browne ML, et al. The use of a self-administered
questionnaire to assess diet four years in the past. Am J Epidemiol. 1988; 127:188-199.
344. Yuan C, Spiegelman D, Rimm EB, et al. Validity of a Dietary Questionnaire Assessed
by Comparison With Multiple Weighed Dietary Records or 24-Hour Recalls. Am J
Epidemiol. 2017; 185:570-584. doi:10.1093/aje/kww104
345. Salvini S, Hunter DJ, Sampson L, et al. Food-based validation of a dietary
questionnaire: the effects of week-to-week variation in food consumption. Int J
Epidemiol. 1989; 18:858-867. doi:10.1093/ije/18.4.858
346. Tabung FK, Smith-Warner SA, Chavarro JE, et al. Development and Validation of an
Empirical Dietary Inflammatory Index. J Nutr. 2016; 146:1560-1570.
doi:10.3945/jn.115.228718
347. Tamimi RM, Baer HJ, Marotti J, et al. Comparison of molecular phenotypes of ductal
carcinoma in situ and invasive breast cancer. Breast Cancer Res. 2008; 10:R67.
doi:10.1186/bcr2128
348. Cheang MC, Chia SK, Voduc D, et al. Ki67 index, HER2 status, and prognosis of
patients with luminal B breast cancer. J Natl Cancer Inst. 2009; 101:736-750.
doi:10.1093/jnci/djp082
349. Cheang MC, Voduc D, Bajdik C, et al. Basal-like breast cancer defined by five
biomarkers has superior prognostic value than triple-negative phenotype. Clin Cancer
Res. 2008; 14:1368-1376. doi:10.1158/1078-0432.CCR-07-1658
350. Van de Rijn M, Perou CM, Tibshirani R, et al. Expression of cytokeratins 17 and 5
identifies a group of breast carcinomas with poor clinical outcome. Am J Pathol. 2002;
161:1991-1996. doi:10.1016/S0002-9440(10)64476-8
195
351. Lee JE, Willett WC, Fuchs CS, et al. Folate intake and risk of colorectal cancer and
adenoma: modification by time. Am J Clin Nutr. 2011; 93:817-825.
doi:10.3945/ajcn.110.007781
352. Martínez ME, Giovannucci E, Spiegelman D, Hunter DJ, Willett WC, Colditz GA.
Leisure-time physical activity, body size, and colon cancer in women. Nurses' Health
Study Research Group. J Natl Cancer Inst. 1997; 89:948-955.
doi:10.1093/jnci/89.13.948
353. Lunn M, McNeil D. Applying Cox regression to competing risks. Biometrics. 1995;
51:524-532.
354. Lin DY, Fleming TR, De Gruttola V. Estimating the proportion of treatment effect
explained by a surrogate marker. Stat Med. 1997;16:1515-1527.
355. Stryker WS, Kaplan LA, Stein EA, Stampfer MJ, Sober A, Willett WC. The relation of
diet, cigarette smoking, and alcohol consumption to plasma beta-carotene and alpha-
tocopherol levels. Am J Epidemiol. 1988; 127:283-296.
356. Jun HJ, Austin SB, Wylie SA, et al. The mediating effect of childhood abuse in sexual
orientation disparities in tobacco and alcohol use during adolescence: results from the
Nurses' Health Study II. Cancer Causes Control. 2010; 21:1817-1828.
doi:10.1007/s10552-010-9609-3
357. McCullough ML, Feskanich D, Stampfer MJ, et al. Diet quality and major chronic
disease risk in men and women: moving toward improved dietary guidance. Am J Clin
Nutr. 2002; 76:1261-1271. doi:10.1093/ajcn/76.6.1261
358. Neveu V, Perez-Jiménez J, Vos F, et al. Phenol-Explorer: an online comprehensive
database on polyphenol contents in foods. Database (Oxford). 2010; 2010:bap024.
doi:10.1093/database/bap024
359. Colditz GA, Hankinson SE. The Nurses' Health Study: lifestyle and health among
women. Nat Rev Cancer. 2005; 5:388-396. doi:10.1038/nrc1608
360. Simone V, D'Avenia M, Argentiero A, et al. Obesity and Breast Cancer: Molecular
Interconnections and Potential Clinical Applications. Oncologist. 2016; 21:404-417.
doi:10.1634/theoncologist.2015-0351
361. Kaaks R, Lukanova A. Energy balance and cancer: the role of insulin and insulin-like
growth factor-I. Proc Nutr Soc. 2001; 60:91-106. doi:10.1079/pns200070
362. Smith TJ, Wolfson JA, Jiao D, et al. Caramel color in soft drinks and exposure to 4-
methylimidazole: a quantitative risk assessment. PLoS One. 2015; 10:e0118138.
doi:10.1371/journal.pone.0118138
363. Leis-Keeling K. Comprehensive Evaluation of Soft Drinks, Effects On Health, and
Nutritional Strategies To Reverse Damage. Journal of the Council on Nutrition of the
American Chiropractic Association. 2010; 33, 15-23.
364. Bradshaw PT, Sagiv SK, Kabat GC, et al. Consumption of sweet foods and breast
cancer risk: a case-control study of women on Long Island, New York. Cancer Causes
196
Control. 2009;20:1509-1515. doi:10.1007/s10552-009-9343-x
365. Kang JH, Peng C, Rhee JJ, et al. Prospective study of a diabetes risk reduction diet
and the risk of breast cancer. Am J Clin Nutr. 2020; 112:1492-1503.
doi:10.1093/ajcn/nqaa268
366. Lajous M, Boutron-Ruault MC, Fabre A, Clavel-Chapelon F, Romieu I. Carbohydrate
intake, glycemic index, glycemic load, and risk of postmenopausal breast cancer in a
prospective study of French women. Am J Clin Nutr. 2008; 87:1384-1391.
doi:10.1093/ajcn/87.5.1384
367. Fowler SP, Williams K, Resendez RG, Hunt KJ, Hazuda HP, Stern MP. Fueling the
obesity epidemic? Artificially sweetened beverage use and long-term weight
gain. Obesity. 2008; 16:1894-1900. doi:10.1038/oby.2008.284
368. Narain A, Kwok CS, Mamas MA. Soft drinks and sweetened beverages and the risk of
cardiovascular disease and mortality: a systematic review and meta-analysis. Int J Clin
Pract. 2016; 70:791-805. doi:10.1111/ijcp.12841
369. Johnson RK, Lichtenstein AH, Anderson CAM, et al. Low-Calorie Sweetened
Beverages and Cardiometabolic Health: A Science Advisory From the American Heart
Association. Circulation. 2018; 138:e126-e140. doi:10.1161/CIR.0000000000000569
370. Abotaleb M, Liskova A, Kubatka P, Büsselberg D. Therapeutic Potential of Plant
Phenolic Acids in the Treatment of Cancer. Biomolecules. 2020; 10:221.
doi:10.3390/biom10020221
371. Tresserra-Rimbau A, Medina-Remón A, Pérez-Jiménez J, et al. Dietary intake and
major food sources of polyphenols in a Spanish population at high cardiovascular risk:
the PREDIMED study. Nutr Metab Cardiovasc Dis. 2013; 23:953-959.
doi:10.1016/j.numecd.2012.10.008
372. Zamora-Ros R, Knaze V, Rothwell JA, et al. Dietary polyphenol intake in Europe: the
European Prospective Investigation into Cancer and Nutrition (EPIC) study. Eur J Nutr.
2016;55:1359-1375. doi:10.1007/s00394-015-0950-x
373. Vauzour D, Rodriguez-Mateos A, Corona G, Oruna-Concha MJ, Spencer JP.
Polyphenols and human health: prevention of disease and mechanisms of
action. Nutrients. 2010; 2:1106-1131. doi:10.3390/nu2111106
374. Rocha LD, Monteiro MC, Teodoro AJ. Anticancer Properties of Hydroxycinnamic Acids
-A Review. Cancer Clin Oncol. 2012;1:109. doi:10.5539/cco.v1n2p109
375. Radtke J, Linseisen J, Wolfram G. Phenolic acid intake of adults in a Bavarian
subgroup of the national food consumption survey. Z Ernahrungswiss. 1998;190-197.
376. Ovaskainen ML, Törrönen R, Koponen JM, et al. Dietary intake and major food
sources of polyphenols in Finnish adults. J Nutr. 2008; 138:562-566.
doi:10.1093/jn/138.3.562
377. Pérez-Jiménez J, Fezeu L, Touvier M, et al. Dietary intake of 337 polyphenols in
French adults. Am J Clin Nutr. 2011; 93:1220-1228. doi:10.3945/ajcn.110.007096
197
378. Weng CJ, Yen GC. Chemopreventive effects of dietary phytochemicals against cancer
invasion and metastasis: phenolic acids, monophenol, polyphenol, and their
derivatives. Cancer Treat Rev. 2012; 38:76-87. doi:10.1016/j.ctrv.2011.03.001
379. Gorzynik-Debicka M, Przychodzen P, Cappello F, et al. Potential Health Benefits of
Olive Oil and Plant Polyphenols. Int J Mol Sci. 2018; 19:686. doi:10.3390/ijms19030686
380. Losada-Echeberría M, Herranz-López M, Micol V, Barrajón-Catalán E. Polyphenols as
Promising Drugs against Main Breast Cancer Signatures. Antioxidants. 2017;6:88.
doi:10.3390/antiox6040088
381. Mocanu MM, Nagy P, Szöllosi J, Vanden Eynde JJ, Mayence A. Chemoprevention of
breast cancer by dietary polyphenols. Molecules. 2015; 20:22578-22620.
doi:10.3390/molecules201219864
382. Johnston KL, Clifford MN, Morgan LM. Coffee acutely modifies gastrointestinal
hormone secretion and glucose tolerance in humans: glycemic effects of chlorogenic
acid and caffeine. Am J Clin Nutr. 2003; 78:728-733. doi:10.1093/ajcn/78.4.728
383. Gunter MJ, Hoover DR, Yu H, et al. Insulin, insulin-like growth factor-I, and risk of
breast cancer in postmenopausal women. J Natl Cancer Inst. 2009;101:48-60.
doi:10.1093/jnci/djn415
384. Garg SK. Green Coffee Bean. In: Nutraceuticals: Efficacy, Safety and Toxicity. 2016.
doi:10.1016/B978-0-12-802147-7.00047-4
385. Nakatani N, Kayano S, Kikuzaki H, Sumino K, Katagiri K, Mitani T. Identification,
quantitative determination, and antioxidative activities of chlorogenic acid isomers in
prune (Prunus domestica L. ). J Agric Food Chem. 2000; 48:5512-5516.
doi:10.1021/jf000422s
386. Manach C, Scalbert A, Morand C, Rémésy C, Jiménez L. Polyphenols: Food sources
and bioavailability. Am J Clin Nutr. 2004;79:727-747. doi:10.1093/ajcn/79.5.727
387. Goldin BR, Woods MN, Spiegelman DL, et al. The effect of dietary fat and fiber on
serum estrogen concentrations in premenopausal women under controlled dietary
conditions. Cancer. 1994; 74:1125-1131. doi:10.1002/1097-
0142(19940801)74:3+<1125::aid-cncr2820741521>3.0.co;2-5
388. Heath AK, Muller DC, van den Brandt PA, et al. Nutrient-wide association study of 92
foods and nutrients and breast cancer risk. Breast Cancer Res. 2020; 22:5.
doi:10.1186/s13058-019-1244-7
389. Martin LJ, Li Q, Melnichouk O, et al. A randomized trial of dietary intervention for
breast cancer prevention. Cancer Res. 2011;71:123-133. doi:10.1158/0008-5472.CAN-
10-1436
390. Cho E, Chen WY, Hunter DJ, et al. Red meat intake and risk of breast cancer among
premenopausal women. Arch Intern Med. 2006; 166:2253-2259.
doi:10.1001/archinte.166.20.2253
391. Boeke CE, Eliassen AH, Chen WY, et al. Dietary fat intake in relation to lethal breast
cancer in two large prospective cohort studies. Breast Cancer Res Treat.
198
2014;146:383-392. doi:10.1007/s10549-014-3005-8
392. Carey LA, Perou CM, Livasy CA, et al. Race, breast cancer subtypes, and survival in
the Carolina Breast Cancer Study. JAMA. 2006; 295:2492-2502.
doi:10.1001/jama.295.21.2492
393. Blows FM, Driver KE, Schmidt MK, et al. Subtyping of breast cancer by
immunohistochemistry to investigate a relationship between subtype and short and long
term survival: a collaborative analysis of data for 10,159 cases from 12 studies. PLoS
Med. 2010; 7:e1000279. doi:10.1371/journal.pmed.1000279
394. Tamimi RM, Colditz GA, Hazra A, et al. Traditional breast cancer risk factors in relation
to molecular subtypes of breast cancer. Breast Cancer Res Treat. 2012; 131:159-167.
doi:10.1007/s10549-011-1702-0
395. Sieri S, Chiodini P, Agnoli C, et al. Dietary fat intake and development of specific
breast cancer subtypes. J Natl Cancer Inst. 2014; 106:dju068. doi:10.1093/jnci/dju068
396. McGee EE, Kim CH, Wang M, et al. Erythrocyte membrane fatty acids and breast
cancer risk by tumor tissue expression of immuno-inflammatory markers and fatty acid
synthase: a nested case-control study. Breast Cancer Res. 2020; 22:78.
doi:10.1186/s13058-020-01316-4
397. Astrup A, Dyerberg J, Elwood P, et al. The role of reducing intakes of saturated fat in
the prevention of cardiovascular disease: where does the evidence stand in 2010? Am
J Clin Nutr. 2011; 93:684-688. doi:10.3945/ajcn.110.004622
398. Bendsen NT, Stender S, Szecsi PB, et al. Effect of industrially produced trans fat on
markers of systemic inflammation: evidence from a randomized trial in women. J Lipid
Res. 2011; 52:1821-1828. doi:10.1194/jlr.M014738
399. Teng KT, Voon PT, Cheng HM, Nesaretnam K. Effects of partially hydrogenated, semi-
saturated, and high oleate vegetable oils on inflammatory markers and lipids. Lipids.
2010; 45:385-392. doi:10.1007/s11745-010-3416-1
400. Overby NC, Sonestedt E, Laaksonen DE, Birgisdottir BE. Dietary fiber and the
glycemic index: a background paper for the Nordic Nutrition Recommendations 2012.
Food Nutr Res. 2013;57:10.3402/fnr.v57i0.20709. doi:10.3402/fnr.v57i0.20709
401. Wolever TM. Is glycaemic index (GI) a valid measure of carbohydrate quality?. Eur J
Clin Nutr. 2013; 67:522-531. doi:10.1038/ejcn.2013.27
402. Augustin LS, Kendall CW, Jenkins DJ, et al. Glycemic index, glycemic load and
glycemic response: An International Scientific Consensus Summit from the International
Carbohydrate Quality Consortium (ICQC). Nutr Metab Cardiovasc Dis. 2015; 25:795-
815. doi:10.1016/j.numecd.2015.05.005
403. Greenwood DC, Threapleton DE, Evans CE, et al. Glycemic index, glycemic load,
carbohydrates, and type 2 diabetes: systematic review and dose-response meta-
analysis of prospective studies. Diabetes Care. 2013; 36:4166-4171. doi:10.2337/dc13-
0325
199
404. Biddinger SB, Ludwig DS. The insulin-like growth factor axis: a potential link between
glycemic index and cancer. Am J Clin Nutr. 2005; 82:277-278.
doi:10.1093/ajcn.82.2.277
405. Samani AA, Yakar S, LeRoith D, Brodt P. The role of the IGF system in cancer growth
and metastasis: overview and recent insights. Endocr Rev. 2007; 28:20-47.
doi:10.1210/er.2006-0001
406. Brand-Miller JC. Glycemic load and chronic disease. Nutr Rev. 2003; 61:S49-S55.
doi:10.1301/nr.2003.may.S49-S55
407. Xie M, Liu J, Tsao R, Wang Z, Sun B, Wang J. Whole Grain Consumption for the
Prevention and Treatment of Breast Cancer. Nutrients. 2019; 11:1769.
doi:10.3390/nu11081769
408. Chandalia M, Garg A, Lutjohann D, von Bergmann K, Grundy SM, Brinkley LJ.
Beneficial effects of high dietary fiber intake in patients with type 2 diabetes mellitus. N
Engl J Med. 2000; 342:1392-1398. doi:10.1056/NEJM200005113421903
409. Montonen J, Boeing H, Fritsche A, et al. Consumption of red meat and whole-grain
bread in relation to biomarkers of obesity, inflammation, glucose metabolism and
oxidative stress. Eur J Nutr. 2013; 52:337-345. doi:10.1007/s00394-012-0340-6
410. Sisti JS, Collins LC, Beck AH, Tamimi RM, Rosner BA, Eliassen AH. Reproductive risk
factors in relation to molecular subtypes of breast cancer: Results from the nurses'
health studies. Int J Cancer. 2016; 138:2346-2356. doi:10.1002/ijc.29968
411. Martínez ME, Cruz GI, Brewster AM, Bondy ML, Thompson PA. What can we learn
about disease etiology from case-case analyses? Lessons from breast cancer. Cancer
Epidemiol Biomarkers Prev. 2010; 19:2710-2714. doi:10.1158/1055-9965.EPI-10-0742
412. Kim H, Caulfield LE, Garcia-Larsen V, et al. Plant-Based Diets and Incident CKD and
Kidney Function. Clin J Am Soc Nephrol. 2019; 14:682-691.
doi:10.2215/CJN.12391018
413. Gómez-Donoso C, Martínez-González MÁ, Martínez JA, et al. A Provegetarian Food
Pattern Emphasizing Preference for Healthy Plant-Derived Foods Reduces the Risk of
Overweight/Obesity in the SUN Cohort. Nutrients. 2019; 11:1553.
doi:10.3390/nu11071553
414. Baden MY, Satija A, Hu FB, Huang T. Change in Plant-Based Diet Quality Is
Associated with Changes in Plasma Adiposity-Associated Biomarker Concentrations in
Women. J Nutr. 2019; 149:676-686. doi:10.1093/jn/nxy301
415. Kim H, Caulfield LE, Rebholz CM. Healthy Plant-Based Diets Are Associated with
Lower Risk of All-Cause Mortality in US Adults. J Nutr. 2018; 148:624-631.
doi:10.1093/jn/nxy019
416. Baden MY, Liu G, Satija A, et al. Changes in Plant-Based Diet Quality and Total and
Cause-Specific Mortality. Circulation. 2019; 140:979-991.
doi:10.1161/CIRCULATIONAHA.119.041014
200
417. Rose DP, Komninou D, Stephenson GD. Obesity, adipocytokines, and insulin
resistance in breast cancer. Obes Rev. 2004; 5:153-165. doi:10.1111/j.1467-
789X.2004.00142.x
418. Fung TT, Hu FB, Hankinson SE, Willett WC, Holmes MD. Low-carbohydrate diets,
dietary approaches to stop hypertension-style diets, and the risk of postmenopausal
breast cancer. Am J Epidemiol. 2011; 174:652-660. doi:10.1093/aje/kwr148
419. Hirko KA, Willett WC, Hankinson SE, et al. Healthy dietary patterns and risk of breast
cancer by molecular subtype. Breast Cancer Res Treat. 2016; 155:579-588.
doi:10.1007/s10549-016-3706-2
420. Fung TT, Hu FB, McCullough ML, Newby PK, Willett WC, Holmes MD. Diet quality is
associated with the risk of estrogen receptor-negative breast cancer in postmenopausal
women. J Nutr. 2006; 136:466-472. doi:10.1093/jn/136.2.466
421. Fung TT, Chiuve SE, Willett WC, Hankinson SE, Hu FB, Holmes MD. Intake of specific
fruits and vegetables in relation to risk of estrogen receptor-negative breast cancer
among postmenopausal women. Breast Cancer Res Treat. 2013; 138:925-930.
doi:10.1007/s10549-013-2484-3
422. Farvid MS, Cho E, Eliassen AH, Chen WY, Willett WC. Lifetime grain consumption and
breast cancer risk. Breast Cancer Res Treat. 2016; 159:335-345. doi:10.1007/s10549-
016-3910-0
423. Sánchez-Quesada C, Romanos-Nanclares A, Navarro AM, et al. Coffee consumption
and breast cancer risk in the SUN project. Eur J Nutr. 2020; 59:3461-3471.
doi:10.1007/s00394-020-02180-w
424. Chen S, Chen Y, Ma S, et al. Dietary fibre intake and risk of breast cancer: A
systematic review and meta-analysis of epidemiological studies. Oncotarget. 2016;
7:80980-80989. doi:10.18632/oncotarget.13140
425. Reynolds A, Mann J, Cummings J, Winter N, Mete E, Te Morenga L. Carbohydrate
quality and human health: a series of systematic reviews and meta-analyses.
Lancet. 2019; 393:434–445. doi: 10.1016/S0140-6736(18)31809-9
426. Maru GB, Hudlikar RR, Kumar G, Gandhi K, Mahimkar MB. Understanding the
molecular mechanisms of cancer prevention by dietary phytochemicals: From
experimental models to clinical trials. World J Biol Chem. 2016; 7:88-99.
doi:10.4331/wjbc. v7.i1.88
427. Liu RH. Health benefits of fruit and vegetables are from additive and synergistic
combinations of phytochemicals. Am J Clin Nutr. 2003; 78:517S-520S.
doi:10.1093/ajcn/78.3.517S
428. Chandran U, McCann SE, Zirpoli G, et al. Intake of energy-dense foods, fast foods,
sugary drinks, and breast cancer risk in African American and European American
women. Nutr Cancer. 2014; 66:1187-1199. doi:10.1080/01635581.2014.951737
201
429. Fiolet T, Srour B, Sellem L, et al. Consumption of ultra-processed foods and cancer
risk: results from NutriNet-Santé prospective cohort. BMJ. 2018; 360: k322.
doi:10.1136/bmj. k322
430. Van den Brandt PA, Schulpen M. Mediterranean diet adherence and risk of
postmenopausal breast cancer: results of a cohort study and meta-analysis. Int J
Cancer. 2017;140:2220-2231. doi:10.1002/ijc.30654
431. Buckland G, Travier N, Cottet V, et al. Adherence to the mediterranean diet and risk of
breast cancer in the European prospective investigation into cancer and nutrition cohort
study. Int J Cancer. 2013;132:2918-2927. doi:10.1002/ijc.27958
432. Cade JE, Taylor EF, Burley VJ, Greenwood DC. Does the Mediterranean dietary
pattern or the Healthy Diet Index influence the risk of breast cancer in a large British
cohort of women? Eur J Clin Nutr. 2011; 65:920-928. doi:10.1038/ejcn.2011.69
433. Trichopoulou A, Bamia C, Lagiou P, Trichopoulos D. Conformity to traditional
Mediterranean diet and breast cancer risk in the Greek EPIC (European Prospective
Investigation into Cancer and Nutrition) cohort. Am J Clin Nutr. 2010; 92:620-625.
doi:10.3945/ajcn.2010.29619
434. Couto E, Sandin S, Löf M, Ursin G, Adami HO, Weiderpass E. Mediterranean dietary
pattern and risk of breast cancer. PLoS One. 2013; 8:e55374.
doi:10.1371/journal.pone.0055374
435. Ghosn, B., Benisi-Kohansal, S., Ebrahimpour-Koujan, S. et al. Association between
healthy lifestyle score and breast cancer. Nutr J. 2020; 19: 4.
https://doi.org/10.1186/s12937-020-0520-9
436. Bertrand KA, Bethea TN, Rosenberg L, et al. Risk factors for estrogen receptor positive
ductal carcinoma in situ of the breast in African American women. Breast. 2020;
49:108-114. doi:10.1016/j.breast.2019.10.009
437. Galceran J, Ameijide A, Carulla M, et al. Cancer incidence in Spain, 2015. Clin Transl
Oncol. 2017;19:799-825. doi:10.1007/s12094-016-1607-9
438. Satija A, Yu E, Willett WC, Hu FB. Understanding nutritional epidemiology and its role
in policy. Adv Nutr. 2015; 6:5-18. doi:10.3945/an.114.007492
439. Martin-Moreno JM, Boyle P, Gorgojo L, et al. Development and validation of a food
frequency questionnaire in Spain. Int J Epidemiol. 1993;22(3):512-519.
440. Rothman KJ, Greenland S, Associate TLL. Modern Epidemiology, 3rd Edition.
Hastings Cent Rep. 2014.
441. Carlos S, De La Fuente-Arrillaga C, Bes-Rastrollo M, et al. Mediterranean Diet and
Health Outcomes in the SUN Cohort. Nutrients. 2018;10:439. doi:10.3390/nu10040439
442. Pérez-Jiménez J, Neveu V, Vos F, Scalbert A. Systematic analysis of the content of
502 Polyphenols in 452 foods and beverages: An application of the phenol-explorer
database. J Agric Food Chem. 2010;58:4959-4969. doi:10.1021/jf100128b
443. Ries LAG, Young JL, Keel GE, Eisner MP, Lin YD, Horner M-J (editors). SEER
Survival Monograph: Cancer Survival Among Adults: U.S. SEER Program, 1988-2001,
202
Patient and Tumor Characteristics. National Cancer Institute, SEER Program, NIH Pub.
No. 07-6215, Bethesda, MD, 2007.
444. Collins LC, Marotti JD, Baer HJ, Tamimi RM. Comparison of estrogen receptor results
from pathology reports with results from central laboratory testing. J Natl Cancer Inst.
2008; 100:218-221. doi:10.1093/jnci/djm270
445. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer
statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36
cancers in 185 countries. CA Cancer J Clin. 2018;0. doi:10.3322/caac.21492
446. Lauby-Secretan B, Scoccianti C, Loomis D, Grosse Y, Bianchini F, Straif K. Body
Fatness and Cancer — Viewpoint of the IARC Working Group. N Engl J Med.
2016;375:794-798. doi:10.1056/NEJMsr1606602
447. Sullivan, K, Fisher, R. NIH Nutrition Research Funding, Fiscal Years 2012-2016. The
FASEB Journal. 2018; S1: 787.2-787.2
448. Gibbons H, Brennan L. Metabolomics as a tool in the identification of dietary
biomarkers. Proc Nutr Soc. 2017; 76:42-53. doi:10.1017/S002966511600032X
449. Dragsted LO, Gao Q, Scalbert A, et al. Validation of biomarkers of food intake-critical
assessment of candidate biomarkers. Genes Nutr. 2018; 13:14. doi:10.1186/s12263-
018-0603-9
450. Greenland P. Dietary Adherence in a Clinical Trial of a Nutritional and Behavioral
Intervention. JAMA. 2019;322:1500. doi:10.1001/jama.2019.15855
203
APPENDICES
204
1. Publications
Title: Sugar-sweetened beverage consumption and incidence of breast cancer: the
Seguimiento Universidad de Navarra (SUN) Project
Journal: European Journal of Nutrition
Impact Factor: 4.663
Status: Published in October 2018
European Journal of Nutrition (2019) 58:2875–2886
https://doi.org/10.1007/s00394-018-1839-2
ORIGINAL CONTRIBUTION
Sugar-sweetened beverage consumption and incidence of breast

cancer: the Seguimiento Universidad de Navarra (SUN) Project
A. Romanos-Nanclares1 · Estefania Toledo1,2,3 · I. Gardeazabal1,4 · J. J. Jiménez-Moleón5,6 ·
M. A. Martínez-González1,2,3,7 · A. Gea1,2,3
Received: 11 June 2018 / Accepted: 27 September 2018 / Published online: 3 October 2018
© Springer-Verlag GmbH Germany, part of Springer Nature 2018
Abstract
Purpose Breast cancer (BC) incidence is increasing worldwide. Higher insulin resistance may potentially lead to an increased
risk of BC. Sugar-sweetened beverages (SSB) are an acknowledged dietary factor that increases insulin resistance. How-
ever, the association between SSB and BC has not been widely explored. We evaluated the association between baseline
consumption of SSB and the incidence of BC among relatively young women in a cohort of Spanish university graduates.
Methods We evaluated 10,713 middle-aged, Spanish female university graduates (median age 33) from the Seguimiento
Universidad de Navarra (SUN) cohort, initially free of BC. SSB consumption was collected at baseline using a validated 136-
item semi-quantitative food-frequency questionnaire. Incidence of BC was confirmed by a trained oncologist using medical
records. We fitted Cox regression models to assess the relationship between baseline categories of SSB consumption and
the incidence of BC during follow-up. We stratified the analyses by menopausal status.
Results During 106,189 person-years follow-up, 100 incident cases of BC were confirmed. Among postmenopausal women,
regular consumption of SSB was associated with a significantly higher incidence of BC (HR 2.12; 95% CI 1.02, 4.41) in the
fully adjusted model, compared to women who never or seldom consumed SSB. No association was found among premeno-
pausal women (HR 1.16; 95% CI 0.66, 2.07).
Conclusions Even though the number of cases was small, in this Mediterranean cohort, we observed a direct association
between SSB consumption and BC risk among postmenopausal women. Nonetheless further larger longitudinal studies are
needed to support this association.
Keywords Sugar-sweetened beverage · Breast cancer · Cohort · Primary prevention
Abbreviations CI Confidence interval

BC Breast cancer BMI Body mass index
SSB Sugar-sweetened beverage MET Metabolic equivalent index
SUN Seguimiento Universidad de Navarra (University SD Standard deviation
of Navarra Follow-up Study)
HR Hazard ratio
OR Odds ratio
5
* Estefania Toledo CIBER Epidemiología y Salud Pública (CIBERESP),
etoledo@unav.es Madrid, Spain
6
1 Department of Preventive Medicine and Publich Health,
University of Navarra, Department of Preventive Medicine
School of Medicine, University of Granada, Instituto de
and Public Health, C/ Irunlarrea, 1, 31008 Pamplona, Spain
Investigación Biosanitaria de Granada ibs-GRANADA,
2
IdiSNA, Navarra Institute for Health Research, Pamplona, Servicio Andaluz de Salud/Universidad de Granada,
Spain Granada, Spain
3 7
CIBERobn Physiopathology of Obesity and Nutrition, Department of Nutrition, Harvard T.H. Chan School
Institute of Health Carlos III (ISCIII), Madrid, Spain of Public Health, Boston, USA
4
Department of Oncology, University of Navarra Clinic,
Pamplona, Spain
13
Vol.:(0123456789)
2876 European Journal of Nutrition (2019) 58:2875–2886
Introduction risk of postmenopausal BC in a prospective cohort [18].

But no other cohort study has replicated this finding so far.
According to the World Health Organization, breast cancer We assessed if SSB consumption was associated with
(BC) is the most commonly diagnosed malignant tumor the incidence of premenopausal BC or postmenopausal BC
after lung cancer in 183 countries or territories and it is among relatively young women in a cohort of Spanish uni-
the first cause of prematurely lost years of life in women in versity graduates.
many countries. In 2012, it accounted for 1.7 million new
cases reported globally [1, 2]. In fact, BC was the main Methods
cause of death in 2015 for women and remained the fifth
cause of the global burden of disease [3]. Study sample
While some genetic factors may determine a predisposi-
tion to develop BC, 90–95% of BC cases are attributable This work was conducted within the “Seguimiento Universi-
to environmental and lifestyle factors [4]. Several well- dad de Navarra” (SUN) Project which is a Spanish dynamic,
established links have been identified between environ- multipurpose and prospective cohort study. Participants are
mental factors and BC incidence, including reproductive all former university students (i.e., university graduates)
factors [2, 5–8] and lifestyles, such as alcohol consumption and the recruitment is permanently open. Participants are
[9–11], overconsumption of food leading to body fatness contacted biennially for follow-up purposes. Details of its
or adult weight gain [12], and lack of physical activity [13, design, recruitment and follow-up methods have been pub-
14]. The role of diet in BC prevention has been widely lished in detail elsewhere [24].
explored during decades [15]. Reducing caloric intake In the SUN project, 13,843 women had been recruited
and better body weight management have been suggested and had completed the baseline questionnaire by Decem-
as leading BC-prevention strategies [16, 17]. Contrarily, ber 2016. To warrant a minimum follow-up time of 2 years,
diets rich in high-calorie foods, such as fatty and sugary we included in the present analysis those women recruited
foods, may lead to increased calorie intake, thereby pro- before March 2014 (n = 13,645). Additionally, women were
moting obesity and leading to an increased risk of cancer excluded if they were lost to follow-up (n = 1286), had a pre-
[18]. But there is still limited evidence about the role of vious history of BC at baseline (n = 102), presented implau-
diet on BC risk, and most foods and food groups remain sible energy intakes using a predefined allowable range of
uncharted. Specifically, despite the potential link between 500–3500 kcal/day for women [25] (n = 1345), or reported
insulin resistance and BC risk [2], there is scarce evidence menopause before 35 years of age (n = 199). Therefore, our
regarding the association between dietary factors acknowl- final sample for the analysis of SSB consumption and inci-
edged as responsible for increased insulin resistance and dence of BC included 10,713 women (Fig. 1).
BC [2]. These factors include sugars, sugary food and The present study was conducted according to the guide-
sugar-sweetened beverages (SSB). lines laid down in the Declaration of Helsinki and all pro-
Frequent consumption of SSB is responsible for cedures involving participants were approved by the Insti-
increased insulin resistance as it has been causally associ- tutional Review Board of the University of Navarra. All
ated with obesity, and particularly in the abdominal region, potential participants were duly informed of their right to
in observational and experimental studies [19, 20], along refuse to participate in the SUN study or to withdraw their
with metabolic syndrome [21], non-alcoholic fatty liver consent to participate at any time without reprisal. Special
disease and type 2 diabetes [20]. Therefore, it is appar- attention was given to the specific information needs of indi-
ent that SSB can lead to endocrine outcomes related to vidual potential candidates, as well as to the methods used to
BC. SSB consumption was associated with a younger age deliver their information and the feedback that may receive
at menarche while sugar-free (diet) beverage and fruit in the future from the research team. After ensuring that
juice consumption was not related to it in a prospective the candidate had understood the information, we sought
study [22]. Moreover, high intake of refined sugar has their potential freely given informed consent through their
been linked to increased mammographic breast density voluntary completion of the baseline questionnaire. These
and, particularly, high consumption of sweet foods among methods were accepted by our Institutional Review Board as
postmenopausal women and higher consumption of SSB to imply an appropriately-obtained informed consent.
among premenopausal have been reported to be associated
with higher mammographic breast density [23]. Assessment of sugar-sweetened beverage
Recently, a higher SSB consumption was reported to consumption
be marginally significantly associated with an increased
Dietary habits at baseline were assessed using a 136-item
semi-quantitative food-frequency questionnaire previously
13
European Journal of Nutrition (2019) 58:2875–2886 2877
Fig. 1 Flow-chart of par- n=22,564 participants

ticipants recruited in the SUN
Project, 1999–2016
n=8,721 men
n=13,843 women
n=198 participants recruited after

March 1st 2014*
n=13,645 women
n=1,286 participants without follow-up
n=12,359 women
n=102 women with previous breast

cancer
n=12,257 women
n=1,345 women with energy intake out

of predefined limits
n=10,912 women
n=199 women with menopause before

the age of 35
n=10,713 women
*To warrant a minimum follow-up of 2 years.
validated in Spain [26, 27], including a specific item on Ascertainment of incident breast cancer cases
SSB consumption. For each food item, there were nine
possible answers (never/seldom, 1–3 servings/month, 1 For the present analysis, incident BC was the primary end-
serving/week, 2–4 servings/week, 5–6 servings/week, 1 point. Participants were asked whether they had received a
serving/day, 2–3 servings/day, 4–6 servings/day, and > 6 medical diagnosis of BC at baseline and during follow-up.
serving/day). A trained team of dietitians using available For the present analysis, if participants reported a medical
information of food composition tables for Spain derived diagnosis of BC at baseline they were considered to have
the nutrient intakes. We did not considered artificially- prevalent BC. Incident BC was defined as women who
sweetened (non-caloric) beverages. develop BC during the follow-up and were free of the dis-
For the principal analysis, participants were classified ease at the beginning of the study.
into three categories of consumption: never/seldom drink- They were also inquired about the date of diagnosis and
ers, > 0 to ≤1 serving/week and > 1 servings/week. Due asked for a copy of their medical records. Then a trained
to the limited consumption of SSBs and as a result of the oncologist confirmed the cases. Moreover, deaths were
small number of incident cases, we categorized frequency reported to the research team by subject’s next of kin, work
of consumption as never/seldom (< 1 serving/month) or associates and postal authorities. If lost during the follow-
any regular consumption (≥ 1 serving/month) when we up or with unidentified causes of death, the National Death
stratified the analyses by menopausal status. We also Index was consulted periodically to identify deceased
repeated the main analysis with the latter categories. cohort members and confirm the cause of death. Confirmed
13
incident cases were represented by both cases adjudicated by reference category and with age as underlying time-variable.
a trained oncologist or considered as a cause of breast cancer Categories of age and period of recruitment were used as
death (n = 100). These trained oncologists were blinded with stratification variables.
respect to dietary exposures. All confirmed cases (n = 100) For all analyses, we fitted an age-adjusted model, and two
jointly with self-reported BC diagnoses (not yet confirmed; multivariable models after adjusting for potential confound-
n = 68) were considered for the analyses on probable inci- ers. Model 1 was adjusted for height (continuous), number
dent cases (n = 168). The main analyses were based only on of relatives with history of BC (0, 1 or 2 relatives), smok-
confirmed BC cases. ing status (never smoker, former smoker, current smoker),
physical activity (MET-h/week, continuous), alcohol intake
Dietary and non-dietary assessments (g/day, continuous), BMI (tertiles), age of menarche (< 10
years, 10–11 years, 12–13 years, 14–16 years, > 16 years),
Mediterranean diet adherence was assessed using the score menopause (no menopause, < 50 years, ≥ 50 years), number
proposed by Trichopoulou et al. [28], which consists of a of pregnancies of more than 6 months (continuous), preg-
9-point Mediterranean-diet scale ranging from 0 to 9 with nancy before the age of 30 years (yes/no), months of breast-
higher scores indicating greater adherence. It includes veg- feeding (continuous), use of hormone replacement therapy
etables, legumes, fruits and nuts, cereal, and fish, alcohol, (yes/no) and its duration (continuous) and years at university
meat, dairy products and the monounsaturated/saturated (continuous).
ratio. The baseline questionnaire also included informa- Model 2 included additional adjustment for diabetes
tion about participants’ socio-demographic characteristics, (yes/no), glycemic index (continuous; excluding SSB for
medical history (prevalent chronic diseases), health-related its calculation) total energy intake (kcal/day, continuous),
habits, lifestyles, anthropometric data and physical activity ultra-processed consumption (continuous), daily servings of
[29, 30]. Ultraproccessed foods were defined as the total coffee (continuous) and decaffeinated coffee (continuous),
amount of the following food items: petit suisse, breakfast and Mediterranean diet (continuous).
cereals, milkshakes, custard, pudding, ice-cream, potato We rerun our analyses for confirmed BC cases by exclud-
chips, cookies, muffins, donuts, croissant, cakes, churros, ing those women with a probable BC.
chocolates, nougat, marzipan, ready-to-consume pies, pizza, To reduce the number of covariates, we fitted multivariate
instant soups and creams, margarine, mayonnaise, sugar- model 3, adjusted for tertiles of a propensity score estimat-
sweetened beverages, diet soft drinks, bottled fruit juice, ing the probability of being exposed to SSB. The probability
processed meat (cooked ham, spicy pork sausage, salami, of being exposed to SSB given covariates in model 2 was
mortadella, foie-gras, black pudding, bacon, hamburger, estimated using binary or multinomial logistic regression.
sausage), and alcoholic drinks produced by fermentation Tests of linear trend were conducted assigning to each
followed by distillation such as whisky, gin, rum or vodka. category of SSB consumption its median and using the
resulting variable as continuous in the above-mentioned
Statistical analysis models.
Next in order, we carried out stratified analysis among
We described the baseline characteristics of the participants pre- or postmenopausal women and their risk of BC. Infor-
across the three categories of SSB consumption. To outline mation on age at menopause was collected in the baseline
baseline characteristics, we used percentages for categorical questionnaire and after 16 years of follow-up. If there was
variables and mean and standard deviations for quantitative lacking information regarding the age at menopause, we
variables. p values were estimated with ANOVA for continu- imputed postmenopausal status according to the 75th per-
ous variables and χ2 test for categorical variables. centile of age at menopause in the study population (52 years
To assess the relationship between baseline total SSB of age) [31]. When assessing premenopausal BC as the out-
consumption (categorized as never/seldom drinkers, ≤ 1 come, we excluded those women who reported having had
serving/week and > 1 servings/week) and the subsequent menopause before study inception and censored follow-up
risk of BC, we used Cox regression models, and included time at the age of 52 years or at the age of menopause. For
confirmed incident cases as the outcome. We calculated postmenopausal BC, we considered women at risk only
person-years of follow-up for each participant from the date those after having turned 52 years old or having had their
of fulfillment of the baseline questionnaire to the date of BC menopause. In the analysis with postmenopausal BC as out-
diagnosis for cases and to the last returned questionnaire come, we additionally adjusted for time since recruitment
for non-cases. Probable incident cases were considered as until the beginning of the time at risk and age at menopause
non-cases and censored at the date of self-reported diagno- (< 50 years, ≥ 50 years).
sis. Hazard ratios (HR) and their 95% CI were calculated As an additional analysis, we performed sensitivity analy-
considering the lowest category of SSB consumption as the ses to explore the robustness of our findings and to account
13
for potential uncertainties in our assumptions with respect to CI 0.66, 2.07; p for trend 0.602) compared to those with the
possible sources of bias or measurement errors. Therefore, we lowest consumption in the fully-adjusted model (Table 3).
further assessed the association between categories of base- On the other hand, postmenopausal women with regular SSB
line SSB consumption and the risk of overall BC and pre- consumption showed a more than twofold higher incidence
menopausal and postmenopausal BC including also probable of BC, (adjusted HR 2.12; 95% CI 1.02, 4.41; p for trend
BC cases and taking into account other energy intake limits, 0.045) in the fully adjusted model in comparison with post-
excluding those beneath the percentile 1 or above the percen- menopausal women in the category of lowest consumption.
tile 99 (P1–P99). Then we conducted our analyses over 11,787 When we adjusted by tertiles of the propensity score based
subjects free of BC at baseline with these latter energy limits. on variables in model 2, the results barely changed (Tables 2,
Analyses were carried out using Stata version 12.0 (Stata 3). In sensitivity analyses, the results barely changed (HR
Corporation). All p values were two-tailed and a p < 0.05 for regular vs. never/seldom SSB consumption among post-
was deemed as significant. menopausal women: 2.12; 95% CI 1.05, 4.28; p for trend
0.037) in the fully adjusted model when we adopted other
Results energy limits (P1–P99) (Fig. 2).
The main baseline characteristics of the 10,713 women

included in our analyses according to categories of baseline Discussion
total SSB consumption are presented in Table 1. The median
age of participants was 33 years and the mean BMI was The results from this prospective study of Mediterranean
22.2 kg/m2 (SD 3.1). Participants in the highest category of university graduates support the hypothesis that a regular
total SSB consumption were younger, were less physically consumption of SSB at baseline (as compared with no or
active, had a higher BMI, higher total energy intake and seldom consumption) was associated with higher risk of BC
ultra-processed food consumption, higher intake of alcohol, among postmenopausal women. Otherwise, our results were
meat, dairy products and fat intake, lower consumption of inconclusive regarding SSB consumption and the risk of BC
fruit, vegetables, fish, and olive oil, and lower adherence to among premenopausal women.
the Mediterranean diet compared with women with never or Some studies have found a direct association between
seldom consumption of SSB. sweet drinks, sweet foods or a higher glycemic load and
postmenopausal BC risk [32–34]. These studies found a
Incidence of breast cancer direct association between frequent consumption of sweets
that included beverages, desserts, foods and spoonsful of
We identified 168 probable incident cases of BC. Out of added sugar and high glycemic index or glycemic load and,
them, we confirmed with medical records 100 new-onset therefore, a potential role of hyperinsulinemia or insulin
cases of BC among 106,189 person-years of follow-up resistance in BC development.
(mean follow-up 10.0 years). Our results are consistent with a previous case–control
When the association between SSB consumption and study that evaluated the consumption of SSBs in relation
overall confirmed BC incidence was assessed adjusting for to BC risk in European American and African American
age and for several dietary and non-dietary potential con- women in the US [35]. They observed a direct association
founders (multivariable model 2), we found that a higher between sugary drinks consumption (> 1 serving/week) and
SSB consumption was monotonically associated in the risk of postmenopausal BC in European American women
point estimates with higher incidence of BC, although this (OR 2.05; 95% CI 1.13–3.70; p for trend 0.02) compared to
association was not statistically significant (Table 2). These those who did not consume any SSBs. More recently, the
results did not differ when we used alternative energy limits Melbourne Collaborative Cohort Study [18] showed direct
(> percentile 1 and < percentile 99) in the sensitivity analy- associations between sugar-sweetened soft drink consump-
ses (Fig. 2), when we excluded women with probable BC tion and obesity-related cancers, particularly, for postmeno-
(without confirmation) or when we considered also probable pausal BC. The highest risk was found for an exposure of
cases as outcome. 1–6 drinks/week (HR 1.21; 95% CI 1.04–1.57), compared
with never consumers. Nonetheless, there is also evidence
Incidence of premenopausal and postmenopausal signaling a direct association between SSB consumption and
breast cancer premenopausal BC [36]. We are not aware of other prospec-
tive cohort studies that have assessed the consumption of
When we divided participants according to menopausal SSBs and the risk of BC.
status, regular consumption of SSBs among premenopausal On the other hand, another study on premenopausal
women was not associated with BC (adjusted HR 1.16; 95% and postmenopausal women found that SSB consumption
13
Table 1 Baseline characteristics of 10,713 women in the SUN cohort according to the categories of sugar-sweetened beverage (SSB) consump-
tion
Variable Consumption of SSB p values*
Never/seldom 1/month to ≤1/week > 1/week
N 4063 4537 2113

Median of SSB consumption (servings/week) 0 1 3
Age (years) 38.8 (11.3) 32.9 (9.6) 31.0 (8.5) < 0.001
Body mass index (kg/m2) 22.5 (3.1) 22.1 (3.0) 22.0 (3.2) < 0.001
Physical activity (METs-h/week) 25.9 (21.8) 23.7 (19.4) 22.4 (19.0) < 0.001
Total energy intake (kcal/day) 2193 (585) 2322 (556) 2440 (550) < 0.001
Alcohol intake (g/day) 4.1 (6.7) 3.5 (4.9) 5.0 (6.0) < 0.001
Years at university 4.9 (1.4) 4.8 (1.3) 4.8 (1.3) 0.01
Height (cm) 163 (6) 164 (6) 164 (6) < 0.001
Number of pregnancies of more than 6 months 0.9 (1.3) 0.6 (1.1) 0.4 (1.0) < 0.001
Breast-feeding (months) 3.2 (5.8) 2.1 (4.6) 1.4 (3.9) < 0.001
Time of hormone replacement therapy (years)a 1.3 (2.4) 1.3 (2.4) 1.2 (2.4) < 0.001
Diabetes (%) 1.87 0.75 0.43 < 0.001
Pregnancy before 30 years (%) 25.8 16.8 12.1 < 0.001
Smoking (%) < 0.001
Former smoker 22.9 25.3 35.5
Current smoker 27.1 18.5 15.4
Number of relatives with BC (%)b 0.57
0 88.8 89.7 89.7
1 9.0 8.5 8.2
2 2.2 1.8 2.0
Age of menarche (%) 0.21
< 10 years 1.3 1.2 0.9
10–11 years 20.2 18.2 18.7
12–13 years 53.8 56.0 54.0
14–16 years 22.3 22.2 23.7
> 16 years 2.5 2.5 27
Menopausal status (%) < 0.001
Premenopausal (%) 81.1 92.5 95.4
Postmenopausal (%) 18.9 7.7 4.6
Age at menopausea < 0.001
< 50 years (%) 47.3 55.5 51.2
≥ 50 years (%) 52.7 44.3 49.4
Macronutrient distribution
Carbohydrate intake (% energy) 43.5 (8.1) 43 (6.8) 43.3 (7) 0.01
Protein intake (% energy) 18.9 (3.8) 18.3 (3) 17.6 (3) < 0.001
Fat intake (% energy) 36.2 (7.2) 37.5 (6.2) 37.7 (6.4) < 0.001
Saturated fatty acids (% energy) 11.8 (3.3) 12.8 (3) 13 (3) < 0.001
Monounsaturated fatty acids (% energy) 15.9 (4.2) 16.2 (3.7) 16.1 (3.7) 0.001
Polyunsaturated fatty acids (% energy) 5.0 (1.6) 5.2 (1.6) 5.3 (1.6) < 0.001
Adherence to the Mediterranean dietc 4.5 (1.6) 3.9 (1.7) 3.7 (1.7) < 0.001
Ultra-processed food consumption (g/day) 2.7 (1.6) 3.2 (1.4) 4.4 (1.7) < 0.001
Coffee consumption (servings/day) 1.2 (1.3) 1.2 (1.2) 1.2 (1.3) 0.63
Decaffeinated coffee consumption (servings/day) 1.3 (2.0) 1.3 (1.9) 1.2 (1.8) 0.09
Fruit consumption (g/day) 403 (312) 358 (283) 315 (277) < 0.001
Vegetable consumption (g/day) 614 (384) 546 (318) 515 (323) < 0.001
Olive oil consumption (g/day) 18 (15) 17 (14) 16 (14) < 0.001
13
Table 1 (continued)
Variable Consumption of SSB p values*
Never/seldom 1/month to ≤1/week > 1/week
Cereal consumption (g/day) 97 (69) 97 (64) 94 (63) 0.23

Legume consumption (g/day) 22 (19) 22 (15) 21 (16) 0.34
Meat/meat product consumption (g/day) 155 (78) 177 (74) 186 (76) < 0.001
Fish and seafood consumption (g/day) 101 (65) 96 (56) 92 (59) < 0.001
Dairy product consumption (g/day) 153 (178) 191 (186) 206 (196) < 0.001
Values are expressed as mean (SD), unless otherwise stated

MET metabolic equivalents
*Results from one-way analysis of variance (ANOVA) for continuous variables and χ2 test for categorical variables
a
Only for postmenopausal women
b
Information from mother, sisters, and both grandmothers was collected
c
Score proposed by Trichopoulou et al. [28]
was directly associated with mammographic density and Strengths and limitations
it is known that a higher mammographic density may lead
to a higher risk of BC [23]. The present study has some potential limitations. First, the
Several biological mechanisms may explain the higher number of incident cases of BC was small which repre-
BC incidence among women with higher SSB consump- sents an advantage for our participants but may endanger
tion. Specifically, intake of sugar seems to yield an our statistical power. Our population is largely constituted
effect on breast tumor cells in vitro [37]. Furthermore, by young women, which diminishes the number of incident
an in vitro study on breast tumor cells has shown that BC cases. Second, the use of self-reported information on
glucose may increase cell proliferation while fructose SSB consumption may imply some degree of misclassi-
may promote cell invasion and migration [37]. In humans, fication which in turn, may bias our results towards the
the consumption of highly glycemic foods, such as SSB, null. Nevertheless, dietary information was collected with
leads to a rapid increase in the production of insulin [23, a previously validated semi-quantitative food-frequency
38, 39] and to a down-regulation of sex-hormone bind- questionnaire. Particularly, for SSB the intra-class cor-
ing globulin and an increase in the insulin-like growth relation coefficient between the food-frequency question-
factor-I (IGF-1) production [40, 41]. This latter protein naire and multiple dietary records was 0.67 and in the
has mitogenic and antiapoptotic effects on cells mainly reproducibility analyses, the correlation coefficient was
mediated by the transmembrane tyrosine kinase recep- 0.69 [27, 47]. Third, information on some cases of BC
tor IGF-1R [42]. Furthermore, insulin has been directly was also based on self-reports (when we included prob-
associated to BC risk, particularly in postmenopausal able cases). To avoid false positives, the main results were
women [43]. Insulin suppresses lipolysis, free fatty acids based on confirmed BC cases. Also, we acknowledge that
are released into the circulation and lead to ectopic lipid we might have missed some BC cases. Fourth, our sample
accumulation in fundamental organs (liver, pancreas and was not representative of the general population. Never-
kidneys) which further contribute to insulin resistance, theless, it is unlikely that underlying mechanisms for the
hyperglycemia, dyslipidemia and hypertension [44]. This association between SSB consumption and BC risk are
condition commonly known as metabolic syndrome inte- different for highly educated women than for women with
grates mechanisms that promote tumor progression and a lower educational level. Our choice to select highly edu-
have been involved in the etiology and progression of cated participants corresponds to the approach known as
several cancers, including BC [45]. In postmenopausal restriction in epidemiology and it was applied to control
women, insulin is also linked to the increased peripheral for confounding by socioeconomic status and educational
conversion of androgens to estrogens which, in turn, stim- level. Restriction is usually assumed in epidemiology as
ulates breast cell proliferation by mutations, estradiol’s an excellent technique for preventing or at least reducing
genotoxic metabolites that generate oxygen free radicals confounding by known factors. Thus, the restriction of the
and DNA mutagenic processes [46].
13
2882
13
Table 2 Hazard ratio (HR) and 95% confidence intervals (CI) of confirmed breast cancer cases according to the categories of baseline SSB consumption in 10,713 women of the SUN Project
Consumption of SSB p for trend
Never/seldom 1/month- ≤1/week > 1/week
N 4063 4537 2113

Persons-years of follow-up 38,699 45,921 21,568
Incidence rate/10,000 person- 10.34 9.36 7.88
years
Age-adjusted HR 1 (ref) 1.32 (0.85, 2.05) 1.33 (0.74, 2.40) 0.312
Multivariable adjusted model 1 (ref) 1.31 (0.84, 2.04) 1.30 (0.72, 2.36) 0.359
1
2
3
Consumption of SSB p for trend
Never/seldom Regular consumption
N 4063 6550
Incident cases 40 60
Persons-years of follow-up 38,699 67,489
Incidence rate/10.000 person- 10.34 8.89
years
Multivariable adjusted model 1 (ref) 1.31 (0.86, 2.01) 0.213
2
European Journal of Nutrition (2019) 58:2875–2886

Multivariable adjusted model 1 (ref) 1.32 (0.86, 2.01) 0.201
3
Never/seldom: no consumption, barely intake throughout the year

Regular consumption: ≥1 serving/month
Model 1 additionally adjusted for height, number of relatives with history of BC (0, 1 or 2 relatives), smoking status (never smoker, former smoker, current smoker), physical activity (METs-h/
week, continuous), alcohol intake (g/day, continuous), BMI (tertiles), age of menarche (< 10 years, 10–11 years, 12–13 years, 14–16 years, > 16 years), menopause (no menopause, < 50 years,
≥ 50 years), number of pregnancies of more than 6 months (continuous), pregnancy before the age of 30 years (yes/no), months of breastfeeding (continuous), use of hormone replacement
therapy (yes/no) and its duration (continuous) and years at university (continuous)
Model 2 additionally adjusted for diabetes (yes/no), glycemic index (continuous), total energy intake (kcal/day, continuous), decaffeinated coffee consumption (continuous), coffee consumption
(continuous), ultra-processed food consumption (continuous) and Mediterranean diet adherence (continuous)
Model 3 adjusted for tertiles of propensity score based on variables in model 2
Fig. 2 Sensitivity analyses.

Hazard ratios and 95% CI
of incident BC according to
menopausal status and different
energy cut-off points. The SUN
project (1999–2016)
◼ Probable cases
● Confirmed cases
Energy limits proposed by Willett [25]: >500-<3500 kcal/day
Alternate energy limits: Percentiles 1-99
selection to university graduates could also improve the is that the consumption of SSB might merely represent a
validity of our study as the high educational level increases marker of an overall unhealthy diet, however, we adjusted
the accuracy of self-reports. Moreover, most cohorts and for the overall dietary pattern and the results remained
other analytical studies are usually non-representative, and basically unchanged. Nevertheless, our study also has
this represents no problem given that the generalization some strengths, starting from its longitudinal design that
should be based on biological plausibility. Another caveat reduces the possibility of reverse causation bias and the
13
Table 3 Hazard ratio (HR) and 95% confidence intervals (CI) of confirmed breast cancer cases for each category of SSB consumption among
premenopausal and postmenopausal women of the SUN Project
Consumption of SSB p value


N 3236 6180
Persons-year of follow-up 26,405 59,035
Incidence rate/10.000 person- year 7.57 6.28
Age-adjusted HR 1 (ref) 1.18 (0.67, 2.06) 0.570
Multivariable adjusted model 1 1 (ref) 1.17 (0.67, 2.06) 0.576
Consumption of SSB p value

N 1610 1235
Persons-year of follow-up 10,644 7054
Incidence rate/10.000 person-year 13.15 28.35
Age-adjusted HR 1 (ref) 2.08 (1.04, 4.13) 0.037
Multivariable adjusted model 2a 1 (ref) 2.12 (1.01, 4.41) 0.045
Multivariable adjusted model 3a 1 (ref) 2.05 (1.02, 4.13) 0.045
Never/seldom: no consumption, barely intake throughout the year

Regular consumption: ≥1 serving/month
Model 1 additionally adjusted for height, number of relatives with history of BC (0, 1 or 2 relatives), smoking status (never smoker, former
smoker, current smoker), physical activity (METs-h/week, continuous), alcohol intake (g/day, continuous), BMI (tertiles), age of menarche (< 10
years, 10–11 years, 12–13 years, 14–16 years, > 16 years), number of pregnancies of more than 6 months (continuous), pregnancy before the age
of 30 years (yes/no), months of breastfeeding (continuous), and years at university (continuous)
Model 2 additionally adjusted for diabetes (yes/no), glycemic index (continuous), total energy intake (kcal/day, continuous), decaffeinated coffee
consumption (continuous), coffee consumption (continuous), ultra-processed food consumption (continuous) and Mediterranean diet adherence
(continuous)
Model 3 adjusted for tertiles of propensity score based on variables in model 2
a
Additionally adjusted for time since recruitment, age at menopause (< 50 years, ≥ 50 years), use of hormone replacement therapy (yes/no) and
its duration (continuous)
adjustment for many potential confounders which have research should be conducted to confirm these results in
been included in the multivariable analyses, although larger prospective studies and to better understand the path-
we cannot completely rule out the possibility of residual ways underlying the risk of breast tumors.
confounding.
Restriction of foods high in energy and sugary drinks Acknowledgements The authors thank the implication and collabora-
tion of the participants in the SUN Project. ARN was supported by the
is one of the evidence-based recommendations suggested Fundación Científica Asociación Española Contra el Cáncer (AECC)
by the World Cancer Research Fund (WCRF) published in (Scientific Foundation of the Spanish Association Against Cancer). We
2017 and included in the 2015–2020 Dietary Guidelines for thank the other members of the SUN Group: Alonso A, Barrio López
Americans [48]. Research on particular dietary factors that MT, Basterra-Gortari FJ, Benito Corchón S, Bes-Rastrollo M, Beunza
JJ, Carlos Chillerón S, Carmona L, Cervantes S, de Irala Estévez J, de
conduce to raised energy intake and weight gain is essential la Fuente Arrillaga C, de la Rosa PA, Delgado Rodríguez M, Dominguz
for understanding the pathways related to cancer prevention. LJ, Donat Vargas CL, Donázar M, Eguaras S, Fernández Montero A,
In this Mediterranean cohort study, a regular consumption Galbete Ciáurriz C, García López M, Goñi Ochandorena E, Guillén
of SSB was associated with an increased risk of develop- Grima F, Hernández-Hernandez A, Llorca J, López del Burgo C, Marí
Sanchís A, Martí del Moral A, Martín Calvo N, Martínez JA, Molero P,
ing BC among postmenopausal women, however, further
13
Núñez-Córdoba JM, Pimenta AM, Rico A, Ruiz-Canela M, Ruiz Zam- 8. Collaborative Group on Hormonal Factors in Breast Cancer
brana A, Sánchez Adán D, Sayón Orea C, Vázquez Ruiz Z. All authors (2002) Breast cancer and breastfeeding: collaborative reanaly-
revised the manuscript and read and approved the final version of it. sis of individual data from 47 epidemiological studies in 30
countries, including 50302 women with breast cancer and 96973
Funding The SUN Project has received funding from the Spanish women without the disease. Lancet 360:187–195
Government-Instituto de Salud Carlos III, and the European Regional 9. Petri AL, Tjønneland A, Gamborg M et al (2004) Alcohol
Development Fund (FEDER) (RD 06/0045, CIBER-OBN, Grants intake, type of beverage, and risk of breast cancer in pre- and
PI10/02658, PI10/02293, PI13/00615, PI14/01668, PI14/01798, postmenopausal women. Alcohol Clin Exp Res 28:1084–1090
PI14/01764, PI17/01795, and G03/140), the Navarra Regional Gov- 10. Romieu I, Scoccianti C, Chajès V et al (2015) Alcohol intake
ernment (45/2011, 122/2014, 41/2016), and the University of Navarra. and breast cancer in the European prospective investigation into
cancer and nutrition. Int J Cancer 137:1921–1930. https://doi.
org/10.1002/ijc.29469
Compliance with ethical standards 11. Smith-Warner SA, Spiegelman D, Yaun SS et al (1998) Alcohol
and breast cancer in women: a pooled analysis of cohort studies.
Ethical standards The present study was conducted according to the JAMA 279:535–540
guidelines laid down in the Declaration of Helsinki and all procedures 12. Fortner RT, Katzke V, Kúhn T et al (2016) Obesity and breast
involving participants were approved by the Institutional Review cancer. Recent Results Cancer Res 208:43–65
Board of the University of Navarra. All potential participants were 13. Lahmann PH, Friedenreich C, Schuit AJ et al (2007) Physical
duly informed of their right to refuse to participate in the SUN study activity and breast cancer risk: the european prospective inves-
or to withdraw their consent to participate at any time without reprisal. tigation into cancer and nutrition. Cancer Epidemiol Biomark
Special attention was given to the specific information needs of indi- Prev 16:36–42
vidual potential candidates as well as to the methods used to deliver 14. Wu Y, Zhang D, Kang S (2013) Physical activity and risk of
their information and the feedback that may receive in the future from breast cancer: a meta-analysis of prospective studies. Breast
the research team. After ensuring that the candidate had understood the Cancer Res Treat 137:869–882. https://doi.org/10.1007/s1054
information, we sought their potential freely given informed consent 9-012-2396-7
through their voluntary completion of the baseline questionnaire. These 15. Doll R, Peto R (1981) The causes of cancer: quantitative estimates
methods were accepted by our Institutional Review Board as to imply of avoidable risks of cancer in the united states today. J Natl Can-
an appropriately-obtained informed consent. cer Inst 66:1192–1308
16. World Cancer Research Fund/American Institute for Cancer
Research. Diet, nutrition, physical activity and cancer: a global
Conflict of interest On behalf of all authors, the corresponding author
perspective. Continuous update project expert report 2018. Avail-
declare no conflict of interest.
able at https://www.wcrf.org/dietandcancer. Accessed 6 June 2018
17. Eccles S, Aboagye EO, Ali S, Anderson a S, Armes J, Berditch-
evski F et al (2013) Critical research gaps and translational priori-
ties for the successful prevention and treatment of breast cancer.
Breast Cancer Res 15:92. https://doi.org/10.1186/bcr3493
References 18. Hodge AM, Bassett JK, Milne RL, English DR, Giles GG (2018)
Consumption of sugar-sweetened and artificially sweetened soft
1. World Health Organization (2012) GLOBOCAN. Cancer fact drinks and risk of obesity-related cancers. Public Health Nutr
sheets: breast cancer. http://globocan.iarc.fr/old/FactSheets/cance 21:1618–1626
rs/breast-new.asp. Accessed 6 Jun 2018 19. Malik VS, Popkin BM, Bray GA, Despres JP, Willett WC, Hu
2. World Cancer Research Fund, American Institute for Cancer FB (2010) Sugar-sweetened beverages and risk of metabolic
Research (2017) Diet, nutrition, physical activity and breast can- syndrome and type 2 diabetes: a meta-analysis. Diabetes Care
cer. http://www.aicr.org/continuous-update-project/repor ts/breas 33:2477–2481. https://doi.org/10.2337/dc10-1079
t-cancer-report-2017.pdf. Accessed 6 Jun 2018 20. Hu FB, Malik VS (2010) Sugar-sweetened beverages and risk
3. Global Burden of Disease Cancer Collaboration C, Fitzmaurice of obesity and type 2 diabetes: epidemiologic evidence. Physiol
C, Allen C et al (2017) Global, regional, and national cancer inci- Behav 100:47–54. https://doi.org/10.1016/j.physbeh.2010.01.036
dence, mortality, years of life lost, years lived with disability, and 21. Barrio-Lopez MT, Martinez-Gonzalez M, Fernandez-Montero A,
disability-adjusted life-years for 32 cancer groups, 1990 to 2015: Beunza JJ, Zazpe I, Bes-Rastrollo M (2013) Prospective study of
a systematic analysis for the global burden of disease study. JAMA changes in sugar-sweetened beverage consumption and the inci-
Oncol 3:524–548. https://doi.org/10.1001/jamaoncol.2016.5688 dence of the metabolic syndrome and its components: the SUN
4. Ligibel JA, Alfano CM, Courneya KS, Demark-Wahnefried W, cohort. Br J Nutr 110:1722–1731. https://doi.org/10.1017/S0007
Burger RA, Chlebowski RT et al (2014) American Society of 114513000822
Clinical Oncology position statement on obesity and cancer. J Clin 22. Carwile JL, Willett WC, Spiegelman D, Hertzmark E, Rich-
Oncol 32:3568–3574. https://doi.org/10.1200/JCO.2014.58.4680 Edwards J, Frazier a L et al (2015) Sugar-sweetened beverage
5. Mørch LS, Skovlund CW, Hannaford PC et al (2017) Contempo- consumption and age at menarche in a prospective study of US
rary hormonal contraception and the risk of breast cancer. N Engl girls. Hum Reprod 30:675–683. https://doi.org/10.1093/humrep/
J Med 377:2228–2239. https://doi.org/10.1056/NEJMoa1700732 deu349
6. Gartlehner G, Patel SV, Feltner C et al (2017) Hormone therapy 23. Duchaine CS, Dumas I, Diorio C (2014) Consumption of
for the primary prevention of chronic conditions in postmeno- sweet foods and mammographic breast density: a cross-
pausal women: evidence report and systematic review for the US sectional study. BMC Public Health 14:554. https ://doi.
preventive services task force. JAMA 318:2234–2249. https://doi. org/10.1186/1471-2458-14-554
org/10.1001/jama.2017.16952 24. Seguí-Gómez M, de la Fuente C, Vázquez Z, de Irala J, Martínez-
7. Kobayashi S, Sugiura H, Ando Y, Shiraki N, Yanagi T, Yamashita González MA (2006) Cohort profile: the “Seguimiento Univer-
H et al (2012) Reproductive history and breast cancer risk. Breast sidad de Navarra” (SUN) study. Int J Epidemiol 35:1417–1422.
Cancer 19:302–308. https://doi.org/10.1007/s12282-012-0384-8 https://doi.org/10.1186/1475-2891-10-121
13
25. Willett WC (2013) Issues in analysis and presentation of dietary 36. Witte JS, Ursin G, Siemiatycki J, Thompson WD, Paganini-Hill A,
data. In: Willett WC (ed) Nutritional epidemiology, 3rd edn. Haile RW (1997) Diet and premenopausal bilateral breast cancer:
Oxford University Press, New York, pp 305–333 a case–control study. Breast Cancer Res Treat 42:243–251. https
26. Martin-Moreno JM, Boyle P, Gorgojo L et al (1993) Development ://doi.org/10.1023/A:1005710211184
and validation of a food frequency questionnaire in Spain. Int J 37. Monzavi-Karbassi B, Hine RJ, Stanley JS, Ramani VP, Carcel-
Epidemiol 22:512–519 Trullols J, Whitehead TL et al (2010) Fructose as a carbon source
27. De la Fuente-Arrillaga C, Vázquez Ruiz Z, Bes-Rastrollo M, induces an aggressive phenotype in MDA-MB-468 breast tumor
Sampson L, Martinez-González MA (2010) Reproducibility of cells. Int J Oncol 37:615–622
an FFQ validated in Spain. Public Health Nutr 13:1364–1372. 38. Liu H, Heaney AP (2011) Refined fructose and cancer. Expert
https://doi.org/10.1017/S1368980009993065 Opin Ther Targets 15:1049–1059. https://doi.org/10.1517/14728
28. Trichopoulou A, Costacou T, Bamia C, Trichopoulos D (2003) 222.2011
Adherence to a Mediterranean diet and survival in a Greek popu- 39. Ludwig DS (2003) Diet and development of the insulin resistance
lation. N Engl J Med 348:2599–2608 syndrome. Asia Pac J Clin Nutr 12:4
29. Bes-Rastrollo M, Pérez Valdivieso JR, Sánchez-Villegas A, 40. Kaaks R (1996) Nutrition, hormones, and breast cancer: is insulin
Alonso Á, Martínez-González M (2005) Validación del peso e the missing link? Cancer Causes Control 7:605–625
índice de masa corporal auto-declarados de los participantes de 41. Kaaks R (2001) Plasma insulin, IGF-I and breast cancer.
una cohorte de graduados universitarios. Rev Esp Obes 3:352– Gynecol Obstet Fertil 29:185–191. https://doi.org/10.1016/S1297
358. https://doi.org/10.3305/nh.2013.28.5.6671 -9589(00)00047-3
30. Martínez-González MA, López-Fontana C, Varo JJ et al (2005) 42. Christopoulos PF, Msaouel P, Koutsilieris M (2015) The role of
Validation of the Spanish version of the physical activity question- the insulin-like growth factor-1 system in breast cancer. Mol Can-
naire used in the Nurses’ Health Study and the Health Profession- cer 14:43. https://doi.org/10.1186/s12943-015-0291-7
als’ Follow-up Study. Public Health Nutr 8:920–927 43. Gunter MJ, Hoover DR, Yu H, Wassertheil-Smoller S, Rohan TE,
31. Shivappa N, Sandin S, Löf M et al (2015) Prospective study of Manson JE et al (2009) Insulin, insulin-like growth factor-I, and
dietary inflammatory index and risk of breast cancer in Swed- risk of breast cancer in postmenopausal women. J Natl Cancer Inst
ish women. Br J Cancer 113:1099–1103. https://doi.org/10.1038/ 101:48–60. https://doi.org/10.1093/jnci/djn415
bjc.2015.304 44. Iyengar NM, Hudis CA, Dannenberg AJ (2015) Obesity and can-
32. Augustin LS, Dal Maso L, La Vecchia C, Parpinel M, Negri E, cer: local and systemic mechanisms. Annu Rev Med 66:297–309.
Vaccarella S et al (2001) Dietary glycemic index and glycemic https://doi.org/10.1146/annurev-med-050913-022228
load, and breast cancer risk: a case–control study. Ann Oncol 45. Bhandari R, Kelley G, Hartley T, Rockett I (2014) Metabolic syn-
12:1533–1538 drome is associated with increased breast cancer risk: a systematic
33. Potischman N, Coates RJ, Swanson CA, Carroll RJ, Daling JR, review with meta-analysis. Int J Breast Cancer 2014:1–13. https
Brogan DR et al (2002) Increased risk of early-stage breast cancer ://doi.org/10.1155/2014/189384
related to consumption of sweet foods among women less than age 46. Shapira N (2017) The potential contribution of dietary factors to
45 in the United States. Cancer Causes Control 13:937–946. https breast cancer prevention. Eur J Cancer Prev 26:385–395. https://
://doi.org/10.1023/A:1021919416101 doi.org/10.1097/CEJ.0000000000000406
34. Bradshaw PT, Sagiv SK, Kabat GC, Satia JA, Britton JA, Tei- 47. Fernández-Ballart JD, Piñol JL, Zazpe I, Corella D, Carrasco P,
telbaum SL et al (2009) Consumption of sweet foods and breast Toledo E et al (2010) Relative validity of a semi-quantitative food-
cancer risk: a case–control study of women on Long Island, frequency questionnaire in an elderly Mediterranean population of
New York. Cancer Causes Control 20:1509–1515. https ://doi. Spain. Br J Nutr 103:1808–1816. https://doi.org/10.1017/S0007
org/10.1007/s10552-009-9343-x 114509993837
35. Chandran U, McCann SE, Zirpoli G, Gong Z, Lin Y, Hong CC 48. U.S. Department of Health and Human Services and U.S. Depart-
et al (2014) Intake of energy-dense foods, fast foods, sugary ment of Agriculture (2015) 2015–2020 dietary guidelines for
drinks, and breast cancer risk in African American and Euro- Americans, 8th edn. https://healt h.gov/dietaryguidelines/2015/
pean American women. Nutr Cancer 66:1187–1199. https://doi. guidelines/. Accessed 6 Jun 2018
org/10.1080/01635581
13
25. Willett WC (2013) Issues in analysis and presentation of dietary 36. Witte JS, Ursin G, Siemiatycki J, Thompson WD, Paganini-Hill A,
data. In: Willett WC (ed) Nutritional epidemiology, 3rd edn. Haile RW (1997) Diet and premenopausal bilateral breast cancer:
Oxford University Press, New York, pp 305–333 a case–control study. Breast Cancer Res Treat 42:243–251. https
26. Martin-Moreno JM, Boyle P, Gorgojo L et al (1993) Development ://doi.org/10.1023/A:1005710211184
and validation of a food frequency questionnaire in Spain. Int J 37. Monzavi-Karbassi B, Hine RJ, Stanley JS, Ramani VP, Carcel-
Epidemiol 22:512–519 Trullols J, Whitehead TL et al (2010) Fructose as a carbon source
27. De la Fuente-Arrillaga C, Vázquez Ruiz Z, Bes-Rastrollo M, induces an aggressive phenotype in MDA-MB-468 breast tumor
Sampson L, Martinez-González MA (2010) Reproducibility of cells. Int J Oncol 37:615–622
an FFQ validated in Spain. Public Health Nutr 13:1364–1372. 38. Liu H, Heaney AP (2011) Refined fructose and cancer. Expert
https://doi.org/10.1017/S1368980009993065 Opin Ther Targets 15:1049–1059. https://doi.org/10.1517/14728
28. Trichopoulou A, Costacou T, Bamia C, Trichopoulos D (2003) 222.2011
Adherence to a Mediterranean diet and survival in a Greek popu- 39. Ludwig DS (2003) Diet and development of the insulin resistance
lation. N Engl J Med 348:2599–2608 syndrome. Asia Pac J Clin Nutr 12:4
29. Bes-Rastrollo M, Pérez Valdivieso JR, Sánchez-Villegas A, 40. Kaaks R (1996) Nutrition, hormones, and breast cancer: is insulin
Alonso Á, Martínez-González M (2005) Validación del peso e the missing link? Cancer Causes Control 7:605–625
índice de masa corporal auto-declarados de los participantes de 41. Kaaks R (2001) Plasma insulin, IGF-I and breast cancer.
una cohorte de graduados universitarios. Rev Esp Obes 3:352– Gynecol Obstet Fertil 29:185–191. https://doi.org/10.1016/S1297
358. https://doi.org/10.3305/nh.2013.28.5.6671 -9589(00)00047-3
30. Martínez-González MA, López-Fontana C, Varo JJ et al (2005) 42. Christopoulos PF, Msaouel P, Koutsilieris M (2015) The role of
Validation of the Spanish version of the physical activity question- the insulin-like growth factor-1 system in breast cancer. Mol Can-
naire used in the Nurses’ Health Study and the Health Profession- cer 14:43. https://doi.org/10.1186/s12943-015-0291-7
als’ Follow-up Study. Public Health Nutr 8:920–927 43. Gunter MJ, Hoover DR, Yu H, Wassertheil-Smoller S, Rohan TE,
31. Shivappa N, Sandin S, Löf M et al (2015) Prospective study of Manson JE et al (2009) Insulin, insulin-like growth factor-I, and
dietary inflammatory index and risk of breast cancer in Swed- risk of breast cancer in postmenopausal women. J Natl Cancer Inst
ish women. Br J Cancer 113:1099–1103. https://doi.org/10.1038/ 101:48–60. https://doi.org/10.1093/jnci/djn415
bjc.2015.304 44. Iyengar NM, Hudis CA, Dannenberg AJ (2015) Obesity and can-
32. Augustin LS, Dal Maso L, La Vecchia C, Parpinel M, Negri E, cer: local and systemic mechanisms. Annu Rev Med 66:297–309.
Vaccarella S et al (2001) Dietary glycemic index and glycemic https://doi.org/10.1146/annurev-med-050913-022228
load, and breast cancer risk: a case–control study. Ann Oncol 45. Bhandari R, Kelley G, Hartley T, Rockett I (2014) Metabolic syn-
12:1533–1538 drome is associated with increased breast cancer risk: a systematic
33. Potischman N, Coates RJ, Swanson CA, Carroll RJ, Daling JR, review with meta-analysis. Int J Breast Cancer 2014:1–13. https
Brogan DR et al (2002) Increased risk of early-stage breast cancer ://doi.org/10.1155/2014/189384
related to consumption of sweet foods among women less than age 46. Shapira N (2017) The potential contribution of dietary factors to
45 in the United States. Cancer Causes Control 13:937–946. https breast cancer prevention. Eur J Cancer Prev 26:385–395. https://
://doi.org/10.1023/A:1021919416101 doi.org/10.1097/CEJ.0000000000000406
34. Bradshaw PT, Sagiv SK, Kabat GC, Satia JA, Britton JA, Tei- 47. Fernández-Ballart JD, Piñol JL, Zazpe I, Corella D, Carrasco P,
telbaum SL et al (2009) Consumption of sweet foods and breast Toledo E et al (2010) Relative validity of a semi-quantitative food-
cancer risk: a case–control study of women on Long Island, frequency questionnaire in an elderly Mediterranean population of
New York. Cancer Causes Control 20:1509–1515. https ://doi. Spain. Br J Nutr 103:1808–1816. https://doi.org/10.1017/S0007
org/10.1007/s10552-009-9343-x 114509993837
35. Chandran U, McCann SE, Zirpoli G, Gong Z, Lin Y, Hong CC 48. U.S. Department of Health and Human Services and U.S. Depart-
et al (2014) Intake of energy-dense foods, fast foods, sugary ment of Agriculture (2015) 2015–2020 dietary guidelines for
drinks, and breast cancer risk in African American and Euro- Americans, 8th edn. https://healt h.gov/dietaryguidelines/2015/
pean American women. Nutr Cancer 66:1187–1199. https://doi. guidelines/. Accessed 6 Jun 2018
org/10.1080/01635581
13
Title: Sugar-sweetened beverages, artificially sweetened beverages, and breast
cancer risk: results from two prospective US cohorts
Journal: The Journal of Nutrition
Submitted: March 22, 2021
Status: Under Review
Title: Phenolic Acid Subclasses, Individual Compounds, and Breast Cancer Risk in a
Mediterranean Cohort: The SUN Project
Journal: Journal of the Academy of Nutrition and Dietetics
Status: Published in January 2020
RESEARCH
Original Research
Phenolic Acid Subclasses, Individual

Compounds, and Breast Cancer Risk in
a Mediterranean Cohort: The SUN Project
Andrea Romanos-Nanclares, MsC, RDN; Cristina Sánchez-Quesada, PhD, MsC; Itziar Gardeazábal, MD, PhD;
Miguel Ángel Martínez-González, MD, PhD, MPH; Alfredo Gea, PharmD, PhD; Estefanía Toledo, MD, PhD, MPH
ARTICLE INFORMATION ABSTRACT

Article history: Background Biological and epidemiological evidence supports an inverse association
Submitted 6 June 2019 of phenolic acids with obesity-related chronic diseases. However, no previous study has
Accepted 7 November 2019 prospectively evaluated the relationship between subclasses and individual compounds
Available online 22 January 2020
of phenolic acids and the risk of postmenopausal breast cancer, one of the most
Keywords: important and prevalent obesity-related cancer sites.
Phenolic acids Objective This study examined associations between subclasses of phenolic acids,
Breast cancer
including hydroxycinnamic and hydroxybenzoic acids intake, and risk of breast cancer.
Chemoprevention
SUN Project Design The Seguimiento Universidad de Navarra (SUN) Project is a dynamic, perma-
Polyphenols nently open prospective cohort which started in 1999.
Participants/setting Participants were 10,812 middle-aged women. All of them were
Supplementary materials:
Figure 2 and Tables 4 and Table 6 are available at
university graduates.
www.jandonline.org Main outcome measures Usual diet was assessed at baseline and after 10 years of
follow-up with a 136-item food frequency questionnaire. Phenolic acid intake was
calculated by matching food consumption with the Phenol-Explorer database on
2212-2672/Copyright ª 2020 by the Academy of
Nutrition and Dietetics. phenolic acids content of each reported food item.
https://doi.org/10.1016/j.jand.2019.11.007 Statistical analysis performed Participants were classified according to tertiles of
subclasses or individual compounds of phenolic acids. Cox regression models were
fitted to estimate multivariable-adjusted hazard ratios and 95% CIs for breast cancer
incidence.
Results Over an average of 11.8 years of follow-up, 101 incident cases of breast cancer
were confirmed. After multivariable adjustment, an inverse association between
hydroxycinnamic acids intake and breast cancer was observed (hazard ratio third tertile
vs first tertile 0.37, 95% CI 0.16 to 0.85; P for trend¼0.029) among postmenopausal
women. Specifically, chlorogenic acids (3-, 4-, and 5- caffeoylquinic acids) showed the
strongest inverse association (hazard ratio third tertile vs first tertile 0.33, 95% CI 0.14 to
0.78; P for trend¼0.012).
Conclusions A higher intake of hydroxycinnamic acids, especially from chlorogenic
acids—present in coffee, fruits, and vegetables—was associated with a lower incidence of
breast cancer among postmenopausal women. Future observational studies are needed
to corroborate these results.
J Acad Nutr Diet. 2020;120(6):1002-1015.
W
ORLDWIDE, BREAST CANCER IS THE MOST
The Continuing Professional Education (CPE) quiz for this article is available frequently diagnosed cancer among women in
for free to Academy members through the MyCDRGo app (available for iOS
and Android devices) and through www.jandonline.org (click on “CPE” in the
the vast majority of countries. Breast cancer
menu and then “Academy Journal CPE Articles”). Log in with your Academy represents a prominent cause of cancer death
of Nutrition and Dietetics or Commission on Dietetic Registration username among women in Europe and North America. Approxi-
and password, click “Journal Article Quiz” on the next page, then click the
“Additional Journal CPE quizzes” button to view a list of available quizzes.
mately 2.1 million women diagnosed with breast cancer
Non-members may take CPE quizzes by sending a request to journal@ was predicted in 2018 worldwide, which accounts for
eatright.org. There is a fee of $45 per quiz (includes quiz and copy of article) almost 1 in 4 cancer cases among women.1 Preventive
for non-member Journal CPE. CPE quizzes are valid for 1 year after the issue strategies may help to address the current burden of breast
date in which the articles are published.
cancer, as well as the predicted increasing incidence of this
1002 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS ª 2020 by the Academy of Nutrition and Dietetics.
RESEARCH
disease.2 According to the World Cancer Research Fund and

the American Institute for Cancer Research, nearly one-third RESEARCH SNAPSHOT
of the most common neoplasms could be prevented by
Research Question: Are subclasses and individual
changing lifestyle and dietary habits in developed
compounds of phenolic acids associated with breast cancer
countries.3
Considerable attention has been focused on chemopre-
risk?
vention as an alternative approach to the control of cancer.4 Key Findings: In this Mediterranean cohort study that
In fact, some biologically active plant compounds may have included 10,812 women from the Seguimiento Universidad
beneficial effects for reducing the risk of chronic diseases, de Navarra Project, a higher intake of hydroxycinnamic acids,
including obesity-related cancers.5 Polyphenols are a large especially from chlorogenic acids—present in coffee, fruits,
and complex family of phytochemicals widespread in plant- and vegetables—was associated with a lower incidence of
based foods, such as fruit, vegetables, tea, coffee, wine, olive
breast cancer among postmenopausal women (both
oil, whole-grain cereals, and cocoa.6 According to their
P<0.05). To date, this is the first cohort which has studied the
chemical structure, polyphenols are often grouped into the
relationship between subclasses and individual compounds
following four main classes: flavonoids, phenolic acids,
lignans, and stilbenes.7 Among these, for example, phenolic
of phenolic acids and breast cancer risk.
acids are further divided into subclasses—hydroxybenzoic
and hydroxycinnamic acids—composed of individual com- present study, those women recruited before October 2015
pounds—3, 4-, and 5-caffeoylquinic acids (CQA)—also known (n¼13,770) were included in order to ensure a minimum
as chlorogenic acids. follow-up time of 2 years. Women were excluded if they were
Various reviews and meta-analyses of observational lost to follow-up (n¼1,295), had a history of breast cancer at
studies have summarized and analyzed the available evi- baseline (n¼104), reported energy intake outside of the
dence on the relationship between polyphenol-rich foods predefined limits (<500 kcal/day and >3,500 kcal/day in
and cancer risk.8-10 While the majority of studies regarding women) (n¼1,353),19 or reported menopause before 35 years
classes of polyphenols have focused on flavonoid intake (one of age (n¼206). Therefore, our final sample included 10,812
of the major contributors to total polyphenol intake),9 very women (Figure 1).
few epidemiological studies have evaluated the associations The study was conducted according to the Declaration of
of dietary phenolic acids and their compounds on chronic Helsinki and the protocol was approved by the Institutional
disease. Indeed, there is only one caseecontrol study that Review Board of the University of Navarra. Voluntary
found an inverse association between phenolic acids and a completion of the baseline self-administrated questionnaire
hormone-related cancer, specifically, prostate cancer.11 was considered to imply informed consent.
Recently, Gardeazabal and colleagues12 reported an inverse
association between intake of phenolic acids and breast
cancer risk among postmenopausal women in the Segui- Phenolic Acid Intake and Dietary Assessment
miento Universidad de Navarra (SUN) Project. Dietary data were collected at baseline and after 10 years of
Phenolic acids contribute to a large proportion of the total follow-up using a 136-item semi-quantitative food-frequency
polyphenol intake consumption. Indeed, on average, phenolic questionnaire (FFQ), previously validated in Spain and reas-
acids may represent about 58% of polyphenol intake across sessed thereafter.20-22 This questionnaire assessed food habits
Europe.13 Current interest in the relationship between these in the previous year. The FFQ inquired about consumption
phenolic acids and breast cancer risk stems from their prom- frequencies in nine close-ended categories (never or seldom; 1
ising protective role against cancer cell growth, oxidative to 3 times per month; once weekly; 2 to 4 times per week; 5 to
stress, inflammation, and diabetes in experimental studies.14-17 6 times per week; once daily; 2 to 3 times per day; 4 to 6 times
Based on the previous results found in the SUN project,12 it per day; and 6 or more times per day) according to the typical
is reasonable to hypothesize a protective association of sub- portion size of each item, which was also specified in the FFQ.
classes (hydroxycinnamic and hydroxybenzoic acids) and Daily food consumption was calculated by multiplying the
intake of individual phenolic acids (chlorogenic acids) with portion size by the consumption frequency for each food item.
the subsequent risk of developing breast cancer. The Phenol-Explorer database (www.phenol-explorer.
eu)6,23 was used to collect data on polyphenolic content in
foods and expressed as in mg/100 g for solid foods and oils or
MATERIALS AND METHODS mg/100 mL for beverages. The agreement between food items
Study Design and Population in the FFQ and the Phenol-Explorer database was evaluated
The SUN Project is a large, ongoing, multipurpose, prospec- with the following steps: (1) foods with merely traces or
tive cohort study of university graduates conducted in without polyphenols were deleted; (2) recipes and processed
Spain.18 Participants are recruited through collaborations foods, which comprised several food components, were
with alumni and professional associations throughout the divided according to their ingredients; (3) retention factors
country. The design is dynamic because the recruitment of from Phenol-Explorer were used to measure the impact of food
participants is permanently open and participants are processing and cooking on polyphenol content values24,25; and
followed-up biennially using questionnaires. It was designed (4) some products that may contain refined wheat flour were
in December 1999 to investigate the role of diet and environ- introduced in the list of foods (ie, pizza, pastries) and phenolic
mental factors in the etiology of different chronic diseases. acid content was obtained from their wheat flour contents.
In the SUN project, 13,999 women had been recruited and Details of the calculation of total and subtypes of polyphenols
completed the baseline questionnaire by July 2018. In the intake have been described previously by Gardeazabal and
June 2020 Volume 120 Number 6 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 1003
RESEARCH
Total participants recruited

n=22,791
Total male participants

n=8,792
Total female participants

n=13,999
Women excluded for the following reasons

(n=3,187):
• Recruited after October 2015 (n=229)
• No follow-up (n=1,295)
• Previous breast cancer (n=104)
• Energy intake outside predefined limits
(n=1,353)
• Menopause before age 35 years (n=206)
Participants included in the current analysis
n=10,812
Figure 1. Flowchart of participants in the Seguimiento Universidad de Navarra Project, 1999 to 2018.
colleagues.12 Based on this estimation, daily food consumption subject’s next of kin, work associates, and postal authorities.
for the items included in the FFQ were calculated and were For women lost to follow-up or with an unidentified cause of
matched with Phenol-Explorer database on individual com- death, the National Death Index was consulted to identify
pounds of phenolic acid content of each reported food (April deceased cohort members and confirm the cause of death.
2018). Total phenolic acid intake was calculated as the sum of When the cause of death was breast cancer, we considered
all subclasses of phenolic acid intakes (hydroxybenzoic acids, this as a confirmed case.
hydroxycinnamic acids, hydroxyphenylacetic acids, and All confirmed cases jointly with self-reported breast cancer
hydroxyphenylpropanoic acids) from all food sources as stated diagnoses (without available confirmation) were considered
in this process. We also checked 3-, 4-, and 5-CQA, often for the analyses as probable incident cases.
grouped as chlorogenic acids, as 98% of the between-person
variability in total hydroxycinnamic acids intake was Assessment of Covariates
explained by these specific acids. Finally, intake of phenolic Total energy intake was derived using the Spanish food
acids, their subclasses, and individual compounds were composition tables.27,28 Adherence to the Mediterranean diet
adjusted for total energy intake using the residual method26 was assessed with the score proposed by Trichopoulou and
and participants were then grouped into tertiles. colleagues.29 The baseline questionnaire also included infor-
mation about participants’ sociodemographic characteristics,
Ascertainment of Breast Cancer Cases medical history, health-related habits, and lifestyles. In addi-
For the present analysis, incident cases of breast cancer were tion, physical activity was evaluated with a previously vali-
defined as the primary end point. In both the baseline and dated questionnaire30 using metabolic equivalents (h/wk
follow-up questionnaires, participants were asked whether scores) for each participant. Self-reported information
they had received a medical diagnosis of breast cancer. regarding anthropometric measurements, such as weight and
Women with a self-reported medical diagnosis of breast height, has also been validated in a subsample of this cohort.31
cancer in the baseline questionnaire were considered as
prevalent cases. If participants self-reported an incident Statistical Analysis
breast cancer during the follow-up and were free of the Baseline characteristics of the participants were described
disease at the beginning of the study, they were asked for a across the three categories of phenolic acids intake. We used
medical record that included the diagnosis and the date. This percentages for categorical variables and means"standard
information was confirmed by a trained oncologist. If the date deviations for quantitative variables. P values were estimated
of the diagnosis was not included in the medical record, we with analysis of variance for continuous variables and c2 test
used the self-reported date given in the questionnaire. When for categorical variables.
the medical record was not yet available to confirm the Person-years of follow-up were calculated for each partic-
diagnosis, breast cancer cases were considered as probable ipant from the date of completion of the baseline question-
incident cases. Deaths were reported to the research team by naire to the date of breast cancer diagnosis, or the date of
1004 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS June 2020 Volume 120 Number 6
RESEARCH
death due to breast cancer cases, or the date of occurrence of was determined, as well as their main subclasses—
the last follow-up questionnaire for noncases. hydroxycinnamic and hydroxybenzoic acids—and individual
To assess the relationship of baseline dietary intake of compounds (chlorogenic acid intakes).19 A series of nested
phenolic acid subclasses or individual compounds of phenolic linear regression models after stepwise-selection algorithm
acids with the subsequent risk of breast cancer, Cox regres- was constructed. The additional contribution of a given food
sion models were used. We estimated hazards ratios (HRs) group (eg, fruits, nuts, vegetables, and other foods) was
and 95% CIs of incident breast cancer during the follow-up for shown in the change of cumulative R2.
the subclasses and individual compounds of phenolic acids Additional analyses after stratifying by smoking status and
tertiles. The lowest tertile was used as the reference category. body mass index were performed in order to assess potential
Categories of age and period of recruitment were used as effect modification by these variables, and we formally tested
stratification variables. the interactions introducing product-terms in the multivari-
To reduce the potential effect of a variation in diet during able Cox regression models.
follow-up, Cox proportional hazard models were fitted with We also performed sensitivity analyses to explore the
repeated dietary measurements using the updated data on robustness of our findings by excluding breast cancer cases
food consumption after 10 years of follow-up. diagnosed in the first 2 years of follow-up or by adding other
Cox regression models were adjusted for several potential potential confounders to the multivariable-adjusted model,
confounders defined a priori. Potential confounders were such as fiber intake.
identified based on available literature rather than deferring Analyses were performed using Stata, version 12.0.36 All P
to statistical criteria.32,33 Potential confounders included values were two-tailed and a P value <0.05 was deemed as
height (continuous), family history of breast cancer (yes/no), significant.
smoking status (never smoker, former smoker, current
smoker), physical activity (metabolic equivalents h/wk, RESULTS
continuous), alcohol intake (g/day, continuous), body mass Reported median phenolic acid intake was 260 mg/day
index (tertiles), age at menarche (categories), menopause (no (interquartile range¼170 to 376 mg/day) and accounted for
menopause, <50 years, #50 years), number of pregnancies of 40% of total polyphenol intake in this Mediterranean cohort.
more than 6 months (continuous), pregnancy before the age Specifically, hydroxycinnamic acids were the main source
of 30 years (yes/no), months of breastfeeding (continuous), (37%) of total polyphenol intake in this sample of 10,812
use of hormone replacement therapy (yes/no) and its dura- women from the SUN cohort, as presented in Figure 2B
tion (continuous), and years at university (continuous). (available at www.jandonline.org).
Additional adjustments for diabetes (yes/no), total energy The main baseline characteristics of participants according
intake (kcal/day, continuous), and adherence to the Medi- to tertiles of phenolic acids intake in the SUN cohort are
terranean diet (continuous) were also included. shown in Table 1. The mean age of participants was 34.7 years
Tests of linear trend were conducted, assigning to each (standard deviation¼10.6 years) and the mean body mass
tertile its median and using the resulting variable as contin- index (calculated as kg/m2) was 22.2 (standard
uous in the aforementioned models. deviation¼3.1). Participants in the highest tertile of phenolic
Proportional hazard assumption was checked with the acid intake were older, were more likely to be physically
GrambscheTherneau test of the scaled Schoenfeld residuals34 active, had lower energy intake, had higher alcohol intake,
and introducing hydroxycinnamic and chlorogenic acids were more likely to be former smokers, and had higher
intake as a time-varying covariate in the model. adherence to the Mediterranean diet.
Subsequently, stratified analyses by menopausal status Overall, phenolic acids along with flavonoids were the main
were performed. Information on age at menopause was contributors to total polyphenol intake in our cohort, ac-
gathered in the baseline questionnaire and after 16 years of counting for 40% and 53%, respectively (Figure 2A; available at
follow-up. For unavailable information on age at menopause, www.jandonline.org). The individual and cumulative contri-
we imputed postmenopausal status according to the 75th butions of different foods to the variability in total phenolic,
percentile of age at menopause in the study population (52 hydroxycinnamic, hydroxybenzoic, and chlorogenic acid intake
years of age).35 When evaluating premenopausal women in according to food groups are presented in Table 2. Food groups
relation to breast cancer as the dependent variable, women that contributed most to total phenolic acid intake were coffee
who reported having had menopause before study inception (48%), fruits (25%), and vegetables (23%). Moreover, 99% of
were excluded. For the rest of the participants, we considered between-person variability in total phenolic acid intake was
time at risk from baseline to age at menopause or date of explained by coffee, fruit, and vegetable consumption only.
turning 52 years, whichever occurred first. Contrarily, for
breast cancer among postmenopausal women, time at risk Subclasses of Phenolic Acid Intake and Risk of Breast
started at study inception for women who were post- Cancer
menopausal and were >52 years at baseline and, if not, at Regarding subclasses of phenolic acids, reported median
self-reported age at menopause or when they turned 52 hydroxycinnamic and hydroxybenzoic intake was 233 mg/
years during follow-up, whichever happened last. In the day (interquartile range¼147 to 352 mg/day) and 15 mg/day
analysis with postmenopausal breast cancer as the depen- (interquartile range¼9 to 24 mg/day), respectively.
dent variable, further adjustment for time since recruitment The main food sources of hydroxycinnamic were coffee
until the beginning of the time at risk and age at menopause (51%), vegetables (26%), and fruits (18%), which explained 99%
(<50 years, #50 years) was included. of the total between-person variability in hydroxycinnamic
Furthermore, the contribution of each food group to the intake. The main food sources for hydroxybenzoic acid intake
between-person variability in total phenolic acid intake were fruits (51%), nuts (29%), and alcohol (9%), which
RESEARCH
Table 1. Baseline characteristics of 10,812 women according to tertiles of phenolic acid intake in the SUNa cohort (University of
Navarra follow-up), Navarra, Spain, 1999 to 2018
Phenolic Acid Intake

Variable Tertile 1 Tertile 2 Tertile 3 P valueb
!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!
Total 3,604 3,604 3,604 —
!!!!!!!!!!!!median (mg/d)!!!!!!!!!!!!!
Phenolic acid intake 135 259 416 <0.001
c
!!!!!!!!!!!!!! mean"SD !!!!!!!!!!!!!!!
Age (y) 31.7"9.9 35.4"10.7 37.0"10.5 <0.001
Body mass index 21.9"3 22.2"3 22.5"3.1 <0.001
Physical activity (METsd-h/wk) 19.9"19 20.2"18.8 20.5"18.6 0.44
Total energy intake (kcal/day) 2,335"575 2,272"573 2,283"568 <0.001
Alcohol intake (g/day) 3.2"5 4.2"6 4.6"6.5 <0.001
Years at university 4.7"1.3 4.9"1.3 4.9"1.4 <0.001
Height (cm) 164"6 163"6 163"6 <0.001
No. of pregnancies of more than 6 mo 0.2"0.6 0.2"0.7 0.3"0.7 <0.001
!!!!!!!!!!!!!!!!n (%)!!!!!!!!!!!!!!!!!
Pregnancy before 30 y 202 (5.6) 266 (7.4) 300 (8.3) <0.001
!!!!!!!!!!!!!!mean"SDc !!!!!!!!!!!!!!!
Breastfeeding (mo) 1.9"4.6 2.5"5.1 2.7"5.3 <0.001
!!!!!!!!!!!!!!!!n (%)!!!!!!!!!!!!!!!!!
Hormone replacement therapye (yes) 109 (40.4) 148 (32.7) 192 (38.6) 0.07
!!!!!!!!!!!!!!mean"SDc !!!!!!!!!!!!!!!
Length of hormone replacement therapye (y) 1.2"2.3 1.3"2.6 1.3"2.2 0.90
!!!!!!!!!!!!!!!!n (%)!!!!!!!!!!!!!!!!!
Diabetes 23 (0.6) 47 (1.3) 48 (1.3) <0.05
Smoking <0.001
Never smoker 2,201 (61.1) 1,858 (51.6) 1,466 (40.7)
Former smoker 684 (19) 770 (26.4) 1,007 (30.6)
Current smoker 686 (19) 950 (21.4) 1,103 (27.9)
Family history of BCf 0.18
No 3,215 (89.2) 3,248 (90.1) 3,206 (89.0)
Yes 389 (10.8) 356 (9.9) 398 (11.1)
Age at menarche <0.05
<10 y 7 (2.6) 4 (0.9) 2 (0.4)
10 to 11 y 51 (18.9) 93 (20.6) 108 (21.7)
12 to 13 y 158 (58.5) 245 (54.2) 268 (53.9)
14 to 16 y 47 (17.4) 105 (23.2) 105 (21.1)
#16 y 7 (2.6) 5 (1.1) 14 (2.8)
(continued on next page)
accounted for 99% of between-person variability in hydrox- contributed most to its total between-person variability in
ybenzoic acid intake (Table 2). intake were coffee, fruits, and vegetables, and the main
In addition, with reference to individual compounds of sources of chlorogenic acid intake were coffee (52%), vege-
hydroxycinnamic acids (chlorogenic acids), those that tables (22%), and fruits (25%).
RESEARCH
Table 1. Baseline characteristics of 10,812 women according to tertiles of phenolic acid intake in the SUNa cohort (University of
Navarra follow-up), Navarra, Spain, 1999 to 2018 (continued)
Phenolic Acid Intake

Variable Tertile 1 Tertile 2 Tertile 3 P valueb
Menopausal status <0.001

Premenopausal 2,892 (80.2) 2,496 (69.3) 2,188 (60.7)
Postmenopausal 712 (19.8) 1,108 (30.7) 1,416 (39.3)
Age at menopausee 0.70
Postmenopausal <50 y 174 (64.4) 293 (64.8) 333 (67)
Postmenopausal #50 y 96 (35.6) 159 (35.2) 164 (33)
c
!!!!!!!!!!!!!mean"SD !!!!!!!!!!!!!!
Mediterranean diet adherence 3.6"1.7 4.3"1.8 4.5"1.8 <0.001
a
SUN¼Seguimiento Universidad de Navarra.
b
Results from one-way analysis of variance for continuous variables and c2 test for categorical variables.
c
SD¼standard deviation.
d
METs¼metabolic equivalents.
e
For postmenopausal women only.
f
Information from mother, sisters, and both grandmothers was collected.
During a median follow-up time of 11.5 years and 115,368 variability (98%) and the main sources of hydroxycinnamic
person-years at risk, 190 probable incident cases of breast acids were, in decreasing order, 4-CQA, 5-CQA, and 3-CQA,
cancer were identified, of which 101 were confirmed by our which are classified as chlorogenic acids. Chlorogenic acids
team of oncologists after reviewing the medical records. were the main individual compounds of hydroxycinnamic
With regard to subclasses of phenolic acid intake, no sig- acid, with a median intake of 162 mg/day (interquartile
nificant association between hydroxycinnamic intake and range¼93 to 226 mg/day). The main food sources of chloro-
overall breast cancer risk was found for confirmed (Table 3) or genic acids were coffee, fruits, and vegetables (Table 2).
probable cases (Table 4; available at www.jandonline.org). A significant inverse association between chlorogenic acid
When we stratified our analyses by menopausal status, an intake and breast cancer was observed only among post-
inverse association between hydroxycinnamic acids intake menopausal women when both confirmed cases (HR T3 vs T1
and postmenopausal breast cancer risk was found for both 0.33, 95% CI 0.14 to 0.78; P for trend¼0.012) (Table 5) and
confirmed (hazard ratio [HR] third tertile [T3] vs first tertile probable cases (HR T3 vs T1 0.42, 95% CI 0.23 to 0.79; P for
[T1] 0.37, 95% CI 0.16 to 0.85; P for trend¼0.029) and probable trend¼0.008)] (Table 6; available at www.jandonline.org)
(HR T3 vs T1 0.43, 95% CI 0.23 to 0.80; P for trend¼0.011) cases. were evaluated. In addition, as a sensitivity analysis, we
Cox proportional hazard models were also fitted with adjusted for fiber intake, although results did not change
repeated measurements using the updated data on food (data not shown).
consumption after 10 years of follow-up (Table 3). Results In addition, the analysis with repeated measures after 10
from these analyses remained basically unchanged, and the years of follow-up showed similar results (HR T3 vs T10.39, 95%
HR for the T3 vs T1 of hydroxycinnamic acids intake was 0.40 CI 0.17 to 0.91; P for trend¼0.036), results that also remained
(95% CI 0.17 to 0.92) among postmenopausal women. These statistically significant after excluding breast cancer cases
compounds were also inversely associated with breast cancer diagnosed in the first 2 years of follow-up (data not shown).
risk when we also excluded women diagnosed with breast Using the likelihood ratio test, the P for interaction was
cancer during the first 2 years of follow-up or adjusted the calculated between consumption of subclasses and individual
main analyses for other potential confounders, such as fiber compounds of phenolic acids and stratification variables for
intake (data not shown). each scenario (body mass index and smoking status). None of
No significant association was observed among premeno- them were significant.
pausal women. Furthermore, no significant association be-
tween intake of hydroxybenzoic acids and overall breast DISCUSSION
cancer risk was found (HR T3 vs T1 0.82, 95% CI 0.48 to 1.38; P The results from this prospective study of Mediterranean
for trend¼0.41; premenopausal HR T3 vs T1 0.76, 95% CI 0.39 university graduates support the hypothesis that high con-
to 1.47; P for trend¼0.64; or postmenopausal women HR T3 sumption of hydroxycinnamic acids (approximately >435
vs T1 0.66, 95% CI 0.24 to 1.80; P for trend¼0.19). mg/day) and chlorogenic acids (>290 mg/day) was associated
with a lower risk of breast cancer among postmenopausal
Individual Compounds of Phenolic Acid and Risk of women.
Breast Cancer
In line with the aforementioned results, we also analyzed Phenolic Acids Intake and Breast Cancer Risk
individual compounds that constitute the hydroxycinnamic Phenolic acids, along with flavonoids, were the main con-
subclass. The main contributors to the between-person tributors to total polyphenol intake in our cohort and coffee
RESEARCH
Table 2. Main sources of variability (cumulative R2a and change in R2) and main sources of quantity (%) of total phenolic,
hydroxycinnamic, hydroxybenzoic, and chlorogenic acid intake according to food groups: The SUNb cohort (University of
Navarra follow-up), Navarra, Spain, 1999 to 2018
Food group Cumulative R2 Change in R2 Sources of quantity (%)
Total phenolic acid intake

Coffeec 0.78 — 48
d
Fruits 0.94 0.17 25
Vegetablese 0.99 0.05 23
Hydroxycinnamic acid intake
Coffee 0.83 — 51
Fruits 0.99 0.08 18
Hydroxybenzoic acid intake
Fruits 0.75 — 51
f
Nuts 0.99 0.24 29
Alcoholg 0.99 0.01 9
Chlorogenic acid intake
Coffee 0.79 — 52
Fruits 0.91 0.13 22
a
Cumulative R2 values were determined from nested regression analyses after stepwise selection.
b
c
Regular and decaffeinated coffee.
d
Cherry, strawberry, apple, peach, grape, kiwi, mango, orange, banana, avocado, olive, canned fruit, and dried fruit.
e
Swiss chard, spinach, artichoke, thistle, leek, tomato, carrot, pumpkin, lettuce, endive, eggplant, zucchini, beans, pepper, cabbage, broccoli, cauliflower, gazpacho, asparagus, and potato
(roasted and fried).
f
Walnuts, peanuts, hazelnuts, and almonds.
g
Wine, beer, other wines, and whiskey.
was by far the principal food source (48%) and the main plant kingdom. Thus, they could have a potential role in the
source of between-person variability (change in R2¼0.78) in prevention of oxidative stress-related diseases, such as can-
total phenolic acids intake. These results are consistent with cer.40 This group of polyphenols is found both covalently
previous studies from Mediterranean countries.37,38 adhered to the plant cell wall polysaccharides and in cyto-
Although a number of in vitro and in vivo studies suggest plasm as soluble forms. In the PREDIMED (Prevención con
that polyphenol intake and phenolic acids intake may influ- Dieta Mediterránea) study,37 hydroxycinnamic acids were the
ence carcinogenesis and tumor development,39 no previous main source of polyphenol intake (33%), a figure close to the
prospective observational study has been conducted to test current study (37%). In the European Prospective Investiga-
potential associations between subclasses and individual tion into Cancer and Nutrition study, Zamora-Ros and col-
phenolic acids consumption and breast cancer risk. leagues13 observed that the hydroxycinnamic acids subclass
Phenolic acids are of great interest because foods rich in was the major contributor to total phenolic acids intake
these molecules, such as coffee, have demonstrated a among 36,037 subjects aged 35 to 74 years from 10 European
decreased risk of breast cancer in animal and cell models.40 countries. More concretely, in the latter study, hydroxycin-
Epidemiological studies have focused on estimating the di- namic acids explained between 84.6% and 95.3% of total
etary intakes and food sources of phenolic acids in different polyphenol intake, depending on the location. Similar results
cohorts,13,41,42 but only one caseecontrol study in Southern were observed by Pérez-Jiménez and colleagues43 in a French
Italy has addressed phenolic acid intake as exposure in study.
relation to a hormone-related tumor, specifically, prostate There are several mechanisms that could explain the
cancer.11 In that caseecontrol study, caffeic acid and ferulic association between hydroxycinnamic acids intake and
acid were both associated with a reduced risk of prostate breast cancer risk, particularly, among postmenopausal
cancer, and caffeic acid and hydroxybenzoic acids were linked women. Hydroxycinnamic acids have been reported to
to a lower risk of advanced prostate cancer. inhibit cancer invasion and metastasis,44 as well as to
modulate apoptosis and ERa expression.40 Inhibition of ERa
Phenolic Acids Subclasses and Breast Cancer Risk expression decreases the peripheral conversion of andro-
Phenolic acids and, particularly, hydroxycinnamic acids, are gens to estrogens, a key mechanism for postmenopausal
important sources of antioxidants due to their ubiquity in the breast cancer.45-47
June 2020 Volume 120 Number 6
Table 3. HRa (95% CI) for the association between tertiles of individual phenolic acid intake and confirmed breast cancer risk in the SUNb cohort
Tertiles of Hydroxycinnamic Acids Intake Tertiles of Hydroxybenzoic Acids Intake

P for
Variable Tertile 1 Tertile 2 Tertile 3 trend Tertile 1 Tertile 2 Tertile 3 P for trend
Overall cases
!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!
Cases 26 34 41 — 33 36 32 —
!!!!!!!!!!!!!!person-years!!!!!!!!!!!!!!! !!!!!!!!!!!!!!person-years!!!!!!!!!!!!!!!
Follow-up 37,851 38,202 39,749 — 40,053 38,615 37,134 —
!!!!!!!!!!per 10,000 person-years!!!!!!!!!!! !!!!!!!!!!per 10,000 person-years!!!!!!!!!!!
Incidence rate 6.9 8.9 10.3 — 8.2 9.3 8.6 —
!!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!! !!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!!
Age-adjusted 1 (refc) 1.02 (0.61 to 1.70) 1.01 (0.61 to 1.66) 0.98 1 (ref) 0.97 (0.60 to 1.56) 0.80 (0.49 to 1.30) 0.33
d
Multivariable-adjusted 1 (ref) 1.00 (0.60 to 1.68) 0.98 (0.60 to 1.63) 0.94 1 (ref) 0.94 (0.59 to 1.52) 0.80 (0.48 to 1.33) 0.38
Multivariable-adjustede 1 (ref) 1.01 (0.60 to 1.69) 1.00 (0.60 to 1.66) 0.99 1 (ref) 0.97 (0.58 to 1.62) 0.82 (0.48 to 1.38) 0.41
Repeated measures 1 (ref) 1.03 (0.61 to 1.73) 1.08 (0.65 to 1.80) 0.75 1 (ref) 1.13 (0.68 to 1.89) 0.87 (0.51 to 1.47) 0.45
Premenopausal women
JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS
!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!
Cases 11 21 25 — 21 16 20 —
Follow-up 33,436 31,374 31,995 — 35,360 32,426 29,019 —
Incidence rate 3.3 6.7 7.8 — 5.9 4.9 6.9 —
!!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!! !!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!!
Age-adjusted 1 (ref) 1.55 (0.74 to 3.22) 1.51 (0.74 3.09) 0.35 1 (ref) 0.69 (0.36 to 1.32) 0.84 (0.45 to 1.55) 0.73
Multivariable-adjustedf 1 (ref) 1.54 (0.74 to 3.22) 1.45 (0.70 to 3.00) 0.43 1 (ref) 0.67 (0.35 to 1.30) 0.80 (0.42 to 1.52) 0.65
Repeated measures 1 (ref) 1.92 (0.90 to 4.07) 1.55 (0.72 to 3.34) 0.39 1 (ref) 0.66 (0.34 to 1.30) 0.82 (0.43 to 1.55) 0.59
RESEARCH
1009
RESEARCH
1010
Table 3. HRa (95% CI) for the association between tertiles of individual phenolic acid intake and confirmed breast cancer risk in the SUNb cohort (continued)
Tertiles of Hydroxycinnamic Acids Intake Tertiles of Hydroxybenzoic Acids Intake

P for
Variable Tertile 1 Tertile 2 Tertile 3 trend Tertile 1 Tertile 2 Tertile 3 P for trend
!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!
Cases 14 10 10 — 9 17 8 —
Follow-up 4,635 7,313 8,359 — 5,050 6,596 8,662 —
Incidence rate 30.2 13.7 12.0 — 17.8 25.8 9.2 —
!!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!! !!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!!
Age-adjusted 1 (ref) 0.46 (0.21 to 1.04) 0.38 (0.17 to 0.86) 0.027 1 (ref) 1.49 (0.66 to 3.36) 0.55 (0.21 to 1.43) 0.09
Multivariable-adjustedg 1 (ref) 0.44 (0.19 to 1.01) 0.40 (0.17 to 0.91) 0.040 1 (ref) 1.41 (0.62 to 3.22) 0.59 (0.22 to 1.58) 0.16
e
Repeated measuresh 1 (ref) 0.51 (0.22 to 1.17) 0.40 (0.17 to 0.92) 0.046 1 (ref) 1.67 (0.69 to 4.02) 0.78 (0.29 to 2.06) 0.18
a
HR¼hazard ratio.
b
c
ref¼reference.
d
Adjusted for age (underlying time variable), height (continuous), family history of breast cancer (yes/no), smoking status (never smoker, former smoker, current smoker), physical activity (metabolic equivalents-h/wk, continuous), alcohol intake (g/day,
continuous), body mass index (tertiles), age at menarche (categories), menopause (no menopause, <50 years, #50 years), number of pregnancies of more than 6 mo (continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding
(continuous), use of hormone replacement therapy (yes/no) and its duration (continuous), and years at university (continuous).
e
Also adjusted for diabetes (yes/no), total energy intake (kcal/day, continuous), and Mediterranean diet adherence (continuous).
f
continuous), body mass index (tertiles), age at menarche (categories), number of pregnancies of more than 6 mo (continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding (continuous), and years at university (continuous).
g
continuous), body mass index (tertiles), age at menarche (categories), menopause (<50 years, #50 years), number of pregnancies of more than 6 mo (continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding (continuous), use
of hormone replacement therapy (yes/no) and its duration (continuous), years at university (continuous), and time since recruitment.
h
Multivariable-adjusted model with repeated measures (updated data at 10 y of follow-up).
RESEARCH
Table 5. HRa (95% CI) for the association between tertiles of chlorogenic acids (5-caffeoylquinic acid, 4-caffeoylquinic acid, and 3
caffeoylquinic acid) intake and confirmed breast cancer risk in the SUNb cohort
Tertiles of Total Chlorogenic Acid Intake

Variable Tertile 1 Tertile 2 Tertile 3 P for trend
Overall confirmed cases

!!!!!!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!!!!!!
Cases 25 35 41 —
!!!!!!!!!!!!!!!!!!!person-years!!!!!!!!!!!!!!!!!!!!
Follow-up 37,564 38,378 39,860 —
!!!!!!!!!!!!!!!per 10,000 person-years!!!!!!!!!!!!!!!!
Incidence rate 6.7 9.1 10.3 —
! HR (95% CI)!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!! !!
Age-adjusted 1 (refc) 1.09 (0.65 to 1.82) 1.08 (0.65 to 1.78) 0.80
d
Multivariable-adjusted 1 (ref) 1.06 (0.63 to 1.79) 1.06 (0.64 to 1.76) 0.83
Multivariable-adjustede 1 (ref) 1.06 (0.63 to 1.79) 1.08 (0.65 to 1.80) 0.79
Repeated measuresh 1 (ref) 1.16 (0.69 to 1.95) 1.10 (0.65 to 1.85) 0.78
Premenopausal women
!!!!!!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!!!!!!
Cases 11 20 26 —
!!!!!!!!!!!!!!!!!!!person-years!!!!!!!!!!!!!!!!!!!!
Follow-up 33,065 31,883 31,858 —
!!!!!!!!!!!!!!!per 10,000 person-years!!!!!!!!!!!!!!!!
! HR (95% CI)!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!! !!
Age-adjusted 1 (ref) 1.51 (0.72 to 3.17) 1.62 (0.79 to 3.30) 0.22
f
h
Repeated measures 1 (ref) 1.71 (0.79 to 3.67) 1.83 (0.86 to 3.89) 0.15
Individual Phenolic Acids Compounds and Breast which explains their ability to downregulate pro-
Cancer Risk inflammatory cytokines through modulation of key tran-
According to Johnston and colleagues,48 chlorogenic acids scription factors.52
may lower food cravings, reduce daily calorie intake, induce The findings of the current study should be considered in
body fat loss by thermogenesis, and enhance glucose toler- light of some limitations. First, unknown or unmeasured
ance and insulin sensitivity by acting as a peroxisome confounders may have affected the results in spite of the
proliferatoreactivated receptor a agonist. This mechanism adjustment for a wide range of potential confounders in the
could partially explain its potential protective effect in post- multivariate models. Also, misclassification of some of the
menopausal women,49 as this group is particularly suscepti- considered confounders may not completely rule out residual
ble to insulin resistance and accumulation of fatty acids. In confounding. Second, variability or lack of information
fact, chlorogenic acids, through their potential antioxidant regarding the specific phenolic content in foods might have
effects, may also reduce the proliferation of new fat cells.50 led to underlying limitations in the estimation of phenolic
Nakatani and colleagues51 suggested that some chlorogenic acids intake. Polyphenol content depends on ripeness at
acids isomers (3-, 4-, and 5-CQA) isolated from prunes may harvest time, environmental factors, processing and storage,
have antioxidant activity, such as scavenging activity on su- and plant variety.52 Furthermore, information about some
peroxide anion radicals, and inhibition of methyl linoleate foods rich in phenolic acids in the FFQ were lacking (eg,
oxidation. This antioxidant activity of chlorogenic acids has spices or herbs)53 or grouped (eg, nuts). Indeed, there are
been observed in in vitro and in vivo studies, and in various possible measurement inaccuracies in the assessment of the
disease models. They also have anti-inflammatory activity, total phenolic intake and their derivatives. Although the FFQ
RESEARCH
Table 5. HRa (95% CI) for the association between tertiles of chlorogenic acids (5-caffeoylquinic acid, 4-caffeoylquinic acid, and 3
caffeoylquinic acid) intake and confirmed breast cancer risk in the SUNb cohort (continued)

!!!!!!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!!!!!!
Cases 14 11 9 —
!!!!!!!!!!!!!!!!!!!person-years!!!!!!!!!!!!!!!!!!!!
Follow-up 4,692 7,010 8,605 —
!!!!!!!!!!!!!!!per 10,000 person-years!!!!!!!!!!!!!!!!
! HR (95% CI)!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!! !!
Multivariable-adjustedg 1 (ref) 0.50 (0.22 to 1.11) 0.35 (0.15 to 0.83) 0.017
e
Repeated measuresh 1 (ref) 0.52 (0.23 to 1.19) 0.39 (0.17 to 0.91) 0.034
a
HR¼hazard ratio.
b
c
ref¼reference.
d
Adjusted for age (underlying time variable), height (continuous), family history of breast cancer (yes/no), smoking status (never smoker, former smoker, current smoker), physical activity
(metabolic equivalents-h/wk, continuous), alcohol intake (g/day, continuous), body mass index (tertiles), age at menarche (categories), menopause (no menopause, <50 y, #50 y), number
of pregnancies of more than 6 mo (continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its
duration (continuous), and years at university (continuous).
e
f
(metabolic equivalents-h/wk, continuous), alcohol intake (g/day, continuous), body mass index (tertiles), age at menarche (categories), number of pregnancies of more than 6 mo
(continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding (continuous), and years at university (continuous).
g
(metabolic equivalents-h/wk, continuous), alcohol intake (g/day, continuous), body mass index (tertiles), age at menarche (categories), menopause (<50 y, #50 y), number of pregnancies
of more than 6 mo (continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its duration
(continuous), years at university (continuous), and time since recruitment.
h
Multivariable-adjusted model with repeated measures (updated data at 10 y of follow-up).
has not been specifically designed to collect data regarding composed of young women, which compromised the ex-
polyphenols, some validation studies have shown that FFQs pected number of incident breast cancer cases. Lastly, no
are accurate enough to estimate polyphenol intake.54 Third, information regarding hormone-dependent subtypes of
misclassification of dietary information obtained with FFQs breast cancer was obtained.
may not be completely ruled out. However, the FFQ was A major strength of the present study is its novelty, as no
validated repeatedly.21 Fourth, baseline history of oral con- previous study has evaluated the association between phenolic
traceptive use was not collected and thus additional adjust- acids subclasses and individual compounds and breast cancer
ment for this information was not possible. Fifth, the sample risk. Other strengths are its prospective design, long follow-up
was not representative of the general population. Nonethe- (more than 16 years), and high retention proportion (91%).
less, it is implausible that underlying mechanisms for the The analyses included 10,812 highly educated women and
association between these phenolic acids and breast cancer comprehensive data on risk factors and confounders for breast
risk may be different for highly educated women than for cancer risk. In order to avoid false positives, the main results
women with a lower educational level. Recruitment of only were based on confirmed breast cancer cases, which increases
highly educated participants follows the approach known as the specificity. In addition, the Phenol-Explorer database was
restriction in epidemiology. The purpose was to control for used because it is the most comprehensive food composition
confounding by socioeconomic status and educational level. platform on polyphenols available.23
By doing so, restriction to university graduates could also
enhance the validity of our study, as the high educational
level may increase the accuracy of self-reports.55 Sixth, the CONCLUSIONS
number of confirmed cases of breast cancer in the SUN cohort The findings of this study suggest that the phenolic acid con-
was low, which also limits the statistical power on associa- tent of the diet, particularly, dietary hydroxycinnamic and
tions of total, subclasses, and individual compounds of chlorogenic acids—present in coffee, fruits, and vegetables—
phenolic acids with breast cancer. The population is largely were associated with lower breast cancer risk among
RESEARCH
postmenopausal women in this Mediterranean cohort. More 21. Martin-Moreno JM, Boyle P, Gorgojo L, et al. Development and
validation of a food frequency questionnaire in Spain. Int J Epidemiol.
in-depth observational studies according to hormone receptor
1993;22(3):512-519.
status and biomarkers of phenolic acid intake are also needed
22. Fernández-Ballart JD, Piñol JL, Zazpe I, et al. Relative validity of a
to corroborate these results and thus obtain a better under- semi-quantitative food-frequency questionnaire in an elderly
standing of which compounds may affect breast cancer risk. Mediterranean population of Spain. Br J Nutr. 2010;103(12):1808-
1816.
References 23. Neveu V, Perez-Jimenez J, Vos F, et al. Phenol-Explorer: An online

1. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global comprehensive database on polyphenol contents in foods. Database.
cancer statistics 2018: GLOBOCAN estimates of incidence and mor- 2010;2010(0):bap024-bap024.
tality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 24. Bognár DA. Tables on weight yield of food and retention factors of
2018;68(6):394-424. food constituents for the calculation of nutrient composition of
2. Sauter ER. Breast cancer prevention: Current approaches and future cooked foods (dishes). http://www.fao.org/uploads/media/bognar_
directions. Eur J Breast Health. 2018;14(2):64-71. bfe-r-02-03.pdf. Published 2002. Accessed December 3, 2019.
3. World Cancer Research Fund, American Institute for Cancer 25. Rothwell JA, Medina-Remón A, Pérez-Jiménez J, et al. Effects of food
Research. Diet, nutrition, physical activity and breast cancer. wcrf. processing on polyphenol contents: A systematic analysis using
org/breast-cancer-2017. Accessed December 17, 2019. Phenol-Explorer data. Mol Nutr Food Res. 2015;59(1):160-170.
4. Kang NJ, Shin SH, Lee HJ, Lee KW. Polyphenols as small molecular 26. Willet W, Howe G, Kushi L. Adjustment for total energy intake in
inhibitors of signaling cascades in carcinogenesis. Pharmacol Ther. epidemiologic studies. Am J Clin Nutr. 1997;65(4 suppl):1220S-
2011;130(3):310-324. 1228S.
5. Del Rio D, Rodriguez-Mateos A, Spencer JPE, Tognolini M, Borges G, 27. Moreiras O. Tablas de Composición de Alimentos: Guía de Prácticas.
Crozier A. Dietary (poly)phenolics in human health: Structures, Madrid: Pirámide; 2015.
bioavailability, and evidence of protective effects against chronic 28. Mataix Verdú FJ; Universidad de Granada, Instituto de Nutrición y
diseases. Antioxid Redox Signal. 2013;18(14):1818-1892. Tecnología de los Alimentos. Tabla de Composición de Alimentos.
6. Pérez-Jiménez J, Neveu V, Vos F, Scalbert A. Systematic analysis of Granada, Spain: Universidad de Granada, Instituto de Nutrición y
the content of 502 polyphenols in 452 foods and beverages: An Tecnología de los Alimentos; 2009.
application of the phenol-explorer database. J Agric Food Chem. 29. Trichopoulou A, Costacou T, Bamia C, Trichopoulos D. Adherence to a
2010;58(8):4959-4969. Mediterranean diet and survival in a Greek population. N Engl J Med.
7. Zamora-Ros R, Touillaud M, Rothwell JA, Romieu I, Scalbert A. 2003;348(26):2599-2608.
Measuring exposure to the polyphenol metabolome in observational 30. Martinez-Gonzalez MA, Lopez-Fontana C, Varo JJ, Sanchez-Villegas A,
epidemiologic studies: Current tools and applications and their Martinez JA. Validation of the Spanish version of the physical activity
limits. Am J Clin Nutr. 2014;100(1):11-26. questionnaire used in the Nurses’ Health Study and the Health
8. Rothwell JA, Knaze V, Zamora-Ros R. Polyphenols: Dietary assess- Professionals’ Follow-up Study. Public Health Nutr. 2005;8(7):920-
ment and role in the prevention of cancers. Curr Opin Clin Nutr Metab 927.
Care. 2017;20(6):512-521. 31. Bes-Rastrollo M, Pérez Valdivieso JR, Sánchez-Villegas A, Alonso Á,
9. Grosso G, Godos J, Lamuela-Raventos R, et al. A comprehensive Martínez-González MÁ. Validación del peso e índice de masa
meta-analysis on dietary flavonoid and lignan intake and cancer risk: corporal auto-declarados de los participantes de una cohorte de
Level of evidence and limitations. Mol Nutr Food Res. 2017;61(4). graduados universitarios. Rev Esp Obes. 2005;3(6):352-358.
10. Lei L, Yang Y, He H, et al. Flavan-3-ols consumption and cancer risk: 32. Hernán MA, Hernández-Diaz S, Werler MM, Mitchell AA. Causal
A meta-analysis of epidemiologic studies. Oncotarget. 2016;7(45): knowledge as a prerequisite for confounding evaluation: An appli-
73573-73592. cation to birth defects epidemiology. Am J Epidemiol. 2002;155(2):
176-184.
11. Russo GI, Campisi D, Di Mauro M, et al. Dietary consumption of
phenolic acids and prostate cancer: A case-control study in sicily, 33. Greenland S, Daniel R, Pearce N. Outcome modelling strategies in
Southern Italy. Molecules. 2017;22(12). epidemiology: Traditional methods and basic alternatives. Int J Epi-
demiol. 2016;45(2):565-575.
12. Gardeazabal I, Romanos-Nanclares A, Martínez-González MA, et al.
Total polyphenol intake and breast cancer risk in the SUN cohort. Br J 34. Royston P, Lambert PC. Flexible Parametric Survival Analysis Using
Nutr. 2019;122(5):542-551. Stata: Beyond the Cox Model. College Station, TX: Stata Press; 2011.
13. Zamora-Ros R, Rothwell JA, Scalbert A, et al. Dietary intakes and food 35. Shivappa N, Sandin S, Lof M, Hebert JR, Adami H-O, Weiderpass E.
sources of phenolic acids in the European Prospective Investigation Prospective study of dietary inflammatory index and risk of breast
into Cancer and Nutrition (EPIC) study. Br J Nutr. 2013;110(8):1500- cancer in Swedish women. Br J Cancer. 2015;113(7):1099-1103.
1511. 36. Stata Statistical Software [computer program]. Release 12. College
14. Nagasaka R, Chotimarkorn C, Shafiqul IM, Hori M, Ozaki H, Ushio H. Station, TX: Stata Corporation LP; 2011.
Anti-inflammatory effects of hydroxycinnamic acid derivatives. 37. Tresserra-Rimbau A, Medina-Remón A, Pérez-Jiménez J, et al. Dietary
Biochem Biophys Res Commun. 2007;358(2):615-619. intake and major food sources of polyphenols in a Spanish popula-
15. De P, Baltas M, Bedos-Belval F. Cinnamic acid derivatives as anti- tion at high cardiovascular risk: The PREDIMED study. Nutr Metab
cancer agents-a review. Curr Med Chem. 2011;18(11):1672-1703. Cardiovasc Dis. 2013;23(10):953-959.
16. Kylli P, Nousiainen P, Biely P, Sipilä J, Tenkanen M, Heinonen M. 38. Zamora-Ros R, Knaze V, Rothwell JA, et al. Dietary polyphenol intake
Antioxidant potential of hydroxycinnamic acid glycoside esters. in Europe: The European Prospective Investigation into Cancer and
J Agric Food Chem. 2008;56(12):4797-4805. Nutrition (EPIC) study. Eur J Nutr. 2016;55(4):1359-1375.
17. Thom E. The effect of chlorogenic acid enriched coffee on glucose 39. Vauzour D, Rodriguez-Mateos A, Corona G, Oruna-Concha MJ,
absorption in healthy volunteers and its effect on body mass when Spencer JPE. Polyphenols and human health: Prevention of disease
used long-term in overweight and obese people. J Int Med Res. and mechanisms of action. Nutrients. 2010;2(11):1106-1131.
2007;35(6):900-908.
40. Rocha LD, Monteiro MC, Teodoro AJ. Anticancer properties of
18. Carlos S, De La Fuente-Arrillaga C, Bes-Rastrollo M, et al. Mediter- hydroxycinnamic acids—A review. Cancer Clin Oncol. 2012;1(2):
ranean diet and health outcomes in the SUN Cohort. Nutrients. p109.
2018;10(4):439.
41. Radtke J, Linseisen J, Wolfram G. Phenolic acid intake of adults in a
19. Willett W. Nutritional Epidemiology. 3rd ed. New York, NY: Oxford Bavarian subgroup of the national food consumption survey.
University Press; 2012. Z Ernahrungswiss. 1998;37(2):190-197.
20. de la Fuente-Arrillaga C, Vázquez Ruiz Z, Bes-Rastrollo M, Sampson L, 42. Ovaskainen M-L, Törrönen R, Koponen JM, et al. Dietary intake and
Martinez-González MA. Reproducibility of an FFQ validated in Spain. major food sources of polyphenols in Finnish adults. J Nutr.
Public Health Nutr. 2010;13(9):1364-1372. 2008;138:562-566.
RESEARCH
43. Pérez-Jiménez J, Fezeu L, Touvier M, et al. Dietary intake of 337 50. Garg SK. Green coffee bean. In: Gupta RC, ed. Nutraceuticals:
polyphenols in French adults. Am J Clin Nutr. 2011;93(6):1220-1228. Efficacy, Safety and Toxicity. London: Academic Press; 2016:
44. Weng CJ, Yen GC. Chemopreventive effects of dietary phytochemi- 653-667.
cals against cancer invasion and metastasis: Phenolic acids, mono- 51. Nakatani N, Kayano SI, Kikuzaki H, Sumino K, Katagiri K, Mitani T.
phenol, polyphenol, and their derivatives. Cancer Treat Rev. Identification, quantitative determination, and antioxidative activ-
2012;38(1):76-87. ities of chlorogenic acid isomers in prune (Prunus domestica L.).
45. Gorzynik-Debicka M, Przychodzen P, Cappello F, et al. Potential J Agric Food Chem. 2000;48(11):5512-5516.
health benefits of olive oil and plant polyphenols. Int J Mol Sci. 52. Manach C, Scalbert A, Morand C, Rémésy C, Jiménez L. Poly-
2018;19(3):686. phenols: Food sources and bioavailability. Am J Clin Nutr.
46. Losada-Echeberría M, Herranz-López M, Micol V, Barrajón-Catalán E. 2004;79(5):727-747.
Polyphenols as promising drugs against main breast cancer signa- 53. Pérez-Jiménez J, Neveu V, Vos F, Scalbert A. Identification of the
tures. Antioxidants. 2017;6(4):88. 100 richest dietary sources of polyphenols: An application of the
47. Mocanu M-M, Nagy P, Szöllo }si J. Chemoprevention of breast cancer Phenol-Explorer database. Eur J Clin Nutr. 2010;64(suppl 3):S112-
by dietary polyphenols. Molecules. 2015;20(12):22578-22620. S120.
48. Johnston KL, Clifford MN, Morgan LM. Coffee acutely modifies 54. Burkholder-Cooley NM, Rajaram SS, Haddad EH, Oda K, Fraser GE,
gastrointestinal hormone secretion and glucose tolerance in Jaceldo-Siegl K. Validating polyphenol intake estimates from a food-
humans: Glycemic effects of chlorogenic acid and caffeine. Am J Clin frequency questionnaire by using repeated 24-h dietary recalls and a
Nutr. 2003;78(4):728-733. unique method-of-triads approach with 2 biomarkers. Am J Clin Nutr.
2017;105:685-694.
49. Gunter MJ, Hoover DR, Yu H, et al. Insulin, insulin-like growth factor-
I, and risk of breast cancer in postmenopausal women. J Natl Cancer 55. Rothman KJ. Epidemiology: An Introduction. 2nd ed. Oxford, UK:
Inst. 2009;101(1):48-60. Oxford University Press; 2012.
RESEARCH
AUTHOR INFORMATION
A. Romanos-Nanclares is a registered dietitian and PhD student, Department of Preventive Medicine and Public Health, School of Medicine,
University of Navarra, Spain, and Navarra Institute for Health Research, Pamplona, Spain. C. Sánchez-Quesada is a PhD, Immunology Division,
Department of Health Sciences, University of Jaén, Jaén, Spain. I. Gardeazábal is a medical doctor and PhD, Department of Preventive Medicine
and Public Health, University of Navarra, Spain, and in the Department of Oncology, University of Navarra Clinic, Navarra, Spain. M. A. Martínez-
González is chair, Department of Preventive Medicine and Public Health, School of Medicine, University of Navarra, Navarra, Spain, Center for
Physiopathology of Obesity and Nutrition, Institute of Health Carlos III, Madrid, Spain, Navarra Institute for Health Research, Pamplona, Spain, and
an adjunct professor, Department of Nutrition, Harvard T. H. Chan School of Public Health, Boston, MA. A. Gea is assistant professor, Department
of Preventive Medicine and Public Health, School of Medicine, University of Navarra, Navarra, Spain, Center for Physiopathology of Obesity and
Nutrition, Institute of Health Carlos III, Madrid, Spain, and Navarra Institute for Health Research, Pamplona, Spain. E. Toledo is associate professor,
Department of Preventive Medicine and Public Health, School of Medicine, University of Navarra, Navarra, Spain, Center for Physiopathology of
Obesity and Nutrition, Institute of Health Carlos III, Madrid, Spain, and Navarra Institute for Health Research, Pamplona, Spain.
Address correspondence to: Estefanía Toledo, MD, MPH, PhD, Department of Preventive Medicine and Public Health, School of Medicine,
University of Navarra, Irunlarrea 1, 31008 Pamplona, Navarra, Spain. Center for Physiopathology of Obesity and Nutrition, Institute of Health
Carlos III, Madrid, Spain, and, Navarra Institute for Health Research, Pamplona, Spain. E-mail: etoledo@unav.es
STATEMENT OF POTENTIAL CONFLICT OF INTEREST
No potential conflict of interest was reported by the authors.
FUNDING/SUPPORT
The SUN Project has received funding from the Spanish Government-Instituto de Salud Carlos III, and the European Regional Development Fund
(FEDER) (RD 06/0045, CIBER-OBN, Grants PI10/02658, PI10/02293, PI13/00615, PI14/01668, PI14/01798, PI14/01764, PI17/01795, and G03/140), the
Navarra Regional Government (45/2011, 122/2014, 41/2016), and the University of Navarra.
ACKNOWLEDGEMENTS
We would like to thank all the participants involved in the SUN Project. A. Romanos-Nanclares would like to thank Fundación Científica Aso-
ciación Española Contra el Cáncer (Scientific Foundation of the Spanish Association Against Cancer) for granting her a predoctoral fellowship. We
thank, with their permission, other members of the SUN Group: A. Alonso, PhD, I. Álvarez-Álvarez, PhD, A. Balaguer, MD, M. Barbagallo, MD, I.
Barrientos, M. T. Barrio-López, PhD, F. J. Basterra-Gortari, PhD, A. Battezzati, MD , P. Bazal, S. Benito, S. Bertoli, MD, M. Bes-Rastrollo, PhD, Y. Beulen,
J. J. Beunza, PhD, P. Buil-Cosiales, PhD, M. Canales, S. Carlos, PhD, L. Carmona, MD, S. Cervantes, PhD, C. Cristobo, J. de Irala, PhD, C. de la Fuente-
Arrillaga, V. de la O, P. A. de la Rosa, M. Delgado-Rodríguez, PhD, J. Díaz-Gutiérrez, J. Díez Espino, PhD, L. Domínguez, PhD, C. Donat-Vargas, PhD,
M. Donazar, PhD, S. Eguaras, PhD, A. Fernández-Montero, PhD, U. Fresán, PhD, C. Galbete, A. García-Arellano, PhD, M. García López, PhD, M.
Gutiérrez-Bedmar, PhD, A. L. Gomes-Domingos, PhD, C. Gómez-Donoso, E. Gómez-Gracia, PhD, E. Goñi, L. Goñi, PhD, F. Guillén, MD, P. Henríquez,
MD, A. Hernández, PhD, M. S. Hershey, M. Hidalgo-Santamaría, E. Hu, F. Lahortiga, PhD, A. Leone, J. Llorca, PhD, C. López del Burgo, PhD, A. Marí,
PhD, I. Marques, PhD, A. Martí, PhD, N. Martín Calvo, PhD, J. M. Martín-Moreno, PhD, J. A. Martínez, PhD, E. H. Martínez-Lapiscina, PhD, R.
Mendonça, PhD, C. Menéndez, M. Molendijk, PhD, P. Molero, PhD, K. Murphy, PhD, M. Muñoz, J. M. Núñez-Córdoba, PhD, R. Pajares, PhD, A.
Papadaki, PhD, N. Parletta, P. Pérez de Ciriza, A. Pérez Cornago, PhD, J. Pérez de Rojas, A. M. Pimenta, PhD, J. Pons, PhD, R. Ramallal, PhD, C.
Razquin, PhD, A. Rico, C. Ruano, L. Ruiz, PhD, M. Ruiz-Canela, PhD, A. Ruiz Zambrana, PhD, E. Salgado, B. San Julián, D. Sánchez, R. Sánchez-Bayona,
A. Sánchez-Tainta, A. Sánchez-Villegas, PhD, S. Santiago, PhD, C. Sayón-Orea, PhD, J. Schlatter, MD, M. Serrano-Martinez, PhD, J. Toledo, PhD, A.
Tortosa, PhD, F. Valencia, MD, Z. Vázquez, D. Zarnowiecki, and I. Zazpe. We especially thank all of the participants in the SUN cohort for their
longstanding and enthusiastic collaboration and our advisors from Harvard T. H. Chan School of Public Health, Walter Willett, PhD, Alberto
Ascherio, PhD, Frank B. Hu, PhD, and Meir J. Stampfer, PhD, who helped us to design the SUN Project.
AUTHOR CONTRIBUTIONS
A. Romanos-Nanclares wrote the draft of the manuscript. A. Romanos-Nanclares, A. Gea, and E. Toledo were responsible for data analysis. M. A.
Martínez-González, A. Gea, and E. Toledo were responsible for project conception, design and development, data acquisition, and interpretation.
C. Sánchez-Quesada, E. Toledo, A. Gea, M. A. Martínez-González, and I. Gardeazábal were responsible for review and editing. All authors revised
the manuscript critically for important intellectual content and approved the final version to be published.
RESEARCH
Figure 2. Main components of total polyphenols (%) according to their chemical structure (A) and their main subclasses and in-
dividual compounds (B) in a sample of 10,812 women from the Seguimiento Universidad de Navarra cohort (1999 to 2018).
1015.e1 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS June 2020 Volume 120 Number 6
Table 4. HRa (95% CI) for the association between tertiles of individual phenolic acid intake and probable breast cancer risk in the SUNb cohort
Tertiles of Total Hydroxycinnamic Acids Intake Tertiles of Total Hydroxybenzoic Acids Intake
Variable Tertile 1 Tertile 2 Tertile 3 P for trend Tertile 1 Tertile 2 Tertile 3 P for trend
Overall probable cases

!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!
Cases 53 61 76 — 64 62 64 —
!!!!!!!!!!!!!! person-years!!!!!!!!!!!!!!! !!!!!!!!!!!!!! person-years!!!!!!!!!!!!!!!
Follow-up 37,676 38,112 39,581 — 39,906 38,478 36,983 —
! per 10,000 person-years!!!!!!!!!
!!!!!!!!! !! ! per 10,000 person-years!!!!!!!!!
!!!!!!!!! !!
Incidence rate 13.8 16.3 19 — 16 16.1 17.3 —
!!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!! !!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!!
Age-adjusted 1 (refc) 0.91 (0.63 to 1.32) 0.95 (0.67 to 1.36) 0.86 1 (ref) 0.87 (0.61 to 1.23) 0.83 (0.58 to 1.18) 0.34
d
Premenopausal women
!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!
Cases 29 38 47 — 44 30 39 —
Follow-up 32,318 29,603 29,666 — 33,872 30,729 26,986 —
!!!!!!!!! !! ! per 10,000 person-years!!!!!!!!!
!!!!!!!!! !!
Incidence rate 9 12.8 15.8 — 13 9.8 14.8 —
!!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!! !!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!!
f
RESEARCH
1015.e2
RESEARCH
1015.e3
Table 4. HRa (95% CI) for the association between tertiles of individual phenolic acid intake and probable breast cancer risk in the SUNb cohort (continued)
Tertiles of Total Hydroxycinnamic Acids Intake Tertiles of Total Hydroxybenzoic Acids Intake
Variable Tertile 1 Tertile 2 Tertile 3 P for trend Tertile 1 Tertile 2 Tertile 3 P for trend
!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!
Case 22 21 19 — 16 27 19 —
Follow-up 4,582 7,269 8,296 — 5,011 6,541 8,594 —
!!!!!!!!! !! ! per 10,000 person-years!!!!!!!!!
!!!!!!!!! !!
Incidence rate 48 28.9 22.9 — 31.9 41.3 22.1 —
!!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!! !!!!!!!!!!!!!!HR (95% CI)!!!!!!!!!!!!!!!
Multivariable model 1g 1 (ref) 0.59 (0.32 to 1.09) 0.45 (0.24 to 0.84) 0.017 1 (ref) 1.32 (0.71 to 2.48) 0.79 (0.40 to 1.57) 0.30
e
Multivariable model 2 1 (ref) 0.59 (0.32 to 1.08) 0.43 (0.23 to 0.80) 0.011 1 (ref) 1.51 (0.78 to 2.95) 0.85 (0.42 to 1.72) 0.33
a
HR¼hazard ratio.
b
c
ref¼reference.
d
continuous), body mass index (tertiles), age at menarche (categories), menopause (no menopause, <50 years, #50 years), number of pregnancies of more than 6 mo (continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding
(continuous), use of hormone replacement therapy (yes/no) and its duration (continuous), and years at university (continuous).
e
f
continuous), body mass index (tertiles), age at menarche (categories), number of pregnancies of more than 6 mo (continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding (continuous), and years at university (continuous).
g
continuous), body mass index (tertiles), age at menarche (categories), menopause (<50 years, #50 years), number of pregnancies of more than 6 mo (continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding (continuous), use
of hormone replacement therapy (yes/no) and its duration (continuous), years at university (continuous), and time since recruitment.
RESEARCH
Table 6. HRsa (95% CIs) for the association between tertiles of chlorogenic acids (5-caffeoylquinic acid, 4-caffeoylquinic acid, and
3-caffeoylquinic acid) intake and probable breast cancer risk in the SUNb cohort

Overall probable cases

!!!!!!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!!!!!!
Cases 53 61 76 —
!!!!!!!!!!!!!!!!!!!person-years!!!!!!!!!!!!!!!!!!!!
Follow-up 37,394 38,279 39,693 —
!!!!!!!!!!!!!!!per 10,000 person-years!!!!!!!!!!!!!!!!
! HR (95% CI)!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!! !!
Age-adjusted 1 (refc) 0.92 (0.63 to 1.33) 0.98 (0.68 to 1.39) 0.94
d
Premenopausal women
!!!!!!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!!!!!!
Cases 30 35 48 —
!!!!!!!!!!!!!!!!!!!person-years!!!!!!!!!!!!!!!!!!!!
Follow-up 31,925 30,052 29,609 —
!!!!!!!!!!!!!!!per 10,000 person-years!!!!!!!!!!!!!!!!
! HR (95% CI)!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!! !!
Multivariable-adjustedf 1 (ref) 1.03 (0.63 to 1.69) 1.27 (0.79 to 2.02) 0.29
e
!!!!!!!!!!!!!!!!!!!!!!!n!!!!!!!!!!!!!!!!!!!!!!!!
Cases 22 21 19 —
!!!!!!!!!!!!!!!!!!!person-years!!!!!!!!!!!!!!!!!!!!
Follow-up 4,638 6,971 8,538 —
!!!!!!!!!!!!!!!per 10,000 person-years!!!!!!!!!!!!!!!!
June 2020 Volume 120 Number 6 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 1015.e4
RESEARCH
Table 6. HRsa (95% CIs) for the association between tertiles of chlorogenic acids (5-caffeoylquinic acid, 4-caffeoylquinic acid, and
3-caffeoylquinic acid) intake and probable breast cancer risk in the SUNb cohort (continued)

! HR (95% CI)!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!! !!
g
a
HR¼hazard ratio.
b
c
ref¼reference.
d
(metabolic equivalents-h/wk, continuous), alcohol intake (g/day, continuous), body mass index (tertiles), age at menarche (categories), menopause (no menopause, <50 y, #50 y), number
of pregnancies of more than 6 mo (continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its
duration (continuous), and years at university (continuous).
e
f
(metabolic equivalents-h/wk, continuous), alcohol intake (g/day, continuous), body mass index (tertiles), age at menarche (categories), number of pregnancies of more than 6 mo
(continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding (continuous), and years at university (continuous).
g
(metabolic equivalents-h/wk, continuous), alcohol intake (g/day, continuous), body mass index (tertiles), age at menarche (categories), menopause (<50 y, #50 y), number of pregnancies
of more than 6 mo (continuous), pregnancy before the age of 30 y (yes/no), months of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its duration
(continuous), years at university (continuous), and time since recruitment.
1015.e5 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS June 2020 Volume 120 Number 6
Title: Carbohydrate quality index and breast cancer risk in a Mediterranean cohort: the
SUN project
Journal: Clinical Nutrition
Impact Factor: 6.77
Status: Published in May 2020
Clinical Nutrition 40 (2021) 137e145
Contents lists available at ScienceDirect
Clinical Nutrition
journal homepage: http://www.elsevier.com/locate/clnu
Original article
Carbohydrate quality index and breast cancer risk in a Mediterranean

cohort: The SUN project
Andrea Romanos-Nanclares a, b, Alfredo Gea a, b, c, *,
!
Miguel Angel Martínez-Gonza !lez a, b, c, d, Itziar Zazpe a, b, c, e, Itziar Gardeazabal f,
a, b
Cesar I. Fernandez-Lazaro , Estefanía Toledo a, b, c
a
University of Navarra, Department of Preventive Medicine and Public Health, Pamplona, Spain
b
Navarra Institute for Health Research (IdisNA), Pamplona, Spain
c
CIBERobn Physiopathology of Obesity and Nutrition, Institute of Health Carlos III (ISCIII), Madrid, Spain
d
Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, USA
e
Department of Nutrition and Food Sciences and Physiology, University of Navarra, Pamplona, Spain
f
Department of Medical Oncology, Marqu! es de Valdecilla University Hospital, Santander, Spain
a r t i c l e i n f o s u m m a r y
Article history: Background & aims: Beyond the quantity of carbohydrate intake, further research is needed on the
Received 7 November 2019 relevance of carbohydrate quality. Thus, we evaluated the association between an a priori defined car-
Accepted 24 April 2020 bohydrate quality index (CQI) and the incidence of breast cancer (BC) in a Mediterranean cohort study.
Methods: We used a validated semi-quantitative 136-item food-frequency questionnaire (FFQ) in a
Keywords: prospective follow-up study of 10,812 middle-aged women. We evaluated at baseline the CQI following 4
Breast cancer
criteria: dietary fiber intake, glycemic index, whole-grain:total-grain carbohydrates ratio and the solid
Carbohydrate quality
carbohydrate:total carbohydrate ratio. Subjects were classified into quartiles according to the final CQI
Glycemic index
Fiber
score.
Whole grain Results: During a median follow-up of 11.8 years, we confirmed 101 incident cases of BC. Our study
Liquid carbohydrates suggests that a higher quality of carbohydrate intake, as measured by the baseline CQI, was associated
with a lower risk of BC [HR Q4 vs. Q1 0.39 (95% CI 0.17, 0.87)]. Particularly, a higher whole-grain:total-grain
carbohydrates ratio was associated with lower risk of BC [HR T3 vs. T1 0.56 (0.34, 0.90)]. When we stratified
by menopausal status, we found an inverse association between CQI and BC in the comparison of
extreme quartiles among premenopausal women.
Conclusions: In this Mediterranean cohort, a better quality of dietary carbohydrate intake showed a
significant inverse association with the incidence of BC, which suggests that strategies for cancer pre-
vention should highlight the quality of this macronutrient.
© 2020 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
1. Introduction century [1]. Among different tumors, breast cancer (BC) is the
most commonly diagnosed and the worldwide leading cause of
Cancer is currently the second leading cause of death globally, cancer deaths among women [1].
and is expected to be the first cause of death in the twenty-first Diet is one modifiable lifestyle factor that has earned much
attention due to the ample and growing evidence of its capacity to
affect BC risk [2]. Carbohydrate intake represents a major source of
energy intake in all populations and it has been considered to
Abbreviations: BMI, body mass index; CI, confidence interval; CQI, Carbohydrate
modulate BC risk [3,4]. Although dietary carbohydrates are the only
Quality Index; FFQ, food frequency questionnaire; GI, Glycemic Index; HR, hazard
ratio; MET, metabolic equivalent score; PREDIMED, Prevencio !n con Dieta Medi- food constituents that directly impact blood glucose and insulin
terra
!nea; SD, standard deviation; SUN, Seguimiento Universidad de Navarra e levels, carbohydrate quality seems to have a more important role
Follow-up University of Navarra. than the total amount of carbohydrate as a percentage of dietary
* Corresponding author. University of Navarra, Department of Preventive Medi- energy in chronic disease prevention [5].
cine and Public Health, C/ Irunlarrea, 1, Pamplona, 31008, Spain.
E-mail address: ageas@unav.es (A. Gea).
https://doi.org/10.1016/j.clnu.2020.04.037
0261-5614/© 2020 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
138 A. Romanos-Nanclares et al. / Clinical Nutrition 40 (2021) 137e145
Several indices for carbohydrate quality ebased on glycemic

index (GI), glycemic load (GL), fiber, whole grains, and other factors
dimensionse have been proposed [5]. On the other hand, associa-
tions between carbohydrate sources such as total sugar, added
sugar, fructose, sucrose, sugary foods or sugar-sweetened bever-
ages [6e9], dietary fiber [10], whole grains [11], GI or GL [12] and BC
have been previously reported.
Nevertheless, a multidimensional approach ei.e. considering at
the same time the GI, fiber content, solid or liquid carbohydrates
and the degree of processinge, would represent an ideal approach
to assess the potential relationship between the nutritional quality
of carbohydrates and BC. Thus, the purpose of the present study
was to examine the association between the overall dietary quality
of carbohydrates and the risk of developing BC in a large prospec-
tive Mediterranean cohort. In secondary analyses, we also assessed
the association between widely-used markers of quality of carbo-
hydrate intake econcretely, GI and GLe and BC risk.
2. Methods
2.1. Study population
The “Seguimiento Universidad de Navarra” (SUN) Project is an

ongoing, multipurpose, dynamic cohort study designed to pro-
spectively assess the association between diet and the occurrence
of major chronic diseases [13]. Participants are all Spanish univer-
sity graduates and the recruitment is permanently open. Details of
cohort design, recruitment strategy and follow-up methods are Fig. 1. Flow-chart of participants recruited in the SUN Project, 1999e2018.
available elsewhere [14]. Information is collected every two years
using self-administered questionnaires sent either by post or
electronic mail. carbohydrate content. Refined grain was determined according to
The present study was conducted according to the guidelines 13 items linked to the consumption of white rice, refined bread,
laid down in the Declaration of Helsinki, and all procedures refined pasta and different bakery products made with refined
involving human subjects were approved by the Institutional Board grains, and then, whole grain consumption was ascertained with
of the University of Navarra. Voluntary fulfillment of the baseline the item “whole bread consumption” (serving size 60 g). The CQI
questionnaire was considered as informed consent. was defined by adding up the quintiles for the following four
Through July 2018, 22,790 participants had been recruited and criteria: dietary fiber intake (g/day, positively weighted), GI
had completed the baseline questionnaire of the SUN Project. We (negatively weighted), ratio whole grains: total grains (positively
considered the data of all women included in the cohort by that weighted), and, lastly, ratio solid carbohydrate: total carbohydrate
time (n ¼ 13,999). To include those women who had the chance to (solid þ liquid carbohydrate) (positively weighted) [21]. For each
answer at least one of the biennial follow-up questionnaires, we of these four components, we categorized participants into
included in the present analysis only those women recruited before quintiles and obtained a value (ranging from 1 to 5) according to
October 2015 (n ¼ 13,770). Out of these women, we excluded those each quintile (only for GI, those in the fifth quintile received 1
from whom we had no information during follow-up (n ¼ 1,295, point and those in the first quintile received 5 points). All criteria
retention rate: 90%), women with prevalent BC (n ¼ 104), women had the same weighting, and the total score potential range was
with a daily energy intake out of the predefined limits (below 500 from 4 to 20, with higher values meaning better quality of car-
or beyond 3500 kcal/day) (n ¼ 1,353) [15] and women with re- bohydrates (Supplementary Table 1).
ported menopause before the age of 35 years (n ¼ 206). Thus, the The CQI is described in detail elsewhere (21,22) and has been
effective sample size was 10,812 women (Fig. 1). previously used in other studies assessing nutritional adequacy,
weight gain, incident obesity and cardiovascular disease [21e24].
2.2. Dietary assessment The GI for food and beverage items was calculated by using
average values from the 2002 International tables of GI and GL
Dietary intake was evaluated using a baseline 136-item semi- values and extended in 2008 [25] with glucose as the reference.
quantitative food-frequency questionnaire (FFQ) repeatedly vali- Additionally, dietary GL was calculated considering the quality
dated in Spain [16,17]. It showed a reasonably good validity for and the amount of carbohydrate [GL ¼ (GI x amount of available
assessing carbohydrate intake [18]. There were 9 options for the carbohydrate)/100] [26]. Finally, both dietary GI and dietary GL
average frequency of consumption (from never/almost never to at were categorized into tertiles.
least 6 times/day). Trained dietitians updated the nutrient data
bank using the available information from the food composition
tables for Spain [19,20]. We used baseline dietary intake data to 2.3. Assessment of other variables
determine the CQI according to the approach previously proposed
by Zazpe et al. [21]. We calculated liquid carbohydrate intake Information about lifestyles, medical history, anthropometric
adding up sugar-sweetened beverages and fruit juices consump- measures, reproductive history, and sociodemographic factors
tion (four items in the FFQ), while solid carbohydrate corre- were gathered in the baseline assessment. BMI was calculated as
sponded to the carbohydrate content of the rest of foods with any [weight (in kilograms)/square of height (in meters2)]. Self-reported
A. Romanos-Nanclares et al. / Clinical Nutrition 40 (2021) 137e145 139
anthropometrics were previously validated in a subsample of the unavailable, we imputed postmenopausal status according to the
cohort [27]. Information on physical activity during leisure-time 75th percentile of age at menopause in the study population (52
was collected at baseline with a specific questionnaire previously years of age) [29]. When assessing premenopausal BC as the
validated in Spain [28]. To assess leisure-time physical activity, an dependent variable, we excluded women who reported having had
activity metabolic equivalent (MET) index was computed by menopause before study outset and censored follow-up time at the
designating a multiple of resting metabolic rate (MET score) to each age of 52 years or at the age of menopause, whichever happened
activity and calculating overall MET-hours per week. first. For postmenopausal BC as outcome, we only considered as
time at risk the time of follow-up after having turned 52 years old
2.4. Outcome assessment or having had their menopause, whichever happened later.
For postmenopausal women, we additionally adjusted the
The primary end point for the present analysis was the diagnosis multivariable models for time since recruitment until the begin-
of incident BC. We requested the medical record from participants ning of the time at risk and age at menopause (<50 years, $ 50
or their families when a diagnosis was self-reported in any of the years).
biennial follow-up questionnaires. An expert oncologist confirmed We also assessed the interaction between the amount of energy
the cases, and those confirmed cases were considered for the main intake from carbohydrates and CQI and assessed the relationship
analyses. When the medical record to certify the diagnosis of BC between the join classification of the tertiles of CQI and the per-
was not yet accessible possibly due to a delay in the delivery of the centage of energy intake from carbohydrates (#50%, >50% of total
documentation, we treated them as probable incident cases. Deaths energy intake) and BC risk.
were reported to our research team by the subject's next of kin, Analyses were performed with STATA version 12 (STATA Corp.,
postal authorities or work associates. We consulted the National TX, USA). All p-values are two-tailed and statistical significance was
Death Index to obtain the vital status and to identify the cause of set at the conventional cut-off of p < 0.05.
death of cohort members in case they were lost to follow-up.
3. Results
2.5. Statistical analysis
Baseline characteristics of the study participants according to
We compared the baseline characteristics of participants ac- quartiles of CQI are presented in Table 1. Higher carbohydrate
cording to quartiles of CQI. We calculated means (SD) or percent- quality intake was observed for older, postmenopausal, and more
ages for each variable across quartiles. physically active women, who have breastfed their offspring for a
We used Cox regression models to assess the relationship be- longer period. They were also more likely to be former smokers and
tween CQI and the risk of BC. Hazard ratios (HRs) with 95% confi- have a higher prevalence of diabetes. Moreover, these women were
dence intervals (CIs) were calculated for the three upper quartiles more prone to follow a special diet, and had better adherence to the
using the lowest quartile of CQI as the reference category. We also Mediterranean diet and higher energy intake. As expected, we
used the CQI as a continuous variable and assessed changes in the observed a higher total energy intake from carbohydrates, solid
outcome associated with a two-point increase. Furthermore, we carbohydrates, fiber, carbohydrates from whole-grains among
repeated Cox regression models using individual components of women with better CQI. Contrarily, CQI was inversely associated
CQI dietary fiber intake, GI, ratio of whole-grain: total-grain car- with GI and consumption of liquid carbohydrates or carbohydrates
bohydrates and ratio solid carbohydrates: (solid from refined grains.
carbohydrates þ liquid carbohydrates) and GL as the main During a median follow-up of 11.8 years and 115,802 person-
exposure. years, we observed 101 confirmed cases of BC. Participants in the
For all the exposures, we fitted two multivariable models after highest quartile of CQI showed a significant lower risk of confirmed
adjusting for the main BC risk factors and other potential con- BC during follow-up than those in the lowest quartile (HR Q4 vs. Q1
founders. Model 1 included height (continuous), years at univer- 0.39, 95% CI 0.17, 0.87) in the fully adjusted model (Table 2).
sity (continuous), family history of BC (yes/no), smoking status Furthermore, we performed sensitivity analysis including also
(never smoker, former smoker, current smoker), physical activity probable incident cases (self-reported in the baseline question-
(METs-h/ week, tertiles), alcohol intake (continuous), BMI naire). Thus, after 11.5 years of follow-up and 115,368 person-years,
(continuous), age at menarche (#11 years, 12e13 years, $14 190 probable incident cases were reported. In the multivariable
years), number of pregnancies of more than 6 months (contin- adjusted model, the highest quartile of carbohydrate quality
uous), pregnancy before the age of 30 years (yes/no), months of compared to the lowest quartile was associated with a relative risk
breastfeeding (continuous), use of hormonal therapy (yes/no) and reduction of 48% for the overall sample of women (HR Q4 vs. Q1 0.52,
its duration (continuous), menopause (no menopause, < 50 years, 95% CI 0.31, 0.89).
$ 50 years). Model 2 was additionally adjusted for energy intake Table 3 shows the incidence of BC during follow-up according to
(continuous), adherence to the Mediterranean diet (kcal/day, the comparisons between the highest vs the lowest tertile of indi-
continuous) and adherence to a special diet (yes/no). vidual components of CQI and GL. A higher baseline GI (HR T3 vs. T1
Tests of linear trend across successive quartiles of CQI (or tertiles 1.45, 95% CI 0.70, 3.01) and GL (HR T3 vs. T1 1.53, 95% CI 0.92, 2.54)
for individual components of CQI and GL) were conducted assign- was associated with a higher risk of BC in the fully-adjusted
ing the median value to each category and treating the resulting multivariable model although no significant differences were
variables as continuous. We also presented the p-value obtained for found. Furthermore, a higher ratio of whole-grain:total-grain car-
the independent variable introduced in the model as a continuous bohydrates (tertile 3) compared to the reference (tertile 1) was
variable. associated with a relative risk reduction of 44% for overall BC.
Stratified analyses by menopausal status were performed. In- Inconclusive results were found for both solid carbohydrates: (solid
formation on age at menopause was collected at baseline and after carbohydrates þ liquid carbohydrates) ratio and for dietary fiber
16 years of follow-up. If the information on age at menopause was intake.
Table 1
Baseline characteristics of participants according to quartiles of the carbohydrate quality index: the Seguimiento Universidad de Navarra (SUN) Project: 1999e2018.
Variable Q1 Q2 Q3 Q4
N 3296 2371 3089 2056

Index range (points) 4e9 10e11 12e14 15e20
Age (years) 32 (10) 34 (10) 35 (11) 37 (11)
Body mass index (kg/m2) 22.0 (3.0) 22.2 (3.0) 22.4 (3.1) 22.4 (3.0)
Physical activity (METs-h/week) 17.5 (16.4) 19.0 (16.4) 21.1 (19.0) 24.6 (23.2)
Alcohol intake (g/d) 4.0 (5.8) 4.1 (6.1) 4.0 (6.0) 4.0 (5.7)
Years at university 4.8 (1.3) 4.8 (1.3) 4.9 (1.4) 4.9 (1.4)
Height (cm) 164 (6) 164 (6) 164 (6) 164 (6)
Pregnancy before 30 years (%)
Yes 20.5 23.5 24.4 25.5
Number of pregnancies of more than 6 months 0.6 (1.1) 0.7 (1.2) 0.71 (1.1) 0.7 (1.2)
Breast-feeding (months) 2.3 (4.9) 2.8 (5.3) 2.8 (5.1) 3.1 (5.6)
Hormone replacement therapy 1
Yes 39.4 42.7 38.5 34.4
Time of hormone replacement therapy (years) 1 1.4 (2.5) 1.1 (2.2) 1.4 (2.5) 1.2 (2.2)
Diabetes (%) 0.7 1.0 1.1 1.9
Smoking (%)
Never smoker 49.4 51.2 52.1 52.1
Current smoker 27.3 24.2 20.8 16.9
Former smoker 22.7 23.8 26.2 30.1
Family history of breast cancer (%) 2
Yes 10.7 9.9 10.6 11.2
Age at menarche (%)
# 11 years 18.8 20.4 20.5 21.8
12e13 years 55.2 54.2 55.2 54.1
$14 years 26.0 25.4 24.3 24.1
Premenopausal (%) 92.4 89.9 87.2 83.8
Postmenopausal (%) 7.6 10.1 12.8 16.2
Age at menopause 1
Postmenopausal < 50 years (%) 57.0 54.0 48.9 53.6
Postmenopausal $ 50 years (%) 43.0 46.0 51.1 46.4
Adherence to the Mediterranean diet 2.5 (1.4) 3.3 (1.5) 3.8 (1.4) 4.6 (1.4)
Special diet
Yes 5.5 6.5 9.4 13.3
Total energy intake (kcal/day) 2152 (564) 2326 (573) 2352 (567) 2422 (548)
Glycemic index 53.0 (4.6) 52.0 (4.4) 51.1 (4.4) 50.0 (4.3)
Carbohydrate intake (% of total energy) 42.2 43.0 43.7 44.8
Solid carbohydrate intake (g/d) 192.0 (66.7) 219.0 (71.4) 227.4 (74.0) 245.1 (71.2)
Liquid carbohydrate intake (g/d) 36.1 (19.6) 32.2 (20.3) 30.9 (20.1) 25.7 (16.6)
Fiber intake (g/d) 19.4 (6.1) 26.3 (7.7) 31.8 (10.5) 41.6 (12.8)
Carbohydrate from whole grains intake (g/d) 0.4 (3.0) 2.6 (7.8) 8.9 (15.1) 21.8 (23.6)
Carbohydrate from refined grains intake (g/d) 86.4 (46.5) 92.1 (51.2) 82.6 (48.7) 67.4 (39.2)
Values are expressed as mean (SD) unless otherwise stated. (1) Only for postmenopausal women. (2) information from mother, sisters, and both grandmothers was collected.
We stratified our analyses by menopausal status (Fig. 2). When energy from carbohydrate intake had the lowest risk of BC after
we compared the highest quartile vs. the lowest quartile of CQI, we adjusting for a wide array of potential confounders [HR T3 vs.T1 for
found a statistically significant inverse association between CQI and those with carbohydrate intake # 50% 0.68 (95% CI 0.35, 1.31)].
BC risk among premenopausal women both in the analyses for
confirmed BC (HR T3 vs. T1 0.32, 95% CI 0.10, 1.00) and when we also 4. Discussion
added the probable cases of BC (HR T3 vs. T1 0.46, 95% CI 0.22, 0.96).
We also performed sensitivity analyses in order to assess the In this prospective cohort study, we observed that those
robustness of our results. We excluded participants with baseline women with best CQI eespecially premenopausal women- had a
diabetes and change implausible energy limits to percentiles 1 and significantly lower risk of BC. The CQI was defined by considering
99 and including either confirmed or probable cases. The results did jointly the dietary fiber, the GI, the solid or liquid form of carbo-
not substantially change in any of these scenarios (Fig. 4). Thus, the hydrates and the degree of processing of carbohydrate-rich foods
inverse association between higher quality of carbohydrates and (whole or refined grains). The total quantity of carbohydrate
risk of BC was robust and remained statistically significant in most intake or the proportion of energy explained by carbohydrates,
sensitivity analyses although the linear trend was marginally sig- their quality and their dietary sources comprise an emerging
nificant in some cases. However, the limited number of confirmed concern for cancer prevention. Nonetheless, GI and GL have been
and probable cases may hinder obtaining any specific conclusion. widely used as markers of carbohydrate quality, although there is
Figure 3 represents the HR of BC according to the joint classi- no agreement on the best standard [30]. Nevertheless, recent
fication by both baseline tertiles of CQI and the percentage of discussion about GI indicate that fiber and whole grain could also
energy from carbohydrates (#50%, >50%). The results suggest an be appropriate markers of carbohydrate quality, even to a better
inverse association between higher CQI and BC, and an inverse extent than GI or GL [31].
association between carbohydrates intake and BC, which was only Several studies have previously used the CQI and has linked it
present when the CQI was high. However, no statistically signifi- to lower risk of cardiovascular disease [22], overweight and
cant interaction was found between carbohydrate quality and obesity [24], and better micronutrient intake adequacy both in
quantity. In any case, participants with higher quality of carbo- elderly [23] and among middle-age participants [21]. In addition,
hydrate intake (CQI in the upper tertile) and with less than 50% of evidence on the association between several of the individual
Table 2
a
Hazard ratios and 95% confidence intervals for confirmed and probable incident breast cancer by quartiles of carbohydrate quality index in the SUN Project (1999e2018).
Confirmed cases Quartiles of CQI For each additional 2 points in the score
Q1 Q2 Q3 Q4 p for trend P for continuous estimate
N 3296 2371 3089 2056

CQI range 4e9 10e11 12e14 15e20
Person-years 36,674 25,770 32,618 20,740
Age-adjusted 1 (ref.) 1.37 (0.82, 2.27) 1.07 (0.64, 1.77) 0.43 (0.20, 0.92) 0.068 0.92 (0.80, 1.03) 0.174
Model 1 1 (ref.) 1.36 (0.82, 2.26) 1.09 (0.65, 1.81) 0.43 (0.20, 0.92) 0.075 0.92 (0.80, 1.03) 0.177
Model 2 1 (ref.) 1.30 (0.77, 2.19) 1.02 (0.59, 1.75) 0.39 (0.17, 0.87) 0.049 0.89 (0.77, 1.02) 0.127
Confirmed þ probable cases
N 3296 2371 3089 2056

CQI range 4e9 10e11 12e14 15e20
Person-years 36,546 25,679 32,463 20,679
Age-adjusted 1 (ref.) 1.02 (0.69, 1.50) 0.99 (0.68, 1.42) 0.58 (0.36, 0.93) 0.057 0.93 (0.84, 1.01) 0.111
Model 1 1 (ref.) 1.02 (0.69, 1.49) 0.99 (0.69, 1.43) 0.58 (0.36, 0.94) 0.062 0.93 (0.84, 1.01) 0.117
Model 2 1 (ref.) 0.97 (0.65, 1.44) 0.93 (0.63, 1.37) 0.52 (0.31, 0.89) 0.034 0.91 (0.81, 1.00) 0.072
Model 1 additionally adjusted for height (continuous), years at university (continuous), family history of BC (yes/no), smoking status (never smoker, former smoker, current
smoker), physical activity (METs-h/ week, tertiles), alcohol intake (continuous), BMI (continuous), age at menarche (# 11 years, 12e13 years, $14 years), number of preg-
nancies of more than 6 months (continuous), pregnancy before the age of 30 years (yes/no), months of breastfeeding (continuous), use of hormonal therapy (yes/no) and its
duration (continuous), menopause (no menopause, < 50 years,$ 50 years). Model 2 additionally adjusted for energy intake (continuous), adherence to the Mediterranean diet
(kcal/day, continuous) and special diet (yes/no).
a
Participants were categorized according to the following item: the ratio of carbohydrates from whole grains to carbohydrates from total grains (whole grains þ refined
grains þ products prepared with refined flours); the GI (negatively weighted); the ratio of solid carbohydrates to total (solid þ liquid) carbohydrates; and finally, total dietary
fiber intake (g/d). For each of these items, participants were categorized into quintiles and received their quintile values (ranging from 1 to 5 or from 5 to 1 for the GI). The four
criteria had the same weight, and we constructed the CQI summing up all values (total score potential range: from 4 to 20).
components included in the multidimensional carbohydrate the consumption of unrefined or unprocessed solid carbohydrates,
quality index and BC has been previously described. A recent such as whole grains may be linked to decreased BC risk [38].
meta-analyses showed a 12% decrease in BC risk with dietary fiber Whole-grains may lessen the post-prandial glucose and insulin
intake by extracting the risk estimate of the highest and lowest responses leading to better glycemic control [39] which in turn may
reported categories of intake from each of the twenty-four reduce BC risk [11].
selected studies [10]. Furthermore, an increase of 10 g/d of di- These dietary components have also been found to be associated
etary fiber intake was associated with a 4% relative risk reduction with reduced levels of inflammatory markers and liver enzymes, and
of BC. Dietary fiber can promote the discharge and decomposition therefore, a decreased risk of cancer [40]. Whole grains are rich in
of harmful and carcinogenic substances in the gut, promote the antioxidants, including vitamins evitamin C and E and b-carotenee
growth of probiotics and inhibit the growth of pathogenic bacte- and minerals eselenium, zinc, copper, and manganesee which are
ria, thereby inhibiting production of carcinogens and promoting elements of enzymes with antioxidant activities and have been
their decomposition in the intestine [10]. It also improves the linked to be inversely associated with BC risk [41]. Finally, whole
phagocytosis of macrophages, blocks nitrosamine synthesis, and grains are an important source of some non-nutrients, such as
reduces estrogen levels [10]. phytoestrogens and polyphenols (phenolic acids and lignans), that
In the last decades, the role of GI in developing chronic con- have antioxidant properties and potential to inhibit cell proliferation
ditions related to diet has received major attention, albeit it is still and angiogenesis, and to induce cell apoptosis [38].
an issue of debate. In a recent meta-analysis, for every additional We acknowledge that the results of the present study might be
10 units/day of GI the risk of BC was increased by 6% in post- interpreted in light of some limitations. First, the use of a self-
menopausal women, whereas no increased risk was observed reported FFQ for dietary assessment could have led to non-
among premenopausal women. Studies have indicated that diets differential misclassification and measurement errors and may
high in GI or GL might be linked with hyperinsulinemia [32], probably underestimate true associations. However, the FFQ is the
higher concentrations of insulin-like growth factor (IGF) 1 [32], most efficient and feasible tool to evaluate food habits in large
type 2 diabetes [33], and inflammatory biomarkers [32], all of epidemiological studies [42]. Furthermore, reliability and validity
which play a role in BC carcinogenesis [12]. Chronically, raised of the FFQ used in our cohort have been evaluated and showed
blood insulin (as a consequence of chronically high blood glucose) good correlation with nutrient intake [17]. Second, we cannot rule
is thought to be the underlying mechanism for GI/GL-related out residual confounding by unknown factors. Nonetheless, we
cancers. Insulin acts as growth factor and also increases the attempted to adjust for a wide array of established or hypothe-
bioactivity of IGFs, such as IGF-1, which can promote tumor sized medical, lifestyle, and dietary risk factors for BC. In this line,
development by inhibiting apoptosis, stimulating cell prolifera- our sample was composed of highly educated participants who
tion and sex-steroid synthesis [34] and promoting angiogenesis are not representative of the general Spanish population which
[35]. Additional conditions linked to chronically high blood could limit the external validity of our results. However, restric-
glucose, such insulin resistance, obesity, and type 2 diabetes may tion to highly educated participants helps to prevent or at least to
also impact cancer risk [36]. lower confounding by known factors, such as educational level
Furthermore, the physical form (solid vs liquid) of carbohydrates and socio-economic status. However, our results need to be
is an important concern. In this sense, prospective investigations in generalized on the basis of biological mechanisms and not on the
several cohorts have confirmed that consumption of liquid carbo- representativeness of our sample. Third, the small number of
hydrates may be positively associated with BC risk [7,8,37], whereas observed BC cases, particularly among postmenopausal women,
Table 3
Hazard ratios and 95% confidence intervals for confirmed incident breast cancer by tertiles of individual components of CQI and glycemic load in the SUN Project (1999e2018).
Tertiles of glycemic load p for trend
T1 T2 T3
N 3604 3604 3604

Person-years 37,774 38,698 39,330
Age-adjusted 1 (ref.) 1.11 (0.67, 1.82) 1.22 (0.76, 1.97) 0.416
Model 1 1 (ref.) 1.12 (0.68, 1.84) 1.24 (0.77, 2.01) 0.381
Model 2 1 (ref.) 1.23 (0.69, 2.17) 1.45 (0.70, 3.01) 0.327
Tertiles of glycemic index p for trend
T1 T2 T3
N 3604 3604 3604

Person-years 37,672 38,807 39,322
Age-adjusted 1 (ref.) 1.16 (0.70, 1.93) 1.46 (0.91, 2.36) 0.115
Model 1 1 (ref.) 1.16 (0.70, 1.93) 1.49 (0.92, 2.41) 0.103
Model 2 1 (ref.) 1.19 (0.71, 2.01) 1.53 (0.92, 2.54) 0.097
Tertiles of solid carbohydrates:(solid carbohydrates þ liquid carbohydrates) ratio
N 3604 3604 3604

Person-years 39,107 38,700 37,995
Age-adjusted 1 (ref.) 1.08 (0.65, 1.79) 1.34 (0.83, 2.15) 0.239
Model 1 1 (ref.) 1.07 (0.64, 1.78) 1.33 (0.82, 2.15) 0.259
Model 2 1 (ref.) 1.07 (0.64, 1.79) 1.33 (0.82, 2.17) 0.253
Tertiles of whole-grain: total-grain carbohydrates ratio
N 6322 886 3604

Person-years 69,336 9480 36,985
Age-adjusted 1 (ref.) 0.66 (0.29, 1.52) 0.57 (0.36, 0.92) 0.026
Model 1 1 (ref.) 0.68 (0.29, 1.57) 0.57 (0.35, 0.91) 0.024
Model 2 1 (ref.) 0.67 (0.29, 1.55) 0.56 (0.34, 0.90) 0.022
Tertiles of dietary fiber intake (g/d)
N 3604 3604 3604

Person-years 38,941 39,243 37,618
Age-adjusted 1 (ref.) 0.80 (0.48, 1.32) 0.98 (0.60, 1.57) 0.942
Model 1 1 (ref.) 0.81 (0.49, 1.35) 1.00 (0.62, 1.62) 0.869
Model 2 1 (ref.) 0.81 (0.46, 1.42) 0.98 (0.50, 1.93) 0.872
Model 1 additionally adjusted for height (continuous), years at university (continuous), family history of BC (yes/no), smoking status (never smoker, former smoker, current
smoker), physical activity (METs-h/ week, tertiles), alcohol intake (continuous), BMI (continuous), age at menarche (#11 years, 12e13 years, $14 years), number of preg-
nancies of more than 6 months (continuous), pregnancy before the age of 30 years (yes/no), months of breastfeeding (continuous), use of hormonal therapy (yes/no) and its
duration (continuous), menopause(no menopause, < 50 years, $ 50 years). Model 2 additionally adjusted for energy intake (continuous), adherence to the Mediterranean diet
(kcal/day, continuous) and special diet (yes/no).
Fig. 2. Hazard ratios and 95% confidence intervals for confirmed (a) and probable (b) incident premenopausal and postmenopausal breast cancer by quartiles of CQI. Adjusted for
height (continuous), years at university (continuous), family history of BC (yes/no), smoking status (never smoker, former smoker, current smoker), physical activity (METs-h/ week,
tertiles), alcohol intake (continuous), BMI (continuous), age at menarche (#11 years, 12e13 years, $14 years), number of pregnancies of more than 6 months (continuous),
pregnancy before the age of 30 years (yes/no), months of breastfeeding (continuous), energy intake (continuous), adherence to the Mediterranean diet (kcal/day, continuous) and
special diet (yes/no). For postmenopausal women we further adjusted for use of hormonal therapy (yes/no), its duration (continuous), menopause (no menopause, <50 years, $ 50
years) and time since recruitment.
Fig. 3. Hazard ratios of incident breast cancer according to the joint classification by tertiles of the carbohydrate quality index (CQI) and percentage of energy from total carbo-
hydrate intake.
Fig. 4. Sensitivity analyses. Hazard ratios (HRs) and 95% confidence intervals (CI) of incident breast cancer for the fourth quartile of CQI compared with the first quartile.
may have led to a low statistical power in some analyses and lead Acknowledgments
to excessively wide confidence intervals. However, after finding a
similar point estimate between pre- and postmenopausal women, The authors thank the implication and collaboration of the
despite not being significant, one could expect the effect to be participants in the SUN Project. ARN was supported by the
similar. Therefore replication of our findings in larger and older Fundacio !n Científica Asociacio ! n Espan~ ola Contra el Ca !ncer (AECC)
cohorts should be warranted. Nonetheless, many of our results (Scientific Foundation of the Spanish Association Against Cancer).
can be integrated in future meta-analyses and contribute to the We thank other members of the SUN Group: Alonso A, Alvarez- !
evidence on this new topic. Fourth, the controversy in defining !
Alvarez I, Balaguer A, Barbagallo M, Barrientos I, Barrio-Lo !pez MT,
whole grain, carbohydrate quality, or restricted carbohydrate diet Basterra-Gortari FJ, Battezzati A, Bazal P, Benito S, Bertoli S, Bes-
could make the comparison with previous results difficult. Fifth, Rastrollo M, Beulen Y, Beunza JJ, Buil-Cosiales P, Canales M, Carlos
the use of self-reported dietary intake was used to define the S, Carmona L, Cervantes S, Cristobo C, de Irala J, de la Fuente-
exposure of the analysis based only in the baseline information, Arrillaga C, de la O V, de la Rosa PA, Delgado-Rodríguez M, Díaz-
and it is possible that participants may have changed their dietary Gutie!rrez J, Díez Espino J, Domínguez L, Donat-Vargas C, Donazar M,
habits and, thus, we may have lost some information on the Eguaras S, Fern! andez-Montero A, Fres! an U, Galbete C, García-Are-
defined CQI. Sixth, our definition of the CQI cannot capture all llano A, García Lo ! pez M, Gutie!rrez-Bedmar M, Gome !z-Domingos
elements of carbohydrate intake in the context of the overall di- AL, Go!mez-Donoso C, Go !mez-Gracia E, Gon ~ i E, Gon~ i L, Guille
!n F,
etary pattern (e.g., pasta and rice are refined grains, but if eaten Henríquez P, Herna !ndez A, Hershey MS, Hidalgo-Santamaría M, Hu
within a Mediterranean diet, they may have a different effect on E, Lahortiga F, Leone A, Llorca J, Lo!pez del Burgo C, Marí A, Marques
health consequences than refined carbohydrates consumed in a I, Martí A, Martín Calvo N, Martín-Moreno JM, Martínez JA, Martí-
framework of an unhealthy dietary pattern). nez-Gonza !lez MA, Martínez-Lapiscina EH, Mendonça R, Mene !ndez
Some strengths of the present study are the use of reliable C, Molendijk M, Molero P, Murphy K, Mun ~ oz M, Nún ~ ez-Co !rdoba JM,
measures and data to ascertain exposures, outcomes and covariates Pajares R, Papadaki A, Parletta N, Pe !rez de Ciriza P, Pe !rez Cornago A,
obtained from a large cohort with a long-term follow-up and high Pe!rez de Rojas J, Pimenta AM, Pons J, Ramallal R, Razquin C, Rico A,
retention, the prospective design, the multiple variables that can be Ruano C, Ruiz L, Ruiz-Canela M, Ruiz Zambrana A, Salgado E, San
included as potential confounders and use of validated question- Juli!
an B, S!anchez D, Sa!nchez-Bayona R, Sa !nchez-Tainta A, Sa !nchez-
naires. Also, BC cases were ascertained using medical records and Villegas A, Santiago S, Sayo !n-Orea C, Schlatter J, Serrano-Martinez
confirmed by trained oncologists. M, Toledo J, Tortosa A, Valencia F, Va !zquez Z, Zarnowiecki D.
In conclusion, in this Mediterranean prospective cohort study, We thank very specially all participants in the SUN cohort for
the CQI, defined in a multidimensional way, showed a significant their long-standing and enthusiastic collaboration and our advisors
inverse association with the incidence of BC during follow-up. from Harvard TH Chan School of Public Health Walter Willett,
These results contribute to highlight the importance that carbo- Alberto Ascherio, Frank B. Hu and Meir J. Stampfer who helped us to
hydrate intake guidelines related to cancer prevention, should be design the SUN Project, the PREDIMED study and the PREDIMED-
focused on improving quality (increasing dietary fiber, whole grains PLUS on-going trial.
consumption, preferring solid carbohydrates and choosing low GI ARN wrote the draft of the manuscript; ARN, AG and ET were
foods), rather than limiting quantity or percentage of total energy responsible for data analysis; MAMG, AG, ET and IZ were respon-
from carbohydrates. sible for project conception, design and development, data acqui-
In this sense, eating less refined grains and more whole grain sition, and interpretation; IZ, IG, C-FL, ET, AG and MAMG were
foods remain an important healthy dietary recommendation in responsible for review and editing. All authors have revised the
order to protect against chronic diseases. In the framework of manuscript critically for important intellectual content and
public health and nutrition policies, our findings should stimulate approved the final version to be published.
the development of dietary guidelines based on macronutrient
quality rather quantity, focusing on carbohydrates, and also to Appendix A. Supplementary data
support the adherence to the traditional Mediterranean dietary
pattern. In this context, the last US Dietary Guidelines published in Supplementary data related to this article can be found at
2010 and the American Institute for Cancer Research, advices to https://doi.org/10.1016/j.clnu.2020.04.037.
limit the consumption of foods rich in refined starchy grains, and to
consume at least half of all grains as whole grains [43,44]. References
[1] Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer
statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide
Funding sources for 36 cancers in 185 countries. CA A Cancer J Clin 2018;68:394e424.
[2] World Cancer Research Fund, American Institute for Cancer Research. Diet,
nutrition, physical activity and breast cancer. https://www.wcrf.org/sites/
The SUN Project has received funding from the Spanish default/files/Breast-Cancer-2017-Report.pdf. [Accessed 31 October 2019].
Government-Instituto de Salud Carlos III, and the European Accessed.
Regional Development Fund (FEDER) (RD 06/0045, CIBER-OBN, [3] Maino Vieytes CA, Taha HM, Burton-Obanla AA, Douglas KG, Arthur AE. Car-
bohydrate nutrition and the risk of cancer. Curr Nutr Rep 2019;8:230e9.
Grants PI13/00615, PI10/02658, PI10/02293, PI13/00615, PI14/ https://doi.org/10.1007/s13668-019-0264-3.
01668, PI14/01798, PI14/01764, PI17/01795 and G03/140), the [4] Giovannucci E, Harlan DM, Archer MC, Bergenstal RM, Gapstur SM, Habel LA,
Navarra Regional Government (27/2011, 45/2011, 122/2014), and et al. Diabetes and cancer: a consensus report. Diabetes Care 2010;60:207e21.
https://doi.org/10.2337/dc10-0666.
the University of Navarra. Andrea Romanos-Nanclares has a per-
[5] Ludwig DS, Hu FB, Tappy L, Brand-Miller J. Dietary carbohydrates: role of
sonal grant from Asociacio
! n Espan
~ ola Contra el Ca
!ncer. quality and quantity in chronic disease. BMJ 2018;361. https://doi.org/
10.1136/bmj.k2340.
[6] Makarem N, Bandera EV, Nicholson JM, Parekh N. Consumption of sugars,
sugary foods, and sugary beverages in relation to cancer risk: a systematic
Conflict of interest review of longitudinal studies. Annu Rev Nutr 2018;38:17e39. https://doi.org/
10.1146/annurev-nutr-082117-051805.
[7] Hodge AM, Bassett JK, Milne RL, English DR, Giles GG. Consumption of sugar-
The authors declare no conflict of interest. sweetened and artificially sweetened soft drinks and risk of obesity-related
cancers. Publ Health Nutr 2018;21:1618e26. https://doi.org/10.1017/ disease in US women. Am J Clin Nutr 2000;71:1455e61. https://doi.org/
S1368980017002555. 10.1093/ajcn/71.6.1455.
[8] Romanos-Nanclares A, Toledo E, Gardeazabal I, Jime !nez-Moleo ! n JJ, Martínez- [27] Bes-Rastrollo M, Pe !rez Valdivieso JR, Sa ! Martínez-
!nchez-Villegas A, Alonso A,
Gonza !lez MA, Gea A. Sugar-sweetened beverage consumption and incidence Gonza ! Validacio
!lez MA. ! n del peso e índice de masa corporal auto-declarados
of breast cancer: the Seguimiento Universidad de Navarra (SUN) Project. Eur J de los participantes de una cohorte de graduados universitarios. Rev Espn ~ ola
Nutr 2018;48:70e225. https://doi.org/10.1007/s00394-018-1839-2. Obes 2005;3:352e8.
[9] Duchaine CS, Dumas I, Diorio C. Consumption of sweet foods and mammo- [28] Martinez-Gonzalez MA, Lopez-Fontana C, Varo JJ, Sanchez-Villegas A,
graphic breast density: a cross-sectional study. BMC Publ Health 2014;14:554. Martinez JA. Validation of the Spanish version of the physical activity ques-
https://doi.org/10.1186/1471-2458-14-554. tionnaire used in the nurses' health study and the health professionals'
[10] Chen S, Chen Y, Ma S, Zheng R, Zhao P, Zhang L, et al. Dietary fibre intake and follow-up study. Publ Health Nutr 2005;8:920e7. doi:S1368980005001230
risk of breast cancer: a systematic review and meta-analysis of epidemio- [pii].
logical studies. Oncotarget 2016. https://doi.org/10.18632/oncotarget.13140. [29] Shivappa N, Sandin S, Lof M, Hebert JR, Adami H-O, Weiderpass E. Prospective
[11] Xiao Y, Ke Y, Wu S, Huang S, Li S, Lv Z, et al. Association between whole grain study of dietary inflammatory index and risk of breast cancer in Swedish
intake and breast cancer risk: a systematic review and meta-analysis of women. Br J Canc 2015;113:1099e103. https://doi.org/10.1038/bjc.2015.304.
observational studies. Nutr J 2018;17:87. https://doi.org/10.1186/s12937- [30] Øverby NC, Sonestedt E, Laaksonen DE, Birgisdottir BE. Dietary fiber and the
018-0394-2. glycemic index: a background paper for the Nordic Nutrition Recommenda-
[12] Schlesinger S, Chan DSM, Vingeliene S, Vieira AR, Abar L, Polemiti E, et al. tions 2012. Food Nutr Res 2013;57. https://doi.org/10.3402/fnr.v57i0.20709.
Carbohydrates, glycemic index, glycemic load, and breast cancer risk: a sys- [31] Wolever TMS. Is glycaemic index (GI) a valid measure of carbohydrate qual-
tematic review and dose-response meta-analysis of prospective studies. Nutr ity? Eur J Clin Nutr 2013;67:522e31. https://doi.org/10.1038/ejcn.2013.27.
Rev 2017;75:420e41. https://doi.org/10.1093/nutrit/nux010. [32] Augustin LSA, Kendall CWC, Jenkins DJA, Willett WC, Astrup A, Barclay AW,
[13] Martínez-Gonz! alez MA, S! anchez-Villegas A, De Irala J, Marti A, Martínez JA. et al. Glycemic index, glycemic load and glycemic response: an International
Mediterranean diet and stroke: objectives and design of the SUN project. Nutr scientific consensus summit from the International Carbohydrate Quality
Neurosci 2002;5:65e73. https://doi.org/10.1080/10284150290007047. Consortium (ICQC). Nutr Metabol Cardiovasc Dis 2015;25:795e815. https://
[14] Carlos S, De La Fuente-Arrillaga C, Bes-Rastrollo M, Razquin C, Rico-Campa " A, doi.org/10.1016/j.numecd.2015.05.005.
Martínez-Gonz! alez M, et al. Mediterranean diet and health outcomes in the [33] Greenwood DC, Threapleton DE, Evans CEL, Cleghorn CL, Nykjaer C,
SUN cohort. Nutrients 2018;10:439. https://doi.org/10.3390/nu10040439. Woodhead C, et al. Glycemic index, glycemic load, carbohydrates, and type 2
[15] Willett W. Nutritional epidemiology. 3rd ed. New York: Oxford University diabetes: systematic review and dose-response meta-analysis of prospective
Press; 2013. studies. Diabetes Care 2013;36:4166e71. https://doi.org/10.2337/dc13-0325.
[16] Martin-Moreno JM, Boyle P, Gorgojo L, Maisonneuve P, Fernandez- [34] Biddinger SB, Ludwig DS. The insulin-like growth factor axis: a potential link
Rodriguez JC, Salvini S, et al. Development and validation of a food frequency between glycemic index and cancer. Am J Clin Nutr 2005;82:277e8.
questionnaire in Spain. Int J Epidemiol 1993;22:512e9. [35] Samani AA, Yakar S, LeRoith D, Brodt P. The role of the IGF system in cancer
[17] De la Fuente-Arrillaga C, Va !zquez Ruiz Z, Bes-Rastrollo M, Sampson L, Mar- growth and metastasis: overview and recent insights. Endocr Rev 2007;28:
tinez-Gonza !lez MA. Reproducibility of an FFQ validated in Spain. Publ Health 20e47. https://doi.org/10.1210/er.2006-0001.
Nutr 2010;13:1364e72. https://doi.org/10.1017/S1368980009993065. [36] Brand-Miller JC. Glycemic load and chronic disease. Nutr Rev 2003;61:49e55.
[18] Ferna!ndez-Ballart JD, Pin~ ol JL, Zazpe I, Corella D, Carrasco P, Toledo E, et al. https://doi.org/10.1301/nr.2003.may.s49-s55.
Relative validity of a semi-quantitative food-frequency questionnaire in an [37] Chazelas E, Srour B, Desmetz E, Kesse-Guyot E, Julia C, Deschamps V, et al.
elderly Mediterranean population of Spain. Br J Nutr 2010;103:1808e16. Sugary drink consumption and risk of cancer: results from NutriNet-Sante !
https://doi.org/10.1017/S0007114509993837. prospective cohort. BMJ 2019;366. https://doi.org/10.1136/bmj.l2408.
[19] Mataix Verdú FJ. Tabla de composicio !n de alimentos. Instituto de Nutricio !n y [38] Xie M, Liu J, Tsao R, Wang Z, Sun B, Wang J. Whole grain consumption for the
Tecnología de Alimentos, Universidad de Granada; 2003. prevention and treatment of breast cancer. Nutrients 2019. https://doi.org/
[20] Moreiras O. Tablas de composicio !n de alimentos : guía de pr!acticas. Pira
!mide; 10.3390/nu11081769.
2015. [39] Chandalia M, Garg A, Lutjohann D, von Bergmann K, Grundy SM, Brinkley LJ.
[21] Zazpe I, S! anchez-Taínta A, Santiago S, de la Fuente-Arrillaga C, Bes- Beneficial effects of high dietary fiber intake in patients with type 2 diabetes
Rastrollo M, Martínez JA, et al. Association between dietary carbohydrate mellitus. N Engl J Med 2002;342:1392e8. https://doi.org/10.1056/
intake quality and micronutrient intake adequacy in a Mediterranean cohort: nejm200005113421903.
the SUN (Seguimiento Universidad de Navarra) Project. Br J Nutr 2014;111: [40] Mok Y, Son DK, Yun YD, Jee SH, Samet JM. g-Glutamyltransferase and cancer
2000e9. https://doi.org/10.1017/s0007114513004364. risk: the Korean cancer prevention study. Int J Canc 2016;138:311e9. https://
[22] Zazpe I, Santiago S, Gea A, Ruiz-Canela M, Carlos S, Bes-Rastrollo M, et al. doi.org/10.1002/ijc.29659.
Association between a dietary carbohydrate index and cardiovascular disease [41] Montonen J, Boeing H, Fritsche A, Schleicher E, Joost HG, Schulze MB, et al.
in the SUN (Seguimiento Universidad de Navarra) Project. Nutr Metabol Consumption of red meat and whole-grain bread in relation to biomarkers of
Cardiovasc Dis 2016;26:1048e56. https://doi.org/10.1016/ obesity, inflammation, glucose metabolism and oxidative stress. Eur J Nutr
j.numecd.2016.07.002. 2013;52:337e45. https://doi.org/10.1007/s00394-012-0340-6.
[23] Sanchez-Tainta A, Zazpe I, Bes-Rastrollo M, Salas-Salvado J, Bullo M, Sorli JV, [42] Satija A, Yu E, Willett WC, Hu FB. Understanding nutritional epidemiology and
et al. Nutritional adequacy according to carbohydrates and fat quality. Eur J its role in policy. Adv Nutr 2015;6:5e18. https://doi.org/10.3945/
Nutr 2016;55:93e106. https://doi.org/10.1007/s00394-014-0828-3. an.114.007492.
[24] Santiago S, Zazpe I, Bes-Rastrollo M, S! anchez-Tainta A, Sayo !n-Orea C, De La [43] World Cancer Research Fund. Food, nutrition, physical activity, and the pre-
Fuente-Arrillaga C, et al. Carbohydrate quality, weight change and incident vention of cancer: a global perspective. The Third Expert Report; 2018.
obesity in a Mediterranean cohort: the SUN Project. Eur J Clin Nutr 2015;69: https://www.wcrf.org/dietandcancer. [Accessed 7 November 2019]. Accessed.
297e302. https://doi.org/10.1038/ejcn.2014.187. [44] US Department of Health and Human Services; US Department of Agriculture.
[25] Atkinson FS, Foster-Powell K, Brand-Miller JC. International tables of glycemic 2015-2020 Dietary guidelines for Americans. 8th ed. Washington, DC: US Dept
index and glycemic load values: 2008. Diabetes Care 2008;31:2281e3. of Health and Human Services; December 2015. http://www.health.gov/
https://doi.org/10.2337/dc08-1239. DietaryGuidelines. [Accessed 29 October 2019]. Accessed.
[26] Liu S, Willett WC, Stampfer MJ, Hu FB, Franz M, Sampson L, et al. A prospective
study of dietary glycemic load, carbohydrate intake, and risk of coronary heart
Title: Healthful and Unhealthful Provegetarian Food Patterns and the Incidence of
Breast Cancer: results from the SUN Project
Journal: Nutrition
Impact Factor: 3.42
Status: Published in June 2020
Nutrition 79!80 (2020) 110884
Contents lists available at ScienceDirect
Nutrition
journal homepage: www.nutritionjrnl.com
Applied nutritional investigation
Healthful and unhealthful provegetarian food patterns and the

incidence of breast cancer: Results from a Mediterranean cohort
Andrea Romanos-Nanclares M.S., B.S. a,b, Estefanía Toledo Ph.D., M.P.H., M.D. a,b,c,*,
!nchez-Bayona Ph.D., M.D. a,d, Cristina Sa
Rodrigo Sa !nchez-Quesada Ph.D., M.S., B.S. e,
!
Miguel Angel !lez Ph.D., M.P.H., M.D. a,b,c,f, Alfredo Gea Ph.D., M.S., B.Pharm. a,b,c
Martínez-Gonza
a
University of Navarra, Department of Preventive Medicine and Public Health, Pamplona, Spain
b
IdiSNA, Navarra Institute for Health Research, Pamplona, Spain
c
CIBERobn Physiopathology of Obesity and Nutrition, Institute of Health Carlos III, Madrid, Spain
d
Department of Oncology, University of Navarra, Clínica Universidad de Navarra, Pamplona, Spain
e
Department of Health Sciences, Faculty of Experimental Sciences, University of Jaen, Jaen, Spain
f
Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, Massachusetts
A R T I C L E I N F O A B S T R A C T
Article History: Objectives: Provegetarian diets, also known as predominantly plant-based (but not vegetarian or vegan) or
Received 21 January 2020 plant-forward diets, have been associated with health benefits. However, a distinction is needed between
Received in revised form 6 April 2020 high- and low-quality provegetarian dietary patterns (PVGs). We sought to examine potential associations
Accepted 14 May 2020
between PVG indices and breast cancer (BC) incidence.
Keywords: Methods: We assessed 10 812 women in the Seguimiento Universidad de Navarra cohort. We calculated an over-
Breast cancer all PVG pattern from a validated semi-quantitative food-frequency questionnaire as proposed by Martínez-Gon-
Diet zalez et al, assigning positive scores (based on quintiles) to plant foods and reversing the quintile scores for
Dietary pattern animal foods. Participants were categorized according to tertiles of the overall score. We also calculated a health-
Epidemiology ful PVG (hPVG) and unhealthful PVG (uPVG) as proposed by Satija et al.
Prospective cohort study Results: After a median of 11.5 years of follow-up, 101 incident BC cases, confirmed by medical records, were
observed. A significant inverse association with BC (comparing tertile 2 vs. tertile 1, HR= 0.55; 95% confi-
dence interval, 0.32!0.95) was identified for a modest overall PVG, but not for hPVG and uPVG separately.
Nevertheless, the highest tertile was not associated with BC.
Conclusions: In this large prospective cohort study, a moderate adherence to a PVG might decrease the risk of BC.
Further studies should replicate and expand these results to other racial, ethnic, and socioeconomic groups.
© 2020 Elsevier Inc. All rights reserved.
Introduction majority of countries (154 of 185 of Organization for Economic Coop-

eration and Development [OECD] countries), and the leading cause of
Among women, breast cancer (BC) is the most commonly diag- death in >100 countries [1]. Worldwide, there were approximately
nosed form of cancer, regardless of age and ethnicity, in the vast 2.1 million newly diagnosed cases of female BC in 2018, which
accounted for almost 1 in every 4 cases of cancer among women [1].
Similarly, in Spain, BC showed the highest risk up to 69 years of age
The Seguimiento Universidad de Navarra project received funding from the Spanish and was the fourth most frequent type of cancer overall (27 747 new
Government-Instituto de Salud Carlos III, and the European Regional Development cases) and the first among women in 2015 [2].
Fund (RD 06/0045, CIBER-OBN, Grants PI13/00615, PI10/02658, PI10/02293, PI13/
According to the World Cancer Research Fund/American Insti-
00615, PI14/01668, PI14/01798, PI14/01764, PI17/01795 and G03/140), the Navarra
Regional Government (27/2011, 45/2011, 122/2014), and the University of Navarra. tute for Cancer Research [3], about one third of the most common
ARN wrote the daft of the manuscript; ARN, AG, MAMG, and ET were responsible for neoplasms could be avoided by changing lifestyle and behavioral
data analysis; AG, MAMG and ET were responsible for project conception and design, (thus, modifiable) factors, such as obesity, physical inactivity, poor
data acquisition, and interpretation; CSQ, RSB, ET, AG and MAMG were responsible for
diet, alcohol consumption, or not breastfeeding.
review and editing. All authors have revised the manuscript critically for important
intellectual content and approved the final version to be published.
Provegetarian dietary patterns (PVGs) are also known as plant-
*Corresponding author. Tel.: +34 948425600 ext: 806224; fax: +34 948425649. based or plant-forward patterns, and do not properly represent
E-mail address: etoledo@unav.es (E. Toledo).
https://doi.org/10.1016/j.nut.2020.110884
0899-9007/© 2020 Elsevier Inc. All rights reserved.
2 A. Romanos-Nanclares et al. / Nutrition 79!80 (2020) 110884
absolute vegetarian diets. PVGs are a gentle and moderate other recommendations, such as the entire exclusion of animal
approach to vegetarianism that incorporates a range of progres- food intake. Thus, in the Prevencio ! n con Dieta Mediterra
!nea (PRE-
sively increasing proportions of plant-derived foods and concomi- DIMED) study, a PVG food pattern was found to be inversely and
tant reductions in animal-derived foods [4]. PVG diets are rich in monotonically associated with all-cause mortality [4].
fruits, vegetables, and whole grains, and are beneficial because More recently, Satija et al. [10] proposed three different ver-
they contain sufficient amounts of synergistic combinations of bio- sions of the PVG diet initially proposed by Martínez-Gonza !lez et al.
active phytochemicals and micronutrients [5]. In contrast, vegetar- [4], also using a gradual approach. The researchers proposed an
ian diets are a subset of selective plant-based diets that exclude overall plant-based diet index (PDI) equivalent to the original PVG,
the intake of some or all animal foods (e.g., meat, fish, milk, or as well as two new scores (healthful and unhealthful PDI) to over-
eggs), and vegan diets completely eliminate the consumption of all come the limitations derived from the fact that all plant foods
animal products [6]. were treated equally in the original PVG. This is an appealing
Built on a comprehensive review of these studies, the Dietary approach because little consideration has been given to the quality
Guidelines for Americans 2015!2020 [7] included a healthy vege- and types of plant foods consumed compared with the proportion
tarian-style dietary pattern compared with a typical American pat- and frequency of animal foods consumed within the plant-based
tern in its recommendations of dietary patterns that can be diet. In this sense, how progressive reductions in animal food
followed for a better health. Vegetarian or vegan diets are gaining intake with accompanying increases in consumption of plant
popularity, but the proportion of people who embrace such diets foods, with a special focus on healthful and unhealthful plant-
are limited in Western countries [8,9] and most people consume based foods, may affect BC is essential to understand. However,
both animal and plant products in different amounts. there is little evidence on the potential association between these
Instead of characterizing plant-based diets as the entire exclu- PVG dietary patterns and the risk of BC.
sion of some or all animal foods, a few recent investigations have In a large cohort of Spanish university graduates, we aimed to
assessed gradations of adherence to a PVG diet [4,6]. These grada- address whether a modified PVG based on an a priori, hypothesis-
tions could have a wider application, because gradual cutbacks in oriented PVG food pattern proposed by Martínez-Gonza !lez et al.
animal food intake may be simpler to adopt and adhere to than [4] was associated with the risk of BC, as well as assess whether
Fig. 1. Flowchart of participants recruited in the Seguimiento Universidad de Navarra (SUN) Project.
A. Romanos-Nanclares et al. / Nutrition 79!80 (2020) 110884 3
two other versions (healthful PVG [hPVG] and unhealthful PVG for hPVG and 30 to 82 for uPVG according to the data from the SUN cohort. The indi-
[uPVG]) [10] could have opposite associations on BC incidence. ces were analyzed as tertiles with energy intake adjustment at the analysis stage.
Breast cancer ascertainment

Methods
The primary end point for the present analysis was the diagnosis of BC. Partici-
Study population pants who self-reported a physician-diagnosed BC in the biennial follow-up question-
naires were contacted to confirm the diagnosis and asked for a copy of their medical
This work was conducted within the Seguimiento Universidad de Navarra records. Then, a trained oncologist confirmed the cases, and those confirmed were
(SUN) project, which is a Spanish, ongoing, dynamic, prospective, and multipur- considered for the current analysis. Only confirmed cases that met the criteria were
pose cohort study established in 1999. Participants were all university graduates included in the analysis. When the medical record to confirm the diagnosis of BC was
who were contacted biennially for follow-up reasons. Design, recruitment, and fol- not yet available, possibly owing to a delay in the delivery of the documentation, we
low-up methods have been published in detail elsewhere [11]. considered the case as a probable incident case. Fatal cases were reported to our
Our analyses included women who had been recruited and had returned the research team by subjects’ next of kin, postal authorities, or work associates. We con-
self-reported semi-quantitative food-frequency questionnaire (FFQ) by July 2018 sulted the National Death Index to identify and obtain the vital status and cause of
(n = 13 999). To ensure a minimum follow-up time of 2 years, we included in the death of cohort members in case they were lost to follow up.
present study those women recruited before October 2015 (n = 13 770). Addition-
ally, women were excluded from the study if they were lost to follow up Assessment of covariates
(n = 1295), had prevalent BC (n = 104), showed a total energy intake outside of the
predefined limits (<500 or >3500 kcal/day; n = 1353) [12], or reported menopause We collected and updated information on sociodemographic and lifestyle
before the age of 35 years (n = 206). Consequently, our final sample for the primary characteristics (e.g., sex, date of birth, years at university, or smoking status),
analysis of PVG food patterns and incidence of BC included 10 812 women (Fig. 1). anthropometrics (e.g., height and weight), dietary intakes, physical activity (vali-
The present research was performed according to the guidelines provided in dated questionnaire) [17], and health status (e.g., family history of BC, use and
the Declaration of Helsinki, and all procedures involving participants were time of hormonal therapy for menopause, or menopausal status). We estimated
approved by the institutional review board of the University of Navarra. Partici- metabolic equivalents (METs) for each participant to generate METs-h/week
pants were properly informed of their right to withdraw participation in the SUN scores. Self-reported weight and height to yield body mass index has been previ-
study or deny their consent to participate at any time without retaliation. Appro- ously validated in a subsample of this cohort [18].
priate attention was given to the specific information needs of individual candi-
dates, combined with the methods used to transfer information and assessments Statistical analysis
received in the future from the research team. The completion of the self-adminis-
tered questionnaire was considered to imply their potential freely given informed For the primary analysis, we fitted Cox proportional-hazards regression mod-
consent, and was accepted by our institutional review board. This study was regis- els to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) to evaluate
tered at clinicaltrials.gov as NCT02669602. separately the associations of tertiles of each index (PVG, uPVG, and hPVG) with
BC incidence in the overall sample of women and by menopausal status. Follow-
up was calculated for each participant from the baseline questionnaire return date
Dietary assessment and plant-based diet indices
until BC diagnosis, death, or end of follow-up.
We used age (in years) as the time scale, with stratification by time period. For all
Dietary data were collected from a 136-item semi-quantitative FFQ repeatedly
analyses, we adjusted for height (cm, continuous), family family history of BC (yes or
used and validated in Spain [13!15]. Participants were asked how often they con-
no), smoking status (never smoker, former smoker, current smoker), physical activity
sumed a defined portion of 136 food items over the previous year [13,14]. For each
(METs-h/ week, continuous), alcohol intake (g/day, continuous), body mass index
food item, there were nine possible answers (never/seldom, 1!3 servings/month,
(tertiles), age at the time of menarche (!11 years, 12!13 years, 13!14 years, "15
1 serving/week, 2!4 servings/week, 5!6 servings/week, 1 serving/day,
years), number of pregnancies of >6 months (continuous), pregnancy before the age
2!3 servings/day, 4!6 servings/day, and >6 servings/day). Food composition tables
of 30 years (yes/no), months of breastfeeding (continuous), years at university (con-
for Spain were used by a trained group of dietitians to derive nutrient intakes. Each
tinuous), total energy intake (kcal/day, continuous), time since recruitment, meno-
food consumption frequency was then converted into servings per day.
pause (no menopause, <50 years, "50 years), use of hormone replacement therapy
For the principal analysis, using the dietary data, we built an overall PVG per the
(yes/no), and its duration (continuous). The latter three covariates were not consid-
methods used by Martínez-Gonza !lez et al. [4], two fitted versions of a hPVG, and an
ered for the analysis among premenopausal women. In the analysis with postmeno-
uPVG, as proposed by Satija et al. [10]. For the present analysis, we distinguished
pausal BC as the outcome we made an additional adjustment for time between
between healthy and less healthy plant foods based on existing epidemiologic knowl-
recruitment until the beginning of the time at risk of postmenopausal BC and age at
edge of association of the foods with type 2 diabetes and cardiovascular disease
the time of menopause (<50 years, "50 years).
[6,10,16], as well as intermediate conditions such as obesity, hypertension, or inflam-
Information on age at the time of menopause was collected in the baseline
mation. The food items included in the PVG food pattern are shown in Supplementary
questionnaire and after 16 years of follow-up. If information regarding age at the
Table 1. To build the score, we adjusted the consumption (g/d) of 10 plant-origin food
time of menopause was missing, postmenopausal status was imputed according
groups (vegetables, fruits, legumes, cereals, potatoes, nuts, vegetable fat, pastries,
to the 75th percentile of age at the time of menopause in the study population
sugar-sweetened beverages (SBB), and coffee) and five animal-origin food groups
(52 years of age) [19]. When assessing premenopausal BC as the outcome, we
(meat and meat products, animal fats for cooking or as a spread, eggs, fish and other
excluded women who reported menopause before study inception and censored
seafood, and dairy products) for total energy intake with the residual method [12].
the follow-up time at the age of 52 years or at the age of menopause, whichever
The energy-adjusted estimates (residuals) were ranked in quintiles and the values
happened first. For postmenopausal BC, we considered women at risk only those
received reversed punctuations in the case of animal products (assigning a value of 5
after having turned 52 years old or after menopause, whichever happened last.
for the first quintiles, 4 for the second quintile, and consecutively until the value of 1
We conducted tests for linear trends by assigning each tertile its median value
was assigned to the fifth quintile). To obtain the final PVG food pattern, quintile values
and treating the resulting variable as continuous. We fitted restricted cubic splines
were summed, and the final score ranged theoretically from 15 to 75. A higher score
to the fully adjusted model. We used Kaplan-Meier curves to describe BC risk
showed a higher contribution of plant foods to the diet. Subsequently, we classified
according to tertiles of adherence to a hPVG or an uPVG. We also compared regres-
adherence to the PVG in tertiles for the overall sample and separately for pre- and
sion coefficients from the different models using either uPVG or hPVG as exposure
postmenopausal women.
to determine whether the association for uPVG was significantly different from
Moreover, with the use of the approaches outlined by Satija et al. [10], we devel-
the association for hPVG [20,21]. Sensitivity analyses by menopausal status and
oped 18 food groups built on nutrient and culinary similarities within the bigger cate-
different energy cutoff points were performed to explore the robustness of our
gories of healthy plant foods (vegetables, fruits, legumes, whole grains, boiled or
findings and account for potential uncertainties in our assumptions with respect
baked potatoes, nuts, vegetable oil, and coffee), less healthy plant foods (refined
to possible sources of bias or measurement errors.
grains, French fries, SSB, pastries, and other vegetable fat [margarine]), and animal
All analyses were performed using Stata 12.0 (Stata Corporation), and statisti-
foods (meat and meat products, animal fats for cooking or as a spread, eggs, fish and
cal significance was set at a two-tailed P value of < 0.05.
other seafood, and dairy products; Suppl. Table 2). These food groups were trans-
formed into energy-adjusted quintiles with the residual method. For hPVG, positive
scores were given to healthy plant food groups, and reverse scores to less healthy Results
plant food groups and animal food groups. Finally, for uPVG, positive scores were
given to less healthy plant food groups, and reverse scores to healthy plant food
Between 1999 and 2017 (median follow-up time: 11.5 years), we
groups and animal food groups. The 18 food group scores for an individual were
summed to obtain the indices, with a theoretical range from 18 (lowest possible identified 190 probable BC cases, of which 101 were confirmed by
score) to 90 (highest possible score). The observed ranges at baseline were 30 to 80 review of the medical records. The remaining cases were self-reported
Table 1
Baseline characteristics of participants according to tertiles of the provegetarian food pattern: The Seguimiento Universidad de Navarra cohort
Provegetarian food pattern Healthful provegetarian Unhealthful provegetarian
Variable T1 T2 T3 T1 T2 T3 T1 T2 T3
N 4162 3209 3441 4041 3593 3178 4089 3364 3359

Adherence score (range) 22!43 44!47 48!66 30!51 52!57 58!80 30!51 52!57 58!82
Age (years) 33.7 (10.2) 34.6 (10.5) 35.9 (11) 32.1 (9.2) 34.8 (10.4) 37.9 (11.5) 37.3 (11.2) 34.3 (10.2) 32.0 (9.3)
Body mass index (kg/m2) 22.2 (3) 22.2 (3) 22.2 (3.1) 22.0 (3.0) 22.3 (3.1) 22.4 (3.0) 22.6 (3.1) 22.2 (3.1) 21.8 (3.0)
Physical activity (METs-h/week) 19.9 (19) 20.3 (18) 20.6 (19.2) 17.7 (16.6) 20.1 (18.3) 23.4 (21.3) 22.8 (20.4) 19.7 (18.8) 17.5 (16.1)
Alcohol intake (g/d) 4.0 (6.2) 4.0 (5.8) 4.1(5.6) 3.6 (5.3) 4.0 (5.8) 4.6 (6.6) 4.1 (5.8) 4.0 (5.9) 4.0 (6.0)
Years at university 4.8 (1.3) 4.8 (1.3) 4.9 (1.4) 4.8 (1.3) 4.8 (1.4) 4.9 (1.4) 4.8 (1.3) 4.9 (1.3) 1.8 (1.3)
Height (cm) 164 (6) 164 (6) 163 (6) 164 (6) 164 (6) 163 (6) 163 (6) 164 (6.1) 164 (6)
Pregnancy before 30 years (%) 23.5 22.8 23.3 19.5 24.2 26.8 28.2 22.3 18.1
Pregnancies >6 months 33.4 33.0 33.0 29.1 33.5 37.9 39.3 31.4 27.4
Breastfeeding (months) 2.7 (5.2) 2.7 (5.3) 2.7 (5.1) 2.3 (5.1) 2.7 (5.1) 3.2 (5.4) 3.3 (5.5) 2.5 (5.0) 2.2 (5.0)
Hormone replacement therapy*
Yes 40.3 35.7 34.8 40.3 39.2 33.7 37.8 36.8 33.9
Time of hormone replacement 1.4 (2.5) 1.2 (2.2) 1.2 (2.4) 1.4 (2.5) 1.4 (2.4) 1.2 (2.3) 1.3 (2.3) 1.3 (2.3) 1.3 (2.5)
therapy (years) *
Diabetes (%) 0.94 1.34 1.05 0.7 1.2 1.5 1.7 0.8 0.7
Smoking (%)
Never smoker 53.9 52.3 48.8 54.5 51.6 48.5 49.0 51.8 55.2
Current smoker 22.2 21.9 24.2 23.9 23.2 20.8 21.1 23.4 24.2
Former smoker 23.9 25.8 26.9 21.5 25.2 30.7 30.0 24.9 20.6
cancer (%)y
Yes 10.8 10.1 10.8 10.4 10.2 11.2 10.8 10.7 10.8
Age at menarche (%)
!11 years 20.4 19.7 20.4 19.1 20.2 21.7 21.9 20.4 17.9
12!13 years 53.7 15.2 55.6 54.5 54.7 55.2 54.3 54.4 55.8
13!14 years 17.2 17.1 16.5 17.6 17.7 15.4 16.5 16.7 17.8
"16 years 8.7 8.0 7.4 8.9 7.5 7.7 7.3 8.5 8.6
Premenopausal (%) 90.6 89.2 86.0 94.0 88.8 82.0 83.8 90.0 93.5
Postmenopausal (%) 9.4 10.8 14.0 6.0 11.2 18.0 16.2 10.0 6.5
Age at menopause*
Postmenopausal <50 years (%) 52.6 53.0 52.9 52.7 54.6 51.7 51.5 56.4 51.4
Postmenopausal "50 years (%) 47.5 47.0 47.1 47.3 45.4 48.3 48.5 43.6 48.6
Adherence to Mediterranean diet 3.5 (1.6) 4.2 (1.7) 4.9 (1.7) 3.1 (1.5) 4.2 (1.5) 5.4 (1.5) 3.1 (1.5) 4.0 (1.7) 5.4 (1.5)
Energy intake (kcal/d) 2396 (539) 2256 (553) 2214 (611) 2299 (571) 2212 (578) 2389 (554) 2193 (541) 2239 (569) 2481 (570)
Carbohydrate (% of energy) 40.7 (7.3) 43.5 (6.8) 46.1 (6.7) 42.1 (6.9) 34.0 (7.2) 45.0 (7.7) 42.2 (7.3) 43.0 (7.3) 44.8 (7.1)
Protein (% of energy) 19.6 (3.4) 18.5 (3.0) 17.0 (3.0) 18.6 (3.3) 18.6 (3.5) 18.0 (3.2) 20.0 (3.2) 18.4 (3.0) 16.5 (2.9)
Total fat (% of energy) 38.6 (6.6) 36.7 (6.5) 35.6 (6.6) 38.3 (5.9) 37.0 (6.8) 35.6 (7.1) 36.5 (6.7) 37.3 (6.7) 37.6 (6.6)
*Only for postmenopausal women.
y
Information from mother, sisters, and both grandmothers was collected.
by the highly-educated participants, and the confirmation is pending model, a higher PVG was not associated with a decreased risk of BC
of the retrieval of the medical records. Only the 101 confirmed cases (HRT3 vs. T1 1.03; 95% CI, 0.67!1.59; P for trend = 0.857). The results
were used for the main analyses. At baseline, the PVG ranged from 22 were unchanged after further adjustment for potential confound-
to 43 in the lowest tertile and 48 to 66 in the highest tertile. ers (HRT3 vs. T1 1.03; 95% CI, 0.67!1.60; P for trend = 0.875). None-
The baseline characteristics of the participants according to the theless, a moderate adherence (range, 44!47) to a PVG compared
tertiles of the PVG, hPVG, and uPVG are described in Table 1. Partici- with the lowest adherence was associated with a significantly rela-
pants with a higher PVG and hPVG were more likely to be older, tive decreased risk of 45% for the overall sample of women (HRT2 vs
more physically active, have a lower energy intake, and lower fat and T1 0.55; 95% CI, 0.32!0.95).
protein (% of energy) intake. Furthermore, these participants also Furthermore, to account for a nonlinear association, we performed
had a higher consumption (g/d) of vegetables, fruits, legumes, cereals, a restricted cubic spline analysis adjusted for the same potential con-
nuts, vegetable fat, and coffee, and have a lower consumption of founding factors as the main Cox regression analyses. Among the over-
meat and meat products, animal fats, eggs, fish and other seafood, all sample of women, we found a significant nonlinear relationship
and dairy products. Nevertheless, participants with the highest between adherence to a PVG and the incidence of BC (P for nonlinear-
adherence to a PVG were more likely to eat more pastries and SSB ity < 0.001; Fig. 2). Particularly, we observed a U-shaped association
compared with those in the lowest adherence group. On the other with increased BC risk in the lower and upper ranges of the PVG. We
hand, participants with the highest adherence to an uPVG were more also repeated the multivariable analyses using the original PVG, and
likely to be younger, less physically active, and have a higher energy the results were similar (HRT3 vs. T1 1.02; 95% CI, 0.65!1.60; P for
intake. trend = 0.897 and HRT2 vs. T1 0.66; 95% CI, 0.40!1.11).
Provegetarian dietary pattern and risk of breast cancer Healthful and unhealthful provegetarian dietary patterns and breast
cancer risk
With more than 115 802 person-years of follow-up, 101
women developed a confirmed case of BC. Associations between a When we analyzed hPVG and uPVG separately (Table 2), we found
PVG and BC risk are presented in Table 2. In the age-adjusted an inverse, nonsignificant association between hPVG and BC
Table 2
Hazard ratio and 95% confidence intervals according to tertiles of the provegetarian food pattern and healthful and unhealthful provegetarian food patterns in the overall
sample of women from the Seguimiento Universidad de Navarra Project
Overall women T1 T2 T3 P for trend

n 4162 3209 3441
Adherence score (range) 22!43 44!47 48!66
Age-adjusted 1 (ref.) 0.55 (0.32!0.95) 1.03 (0.67!1.59) 0.857
Multivariable adjusted 1 (ref.) 0.55 (0.32!0.95) 1.03 (0.67!1.60) 0.875
Healthful provegetarian food pattern
n 4041 3593 3178
Age-adjusted 1 (ref.) 0.93 (0.58!1.48) 0.80 (0.49!1.32) 0.382
Unhealthful provegetarian food pattern
n 4089 3365 3359
Age-adjusted 1 (ref.) 1.44 (0.89!2.30) 1.41 (0.86!2.31) 0.161
Ref, reference
Multivariable adjusted for height, family history of breast cancer (yes or no), smoking status (never smoker, former smoker, current smoker), physical activity (METs-h/ week,
continuous), alcohol intake (g/day, continuous), body mass index (tertiles), age at the time of menarche (!11 years, 12!13 years, 13!14 years, "15 years), menopause (no
menopause, <50 years, "50 years), number of pregnancies >6 months (continuous), pregnancy before the age of 30 years (yes/no), months of breastfeeding (continuous),
use of hormone replacement therapy (yes/no) and its duration (continuous), years at university (continuous), total energy intake (kcal/day, continuous).
P value when assigned the median value to each quartile and entered as a continuous variable in the model.
Fig. 2. Restricted cubic splines for the hazard ratio (HR) and 95% confidence intervals (CIs) of adherence to the provegetarian (PVG) food pattern and incidence of breast can-
cer in the Seguimiento Universidad de Navarra (SUN) cohort. Black line represents the HR and the dashed lines represent the 95% CIs. A PVG value of 40 was the reference
value (median of tertile 1). Adjusted for height, family history of breast cancer (yes or no), smoking status (never smoker, former smoker, current smoker), physical activity
(metabolic equivalents [METs]-h/ week, continuous), alcohol intake (g/day, continuous), body mass index (tertiles), age at menarche (!11 years, 12!13 years, 13!14 years,
"15 years), menopause (no menopause, <50 years, "50 years), number of pregnancies of more than 6 months (continuous), pregnancy before the age of 30 years (yes/no),
months of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its duration (continuous), years at university (continuous), total energy intake (kcal/
day, continuous). P for non-linearity = 0.0003.
incidence in the age-adjusted model (HRT3 vs. T1 0.80; 95% CI, (Table 3), the results were nonsignificant after adjusting for multi-
0.49!1.32; P for trend = 0.382) and the multivariable adjusted model ple confounders, both for premenopausal and postmenopausal
(HRT3 vs. T1 0.83; 95% CI, 0.50!1.37; P for trend = 0.466; Table 2; women. The HR for the association between adherence to a PVG
Fig. 3). A nonsignificant higher risk of BC was also found with a higher food pattern and BC among premenopausal women was 0.89 (95%
adherence to an uPVG food pattern (HRT3 vs. T1 1.31; 95% CI, CI, 0.50!1.58 for T3 vs. T1; P for trend = 0.441) and 1.41 (95% CI,
0.79!2.19; P for trend = 0.282) in the multivariable adjusted model. 0.61!3.24 for T3 vs. T1; P for trend = 0.431) for postmenopausal
When we compared regression coefficients for the comparison women. Nonetheless, moderate adherence to a PVG compared
of the second versus first and third versus first tertiles across the with the lowest adherence seemed to be inversely associated with
models of uPVG and hPVG, we found P values of 0.279 and 0.225, BC for premenopausal women (HRT2 vs. T1 0.40; 95% CI, 0.19!0.86).
respectively. When we stratified our results for menopausal status Moreover, we found evidence of deviation from linearity for a PVG
Fig. 3. Hazard ratios (HR) and 95% confidence intervals (CI) of breast cancer incidence according to (A) a healthful and (B) an unhealthful provegetarian food pattern in the
Seguimiento Universidad de Navarra (SUN) cohort. Adjusted for height, family history of breast cancer (yes or no), smoking status (never smoker, former smoker, current
smoker), physical activity (metabolic equivalents [METs]-h/ week, continuous), alcohol intake (g/day, continuous), BMI (tertiles), age at menarche (!11 years, 12!13 years,
13!14 years, "15 years), menopause (no menopause, < 50 years, " 50 years), number of pregnancies of more than 6 months (continuous), pregnancy before the age of
30 years (yes/no), months of breastfeeding (continuous), use of hormone replacement therapy (yes/no) and its duration (continuous), years at university (continuous), total
energy intake (kcal/day, continuous).
Table 3
Hazard ratio and 95% confidence interval according to tertiles of a provegetarian food pattern and a healthful and unhealthful provegetarian food pattern in the overall sample
of women from the Seguimiento Universidad de Navarra cohort
Premenopausal women T1 T2 T3 P for trend

n 3742 2834 2927
Age-adjusted 1 (ref.) 0.42 (0.20!0.90) 0.91 (0.51!1.61) 0.483
Healthful provegetarian
n 3190 3294 3019
Age-adjusted 1 (ref.) 1.01 (0.55!1.87) 0.74 (0.37!1.45) 0.366
Unhealthful provegetarian
n 3375 3001 3127
Adherence score (range) 30-51 52-57 58-82
Cases/person years 20/30131 19/29254 18/32451
Age-adjusted 1 (ref.) 1.19 (0.63!2.25) 1.14 (0.60!2.18) 0.700
n 1058 1133 898
Age-adjusted 1 (ref.) 0.90 (0.39!2.09) 1.33 (0.59!3.03) 0.498
Multivariable adjusted1 1 (ref.) 0.96 (0.41!2.24) 1.41 (0.61!3.24) 0.431
Healthful provegetarian
N 1128 1044 917
Age-adjusted 1 (ref.) 1.10 (0.52!2.35) 0.63 (0.25!1.58) 0.347
Multivariable adjusted* 1 (ref.) 1.17 (0.54!2.54) 0.71 (0.28!1.84) 0.525
Unhealthful provegetarian
n 1150 1003 936
Age-adjusted 1 (ref.) 2.50 (1.02!6.15) 2.27 (0.89!5.82) 0.114
Multivariable adjusted* 1 (ref.) 2.57 (1.03!6.43) 2.12 (0.76!5.68) 0.177
Ref, reference
Multivariable adjusted for height, family history of breast cancer (yes or no), smoking status (never smoker, former smoker, current smoker), physical activity (METs-h/ week,
continuous), alcohol intake (g/day, continuous), body mass index (tertiles), age at the time of menarche (!11 years, 12!13 years, 13!14 years, "15 years), number of preg-
nancies >6 months (continuous), pregnancy before the age of 30 years (yes/no), months of breastfeeding (continuous), years at university (continuous), and total energy
intake (kcal/day, continuous).
*Multivariable model additionally adjusted for menopause (<50 years, "50 years), use of hormone replacement therapy (yes/no) and its duration (continuous).P value when
assigned the median value to each quartile and entered as a continuous variable in the model.
food pattern in both premenopausal and postmenopausal BC inci- observed between a plant-based dietary score and BC incidence,
dence with a P value of 0.0167 and 0.0443, respectively (Suppl. Fig. neither for premenopausal nor for postmenopausal women. In the
E1). prospective study by Martínez-Gonza !lez et al. [4], researchers
We found no significant associations between a higher adher- found a reduced risk in all-cause mortality among omnivorous par-
ence to an uPVG and BC risk among premenopausal (HRT3 vs. T1 ticipants at high cardiovascular risk with a higher adherence to a
1.09; 95% CI, 0.56!2.13; P for trend = 0.807) or postmenopausal proplant-based score. The association with cancer deaths, despite
(HRT3 vs T1 2.12; 95% CI, 0.76!5.68; P for trend = 0.177) women. In being inverse, was not statistically significant, and only 130 cancer
the latter group, a moderate adherence to an uPVG was associated deaths were reported. A lack of statistical power could have limited
with a significant increased risk of BC compared with the reference the results.
category (HRT2 vs. T1 2.57; 95% CI, 1.03!6.43). When adherence to The summary of existing evidence on the association between a
an hPVG was evaluated, inverse although nonsignificant associa- pure vegetarian diet and the risk of BC showed scant findings from
tions with BC risk were observed among premenopausal (HRT3 vs. prospective cohort studies. Thus, unstable vegetarian diets may be
T1 0.79; 95% CI, 0.40!1.57; P for trend = 0.503) and postmeno- unfavorable in terms of nutritional adequacy, and the exclusion of
pausal (HRT3 vs. T1 0.71; 95% CI, 0.28!1.84; P for trend = 0.525) key food groups from the diet, such as meat, fish, and animal!derived
women. We also performed analyses with probable incident BC foods, may likely result in nutrient deficiencies [24].
cases, but the results hardly changed (data not shown). In the sen- When adopting a modest approach, favoring a pattern com-
sitivity analyses, Supplementary Figure E2 shows that the confi- posed of both plants and some limited animal products could pos-
dence intervals were widely overlapping regardless of the energy sibly portray a beneficial impact in BC prevention. Nonetheless,
limits considered. less healthy plant foods, such as processed and ultraprocessed
foods, may have a significant impact on the overall effect of a diet
Discussion because they tend to be richer in energy, sodium, unhealthy fat,
and sugar, and poor in fiber and micronutrients. In fact, some anti-
In this large prospective cohort study, we aimed to apply the oxidants can become prooxidants when processed at high temper-
concept of provegetarian diets developed by Martinez-Gonzalez atures, frying, or grilling. The frequent consumption of SSB have
et al. [4] to BC risk. Our results showed, among middle-aged been associated with a lower satiety effect [25], general and
women, a significant nonlinear relationship between adherence to abdominal obesity [26], a major risk for postmenopausal BC [27],
a PVG and the incidence of BC. Among premenopausal or postmen- fatty liver, type 2 diabetes [28], metabolic syndrome [29], and
opausal women, we found no statistically significant linear or non- resultant pathometabolic or endocrine outcomes that are related
linear association between adherence to a PVG, hPVG, or uPVG and to BC [30,31]. SSB may also reduce the age at the time of menarche
the risk of BC. The present results support a relative reduction in [32] and increase breast density [33], which are well-known fac-
the risk of BC of 45% associated with a moderate adherence to a tors for BC development. Although the results were nonsignificant,
PVG, which may represent not an entirely vegetarian diet but a higher adherence of an unhealthy PVG food pattern could
rather a pattern emphasizing plants, some animal products, and adversely increase BC risk due to its high consumption of proc-
limiting processed sugary foods and drinks. The proposed modest essed foods (SSB, pastries, and refined cereals) and therefore its
changes toward a PVG dietary pattern seem feasible, affordable, high glycemic index and load, and low fiber, micronutrient, and
and realistic because the observed pattern is represented in a hefty higher calorie content. In humans, the consumption of highly gly-
proportion in our cohort study. cemic foods, such as SSB, bottled fruit juices, or pastries, leads to a
Overall, our findings underline the broad variation in nutri- rapid increase in the production of insulin [33!35], a downregula-
tional quality of plant foods; thus, analyzing the quality of plant tion of sex-hormone binding globulin [36], and an increase in the
foods consumed in plant-rich diets for BC prevention is critical. insulin-like growth factor-I production [37], all of which have been
Only a few prospective studies have examined the association linked to a higher cancer risk.
between healthful and unhealthful plant-based diets and chronic Following a moderate version of a PVG by avoiding processed
diseases, particularly, type 2 diabetes [10], cardiovascular disease plant-foods and incorporating fruits, vegetables, whole grains,
[6,16], and chronic kidney disease [22]. Satija et al. [6] observed healthy fats, and small amounts of animal foods could enhance gly-
varying cardiometabolic effects of these diet indices, stressing the cemic control, improve lipid profile, reduce blood pressure,
importance of defining plant-based diets in terms of the quality of improve gut microbial profile and insulin sensitivity, and decrease
plant foods. To date, the authors have not explored the association chronic inflammation [6]. Such a diet may modulate BC risk by
between cancer risk and these scores. lowering inflammation signals, circulating concentrations of estro-
To our knowledge, the association between healthful and gens [38], or increasing sex-hormone binding globulin levels [39].
unhealthful PVG food patterns and BC incidence have never been They also contain a wide range of secondary metabolites essential
explored before, so a direct comparison between our results and for plant physiology, such as polyphenols that can scavenge and
those of other epidemiologic studies was not feasible. Nonetheless, neutralize free oxygen and nitrogen species [40], reduce oxidative
authors from the NutriNet-sante ! cohort [23] investigated the pro- stress, postprandial glucose, and insulin responses, leading to bet-
spective association between a plant-based dietary score, which ter glycemic control [41]. In fact, women with higher circulating
was built according to the algorithm proposed by Martínez- levels of alpha-carotene, beta-carotene, lutein + zeoxanthin, lyco-
Gonza !lez et al. [4] and overall cancer risk, although they did not pene, and total carotenoids may show a lower risk of BC [42,43].
further analyze the effect of healthful or unhealthful plant-based Our research adds to the evidence base by highlighting that a
diets. The researchers found an overall decreased cancer incidence modest overall PVG diet is associated with a lower risk of BC,
among participants with a higher plant-based dietary score stressing the importance of the quality of the plant foods con-
(HRT3 vs. T1 0.85; 95% CI, 0.76!0.97; P for trend = 0.02). When par- sumed and the balance with animal food sources. From a public
ticular tumors were assessed, digestive and lung cancer were also health point of view, dietary recommendations focused on the pro-
found to be decreased among participants with a high consump- motion of moderate adherence to a PVG together with the promo-
tion of plant products along with low intakes of animal products. tion of eating a variety of minimally processed and, ideally, local,
Consistent with our results, no particular linear association was plant-based foods may also have less environmental impact,
because plant-based food systems use less resources than food sys- Conclusions
tems that are densely reliant on animal foods [44,45].
The current study has some limitations. First, we are aware that The results of this study highlight that a moderate PVG adher-
the number of incident BC cases is limited, which may compromise ence among women was inversely associated with the risk of BC.
our statistical power and lead to excessively wide CIs. Neverthe- Nevertheless, the present study should be replicated and expanded
less, age-standardized BC incidence is compatible with the esti- to other racial, ethnic, and socioeconomic groups to better under-
mated age-standardized BC incidence for the Spanish population stand the underlying mechanisms that may be involved in the
[2]. Moreover, we cannot rule out chance as an explanation for the potential associations between healthful or unhealthful plant-
overall result. However, many of our results, despite not being sta- based diet indices and BC.
tistically significant, can be integrated in future meta-analyses and
contribute to the evidence on this interesting topic. Also, our popu- Acknowledgments
lation is largely composed of young women, which diminishes the
number of incident BC cases. The authors thank the implication and collaboration of the partici-
Second, the use of self-reported dietary intake was used to pants in the SUN Project. ARN was supported by the Fundacio ! n Cien-
define the exposure of the analysis. However, of note, the FFQ has tífica Asociacio ! n Espan
~ ola Contra el Ca!ncer (Scientific Foundation of
been repeatedly validated [13!15,17] and the inclusion of univer- the Spanish Association Against Cancer). The authors also thank the
sity graduates may have contributed to an increased accuracy of other members of the SUN Group: Alonso A, Alvarez- ! !
Alvarez I, Bala-
self-reported information. Notwithstanding, there could be some guer A, Barbagallo M, Barrientos I, Barrio-Lo ! pez MT, Basterra-Gortari
degree of nondifferential misclassification in the dietary assess- FJ, Battezzati A, Bazal P, Benito S, Bertoli S, Bes-Rastrollo M, Beulen Y,
ment, which might have biased our results towards the null. In Beunza JJ, Buil-Cosiales P, Canales M, Carlos S, Carmona L, Cervantes S,
addition, there may be measurement errors as a result of using dif- Cristobo C, de Irala J, de la Fuente-Arrillaga C, de la O V, de la Rosa PA,
ferent food composition databases, which may not be complete for Delgado-Rodríguez M, Díaz-Gutie !rrez J, Díez Espino J, Domínguez L,
the whole range of consumed foods. Third, the distinction between Donat-Vargas C, Donazar M, Eguaras S, Ferna !ndez-Montero A, Fresa !n
healthy and less healthy plant foods is based on existing knowl- U, Galbete C, García-Arellano A, García Lo ! pez M, Gutie !rrez-Bedmar M,
edge of the association of foods with type 2 diabetes and cardiovas- Gome !z-Domingos AL, Go !mez-Donoso C, Go ! mez-Gracia E, Gon ~ i E,
cular disease [7,10,16], as well as intermediate conditions such as Gon ~ i L, Guille
!n F, Henríquez P, Herna !ndez A, Hershey MS, Hidalgo-
obesity, hypertension, or inflammation and not specifically with Santamaría M, Hu E, Lahortiga F, Leone A, Llorca J, Lo !pez del Burgo C,
BC. Nevertheless, the present methods have been previously Marí A, Marques I, Martí A, Martín Calvo N, Martín-Moreno JM, Martí-
explored in remarkable and solid cohort studies with longer fol- nez JA, Martínez-Lapiscina EH, Mendonça R, Mene !ndez C, Molendijk
low-up periods and larger numbers of participants and disease M, Molero P, Murphy K, Mun ~ oz M, Nu !n~ ez-Co !rdoba JM, Pajares R,
cases. On the other hand, our cohort study recently showed a bet- Papadaki A, Parletta N, Pe !rez de Ciriza P, Pe !rez Cornago A, Pe !rez de
ter conformity with a healthy provegetarian diet associated with a Rojas J, Pimenta AM, Pons J, Ramallal R, Razquin C, Rico A, Ruano C,
reduced long-term risk of being overweight or obese [46]. Ruiz L, Ruiz-Canela M, Ruiz Zambrana A, Salgado E, San Juli!an B,
Fourth, information about cases of BC was also self-reported; S!anchez D, Sa !nchez-Tainta A, Sa !nchez-Villegas A, Santiago S, Sayo ! n-
nevertheless, the main results of the analyses were based on con- Orea C, Schlatter J, Serrano-Martinez M, Toledo J, Tortosa A, Valencia F,
firmed BC cases to avoid false positives. Fifth, our highly educated V!azquez Z, Zarnowiecki D, Zazpe I.
participants are not representative of the general Spanish popula- The authors especially thank all participants in the SUN cohort
tion, which could limit the external validity of our results. None- for their long-standing and enthusiastic collaboration, as well as
theless, our choice to select these highly educated participants the advisors from Harvard TH Chan School of Public Health Walter
coincides with the approach known as restriction in epidemiology, Willett, Alberto Ascherio, Frank B. Hu, and Meir J. Stampfer who
which helps prevent or at least reduce confounding by known fac- helped design the SUN Project, the PREDIMED study, and the PRE-
tors, such as educational level and socioeconomic status. They DIMED-PLUS ongoing trial.
might also provide better quality self-reported information,
improve the internal validity of our study, and improve the reten-
tion in the cohort. Sixth, we evaluated these food patterns at a par- Supplementary materials
ticular timepoint, making fully capturing the relationship of a
time-varying exposure, such as diet, on the development of BC dif- Supplementary material associated with this article can be
ficult given the reasonably long induction period. Finally, the PVG, found in the online version at doi:10.1016/j.nut.2020.110884.
uPVG, and hPVG included dairy products, fish, seafood, and poul-
tries, assuming they have negative effects on BC risk, even though References
there was no convincing evidence for these effects [27]. Therefore,
we acknowledge that the truthfulness of an a priori index approach [1] Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer
statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide
is limited by the current level of dietary knowledge concerning the for 36 cancers in 185 countries. CA Cancer J Clin 2018;68:394–424.
diet!BC relationship, as well as uncertainties accompanying the [2] Galceran J, Ameijide A, Carulla M, Mateos A, Quiro ! s JR, Rojas D, et al. Cancer
index development. The aim behind the PVG, hPVG, and uPVG con- incidence in Spain, 2015. Clin Transl Oncol 2017;19:799–825.
[3] World Cancer Research Fund/American Institute for Cancer Research. Recom-
struction was not to get the best predictive scores for BC but to cor-
mendations and public health and policy implications; 2018. Available at:
rectly test their association with BC risk. https://www.wcrf.org/sites/default/files/Recommendations.pdf. Accessed 21
Our study has some important strengths, such as its dynamic January 2020.
[4] Martínez-Gonza !lez MA, Sa
!nchez-Tainta A, Corella D, Salas-Salvado ! J, Ros E,
and prospective design (avoiding reverse causation bias), exhaus-
Aro! s F, et al. A provegetarian food pattern and reduction in total mortality in
tive data collection, and data analysis. Other strengths are its large the Prevencio !n con Dieta Mediterr!anea (PREDIMED) study. Am J Clin Nutr
sample size, high retention, long follow-up period, and the wide 2014;100:320S–8.
array of potential confounders included in the multivariable analy- [5] Liu RH. Health benefits of fruit and vegetables are from additive and synergis-
tic combinations of phytochemicals. Am J Clin Nutr 2003;78:517S–20.
ses, reducing the room for residual confounding in our results. [6] Satija A, Hu FB. Plant-based diets and cardiovascular health. Trends Cardiovasc
Med 2018;28:437–41.
[7] U.S. Department of Health and Human Services and U.S. Department of Agri- overweight and obesity: The University of Navarra Follow-Up (SUN) cohort
culture. 2015!2020 Dietary guidelines for Americans. 8th Edition. Available study. Am J Clin Nutr 2016;104:1433–40.
at: http://health.gov/dietaryguidelines/2015/guidelines/. Accessed 21 Janurary [27] World Cancer Research Fund. American Institute for Cancer Research. Diet
2020. Nutr Phys Activity Breast Cancer 2017:117.
[8] Leitzmann C. Vegetarian nutrition: Past, present, future. Am J Clin Nutr [28] Malik VS, Popkin BM, Bray GA, Despres JP, Willett WC, Hu FB. Sugar-sweetened
2014;100:496S–502. beverages and risk of metabolic syndrome and type 2 diabetes: A meta-analy-
[9] Alle
"s B, Baudry J, Me !jean C, Touvier M, Pe !neau S, Hercberg S, et al. Comparison sis. Diabetes Care 2010;33:2477–81.
of sociodemographic and nutritional characteristics between self-reported [29] Barrio-Lopez MT, Martinez-Gonzalez Ma, Fernandez-Montero A, Beunza JJ,
vegetarians, vegans, and meat-eaters from the nutrinet-sante ! study. Nutrients Zazpe I, Bes-Rastrollo M. Prospective study of changes in sugar-sweetened
2017;9:1023. beverage consumption and the incidence of the metabolic syndrome and its
[10] Satija A, Bhupathiraju SN, Rimm EB, Spiegelman D, Chiuve SE, Borgi L, et al. components: The SUN cohort. Br J Nutr 2013;110:1722–31.
Plant-based dietary patterns and incidence of type 2 siabetes in U.S. men and [30] Romanos-Nanclares A, Toledo E, Gardeazabal I, Jime !nez-Moleo! n JJ, Martínez-
women: Results from three prospective cohort studies. PLoS Med 2016;13: Gonz!alez MA, Gea A. Sugar-sweetened beverage consumption and incidence
e1002039. of breast cancer: The Seguimiento Universidad de Navarra (SUN) Project. Eur J
[11] Carlos S, De La Fuente-Arrillaga C, Bes-Rastrollo M, Razquin C, Rico-Camp"a A, Nutr 2018;48:70–225.
Martínez-Gonza !lez M, et al. Mediterranean diet and health outcomes in the [31] Chandran U, McCann SE, Zirpoli G, Gong Z, Lin Y, Hong CC, et al. Intake of
SUN cohort. Nutrients 2018;10:439. energy-dense foods, fast foods, sugary drinks, and breast cancer risk in African
[12] Willett W. Nutritional epidemiology. 3rd. ed. New York: Oxford University American and European American women. Nutr Cancer 2014;66:1187–99.
Press; 2012. [32] Carwile JL, Willett WC, Spiegelman D, Hertzmark E, Rich-Edwards J, Frazier aL,
[13] de la Fuente-Arrillaga C, V! azquez Ruiz Z, Bes-Rastrollo M, Sampson L, Marti- et al. Sugar-sweetened beverage consumption and age at menarche in a pro-
nez-Gonza !lez MA. Reproducibility of an FFQ validated in Spain. Public Health spective study of U.S. girls. Hum Reprod 2015;30:675–83.
Nutr 2010;13:1364–72. [33] Duchaine CS, Dumas I, Diorio C. Consumption of sweet foods and mammographic
[14] Martin-Moreno JM, Boyle P, Gorgojo L, Maisonneuve P, Fernandez-Rodriguez breast density: A cross-sectional study. BMC Public Health 2014;14:554.
JC, Salvini S, et al. Development and validation of a food frequency question- [34] Liu H, Heaney AP. Refined fructose and cancer. Expert Opin Ther Targets
naire in Spain. Int J Epidemiol 1993;22:512–9. 2011;15:1049–59.
[15] Fern! andez-Ballart JD, Pin~ ol JL, Zazpe I, Corella D, Carrasco P, Toledo E, et al. Rel- [35] Ludwig DS. Diet and development of the insulin resistance syndrome. Asia Pac
ative validity of a semi-quantitative food-frequency questionnaire in an J Clin Nutr 2003;12:S4.
elderly Mediterranean population of Spain. Br J Nutr 2010;103:1808–16. [36] Dimou NL, Papadimitriou N, Gill D, Christakoudi S, Murphy N, Gunter MJ, et al.
[16] Satija A, Bhupathiraju SN, Spiegelman D, Chiuve SE, Manson JAE, Willett W, Sex hormone binding globulin and risk of breast cancer: A Mendelian random-
et al. Healthful and unhealthful plant-based diets and the risk of coronary ization study. Int J Epidemiol 2019;48:807–16.
heart disease in U.S. adults. J Am Coll Cardiol 2017;70:411–22. [37] Kaaks R, Johnson T, Tikk K, Sookthai D, Tjønneland A, Roswall N, et al. Insulin-like
[17] Martínez-Gonza !lez MA, Lo ! pez-Fontana C, Varo JJ, Sa !nchez-Villegas A, Martinez growth factor I and risk of breast cancer by age and hormone receptor status - A
JA. Validation of the Spanish version of the physical activity questionnaire prospective study within the EPIC cohort. Int J Cancer 2014;134:2683–90.
used in the Nurses’ Health Study and the Health Professionals’ Follow-up [38] Carruba G, Granata OM, Pala V, Campisi I, Agostara B, Cusimano R, et al. A tradi-
Study. Public Health Nutr 2005;8:920–7. tional Mediterranean diet decreases endogenous estrogens in healthy post-
[18] Bes-Rastrollo M, Pe !rez Valdivieso JR, S! anchez-Villegas A, Alonso A,! Martínez- menopausal women. Nutr Cancer 2006;56:253–9.
Gonz! alez MA.! Validacio! n del peso e índice de masa corporal auto-declarados [39] Wu AH, Yu MC, Tseng CC, Stanczyk FZ, Pike MC. Dietary patterns and breast
de los participantes de una cohorte de graduados universitarios. Rev Esp Obes cancer risk in Asian American women. Am J Clin Nutr 2009;89:1145–54.
2005;3:352–8. [40] Pandey KB, Rizvi SI. Plant polyphenols as dietary antioxidants in human health
[19] Shivappa N, Sandin S, Lof M, Hebert JR, Adami H-O, Weiderpass E. Prospective and disease. Oxid Med Cell Longev 2009;2:270–8.
study of dietary inflammatory index and risk of breast cancer in Swedish [41] Chandalia M, Garg A, Lutjohann D, von Bergmann K, Grundy SM, Brinkley LJ.
women. Br J Cancer 2015;113:1099–103. Beneficial effects of high dietary fiber intake in patients with type 2 diabetes
[20] Clogg C.C., Petkova E., Haritou A.Statistical methods for comparing regression mellitus. N Engl J Med 2000;342:1392–8.
coefficients between models1995;100:1261!93. [42] Aune D, Chan DSM, Vieira AR, Navarro Rosenblatt DA, Vieira R, Greenwood DC,
[21] Weesie J. Seemingly unrelated estimation and the cluster-adjusted sandwich et al. Fruits, vegetables and breast cancer risk: A systematic review and meta-
estimator. Stata Tech Bulitin 1999. analysis of prospective studies. Breast Cancer Res Treat 2012;134:479–93.
[22] Kim H, Caulfield LE, Garcia-Larsen V, Steffen LM, Grams ME, Coresh J, et al. [43] Eliassen AH, Hendrickson SJ, Brinton LA, Buring JE, Campos H, Dai Q, et al. Cir-
Plant-based diets and incident CKD and kidney function. Clin J Am Soc Nephrol culating carotenoids and risk of breast cancer: Pooled analysis of 8 prospective
2019;14:682–91. studies. J Natl Cancer Inst 2012;104:1905–16.
[23] Kane!Diallo A, Srour B, Sellem L, Deschasaux M, Latino!Martel P, Hercberg S, [44] Pimentel D, Pimentel M. Sustainability of meat-based and plant-based diets
et al. Association between a pro plant!based dietary score and cancer risk in and the environment. Am J Clin Nutr 2003;78:660S–3.
the prospective NutriNet!sante ! cohort. Int J Cancer 2018;143:2168–76. [45] Willett W, Rockstro € m J, Loken B, Springmann M, Lang T, Vermeulen S, et al.
[24] Godos J, Bella F, Sciacca S, Galvano F, Grosso G. Vegetarianism and breast, colo- Food in the anthropocene: The EAT!Lancet Commission on healthy diets from
rectal and prostate cancer risk: An overview and meta-analysis of cohort stud- sustainable food systems. Lancet 2019;6736:3–49.
ies. J Hum Nutr Diet 2017;30:349–59. [46] Go! mez-Donoso C, Martínez-Gonza ! Martínez JA, Gea A, Sanz-Serrano J,
!lez MA,
[25] Fardet A. Minimally processed foods are more satiating and less hyperglycemic Perez-Cueto FJA, et al. A provegetarian food pattern emphasizing preference
than ultra-processed foods: A preliminary study with 98 ready-to-eat foods. for healthy plant-derived foods reduces the risk of overweight/obesity in the
Food Funct 2016;7:2338–46. SUN cohort. Nutrients 2019;11:1553.
[26] De Deus Mendonça R, Pimenta AM, Gea A, De La Fuente-Arrillaga C, Martinez-
Gonzalez MA, Lopes ACS, et al. Ultraprocessed food consumption and risk of
Title: Healthful and unhealthful plant-based diets and risk of breast cancer in U.S.
women: results from the Nurses' Health Studies
Journal: Cancer Epidemiology, Biomarkers and Prevention
Submitted: March 18, 2021
Status: Under Review
Healthful and unhealthful plant-based diets and risk of breast cancer in U.S. women: results from the
Nurses’ Health Studies
Authors: Andrea Romanos-Nanclares, MS, BS 1,2, Walter C Willett, MD, DrPH 3,4,5, Bernard Rosner, PhD 3,6,
Laura C Collins, MD 7, Frank B Hu, MD, PhD 3,4,5, Estefania Toledo, MD, PhD 1,2,8, A. Heather Eliassen, ScD 3,5
1
Department of Preventive Medicine and Public Health, School of Medicine, University of Navarra, Pamplona,
Spain
2
IdiSNA, Navarra Institute for Health Research, Pamplona, Spain
3
Channing Division of Network Medicine, Department of Medicine, Brigham & Women’s Hospital, and Harvard
Medical School, Boston, MA 02115
4
Department of Nutrition, Harvard T. H. Chan School of Public Health, Boston, MA 02115
5
Department of Epidemiology, Harvard T. H. Chan School of Public Health, Boston, MA 02115
6
Department of Biostatistics, Harvard T. H. Chan School of Public Health, Boston, MA 02115
7
Department of Pathology, Beth Israel Deaconess Medical Center and Harvard Medical School, 330 Brookline
Ave., Boston, MA 02215, USA
8
Centro de Investigacion Biomedica en Red Fisiopatologia de La Obesidad y La Nutrición (CIBEROBN),
Institute of Health Carlos III, Madrid, Spain
Corresponding author: Andrea Romanos-Nanclares. Department of Preventive Medicine and Public Health,
University of Navarra, Research Building, 2nd Floor (office nº 2480); 1 Irunlarrea Street, Pamplona, Navarre,
31008; Spain; Phone number: +34 948 42 56 00 (806635); E-mail: nharo@channing.harvard.edu
The authors declare no potential conflicts of interest.

ABSTRACT
Background
Plant-based diets have been associated with lower risk of various diseases, including type 2 diabetes,
cardiovascular disease and other cardiometabolic risk factors. However, the association between plant-based
diet quality and breast cancer remains unclear.
Methods
We prospectively followed 76,690 women from the Nurses' Health Study (NHS,1984-2016) and 93,295 women
from the Nurses' Health Study II (NHSII,1991-2017). Adherence to an overall plant-based diet index (PDI), a
healthful PDI (hPDI) and an unhealthful PDI (uPDI) was assessed using previously developed indices. Cox
proportional hazards models were used to estimate HR and 95% CIs for incident invasive breast cancer.
Results
Over 4,841,083 person-years of follow-up, we documented 12,482 incident invasive breast cancer cases.
Women with greater adherence to PDI and hPDI were at modestly lower risk of breast cancer [(HR Q5vsQ1 0.89;
95% CI 0.84, 0.95); (HR Q5vsQ1 0.89; 95% CI 0.83, 0.94)]. We observed significant heterogeneity by estrogen
receptor (ER) status, with the strongest inverse association between hPDI and breast cancer observed with ER-
negative tumors [HR Q5vsQ1 0.77; 95% CI 0.65, 0.90; p-trend=<0.01]. When we evaluated the independent
associations of the 3 food categories that constituted the diet indices with breast cancer risk, we found an
inverse association between extreme quintiles of healthy plant foods and ER-negative breast cancer [HR 0.74;
95% CI 0.61, 0.88; p-trend=<0.01].
Conclusions
This study provides evidence that adherence to a healthful plant-based diet may reduce the risk of breast
cancer, especially those that are more likely to be aggressive tumors.
INTRODUCTION
Plant-based diets have been associated with lower risk of chronic diseases such as obesity, type 2
1–6
diabetes, cardiovascular disease and some cancers and are recommended for both health and
7
environmental benefits . While vegetarian or vegan diets are gaining popularity, the proportion of people that
8,9
embrace such diets is limited in Western countries and most people consume both animal and plant products
in different amounts. Instead of characterizing plant-based diets as the entire exclusion of some or all animal
10,11
foods, a few recent investigations have assessed gradations of adherence to a plant-based dietary pattern .
They could have a wider application, as gradual cutbacks in animal food intake may be simpler to adopt and
adhere to than other recommendations, such as the entire exclusion of animal food intake.
Recently, Satija et al.12 proposed three different approaches of plant-based dietary indices. An overall
10
plant-based diet index (PDI) was created, equivalent to the original provegetarian dietary pattern , and two
additional scores, a healthful PDI (hPDI), and an unhealthful PDI (uPDI). The hPDI and uPDI overcome the
limitations derived from the fact that all plant foods were treated equally in the original provegetarian diet, but the
nutritional quality is not equivalent across all plant foods.
The association between plant-based diet quality and breast cancer remains unclear 6,13. In the NutriNet-
13
Santé study , higher intakes of plant-based products along with lower intakes of animal foods was associated
with a lower cancer risk, though no association was found for breast cancer. Moreover, in the SUN project 6, a
moderate, but not greater, adherence to a provegetarian dietary pattern, was associated with a decreased risk of
breast cancer. No further associations were found when the authors distinguished between healthful and
unhealthful provegetarian food patterns.
In this context, the objective of the present prospective study was to examine the associations between
plant-based diet indices and breast cancer incidence characterized by hormone receptor status and molecular
subtypes in the NHS and NHSII.
METHODS
Study population
The Nurses’ Health Study (NHS) is an ongoing study of 121,700 female nurses aged 30-55 years in
1976, and NHSII has followed 116,429 female nurses (aged 25-42 years) since 1989. Every two years,
participants have provided information on health-related factors and medical history. Women were followed from
1984 in the NHS and from 1991 in the NHSII to 2016 in NHS and to 2017 in NHSII. We excluded women who
died prior to baseline, had prevalent cancer, had missing dietary information, or reported implausible total energy
intake (<600 or >3500 kcal/day); leaving 76,690 women from NHS, and 93,295 from NHSII. The study protocol
was approved by the institutional review boards of the Brigham and Women’s Hospital and Harvard T.H. Chan
School of Public Health, and those of participating registries as required.
Dietary assessment and Plant-Based Diet indices

Diet was assessed with food-frequency questionnaires (FFQs) administered in the NHS in 1980, 1984,
1986 and every 4 years thereafter, and in the NHSII in 1991 and every 4 years thereafter. The number of FFQ
food items have evolved: in NHS, there were 61 items in 1980, 116 items in 1984 and 1986 and 130+ items
thereafter; in NHSII, the FFQ from 1991 had 130+ items. Given the complexity and broad array of foods included
in our plant-based dietary patterns, we used the 1984 FFQ as baseline questionnaire. FFQs include foods with a
portion size, and participants were asked to specify the food-specific average consumption during the previous
year (9 choices from “almost never” to “>6/day”). Participants’ nutrient intakes were calculated by multiplying the
nutrient content of a food serving (based on updated USDA databases and other sources) and consumption
frequency.
As described in a prior publication 12, we derived three versions of a plant-based diet using the FFQ data:
an overall PDI, a hPDI, and an uPDI. Briefly, we included 18 food groups based on nutrient and culinary
similarities within three broad categories: healthy plant foods, less healthy plant foods, and animal foods.
Healthy plant food groups included whole grains, fruits, vegetables, nuts, legumes, vegetable oils, and
tea/coffee; less healthy plant food groups included fruit juices, sugar-sweetened beverages, refined grains,
potatoes, and sweets; and animal food groups included animal fats, dairy, eggs, fish/seafood, meat, and
miscellaneous animal-based foods. Healthy and less healthy plant foods were identified using current knowledge
of associations of the foods with type 2 diabetes, cardiovascular disease, some cancers, and intermediate
conditions (i.e., obesity, hypertension or inflammation). The 18 food groups were ranked into quintiles and given
positive or reverse scores. With positive scores, participants in the highest quintile of a food group received a
score of 5, following on through to participants in the lowest quintile who received a score of 1. With reverse
scores, this pattern of scoring was opposite, with a score of 5 for the low quintile. For the overall PDI, all plant
food groups were given positive scores, while all animal food groups were given reverse scores. For the hPDI,
positive scores were given to healthy plant food groups, and reverse scores to less healthy plant food groups
and animal food groups. Finally, for uPDI the opposite pattern of scoring was applied. The 18 food group scores
were summed to obtain the indices, ranging from 18 to 90. Pearson correlation coefficients between hPDI and
uPDI were -0.36 and -0.33 for NHS and NHSII, respectively12. The PDI showed low correlations with hPDI
(r=0.21 in NHS and r=0.26 in NHSII) and with uPDI (r=-0.11 in NHS and r=-0.21 in NHSII)12.
Given that alcoholic beverages have different directions of association for various health outcomes, and
margarine’s fatty acid composition has changed over time from high trans to high unsaturated fats, we did not
include these foods in the indices; we adjusted for alcohol in the main analysis.
Assessment of breast cancer

We first identified incident breast cancer cases through self-report on the biennial questionnaires. We
requested permission from women reporting breast cancer to review hospital records and pathology reports for
diagnosis confirmation and ascertainment of invasive vs. in situ, and estrogen receptor (ER), progesterone
receptor (PR) and human epidermal growth factor receptor 2 (HER2) status. Given the high confirmation rate of
reported breast cancer cases in the NHS and NHSII (>99%), we included both breast cancer cases confirmed
via medical record review and self-reported cases confirmed by the nurse but lacking medical record. For
deceased cases, the next of kin was contacted for this permission; deaths were reported by family members or
by the postal service in response to follow-up questionnaires, or they were identified through the National Death
Index.
Tissue Microarrays (TMAs), Immunohistochemical Analysis and Subtype Classification
We have already reported details of breast cancer tissue block collection and TMA construction1. Briefly,
we collected archived formalin-fixed paraffin-embedded breast cancer blocks from participants with incident
breast cancer diagnosed up through 2006. For molecular subtype classification, immunohistochemical staining
information was available for the markers of ER, PR, HER2, cytokeratins 5/6 and epidermal growth factor
receptor. Further staining for the proliferative marker Ki-67 was achieved in NHS cases; Ki-67 data were not
available for NHSII cases. Cases with TMAs were very similar to all suitable invasive cases in terms of
demographics, breast cancer risk factors and tumor characteristics.
Definitions that correlated with gene expression profile classifications were used for tumor molecular
2–7
subtyping for a subgroup of cases . If Ki-67 expression data was missing (NHSII tumors), histological grade
was used instead. Hence, luminal A tumors were ER-positive and/or PR-positive, HER2-negative, and Ki-67-
negative (or histologic grade 1 or 2). Luminal B tumors were either (a) ER-positive and/or PR-positive and
HER2-positive or (b) ER-positive and/or PR-positive, HER2-negative, and Ki-67-positive (or histologic grade 3).
HER2-enriched tumors were ER-negative, PR-negative and HER2-positive. Basal-like tumors were ER-negative,
PR-negative, HER2-negative and CK 5/6-positive and/or EGFR-positive. For evaluating ER-positive vs. ER-
negative tumors, ER status was determined primarily from TMA slides, and if not available, secondarily from
pathology reports.
Statistical analysis
Data from the NHS and NHSII were pooled. Person-time for each participant was calculated from the
date of return of the baseline questionnaire until the date of breast cancer diagnosis, other cancers (excluding
non-melanoma skin cancers), death, or the end of follow-up (2016 (NHS) or 2017 (NHSII) for the main analysis
and 2006 for molecular subtype analysis), whichever occurred first.
Indices were cumulatively averaged over follow-up to better capture long-term diet. We evaluated the
association between PDI, hPDI and uPDI quintiles and incident breast cancer using multivariable-adjusted time-
varying Cox proportional hazards regression models, stratified by age, 2-year time-period at risk and cohort. We
tested for linear trends by evaluating the quintile median values as a continuous variable.
Covariates included race, socioeconomic status, age at menarche, age at menopause, postmenopausal
hormone use, oral contraceptive use, parity, age at first birth, breastfeeding history, height, alcohol intake, total
caloric intake, physical activity, and BMI at age 18 years. As change in weight from age 18, total carotenoids
intake and dietary fiber may be intermediates between diet and breast cancer, we additionally adjusted for them
in separate models.
To take advantage of repeated diet assessments in these cohorts and evaluate the latency between
these indices and breast cancer incidence, we conducted latency analyses, whereby we created different
regression models based on dietary data collected at distinct time points. In the simple update model, index
scores reported on the most recent FFQ before each follow-up interval was used; in the latency models, we
used index scores reported at different latencies (i.e., 4–8, 8–12, 12–16, 16-20 y) before breast cancer diagnosis
14
. Furthermore, we added interaction terms and used the Wald test to test whether the indices and breast
15
cancer association differed by menopausal status, current BMI and physical activity. Mediation analyses were
performed in order to assess the extent to which associations may be potentially mediated by weight gain from
age 18 years and estimated the mediation proportion (the proportion of the observed association attributable to a
mediator)16,17. To evaluate whether associations differed by molecular subtype or ER status, we used the Lunn-
18
McNeil approach to derive the p for heterogeneity . Statistical tests were two-sided with P-values<0.05
indicating statistical significance. All analyses were performed using SAS for UNIX version 9.4 (Cary, NC).
Adjustments were not made for multiple comparisons as our analysis was guided by strong, biologically driven a
priori hypotheses and we interpreted the results considering biological plausibility, coherence and consistency.
RESULTS
During 4,841,083 person-years of follow-up, 12,482 participants developed invasive breast cancer (8,220
cases in NHS and 4,262 cases in NHSII). Women with higher PDI were more likely to have lower BMI, lower
weight change from age 18, lower prevalence of diabetes, less alcohol intake, higher physical activity and
energy intake, lower protein intake (% of energy) and higher carbohydrate intake (% of energy) (Table 1).
In multivariable-adjusted models (Table 2), a higher adherence to a PDI was significantly inversely
associated with breast cancer (HR Q5 vs. Q1 0.89; 95% CI 0.84, 0.95; p-trend=<0.01). Further adjustment for
weight change since age 18 (data not shown) slightly attenuated the results (HR Q5 vs Q10.93; 95% CI, 0.87,
0.99; p-trend =0.01). When we analyzed hPDI and uPDI separately, we found a modest inverse association
between a hPDI and breast cancer incidence (HR Q5vs.Q1 0.89;95% CI 0.83, 0.94; p-trend=<0.01). Only a modest
attenuation with addition of weight change (HR Q5 vs Q1 0.91; 95% CI, 0.86, 0.97; p-trend=<0.01) or total
carotenoid intake (HR Q5 vs Q1 0.91; 95% CI, 0.85, 0.98; p-trend=<0.01) was identified for the hPDI. Further
adjustment for dietary fiber, but not weight change or total carotenoid intake, did not further changed the
scenario (HR Q5 vs Q1 0.92; 95% CI, 0.85,0.99; p-trend=0.01).
We observed a significant heterogeneity by ER status in the hPDI (p-heterogeneity=0.01) and the uPDI
(p-heterogeneity=0.01) (Table 3). We detected an inverse association between a higher hPDI and ER-negative
breast cancer (HR Q5vsQ1 0.77; 95%CI, 0.65, 0.90; p-trend=<001) and a positive association with uPDI (HR Q5vsQ1
1.28; 95%CI, 1.08, 1.51; p-trend=<0.01). This association was further observed when we performed
multivariable spline analysis (Figure 1). Further adjustment for dietary fiber (HRQ5vsQ1 0.83; 95% CI, 0.68, 1.00;
p-trend=0.02), total carotenoid intake (HRQ5vsQ1 0.82; 95% CI, 0.69, 0.98; p-trend=<0.01) or mutual adjustment
for fruits and vegetables (HRQ5vsQ1 0.85; 95% CI, 0.70, 1.03; p-trend=0.03), slightly attenuated the association
(data not shown). For the uPDI, adjustment for dietary fiber or total carotenoid intake attenuated the association
and made it not significant. Although we observed no significant heterogeneity by molecular subtypes (Table 3),
each 10-unit increase in the hPDI was associated with lower risk of HER2-enriched (HR 0.80; 95% CI, 0.65,
0.99) and basal-like tumors (HR 0.79; 95% CI, 0.65, 0.96).
In ancillary analyses (eTable 1), we entered variables for the three food categories together into the fully
adjusted model in place of the indices. We found an inverse association between extreme quintiles of healthy
plant-based foods and ER-negative breast cancer (HR 0.74; 95% CI 0.61, 0.88; p-trend=<0.01).
We assessed effect modification by menopausal status, physical activity and current BMI (Figure 2).
Although we did not find any significant interaction, among women with higher physical activity (≥21MET-
h/week) and lower BMI (<25 kg/m2), higher adherence to the PDI was inversely associated with ER-negative
breast cancer [(HR per 10-unit increase 0.78; 95%CI 0.65, 0.95) ;(HR 0.81; 95%CI 0.70, 0.93)]. Moreover, we observed
a significant inverse association between each 10-unit increase in the hPDI and ER-negative breast cancer
among lean women (HR 0.80; 95% CI 0.71, 0.90).
Associations were similar by menopausal status; however, the association was stronger for
postmenopausal women (HR 0.86; 95%CI 0.78, 0.95). Furthermore, each 10-unit increase in the uPDI was
positively associated with a higher risk of breast cancer among those with low physical activity (HR 1.16; 95%CI
1.06,1.27) and lean women (HR 1.19; 95%CI 1.06, 1.33). Among postmenopausal women (data not shown),
there was a suggestion of stronger associations among never and past users of postmenopausal hormones
(HRQ5vsQ1 0.69; 95%CI 0.54,0.90; p-trend=<0.01) than current users. Nonetheless, p for interaction was non-
significant.
We evaluated the extent to which the inverse association with higher PDI and hPDI may be mediated by
less weight gain from age 18 (data not shown). The calculated mediation proportion was 11.0% (95%CI=3.4% -
30.2%; p=<0.01), indicating that less weight gain could statistically explain 11.0% of the inverse association with
PDI. Moreover, the proportion of hPDI effect mediated by weight change since age 18 was 7.3% (95%CI=3.0% -
16.5%; p=<0.01). Results from latency analysis (eTable 2) indicated the highest vs. the lowest adherence to a
hPDI 0-4 and 4-8 years before diagnosis was associated with lower risk of ER-negative breast cancer [(HR Q5 vs.
Q1 0.85; 95% CI, 0.72, 1.01; p-trend=0.01); (HR Q5 vs. Q1 0.80; 95% CI, 0.66, 0.96; p-trend=0.01)].
DISCUSSION
In two large prospective cohort studies in the US, we found that a higher hPDI score, a measure of
adherence to a high-quality plant-based diet, was associated with lower risk of total breast cancer, independent
of weight change, total carotenoid intake and dietary fiber. The association was most evident in relation to ER-
19–21
negative tumors for which non-hormonal exposures may be most important . Less weight gain could
statistically explain 7.3% of the inverse association with hPDI.
Only few prospective studies have assessed the association between hPDI and uPDI defined by Satija et
12,22–27
al., . Because of the impracticality of assessment of a pure vegetarian diet in these cohorts and the fact
that it may not easily be embraced by many individuals, consuming preferentially plant-derived foods would be a
more comprehensible message. In this context, these scores are designed to understand common dietary
patterns that incorporates a range of progressively increasing proportions of plant foods and accompanying
reductions in animal-foods.
Previous studies that examined associations between these plant-based diets and breast cancer are not
6,13
consistent in the literature . In the SUN project, a moderate, but not greater, adherence to a provegetarian
dietary pattern, was associated with a decreased risk of breast cancer, otherwise no particular associations were
found when assessing healthful and unhealthful provegetarian dietary patterns separately 6. Furthermore, a
higher pro plant-based dietary score was associated with decreased risks of overall cancer though no
association was found for breast cancer 13 which may partly due to the restricted number of cases (n=487 breast
cancer cases, only 13.6% representing ER‐/PR‐and ER-/PR+).
In 2007, the World Cancer Research Fund supported that there was insufficient evidence to make a
judgment about the association between dietary patterns and the risk of breast cancer 28. Subsequently, in 2010,
a systematic review and meta-analysis showed that a Prudent dietary pattern characterized by high intakes of
fruit, vegetables, whole grains, low-fat dairy products, fish and poultry, was linked to a 11% breast cancer risk
reduction29. In a recent systematic review and meta-analysis of 32 observational studies 30
, a Western and a
Prudent diet patterns were associated with a 14% increased and a 18% reduced risk of breast cancer,
respectively.
Hormone receptor status in an important diagnostic and prognostic characteristic of breast tumor and,
therefore, deserves consideration. In previous analyses within the NHS, the Prudent dietary pattern, the Dietary
Approaches to Stop Hypertension 31,32, the Mediterranean Diet 33 and a diabetes risk reduction diet 34 have been
significantly associated with a decreased risk of ER-negative breast tumors. Thus, results are in close
agreement with previous findings, which could be expected, as most healthy plant foods positively weighted in
35–38 39 40 34,41
the hPDI (e.g., vegetables, fruits , whole grains , olive oil or coffee ) have been associated with
lower risk of breast cancer in prospective cohort studies, including our own. Moreover, higher intakes of dietary
42,43
fiber and carotenoids, particularly α-carotene, β-carotene, and lutein/zeaxanthin as shown in a pooled
44
analysis of 18 prospective cohort studies, were inversely associated with risk of ER-negative breast cancer .
These compounds have been hypothesized to reduce cancer risk through antioxidant, anti-inflammatory or
45
antiproliferative activity and by decreasing concentrations of circulating estrogens and androgens, and
46
reducing insulin-resistance . However, adjustment for carotenoid intake, dietary fiber or fruits and vegetables
slightly attenuated the observed associations with regard to hPDI adherence and ER-negative breast tumors,
indicating that other constituents in plant-based foods also account for the observed findings. In this sense, other
phytochemicals present in fruits, vegetables, whole grains and legumes, such as flavonoids or other phenolic
compounds, or an interaction among several phytochemicals, could be responsible for the observed association
47
. Future studies should explore the associations between additional phytochemical components and breast
cancer risk. On the other hand, less healthy plant-based foods positively weighted in the uPDI (refined grains,
48 49–51 52
pastries , sugary drinks or processed foods ) have been associated with higher risk of breast cancer in
other studies.
In our analyses, the consistent opposite associations of hPDI and uPDI also strengthened the value of
considering plant-based dietary quality and increasing intake of healthy plant foods while lessening consumption
of less healthy plant foods. A healthful version of a PDI represents many advantages such as the representation
of a healthy plant-based diet without complete exclusion of animal foods. In a previous analysis from the NHS
27
and the Health Professionals Follow-Up Study where changes in plant-based diet quality and mortality were
evaluated, a 10-point increase in hPDI could be reached by increasing healthy plant foods (i.e., fruits, vegetables
and whole grains) by about 3 servings/day and decreasing less healthy plant foods (i.e., refined grains and
sugary beverages) and some animal foods (i.e., processed meat) by approximately 2 servings/day. Such an
approach is preferable because it is flexible and allows individuals to make gentle changes to their diets.
Besides, excluding all animal foods might not be convenient for all populations as there is limited-suggestive
evidence that moderate intakes of some animal foods such as fish and dairy may be associated with lower
breast cancer risk 53.
Strengths and limitations of study
The strengths of this study include the two large prospective cohorts, the large sample size, the long
follow-up period and low attrition. Detailed collection of updated dietary, lifestyle and medical information (such
as tissue information for the determination of molecular subtypes) over several decades permitted the evaluation
of quality of plant-based diets and the adjustment for a widely recognized confounders of these associations.
However, several limitations should be acknowledged. First, generalizability may be limited because participants
in our study were all health professionals and were predominantly white. However, the high educational status of
our participants should be considered as a strength because it permitted us to gather detailed and accurate
information on diet, lifestyle, and other health variables and minimize confounding by socioeconomic status.
Second, whilst we controlled for a wide variety of lifestyle factors and excluded participants with cancer, or
implausible energy intakes, the possibility of residual confounding cannot be excluded due to the observational
nature of the study. Third, our dietary assessment was based on self-reported questionnaires, which inevitably
produce measurement errors; however, these would likely be non-differential in relation to risk of breast cancer
and therefore, have caused underestimation of associations. Nonetheless, the FFQs used in the current study
54,55
were extensively validated against diet records and biomarkers . Fourth, one challenge in examining ER-
negative breast cancer in epidemiologic studies is that it only accounts for 15% to 20% of breast cancer 56. Thus,
further analyses exploring molecular subtypes beyond ER status are warranted given our limited power.
CONCLUSIONS
A healthful plant-based diet was significantly associated with lower risk of total breast cancer,
independently of total carotenoid intake, dietary fiber and weight change, and specifically for ER-negative
tumors. Our results support dietary guidelines that emphasize increasing intake of healthy plant-based foods for
breast cancer prevention. While there is some mechanistic support for the associations with breast cancer
subtypes, further confirmatory studies are warranted.
Acknowledgments
We would like to thank the participants of the NHS and NHSII and the following state cancer registries for their
help: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND,
OH, OK, OR, PA, RI, SC, TN, TX, VA, WA, WY. The authors assume full responsibility for analyses and
interpretation of these data.
Author Contributors:
The authors’ responsibilities were as follows—ARN, WCW, BAR, and AHE: designed and conducted the
research; ARN, AHE, and BAR: analyzed the data or performed the statistical analysis; ARN: had primary
responsibility for the final content; and all authors wrote the paper and read and approved the final manuscript.
The authors report no conflicts of interest.
Funding/Support: This study was supported by grants UM1 CA186107, U01 CA176726, P01 CA87969, and
R01 CA50385 from the National Institutes of Health, and the Breast Cancer Research Foundation. ARN was
supported by a fellowship from the Spanish Association Against Cancer Scientific Foundation (FC AECC).
The funding sources did not participate in the design or conduct of the study; collection, management, analysis
or interpretation of the data; or preparation, review, or approval of the manuscript.
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study;
collection, management, analysis, and interpretation of the data; preparation, review or approval of the
manuscript; and decision to submit the manuscript for publication.
REFERENCES
1. Hemler EC, Hu FB. Plant-Based Diets for Personal, Population, and Planetary Health. Adv Nutr. 2019.
doi:10.1093/advances/nmy117
2. Huang T, Yang B, Zheng J, Li G, Wahlqvist ML, Li D. Cardiovascular disease mortality and cancer incidence in
vegetarians: A meta-analysis and systematic review. Ann Nutr Metab. 2012. doi:10.1159/000337301
3. Kwok CS, Umar S, Myint PK, Mamas MA, Loke YK. Vegetarian diet, Seventh Day Adventists and risk of
cardiovascular mortality: A systematic review and meta-analysis. Int J Cardiol. 2014.
doi:10.1016/j.ijcard.2014.07.080
4. Tonstad S, Stewart K, Oda K, Batech M, Herring RP, Fraser GE. Vegetarian diets and incidence of diabetes in
the Adventist Health Study-2. Nutr Metab Cardiovasc Dis. 2013. doi:10.1016/j.numecd.2011.07.004
5. Yokoyama Y, Nishimura K, Barnard ND, et al. Vegetarian diets and blood pressure ameta-analysis. JAMA
Intern Med. 2014. doi:10.1001/jamainternmed.2013.14547
6. Romanos-Nanclares A, Toledo E, Sánchez-Bayona R, Sánchez-Quesada C, Martínez-González MÁ, Gea A.
Healthful and Unhealthful Provegetarian Food Patterns and the Incidence of Breast Cancer: results from a
Mediterranean cohort. Nutrition. 2020. doi:10.1016/j.nut.2020.110884
7. Willett W, Rockström J, Loken B, et al. Food in the Anthropocene: the EAT–Lancet Commission on healthy
diets from sustainable food systems. Lancet. 2019;6736(18):3–49. doi:10.1016/S0140-6736(18)31788-4
8. Leitzmann C. Vegetarian nutrition: Past, present, future. Am J Clin Nutr.; 2014. doi:10.3945/ajcn.113.071365
9. Allès B, Baudry J, Méjean C, et al. Comparison of sociodemographic and nutritional characteristics between
self-reported vegetarians, vegans, and meat-eaters from the nutrinet-santé study. Nutrients. 2017.
doi:10.3390/nu9091023
10. Martínez-González MA, Sánchez-Tainta A, Corella D, et al. A provegetarian food pattern and reduction in total
mortality in the Prevención con Dieta Mediterránea (PREDIMED) study. Am J Clin Nutr.
2014;100(suppl_1):320S-328S. doi:10.3945/ajcn.113.071431
11. Satija A, Hu FB. Plant-based diets and cardiovascular health. Trends Cardiovasc Med. 2018.
doi:10.1016/j.tcm.2018.02.004
12. Satija A, Bhupathiraju SN, Rimm EB, et al. Plant-Based Dietary Patterns and Incidence of Type 2 Diabetes in
US Men and Women: Results from Three Prospective Cohort Studies. PLoS Med. 2016.
doi:10.1371/journal.pmed.1002039
13. Kane‐ Diallo A, Srour B, Sellem L, et al. Association between a pro plant‐ based dietary score and cancer risk
in the prospective NutriNet‐ santé cohort. Int J Cancer. 2018. doi:10.1002/ijc.31593
14. Lee JE, Willett WC, Fuchs CS, et al. Folate intake and risk of colorectal cancer and adenoma: Modification by
time. Am J Clin Nutr. 2011. doi:10.3945/ajcn.110.007781
15. Lin DY, Fleming TR, De Gruttola V. Estimating the proportion of treatment effect explained by a surrogate
marker. Stat Med. 1997;16(13):1515-1527.
16. Stryker WS, Kaplan LA, Stein EA, Stampfer MJ, Sober A, Willett WC. The relation of diet, cigarette smoking,
and alcohol consumption to plasma beta-carotene and alpha-tocopherol levels. Am J Epidemiol. 1988.
17. Jun HJ, Austin SB, Wylie SA, et al. The mediating effect of childhood abuse in sexual orientation disparities in
tobacco and alcohol use during adolescence: Results from the Nurses’ Health Study II. Cancer Causes
Control. 2010. doi:10.1007/s10552-010-9609-3
18. Lunn M, McNeil D. Applying Cox Regression to Competing Risks. Biometrics. 1995. doi:10.2307/2532940
19. Sisti JS, Collins LC, Beck AH, Tamimi RM, Rosner BA, Eliassen AH. Reproductive risk factors in relation to
molecular subtypes of breast cancer: Results from the nurses’ health studies. Int J Cancer. 2016.
doi:10.1002/ijc.29968
20. Barnard ME, Boeke CE, Tamimi RM. Established breast cancer risk factors and risk of intrinsic tumor
subtypes. Biochim Biophys Acta - Rev Cancer. 2015. doi:10.1016/j.bbcan.2015.06.002
21. Martínez ME, Cruz GI, Brewster AM, Bondy ML, Thompson PA. What can we learn about disease etiology
from case-case analyses? Lessons from breast cancer. Cancer Epidemiol Biomarkers Prev. 2010.
doi:10.1158/1055-9965.EPI-10-0742
22. Satija A, Bhupathiraju SN, Spiegelman D, et al. Healthful and Unhealthful Plant-Based Diets and the Risk of
Coronary Heart Disease in U.S. Adults. J Am Coll Cardiol. 2017. doi:10.1016/j.jacc.2017.05.047
23. Kim H, Caulfield LE, Garcia-Larsen V, et al. Plant-Based Diets and Incident CKD and Kidney Function. Clin J
Am Soc Nephrol. 2019. doi:10.2215/CJN.12391018
24. Gómez-Donoso C, Martínez-González MÁ, Martínez JA, et al. A provegetarian food pattern emphasizing
preference for healthy plant-derived foods reduces the risk of overweight/obesity in the SUN cohort. Nutrients.
2019. doi:10.3390/nu11071553
25. Baden MY, Satija A, Hu FB, Huang T. Change in plant-based diet quality is associated with changes in plasma
adiposity-associated biomarker concentrations in women. J Nutr. 2019. doi:10.1093/jn/nxy301
26. Kim H, Caulfield LE, Rebholz CM. Healthy plant-based diets are associated with lower risk of all-cause
mortality in US Adults. J Nutr. 2018. doi:10.1093/jn/nxy019
27. Baden MY, Liu G, Satija A, et al. Changes in Plant-Based Diet Quality and Total and Cause-Specific Mortality.
Circulation. 2019. doi:10.1161/CIRCULATIONAHA.119.041014
28. Wiseman M. The Second World Cancer Research Fund/American Institute for Cancer Research expert report.
Food, Nutrition, Physical Activity, and the Prevention of Cancer: A Global Perspective. In: Proceedings of the
Nutrition Society. ; 2008. doi:10.1017/S002966510800712X
29. Brennan SF, Cantwell MM, Cardwell CR, Velentzis LS, Woodside J V. Dietary patterns and breast cancer risk:
a systematic review and meta-analysis. Am J Clin Nutr. 2010;91(5):1294-1302. doi:10.3945/ajcn.2009.28796
30. Xiao Y, Xia J, Li L, et al. Associations between dietary patterns and the risk of breast cancer: A systematic
review and meta-analysis of observational studies. Breast Cancer Res. 2019. doi:10.1186/s13058-019-1096-1
31. Fung TT, Hu FB, Hankinson SE, Willett WC, Holmes MD. Low-carbohydrate diets, dietary approaches to stop
hypertension-style diets, and the risk of postmenopausal breast cancer. Am J Epidemiol. 2011.
doi:10.1093/aje/kwr148
32. Hirko KA, Willett WC, Hankinson SE, et al. Healthy dietary patterns and risk of breast cancer by molecular
subtype. Breast Cancer Res Treat. 2016. doi:10.1007/s10549-016-3706-2
33. Fung TT, Hu FB, McCullough ML, Newby PK, Willett WC, Holmes MD. Diet Quality Is Associated with the Risk
of Estrogen Receptor–Negative Breast Cancer in Postmenopausal Women. J Nutr. 2006.
doi:10.1093/jn/136.2.466
34. Kang JH, Peng C, Rhee JJ, et al. Prospective study of a diabetes risk reduction diet and the risk of breast
cancer. Am J Clin Nutr. October 2020. doi:10.1093/ajcn/nqaa268
35. Jung S, Spiegelman D, Baglietto L, et al. Fruit and vegetable intake and risk of breast cancer by hormone
receptor status. J Natl Cancer Inst. 2013. doi:10.1093/jnci/djs635
36. Fung TT, Chiuve SE, Willett WC, Hankinson SE, Hu FB, Holmes MD. Intake of specific fruits and vegetables in
relation to risk of estrogen receptor-negative breast cancer among postmenopausal women. Breast Cancer
Res Treat. 2013. doi:10.1007/s10549-013-2484-3
37. Farvid MS, Chen WY, Rosner BA, Tamimi RM, Willett WC, Eliassen AH. Fruit and vegetable consumption and
breast cancer incidence: Repeated measures over 30 years of follow-up. Int J Cancer. 2018.
doi:10.1002/ijc.31653
38. Farvid MS, Chen WY, Rosner BA, Tamimi RM, Willett WC, Eliassen AH. Fruit and vegetable consumption and
breast cancer incidence: Repeated measures over 30 years of follow-up. Int J Cancer. 2019.
doi:10.1002/ijc.31653
39. Farvid MS, Cho E, Eliassen AH, Chen WY, Willett WC. Lifetime grain consumption and breast cancer risk.
Breast Cancer Res Treat. 2016. doi:10.1007/s10549-016-3910-0
40. Toledo E, Salas-Salvado J, Donat-Vargas C, et al. Mediterranean Diet and Invasive Breast Cancer Risk
Among Women at High Cardiovascular Risk in the PREDIMED Trial: A Randomized Clinical Trial. JAMA Intern
Med. 2015;175(11):1752-1760. doi:10.1001/jamainternmed.2015.4838
41. Sánchez-Quesada C, Romanos-Nanclares A, Navarro AM, et al. Coffee consumption and breast cancer risk in
the SUN project. Eur J Nutr. 2020. doi:10.1007/s00394-020-02180-w

42. Chen S, Chen Y, Ma S, et al. Dietary fibre intake and risk of breast cancer: A systematic review and meta-
analysis of epidemiological studies. Oncotarget. 2016. doi:10.18632/oncotarget.13140
43. Reynolds A, Mann J, Cummings J, Winter N, Mete E, Te Morenga L. Carbohydrate quality and human health:
a series of systematic reviews and meta-analyses. Lancet. 2019. doi:10.1016/S0140-6736(18)31809-9
44. Zhang X, Spiegelman D, Baglietto L, et al. Carotenoid intakes and risk of breast cancer defined by estrogen
receptor and progesterone receptor status: A pooled analysis of 18 prospective cohort studies. Am J Clin Nutr.
2012. doi:10.3945/ajcn.111.014415
45. Bertram JS. Dietary carotenoids, connexins and cancer: What is the connection? In: Biochemical Society
Transactions. ; 2004. doi:10.1042/BST0320985
46. Maru GB. Understanding the molecular mechanisms of cancer prevention by dietary phytochemicals: From
experimental models to clinical trials. World J Biol Chem. 2016. doi:10.4331/wjbc.v7.i1.88
47. Liu RH. Health benefits of fruit and vegetables are from additive and synergistic combinations of
phytochemicals. In: American Journal of Clinical Nutrition. ; 2003. doi:10.1093/ajcn/78.3.517S
48. Chandran U, McCann SE, Zirpoli G, et al. Intake of energy-dense foods, fast foods, sugary drinks, and breast
cancer risk in African American and European American women. Nutr Cancer. 2014;66(7):1187-1199.
doi:10.1080/01635581.2014.951737
49. Romanos-Nanclares A, Toledo E, Gardeazabal I, Jiménez-Moleón JJ, Martínez-González MA, Gea A. Sugar-
sweetened beverage consumption and incidence of breast cancer: the Seguimiento Universidad de Navarra
(SUN) Project. Eur J Nutr. 2018;48:70-225. doi:10.1007/s00394-018-1839-2
50. Chazelas E, Srour B, Desmetz E, et al. Sugary drink consumption and risk of cancer: Results from NutriNet-
Santé prospective cohort. BMJ. 2019. doi:10.1136/bmj.l2408
51. Hodge AM, Bassett JK, Milne RL, English DR, Giles GG. Consumption of sugar-sweetened and artificially
sweetened soft drinks and risk of obesity-related cancers. Public Health Nutrition. 2018:1-9.
52. Fiolet T, Srour B, Sellem L, et al. Consumption of ultra-processed foods and cancer risk: Results from
NutriNet-Santé prospective cohort. BMJ. 2018. doi:10.1136/bmj.k322
53. World Cancer Research Fund, American Institute for Cancer Research. Diet, nutrition, physical activity and
breast cancer. 2017:117. wcrf.org/breast-cancer–2017.
54. Salvini S, Hunter DJ, Sampson L, et al. Food-based validation of a dietary questionnaire: The effects of week-
to-week variation in food consumption. Int J Epidemiol. 1989. doi:10.1093/ije/18.4.858
55. Willett WC, Sampson L, Stampfer MJ, et al. Reproducibility and validity of a semiquantitative food frequency
questionnaire. Am J Epidemiol. 1985. doi:10.1093/oxfordjournals.aje.a114086
56. Ries LAG, Young JL, Keel GE, Eisner MP, Lin YD, Horner M-J (editors). SEER Survival Monograph: Cancer
Survival Among Adults: U.S. SEER Program, 1988-2001, Patient and Tumor Characteristics. National Cancer
Institute, SEER Program, NIH Pub. No. 07-6215, Bethesda, MD, 2007.
Table 1: Age and age-standardized baseline characteristics of women according to quintiles of an overall plant-based diet index (PDI) in the Nurses’ Health Study (NHS) and
NHSII.
NHS (1984) NHSII (1991)
Q1 (n=14,629) Q3 (n=13,832) Q5 (n=15,857) Q1 (n=20,408) Q3 (n=16,541) Q5 (n=17,944)
Median PDI 46 54 62 47 55 63
a
Age, years 50.2 (7) 50.7 (7.2) 51.7 (7.3) 36.5 (4.8) 36.6 (4.7) 36.9 (4.5)
Body mass index, kg/m2 25.7 (5.1) 25.1 (4.7) 24.5 (4.4) 25.2 (5.7) 24.5 (5.2) 23.7 (4.8)
2
Body mass index at age 18 years, kg/m 21.7 (3.2) 21.4 (3.0) 21.1 (2.8) 21.6 (3.6) 21.3 (3.3) 20.9 (3.0)
Weight change from age 18 years, kg 10.7 (12.3) 9.8 (11.1) 9.0 (10.5) 9.8 (12.4) 8.7 (11.2) 7.7 (10.4)
Height, inches 64.5 (2.4) 64.5 (2.4) 64.5 (2.4) 64.9 (2.6) 64.9 (2.6) 64.9 (2.6)
Self-reported African heritage (%) 1.5 1.4 1.1 2.0 1.4 1.2
Self-reported history of diabetes (%) 3.8 3.3 2.3 1.2 1.0 0.7
Family history of breast cancer (%) 7.8 8.1 8.1 5.8 6.1 6.3
Personal history of benign breast disease (%) 29.5 30.2 31.0 9.1 9.5 9.6
Age at menarche <12 years (%) 23.5 21.6 22.5 24.8 24.0 24.8
Oral contraceptives, ever (%) 50.5 49.3 48.3 85.6 84.7 83.4
Parous (%) 92.0 92.9 93.3 69.3 75.9 77.5
Parity, nb 3.1 (1.5) 3.2 (1.5) 3.2 (1.5) 2.1 (0.9) 2.1 (0.9) 2.2. (0.9)
Breastfeeding, ≤6 months (%)b 35.8 36.6 36.1 19.7 16.9 14.1
Postmenopausal (%) 48.4 48.6 48.8 3.1 3.2 3.4
Postmenopausal hormone use, never (%)c 52.9 52.7 52.8 6.1 6.9 7.8
Physical activity, METs-h/week 11.4 (13.2) 11.9 (13.1) 12.9 (13.4) 17.8 (24.4) 20.6 (26.8) 24.9 (31.5)
Alcohol intake (g/day) 8.6 (13.6) 6.7 (11) 5.8 (9.3) 3.3 (7.1) 3 (5.8) 3.2 (5.6)
Total energy intake (kcal/day) 1458 (451) 1723 (488) 2056 (519) 1478 (464) 1792 (504) 2134 (527)
Saturated fat (% of energy) 13.9 (2.9) 12.5 (2.3) 11.2 (2.1) 12.6 (2.5) 11.1 (2.2) 9.8 (2.1)
Monounsaturated fat (% of energy) 13.4 (2.6) 12.7 (2.3) 12 (2.2) 12.7 (2.6) 11.9 (2.4) 11.2 (2.3)
Polyunsaturated fat (% of energy) 6.5 (1.9) 6.6 (1.7) 6.8 (1.6) 5.7 (1.5) 5.6 (1.4) 5.6 (1.3)
Trans fat (% of energy) 1.8 (0.6) 1.9 (0.6) 1.9 (0.6) 1.7 (0.7) 1.6 (0.6) 1.5 (0.5)
Protein intake (% of energy) 19.7 (3.9) 17.7 (3.1) 16.2 (2.5) 21 (3.8) 19.2 (3.2) 17.6 (2.9)
Carbohydrate intake (% of energy) 40.5 (8) 46.7 (6.8) 51.2 (6.5) 44.8 (7.4) 50.1 (6.5) 54.7 (6.6)
Healthy plant foods (servings/day)
Whole grains 0.7 (0.9) 1.1 (1.1) 1.6 (1.2) 0.9 (0.9) 1.4 (1.1) 2.1 (1.3)
Fruits 0.9 (0.9) 1.4 (1) 1.9 (1.1) 0.7 (0.7) 1.2 (0.9) 1.8 (1.1)
Vegetables 2.3 (1.4) 2.9 (1.5) 3.7 (1.8) 2.1 (1.4) 2.9 (1.8) 4.1 (2.1)
Nuts 0.1 (0.2) 0.2 (0.3) 0.3 (0.4) 0.1 (0.1) 0.2 (0.2) 0.3 (0.3)
Legumes 0.3 (0.2) 0.4 (0.3) 0.5 (0.3) 0.2 (0.2) 0.4 (0.3) 0.6 (0.4)
Vegetable oil 0.2 (0.3) 0.3 (0.4) 0.4 (0.4) 0.2 (0.3) 0.3 (0.4) 0.5 (0.5)
Tea and coffee 2.6 (1.9) 3.1 (1.9) 3.6 (2) 1.8 (1.8) 2.2 (1.9) 2.7 (2)
Less healthy plant foods (servings/day)
Fruit juices 0.4 (0.6) 0.7 (0.7) 1 (0.8) 0.4 (0.6) 0.6 (0.7) 1.1 (1)
Refined grains 1.2 (1.1) 1.6 (1.3) 2.2 (1.5) 1.2 (0.9) 1.6 (1) 2 (1.1)
Potatoes 0.4 (0.3) 0.5 (0.4) 0.7 (0.4) 0.4 (0.3) 0.5 (0.4) 0.7 (0.4)
Sugar-sweetened beverages 0.2 (0.6) 0.3 (0.6) 0.4 (0.6) 0.4 (0.8) 0.5 (0.9) 0.6 (0.9)
Sweets and desserts 0.8 (0.9) 1.2 (1.2) 1.8 (1.4) 0.8 (0.9) 1.2 (1.1) 1.6 (1.2)
Animal foods (servings/day)
Animal fat 0.5 (0.9) 0.4 (0.8) 0.3 (0.6) 0.2 (0.5) 0.2 (0.4) 0.1 (0.4)
Dairy 2 (1.4) 2 (1.4) 2 (1.3) 2.3 (1.5) 2.3 (1.5) 2.3 (1.4)
Eggs 0.4 (0.4) 0.3 (0.3) 0.3 (0.3) 0.2 (0.2) 0.2 (0.2) 0.2 (0.2)
Fish and seafood 0.3 (0.3) 0.3 (0.3) 0.3 (0.2) 0.3 (0.2) 0.3 (0.2) 0.3 (0.3)
Meat 1.7 (0.8) 1.7 (0.8) 1.7 (0.8) 1.6 (0.8) 1.7 (0.8) 1.6 (0.9)
Miscellaneous animal-based foods 0.5 (0.4) 0.4 (0.4) 0.4 (0.4) 0.4 (0.4) 0.4 (0.3) 0.4 (0.3)
Food groups (servings/day)
Healthy plant foods 7.1 (3) 9.3 (3.3) 12.1 (3.7) 6 (2.9) 8.6 (3.5) 12 (4.1)
Less healthy plant foods 3 (1.9) 4.4 (2.3) 6.1 (2.7) 3.1 (1.8) 4.4 (2.2) 5.9 (2.4)
Animal foods 5.4 (2.3) 5.1 (2.1) 4.9 (1.9) 5 (2.1) 5 (2.1) 4.8 (2)
Q= Quintile; NHS= Nurses’ Health Study; NHSII= Nurses’ Health Study II; PDI= plant-based diet index; METs-h/week=metabolic equivalent task hours per week.
Values are means (standard deviations) for continuous variables and percentages for categorical variables. All variables except age are standardized to the age distribution of
the study population.
Each food intakes expressed as servings/day. The SI equivalent of 1 inch is equivalent to 2.54 cm.
a
Value is not age adjusted; b Among parous women only; c Among postmenopausal.
Table 2. Age-adjusted and multivariable-adjusted hazard ratios (95% confidence intervals) for total breast cancer according to quintiles of plant-based diet indices (PDI, hPDI,
uPDI) in the NHS and NHSII.
HR
Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5 P Trend (95% CI)
per 10 units
Plant-based dietary index (PDI)
Median 47 51 54 57 61
Cases/PY 1671/472789 1603/469817 1656/475439 1657/479745 1633/479292
NHS
Age-adjusted 1.00 0.96 (0.90, 1.03) 0.98 (0.91, 1.05) 0.96 (0.90, 1.03) 0.94 (0.87, 1.00) 0.08 0.95 (0.92, 0.99)
MV 1.00 0.95 (0.89, 1.02) 0.96 (0.89, 1.03) 0.94 (0.87, 1.01) 0.91 (0.84, 0.98) 0.01 0.93 (0.89, 0.97)
Median 47 51 54 58 62
Cases/PY 869/494773 868/484733 829/504428 843/493618 853/486450
NHSII
Age-adjusted 1.00 1.01 (0.92, 1.11) 0.91 (0.83, 1.00) 0.94 (0.86, 1.04) 0.95 (0.87, 1.05) 0.14 0.97 (0.92, 1.02)
MV 1.00 0.97 (0.89, 1.07) 0.87 (0.78, 0.95) 0.88 (0.80, 0.97) 0.87 (0.79, 0.97) <0.01 0.92 (0.86, 0.98)
Median 47 51 54 58 62
2471/954550 2485/979867 2500/973362
Pooled
Cases/PY 2540/967562 2486/965742

Age-adjusted 1.00 0.98 (0.92, 1.03) 0.95 (0.90, 1.01) 0.95 (0.90, 1.01) 0.94 (0.89, 0.99) 0.02 0.96 (0.93, 0.99)
MV 1.00 0.96 (0.91, 1.01) 0.93 (0.87, 0.98) 0.92 (0.86, 0.97) 0.89 (0.84, 0.95) <0.01 0.92 (0.89, 0.96)
Healthful Plant-based dietary index (hPDI)
Median 47 52 55 58 63
Cases/PY 1588/476504 1680/479159 1633/471599 1645/468759 1674/481061
NHS
Age-adjusted 1.00 1.02 (0.95, 1.10) 0.99 (0.92, 1.06) 0.99 (0.92, 1.06) 0.96 (0.90, 1.03) 0.15 0.98 (0.94, 1.01)
MV 1.00 1.01 (0.94, 1.08) 0.96 (0.89, 1.03) 0.95 (0.88, 1.02) 0.92 (0.85, 0.99) 0.01 0.96 (0.92, 0.99)
Median 46 51 55 58 64
Cases/PY 825/495153 839/480274 879/497668 895/508067 824/482839
NHSII
Age-adjusted 1.00 1.00 (0.91, 1.10) 0.99 (0.90, 1.09) 0.97 (0.88, 1.06) 0.89 (0.81, 0.98) 0.01 0.95 (0.91, 1.00)
MV 1.00 0.97 (0.88, 1.07) 0.95 (0.86, 1.04) 0.92 (0.83, 1.02) 0.83 (0.75, 0.92) <0.01 0.92 (0.87, 0.97)
Median 46 51 55 58 64
2519/959433 2512/969268 2540/976826
Pooled
Cases/PY 2413/971657 2498/963900

Age-adjusted 1.00 1.02 (0.96, 1.07) 0.99 (0.94, 1.05) 0.98 (0.93, 1.04) 0.94 (0.88, 0.99) <0.01 0.97 (0.94, 0.99)
MV 1.00 0.99 (0.94, 1.05) 0.96 (0.90, 1.01) 0.94 (0.88, 0.99) 0.89 (0.83, 0.94) <0.01 0.94 (0.91, 0.97)
Unhealthful Plant-based dietary index (uPDI)
Median 46 51 55 58 64
Cases/PY 1722/478657 1682/475096 1613/473993 1613/476991 1590/472345
NHS
Age-adjusted 1.00 0.99 (0.93, 1.06) 0.96 (0.90, 1.03) 0.96 (0.90, 1.03) 0.98 (0.91, 1.04) 0.30 0.98 (0.95, 1.01)
MV 1.00 0.99 (0.93, 1.06) 0.96 (0.90, 1.03) 0.97 (0.90, 1.05) 1.00 (0.93, 1.08) 0.87 0.99 (0.96, 1.03)
Median 46 51 55 59 64
NHSII
Cases/PY 884/493238 844/491431 852/497283 859/485395 823/496654

Age-adjusted 1.00 0.99 (0.90, 1.09) 1.01 (0.91, 1.11) 1.06 (0.96, 1.16) 1.02 (0.93, 1.12) 0.35 1.02 (0.98, 1.06)
MV 1.00 1.00 (0.91, 1.10) 1.04 (0.94, 1.14) 1.10 (1.00, 1.22) 1.10 (0.99, 1.22) 0.03 1.06 (1.00, 1.11)
Median 46 51 55 59 64
2526/966528 2465/971276 2472/962386
Pooled
Cases/PY 2606/971894 2413/968999

Age-adjusted 1.00 0.99 (0.94, 1.05) 0.98 (0.92, 1.03) 0.99 (0.94, 1.05) 0.99 (0.94, 1.05) 0.80 0.99 (0.97, 1.02)
MV 1.00 1.00 (0.94, 1.05) 0.99 (0.93, 1.05) 1.02 (0.96, 1.08) 1.04 (0.97, 1.10) 0.20 1.02 (0.99, 1.05)
Multivariable model (MV) stratified by age in months and calendar year, adjusted for race (Non-Hispanic Caucasian, African-American, Asian-American, Hispanic Caucasian),
age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current
user– estrogen only for <5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin user for ≥5 years,
current user of other types), oral contraceptive use history (never, ever), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children
before age 25, 2+ children ≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no),
history of benign breast disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5-9, 10-14, 15+ g/day),
cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass index at age 18 years (<20.0, 20.0-21.9, 22.0-23.9, 24.0-
26.9, ≥27.0) and neighborhood-based socioeconomic status indicator (continuous).
Table 3. Multivariable-adjusted hazard ratios (95% confidence intervals) for the association between quintiles (Q) of cumulatively updated PDI, hPDI and uPDI and breast
cancer tumor subtypes in the NHS and NHSII.
P for HR
Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5 P Trend heteroge (95% CI)
neity per-10 units
ESTROGEN RECEPTOR STATUS (NHS with follow-up of 1984-2016 and NHSII with follow-up of 1991-2017)
Cases/PY 1609/968461 1599/955406 1659/980656 1636/974166 1632/966567
ER+
MV 1.00 0.97 (0.91, 1.04) 0.97 (0.90, 1.04) 0.94 (0.88, 1.02) 0.93 (0.86, 1.00) 0.05 0.95 (0.91, 0.99)
Cases/PY 368/969664 358/956537 334/981914 345/975387 345/967764
ER-
MV 1.00 0.96 (0.82, 1.11) 0.86 (0.74, 1.01) 0.87 (0.74, 1.02) 0.85 (0.72, 1.00) 0.03 0.25a 0.88 (0.80, 0.96)
ER + Cases/PY 1551/972520 1577/960328 1671/970065 1656/977670 1680/964673
MV 1.00 0.96 (0.89, 1.03) 0.98 (0.91, 1.05) 0.94 (0.87, 1.01) 0.91 (0.85, 0.99) 0.02 0.96 (0.92, 1.00)
ER- Cases/PY 377/973605 390/961476 325/971282 336/978887 322/966017
MV 1.00 0.99 (0.86, 1.15) 0.81 (0.70, 0.95) 0.82 (0.70, 0.96) 0.77 (0.65, 0.90) <0.01 0.01a 0.85 (0.78, 0.92)
ER + Cases/PY 1750/972733 1690/967320 1577/972116 1594/963230 1524/969856
MV 1.00 0.99 (0.93, 1.06) 0.95 (0.88, 1.02) 0.99 (0.92, 1.06) 1.00 (0.93, 1.08) 0.99 0.99 (0.96, 1.03)
ER- Cases/PY 305/974085 342/968560 377/973277 367/964382 359/970963
MV 1.00 1.17 (1.00, 1.36) 1.29 (1.10, 1.50) 1.26 (1.07, 1.48) 1.28 (1.08, 1.51) <0.01 0.01a 1.12 (1.04, 1.21)
MOLECULAR SUBTYPES (NHS with follow-up of 1984-2006 and NHSII with follow-up of 1991-2005)
Cases/PY 571/661577 573/648175 611/670335 592/667358 572/658149
Luminal Ab
MV 1.00 0.99 (0.88, 1.11) 1.00 (0.89, 1.12) 0.95 (0.84, 1.08) 0.91 (0.80, 1.03) 0.11 0.96 (0.89, 1.03)
b Cases/PY 271/661825 228/648509 267/670634 270/667639 281/658411
Luminal B
MV 1.00 0.86 (0.72, 1.03) 0.98 (0.82, 1.17) 1.00 (0.84, 1.20) 1.04 (0.86, 1.26) 0.39 1.00 (0.89, 1.12)
Cases/PY 46/662033 47/648657 61/670826 57/667847 59/658604
HER-2b
MV 1.00 0.97 (0.64, 1.47) 1.22 (0.82, 1.82) 1.10 (0.73, 1.66) 1.10 (0.72, 1.69) 0.57 0.99 (0.77, 1.26)
Cases/PY 65/662011 73/648631 53/670829 54/667847 54/658617
Basal-likeb
MV 1.00 1.10 (0.78, 1.55) 0.72 (0.50, 1.05) 0.70 (0.48, 1.03) 0.69 (0.46, 1.02) 0.01 0.04a 0.77 (0.62, 0.96)
Cases/PY 561/664554 523/647388 626/667302 604/668016 605/658333
Luminal Ab
MV 1.00 0.88 (0.78, 1.00) 0.99 (0.88, 1.11) 0.93 (0.82, 1.05) 0.88 (0.77, 1.00) 0.12 0.95 (0.89, 1.01)
Cases/PY 254/664818 247/647610 265/667608 265/668332 286/658649
Luminal Bb
MV 1.00 0.94 (0.79, 1.13) 0.96 (0.80, 1.15) 0.96 (0.80, 1.15) 1.00 (0.83, 1.20) 0.99 1.01 (0.92, 1.11)
Cases/PY 56/664998 66/647780 62/667794 41/668527 45/658868
HER-2b
MV 1.00 1.10 (0.76, 1.58) 1.05 (0.72, 1.53) 0.69 (0.45, 1.06) 0.72 (0.47, 1.11) 0.04 0.80 (0.65, 0.99)
Cases/PY 69/664993 56/647785 60/667792 62/668505 52/658861
Basal-likeb
MV 1.00 0.82 (0.57, 1.17) 0.85 (0.59, 1.21) 0.86 (0.60, 1.24) 0.70 (0.47, 1.04) 0.13 0.19a 0.79 (0.65, 0.96)
Cases/PY 621/663923 611/657264 576/664708 575/656569 536/663131
Luminal Ab
MV 1.00 1.02 (0.91, 1.15) 0.99 (0.88, 1.11) 1.02 (0.90, 1.15) 0.99 (0.87, 1.13) 0.91 0.98 (0.92, 1.04)
b Cases/PY 278/664221 279/657567 248/665003 239/656848 273/663379
Luminal B
MV 1.00 1.01 (0.85, 1.19) 0.90 (0.75, 1.07) 0.88 (0.73, 1.05) 1.03 (0.86, 1.24) 0.81 0.98 (0.89, 1.07)
Cases/PY 46/664428 45/657782 64/665164 54/657032 61/663561
HER-2b
MV 1.00 0.97 (0.64, 1.47) 1.43 (0.97, 2.12) 1.22 (0.81, 1.85) 1.49 (0.98, 2.27) 0.04 1.20 (0.98, 1.46)
Cases/PY 59/664419 60/657757 59/665173 57/657036 64/663551
Basal-likeb
MV 1.00 1.04 (0.72, 1.49) 1.01 (0.70, 1.46) 0.98 (0.67, 1.44) 1.13 (0.77, 1.66) 0.66 0.21a 1.09 (0.91, 1.31)
PY= Person-years; ER= estrogen receptor; HER= human epidermal growth factor receptor 2.
Multivariable model (MV) stratified by cohort, age in months and calendar year, adjusted for race (Non-Hispanic Caucasian, African, Asian, Hispanic Caucasian), age at
menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current user–
estrogen only for <5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin user for ≥5 years,
current user of other types), oral contraceptive use history (never, ever), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+
children before age 25, 2+ children ≥25 years of age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes
or no), history of benign breast disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5-9, 10-14, 15+
g/day), cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass index at age 18 years (<20.0, 20.0-21.9, 22.0-
23.9, 24.0-26.9, ≥27.0) and socioeconomic status (continuous).
a
For testing heterogeneity by subtype, we used the Lunn-McNeil approach, for Multivariable model (MV).
b
Due to smaller sample sizes in analyses, to ensure that models would run, covariate categorizations were simplified.
Figure 1. Multivariable spline analysis of the association between adherence to a healthful and unhealthful plant-based dietary indexes and risk of incident
breast cancer in the NHS (1984-2016) and NHSII (1991-2017).
Figure 2. Pooled hazard ratios of estrogen receptor negative breast cancer per 10-units increment in the three dietary indices (PDI, hPDI and uPDI) across
subgroups (physical activity, current BMI and menopausal status).
Figure 1. Multivariable spline analysis of the association between adherence to a healthful and
unhealthful plant-based dietary indexes and risk of incident breast cancer in the NHS (1984-2016) and
NHSII (1991-2017).
Hazard Ratio for ER- breast cancer

Hazard Ratio for ER+ breast cancer

ER= estrogen receptor; Model stratified by cohort, age in months and 2-year period at risk, adjusted for race (Non-Hispanic Caucasian, African, Asian, Hispanic
Caucasian), cumulatively updated total caloric intake (kcal/day, quintiles), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal,
<45, 45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current user– estrogen only for <5 years, current user – estrogen only for ≥5
years, current estrogen + progestin user for < 5 years, current estrogen + progestin user for ≥5 years, current user of other types), parity and age at first birth
(nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+children ≥25 years of age), breastfeeding history (never, breastfed
for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast disease (yes or no), body mass index ate age 18
years (<20.0, 20.0-21.9, 22.0-23.9, 24.0-26.9, ≥27.0), oral contraceptives use (never/ever), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75+ m),
cumulatively updated alcohol intake (0, <5, 5-9, 10-14, 15+ g/day), physical activity (linear METs-h/week) and neighborhood-based socioeconomic status
indicator (continuous). For ER negative breast cancer cases, the p values for test of curvature for healthful plant-based diet index (hPDI)= 0.95 and for
unhealthful plant-based diet index (uPDI)=0.29. For ER positive cases, the p values for test of curvature for hPDI=0.35 and uPDI=0.15.
Figure 2. Pooled hazard ratios of estrogen receptor negative breast cancer per 10-units
increment in the three dietary indices (PDI, hPDI and uPDI) across subgroups (physical
activity, current BMI and menopausal status).
The hazard ratios (HRs) and p values for women were obtained after combining all 2 cohorts (NHS; NHSII). Stratified by age in months,
cohort and calendar year, adjusted for race (Non-Hispanic Caucasian, African, Asian, Hispanic Caucasian), age at menarche (<12, 12, 13,
14, >14 years), age at menopause (premenopausal, <45, 45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current
user– estrogen only for <5 years, current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current
estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at first birth
(nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age), breastfeeding
history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast
disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5-9, 10-14, 15+
g/day), cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass index at age 18
years (<20.0, 20.0-21.9, 22.0-23.9, 24.0-26.9, ≥27.0) and socioeconomic status (continuous).For the analysis of BMI, we further adjusted for
current body mass index (linear, kg/m 2).To test whether the PDI, hPDI, uPDI and breast cancer association differed by current BMI, physical
activity, or menopausal status we added interaction terms and used the Wald test. BMI= body mass index; CI= confidence interval; MET=
metabolic equivalent task.
Supplementary material
eTable 1: Risk of ER-negative breast cancer according to quintiles of food categories in pooled
data from NHS and NHSII
eTable 2. Risk of estrogen receptor negative breast cancer according to PDI, hPDI, uPDI with
different latency periods using pooled data from NHS and NHSII.
eTable 1. Risk of ER-negative breast cancer according to quintiles of food
categories in pooled data from NHS and NHSII.
Healthful plant food Unhealthful plant
Animal food group
group food group
ER – breast Cancer (n=1,750)
Quintile 1 1.00 (ref) 1.00 (ref) 1.00 (ref)
Quintile 2 1.02 (0.88, 1.18) 0.90 (0.77, 1.06) 1.15 (0.98, 1.35)
Quintile 3 0.99 (0.85, 1.16) 0.96 (0.81, 1.14) 1.03 (0.86, 1.22)
Quintile 4 0.84 (0.71, 0.99) 0.98 (0.81, 1.18) 1.08 (0.89, 1.30)
Quintile 5 0.74 (0.61, 0.88) 0.94 (0.76, 1.16) 1.11 (0.90, 1.36)
p trend <0.01 0.83 0.54
Model stratified by age in months, cohort and calendar year, adjusted for race (Non-Hispanic Caucasian, African,
Asian, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45,
45-49, 50-52, 53+), postmenopausal hormone use (never user, past user, current user– estrogen only for <5 years,
current user – estrogen only for ≥5 years, current estrogen + progestin user for < 5 years, current estrogen + progestin
user for ≥5 years, current user of other types), oral contraceptive use history (never, ever), parity and age at first birth
(nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of
age), breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer
quintiles), physical activity (linear MET-hours/week), body mass index at age 18 years (<20.0, 20.0-21.9, 22.0-23.9,
24.0-26.9, ≥27.0) and socioeconomic status (continuous).
The 3 food categories (healthy and less healthy plant foods, and animal foods) were simultaneously included in the
same model.
eTable 2. Risk of estrogen receptor negative breast cancer according to PDI, hPDI, uPDI with different
latency periods using pooled data from NHS and NHSII.
Simple update
4-8 y lag 8-12 y lag 12-16 y lag 16-20 y lag
(0-4 y lag)
Number of
cases (n) / 1,542/3,955,108 1,296/3,314,002 1,051/2,728,313 777/2,142,682 484/1,560,331
person-years
Plant-based diet index
Quintile 1 1.00 1.00 1.00 1.00 1.00
Quintile 2 1.10 (0.94, 1.29) 0.93 (0.78, 1.11) 0.85 (0.70, 1.03) 0.89 (0.71, 1.11) 0.95 (0.72, 1.25)
Quintile 3 0.93 (0.79, 1.10) 0.97 (0.81, 1.15) 0.87 (0.72, 1.06) 1.03 (0.82, 1.28) 0.88 (0.66, 1.18)
Quintile 4 0.99 (0.83, 1.16) 0.91 (0.76, 1.09) 0.85 (0.69, 1.03) 0.93 (0.73, 1.17) 0.84 (0.62, 1.14)
Quintile 5 0.82 (0.69, 0.99) 0.92 (0.76, 1.11) 0.88 (0.72, 1.09) 0.90 (0.70, 1.15) 0.97 (0.71, 1.31)
p trend 0.02 0.35 0.28 0.52 0.63
Healthful plant-based diet index
Quintile 1 1.00 1.00 1.00 1.00 1.00
Quintile 2 1.07 (0.91, 1.25) 1.06 (0.89, 1.26) 1.08 (0.89, 1.31) 0.90 (0.72, 1.12) 0.94 (0.70, 1.26)
Quintile 3 0.98 (0.84, 1.15) 0.95 (0.80, 1.13) 1.04 (0.86, 1.26) 0.77 (0.61, 0.96) 0.94 (0.70, 1.25)
Quintile 4 0.89 (0.76, 1.05) 0.94 (0.79, 1.12) 0.94 (0.77, 1.15) 0.80 (0.63, 1.01) 1.00 (0.74, 1.35)
Quintile 5 0.85 (0.72, 1.01) 0.80 (0.66, 0.96) 0.91 (0.74, 1.12) 0.88 (0.70, 1.11) 0.98 (0.72, 1.32)
p trend 0.01 0.01 0.20 0.18 0.99
Unhealthful plant-based diet index
Quintile 1 1.00 1.00 1.00 1.00 1.00
Quintile 2 0.99 (0.84, 1.17) 1.04 (0.87, 1.25) 1.09 (0.89, 1.33) 1.11 (0.88, 1.40) 1.17 (0.87, 1.58)
Quintile 3 1.12 (0.95, 1.31) 1.19 (0.99, 1.42) 1.29 (1.06, 1.57) 1.04 (0.82, 1.31) 1.30 (0.97, 1.73)
Quintile 4 1.03 (0.87, 1.22) 1.19 (0.99, 1.43) 1.08 (0.88, 1.33) 1.06 (0.84, 1.35) 0.95 (0.69, 1.30)
Quintile 5 1.01 (0.85, 1.21) 1.22 (1.01, 1.48) 1.20 (0.97, 1.49) 1.29 (1.01, 1.64) 1.17 (0.85, 1.60)
p trend 0.80 0.02 0.13 0.07 0.68
All values are hazard ratios with 95% CIs in parentheses, stratified by age in months, cohort and calendar year, adjusted for race (Non-Hispanic
Caucasian, African, Asian, Hispanic Caucasian), age at menarche (<12, 12, 13, 14, >14 years), age at menopause (premenopausal, <45, 45-49, 50-52,
53+), postmenopausal hormone use (never user, past user, current user– estrogen only for <5 years, current user – estrogen only for ≥5 years, current
estrogen + progestin user for < 5 years, current estrogen + progestin user for ≥5 years, current user of other types), oral contraceptive use history (never,
ever), parity and age at first birth (nulliparous, 1 child before age 25, 1 child at ≥25 years of age, 2+ children before age 25, 2+ children ≥25 years of age),
breastfeeding history (never, breastfed for ≤ 6 months, breastfed for > 6 months), family history of breast cancer (yes or no), history of benign breast
disease (yes or no), height (<1.60, 1.60-1.64, 1.65-1.69, 1.70-1.74, 1.75 + m), cumulatively updated alcohol intake (0, <5, 5-9, 10-14, 15+ g/day),
cumulatively updated total caloric intake (kcal/day, quintiles), physical activity (linear MET-hours/week), body mass index at age 18 years (<20.0, 20.0-
21.9, 22.0-23.9, 24.0-26.9, ≥27.0) and socioeconomic status (continuous).
2. SUN Project questionnaires
C-8 : QUINTO CUESTIONARIO PROYECTO SUN
Fecha de Nacimiento Peso actual (Kg.)
Ninguno 1 2 3 4 5 6 7 8 9 10 ó más
¿Cúantos hijos e hijas tienes?
Si tienes miopía, hipermetropía/vista cansada o astigmatismo, escribe el nº de dioptrías que tienes
Miopía Hipermetropía/Vista cansada Astigmatismo
Ojo Derecho Ojo Izquierdo Ojo Derecho Ojo Izquierdo Ojo Derecho Ojo Izquierdo ID
0 0 0 0 0 0
1 1 1 1 1 1
En los últimos 2 años, ¿has perdido peso intencionadamente?
2 2 2 2 2 2
No he cambiado He perdido 1-2 kg. He perdido 3-4 kg. He perdido 5-10 kg.
3 3 3 3 3 3
He perdido más de 10 kg.
4 4 4 4 4 4
En los últimos 2 años, ¿has perdido peso NO voluntariamente (Ej.: debido a la enfermedad, estrés o depresión)?
5 5 5 5 5 5
No he perdido He perdido 1-2 kg. He perdido 3-4 kg. He perdido 5-10 kg.
6 6 6 6 6 6
He perdido más de 10 kg.
7 7 7 7 7 7
Nunca Al mes A la semana Al día
o casi
1-3 1 2-4 5-6 1 2-3 4-6 6+
8 8 8 8 8 8
nunca
9 9 9 9 9 9
¿Con qué frecuencia haces comidas fuera de casa?
¿Con qué frecuencia estas comidas son en lugares
de fast-food (pizzerías, hamburgueserías)?
¿Cuánto tiempo por término medio dedicas al día a trabajo de cerca (lectura, estudio, ...) un día típico de trabajo entre semana?
Nunca 30 min 30-60 min. 1 hora 2 horas 3 horas 4 horas 5 horas 6 horas 7 horas 8 horas 9 ó más
¿Cuánto tiempo por término medio dedicas al día a trabajo de cerca (lectura, estudio, ...) un día típico de fin de semana?
Nunca 30 min 30-60 min. 1 hora 2 horas 3 horas 4 horas 5 horas 6 horas 7 horas 8 horas 9 ó más
¿Has fumado algún cigarrillo en las últimas 4 semanas? No, nunca he fumado
No, dejé de fumar desde hace... < 1 año 1-2 años 3-5 años 6-9 años 10+ años
Sí, fumo desde hace... < 1 año 1-2 años 3-5 años 6-9 años 10+ años
Nº de cigarrillos/día... Ocasionalmente 1-4 5-14 15-24 25-34 35-44 45+
De modo habitual…
¿Cuántos kms. viajas en coche o moto al año? < 1.500 1.501-5.000 5.001-10.000 10.001-20.000 >20.000
Cuando vas en coche ¿usas el cinturón de seguridad? Nunca Casi nunca No siempre Siempre
Cuando vas en moto ¿usas el casco? Nunca A veces Siempre No voy en moto
Cuando vas en bicicleta ¿usas el casco? Nunca A veces Siempre No voy en bicicleta
(Dobla por esta línea)

Desde el último cuestionario que nos contestaste ¿has sido diagnosticado por un médico por PRIMERA VEZ de alguna de estas
condiciones o has pasado por alguna de las siguientes circunstancias? (recuerda que tienes la fecha de respuesta del último
cuestionario que nos enviaste en la etiqueta pegada junto al nº de identificación en la parte superior derecha de la primera cara.
Así pues, la fecha que escribas en las siguientes preguntas debe ser posterior a la que te indicamos):
Fecha aproximada Fecha aproximada
NO SÍ diagnóstico NO SÍ diagnóstico
Mes Año Mes Año
Accidente de tráfico con hospitalización de más de 24 horas Operación de cataratas
Otro accidente de tráfico sin hospitalización, pero con baja laboral Diagnóstico de cataratas sin operación
Accidente deportivo con lesión (que requirió tratamiento médico) Aumento de más de media dioptría en la miopía
Otro tipo de accidente con lesión, incluyendo cualquier fractura Glaucoma
Hipertensión arterial (más de 8,5 de mínima o más de 13 de Degeneración macular de retina

máxima) Bronquitis crónica o enfisema
Osteoporosis Asma
Colesterol alto (más de 200 mg/dl) Ulcera gástrica o duodenal
Infarto de miocardio Cólico nefrítico
Angina de pecho Cálculos en la vesícula
Cirugía coronaria (“by-pass”) Anorexia nerviosa o bulimia
Fibrilación auricular Diagnóstico de ansiedad
Aneurisma de aorta Diagnóstico médico de depresión
Insuficiencia cardiaca Pólipos en colon o recto
Trombosis venosa periférica Tumor maligno (señalar el tipo)
Embolia pulmonar
Diabetes (excluye diabetes gestacional) Tumor de
Diabetes en el embarazo (gestacional)
Accidente cerebral vascular (trombosis, embolia o hemorragia) Otra enfermedad (incluye tumores benignos)
Claudicación intermitente (insuficiencia arterial periférica)
Embarazo (indica fecha parto o fecha prevista de parto)
(especificar)
Consulta al médico por dificultad para lograr el embarazo
En general, dirías que tu salud es:
Excelente Muy buena Buena Regular Mala
¿Cómo dirías que es tu salud actual, comparada con la de hace 2 años, cuando nos respondiste por última vez?
Mucho mejor que antes Algo mejor que antes Más o menos igual Algo peor ahora Mucho peor ahora
© Departamento de Medicina Preventiva y Salud Pública UNIVERSIDAD DE NAVARRA

Sí, me limita Sí, me limita No, no me
Las siguientes preguntas se refieren a actividades o cosas que podrías hacer en un día normal. Tu salud… mucho un poco limita nada
¿Te limita para hacer esfuerzos intensos tales como correr, levantar objetos pesados o participar en deportes agotadores?
¿Te limita para hacer esfuerzos moderados como mover una mesa, pasar la aspiradora, jugar a bolos o caminar más de una hora?
¿Te limita para coger o llevar la bolsa de la compra?
¿Te limita para subir varios pisos por la escalera?
¿Te limita para subir un solo piso por la escalera?
¿Te limita para agacharte o arrodillarte?
¿Te limita para caminar un kilómetro o más?
¿Te limita para caminar varias manzanas (varios centenares de metros)?
¿Te limita para caminar una sola manzana (unos 100 metros)?
¿Te limita para bañarte o vestirte por ti mismo?
Durante las últimas 4 semanas… Sí No

¿Tuviste que reducir el tiempo dedicado al trabajo o a tus actividades cotidianas, a causa de tu salud física?
¿Hiciste menos de lo que hubieras querido hacer, a causa de tu salud física?
¿Tuviste que dejar de hacer algunas tareas en tu trabajo o actividades cotidianas, a causa de tu salud física?
¿Tuviste dificultad para hacer tu trabajo o tus actividades cotidianas (por ejemplo, te costó más de lo normal), a causa de tu salud física?
¿Tuviste que reducir el tiempo dedicado al trabajo o a tus actividades cotidianas, a causa de algún problema emocional (como estar triste, deprimido o nervioso?
¿Hiciste menos de lo que hubieras querido hacer, a causa de algún problema emocional (como estar triste, deprimido o nervioso)?
¿No hiciste tu trabajo o actividades cotidianas tan cuidadosamente como de costumbre, a causa de algún problema emocional (como estar triste, deprimido o nervioso)?
Durante las últimas 4 semanas… Nada Un poco Regular Bastante Mucho
¿Hasta qué punto tu salud física o los problemas emocionales han dificultado tus actividades sociales habituales
con la familia, los amigos, los vecinos u otras personas?
¿Tuviste dolor en alguna parte del cuerpo durante las últimas 4 semanas?
No, ninguno Sí, muy poco Sí, un poco Sí, moderado Sí, mucho Sí, muchísimo
¿Hasta qué punto el dolor te ha dificultado tu trabajo habitual? (incluido el trabajo fuera de casa y las tareas domésticas)
Nada Un poco Regular Bastante Mucho
(Dobla por esta línea)
Casi Muchas Algunas Sólo alguna

Durante las 4 últimas semanas… Siempre siempre veces veces vez Nunca
¿Cuánto tiempo te sentiste lleno de vitalidad?

¿Cuánto tiempo estuviste muy nervioso?
¿Cuánto tiempo te sentiste tan bajo de moral que nada podía animarte?
¿Cuánto tiempo te sentiste calmado y tranquilo?
¿Cuánto tiempo tuviste mucha energía?
¿Cuánto tiempo te sentiste desanimado y triste?
¿Cuánto tiempo te sentiste agotado?
¿Cuánto tiempo te sentiste feliz?
¿Cuánto tiempo te sentiste cansado?
Totalmente Bastante Bastante Totalmente

Por favor, di si te parece cierta o falsa cada una de las siguientes frases: cierta cierta Nunca falsa falsa
Creo que me pongo enfermo más facilmente que otras personas

Me considero una persona con fuerza de voluntad
Estoy tan sano como cualquiera
Creo que mi salud va a empeorar
Mi salud es excelente
Durante las últimas 4 semanas. ¿con qué frecuencia la salud física o los problemas emocionales te han dificultado tus actividades sociales?
(como visitar a los amigos o familiares)
Siempre Casi siempre Algunas veces Solo algunas veces Nunca
Por favor, indica a continuación los datos de las últimas mediciones que te hayan hecho y que tengan menos de 2 años de antigüedad:
Perímetro AZUCAR
cintura TENSIÓN ARTERIAL Glucemia basal LÍPIDOS
(Cm.) (mmHg) (mg/dl) (mg/dl)
Alta Baja Colesterol
Glucosa LDL HDL Triglicéridos
(sistólica) (diastólica) Total

CUESTIONARIO C_10 PROYECTO SUN
INSTRUCCIONES
Rellena exclusivamente a lápiz No arrugues la hoja Marca así
Borra completamente las marcas erróneas Marca correctamente las casillas Así no marques
No escribas sobre el margen derecho Marca una sola casilla por respuesta usa lápiz
Fecha de Nacimiento Peso (Kg.) Talla (Cm.) Estado civil

ID
soltero/a
DÍA MES AÑO
casado/a
0 0 0 0 0 0
viudo/a
1 1 1 1 1 1 0 0 0 0 0 0
separado/a
0 0 0 0 0 0 2 2 2 2 2 1 1 1 1 1 1
otros
1 1 1 1 1 1 3 3 3 3 2 2 2 2 2 2
2 2 2 2 2 4 4 4 4 3 3 3 3 3 3
3 3 3 3 3 5 5 5 5 4 4 4 4 4 4
4 4 4 4 6 6 6 6 5 5 5 5 5 5
5 5 5 5 7 7 7 7 6 6 6 6 6 6
6 6 6 6 8 8 8 8 7 7 7 7 7 7
7 7 7 7 9 9 9 9 8 8 8 8 8 8
8 8 8 8 9 9 9 9 9 9
9 9 9 9
Desde el último cuestionario, marca si has sido diagnosticado/a por un médico de alguna de las siguientes enfermedades o has pasado por
alguna de las siguientes circunstancias:
Fecha Fecha
aproximada aproximada
diagnóstico diagnóstico
NO SÍ Mes Año NO SÍ Mes Año
Accidente de tráfico con hospitalización de más de 24 h. Operación de cataratas
Otro accidente de tráfico sin hospitalización, con baja laboral Diagnóstico de cataratas sin operación
Hipertensión arterial ( > 8,5 de mínima ó > 13 de máxima) Degeneración macular de retina
Osteoporosis Bronquitis crónica o enfisema
Colesterol alto (más de 200 mg/dl) Asma
Infarto de miocardio Úlcera gástrica o duodenal
Angina de pecho Cólico nefrítico
Cirugía coronaria (“by-pass”) Cálculos en la vesícula
Fibrilación auricular Anorexia nerviosa o bulimia
Aneurisma de aorta Diagnóstico de ansiedad
Insuficiencia cardíaca Diagnóstico médico de depresión
Embolia pulmonar Pólipos en colon o recto
Diabetes (excluye diabetes gestacional) Tumor de
Diabetes gestacional
Accidente cerebral vascular (trombosis, embolia o hemorragia)
Claudicación intermitente (insuficiencia arterial periférica) Otra enfermedad o circunstancia
Embarazo (indica fecha parto o fecha prevista de parto) (incluye tumores benignos) (especificar)
Consulta al médico por dificultad para lograr embarazo
FIN
¡GRACIAS POR SEGUIR PARTICIPANDO!
UNIVERSIDAD DE NAVARRA Departamento de Medicina Preventiva y Salud Pública

POR FAVOR, NO RELLENES ESTA PARTE DEL CUESTIONARIO, ES SÓLO PARA FINES DE CODIFICACIÓN.
ENFERMEDAD O CIRCUNSTANCIA MES AÑO
Accidente de tráfico con hospitalización de más de 24 h.
ENE FEB MAR ABR MAY JUN JUL AGO SEP OCT NOV DIC 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025
Otro accidente de tráfico sin hospitalización, con baja laboral

Accidente deportivo con lesión (que requirió tratamiento médico)

Otro tipo de accidente con lesión, incluyendo cualquier fractura

Hipertensión arterial ( > 8,5 de mínima ó > 13 de máxima)

Osteoporosis
Colesterol alto (más de 200 mg/dl)
Infarto de miocardio
Angina de pecho
Cirugía coronaria (“by-pass”)
Fibrilación auricular
Aneurisma de aorta
Insuficiencia cardíaca
Embolia pulmonar
Trombosis venosa periférica
Diabetes (excluye diabetes gestacional)




Operación de cataratas ENE FEB MAR ABR MAY JUN JUL AGO SEP OCT NOV DIC 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025
Diagnóstico de cataratas sin operación

Aumento de más de media dioptría en la miopía

Glaucoma
Degeneración macular de retina
Bronquitis crónica o enfisema

Asma
Úlcera gástrica o duodenal
Cólico nefrítico
Cálculos en la vesícula
Anorexia nerviosa o bulimia
Diagnóstico de ansiedad
Diagnóstico médico de depresión

Pólipos en colon o recto

Tumor maligno
Otra enfermedad o circunstancia (incluye tumores benignos)
Código tipo de tumor Código otra enfermedad o circunstancia

0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9
Fecha de Nacimiento Peso (Kg.) ID

INSTRUCCIONES
DÍA MES AÑO
Rellena exclusivamente a lápiz 0 0 0 0 0 0 0 0 0

Borra completamente las marcas erróneas 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1
No escribas sobre el margen derecho 1 1 1 1 1 1 2 2 2 2 2 2 2 2
No arrugues la hoja 2 2 2 2 2 3 3 3 3 3 3 3 3
Marca correctamente las casillas 3 3 3 3 3 4 4 4 4 4 4 4 4
Marca una sola casilla por respuesta 4 4 4 4 5 5 5 5 5 5 5 5
5 5 5 5 6 6 6 6 6 6 6 6
usa lápiz
6 6 6 6 7 7 7 7 7 7 7 7
7 7 7 7 8 8 8 8 8 8 8 8
Marca así
Así no marques 8 8 8 8 9 9 9 9 9 9 9 9
9 9 9 9
Consumo medio durante el año pasado

Nunca Al mes A la semana Al día
o casi
nunca 1-3 1 2-4 5-6 1 2-3 4-6 6+
¿Con qué frecuencia comes en lugares de fast-food (pizzerías, hamburgueserías)?
¿Con qué frecuencia tomas bebidas carbonatadas azucaradas (refrescos, colas, tónicas,
bitter...) excluyendo las "light" (1 lata, 33cl.)?
¿Con qué frecuencia tomas bebidas "light" o “zero” (1 lata, 33cl.)?
¿Con qué frecuencia tomas zumos envasados o embotellados ?
Desde el último cuestionario, marca si has sido diagnosticado/a por un médico de alguna de las siguientes enfermedades o circunstancias (anota
la fecha aproximada del diagnóstico):
Otro accidente de tráfico sin hospitalización, con baja laboral Diagnóstico de cataratas sin operación
Hipertensión arterial ( > 8,5 de mínima ó > 13 de máxima) Degeneración macular de retina
Osteoporosis Bronquitis crónica o enfisema
Colesterol alto (más de 200 mg/dl) Asma
Infarto de miocardio Úlcera gástrica o duodenal
Angina de pecho Cólico nefrítico
Cirugía coronaria (“by-pass”) Cálculos en la vesícula
Fibrilación auricular Anorexia nerviosa o bulimia
Aneurisma de aorta Diagnóstico de ansiedad
Insuficiencia cardíaca Diagnóstico médico de depresión
Embolia pulmonar Pólipos en colon o recto
Claudicación intermitente (insuficiencia arterial periférica) Otra enfermedad o circunstancia
Embarazo (indica fecha parto o fecha prevista de parto) (incluye tumores benignos) (especificar)
Indica si has sido diagnosticado/a alguna vez por un médico de alguna de las siguientes enfermedades o circunstancias: (anota la fecha aproximada
del diagnóstico)
Pérdida de memoria Reuma
Demencia Artrosis
Alzheimer Artritis
Parkinson
FIN
¡GRACIAS POR SEGUIR PARTICIPANDO!
UNIVERSIDAD DE NAVARRA Departamento de Medicina Preventiva y Salud Pública
POR FAVOR, NO RELLENES ESTA PARTE DEL CUESTIONARIO, ES SÓLO PARA FINES DE CODIFICACIÓN.
ENFERMEDAD O CIRCUNSTANCIA MES AÑO

ENE FEB MAR ABR MAY JUN JUL AGO SEP OCT NOV DIC 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Accidente de tráfico con hospitalización
Otro accidente de tráfico sin hospitalización
Accidente deportivo con lesión
Otro tipo de accidente con lesión
Hipertensión arterial
Osteoporosis
Colesterol alto (más de 200 mg/dl)
Infarto de miocardio
Angina de pecho
Cirugía coronaria (“by-pass”)
Fibrilación auricular
Aneurisma de aorta
Insuficiencia cardíaca
Embolia pulmonar
Trombosis venosa periférica
Accidente cerebral vascular
Claudicación intermitente
Embarazo
Consulta dificultad embarazo
Operación de cataratas
Diagnóstico de cataratas sin operación
Aumento de más de media dioptría en la miopía ENE FEB MAR ABR MAY JUN JUL AGO SEP OCT NOV DIC 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Glaucoma
Degeneración macular de retina
Bronquitis crónica o enfisema
Asma
Úlcera gástrica o duodenal
Cólico nefrítico
Cálculos en la vesícula
Anorexia nerviosa o bulimia
Diagnóstico de ansiedad
Diagnóstico médico de depresión
Pólipos en colon o recto
Tumor maligno
Otra enfermedad o circunstancia
Código tipo de tumor Código otra enfermedad o circunstancia

0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9
Pérdida de
memoria Demencia Alzheimer Parkinson Reuma Artrosis Artritis
MES AÑO MES AÑO MES AÑO MES AÑO MES AÑO MES AÑO MES AÑO
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4
5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6
7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7
8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8
9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9
Fecha nacimiento Peso actual

Día Mes Año (Kg.) Soltero/a Casado/a Viudo/a Separado/a Otros
Estado civil
ID
0 0 0 0 0 0 0 0 0 Ninguno 1 2 3 4 5 6 7 8 9 10 ó más 0 0 0 0 0 0
Nº de hijos
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
2 2 2 2 2 2 2 2 ¿Has fumado algún cigarrillo en las últimas 4 semanas? 2 2 2 2 2 2
3 3 3 3 3 3 3 3 3 3 3 3 3
No, nunca he fumado
4 4 4 4 4 4 4 4 4 4 4 4
5 5 5 5 5 5 < 1 año 1-2 años 3-5 6-9 10 ó más 5 5 5 5 5 5
No, dejé de fumar desde hace...
6 6 6 6 6 6 6 6 6 6 6 6
7 7 7 7 7 7 < 1 año 1-2 años 3-5 6-9 10 ó más 7 7 7 7 7 7
Sí, fumo desde hace...
8 8 8 8 8 8 8 8 8 8 8 8
Ocasio-
9 9 9 9 9 9 nalmente 1-4 5-14 15-24 25-34 35-44 45+ 9 9 9 9 9 9
Nº de cigarrillos/día...
¿Haces ejercicio? No (pasa a la siguiente pregunta) Sí
FRECUENCIA MEDIA DURANTE LA SEMANA

¿Cuánto tiempo por término medio dedicaste MESES AL AÑO
MINUTOS / SEMANA HORAS / SEMANA
a las siguientes actividades en el último año? NUNCA 1-4 5 - 19 20 - 59 1 - 1,5 2-3 4-6 7 - 10 ≥ 11 <3 3-6 >6
Andar o pasear fuera de casa (incluye golf)

Correr o hacer jogging despacio
Correr más competitivo y rápido (atletismo, etc.)
Pasear en bicicleta
Bicicleta estática
Nadar
Tenis, frontón, squash, otros de raqueta o pala
Fútbol, futbito
Otros de equipo (baloncesto, balonmano…)
Baile, danza, aerobic
Excursiones al monte, escalada
Gimnasia
Cuidado del jardín y/o piscina, bricolaje, etc.
Esquí, patinaje
Judo, karate u otras artes marciales
Vela
Otras actividades físicas-deporte no mencionadas
¿Qué medicación o suplementos dietéticos (incluyendo vitaminas, fibras, salvado, etc.) consumes de modo HABITUAL?:
No consumo medicación o suplementos, salvo muy de vez en cuando

Consumo a diario la siguiente medicación o suplementos (si necesitas más espacio, añade una hoja por favor)
FÁRMACO Nº veces al día Y DOSIS FÁRMACO Nº veces al día Y DOSIS
1. 4.
2. 5.
3. 6.
CONSUMO MEDIO DURANTE EL AÑO PASADO

NUNCA AL MES A LA SEMANA AL DÍA
Por favor, marca una única opción para cada alimento. O CASI
NUNCA 1-3 1 2-4 5-6 1 2-3 4-6 6+
Mantequilla, margarina o nata (1 ración: 12 g.)

Aceite de oliva para cocinar, aliñar ensaladas o tomar con el pan (1 ración: 1 cucharada)
Sofrito (salsa de tomate con ajo, cebolla) cocinados con aceite de oliva
Verduras u hortalizas (1 ración 200 g.)
Legumbres (lentejas, garbanzos, alubias, etc –1 ración: 60 g. en seco)
Azúcar que añades a las bebidas (café, té) o alimentos (yogur, cuajada), 1 cucharadita de postre
Ajo
Especias
Arroz integral (60 g. en crudo)
Pasta integral (60 g. en crudo)
SUMCO 00000-13
Cereales integrales: muesli, copos avena, tipo all-bran (30 g.)

Piezas de fruta (1 unidad)
Frutos secos (1 ración: 30 g.)

Continua de la página anterior CONSUMO MEDIO DURANTE EL AÑO PASADO
NUNCA AL MES A LA SEMANA AL DÍA
Por favor, marca una única opción para cada alimento. O CASI
NUNCA 1-3 1 2-4 5-6 1 2-3 4-6 6+
Carnes rojas, salchichas o embutidos (1 ración 100-150 g.)

Carne de pollo, pavo, conejo (1 ración 100-150 g.)
Pescado-mariscos (1 rac. de pescado: 100-150 g.; 1 rac. de mariscos: 4-5 piezas ó 200 g.)
Alimentos precocinados (1 ración)
Repostería comercial (no casera), incluyendo galletas, pasteles, flanes y postres dulces
Zumo natural (1 vaso)
Bebidas carbonatadas y/o azucaradas (refrescos, colas, tónicas, bitter) (1 lata, 33 cl.)
Bebidas “light” o “zero” (1 lata, 33 cl.)
Bebidas energéticas (Red Bull, Burn, etc.) (1 lata, 25 cl.)
Cerveza (1 lata, 33 cl.)
Cerveza sin alcohol (1 lata, 33 cl.)
Vino tinto (1 vaso, 100 ml.)
Otro tipo de vino (1 vaso, 100 ml.)
Vino en las comidas (1 vaso, 100 ml.)
Licores destilados: whisky, ginebra, coñac, anís (1 copa, 50 ml.)
Por término medio en una semana típica, ¿cuántos días/semana bebes alcohol (vino, cerveza o licores destilados), incluyendo el que tomas en las comidas?
Nunca o casi nunca 1 2 3 4 5 6 7
¿Cuántos días/semana bebes vino en la comida?
Nunca o casi nunca 1 2 3 4 5 6 7
¿Cuál fue el máximo número de bebidas alcohólicas (sumado vino, cerveza y licor) que tomaste un día entre semana, un día de fin de semana o un
día especial (celebración, boda, festividad)?
Ninguna 1-2 3-5 6-9 10-14 15 ó más
¿Qué porcentaje del aceite de oliva que consumes es virgen extra? No consumo < 25% 25-50% 51-75 % > 75%
¿Reutilizas el aceite de fritura? Nunca A veces Siempre
¿Cuántas veces reutilizas el mismo aceite? 1-2 3-5 6 ó más
Desde el último cuestionario que nos contestaste ¿Has sido diagnosticado por un médico por PRIMERA VEZ de alguna de estas condiciones? o ¿Has
pasado por alguna de las siguientes circunstancias? (recuerda que tienes la fecha de respuesta del último cuestionario que nos enviaste en la etiqueta
pegada junto al nº de identificación en la parte superior derecha de la primera cara. Así pues, la fecha que escribas en las siguientes preguntas debe
ser posterior a la que te indicamos):
Año diagnóstico Año diagnóstico
Accidente de tráfico con hospitalización de más de 24 horas Degeneración macular de retina
Otro accidente de tráfico sin hospitalización, pero con baja laboral Bronquitis crónica o enfisema
Accidente deportivo con lesión (que requirió tratamiento médico) Asma
Otro tipo de accidente con lesión, incluyendo cualquier fractura Ulcera gástrica o duodenal
Hipertensión arterial ( más de 8,5 de mínima o más de 13 de máxima) Cólico nefrítico
Osteoporosis Cálculos en la vesícula
Colesterol alto (más de 240 mg/dl) Anorexia nerviosa o bulimia
Infarto de miocardio Diagnóstico de ansiedad
Angina de pecho Diagnóstico médico de depresión
Cirugía coronaria (“by-pass”) Pólipos en colon o recto
Fibrilación auricular Pérdida de memoria
Aneurisma de aorta Demencia
Insuficiencia cardiaca Alzheimer
Embolia pulmonar Parkinson
Trombosis venosa periférica Reuma
Diabetes (excluye diabetes gestacional) Artrosis
Diabetes gestacional Artritis
Accidente cerebral vascular (trombosis, embolia o hemorragia) ¿Has sido diagnosticado por un médico
Claudicación intermitente (insuficiencia arterial periférica) de pérdida de memoria?
Embarazo (indica fecha parto o fecha prevista de parto) Tumor maligno
Consulta al médico por dificultad para lograr embarazo Otra enfermedad (incluye tumores benignos)
Diagnóstico de cataratas sin operación (especificar)
Glaucoma
SUMCO 00000-13 (Ret.)
¿Has sido diagnosticado/a por un médico de enfermedad periodontal o gingivitis? No Sí Mes: Año:
¿Has notado últimamente pérdida de memoria? No Sí Mes: Año:
¡GRACIAS POR TU COLABORACIÓN!

4. Peso mínimo 5. Peso máximo
3. Peso en los 2 últimos en los 2 últimos
2. Fecha nacimiento actual (kg.) años (kg.) años (kg.)
1. ID
0 0 0 0 0 0
6. ¿Has perdido peso intencionadamente? Sí (indica la pérdida y el número de veces) No (pasa a Pregunta 8)
1 1 1 1 1 1
Pérdida de peso Nº de veces
2 2 2 2 2 2
1-2 kg. 0 1-2 3-4 5-6 7 ó más
3 3 3 3 3 3
3-4 kg. 0 1-2 3-4 5-6 7 ó más
4 4 4 4 4 4
5-10 kg. 0 1-2 3-4 5-6 7 ó más
5 5 5 5 5 5
> 10 kg. 0 1-2 3-4 5-6 7 ó más
6 6 6 6 6 6
7. ¿Qué tipo de método usaste? (puedes elegir varias opciones)
7 7 7 7 7 7
No he perdido peso intencionadamente Aumento de ejercicio físico Saltarte comidas/ayuno
8 8 8 8 8 8
Pérdida involuntaria (enfermedad, estrés, depresión) Cirugía bariátrica/bypass gástrico Medicación
9 9 9 9 9 9
Restricción energética (dieta hipocalórica) Balón gástrico Otros
8. ¿Recibiste lactancia materna?
No Sí (meses) 9. ¿Conoces tu peso al nacer? (En gramos)
10. ¿Has fumado algún cigarrillo en las últimas 4 semanas?
< 1 año 1-2 años 3-5 años 6-9 años 10 ó más años
No, nunca he fumado No, dejé de fumar hace…
Ocasio-
< 1 año 1-2 años 3-5 6-9 10 ó más nalmente 1-4 5-14 15-24 25-34 35-44 45+
Sí, fumo desde hace… Nº de cigarrillos/día...
11. ¿Cuál es el nivel más alto de estudios que has completado? 12. ¿Has terminado alguna de estas carreras?
No he acabado ninguna diplomatura ni licenciatura (pasa a pregunta 13) Medicina Enfermería
Doctorado Máster Farmacia Nutrición-Dietética
Licenciatura - Escuela técnica superior - Grado Biológicas Otra carrera biosanitaria
Diplomatura (ingeniería técnica - escuelas universitarias) Ninguna carrera biosanitaria Psicología
13. Tiempo medio de sueño en día típico entre semana 14. Tiempo medio de sueño en día típico de fin de semana
Dormir por la noche (horas) Dormir por la noche (horas)
Dormir la siesta (minutos) Dormir la siesta (minutos)
15. ¿Cuánto tardas en dormirte?
Tiempo que tardas en dormirte (minutos) Tiempo que tardas en dormirte si te despiertas por la noche (minutos)
En caso de necesidad dobla por esta línea
16. ¿Trabajas a turnos? No Sí Mañanas Tardes Noches Rotatorio Fines de semana Otros
17. ¿Qué red/es social/es utilizas? Facebook Twitter ResearchGate Academia Linkedin Otras Ninguna
18. ¿Qué tiempo semanal dedicas a las redes sociales? (horas/semana)
Se presentan debajo afirmaciones que uno podría hacer de sí mismo. Debes describirte a ti mismo tan sinceramente como puedas eligiendo una
de las 4 opciones siguientes: 0= Muy en desacuerdo; 1= Moderadamente en desacuerdo; 2= Moderadamente de acuerdo; 3= Muy de acuerdo
0 1 2 3
19. La gente me describiría como imprudente
0 1 2 3
20. Me parece que actúo totalmente por impulso
0 1 2 3
21. Aunque sabría hacerlo mejor, no puedo dejar de tomar decisiones precipitadas
0 1 2 3
22. A menudo siento que nada de lo que hago importa realmente
0 1 2 3
23. Los demás me ven como irresponsable
0 1 2 3
24. No soy muy bueno/a planificando con antelación
0 1 2 3
25. Con frecuencia mis pensamientos no tienen sentido para los demás
0 1 2 3
26. Me preocupo por casi todo
0 1 2 3
27. Me emociono fácilmente, a menudo por motivos muy pequeños
0 1 2 3
28. Estar solo/a en la vida me da más miedo que cualquier otra cosa
0 1 2 3
29. Me empeño en una manera de hacer las cosas, incluso cuando está claro que no funciona
0 1 2 3
30. He visto cosas que no estaban realmente allí
0 1 2 3
31. Me mantengo alejado/a de las relaciones sentimentales
0 1 2 3
32. No me interesa hacer amigos
0 1 2 3
33. Me irrito fácilmente por todo tipo de cosas
0 1 2 3
34. No me gusta intimar demasiado con los demás
0 1 2 3
35. No es tan importante si hiero los sentimientos de otras personas
0 1 2 3
36. Rara vez me entusiasmo con nada
0 1 2 3
37. Ansío llamar la atención
0 1 2 3
38. A menudo he de tratar con gente que es menos importante que yo
0 1 2 3
39. A menudo tengo pensamientos que tienen sentido para mí pero que otros ven extraños
0 1 2 3
40. Utilizo a la gente para conseguir lo que quiero
0 1 2 3
41. A menudo me quedo ensimismado/a y, cuando de pronto reacciono, veo que ha pasado un buen rato
0 1 2 3
42. Las cosas que me rodean a menudo me parecen irreales, o más reales de lo normal
0 1 2 3
43. Me resulta fácil aprovecharme de los demás
Certificado: 900000 ©Copyright 2015 Dara Informática SLU. Todos los derechos reservados. http://www.dara.es/omr

44. Desde el último cuestionario que nos contestaste ¿has sido diagnosticado por un médico por PRIMERA VEZ de alguna de estas condiciones?
o ¿has pasado por alguna de las siguientes circunstancias? (recuerda que tienes la fecha de respuesta del último cuestionario que nos enviaste
en la etiqueta pegada junto al nº de identificación en la parte superior derecha de la primera cara. Así pues, la fecha que escribas en las
siguientes preguntas debe ser posterior a la que te indicamos):
Año diagnóstico Año diagnóstico
Accidente de tráfico con hospitalización de más de 24 horas Bronquitis crónica o enfisema
Otro accidente de tráfico sin hospitalización, pero con baja laboral Asma
Accidente deportivo con lesión (que requirió tratamiento médico) Ulcera gástrica o duodenal
Otro tipo de accidente con lesión, incluyendo cualquier fractura Cólico nefrítico
Hipertensión arterial (más de 8,5 de mínima o más de 13 de máxima) Cálculos en la vesícula
Osteoporosis Anorexia nerviosa o bulimia
Colesterol alto (más de 240 mg/dl) Diagnóstico médico de ansiedad
Infarto de miocardio Diagnóstico médico de depresión
Angina de pecho Pólipos en colon o recto
Cirugía coronaria (“by-pass”) Pérdida de memoria
Fibrilación auricular Demencia
Aneurisma de aorta Alzheimer
Insuficiencia cardiaca Parkinson
Embolia pulmonar Reuma
Trombosis venosa periférica Artrosis
Diabetes (excluye diabetes gestacional) Artritis
Diabetes gestacional Artritis reumatoide
Accidente cerebral vascular (trombosis, embolia o hemorragia) Operaciones (cadera, rodilla, etc.)
Claudicación intermitente (insuficiencia arterial periférica) Gingivitis (sangrado de encías, inflamación)
Embarazo (indica fecha parto o fecha prevista de parto) Tumor maligno (especificar)
Diagnóstico de cataratas sin operación Otra enfermedad (incluye tumores benignos)
Glaucoma
Degeneración macular de retina (especificar)
45. ¿Has sido diagnosticado/a alguna vez de migrañas? No Sí Mes: Año:

La siguiente pregunta solo deben responderla las mujeres
46. ¿Te han desaparecido las reglas? No Sí ¿Cuándo? Mes: Año:
47. ¿Qué medicación o suplementos dietéticos (incluyendo vitaminas, fibras, salvado, etc.) consumes de modo HABITUAL?:
No consumo medicación o suplementos
Consumo a diario la siguiente medicación o suplementos (si necesitas más espacio, añade una hoja por favor)
FÁRMACO Nº veces al día Y DOSIS FÁRMACO Nº veces al día Y DOSIS
1. 4.
2. 5.
3. 6.
48. Alimentación (por favor, marca una única opción para cada alimento) CONSUMO MEDIO DURANTE EL AÑO PASADO
Para cada alimento, marca el recuadro que indica la frecuencia de consumo NUNCA AL MES A LA SEMANA AL DÍA
por término medio durante el año pasado. O CASI
Se asume el tamaño de una ración típica. NUNCA 1 - 3 1 2-4 5-6 1 2-3 4-6 6+
Pan integral (3 rebanadas, 60 g)

Arroz integral (ración de 60 g en seco)
Pasta integral (ración de 60 g en seco)
Cereales integrales de desayuno (muesli, avena, all-bran) (30 g)
Alimentos funcionales (leches enriquecidas, margarinas con fitoesteroles, cereales enriquecidos)
Bebidas carbonatadas y/o azucaradas (refrescos, colas, tónicas, bitter) excluyendo las bebidas “light” (1 lata: 33 cl.)
Refrescos light o zero (1 lata: 33 cl.)
Fruta (una pieza o ración)
Verduras y hortalizas (1 ración: 200 g)
¿Con qué frecuencia realizas comidas en lugares de fast-food (pizzerías, hamburgueserías)?
Azúcar que añades a bebidas o alimentos (1 cucharadita)
Edulcorantes no calóricos
Sal en las comidas (después de cocinar)
Certificado: 900000 ©Copyright 2015 Dara Informática SLU. Todos los derechos reservados. http://www.dara.es/omr
CUESTIONARIO BREVE. ESTUDIO SUN
1. Fecha de Nacimiento 2. Peso actual

Día Mes Año (Kg.)
ID
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
2 2 2 2 2 2 2 2 2 2 2 2 2
3 3 3 3 3 3 3 Fecha 3 3 3 3 3 3
4 4 4 4 4 4 4 4 4 4 4 4
5 5 5 5 5 5 5 5 5 5 5 5
6 6 6 6 6 6 6 6 6 6 6 6
7 7 7 7 7 7 7 7 7 7 7 7
8 8 8 8 8 8 8 8 8 8 8 8
9 9 9 9 9 9 9 9 9 9 9 9
3. Desde , marque si ha sido diagnosticado de alguna de las siguientes enfermedades:
Fecha Fecha
aproximada aproximada
diagnóstico diagnóstico
NO SÍ N O SÍ
Mes Año Mes Año
Otro accidente de tráfico sin hospitalización, pero con baja Diagnóstico de cataratas sin operación
laboral Aumento de más de media dioptría en la miopía
Accidente deportivo con lesión (que requirió tratamiento médico) Glaucoma
Otro tipo de accidente con lesión, incluyendo cualquier fractura Degeneración macular de retina
Hipertensión arterial ( más de 8,5 de mínima o más de 13 Bronquitis crónica o enfisema

de máxima) Asma
Osteoporosis Úlcera gástrica o duodenal
Colesterol alto (más de 240 mg/dl) Cólico nefrítico
Infarto de miocardio Cálculos en la vesícula
Angina de pecho Anorexia nerviosa o bulimia
Cirugía coronaria (“by-pass”) Diagnóstico de ansiedad
Fibrilación auricular Diagnóstico médico de depresión
Aneurisma de aorta Pólipos en colon o recto
Insuficiencia cardiaca Tumor maligno (señalar el tipo)
Embolia pulmonar
Trombosis venosa periférica Tumor de
Diabetes gestacional Otra enfermedad (incluye tumores benignos)
(especificar)
4. Si toma algún fármaco de manera habitual puede indicarlo a continuación:
Fármaco 1 F á r ma c o 2 Fárm aco 3
SUMCO 00000-06

GRACIAS POR SEGUIR PARTICIPANDO
Para más información:

http://www.unav.es/preventiva/sun
Para cualquier comentario o sugerencia dispone del siguiente recuadro:
SUMCO 00000-06 (Ret.)
Sólo si ha cambiado de dirección, debe indicarla en la hoja de color que le adjuntamos.

Si sigue teniendo la misma dirección solamente debe rellenar este documento.
3. Nurses’ Health Studies questionnaires
Archive of Nurses’ Health Studies questionnaires. The table below contains links to PDFs of all of
the questionnaires from the Nurses’ Health Study and Nurses’ Health Study II.
The standard questionnaire (includes questions on many aspects of lifestyle and on health status) is
available in a regular-size font (Long questionnaire) and in large print (Booklet version). The shorter
version of the questionnaire is primarily focused on health status (Short questionnaire).
Study Year Questionnaire format Special Study Year Questionnaire format Special
2016 Long Short 2017 Long
2014 Long Short 2015 Long Short
2012 Long Short Booklet 2013 Long Short
Nurses’ Health Study II

Nurses’ Health Study

2001 Mom 2001 Long Short Mom Stress
1998 Long Short Booklet 1998 Diet
1988 Long Short
1986 Long Short
1984 Long Short
1982 Long Short
1980 Long
Below you will find examples of the NHS questionnaires (1984,1986) and NHSII
questionnaires (1991,1993).
HARVARD MEDICAL SCHOOL NURSES' HEALTH STUDY
PLEASE REPLY TO:

Channing Laboratory
180 Longwood Ave.
Boston, Mass. 02115
617-732-2279
- -- Dear Colleague: ---

•
It is now eight years since you responded to the first Nurses' Health Study Questionnaire and
my co-investigators and I are most grateful for your continued cooperation and participation in this
major research project.
In recent analyses we compared risk factors for breast cancer in premenopausal women with those
among postmenopausal women, and found no important differences. Also, we found no overall
relationship between postmenopausal estrogens and risk of breast cancer.* Although these findings
are reassuring, continued data collection is necessary to observe possible health effects of these
preparations over longer periods of time. Soon we will complete analyses of postmenopausal
hormones and coronary heart disease and will include summaries of these and other results with our
1985 informational mailing to you.
Since 1982, more than 70,000 participants have sent toenail specimens. These have now been
cataloged and we will shortly begin analyzing selected specimens for trace element content. These
analyses will add a unique and invaluable dimension to the Nurses' Health Study.
This year's questionnaire includes a comprehensive diet assessment to update our information on
your food consumption. This is made possible through a special grant from The National Institutes
of Health recognizing both the importance of diet in health and the high quality of the dietary
information previously reported by participants in this study. Although the diet section involves some
extra effort, we hope that you will give it your most careful attention since future national dietary
recommendations will likely be based heavily on your accurate and complete responses.
This questionnaire again uses an optically scannable format to assure more efficient and accurate
processing of forms. It is therefore important to use an ordinary (No.2} pencil and to make no stray
--- marks on the form. If you wish to make additional comments, please write them on a separate sheet.
All information provided will remain confidential and be used for medical statistical purposes only.
During the past year many participants have rect'ived promotional mailings regarding diet and health
and asked whether we had released their names and addresses. I can assure you that we never
have and never will release your name and address to any other individual or group.
Through the Nurses' Health Study questionnaires we are learning about ways to prevent cancer,
heart disease and other important illnesses. Again, I am very grateful for your critical contribution
toward this goal.
Sincerely,
Frank E. Speizer, M.D.

Principal Investigator
•References: 1. A Prospective Cohort Study of Postmenopausal Female Hormone Use and Risk of Breast Cancer.
Am. J . Epidemiol. 1983; 118: 416
2. A Case-control Study of Risk Indicators Among Women With Premenopausal and Early Postmenopausal
Breast Cancer. Cancer 1984; 53: 1020- 4
Adv1sory Bonrd. Research Group:

Thelma M . Schorr, R.N. Rosa Bova Barbara Egan Laura Sampson, R.D.
President, American Journal of Nursing. Marilyn Browne, M .S. Ann Harvey, R.N. Frank E. Speizer. M.D.
Richard McKibben, Ph.D. Graham Colditz. M .D. Charles Hennekens. M .D. Meir Stampfer. M .D.
Policy Analyst Meryl Dannenberg, B.S Charlotte Petro Harry Taplin, M .A.
American Nurses· Association David Dysert Bernard Rosner, Ph .D. Walter Willett, M .D.
PLEASE REPLY TO:

Channing Laboratory
180 Longwood Ave.
Boston, Mass. 02115
617-732-2279
- -- Dear Colleague: ---

•
It is now eight years since you responded to the first Nurses' Health Study Questionnaire and
my co-investigators and I are most grateful for your continued cooperation and participation in this
major research project.
In recent analyses we compared risk factors for breast cancer in premenopausal women with those
among postmenopausal women, and found no important differences. Also, we found no overall
relationship between postmenopausal estrogens and risk of breast cancer.* Although these findings
are reassuring, continued data collection is necessary to observe possible health effects of these
preparations over longer periods of time. Soon we will complete analyses of postmenopausal
hormones and coronary heart disease and will include summaries of these and other results with our
1985 informational mailing to you.
Since 1982, more than 70,000 participants have sent toenail specimens. These have now been
cataloged and we will shortly begin analyzing selected specimens for trace element content. These
analyses will add a unique and invaluable dimension to the Nurses' Health Study.
This year's questionnaire includes a comprehensive diet assessment to update our information on
your food consumption. This is made possible through a special grant from The National Institutes
of Health recognizing both the importance of diet in health and the high quality of the dietary
information previously reported by participants in this study. Although the diet section involves some
extra effort, we hope that you will give it your most careful attention since future national dietary
recommendations will likely be based heavily on your accurate and complete responses.
This questionnaire again uses an optically scannable format to assure more efficient and accurate
processing of forms. It is therefore important to use an ordinary (No.2} pencil and to make no stray
--- marks on the form. If you wish to make additional comments, please write them on a separate sheet.
All information provided will remain confidential and be used for medical statistical purposes only.
During the past year many participants have rect'ived promotional mailings regarding diet and health
and asked whether we had released their names and addresses. I can assure you that we never
have and never will release your name and address to any other individual or group.
Through the Nurses' Health Study questionnaires we are learning about ways to prevent cancer,
heart disease and other important illnesses. Again, I am very grateful for your critical contribution
toward this goal.
Sincerely,
Frank E. Speizer, M.D.

•References: 1. A Prospective Cohort Study of Postmenopausal Female Hormone Use and Risk of Breast Cancer.
Am. J . Epidemiol. 1983; 118: 416
2. A Case-control Study of Risk Indicators Among Women With Premenopausal and Early Postmenopausal
Breast Cancer. Cancer 1984; 53: 1020- 4
Adv1sory Bonrd. Research Group:

Thelma M . Schorr, R.N. Rosa Bova Barbara Egan Laura Sampson, R.D.
President, American Journal of Nursing. Marilyn Browne, M .S. Ann Harvey, R.N. Frank E. Speizer. M.D.
Richard McKibben, Ph.D. Graham Colditz. M .D. Charles Hennekens. M .D. Meir Stampfer. M .D.
Policy Analyst Meryl Dannenberg, B.S Charlotte Petro Harry Taplin, M .A.
American Nurses· Association David Dysert Bernard Rosner, Ph .D. Walter Willett, M .D.
INSTRUCTIONS
Please use an ordinary pencil to answer all questions by completely

filling in the appropriate response circle, or by writing the requested
information if a space is provided. Note that some questions ask for
information since June 1982, some ask for current status, and some
ask about events over longer periods. Because this form is meant to
be read by optical-scanning equipment, it is important for you to
make no stray marks and to keep any write-in responses within the
provided spaces. Should you need to change a response, erase the
incorrect mark completely. If you have comments, please write them on
a separate piece of paper.
EXAMPLE 1: Do you live in the United States? 0 NO YES

t
Fill circle completely, do not mark this way:
I-
EXAMPLE 2: Where were you born? OR£60N
Keep hand-writing within borders of the response box.
EXAMPLE 3: DATE OF BIRTH AND CURRENT WEIGHT:
a) Write in birthdate and

11. DATE OF BIRTH: 12. CURRENT
weight i n the boxes at DAY WEIGHT:
MONTH YEAR and
the top of each grid. 0 s 0 9 2. I I 1/ 0 140
For exam pie, May 9
•0 0 •
@) ® ®® @@
• -
pounds:
1921 WO uld be
(i)
(1) 0)
0 0
Ql
• • •0 0
(l) 0
and fill
® QJ 0 Q' (3) (3!
® ® circles
b) Below each number,
0 0 (1 (9 0 •0 that cor-
fill in the circle that
correspo nds to that
-• 0
® (6)
\.6
®®®
(jj'\
®
respond
to 140
"D (?) 7)
-
7 7
number ( a) a .a a)
• 9 9_, !)
Thank you for completing the 1984 NURSES' HEALTH STUDY

Questionnaire.
t Please tear off the cover letter and return both
r
t parts (pages 1-6) in the enclosed pre-paid envelope.
t
NURSES' HEALTH STUDY QUESTIONNAIRE
111- ;:
'·
P
--
I I
-
I --
- --· ---=- --
-----=-----
. -
- . -· -·
---,.---- r -.- -
. - .
--
PAGE 1
r ··1
HARVARD MEDICAL SCHOOL
. -. . · .·
-- '
1. SINCE JUNE 1982 HAVE YOU HAD A PREGNANCY 5. DID YOUR FATHER EVER HAVE A MYOCARDIAL INFARCTION?
LASTING SIX MONTHS OR MORE?
NO YES )lo IF YES. AT WHAT AGE DID THE HEART ATTACK
(OR DON 'T OCCUR?
NO YES +IF YES. WHEN WAS THE DELIVERY? KNOW)
+ 1982 1983 1984 1985 BEFORE AGE 50 AFTER AGE 60

AGE 50 TO 60 AGE UNKNOWN
2. WHAT IS THE TOTAL NUMBER OF YEARS YOU HAVE
REGULARLY WORKED IN OPERATING ROOMS? (COUNT 6. DID YOUR MOTHER EVER HAVE A MYOCARDIAL INFARCTION?
ONLY THE TIME SINCE YOUR TRAINING.) r-
NO YES )lo IF YES. AT WHAT AGE DID THE HEART ATTACK
(OR DON'T OCCUR?
NONE 1-4 YEARS 10-14 YEARS KNOW)
LESS 1 HAN 5 9 YEARS 1 5 OR MORE YEARS BEFORE AGE 50 AFTER AGE 60
E YEAR
AGE 50TO 60 AGE UNKNOWN
3. ON AVERAGE HOW MANY DAYS EACH MONTH DO
YOU TAKE ASPIRIN? (INCLUDE ANACIN, BUFFERIN, 7. DO YOU CURRENTLY SMOKE CIGARETTES?
etc. DO NOT INCLUDE TYLENOL OR OTHER ASPIR IN-
FREE PRODUCTS.) NO YES YES. a) HOW MANY DO YOU SMOKE PER DAY7
NEVER 5- 14 DAYS A MONTI-I , 4 16 24 35 44
1 4 DAYS 15-21 DAYS A MONTH 5 14 25-34 45 OR
A MONTH MORE
22 OR MORE DAYS
/b) WHAT SPECIFIC BRAND DO YOU
)t SMOKE7 (e.g. "MARLBORO liGHTS 100's"l :-
4. ON DAYS THAT YOU DO TAKE ASPIRIN, HOW MANY
DO YOU USUALLY TAKE? I
'
(NEVER TAKE IT) 3- 4 ASPIRIN BRAND AND TYPE
1 5- 6 ASPIRIN I
2 ASPIRIN 7 OR MORE ASPIRIN 6 7 '

tO
8. HAVE YOUR MENSTRUAL PERIODS CEASED PERMANENTLY?

NO YES . I F YES, I
lOR No·r
SliREI '
(a) AT WHAT AGE? b) FOR WHAT REASON?
I
(AGE) SURGERY ._IF DUE TO SURGERY. WERE YOUR OVARIES REMOVED? 'j
YES.BOTH NOT
REMOVED
- ,'
I
ONE OhiLY
,
I
RADIATION
,
I
.. , I
j
I
NATURAL ._IF YOU HAD A NATURAL MENOPAUSE (NON-SURGICAL)
MENOPAUSE HAVE YOU SUBSEQUENTLY HAD SURGERY TO
-
,j
REMOVE OVARIES OR UTERUS? ,- ,j
UTERUS !-
,j
NO ONE OVARY REMOVED

BOTH OVARIES REMOVED
REMOVED!- ,I
j
I
9. SINCE JUNE 1982. HAVE YOU USED FEMALE HORMONES? (OTHER THAN ORAL CONTRACEPTIVES)
,j
,j
NO YES YES, a) ARE YOU CURRENTLY USING THEM? NO NO I CURRENT!":(

,j
b) HOW MANY MONTHS HAVE U USED TH ... ,,.

24-MONTH PERIOD BETWEEN JUNE 1982 A11.1n
1-4mo
5- 9 rno
10 14f!lo
15 19 mo.
20 TO
24 mo ,j
c) WHAT TYPE OF HORMONE

HAVE YOU USED THE LONGEST
PHEMARIN OTHEH CONJUGATED ESTROGEN I g
',
DURING THIS PERIOD7 PROGESTERONE COMBINATION E.STROGEN AND PROGESTERONE
SEQUENTIAL ESTROGEN AND PROGESTERONE
,
J
__
d) WAS THIS AN ORAL OR VAGINAL PREPARATION?

ESTROGEN AND TESTOSTERONE
ORAL VAGINAL
,j
e) IF THIS IS CONJUGATED ESTROGEN .30 mg./DAY OR LESS (GREEN) MORE THAN 1 25 mg DAY j
(eO.gU.
Y WKEH?AT DOSE DID .625 mg /DAY (BROWN) DOSE UNKNOWN OR
,
f) IF THIS IS CONJUGATED ESTROGEN (e.g. PREMARIN),
1 26 mg /DAY (YELLOW) USED VAGINAL CREAM
- ,j
j
HAVE YOU USUALLY TAKEN IT DAILY OR CYCLICALLY?
(CYCLICALLY OMITS SOME DAYS EACH MONTH) DAILY
CYCLICALLY
UNKNOWN
,
j
10. HAVE YOU EVER HAD ANY OF THE FOLLOWING: BE- YEAR OF OCCURRENCE
(IF YES, PLEASE MARK THE YEAR OF OCCURRENCE) ..,.I 1976 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985
a) CORONARY ARTERY SURGERY> NO l

l
c) COLON POLYPS? NO
-- NCS Trnns-Optic' EBO 1-14222:321 II PAGE 2
11. DATE OF BIRTH: 12. CURRENT

ONTH DAY YEAR WEIGHT:
--
THIS IS YOUR IDENTIFICATION NUMBER:
USED FOR MAINTAINING CO
@ @ 6
,'
0 7
l G) 1 0 2 4 6 6 7
-- 0
G)
0
,
,
2
2
4
4
5
5
II 8
B
9
G) t 2 4 6 6 1
@ ® 1 3
-- 0
-
@ G) 2 6 6 7
(i) 0 I .2 3 4 6 8
0 (!) IF YES.
G) 0 (!) PLEASE MARK MONTH OF DIAGNOSIS YEAR OF DIAGNOSIS
THE DATE OF
13. SINCE JUNE 1982 HAVE YOU HAD ANY OF z (X) a: a:
<
z (.!) Q.
w t) 0> 1982 1983 1984 1985
THE FOLLOWING ILLNESSES DIAGNOSED:
DIAGNOSIS ..,< :! <
Q. ::;,
-, -,
::;)
<( VI 0 2
-- ELEVATED CHOLESTEROL ? 0 NO
YES>-f..J
GH BLOOD ESSURE ? Q NO
WHILE
MYOCARDIAL INFARCTION (HEART ATT,\ CKI 7 0 NO 0 Yes.-Q)0@@@)Q)Q)@@@®® 19 @ @> ®

WERE YOU HOSPITALIZED? ---i)lr..,..Q NO
ANGINA PECTORIS 7
- STROKE (CVA) 7
PULMONARY EMBOLUS? NO
-- PERIPIIERAL VENOUS THROMBOSIS ? NO

@
-I
FIBROCYSTIC BREAST DISEASE? NO
NO YES
- 2
-
0 OTHER BENIGN BREAST DISEASE ? NO
Q YES
--
I
- -----"
BREAST CANCER ? NO @
- o 3 CANCER OF THE CERVIX IINCLUDE IN SITUJ 7 NO
- a t CANCER OF THE UTERUS ? NO
- 1 2 CANCER OF THE OVARY ? NO
- s 3 e CANCER OF THE COLON OR RECTUM ? NO
-_, J
8
It - - - -
CANCER OF THE LUNG ?
MELANOMA 7
- tt 9 BASAL CELL SKIN CANCER ?
- , ----------
SQUAMOUS CEtL SKIN CANCER? NO
-- " 8 OTHER CANCER ? NO
RHEUMATOID ARTHRITIS ?
- 11 I 1 GOUT?
---"
- 2 2 l 011-IER ARTHRITIS ? NO
- 3 3 s SYSTEMIC LUPUS ERYTHEMATOSIS 1sr E) ? 0 NO
- .c 4 4 CHOLECYSTECTOMY (GALL HlADOER REMOVAl! ? 0 NO
-
- 6 e e
- 1 7 7 GALL STONES, NOT REMOVED ? 0 NO @ @
- 8 0 SYMPTOMS ONLY
ULTRASOUND
0 0 b) WHEN DID GALL BE- 1982 0SYMPTOMS
NO
STONE SYMPTOMS START?
- 1
- 2 2. 2 OSTEOPOROSIS? 0
- s s a
r----------- -- - ------=
FRACTURE OF HIP OR fOREARM ? NO
NO
, 6 5 B
GASTRIC OR DUODENAL ULCER ?
ULCERATIVE COLITlS ? NO
,
I
I G 6 OTHER MAJOR ILLNESS 7 NO
,'
I
7
8
7
8
7
s
IF YES, PLEASE
SPECIFY ILLNESS
I
9 PLEASE GO TO PAGE 3 AND BEGIN BY WRITING IN YOUR 10 NUMBER

NURSES' HEALTH STUDY QUESTIONNAIRE PAGE 3 HARVARD MEDICAL SCHOOL -
14. PLEASE WRITE IN YOUR IDENTIFICATION NUMBER :.
(YOUR ID NUMBER IS PRINTED AT THE TOP OF PAGE 2)
' 2
15. DO YOU REGULARLY TAKE MULTIPLE VITAMINS? 2 3 4 8
Q NO ( YES • IF YES, a) HOW MANY DO YOU TAKE b) WHAT SPECIFIC BRAND DO 0 1 2 3 5 7 .1!
PER WEEK? YOU USUALLY USE? 2 3 5 6 1 a
2 OR Lt:SS Q 6-9
3-5 Q 10 OR MORE SP.ECIFY EXACT BRAND
16. NOT COUNTING MULTIPLE VITAMINS¥ DO YOU TAKE ANY OF THE FOLLOWING PREPARATIONS:
a) VITAMIN A?
( NO Q YES, SEASONAL ONLY} IF YES WHAT DOSE ; Q l!ESS THAN Q 8,000 to Q 13,000 to Q 23,000 IU DON'T
'PER DAY? 8,000 IU 1 .Z,OOO IU 22,000 IU OR MORE KNOW
Q
•
b) VITAMIN C?
YES , MOST MONTHS
-
Q NO Q YES, SEASONAL ONLY} IF YES. WHAT DOSE
PER DAY?
•0 LESS 1 HAN
400 n1g.
Q 400 to
700 mg. 0 750 to
1250 rng 0 1300 mg.
OR MORE
Q DON'T
KNOW
Q YES. MOST MONTHS
+
c) VITAMIN Be?
( NO Q YES IF YES, WHAT DOSE PER DAY? LESS THAN
10 mg.
Q 10
39 mg
Lo 0 40
79 mg.
Q 80 mg
OR MORE
0 DON'T
KNOW
d} VITAMIN E?
0 NJ>
,
Q YES If YES, WHAT DOSE PER DAY? LESS THAN
\ 00 IU
Q 1 00 to
250 IU
Q 300 to
500 IU 0 600 IU
OR MORF
Q DON'T
KNOW
e) SELENIUM?
0 0 YES If YES, WHAT DOSE PER DAY? LESS THAN

80 rncg.
0 80 to
130 meg
0 140 to
250 meg.
0 260 meg.
OR MORE
Q DON' T
KNOW
f) IRON?
0 ( YES --=:...,.IF YES. WHAT DOSE PER I:>AV? _.....,,_,....0 LESS

51
THAN 0 51 to
200 mg.
Q 201 to
400 rng,
Q 401 mg
OR MORE
Q pON'T
KNOW
g) ZINC? j
0 Q YES )lo IF YES, WHAT DOSE PER DAY? -....,:)to..., Q lESS THAN Q 25 to
0 75 to Q 101 mg. 0 DON'T I
25 mg 74 mg. 1 00 mg OR MORE KNOW
'
I
h) CALCIUM? !Include Colohtm

· in Oolomllll.)
Q NO 0 YES IF YES, WHAT DOSE PER DAY? _ _;,.._. Q LESS THAN Q 400 to Q 90'1 to Q 1301 mg Q DON'T
400 mg. 900 mg. 1300 l'fl9· OR MORE KNOW
l
i) ARE THERE OTHER SUPPLE- 0 FOLIC ACID Q CHROMIUM Q IODINE Q BETA-
MENTS THAT YOU TAKE ON Q CAROTENE
Q VITAMIN D ( ' COPPER LECITHIN
A REGULAR BASIS? PLEASE
MARK IF YES: 0 B-COMPLEX
VITAMINS
( MAGNESIUM Q RUTIN
PLEASIO SPECIFY
17. FOR EACH FOOD USTED, FILL IN THE CIRCLE INDICATING

HOW OFTEN ON AVERAGE YOU HAVE USED THE AMOUNT AVERAGE USE LAST YEAR
SPECIFJED DURING THE PAST YEAR. NEVER, 1-3 1 2-4 5-6 1 2- 3 4-5 6+
OR LESS PER PER PER PER PER PER PER PER
F_O_O_D_S________. - j THAN ONCE
.--------'-----0-A-IR_Y
__ PER MONTH MO. WEEK WEEK WEEK DAY DAY DAY DAY
SKIM OR LOW FAT MILK (8 oz.. GLA.SS)

WHOLE MILK {8 oz GLASS)
CREAM, e.g COFFEE, WHIPPED (TBSl
SOUR CREAM (TBS)
tCE CREAM ( CUPI

YOGURT t 1 CUP)
COTIAGE OR RICOTIA CHEESE ('•; CUP)
CREAM CHEESE (1 oz.)
OTHER CHEESE. e g AMERICAN, CHEDDAR. etc.

PLAIN OR AS PART Of A DISH ( 1 SLICE
OR 1 SERVING) 0
-------------------
MARGARINE (PAT). ADDED TO FOOD OR BREAD;
EXCLUDE USE IN COOKING 0 0 w 0 0 @ 0 0 0
PLEASE TURN BU I I ER (Pf\T), ADDED TO FOOD OR BREAD;
TO PAGE 4 EXCLUDE USE IN COOKING
•
- PAGE 4
17. (CONTINUED) PLEASE FILL IN YOUR AVERAGE USE,

DURING THE PAST YEAR, OF EACH SPECIFIED FOOD. NEVER, 1- 3 1 2-4 5-6 1 2-3 4 -5 6+
THAN ONCE WEEK WEEK WEEK DAY DAY DAY DAY
PER M ONTH MO.
- FRUITS
- -
RAiSINS I 1 oz. OR SMALL PACK) OR GRAPES
PRUNES ('h CUP) - 0 0 w 0 0 @ 0 0-:.-: 0 -..
-- 1----=---+----=r-:\='" - -
BANANAS ( 1) 00 w O O ® v 0 0 '·· ·
--
PLEASE TRY TO
AVERAGE YOUR --- a·-f'- o w -o l- o -\::.1 o o -
SEASON AL USE
CANTALOUPE ( MELON)
---'-------
WATER M ELON I 1 SLICE)
- 0 0- V \!:!:!) 0 =-1---'@o)='_
0 \.,..
--;-,-;-4.-+ '
U
--
Of FOODS OVER
THE ENTIRE YEAR . FRESH APPLES OR PEARS ( 1l 00 8 00 ® 000'
FOR EXAMPLE, IF ·-
APPLE JUICE OR CIDER (SMALL GLASS) ___ _Q 0 ® o 'J Q_ -
--
A FOOD SUCH AS
CANTALOUPE IS ORANGES {11 _ ,_Q_ o _o_ o ® ® n o o _ I·
EATEN 4 TIMES A
ORANGE JUICE (SMALL GLASS) _ 1---':o- o e Q_ o ® o o o .
--
WEEK DURING THE
A PPROXIMATE 3
M ONTHS THAT IT IS I-
GRAPEFRUFr {lh) 0 _0_1_® __0 0 @ Q_ 0 ___ .
GRAPEFRUIT JUICE (SMALL GLASS! ·- 0 ® 0 0 @) 0 0 0
-- IN SEASON, THEN
THE AV ERAGE USE
WOULD BE ONCE
OTHER FRUIT JUICES GLASS)
STRAWBER RIES, FRESH, FROZEN OR CAN NED I
_
- _
0 0 __ 1\Ai\ 0
o__ .-Q_
0
Q
/";;\
@
0 ·- (l'l 0 -·,
-- PER WEEK.
BLUEBERRIES . FRESH. FROZEN OFI CANNED ( V2 CUP)
PEACHES, APR ICOTS OR PLUMS (1 FRESH,
0 1U
·---:o: a·- fW\
'-::/
l J 0 @
a-.
Q
o o
-Q () -
OR 1• CUP CANNED) , - 7 1;::- .- --c 1-
-
--
NEVER, 1-3 1 2- 4 5- 6 1 2-3 4 -6 6+
OR LESS PER PER PEA PEA PEA PER PER PER
.-----------------------------------1 THAN ONCE
PER M ONTH MO. WEEK WEEK WEEK DAY DAY DAY DAY
VEGETABLES
-- TOMATOES ( 1l
1 OMATO JUICE (SMALL GLASS)
0 0 u 1-0 ___®_ l_c; :· 0 _Q
n ® -cs n
-- TOMATO SAUCE ( CUP) o g SPAGHE I II SAUC£: Q_ o ® __ o o Gl Q_ o _ 1,
RED CHILl SAUCE ( 1 TB.Sl 0 0 Q_ I___Q 0 _Q_ -

-- TOF U OR SOYBEANS (3 4 oz)
51 RING BEANS CUP)
- 0 0 ® 0
__@ n
0
o
® Q 0
® o
0
.Q_
·- BROCCOLI I CUPI
CABBAGE OR COLE SLAW ( CUP)
1-·0
Q
0 -® 0 0
Q @) Q _0 _ g)_
0 - @=-'
--
() _
CAULIFLOWER (1 CUP) __ Q {) @5 Q Q @ Q 0 Q _
BRUSSELS SPROUTS l h CUP) Q 0 "_W_ _Q Q $ _ Q _0 Q
-- ,_
CARROTS 11 W IIOLE OR CUP COOKED)
CORN ( 1 EAR Ofl 112 CUP FROZEN OA CANNED! Q -Q @)
A -
- - c::-
( ) ()
"?) l'j)\. 1" 0
1
"" \!!)
@ Q
1
\ j.- - " ' - -

U Q.. -
-- -
PEAS. OR LIMA BEANS ( CUP FRESH , FROZEN. CANNWI
MIXED VEGETABLES (112 CUPI
. 1- ;.---;;,-
v
Q t_)
C) _@
.'---
,.... -;;s: 1 -o-.:----
(:5. 1- -
_Q
-o-- D
I_ : - _
)
r--\
u
....
j )
,.......,
U-"""'"
/-\ · -
-
-
BEANS OR LENTILS, BAKED OR DRIED (1 CUP} Q () ® () _Q o u t)
-- ,_ YELLOW (WINTER) SQUASH 1112 CUP)

I-
Q ,_Q
r"'
U
0
@)
\.
Q -0 - I-@- -Q-
. . . ---
1
f t.\
r't fJ
--
EGGPLANT, ZUCCHI NI OR OTHER SUMMER \0.,.
SQUASH ( CUP) _ ,- · _ _
1-
YAMS OR SWEET POTATOES l'h CUP) Q -0
_Q--= :--==o
-@) -1
1
I·
·0 Q
r\
_@ __0
<""o '·
9 -'·
SPI NACH, COOKED (1 CUP) .w . ..:. -.::_
I:
- SPINACH, RAW AS IN SALAD

KALE, MUSTARD OR CHARD GREI:NS { CUP) Q CI
® Q 0- :=--@ _( _ Q
() .J?J
Q ·•-
-
-- ICEBERG OR HEAD LETTUCE (SERVING)
ROMAINE OR LEAF LETTUCE (SERVING}
·- r., Q
Q
{W) Q
{
Q (o)
t rio
a· (
.__.t' ·-·
U
"--';"""'" 1
•.
-
1vv : _ '· r ...
1-
CELERY {4" STICK)
.- I"""?"'
U
Q
CJ
.- Q- fi)\.
SJ.
f"""\ '-··-"=
(,__) 0
MUSHROOMS (ONEl FRESH, COOKED, OR CANNED Q Q- w 1- f- Q !'o - t
1
C I·
-- 1
BEETS (lh CUP)
1- ·
ALFALFA SPROUTS (' CLIP!
Q
Q
1
Q - @- r-Q
\_•YJ
- I-Q
V
Q- _Q
1 .J .
:- f')
Q_
--
.:.!' _}_
1- GARLIC. FRESH OR POWDERED ( t CLOVE OR SHAKE)
- 0 o:...=--- /WI 0 1- rno..
o
\!!./
0 0 l----:;;.0:::----tf-
-- NEVER,
OR LESS
1- 3
PER
,
PER
2-4
PER
5-6
PER
1
PEA
2-3
PER
4- 5
PER
6+
PEA
-- ,....- - - -- - - - - -- - - - - - - - - - -- -- - - - - - - - i T H A N ONCE
_EGGS (1)
EGGS, MEATS, ETC.
PER MONTH
0
MO.
_Q
WEEK WEEK W EEK
@ 0 0
DAY
@
DAY
_Q
DAY DAY
--
_
CHICKEN OR TURKEY, WITI I SKIN (4 - 6 oz.l _ __0 Q @ Q _Q
-='-_1--'@='
o_ 1
_0_ 0 Q _
Q _Q_
1
Ct·!ICKEN 9R TUfiKEY, WITHOUT SKif'J (4- 6 oz.} _ _Q Q @ Q _1• _
-
__
1 11
BACON (2 SLICES! - -- -- - - - -+-0

::---- 0 Q_._Q @ ·- 0 0
- PLEASE GO
TO PAGE 5 HOTOOGS(1) Q Q @ 0 0 @ C Q Q
PAGE 5 NCS Trans-Optic EB01·1 4 2 11 :32 1
--
17. (CONTINUED) PLEASE FILL IN YOUR AVERAGE USE, NEVER.
DURING THE PAST YEAR, OF EACH SPECIFIED FOOD. 1 -3 1 2- 4 6-6 1 2- 3 4- 5 6+
OR LESS
THAN ONCE
PER MONTH
PER
M O.
PER PER
W EEK WEEK W EEK
PER PER
DAY
PER
DAY
PER
DAY
PER
DAY l
MEATS(CONTINUEO) l
PROCESSED MEATS, e g S AUSAGE, SALAMI,
BOLOGNA, etc. (PIECE OR SLICE}
u
LIVER (3- 4 oz.}
HAMBURGER (1 PATIY)
BEEF, PORK, OR LAMB AS A SANDWICH OR MIXED

DISH, e .g. STEW, CASSEROLE, LASAGNE. etc.
BEEF. PORK. OR LAMB AS A MAIN DISH. e.g STEAK.

ROAST, HAM etc. 14-6 oz.)
0 0
CANNED TUNA FISH (3-4 oz.) ®
DARK MEAT FISH. e.g. MACKEREL. SALMON . @) 0
SARDINES BLUEFISH, SWORD FISH (3 -5 oz)
OTHER FISH (3 - 5 oz.}

------------------
SHRIMP. LOBSTER, SCALLOPS AS A MAIN DISH
NEVER. 1-3 1 2-4 6- 6 2-3 4 -5 6+

OR
THAN vr
cj PER PER P£ 8 P£R PER PEA
r-----B- R
_E_A_D
_S- -,C--ER--EA--lS--
,S-T
_A__
R_
C_H_E_S____ MO. WEEK WEEK WEEK DAY DAY
COLD BREAKFAST CEREAl (1 CUP)

COOKED OATMEAL (1 CUP)
--
1-
OTHER COOKED BREAKFAST CEREAL {1 CUP)
WHITE BREAD (SLIC INCLUDING PITA BREAD
--
DARK BREAD (SLICE)
ENGLISH MUFFINS, BAGELS. OR ROLLS (1)
'--- -- - - - -
-
MUFFINS OR BISCUITS (1)
BROWN RICE (1 CUP)
WHITE RICE (1 CUP)
PASTA. e.g. SPAGHETTI , NOODLES, etc. ( 1 CUP} -
OTHER GRAINS, e g BULGAR, KASHA.
COUSCOUS. etc 11 CUP!
PANCAKES OR WAFFLES {SERVING)

- -
FRENCH FRIED PO;rATOES (4 oz.)
POTATOES, BAKED, BOILED ( 1) OR MASHED ( 1 CUP)
POTATO GHIPS OR CORN CHIPS (SMAL L BA G OR 1 oz.)
CRACKERS . TRISKETS, WHEAT TH INS i)
PIZZA (2 SLICES)
1 5- 6 4-5 6+
PER PER PER PEA
WEEK WEEK DAY DAY
BEVERAGES
CARBONATED LOW CALORIE COLA. e g. TAB W ITH CAFFEINE
BEVERAGES low Calor ie
LOW CALORIE CAFFEINE·FAEE COLA. e.g PEPSI FREE
(sugar-free)
types OTHER LOW CALORIE CARBONATED BEVERAGE. e.g
CONSI DER THE FRESCA. DIET 7 UP, DIET GINGER ALE
SERVING SIZE
COKE. PEPSI , OR OTHER COLA WITH SUGAR
AS 1 GLASS.
BOTTLE OR CAN Regular types CAFFEINE FREE COKE, PEP.SL OR OTHER COLA
FOR THESE (not sugar · WITH SUGAR
CARBONAT ED
BEVERAGES.
free)
OTHER CARBONATED BEVERAGE WITH SUGAR, 0 @ 0
e .g. 7- UP. GINGER ALE
OT HER HAWAIIAN P.UNCH. LEMONADE. OR OTHER NON·

BEVERAGES CARBONATED FRUIT DRINKS ( 1 GLASS, BOn LE, CAN)
DECAFFEINATED COFFEE ( 1 CUP)

-
COFFEE 11 CUP)
------
TEA (1 CUP), NOT HERBAL TEAS
--
I--
BEER I1 GLASS, BOffiE. _ ______
___;,_CAN)
RED WINE (4 oz. GLASS)

--
WHITE WINE {4 oz. GLASS!
LIQUOR . e .g WHISKEY, GIN, etc. l 1 DRINK OR SHOT)
PLEASE TURN 1
TO PAGE 6
- PAGE 6 Copyright 0 1984 Brigha m a nd Women's Hospital.
All Rights Reserved Worl dwide.
t
r, .... IDENTIFICATION NUMBER: - - - - - -- -- - - -- A
.... 17. (CONTINUED) PLEASE FILL IN YOUR

AVERAGE USE DURING THE PAST YEAR,
OF EACH SPECIFIED FOOD.
NEVER, 1-3 1 2- 4 5- 6 1 2- 3 4-5 6+
..
THAN ONCE
PER MONTH MO. WEEK WEEK WEEK DAY DAY DAY DAY
SWEETS, BAKED GOODS. MISCELLANEOUS
.. CHOCOLATE (BARS OR PIECES) e.g. HERSHEY'S, M & M' S
.. CANDY BARS. e g SNICKERS. MILKY WAY. REESES
CANDY WITHOUT CHOCOLATE ( 1 oz.)
.. COOKIES HOME BAKED 11)
COOKIES READY MADE ( 1)
.. BROWNIES ( 1)
-'---
3 3
, ..
DOUGHNUTS I 1)
CAKE. HOME BAKED (SLICE)
....
I- . . CAKE. READY MADE (SLICE)
SWEET ROLL, COFFEE CAKE OR OTHER PASTRY,
HOME BAKED (SERVING)
0
.... SWEET ROLL. COFFEE CAKE OR OTHER PASTRY.
READY MADE (SERVING)
.... PIE, HOMEMADE {SLICE)

PIE, READY MADE (SLICE)
JAMS. JELLIES. PRESERVES, SYRUP, OR HONEY ( 1 TBSI
.... PEANUT BUITEA (TBS)

POPCORN (1 CUP)
-
.... NUTS (SMALL PACKET OR 1 oz.)
----
BRAN, ADDED TO FOOD ( l TBS)
.. WHEAT GERM t 1 TBS)

CHOWDER OR CREAM SOUP II CUP)
-
--..
011 AND VINEGAR DRESSI NG e.g ITALIAN { 1 TBSI
MAYONNAISE OR OTHER CREAMY
----
SALAD DRESSING i 1 TBS)
.. MUSTARD DRY OR PREPARED ( 1 tsp)
PEPPER ( 1 SHAKE)
SALT (1 SHAKE}
____. ___
18. WHAT DO YOU DO WITH THE VISIBLE FAT ON YOUR 24. HOW MANY TEASPOONS OF
.. MEAT? SUGAR DO YOU ADO TO
.. EAT MOST OF THE FAT 0 EAT AS U TILE A S POSSI BLE .... -- e
.. Q EAT SOME OF THE FAT (DON'T EAT M EAT} 25. WHAT TYPE
OF COOKING
.... 19. WHAT KIND OF FAT DO YOU USUALLY USE FOR

FRYING AND SAUTEING? (EXCLUDE "PAM" -TYPE
SPRAY)
OIL DO YOU
USUALLY S PECIFY TYPE AND BRAND
26. WHAT KIND OF

.. 0 REAL BUTTER VEGETABLE OIL 0 LARD
CO LD BREAKfAST
CEREAL DO YOU
- 0 M ARGARINE VEGETABLE SHORTEN IN G USUALLY
SPECIFY TY PE AN D BRAND
. . 20. WHAT KIND OF FAT DO YOU USUALLY USE FOR

BAKING? 3
REAL BUTTER VEGETABLE OI L Q LARD
- MARGARINE VEGETAB LE SHORTENIN G INCLUDE FOR EXAMPLE: PATE, TORTILLAS. YEAST. CREAM SAUCE.
CUSTARD, HORSERADISH, PARSNIPS. RHUBARB. RADISHES. FAVA
..
. . 21 . WHAT FORM OF MARGARINE DO YOU USUALLY
USE?
BEANS, CARROT JUICE. COCONUT, AVOCADO, MANGO, PAPAYA •
-..
DRIED APRICOTS. DATES, FIGS .
NONE STICK TUB 0 DIET FO RM
(DO NOT INCLUDE DRY SPICES AND DO NOT LIST SOMETHING
FORM FORM (LOW CALO RIE}
9 0
THAT HAS BEEN liSTED IN THE PREVIOUS SECTIONS.)
-..
. . 22. HOW OFTEN DO YOU EAT FOOD THAT IS FRIED AT
OTHER FOODS THAT YOU USUALLY USUAL SERVINGS
HOME? (EXCLUDE THE USE OF "PAM" -TYPE SPRAY) USE AT LEAST ONCE PER WEEK SERVING SIZE PER WEEK
.. QDAILY 0 4-6 TIM ES PER WEEK
- Q 1-3 TIMES PER WEEK 0 LESS THAN ONCE A W EEK

(a)
..
. . 23. HOW OFTEN DO YOU EAT FRIED FOOD AWAY FROM
HOME? (e.g . FRENCH FRIES, FRIED CHICKEN, FRIED
(b)
-.. FISH)
.. DAILY
1-3 TIMES PER WEEK
Q
Q
4 -6 TIMES PER WEEK
LESS THAN ONCE A WEEK I d)
PLEASE REPLY TO:
Channing Laboratory
180 Longwood Ave.
Boston. Mass. 02115
617-732-2279
Dear Colleague:
It is now ten years since you completed the first Nurses' Health Study Questionnaire in 1976.
---- Your participation and that of over 120,000 other R.N.'s has made this study the largest pros-
pective investigation specifically directed to the health of women. I am most grateful for the
detailed information you have provided over these years. Whether or not you are still active in
nursing, your continued participation is extremely valuable.
Several important findings have already emerged from the study. In 1976, there was concern
that oral contraceptives and post-menopausal estrogens might increase the risk of breast cancer.
Reassuringly, neither of these hormonal preparations affect the risk of breast cancer ( 1,2).
Continued monitoring is necessary to determine whether this remains true over longer periods.
The use of postmenopausal estrogens was associated with a reduced risk of myocardial infarction
(MI) (3). Current use of oral contraceptives increased the risk of Ml (4) but past use carried no
extra risk (5). As expected, cigarette smoking was clearly a major risk factor for Ml (6).
We are examining the effects of hair-dye use and diet on breast cancer; details will be sent
to you on our 1987 newsletter. Other analyses in progress include the relationships of diet with
risk of other cancers and heart disease, and the health effects of exercise, to name a few. We
particularly appreciate the special efforts of those who provided additional detailed information
on dietary intake, or who gave permission to review their medical records for further diagnostic
details. This added information has documented the accuracy of reported dietary intake (7) and
medical events (8).
Please complete and return the enclosed questionnaire at your earliest convenience. As always,
information will be kept strictly confidential and used for medical statistical purposes only.
Again, I would like to express my deepest gratitude for the contribution you have made to this
study. Already this has yielded useful information, and we are confident that findings during the
---- next several years will provide important guidance for maintaining optimal health.
Sincerely,
Research Group:
---- Frank E. Speizer, M.D.
Graham Colditz, MD.
Karen Corsano. M.A.
Meryl Dannenberg, B.S.
David Dysert
Gertrude Geller, R.N.
Barbara Egan
Chartes Hennekens, M.D.
Maureen Ireland
Susan Newman *References:
Bemard Rosner, Ph.D.
l aura Sampson, R.D.
Frank E. Speizer, MD. 1. Oral contraceptives and breast cancer. a prospective cohort study. JAMA 1986; 255:58-61.
Meir Stampfer, MD. 2 A prospective cohort study of postmenopausal female hormone use and risk of breast cancer. Am. J. Epidemiol
Harry Taplin, M.A. 1983; 118:416.
Walter Willett. M .D. 3. A prospective study of postmenopausal estrogen therapy and coronary heart disease. N. Engl. J . Med. 1985;
313:1044-9.
Advisory Soard: 4. Oral contraceptive use in relation to nonfatal myocardial infarction. Am. J . Epidemiol 1980; 111:59-66.
5. Past use of oral contraceptives and risk of coronary heart disease. Atn. J . Epidemiol 1985; 122:547.
Thelma M Schorr, R.N. 6. Cigarette smoking and non-fatal myocardial infarction in women. Am. J. Epidemiol 1981; 113:575-82
President, American Journal of Nursing.
Susan S. Hartley, Ph.D.
7. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am. J. Epidemiol 1985; 122:51-65.
Director of Statistics and Data Analysis a Validation of questionnaire information on risk factors and disease outcomes in a prospective cohort study of
American Nurses' Association women. Am. J. Epiderniol 1986; 123:894-900.
I
INSTRUCTIONS
(
Please use an ordinary pencil to answer all questions by completely

filling in the appropriate response circle, or by writing the requested
information if a space is provided. Note that some questions ask for
information since June 1984, some ask for current status, and some ask
about events over longer periods. Because this form is meant to be read
by optical-scanning equipment, it is important for you to make no stray
marks and to keep any write-in responses within the provided spaces.
Should you need to change a response, erase the incorrect mark
completely. If you have comments, please write them on a separate
piece of paper.
EXAMPLE 1: Do you live in the United States? QNO .YES
Fill circle completely, do not mark this way:
EXAMPLE 2: Where were you bom? OR£GDN

Keep hand-writing within borders of the response box.
EXAMPLE 3: DATE OF BIRTH AND CURRENT WEIGHT:
a) Write in birthdate and 2. CURRENT

1. DATE OF BIRTH:
weight in the boxes at WEIGHT:
and
the top of each grid. I
140
For example, May 9, ® pounds:
192 1 would be G)
®
b) Below each number, and fill
'+--- circles
fill in the circle that 0 that cor-
corresponds to that r---++--.1• respond
number ® to 140
0
®
Thank you for completing the 1986 NURSES' HEALTH STUDY Questionnaire.
Please tear off the cover letter and return

questionnaire in the enclosed pre-paid envelope.
NURSES' HEALTH STUDY QUESTIONNAIRE Page 1 HARVARD MEDICAL SCHOOL --
1. DATE OF BIRTH
MONTH DAY YEAR
2. CURRENT
WEIGHT (lbs.)
3. Have your menstrual periods ceased permanently?
0 Yes _... a) For what reason 7 b) At what age?
--
@@ @@ @ @@
0No
l
4 0 SURGERY --+ If due to surgery.
were your ovaries removed?
@ @ -
age--+
-
00 00 00 CD 0 CD To #4 0 Yes, both 0Not CD - - (1
® ®® ®® ® ®® Oone only
removed
® ® -
0 00 ®0
@ @ @0
0 0®
4) 0@ 0 RADIATION OR CHEMOTHERAPY
®
0 - -
- (3
;::.
(4
® ® -
;::::
® ® ®® - ®
5) ®
® - -
NATURAL • If natural (non-surgical)
® ® ®® ®® 0 menopause. have you had subsequent (i
3) 0 0 -®
(7 00 surgery to remove ovaries or uterus?
0 -
S)®
® ®
®
® ®
®®
®®
0No
0Uterus
0 One ovary removed
0 Both ovaries removed l9-
-
--
(8
;::::
removed
--
I ;
4. Since June 1984, have you used female hormones (other than oral contraceptives)?
0 Yes Are you currently using them? 0 Yes. currently 0 No, not currently
0No b) How many months have you used them during the 24-month period between June 1984 and June 1986?
01-4 mo. 05-9 mo. 010-14 mo. 0 15-19 mo. 020-24 mo.
--
c) What type of hormone have you used the longest during this period?
0 Premarin or other conjugated estrogen alone. 0 Estrogen & progesterone (in one pill)
--
0 Progesterone (e.g.
0
alone
Sequential Estrogen & Progesterone
0
0
Estrogen & Testosterone
Other, Please specify .. -
-
d) Was this an oral or vaginal preparation? Ooral 0 Vaginal 0 0 ®0 ® 0 0 ® -
@ @
e) If this was conjugated estrogen (e.g. Premarin}, what dose did you usually take?
0 .30 mg./day or less (Green) 0 9 mg./day (White) 0 More than 1.25 mg./day
--
0 .625 mg./day (Brown) 0 1.25 mg./day (Yellow) 0 Dose unknown
f) If this was conjugated estrogen (e.g. Premarin), did you take it daily or cyclically?
0 Used vaginal cream
--
' (cyclically omits some days each month)
5. How many months in total (all births combined) did you breast feed?
Ooaily 0 Cyclically 0Unknown
--
0Did not 0 Less than 1 month 01-3 mo. 04-6 mo. 07-11 mo. 0 12-17 mo. --
breast feed
018--23 1110. 024-35 mo.
6. a) What was birth weight of your heaviest child? 0 No children
036-47 mo.
Q less than 81/:o lbs.
048 or more
Os'h-91f2
0 Cannot remember
0 9'12-10112 0 10112 lbs+
--
b) What was your age at delivery of first child over 9112 lbs? 0 None over 91fz lbs. 0
c) What was your age at detivery of your heaviest child? 0 No children 0
Age 24 ot less
Age 24 or less
025-29
025-29
030-34
030-34
-- 0Age 35+
OAge 35+
7. Do you currently smoke cigarettes? ®0®®@®®0®® -
0 Yes How many per day? 01-4 05-14 0 15-24 0 25-34 0 35-44 0 45 or ®0®®0®®0®® -
®CD®®@®®®®® -
more ff.
0No
1 b) What specific brand do you smoke?

(e.g. "Marlboro Lights 100's")
specify exact
brand and
type ..• --
8. On average each day, how often do you apply lipstick?O Never
(not gloss or lip balm) use 1t
0Less than
once a day
01-3
times/day
04-5
times/day --06 or more
times/day
9. During the past year, what was your average time per week AVERAGE TOTAL TIME PER WEEK --
0 -
spent at each of the following activities: 1-4 5-19 20-59 ONE 1-1'12 2-3 4-6 7-10 11+
ZERO
Min. Min. Min. HOUR Hrs. Hrs. Hrs. Hrs. Hrs.
Walking or hiking outdoors (include walking at golf)
Jogging (slower than 10 minutes/mile)
0
0
0
0
0
0
0
0
0
0
0
0 0 0
0 0 -
-
0
0
0 0
0 0
Running (10 minutes/mile or faster} 0 0 0 0 0 0 0 0 - 0 0 0
Bicycling (include stationary machine) 0 0 0 0 0 0 0 0 - 0 0 0
Lap Swimming 0 0 0 0 0 0 0 0 - 0 0 0
Tennis 0 0 0 0 0 0 0 0 - 0 0 0
Calisthenics/ Aerobics/ Aerobic Dance/ Rowing Machine
Squash or Racquet Ball 0
0
0
0 0
0
0
0 0
0
0
0 0
0 0 -
0
0 0
0
- 0 0
0
10. What is your usual walking pace? 0 Easy, casual
(less than 2 mph}
0 Normal. average
(2-2.9 mph)
0Bnsk pace
(3-3.9 mph) --
0 Very brisk/striding
(4 mph or faster)
11. How many flights of stairs (not individual steps) do you climb daily?
0 2 flights or less 03-4 05-9 010-14 0 15 or more fhghts
--
12. Please count the number of moles on your left arm from your shoulder to your wrist, of this size or larger:e (3 mm+ diameter)
0None 01-2 moles 03-5 06-9 010-14 015-20 0 21 or more moles
--
13. a) Indicate total hours of actual sleep in a 24-hour period: 0 5 hrs. or less 0 6 hrs. 0 7 hrs. 0 + -
8 hrs. 0 9 hrs. 0 10 hrs. 0 11
b) What is your usual sleeping position? Oon back Oon side Oon front @) ® 0 ® - @ fC'
c) Do you snore? Regularly ( ) Occasionally ()Never
-
NCS Trans-Optic® EPO 1·23023:321 Copyright (r 1986 Brigham and Women's Hospital.
A9114 Page 2 All Reserved Worldwide.
I
,_- This is your identification number:
,_ ..2Jl:!J 0 (!) (!) (9 (!. 0 (.!) Q> @>I® G)

(used for maintaining confidentiality)
0 ®
-- ®0®®0®®0®® ®®®® ®
®0®®0®®0®® ®®@ ®
0
6)
®
®
®
®
@
0
®
®
@
®
0
®
®
®
(9)
(ff)
-- ®<D®®@®®0®® ® ® ® 0 @ 0 ® ® @ ® ® 0 ® -
(9
®CD®®@®®®®® ® ® ® 0 @ 0 ® ® @ ® @ 0 ® (9
-- ®0®®0®®0®® @®®0 @
YEAR OF DIAGNOSIS
CD ® ® @ ® (6) 0 ® ®
--
14. Since June 1984 have you had any of 17. Do you currently take multiple vitamins?
the following physician-diagnosed illnesses? BEFORE JUNE 84 JUNE 86 AFTER
JUNE 1 TO TO JUNE 1 -
Mark here for yes-.._ 1984 MAY 85 MAY 88 1986 How many do you take per week1
-- It DIABETES MELLITUS 0-rt

®-4
0 0 0 0 0No 02or
less
03-5 06-9 O 10 or
more
1
•' ELEVATED CHOLESTEROL 0 0 0 0
-- ·'
"i'
'I HIGH BLOOD PRFSSURE ®-+ 0 0 0 0 ,-b) \{\/hat specific brand do you usually use? ® a
·' MYOCARDIAL INFARCTION (Heart
@-t ... @ IS'
,, 0 0 0 0
--
'I
attack)
'I 4 Were you hospitalized? • 0Yes 0No
SPECIFY EXACT BRAND AND TYPE
©
ANGINA PECTORIS @--t 0 0 ® l<b
--
·: 0!0
1: 4Did you have angiogram or stress 0Yes 0No 18. Not counting multigle vitamins, do you take any
of the foflowing preparations: ®
•' CORONARY ARTERY BYPASS
--
"t
•'
'l or ANGIOPLASTY ®-+ 0 0 0 0 .r-PREPARATION _£AMOUNT PER DAY G)®ca
,I
@@co
r
STROKE ( CVA ) ®-+ 0 Vitamin A? 0 less than 8.000 IU per day
--
0 0 0
-!
.-
PULMONARY EMBOLUS 0 8.000-12,000 IU
::.
G)G)(1
"t ®-+ 0 0 0 0 0 Yes. seasonal only
,I
FIBROCYSTIC BREAST DISEASE ®-+ 0 0 0 0 0 Yes. most months 0 13.000-22.000 JU O®®C2
-- •'
"l
don't
,T
'l 4 Was this confirmed by a breast biopsy? • 0Yes 0No 0No+ 0 23.000 or more know ®®
., @@
r
OTHER BENIGN BREAST DISEASE ®-+ 0 0 0 0 Vitamin C? 0 less than 400 mg. per day
--., .•
,I
4Was thts confinned by a breast btopsy?
BREAST CANCER ®-+
• 0Yes
0 0
0No
0 0
I 0
0
Yes. seasonal only
Yes. most months
0 400-700 mg.
0750-1250 mg.
® @(6)
;::::
--
't
don't • ......
•'
'I CANCER OF THE CERVIX !include 1n·situl 0-+ 0 0 0 0 0No+ 0 1300 mg. or more know 0
·' CANCER OF THE UTERUS (enclornettiuml ®--. 0 0 0 0 Vitamin 8-67 0 less than 10 mg. per day (!)®(8
--
"t
;::::
•! CANCER OF THE OVARY ®-t 0 0 0 0 0Yes • 010-39 rng. @@(!,
-- ., .;
·'
COLON POLYPS (bentgn)
CANCER OF THE COLON OR RECTUM@-+
®-+ 0
0
0
0
0
0
0
0
0No
l
040-79 mg.
0 80 mg. or
doo't
know
0 ®@@
G)@c1
-- ., 'I
•'
,,
CANCER OF THE LUNG ®-+ 0 0 0 0
1-
Vitamin E? 0 less than 100 IU per day
-::::
MELANOMA ®-+ 0 0 0 0 0Yes 0100-250 IU

-
--
r
BASAL CELL SKIN CANCER ®-+ 0 0 0 0 0No 0300-5001U 0 @@(4

don"t
SQUAMOUS CELL SKIN CANCER ®-+ 0 0 0 0 l 0 600 IU or more know ® 5 (S
-- ,I
OTHER CANCER
4 Spectfy site of
®-+ 0 0 OtO
1-
Selenium?
0Yes
0 less than 80 meg per day
0 80-130 meg
@@s
00(7
;::.
;:.
-- .
'I •
,I
'l other cancer r 0No 0140-250 meg. d on.0
t
®®ci
-::::
l know ®®cs
-
FRACTURE OF HIP OR FOREARM 0 260 meg. or more

®-+ 0 0 0 0 -
--
'I
IJ 4 Please spectfy site and circumstances on a separate sheet.

OSTEOPOROSIS ®-+ 0 0 0 0 Iron? 0 less than 5 1 mg. per day
-- •1 RHEUM ATOlD ARTHRITIS (Phymcmn

GOUT
dwgJOsad)
®-+
®---. 0
0
0
0
0
0
0
0
0Yes
0No
r 051-200 mg.
0 20 1-400 mg. 0
@G) G)
-- 'J
lj
,,
SYSTEMIC LUPUS ERYTHEMATOSUS (SLE) @-4
OTHER ARTHRITIS
0
0
0
00-+
0
0 0
0 ,-
Zinc?
l 0 40 1 mg. or more
0 less than 25 mg. per day
don' t
know
l@@ W
®®®
-- ,,-.
••
·' GALL STONES
4 a) D1d you have symptoms?
®-t 010
. 0Yes
0
0No
0 0Yes
0No
• 025-74 mg
075-100 mg.
@@(5
0 (6) @(6
-- IJ b) How diagnosed?
CHOLECYSTECTOMY
• 0 X-Ray/ultra-sound 00ther
®-+ 0 0 0 0
l
Calcium (include
0101 mg. or more
0 less than 400 mg. per day
don"t -
know (i)(D(7
® ® (a ..::...
--
'j
·' GASTRIC or DUODENAL ULCER ®-+ 0

1
010 0 dolomite, Turns, etc.)? 0 400-900 mg. ®® s
·'' I ULCERATIVE COUTIS ®---. 0 0 0 0 0Yes •• 0901-1300 mg. ®
--
lj
CATARACT EXTRACTION
OTHER MAJOR ILLNESS
0-+
®-+
0
0
0
0
0
0
0
0
0No-"l.
' Mark if 0Potassium
0 1301 mg. or more
0 Chrom1um 0 lod1ne
don.0
t
know ®
0 Beta-Carotene
--
,!
4 Specify illness r
you take
any of
these
0 Vitamtn D
• 0 B·Complex
Ocopper
0 MagneSium 0 Ruttn
0 Lec1thin 0 Folic Acid
0 Brewers yeast
-- _)No
15. Glaucoma?
........
0 Yes-+ 0betore ·so
If yes, mark year of diagnosis.
0'80.'81 0'82-"83 0'84-'86 O·as or dfter
19.
Current
a) SYSTOLIC 0< 120 11ll'nlig.
Qunknown 0120-139
0140-149
0160-159
0160-169
Ono-+
-- 16. Macular degeneration? tf yes, mark year of diagnosis.
usual
blood b) DIASTOUC 0<75
pressure: o .u.nknown rns-84
Qss-sg
so-94 r
095-104
()100
-
T
()Yes ... ,Obefore '80 ,0 '80-"81 0'82-'83 '84-'85 c

'86 or after r
NURSE'S HEALTH STUDY Page 3 HARVARD MEDlCAL SCHOOL
20. Please enter your ID number (found at the top of page 2)----_.1 J I I I I 1-1 I I --
21. For each food listed, fill in the circle indicating
how often on average you have used the amount AVERAGE USE DURING PAST YEAR --
jill
specified during the past year.
DAIRY FOODS
Never,
or less
1-3
than once per
per month mo
per
12-4 5-6
per per
week week week
1
per
day
2-3
per
day
4-5
per
day
6+
per
day 0 000 -
-
lr
0 @@@ -
I'-.
Skim or low-fat milk (8 oz glass) 0 0 ® 0 0 ® 0 0 0
Whole milk (8 oz glass)
Cream, e.g. in coffee. or whipped cream ( 1 Tbs)
0
0
0
0
@)
@)
0
0
0
0
®
®
0
0
0
0
0
0 iQ ®®0 -
- <!)@G)
Sour cream ( 1 Tbs) 0 0 e 0 0 ® 0 0 0 ,Q. ®00 -

Non-datry coffee whitener ( 1 tsp} 0 0 @) 0 0 ® 0 0 0 Q ®@@ -
Sherbet or ice milk (V2 cup) 0 0 @) 0 0 ® 0 0 0 .j ®®® -
Ice cream (lh cup) 0 0 @) 0 0 ® 0 0 0 0 ®®® -
Cottage or ricotta cheese (lh cup) 0 0 @> 0 0 ® 0 0 0 A
\ 0®0 -
Cream cheese ( 1 oz) 0 0 @) 0 0 ® 0 0 0 ' ®®® -
Other cheese, e.g. Amencan, cheddar, etc. 0 0 @) 0 0 ® 0 0 0 6®®® -
plain or as part of a drsh ( 1 slice or 1 oz serving)
Margarine. added to food or bread ( 1 pat);
0 0 @) 0 0 ® 0 0 0 c5 .ooo -
-
exclude use rn cookrng
Butter, added to food or bread ( 1 pat);
exclude use in cooking 0 0 @> 0 0 ® 0 0 0
·""'-
Q ®®@ -
-."'
Yogurt ( 1 cup) 0 0 (W) 0 0 ® ( 0 0 l5 C.D<D<D -

, ®®® -
0@® -
Never. 1-3 1 2-4 5-6 2-3 4-5 6+
or less
0 0@@ -
per per per per per per per per
than once
FRUITS per month mo week week week day day day day
Raisins ( 1 oz or small pack) or grapes 0 0 @) 0 ( ) ® 0 0 () iO ®®® -
Avocado (1/2 fruit) 0 0 @) 0 0 ® 0 0 0 0 ®®® -
Q 000 -
..
Bananas (1) 0 0 @) 0 0 ® 0 0 0
Please try to Cantaloupe (% melon} 0 0 @) 0 0 ® 0 0 0 Q ®®® -
0 ®®® -
average your
seasonal use Watermelon (1 slice) 0 0 @> 0 0 ® 0 0 0
of foods over
the entire year.
For example, if
1-
Fresh apples ( 1)
Applesauce ('12 cup)
0
0
0 @
0 '®
0
0
0
0
®
®
0
()
0
0
0
0
Q
Q®®® -
- •
a food such as Apple juice or cider (small glass) 0 0 @) 0 0 ® 0 0 0 Q CD<D0 -

cantaloupe is
eaten 4 times a
Fresh pears ( 1) 0 0 ® 0 0 ® 0 0 0 Q®®® -
week during the Canned pears (lf2 cup) 0 0 @) 0 0 ® 0 0 0 0
approximate 3
months that it is
Fresh peaches. aprieots or plums ( 1) 0 0 @) 0 0 ® 0 0 0 0 00 -·· -
Canned peaches. apricots (112 cup) 0 0 @) 0 0 ® 0 0 0 t\.J ®®® -
-
in season, then
the average use
would be once
Other canned fruit. fruit cocktatl (1/2 cup) 0 0 @ 0 0 ® 0 0 0 0 ®®®
per week.
Strawberries, fresh or frozen (lh cup) 0 0 @) 0 0 ® 0 0 0 IU 000 -
Blueberries. fresh or frozen (112 cup) 0 0 @) 0 0 ® 0 0 0 Q®®@-
Prunes. dried (5) or canned
-
cup) 0 0 @) 0 0 ® 0 0 0
f..,
0
-
®®®1-
Oranges (1) 0 0 @) 0 0 ® 0 0 0 0 @@ -
Orange jurce (small glass}
Grapefrutt (1/2)
0
0
0
0
@)
@
0
0
0
0
®
@
0
0
0
0
0
0
Q -
G) G)
®® -
Grapefruit juice (small glass) 0 0 @) 0 0 @ 0 0 0 v ®® -
Other fruit juices (small glass) 0 0 @) 0 0 ® 0 0 0 0 00 -
®® -
Consider the
. .
servrng stze BEVERAGES
Never,
or less
than once
1-3
per
per month mo
1
per
2-4 5-6
per per
week week week
1
per
day
2-3
per
day
4-5
per
day
6+
per
day
-
0 ®CD -
@®
as 1 glass, Low calorie soda. e.g. Peps1 Free. Diet 7 -up, etc. 0 0 @) 0 0 (o) 0 0 0 Q ®®-
bottle or can
•{ Soda with sugar. e.g. Coke. Pepsi, 7 -up. etc. 0 0 @) 0 0 @ 0 0 0 0 ®®-
@@ -
for these
carbonated Hawaiian Punch. lemonade, or other non-
-
•
beverages. carbonated fruit dnnks ( 1 glass, bottle, can)
0 0 @) 0 0 @ 0 0 0 Q
Decaffeinated coffee or non-caffeine tea {1 cup)
Coffee ( 1 cup)
0
0
0
0
@)
@)
0
0·
0
0
@
®
0
0
0
0
0
0
0
0
-
-
Tea with caffe1ne (1 cup)
Beer ( 1 glass. bottle. can)
0
0
0
0
@)
@)
0
0'
0
0
@
®
0
0
0
0
0
0
0
0
-
-
w
Red wine (4 oz glass}

Whrte wtne (4 oz glass)
0
0
0
0
@)
@)
0
0
0
0
@
®
0
0
0
0
0
0
Q
Q
-
-
Please turn
to page 4
Liquor. e.g. whiskey. gin, etc. ( 1 dnnk or shot)
Plarn water ( 1 cup)
0
0
0
0
@)
\
" )
0
0
0
0
®
®
0
0
0
()
0
0
0
0 ;oo-
-
-- Page 4
- 21. (Continued) Please fill 1n

•
-
your average use, Never, 1-3 1 2-4 5-6 1 2-3 4-5 6+
during the past year, o f each specified food. or less
I -
than once per per per per per per per per
/"'""\
VEGETABLES per month mo week week week day day day day
-- String beans (1h cup)

Broccoli ('h cup)
0
0
0
0
@
@
0
0
0
0
®
®
0
0
0
0
0
0
-- Sauerkraut (112 cup)
Coleslaw, t,mcooked cabbage ('h cup)
0
0
0
0
@)
® 0
0 0
0
®
®
0
0
0
0
0
0
;::
t5
-- Cooked cabbage ('h cup}
Cauliflower ('h cup)
0
0
0
0
@ 0
@ 0
0
0
@
®
0
0
0
0
0
0
-- Brussel sprouts ( h cup)
1
Carrots, raw ('h carrot)

0
0
0
0
@) 0
@) 0
0
0
@
@
0
0
()
0
0
0
IY
0
,Q
-- Carrots, cooked (112 cup)
Corn { 1 ear or 1h cup frozen or canned)
0
0
0
0
@) 0
@) 0
0
0
@
@
0
0
0
0
0
0
-- Peas, or lima beans (Vz cup fresh, frozen, canned)
Mixed vegetables (112 cup)
0
0
0
0
@)
@)
0
0
0
0
®
®
0
0
0
0
0
0
-- Beans or lentils, baked or drie<:l (112 cup),
Alfalfa sprouts (1/2 cup)
0
0
0
0
@
®
0
0
0
0
@
®
0
0
0
0
0
0
-- Celery (4-inch st1ck) 0 0 ® 0 0 ®
.,,
0 0 0 -
-- Mushrooms, fresh, cooked, or canned (one) 0 0 @ 0 0 ® 0 0 0 -
Dark yellow/orange (winter) squash ('h cup) 0 @) 0 0 ® 0 0 0
--
0
Eggplant, zucchini. or other summer squash (1/2 cup) 0 0 @) 0 0 ® 0 0 0 -
Yams or sweet potatoes ( 1h cup) 0 0 @) 0 0 ® 0 0 0
-- Spinach, cooked 12 cup)
(1
Sp1nach. raw as in salad ( 1 serv1ng)
0
0
0
0
@)
@
0
0
0
0
®
®
0
0
0
0
0
0
:::::
-- Kale. mustard or chard greens

Iceberg or head lettuce (serving)
Ch cup) 0
0
0
0
@)
@)
0
0
0
0
®
®
0
0
0
0
0
0
lr'\
R
-- -
Romaine or leaf lettuce (serving)
Green pepper (Vz pepper)
0
0
0
0
@)
@)
0
0
0
0
@
®
0
0
0
0
0
0
-n 6
--
1\. ..1
Cucumber ( / 4 cucumber) 0 0 @) 0 0 ® 0 0 0
, - Tomatoes ( 1) 0 0 @) 0 0 ® 0 0 0 IQ
-- Tomato JUtce (small glass)

Tomato sauce. e.g. in spaghetti sauce (1;2 cup)
0 0
0
@)
@
0
0
0
0
®
®
0
0
0
0
0
0
R
t50
--
0
Red chili sauce ( 1 Tbs) 0 0 @) 0 0 ® 0 0 0
Tofu or soybeans (3-4 oz} 0 0 @) 0 ( ) ® 0 () 0 0
-- Never,
or less
1-3 1 2-4 5-6 1 2-3 4-5 6+
-- Eggs { 1)
EGGS, MEATS, ETC.
than once per
per month mo
0 0
per per per
week week week
@) 0 ()
per
day
®
per
day
0
per
day
0
per
day
0
1J
-- Ch1cken or turkey, with skin (4-6 oz) 0 0 @) 0 0 ® 0 0 0

-
1-
-
1-
-- Chicken or turkey, without sl<tn (4-6 oz) 0 0 @) 0 0 ® 0 0 0 _-.
--
Processed meats, e.g. sausage, salami.
bologna. etc. (p1ece or slice)
0 0 @ 0 0 ® 0 0 0 -
Bacon (2 shces) 0 0 @) 0 0 ® 0 0 0
-- Hot dogs (1)
. Hamburger ( 1 patty)
0
0
0
0
@
@)
0
0
0
0
®
®
0
0
0
0
0
0
6
IQ
-- Beef. pork. or lamb as a sandwich or mixed
d1sh. e.g. stew. casserole. lasagne. etc.
0 0 @) 0 0 ® 0 0 0 0
-- Beef. pork, or lamb as a ma1n d1sh. e.g. steak.
roast. ham. Me. (4-6 oz)
0 0 @) 0 0 ® 0 0 0
-- 1-
Canned tuna fish (3-4 oz)
Dark meat fish, e.g. mackerel. salmon.
0 0 @) 0 0 ® 0 0 0 b
-- sardines. bluefish. swordfish (3-5 oz)
Other fish (3-5 oz)
0
0
0 @)
@)
0
0
0
0
®
®
0
0
0
0
0
0
00
Q0
--
0
lobster. scallops as a main dish 0 0 @) 0 0 ® 0 0 0 IC
-- -
Liver: beef. calf or pork (4 oz)
L1ver: chicken or turkey (1 oz)
0 Never
0Never
0 Less than 1/month
0Less than 1/month .
0
0
1/mo
1/mo
2-3/mo
2-3/mo 0
0 1/week or more
1/week or more
-
- Please go
to page 5
How often do you eat meat that was charred during cooking? (e.g. during barbequing or broiling)
0 Never 0 Less than 1/month 0 1/month 0 2 -3/mo 0 1/week 0 2+/week
I
I
I
NURSE'S HEALTH STUDY Page 5 HARVARD MEDICAL SCHOOL -

21. (Continued) Please f ill in your average
during the past year, of each
-
use,
specified f ood.
Never,
or less
. - - . : . . . . - - - - - - - - - - - - - - - ! t h a n once
1-3
per
1
per
2-4
per
5-6
per
1
per
2-3
per
4-5
per
6+
per
..
BREADS, CEREALS, STARCHES per month mo week week week day day day day
Cold breakfast cereal (1 cup) (O._) ..

Cooked oatmeal (1 cup) 0 (w) 0 0 ® 0
Other cooked breakfast cereal (1 r•Jp) , 0
White bread (slice). including p1ta bread ® () ® Q 0 0 [01..
Dark bread '--''· -' 0 0 0 0 §) 0 0 ')[ 1
1
English muffins. bagels. or rolls (1) () @ 0 0 @ 0 .0 IQ)-1-• •
1
Muffins or biscuits (1) () (@ 0 0 (o) () ( ) 0 Ul..
Brown rice (1 cup .J\ .I () rn .. ..
White rice ( 1 cup cooked) () §) 0 ® 0 0 ..
Pasta. e.g. spaghetti. noodles. etc. ( 1 cup cooked) 0 0 ® 0 0 0 ,
1
Other 1,, hulgur. kasha, etc. ( 1 r-un -' -n 0 () , 1 0 0 £> 0 0 Q

Pancakes or waffles (serving) 0 ® 0 (§ 0 0 :-t-+n . .
French fried potatoes (4 oz) 0 @ 0 (o) 0 0 () () . .
I--P
- o_ta=to= --
ci,1_'P_s_o=t:_co_rn_c_h_ip""'s""='(s'-m_a._
ll_bp_g_o_r....:1__.o
..._z_)- -=-+'1,_0
--'--1-'-'-()
---''';tt_@
""'-+ __ o. . . .
-"- () () Q 4
..
Crackers. P.n TriscUits. Wheat Thins, etc. (4) 0 0 (iN) 0 0 ® 0 Q ..
____________________ ..
Never, 1-3 1 2-4 5-6 1 2-3 4-5 6+
ISV\tt:t: 1::s, BAKED GOODS, MISCELLANEOUS t

Chocolate PI Pn Hershey's M&M's, etc. (1 oz)
h
or less
(J
per
mo
0
per
@
per per
week week week
C) 0
per
day
per
day
0
per
day
0
per
day
0 (:
-
Candy bars. P._g_ Snickers, Milky Way, Reeses. etc. (1) 0 0 @ Q 0 (D) 0 0 0 ..... .,),Il • •
Candy without chocolate (1 oz) 0 Q (iN) 0 () @ 1Qj1
Cookies. home-baked (1) l)j1
______________
I Brownies ( 1) 0 0 @) 0 0 ® 0 0 -
1-n-,,n-,,.......ln'-IU,t_s(i-,)
__;_:_, - - - - - - - ---1---=:---+ C> _®__ 0 0 0 0 :r
Cake, horne-baked ( 1 slice) 0 @ 0 0 (o: 0 0 0 ;011111
Cake, ready-made (1 slice) 0 0 @) 0 0 0 0 0
________________ ..
Pie. ready-made (1 slice) @) 0 0 (o) 0 0 () ..
..
Jams. jellies. ""' . :ls. syrup, or honey (1 Tbs) 0 () @) 0 " :0 0 ..
Peanut butter ( 1 Tbs) 0 0 0 (o: Q () ..
.. ..
1. .
t - - -- - - -- - - -- - - -- -- -- - - -- - - --1-- -1 - - - ---1-- ..
Popcorn (1 cup) 0 0 0 ® 0 0 10)1..
rB_r_an_,_a_
dd_e_d_to__
fo_o_d _;_(1_ T_b_s.;_
) ______________ ..
Wheat ger,,l (1 Tbs) 0 0 0 0 @ () 0 0 '-"l
Chowder or cream soup (1 cup) 0 0 G 0 0 0 0 ()
Oil and vinegar dressing, e.g. Italian, etc. (1 Tt:- :: () 0 @) () () @ () 0 0
Mayonnaise or other "' salad dressing (1 Tbs)
CCliiiY 0 0 e 0 0 ® 0 0 0 1{)1--
__________ 0
___________________ ..
Soy or Worcestersh1re sauce (1 Tbs) () () '-:jy' () () @ Q () Q v),l,••
1
Home-made soup with bouillon cubes (1 rtln\ 0 0 (W)-t----:0:=--t--=--

,_H
_ o_m_e..... _bo_u_il--'
lo_n
- _cu_b_e_s . 0 () :01. .
Ready-made soup from can. package or 0 0 _ o o ® o I v J
restaurant (1 cup) ..
22. HOW OFTEN DO YOU EAT BEEF, PORK or LAMB cool<ed these ways?
..
..
a) Roasted ()Never ()Less than 1x/month 0 1-3x/month 01 Weekly -02-4x/week 05+/week ,....... . .
b) Panfried ()Never 0Lessthan 1x/month ..
c) Broiled 0 Never ()Less than 1x/month 0 1-3x/monlh () 1 Weekly 0 2-4x/week 0 6+/week 1111
Please turn d) Barbequed ()Never DLess than 1x/month 0 1-3x/month _Q 1 Weekly 02- 4x/week Q5+/week C ..
to"' 6 e) Boiled or stewed ()Never QLess than 1x/month 0 1-3x/month Q 1 V'.. 'iy 02-4x/week 05+/week """)I-
I
-- NCS Trans-Optic ·ii- EPO 1·23145:321
23. What do you do with the visible fat on your meat?

A9114 Page 6
Copyright 1986 Brigham and Women's Hospital.
All Rights Reserved Worldwide.
29. Teaspoons of sugar added to your food and t - - - - "

I - 0 Eat most of the fat 0 Eat as little as possible beverages each day (e.g. coffee, cereal, etc.)
TSP
I - 0 Eat about half of the fat 0 (Don't eat meat)
30. Usual type of cold
I - 24. Kind of fat most often used at home for baking: breakfast cereal:
- 0 Real butter 0 Vegetable Oil 0 Lard
SPECIFY BRAND AND TYPE
- 0 Marganne 0 Solid vegetable shortening 0 None
31. Usual type of
- 25. Kind of fat usually used for frying and sauteing (Exclude cooking oil:
- "Pam"-type spray)
SPECIFY BRAND AND TYPE
- 0 Real butter 0 Vegetable Oil 0Lard
32. How often are your midday meals prepared at home?
- 0 Margarine 0 Solid vegetable shortening 0None
- 26. What form of margarine do you usually use? 0 Never 0 1-2/week 0 3-4 0 5-6 0 7 days
- 0 None 0 Stick form 33. How often are your evening meals prepared at home?
- 0 Tub form 0 Diet form (low calorie or spread) SPECIFY BRAND AND TYPE 0 Never 0 1-2/week 0 3- 4 0 5-6 0 7 days
- 27. How many shakes of salt do you add 34. How often do you eat food that is fried at home
- to your food at the table each day? SHAKES
(exclude "Pam" -type spray):
- 28. How much salt is added during cooking to these 0 Less than 1/week 0 1-3/wk 0 4-6 0 Daily
-- home made foods per serving:
a) Meat
..-No--ne__,.-,/-S-tsp--.-,-
0
/4_tsp__,.-1!-2+
0
__ts...,P 35. How often do you eat food that is fried away
from home (e.g. Fried chicken, fish etc.)
I - b) Vegetables 0 0 0 0 0 Less than 1/week 0 1-3/wk 0 4-6 0 Daily
- c) Staple foods (e.g. rice 36. In a typicaJ week, on how many days do
- pasta. potatoes. etc.) 0 0 0 0 you have any form of alcoholic beverages?
----'-
I - d) Soup 0 0 0 0 0 None 0 1-2 days 0 3-4 0 5-6 0 7 days/week

..----------------------------------------------.. .--. ..
-- 37. Please think about the years you were in High School (i.e. about ages 13-18).
How often did you eat the specified amounts of these foods? We understand this is difficult, but
please make your best estimate:
-- Never,
or less
than once
month
1-3
per
1 2-4 5-6 1 2-3 4-5 6+
Skim or low-fat milk (8 oz glass)

- Whole milk (8 oz glass)
- Milk shake 1)
- - Ice cream ('h cup)
- Hard cheese 1 slice or 1 serving)
- 1 Margarine (1 pat)
- Real butter ( 1 pat)
- Apples (1)
- Orange juice (small glass)
- Cabbage. coleslaw (112
- Broccoli or caultflower (1/:! cup}
- Carrots ( 1 raw or 'h cup cooked)
- cup cooked)
- Eggs (1)
..
- 1 Hot dogs (1)
Other beef. pork, lamb ( 1 serving)
- Fish, tuna fish 3-6 oz)
l - Bread (slice)
Rice ( 1 cup cooked)
- Potatoes: baked, boiled, mashed ( 1 cup)
I - French Fried potatoes (4
- 1 Cold cereal ( 1 cup)
Cooktes
- Vitamin pills or capsule
- 38. When you were 18-20 years old, how many drinks of beer, wine and/or liquor did you have per week?
0
..------------------------------------------------------------------------------------------------------------
- 0None 0Less than 1/week
39. (Optional) If a tape measure is convenient,

Oonce per week 02-3/wk
WAIST (Measure while standing.

04-6/wk 07-10/wk 10+/wl<
-
-
please record these measurements -------+
to the nearest 1/4 inch (do not estimate).
relaxed.) L.-1-----'1'------'11 inches
: HIP (Largest crrcumference} ._l_ __.._l_____.ll inches
-
UPPER ARM (Circumference with arm hang1ng
straight down. halfway between
QA inches
- shoulder and elbow) 1 1 1·
-
------------------------------------------------------------------------------------------------------------
THANK YOU! Please return fonns in prepaid retum envelope. Nurse's Health Study, 180 Longwood Ave., Boston, MA 02115
I
1
I
I
HARVARD HARVARD
MEDICAL
SCHOOL
STUDY II SCHOOL
of
PUBLIC HEALTH
• Harvard School of Public Health • 677 Huntington Avenue • Boston, Massachusetts 02115 • (617) 432-2279 •
------------ -
Dear Colleague:
On behalf of our research group, I again want to express my thanks for your participation
in the Nurses' Health Study II. We have now finished the processing of the baseline
questionnaire that you returned almost two years ago. In total, 116,680 nurses fully
completed the forms and thus comprise the population for this important study of lifestyle
factors, diet, and oral contraceptive use in relation to breast cancer and other important
health issues among women. ·rhe accuracy and completeness of the information you provided
is truly impressive, and we are confident that this study wHl provide answers to many
critical questions.
The enclosed questionnaire marks the beginning of the follow-up phase of the study. You
will note that we ask about your current status for many of the same questions that we
posed earlier. We also ask about new medical diagnoses and conditions that have occurred
since September, 1989. This date was chosen because the vast majority of participants
completed the initial questionnaire during that month or shortly thereafter. We have also
included a complete dietary assessment using a questionnaire developed and validated as
part of the Nurses' Health Study I. This will provide important information about the effects
of diet on medical conditions that occur in later life.
We hope that you give this questionnaire the same attention and care that you did in
completing the baseline form. The validity of this major research undertaking depends
directly on complete and accurate follow-up information for all study members. We know
that some participants are no longer in active nursing. However, the continued participation
of all study members is critical regardless of current employment status. As always, the
information you provide is strictly confidential and will be used only for medical statistical
purposes.
Thank you again for your invaluable participation in this study. We will be sending you the
RESEARCH
next edition of our newsletter in June of 1992 to update you on the progress of the
GROUP
investigation.
Stefante Bechtel. B.A. Sincerely,
Gary Chase. BS.
Graham Cold1tz. M.D.
Sue Hankinson, R N.
Davie! Hunter. M.D.
Walter Willett, M.D.
I JoAnn Manson, M.D.
Professor of Epidemiology and Nutrition
I Laura Newcomer. B.S
I Frank Spetzer. M.D. P.S. We would be extremely grateful if you could return the questionnaire within the next 2 weeks;
I Me•r Stampfer. M.D. due to federal reductions in research support, our budget for rernailings is very limited.
Lisa Troy, B.A.
NURSES' HEALTH STUDY II HARVARD UNIVERSITY
INSTRUCTIONS
Please use an ordinary No. 2 pencil to answer all questions. Fill in the appropriate response
circles completely, or write the requested information in the boxes provided. Note that some
questions ask for information since September 1989, some ask for current status, and some
ask about events over longer periods. The form is designed to be read by optical-scanning
equipment, so it is important that you make NO STRAY MARKS and keep any write-in
responses within the spaces provided. Should you need to change a response, erase the incorrect
mark completely. If you have comments, please write them on a separate piece of paper.
CURRENT
WEIGHT
EXAMPLE 1. Write your weight in POUND'
the boxes... l4- 0
@ ® •
...and fill in the circle (i) (i)
corresponding to the 0 0 0
figure at the head of ® ® 0
each column. @ @
® ® ®
Please fill in the ® ® ®
circle completely, do 0 0
not marl< this way: ® ®
® ® Q 0 ® ®
EXAMPLE 2: Keep handwriting within borders of the response box.

What specific brand do you smoke? Specify exact brand and type:
=-==--::1)
(e.g., "Marlboro Lights 100's")
A'A.RtBORO Lf,/ff1S too's
EXAMPLE 3: Mark "Yes" bubble and Year of Diagnosis bubble

for each illness you have had diagnosed.
16. Since SEPTEMBER 1989 have you
had any of these physician-diagnosed
Bt;FORE . 89 AFTER
illnesses? SEPT TO JUNE l
LEAVE BLANK FOR "NO". MARK HERE FOR -ves"---L 1989 MAY 91 t981
Thank you for completing the 19.91

Nurses' Health Study II Questionnaire
Please tear off the cover letter (to preserve confidentiality} and return
the questionnaife in the enclosed postage paid envelope.
- NURSES' HEALTH STUDY II HARVARD UNIVERSITY
-- PLEASE USE
PENCIL! 2.
PAGE 1
What was the natural color of your hair at age 18? ®

-- 1.
WEIGHT
3.
0Red 0Bionde 0 Light Brown
Do you currently smoke cigarettes?
0Dark Brown 0Biack
®
--
POUNDS
0 Yes --+ a) How many per day? 01-4 05-14 015-24 025-34 035-44 045 or more 0
® ® ® 0No b) What specific brand do you smoke? ®
--
Specify exact brand and type:
CD CD (i) (e.g.. "Marlboro Lights 100's")
® ® ® ..
-- ® ® ®
0 0 @
4. Are you CURRENTLY pregnant?
0Yes 0No
@
-- ®
®
®
®
®
®
5. Since September 1989, have you been pregnant? (Do not include current pregnancy or those
ending before Sept. 1, 1989.)
'®
-- 0
®
0
®
0 Yes
rr- 0No
-+a) Nurnber of pregnancies lasting LESS THAN G months Ozero 01 02 03 or more
b) Number of pregnancies lasting 6 months or more Ozero 01 02 03
@
®
-
-- 6. ( ) ® ® •
-
c) Did you GIVE BIRTH to twins or triplets?
Have you tried to become pregnant for more than one year without success since September 1989?
0No 0Twins 0 Triplets {§)
®
-- 0Yes
0No
What was the cause? {Mark all that apply.)
0 Tubal blockage 0 Ovulatory disordbr 0 Endometriosis 0 Cerv1cal mucous factors
0
- 7.
0 Srl")use 0 Not investigated
Have your m enstrual periods ceased
0Not found 00ther
AGE
...
PERMANENTLY?
-- 0 No. Premenopausal
0 Yes: No menstrual periods
"
....
,' a) Age
®
natural ®
b) For what reason did your periods cease?
0 SURGERY It due to surgt:.ry, wer(; yoUI ovanes
0
@
-- 0 Yes: Had menopause but now have
periods induced by hormones
periods
ceased?
® ®
CD CD removed?
Yes. both 0 0 Only uterus removed
@
@
-- 8. 0Not sure
SINCE SEPTEM.B ER 1989, have you used f em ale
®
@
®
0
Oone only
0 RADIATION or Cl iEMOTHERAPY
@
-- replacement hormones (other than oral contraceptives)?

0 No a) How many months have you used them
® ®
® ®
0 NATURAL· II natural (non-surgical} menopause.
have you had subsequent surgery to remove
-- - 0 Past
0 Currently
since September 1989?
0 1-4 mo. 0 5-9 0 10- 14
0
®
ovaries 01 uterus?
No 0 0
One ovary removed
-- 16 19 0 0
20+ mot1ths ®
b) What type of hormone have you used the longest during this period?
0 Uterus removed 0 Both ovaries removed
@
-- 0 Oral Premar1n or other c.OnJliQElted estrogen
0 Oral conjugated estrogen and progesterone (e.g. Provera)
0 Oral progesterone alone (e.g.. Provera)
0 Vaginal estrogen Plooso spoc•fv othor hormone:
®
®
-- 0 Patch estrogen alone
0 Patch estrrgen and progesteronC"
0 Other
(e.g. non COilJUOated e >tro::Jen)
-- c) If this was oral conjugated estrogen (e.g. Premarin) what dose did you usually take?
0 Nat used 0 Dos€ unknown 0 .30 mg/clay or lAss 0 625 mg/day 0 .9 mg/day or more
- d) Dose of progesterone
(if taken)?
0 Not used 0 Dose unknown 0 < 5 mg 0 5-9 rng 0 10 mg 0 More than 10 mg
9. Do y ou CURRENTLY use any of these forms of contraception? (Marl< all that apply.)
-- 0 None
0 Condom
0
Oral contraceptive
0
Intrauterine device
0 NorplanL
0 Rhythm/NFP
0 Diaphragm/Cervical cap
0 Vasectomy
0 Tubal hgar1on
0 Sponge
0 Foam or jelly
00ther
- 10. Have you EVER used oral contraceptives (OC's) for 2 months or ®
-- ! more for any reason (contraception, acne, etc.)?
0 Yes 0 No· Go to Question 12
®@®®0®®0®0®0®@@®®0®®
®@®®0®®0®®®0®®0®®0®®
-
-
11. SINCE SEPTEMBER 1989, have you used oral contraceptives?
®CD®®@®®0®®®0®®@®®0®®
2>0®®®® ®'
-- 0 Yes
0No
a) How many months have you used OC' s since September 1989?
01 or less 1"1onths Q2-4 05-9 0 10 14 015 19 020 or more months
®CD®®0®®0®®®
@0®®0®®0®®®
--
I'
b) Please indicate the brand and type of OC used longest during this
time period. Refer to the OC Brand Code Sheet enclosed with this
-- 12. questionnaire and wrjte the code in this box.
Since September 1989, how many months have you worked ROTATING night shifts (at least 3 nights/month in @I
-- addition to other days and evenings in that month)?
0None 0 1-4 mo. 05-9 010-14 015-19 020+ months
13.
-- How many times per week do you engage in physical activity long enough to perspire heavily (including swimming)?
0 Less than once/week 0 Once/week 0 2-3 times/week 0 4-6 times/week 0 7 or more times/week
@
-- 14. How m any FLIGHTS of stairs (not individual steps) do you climb daily?
0 2 flights or less 03-4 05-9 010-14 0 1 5 or more flights
®
- Copyright 1991. The President and Fellows <'f Harvard College. Marl< by NCS EP-41537:321 A9114 Printed in U.S.A. •
PAGE 2
15. Is this your correct date of birth?-
r0 Yes 0 No-+ If no, please ® ® ® ® ®
+ write correct date. MONTH DAY YEAR @ 0 ® @ @
16. Since SEPTEMBER 1989 have you @ ® ® ® ®
had any of these physician-diagnosed
illnesses?
LEAVE BLANK FOR "NO" MARK HERE FOR ·'Yes·• THIS IS YOUR ID
. Have you EVER had any of
these physician-diagnosed
BEFORE SEPT 89 AmR
illnesses? SEPT m JUNE 1
LEAVE BlANK FOR
HERE FOR "YES"
MARK - - ---,
1989 MAY 81 1991 1\c-J®@-
V infection
Kidney stones
-
Urinary tract infection ®®-
us cell skin cancer Pneumonia ®®-
- - Frbrocyst1c/other benign breast dis. Hirsutism f acial ha1r) 00-
Conftrmed by breast btopsy?@ No 0 Yes Multiple sclerosis- 1st Ox \.:./1\.::,1-® ®-
Conftrmed asp1rat1on? No ®-
-
:.lt'-=1'®
Breast cancer Asthma- 1st 1cian Ox

Other cancer: Ulcerative colitis s
Specify stte of 18. In how many months did you practice breast
other cance1- . self-examination in the past year?
Hrgh blood pressure {excluding 0 .::::O
:L.:.:No:: :. n:.: :e @) @ -
dunn 19. Since September 1989,
YES. FOR
CD CD CD
YES, FOR 0 t=;'\ a\ t=;'\
-
Diabetes: Gestational have YOU had: NO SCREENING SYMPTOMS 0 \b \b -
Mammogram @ 0 0 ')®®® -
Elevated cholesterol Breast exam by clinician ® 0 0 @@® -
Colonoscopy/Sigmoidoscopy @ 0 0 0 ®@ ® -
arthritis, doctor chRgnr)sf.lrJ 20. Mark if xou use: 0 Insulin 0 Oral Hypoglycemic s @@ -
--7 .... factor tJve/Unknown 21. Regarding YOUR infancy: 0 0 CD-
Other arthritis a) W ere you breast fed? ® ® ®-
I Colon or rectal 0 Yes • Number of months?
I
®®®- ·
0 No 0 Unknown 0 or less @@) -
0 Don't know 0 48month 09+tll011ths ®@ -
- Gall stones b) Your birthweight in pounds: 0 Unknown ®-
a) Dtd you have svmpwms? ®No 0 Yes 0 <5.5lbs. 0 5 .5·69 0 7-8.4 O a.5-9.9 010tlbs.
b) How dt 1gnosed? 0 X· ray or 0 Other -
LJilrasound 0 rull-term 0 2I Weeks prornaMe 0 Multiple btrlh ,-
Toxernia/Pre-eclampsia of pregnancy 22. Please indicate the name of someone at a DIFFERENT -
(raised blood re and pro · PERMANENT ADDRESS to whom we might write in -
- · Other major illness or surgery since the event we are unable to contact you: -
Sept. 1989 specify other major illness or surgery: Name: __________________________________________________
Address: _ __ _ _ _ _ _ _ _ _ __ _ _ _ __
23. DURING THE PAST YEAR, what was your average time PER
WEEK spent at each of the following recreational activities?
--
Walk11 or h1 outdoorc: elude
than 10 minutes/mi
rr to work
--
include stationary machine
--
Calisthenics Aerobics/ Aerobic Dance
Tennis, Squash. Racquetball
Machine --
..
-
24. On average, how many HOURS PER WEEK do you spend: z.ERO oNE 2-s &-10 11-20 21-40 41·60 61-90 ovER 90 -
HRS. HOUR HRS. HRS. HRS. HRS. HRS. HRS. , HRS. -
--
I
or wal around at work or away from home?
or wal around at home? hrs./week
--
-
PAGE 3 D91NHS2 Printed in U.S.A.
Please copy your 10 from page 2 to here.
ID: DDDDDD D
b) What specific brand do you usually use?
26. Not countir"!9 multi-vitamins, do you take any of following preparations:

'
a) Vitamin A? How many 0 0- .I yr. 0 2-4 years

years?
0No 0 Yes, seasonal only
If YES,
0 Yes. most months
How many
b) Vitamin C? years?
Oo 1 yr.
0No 0 Yes, seasonal only
0 Yes. most months What dose 0 Less than 0400 to O 750 to 01300 mg.
per day? 400 mg. 700 mg. 1250 mg. or more
c) Vitamin E?
How many years? Qo-1 yr 02 4 years 05 9 years 0 10i· years
0No 0Yes If YES, What d ose p er day ? Qless than 0 100 to Q300 to Q600 IU or
100 IU 2501U 5001U more
d) Vitamin 8 6 7
How many years? 0 0-1 yr. 02-4 years 05-9 years 0 10+ years
Don't
Oknow
ODon't
e:
0No QYes If YES, What dose per day? OLess than 010 to 040 to 080 mg or
10 mg. 39 mg. 79 mg. more know
e) Selenium? Don't
How many years? Q0-1 yr 02-4 years 05-9 years 0 10+ years 0know
QNo QYes ==:··'·· If YES,
-
What dose per day? QLess 1hfln oao to 0140 to 0260 meg QDon't
80 mcq 130 meg. 250 meg. or rnore know
Don't
f) Iron How many years? 0 0 - 1 yr. 02·4 years 05 9 years 0 10+ years 0know
0No QYes If YES, What dose per day? 0 Less than 051 to 0201 to Q401 mg QDon't
200 mg. 400 mg. or more know
mg. Don't
g) Zinc
How m any y ears?
0-l yr 0 02-4 years 05-9 years 0 10+ years 0know
0No 0Yes If YES, What dose per day? QLess than 025 to Q75 to 0 101 mg. ODon't
25 mg. 71.1 mg. 100 mg. or more know
h) Calcium Don't
(includfJ Calcium in Dolomite
nnd Tums, ntc.) How many years? Qo
02-4 years 05 9 years
1 yr. years 0 0know j
QNo QYes If VES, What dose per day? Less than Q400 to O 0901 to 01301 mg. QDon'tQ
400 mg. 900 mg. 1300 mg. or more know
i) Are there other Bett.I-
0
Vrtamrn D Cod hver oil 0Other
Frsh oil 0Ntar.in0 Ocarotene Other (please specify) 0
supplements that you 0
s-rornpl Folic acrd 0Brewer's Olodine 0 0Magnesrum
take on a regular basis?
---- yeast
21. For each food listed, fill in the circle indicating how often on AVERAGE USE LAST YEAR
average you have used the amount specified during the NEVER ®
past year. OR LESS
TH.AN ONCE
0
0
-- Cream. e.g. coffee. whipped (Tbs)
Sour cream (Tb$)
0
0
-- ®®® Non-dairy coffee whitener (tsp.)
---------
Sherbet. rce milk or frozen yogurt ( 1/2 cup)
0
0
00 Ice cream 1/2 cup} 0
-- 3@@2@® Yogurt (1 cup)
Cottage or riootta
0
0
-- u®®
'--+-.....;
4@@ Cream cheese ( 1 oz.) 0
Otl1er cheese, e.g. Arnerican. cheddar, etc.. plain or 0
-- 000
6@@ as part of a dish ( 1 slice or 1 oz. serving)
Margarine (pat). added to food or bread; exclude use 0
-- ®®®
®®®
.
1n
Butter (pat), added to food or b read: exclude use 0 0 @) 0 0 0
I
(
-- in
I
I
--
PLEASE
TURN TO
a) What form of margarine do you usually use?
0Nore Form? Osock 0 .ub Qsqueeze (liqu1d)
Whnt specific brand and type (e,g .• Parkey Com Oil Spread)?
- PAGE 4 ..J. Type? 0 Regular 0 Light 0 Extra L•gl'lt

27. (Continued) Please fill in your average use, during the past
PAGE 4
NEVER
- I
year, of each specified food. OR LESS 1.3 t 2-4 6-6 I 1 2-3 4-5 6+ I
- --
- -
FRUITS',. - --:-:-
-
-- -
THAN ONCE PER
PER
PfR MONTH MONTH WE£K
PER
WEEK
PER
WEEK
PER
DAY
PER
DAY
PER
DAY
PER
DAY -J
o-
-
- -
r-
-0
-
,..
Please try to
0 0 ,) r r """'
average your
Ra1slns (1 oz. or small pack) or grapes \::/ - '-- J
Prunes (7 prunes or 112 cup) 0 0 ® 0 0 ® 0 0
seasonal use of 0
Bananas (1) 0 0 ® 0 0 ® 0 0
foods over the ® 0 0 ® 0 0 0
Cantaloupe (1 I 4 melon) 0 0
entire year. For Avocado (112 fruit or 112 cup) 0 0 @) 0 0 @ o. 0 0
example, if a food ® 0
Fresh apples or pears (1) 0 0 @) 0 0 0 0
such as cantaloupe 0 @) 0 0 @ 0 0 0
Apple jurce or cider (small glass} 0
is eaten 4 times a ® 0 0 0
Oranges (1) 0 0 @) 0 0
week during the Orange juice (small g)ass) 0 0 @) 0 0 ® 0 0 0
approximate 3 ®- 0 0
Grapefruit ( 1/2) 0- 0 @> 0 0 i-0
months that it is in
season, then the
Grapefruit juice (small glass)
Other fruit juices (small glass)
0
0
0
0
@)
@)
0
0
0
0
@
®
0
0
0
0
0
0
--
average use
would be once per
Strawberries. fresh, frozen or canned (112 cup}
Blueberries. fresh, frozen or canned ( 112 cup)
0
0
0
0
@)
@)
0
0
0
0
@
®
0
0
0
0
0
0 1
--
week. Peaches, apncots or elums ( 1 fresh. or 1 12 CUE_ canned) 0 0 @) 0 0 @ 0 0 0 i
--
NEVER
OR LESS 1-3 1 2-4 5-6 1
f
2-3
I
4-5 6+
--
_.------,.----! THAN ONCE
--- II... ' t ..._ ...

PER
PER
'• :P£RMONTH MONTH WEEK
-
PER
WEEK
PER
WEEK
PER
DAY
PER
DAY
PER
DAY
PER
DAY --
-0
-
) 0 @
Tomatoes (1) l'\1
- - -
--
'-" J
Tornata juice. V8 (small glass)

- 0 0 @) 0 0 ® 0 0
Tomato sauce ( 112 cup) e.g. spaghetti sauce
- 0 0 @) 0 0 ® 0 0 0
Red chili sauce ( 1 Tbs)
Tofu or soybeans (3·4 oz.)
0
0
0
0
@)
® 0
0 0
0
@
@
0
0
0
0
0
0
-
String beans ( 112 cup) 0 0 ® 0 0 ® 0 0 0
Broccoli ( 112 cup)
-
0 0 @) 0 0 ®- 0 0 0
Cabbage or aole slaw ( 112 cup) 0 0 @) 0 0 ® 0 0 0
Cauliflower ( 112 cup) 0 0 @) 0
- 0 @ 0 0 0
Brussels sprouts ( 112 cup)
- 0 0 @) 0 0 ® 0 0 0
Carrots, raw ( 112 carrot or 2-4 sticks) 0 0 @) 0 0 @ 0 0 0
Carrots, cooked ( 112 cup) or carrot ju1ce (2-3 oz.)
- 0 0 ® 0 0 ® 0 0 0 1
Beets - not greens ( 112 cup) 0 0 @) 0 0 @ 0 0 0
Corn ( 1 ear or 112 cup frozen or canned) 0 0 @) 0 0 ® 0 0 0 •
p eas or lima beans (1 12 cup fresh. frozen, canned) 0 0 @) 0 0 @ 0 0 0
M ixed vegetables (112 cup) 0 0 @) 0 0 @ 0 0 0
Beans or lentils. baked or dned ( 112 cup) 0 0 @) 0 0 ® 0 0 0
Dark orange (winter) squash ( 1/2 cup) 0 0 @) 0
- 0 @ 0 0 0
Eggplant. zucchini or other summer
squash ( 1.[2 cup)
0
.-
0 ® 0 0 ® 0 0 0 --
--
• -
Yams or sweet potatoes ( 112 cup) 0 0 @) 0 0 @

-0 0 0
Spinach. cooked ( 112 cup) 0 0 @) 0 0 @ 0 0 0
Sp1nach, raw as in salad (serving)
Kale. mustard or chard greens ( 112 cup)
0
0
0
0
@)
®
0
0
0
0
@
®
0
0
0
0
0
0
--
Iceberg or head lettuce (serving)
Romaine or leaf lettuce (serving)
0
0
0
0
@)
@)
0
0
0
0
®
@
0
0
0
0
0
0
-
Celery (4" stick)
Green peppers (3 slices or 1I 4 pepper)
0
0
0
0
®
@>
0
0
0
0
®
®
0
0
0
0
0 !
0
--
Onions as a garn1sh, or in salad (1 slice) 0 0 ® 0 0 @ 0 0 ot --
On1ons as a vegetable, nngs or soup ( 1 on1on) 0 0 @) 0 0 @ 0 0 0
--
NEVER
OR LESS
ONC.E
1-3
PER
1
PER
2-4
PER
WEEK
5-6
PER
WEEK
1
PER
DAY
2-3
PER
DAY
4-5
PER
DAY
6+
PER
DAY
-
Eggs (1) ..) t
Cf:licken with skin (4-6 oz.)
•
PLEASE
Chicken with0ut skin (4-6 oz.)
Turkey, oz. or 2
0-
G-
TURN TO Hot dogs (1) - I
PAGE 5 Bacon (2 slices)
I
-- 21. (Continued) Please fill in your average use, during the past
PAGE 5 Mork Reflex by NCS EP-41538:321 A9114
I -
j
NEVER
yearc of each specified food. OR LESS 1-3 1 2-4 5-6 1 2-3 4-5 6+ ®
I -
r -----' lliAN ONCE PER PER PER PER PER PER PER
MONTH
- WEEK ®
1 - Processed meats, e.g. sausage, salami. bologna, etc. J 0
-- - (piece or shce)
Hamburger ( 1 patty)
-- Beef, pork. or lamb as a sandwich or mixed dish.

e.g. stew. casserole. lasagna. etc.
0
-- Pork as a main dish. e.g. ham or chops (4-6 oz.)

Beef or lamb as a main dish. e.g. steal<. roast
0
0
-- (4-6 oz.)
Canned tuna fish (3-4 oz.) 0
-- Dark meat fish, e.g. mackerel, salmon. sardines.
bluefish, swordfish (3-5
0
--
UL. ,
Other fish (3-5 oz.} 0

Shrimp, lobster. scallops as a ma1n dish _________________ 0
--
-- NEVER
OR LESS
ONCE I 1-3
PER 1 PER
1
l 2-4
PER
5-6
PER
L 1
PER
2-3
PER
4-5
PER PER
6+ ®
- PERl MO.Nlll DAY .;.. DAY T
--
WEEK W_EEK WEE.K OA·......,
Cold breakfast cereal (1 cup) ( I (_ _; (_ (JC (J ' 1 0
Cooked oatmeal/cooked oat bran (1 cup) 0 0 ® 0 0 @ 0 0 0 0
-- Other cooked breakfast cereal (1 cup}
Wh1te bread (slice). including pita bread
0
0
0
0
@) 0
@) 0
0
0
@
@
0
0
0
0
0
0
0
0
-- Dark bread (slice), including wheat pita bread
English muffins, bagels. or rolls ( 1)
0
0
0
0
@) 0
@) 0
0
0
® 0
@ 0
0
0
0
0
0
0
-- Muffins or b1scuits (1) 0 0 @) 0 0 @ 0 0 0
0
0
0
--
Brown rice (1 cup)
White rice (1 cup) 0 @) 0 0 0 @) 0 0 0
Pasta. e.g. spaghetti, noodles. etc. (1 cup) -----------+--0=--+-0-=---+-@)-=w'--i- 0 0 @ 0 0 0 0
-- Tortillas (1) 0 0 @) 0 0 @ 0 0 0 0
-- - Other grains. e.g. bulgur. kasha. couscous. etc. ( 1 cup)
Pancakes or wafnes (serving)
0 0 @) 0
0 0 ® 0
0
0
@ 0
@ 0
0
0
0
0
0
0
-- French fned potatoes (4 oz.)
Potatoes, baked. boiled ·(1) or mashed ( 1 cup)
0 0 @) 0
0 0 @) 0
0
0
@ 0
@ 0
0
0
0
0
0
0
-- Potato chips or corn chips (small bag or 1 oz."'-
Crackers, Tnscuits. Wheat Thins (5)
)
0 0 @) 0 0 @ 0 0 0
0
0
-
-- Pizza (2 slices) 0 0 @) 0 0 @ 0 0 0 0
-- NEVER
OR ... 1• 3 I 1 2-4 .
56 1 2-3 4-5 6+ ®
-- CARBONATED .I THAN ONCE
PER:M6_N
PER
MONTH
PER PER
WEEK..J..WEEf< t
PER I PER
WEEK....,_ DAV DAV
PER -1- PER
DAY DAY -!
- - -0
....
-0 0- @
,.-
Low-calorre cola. e.g. Diet Coke with caffe1ne (
'-.., -@) I
-
("\ L, ( I
0
-- BEVERAGES CALORIE
Consider the
LOW·
(sugar tree)
Low-calorie caffeine-free cola
Other low-calone carbonated beverage. e.g. Ftesca.
-- 0
0
0
0 ® 0 0
0
@ 0
0
0 0
0
0
glass. bottle or
can for these
--
servrng srze as 1 TYPES
REGUlAR
D1et 7-Up. diet g1nger ale
Coke, Pepsi. or other cola with sugar 0 0 @) 0 0 @ 0 0 0 0
carbonated
beverages -- TYPES
(not s·.,. • free)
Caffeine Free Coke. Pepsi, or other cela with sugar
Other carbonated beverage with sugar, e.g. 7-Up
0
0
0
0
@)
®
0
0
0
0
@
@
0
0
J
0
0
0
0
0
-- OTHER
BEVERAGES
Hawaiian Punch. lemonade. or other non-carbonated
fruit drinks ( 1 glass. bottle. can)
0 0 ® 0 0 @ 0 0 0 (J
-- Regular Beer ( 1 glass. bottle, can)

Light Beer. e.g.. Bud Light ( 1 glass, bo.ttle. can)
0
0
0
0
@
®
0
0
0
0
@ 0
@)0
0
0
0
0
0
0
-
-- Red wine (4 oz. glass)
White wine (4 oz. glass)
0
0
0
0
@)
@)
0
0
0
0
@ 0
® 0
0
0
0
0
0
0
-- Liquor, e.g. whiskey. gin. etc. (1 drink or shot)
Plain water. bottled or tap ('1 cup or glass)
0
0
0
0
@)
®
0
0
0
0
@ 0
® 0
0
0
0
0
0
0
-- Tea ( 1 cup). not herbal tea
Decaffemated coffee ( 1 cup)
0
0
0
0
@)
@)
0
0
0
0
@ 0
@ 0
0
0
0
0
0
0
-- 27b.
Usual method of } Decaffeinated: 0
Coffee with caffeine (1 cup)
Mainl, filtered 0 Mainly instant 0
0
Mainly espresso or perc.
0 @) 0 0 0
0 No usual method/don't know/don't use
@ 0 0 0
0
-
preparing coffee 0
Caffeinated: Mainly filtered 0 Mainly instant 0 Mainly espresso or perc. 0 No usual method/don't know/don't use 0
21. {Continued) Please fill in your average use, during
PAGE 6
NEVER
--
the past year, of each specified food. OR LESS 1-3 1 2-4 5-6 1 2-3 4-5 6+
®-
- : .. - sv\lEEis,sAI<Eo Gooos,
'
Chocolate (bar or packet)' e.g. Hershey's, M & M 's

" .
THAN ONCE
-P ONT MON
0
PER
0
PER
EEl<
@)
PER
WE
0
PER
WEEK
0
PER
DAY
@
PER
DAY
0
PER
DAY
0
PER
DAY
0 o-
-
Candy bars, e.g. Snickers. Milky Way, Reeses 0 0 @) 0 0 @ 0 o- 0 0
Candy without chocolate ( 1 oz.) 0 0 @) 0 0 ® 0 o- 0 0
Cookies. home baked ( 1) 0 0 @) 0 0 ® 0 o- 0 0
Cookies, ready made ( 1) 0 0 @> 0 0 ® 0 o- 0 0
Brownies ( 1) 0 0 @) 0 0 @ 0 o- 0 0
Doughnuts ( 1) 0 0 ® 0 0 @ 0 0 0 o-
Cake, home baked (slice) 0 0 ® 0 0 ® 0 0 0 o-
Cake. ready made (slice) 0 0 ® 0 0 ® 0 0 0 o-
Pie, homemade (slice) 0 0 @) 0 0 @ 0 0 0 o-
Pie. ready m ade (slice.) 0 0 ® 0 0 ® 0 0 0 o-
Sweet roll. coffee cake or other pastry, home baked 0 0 @) 0 0 @ 0 0 0 o-
(serving)
Sweet roll. coffee cake or other pastry, o o ® o o ® o o o o--
ready made (ser\ling)
- -
Jams. jellies, preserves. syrup. or honey ( 1 T bs) 0 0 @) 0 o ® o o o o--
Peanut butter ( 1 Tbs) 0 0 @> 0 0 @ 0 0 0 o-
-
Popcorn ( 1 cup) 0 0 @) 0 0 @ 0 0 0 o-
Peanuts (small packet or 1 0 0 @) 0 0 @) 0 0 0 o-
Other nuts (small packet or 1 oz.) 0 0 @> 0 0 ®
-- 0 0 0 o-
Oat bran. added to food ( 1 Tbs)
- 0 0 @) 0 0 ® 0 0 0 o-
Other bran. added to food ( 1 Tbs) 0 0 @> 0 0 @ 0 0 0 o-
Wheat germ ( 1 T bs)..;.._ _____ 0 0 @)
- 0 0 @ 0 0 0 o-
Chowder or cream soup ( 1 cup) 0 0 -® 0 0 @ 0 0 0 o-
Olive oil salad dressings ( 1 T bs) 0 0
--
@) 0 0 @ 0 0 0- o- -
0 0 0 0 o-
-
Other oil and vinegar dressing. e.g. Italian ( 1 Tbs)

- 0 ® 0 0 @ -
(j o- -
- .
M ayonnaise or other creamy salad dressing ( 1 T bs) 0 0 @) 0 0 ® 0 0
Salt added at table ( 1 shake) 0 0 ® 0 0 ® 0 0 0 o-
0 0 0 o-
-
Garhc ( 1 clove or 4 shakes) 0 @) 0 0 ®
28. beef, calf or pork L4_oz.) 0Nevet· 0 Less !han 1/mo. 0 '1/mo 02·3 lllO. 0 1/weel< ur mo1e -
-
@-
Liver: chicken or turkey ( 1 oz.) 0 Never 0 Less than 1/mo. 0 1/mo. 02-3 mo. 0 1/week or more ®-
29. Which cold breakfast cereal do you Specify brnnd 11nd type (e.g. Kellogg s Nutrigrain Nuggots)
®0®®@®®0®®® -
usually eat?
0 Don't eat cold breakfast cereal
--
@(i)0@@@@0®®
® (!) ® ®@ ® ® 0@ ®
--
30. How many teaspoons of sugar do you add to your @0 0 0 0 0 0 0 0 0 @-
beverages or food tsp. ®0®00®®0®® -
31. When you have beef or lamb as a main dish, how well done is the meat cooked? -
0 Rare 0 Mecf1um rare 0 Medium 0 Med1um well 0 Well 0 Don t know/not eaten @) -
32. How often do you eat meat that w as charred during c ooking? (e.g. during or broiling) @> -
0 Never 0 LPss than ·1 /mo. 0 1/mo. 0 2 3/mo. 0 1/week 0 2-t jwt;ek -
33. How much of the visible fat on your beef, pork or lamb do you remove before eating? @) -
0 Remove all visible fat 0 Remove most 0 Remove small part of fat 0 Remove none 0 Don't eat meat -
34. What kind of fat do you usually use for frying and sauteing at home? (Exclude " Pam" -type spray) @-
0 Real butter 0 Marganne 0 Vegetable 011 0 Vegetable shorten1n 0 Lard
35. What l<ind of fat do you usually use for baking at home?
-
0 Real butter 0 Margarine 0 Vegetable otl 0 Vegetable shortening 0 Lard
36. How often do you eat food that is fried at home? (Exclude " Pam" -type spray)
0 Less than once week 0 1 3 t1mes pet wee" 0 4-6 tunes per week 0 Daily
1
37. How often do you eat fried food away from home? {e.g. f rench fries, tried chicl<en, fried fish)
-
0 Less than once a week 0 1-3 times per week 0 4-6 times per week 0Daily
38. What type of cooking oil do you usually use at Specify brand and typo
®CV®®0®®0®®
39.
home {e.g. Mazola Corn Oil)?
How does your current diet compare to your usual diet over the past five years?
®0®®0®®0®® -
0 Almost the same 0 Slightly changed 0 Moderately changed 0 Greatly changed
HANK YOU! Please return the q uestionnaire ih the
enclosed postage-paid envelope to:
Walter Willett, M.D. 677 Huntington Avenue
Nurses' Health Study II Boston, MA 02115
--
• Copyright t: 1991. President and Fellows of Harvard College. All Rights Reserved Worldwide.
-
•
Ill
Ill
II
II
-
.....
HARVARD HARVARD
MHI>ICAL SCHOOL of
SCHOOL PUBLIC HEALTH
- • Harvard School of Public Health • 6n Huntington Avenue • Boston, Massachusetts 02115 • (617) 432-2279 •
-
...
...
--
Decu Colleague:
.... On behalf of our research group, I again want to express my gratitude for your
participation in the Nurses' Health Study IT. The accuracy and completeness of
:r
..... the information you provide is truly impressive, and we are confident that this
study will provide answers to many critical questions about lifestyle factors, diet,
and oral contraceptive use. Analyses of these factors in relation to breast cancer
and several other diagnoses will begin soon. We have already begun to analyze
....
information on several common conditions and will report findings to you in our
next newsletter.
- The enclosed questionnaire continues our every-other-year follow-up . You will

note that we ask about your current status for many of the same questions that
I we posed earlier. We also ask about new medical diagnoses and conditions.
...
-.... We hope that you give this questionnaire the same attention and care that you
did in completing the earlier forms. The validity of this major research
.... Research Group
....
undertaking depends directly on complete and accurate follow-up information
Walter C. Willett, M .D.
.... Pnnc1pa Invest gator for all study members. We know that some participants are no longer in active
Stefanie Bechtel, B.A. nursing. However, your continued participation is critical regardless of current
Kim Boulger, B.A . employment status. As always, the information you provide is strictly
.. Lisa Chasao-Taber, M.P.H.
confidential and will be used only for medical statistical purposes .
Gary Chase, B.S.
Joyce Clifford, R.N., M .S.N .
Thank you again for your invaluable participation in this study. We will be
Graham Colditz, M .D.
•• Karen Corsano, M .A .
sending you the next edition of our newsletter in June of 1994 to update you on
II Gary Curhan, M .D. the progress of the investigation.
(\ I Marlene Goldman, Sc.D.
---
"" I Francine Grodstein, Sc.D. Sincerely,
---
" Sue Hankinson, R.N .. Sc.D.
David Hunter, M .D.
....
lchiro Kawachi, M .D.
-
.....
.....
Lisa litinf R.D.
JoAnn Manson, M .D •
Walter Willett, M.D .
Rachel Meyer. B.A. Professor of Epide1niology and N utritim\
Jennifer Nelson, B.A.
•
,...- Janet Rich-Edwards, M.P.H.
Helaine Rockette, R.D., M.S.
Laura Sampson, R.D.• M.S.
P.S. Your updated questionnaire information is needed to maintain the validity of
this study. Your reply within the next two weeks would be greatly appreciated.
Caren Solomon, M.D.
Franl< Speizer, M .D.
Donna Spiegelman, Sc.D.
Meir Stampfer, M .D.
II Lisa Troy, B.A .
....
.... II Diana Walsh, Ph.D .
Lori Ward
-....
......
Ill
Ill
Anne Wolf, M .S.
HARVARD UNIVERSITY NURSES' HEALTH STUDY II
INSTRUCTIONS
Please use an ordinary No. 2 pencil to answer all questions. Fill in the appropriate
response circles completely, or write the requested information in the boxes provided.
Note that some questions ask for information since June 1991, some ask for current status,
and some ask about events over longer periods. The form is designed to be read by
optical-scanning equipment, so it is important that you make NO STRAY MARKS and
keep any write-in responses within the spaces provided. Should you need to change a
response, erase the incorrect mark completely. If you have comments, please write them
on a separate piece of paper.
1. Current
EXAMPLE 1: Write in your weight in Weight
the boxes ... 1111':1111 NOTE: It is important that

... and fill in the circle gou write in gour
corresponding to the e CD weight in addition to
2 2
figure at the head of completing the
3 ® •
each column. 4 • 4
corresponding
Please fill in the circle 5 0 5 circles. This allows
®
completely, do not mark
this way:
6
0
6
7
us to confirm that
the correct circles
-
8
./ X • 9 have been darkened.
EXAMPLE 2: 1
Mark ' Yes" circle and Year of Diagnosis circle for each illness you
have had diagnosed.
11 . Since June 1991, have you had any of YEAR OF

these physician-diagnosed illnesses? DIAGNOSIS
LEAVE BlANK FOR
MARK HERE FOR "YES"
Stroke (CVA} or TIA

M elanoma
Basal cell skin cancer
Thank you for completing the 1993 Nurses' Health Study II

Questionnaire.
Please tear off the cover letter (to preserve confidentiality) and
return the questionnaire in the enclosed postage paid envelope.
If you need to make any changes or corrections to your

name/address you may do so on the cover letter and enclose it
with your completed questionnaire.
S' HEALTH STUDY II
PLEASE USE
1
2. We would like to update your pregnancy history from the time of the first
HA VARD UNIVERSITY
2 1 -
-
PENCIL!
CURRENT questionnaire in 1989 to the present. 2 -
WEIGHT a) Since September 1, 1989 have you been pregnant? a 4 -

POUNDS No -go to question 3 Yes a -
b) Are you currently pregnant? b p -
® No Yes- continue with part c. but do not fill in a bubble for current pregnancy
--
- c -
CD c) For each pregnancy ending after 1, 1989. fill in a response bubble for the year during which
'2' each pregnancy ended.
--
3
4 Calendar Year
Pregnancies lasting
6 months or more
Birth'.
Pregnancies lasting
less than 6 months
(Include miscarriages/ induced abortions)
--
--
'" ···- .- . ' . - -. .
® 5 9/1/89-12/31/89 1
6 1990 2
7 1991 3
--:-----
QU
1992
1993
----- 4
5
--
1994 6 7 -
3 Do you CURRENTLY use any of these forms of contraception? (Mark all that apply.)
-----------------------------1 8 9 -
r None Oral contraceptive Norplant Diaphragm/Cervical cap Tubal ligation Foam/Jelly/Sponge 10 X -
(__, Condom Intrauterine device Rhythm/NFP Vasectomy Depo Provera Other
-
3 -
SINCE JUNE 1991, have you used oral contraceptives (OC's)? 0 1 2 3 4 5 ol -
5 What is the current usual length of your menstrual cycle (interval from first day of period to first day of next period)?
r < 21 days 21-25 26-31 32 39 40-50 51-i days at too trregular to est1mate -
6. What is the current usual pattern of your menstrual cycles (when not pregnant or lactating)? --------------------------r-
6 -
'-I Extremely regular (no more than 1 2 days before or after expected) Vr:;ry raguln (within 3-4 clc;ys) -
r Regular (within 5 7 days) Usually irrc.gular Always irregular No periods -
7. Have your menstrual periods ceased
PERMANENTLY?
( No: Premenopausal a) Age +
. . . . .b) For what reason did your periods cease?
SURGERY: If due to surgery, were your ovaries
8-
1
b -
-
(
L
Yes: No menstrual periods _ _ __
Yes: Had menopause but now have
natural
periods
0
1
0
1
removed?
Yes, both Only uterus removed
-
n •
(
periods induced by hormones
Not sure
ceased?
2
3
2
3
One only
RADIATION or CHEMOTHERAPY
--
8. SINCE JUNE 1991, have you used female 5
4
6
NATURAL: If natural (non-surgical) menopause,
have you had subsequent surgery lo remove
--
replacement hormones (other t han oral contraceptives)?
_ No
6 6
7 No
ovaries or uterus?
One ovary removed
--
--
a) How many months have you
•
used them since JUNE 19917
( Yes. - - 1·4mo. -6-9 10-14 8 Uterus removed Both ovaries removed
cu rrently
-·
l Yes,
15·19 20+ months 9
--
discontinued b) Mark the types of hormones --
you have used the longest during this period. 0 1 2 3 4 5 6 7 8 9 8 -
Estrogen: Oral Premarin Estrace Ogen Patch Estrogen 012346o789a •
Vaginal Estrogen Other Estrogen 0123456789b .

- - -
Progesterone/Progestin {e.g., Provera): Ornl Vaginal Other (specify below)
- -- •
•
- - < -
Otltertype of honnones used, please specify: c:::==-====::::::>

- ·- - -- - - •
c) If you used oral conjugated estrogen (e.g., Premarin) what dose did you usually take? c •
.30 mg/day or less (Green) .625 mg/day (Brown) .9 mg/day {While) 1.25 rng/day !Yellow) •
......___
More lhan 1.25 mg/dc.. y
- - Dose unknown Did not take oral conjugated estrogen •
d) tf you used oral Medroxy Progesterone (e.g., Provera, Cycrin), what dose did you usually take? d .
5 mg 5·9 mg 10 mg More than 10 mg

e) What was your pattern of hormone use (Days per Month)?
Dose unknown Not used
-
e •
-•
Oral or Patch Estrogen: Days per Month Not used <1 day/rna 1·8 days 9· 18 19-26 27+ days/mo 1 •
Progesterone: Days Month Not used <1 day/mo
-----------
1-8 days 9-18 19-26 27+ days/mo 2 •
-
9. Have you had a tubal ligation? 9 •
() No 8 •
C Yes At what age? <25 25-29 30·34 35-39 40·44 45+

••
-- 1 1 1 1 1 1 1 2 3
PAGF. 2
4 12. Have you EVER had any of

NHS II 93
12
- 2 2 2 2 2 2 2 5 6 7 8 these physician-diagnosed
- 4 4 4 4 4 4 4 9 10 11 12 illnesses?
- 8 8 8 8 8 8 8 93 94 95
LEAVE BLANK FOR •NO". MARK - -----.1
p p p p p P A B C 0 HERE FOR NYESW •
- p
- ._ THIS IS YOUR ID Ectopic pregnancy 'f
- 1 0. Is this your correct date of birth7 ._. High blood pressure 2
-- (pregnancy-related)
-- 11. r Yes No . . .
If no, please
write correct date. Month I I Dey Voer
Toxemia/Pre-eclampsia
of pregnancy
3
Since June 1991, have you had any YEAR OF. SLE (systemic lupus) 4
-- of these physician-diagnosed
illnesses?
DIAGNOSIS
Before 91 After
Active TB (X-ray confirmed)
Graves' Disease
5
6
-- LEAVE. BLANK r-oR , MARK HERE FOR "YES"
Myocardial infarction (heart attack)

'+
Juqe 1 June 1
1991 May 93 1993
1
Other Hyperthyroidism
Hypothyroidism
7
8
-- r- Angina pectoris
. . . . Confirmed by angiography? N No v Yes
2
a
Thyro1d nodule (benign)
Mitral valve prolapse
9
10
-- Stroke !CVAl
Melanoma
01 TIA
•
3
4
Confirmed by echocardiogram7
Herniated lumbar disk y
s
11
-- Basal cell skin cancer s

Confirmed by CT or MRI?
-- Squamous cell skin cancer

FibrocystiC/other benign breast dis.
6
7
Other chronic back problem
Endometriosis - 1st Ox 13
-- Confirmed by breast biopsy?

Confirmed by aspiration?
N
N
No
No
v Yes
v Yes b
Confirmed by laparoscopy7
No
N v Yes 0
-- Breast cancer
Other cancer:
8
9
Uterine fibroid(s) - 1st Dx
Confirmed by pelvic exam?
y
N No v Yes
14
11
-- Specify site of
other +
Confirmed by ultrasound/hysterectomy? N No
13. Current Medication (mark if used regularly)
v
·
Yes b
13
--- - High blood pressure

(excluding during pregnancy)
10 No regular med1cat1on
Acetaminophen, 2+ times/week (e.g., Tylenol)
-
Diabetes: Gestational v 11 Aspirin, (e.g., Anacin, Buffenn, Atka-Seltzer, etc.)
-- Diabetes: Not pregnancy-related

Elevated cholesterol
--t-
12
13
DayS/week: <1 day 1•2 3- 4
Other anti-inflammatory analgesics, 2+ times/week (e.g .,
5 - 7 days D
-- r- Deep vein thrombosis/Pul. embolism

Rheumatoid arthritis, doctor diagnosed v
14
15
Ibuprofen, lndocin, Naprosyn, Advil)
Thiazide diuretic (e.g .. Hygroton, Dyazide, HCTZ, Diuril)
--
. . . Rheumatoid factor
Other ar1hntis
Negative/ Unknown Positive 8
16
Any anti-hypertensive medication
Thyroid hormone replacement (e.g, Synthroid, Levothroid)
-- Colon or rectal polyp (benign)

Gastric o r duodenal ulcer
17 Minor Tranquilizers (e.g., Valium, Xanax, Ativan, Librium)
-- Cholecystectomy 19 Major Tranquilizers (e.g., Stelazine, Thorazine, Haldol,

Prolixin, MeiJaril, Trilafon)
-
--- •Gall stones
a) Did you have symptoms? N No v Yes 8
14. Have you ever taken any of the following medications?
a) Tetracycline:
t4
-- How dlaqnosecl?
Polycysuc ovaries
X-ray or ultrasound Other b
21
0 Yes
No
+ For how long?
Age@ 1st use:
<1 yr.
<15
1-2
15-19
3-4
20-29
5+yrs
30+
1
-- Prcmenstruill syndrome (PMS)

Vaginal yeast infection
22
23
b) Oral Acutaine:
Yes For how long? <1 yr. 1-2 3-4 5+yrs
b
--
Kidney stones
Pneumonia, X-ray confirmed
24 No Age@1stuse: <15 15-19 20-29 30+
c) Tricyclic antidepressants (e.g., Elavil, Norpramin, Tofranil,
2
-- --
Multiple sclerosis
--
Hydatidiform mole (of pregnancy)
26
27
Pamelor, Sinequan, Vivactil, Surmontil):
Yes for how long? <1 yr. 1-2 3-4 5+yrs
c
1
-- Asthma, Physician Ox
Ulcerative colitis/Crohn's
28
29
No Age®1stuse: <15
d) Prozac (Fiuoxetine) or Zoloft {Sertraline):
15-19 20-29 30+ 2
d
--Migraine headaches
-
Other major illness or surgery since
30
31
Yes +
No
for how long? <1 yr. 1-2 3-4 5+yrs 1
I -
-June, 1991 Please •pacify other m ajor illness or surgery: 15. a) Your TB skin test since 19897
Pas Neg Not done BCG prior to 1989
t5
11
I - b) If ever positive, conversion date: b
I - 0123156789 f 2 3 4 £ 6 7 8 9
-c) Before 1989 1989+ Never positive
J -- z () 0 1
,
2
2
3 4 5
3 4 5
8
6
7
7
8
8
9
9
0
0
1
1
z
2
3
3
4
4
5
5
6 7
6 7
8
8
9
9
If ever positive, were you treated with INH7
Yes No Never positive
16. Since June 1991 , have you tried to become pregnant for more than one year without success?
3
-
8-
16 -
Yes ..... What was the cause? {Mark all that apply.)
Tubal blockage
Spouse/Partner
Ovulatory disorder
Not investigated
Endometriosis
Not found
Cervical mucous factors
Other
--
17 Have you ever taken Clomid (Clomiphene) or Pergonal to induce ovulation1
-
a-
17 -
Yes a) In how m any months was Clomid used: 0 months 2-3 4-5 6-11 12+ m onths
___
o _________b_l_l_n_h_o_w___ ___
o_nt_h_s_w
___ ___ ___o_n_th_s___ ___ ___ ___ 12+ months b -
18. Have you ever had a miscarriage or induced abortion before the sixth month of pregnancy?
--
a-
18
Miscarriage: No Yes· at what age(s) <18 18-20 ._, 21 -23 24-26 27-29 30-34 35+
Induced Abortion: No Yes: at what age(s) <18 18-20 21-23 24-26 27-29 30-34 35+ b -
19. Since June 1991, how many months have you worked ROTATING night shifts (at least 3 nights/month in addition to 19 -
-
other days and evenings in that month)?
None 1-4 months 5-9 10-14 15-19 20+ months
--
20. Which best describes your current employment status? 20 -
Inpatient or ER Nurse
Nursing Administration
Outpatient/Community
Other Nursing
OR Nurse
Non-nursing employment
Nursing Education
Fullt1me Homemaker
Student
Disabled
--
21 . How many times per week do you engage in physical activity long enough to perspire heavily (including swimming)?
Less than once/week Once/week 2-3 times/week 4-6 times/w eek 7 or more times/week 21 -
-
22. In how many months did you practice breast self-examination in the past year?
None One month 2-3 46 7-11 12 months 22 -
-
23. Since June 1991, have you had: No Yes, for screening Yes, for symptoms 23. -
Mammogram
Breast exam by clinician
--
Colonoscopy /Sigmoidoscopy
Pap smear
--
24. How many months in total (all births combined) did you breast feed? 24 -
Didnotbreastfeed
,. No ch ildren
_. <1 month
18-23 mo.
1-3mo.
24-35 mo.
4-6mo.
r-. 36-47 mo.
, 7-11 mo.
48+ mo.
.._ 12-17mo.
Cann ot remember
--
25 Between the ages of 18 and 30 (excluding illness and pregnancy-related changes): 25
a) What was your: Minimum weight lbs. Maximum weight lbs. a-
b) Between the ages of 18 and 30, how many times did you lose each of the following amounts of w eight on purpose?
5-9 pounds: 0 t imes 1-2 times 3·4 times , 5-6 times 7+ times MX
--
10-19 pounds:
20-49 pounds:
0 0 times
0 times
0 1-2 times
1-2 times
3-4 times
3-4 times
5-6 times
5-6 times
7+ times
7+ times
b
--
26.
50+ pounds: 0 times 1-2 t imes 3-4 times 5-6 times
Within the last 4 years (excluding illness and pregnancy-related changes):
7+ times
26
--
a) What was you::r.::. . .______...:.M:.:.•:.:.· - -====:...:':.:b=s:.__
. _____________________ a-
MN-
b) Within the last 4 years, how many times did you lose each of the following amounts of w eight on purpose?
5-9 pounds: 0 times 1-2 times 3-4 times 5-6 times 7+ times
-
MX -
10-19 pounds: 0 times 1-2 times 3-4 t imes 5-6 t imes 7+ times b -
20-49 pounds:
50+ pounds: 0 times
0 times
1-2 ti mes
1 2 t tmes
( 3-4 times 0 5-6 times
3-4 times
( 7 + t imes
5-6 times 7+ times .
--
-------------------
c) What primary method(s) did you use for your most recent w eight loss of 10 or more pounds? (Mark all that apply)
Did not lose 10 or more pounds Diet pills Increased exercise c--
Low calorie diet
Low fat diet
Commercial weight loss program
Gastric surgery/intestinal bypass
Decreased alcohol intake
Resumed/increased smoking
--
Skipped meals/fasted Other
Weight loss was unintentional (e.g., illness, unusual stress, depression)
--
27 Do you currently smoke cigarettes?
a-
'1:1 -
No Yes How many per day? 1-4 5-14 15-24 25-34 35-44 45 or more
28. What was the cup size of your bra when you were 20 years old? (Estimate if you did not wear a bra.) 28 -
A or smaller
29. How many biological sisters do you have?
B C D or larger
-
29 -
0 1 2 3 4 5 or more I -
30. Did your mother or any of your sisters have ovarian cancer? 30 -
______YI_e_s__ __o_th_e_r_______S_i_st_e_r____ ____________________________________________ ..
31. When your mother was pregnant with you, did she take DES (Diethylstilbestrol) or other hormones? Please Continue a1 l. .
L.. Don't know ---.. No Yes • -.. DES , ___, Other hormones on Psge 4 a 1 -
I
-- 32.
In which state were you born?
II -
I -
In which state did you live at age 157
In which state did you live at age 307 : : : : : : .t
I - 33. During summers how many times per week were you outdoors in a swimsuit:
- a) as a teenager? < 1/week 1,'week 2/week Several/week Daily
- b ) in the past summer? <1/week 1/week 2 week Several/week Daily
- 34 When you were outside at the pool or beach, what percent of the time did you wear sunscreen:
- a) as a teenager? Not m sun 0% 25to 50% 75% 100' 'c
- b) in the past summer? Not in sun 0% 25% 50% 75% 100%
- 35 Is your biological mother still living?
- Yes No a) At what age did she die? <50 50-59 60-69 70-79 80+
- b) Was this due to: Heart Disease
-
Cancer
- Trauma/Accident/Suicide Other b (Southern!
- 36. Is your biological father still living? 36 co
•
- Yes No a) At what age did he die? <50 CT

-
-- 37 What is your current marital status?
b) Was this due to: Heart Disease Cancer Tra Ll malAccidenl/Su icide Other b
37
DE
DC
- M arried 01vorced/Separated Widowed Never Married FL
•
- 38. What is your current living arrangement? 38 GA
- Alone Wi th husband/partner With other family Other HI
- -------------------
39. Do you currently take a multi-vitamin? (Please report additional individual vitamins in question 40.)
- No 39 10
- 3-5 6-9 10 or more

0
o
1
1
2
2
3
3
4
4
5
5
6
6
7
7
s
8
9
9
IL
IN
I::
2 3 4 5 6 7 8 9 b lA
a 3 s 6 7 e 9 KS
40 KY
LA
) Beta- r ,' No 0 Less than 8,000 IU per day C 8,000 to 12,000 IU BC ME
-- carotene?
b) Vitamin A? ( 1 No (
--
Yes - - •
- Q
Q
13,000 to 22,000 IU
Less than 8,000 IU per day
n 23,000 IU or more
J 8,000 to 12,000 IU
0 Amount unknown
A
MD
MA
--- (excluding
carotene)
c) Vitamin C7 ( No ( Yes, seasonal only
0
0
13,000 to 22,000 IU
Less than 400 mg.
v 23,000 IU
0
01
400 to 700 mg.

more Amount unknown
c
Ml
MN
-- d) Vitamin E7
•
0 No
( Yes, most months
lJ Yes
0 750 to 1,250 mg.
Less than 100 IU
0 1300 mg. or more
0 100 to 250 IU
0 Amount unknown
e
MS
MO
-- +
Q No 0 Yes
-- Q
Q
300 to 500 IU
Less than 400 mg per day
0 600 IU or more
0
_______
400 to 800 mg.
_____
Amount unknown
__..;;;;::;...._ MT
e) Calcium? CL NE
-- (elemental)
f) Folic acid?
+
0 No 0 Yes
0
0
900-1,200 mg.
Less than 100 mg.
0 1,300 mg. or more
0 -
100 to 300 mg.
Amount unknown
FA
NV
NH
-- -- ------------ -- - ___ f'\ 301 500 mg

..;..:
41. Please indicate the name of someone at a DIEFEBENT PERMANENT ADDRESS to whom we might write,
0 501 mg. or more n Amount
41
NJ
NM
-- in the event we are unable to contact you: 42 NY
-- Name: __________________________________________________________________________ w NC
NO
-- Address: __________________________________________
WAIST HIP
---'---1
AD OH
01<
-- 42 Question 42, which should only be answered if
a tape measure is convenient, asks about body
measurements. This information will be more
Inches Fraction Inches Fraction OR
PA
-- accurate if you follow these suggestions:

..,. Make measurements while standing
PR
Rl
-- ..,. Avoid measuring over bulky clothing

..,. Try to record answers to the nearest 1/4 inch
- 214
11/4
2/4
sc
so
-- {do not estimate)
If a tape measure is not available,

Hip: Waist
Measure
3 3 @ 3
4
3
4
'3/4 TN
TX
--
Measure the
please leave blank. largest at navel
5 5 UT
circumference
around hips 6 6 VA c
I -
- (including
buttocks) VT
WA
I - Mark Refle x® b y NCS EM-1511 72:321 A2102 Printed in U.S.A. wv
I - WI
I - PLEASE GO TO PAGE 5 AND BEGIN BY WRITING YOUR ID NUMBER FROM PAGE 2 WY
- Non US
-- HEAI..!l'H STUDY II
Please copy your 10 from page 2 to here. o 1 2 3 4
PAGES
s o 7 s 9 o 1 2 3 4 5 G 7
HARVARD UNIVERSITY
a 9 o 2 3 4 s 6 8 9
--
If
ID: - o 1 2 3 4 s G 1 a 9 o 1 2 3 4 s s 7 a 9
For Othc Use Only
A II
o 1 2 a 4 s s 7 s s o 2 3 4 s 6 7 s 9 I
- Many participants have pointed out that stress, personal and family relationships, and other aspects of quality I
- of life are important factors relating to health. \Ve have added the following questions to learn more about
- these aTeas. (As always, all of your responses will remain strictly coniidential.)
- 43. These questions are about how you feel and how things have been with you during the past 4 weeks.
- For each question, please give the one answer that comes closest to the way you have been feeling.
- How much of the time during the past 4 weeks... A tittle
of the
None
of the :rm: •
- (Mark one response on each line.) time time time time time time
- Did you feel full of pep?
- Have you been a very nervous person_?_ ______________-7-:.__- l------"=----1--.::..__- __
- Have you felt so down in the dumps nothing could cheer you up?
- Have you felt calm and peaceful?
- Did you have a lot of energy?
- Have you felt downhearted and blue?
- Did you feel worn out?
- Have you been a happy person?
-
-
Did you feel tired?
41. During tho past 4 weeks, how much of the time has your ph-ysical health or have emotional problems interfered --.---,.- -
- with your social activities (like visiting with friends, relatives, etc.)?
- All of the time Most of the ttme Some of the ttme A little of the timP None of the time
- 45. Please choose the answer that best describes how true or false each of the
-- following statements is for you. (Mark one response on each line.)
Over the past 4 weeks, I have felt about the same as I have felt during the past year
Definitely Mostly
True True
Not
Sure
Mostly Definitely
False False
-- I seem to get sick a little easier than other people

1am as healthy as anybody I know
-------------------------
-- 1 expect my health to get worse
My health is excellent
- 46
--- The following items are about activities you might do during a typical day. Does your health
now limit you in these activities? If so, how much? (Mark one response on each line.) Yes. limited
A lot
Yes. limited No, Not limited
A little At All
-- Vigorous activities, such as running, lifting heavy objects, participating In strenuous sports
Moderate activities, such as moving a table, pushing a vacuum, bowling, or golfing
---1
-- Lifting or carrying groceries

Climbing several flights of stairs
•
-- Climbing one flight of stairs

Bending, kneeling, or stooping
-- Walking more than a mile
Walking several blocks
-- Walking one block
---------
-- 47 Bathing or yourself
During the past 4 weeks, have you had any of the following problems with your work or other regular daily activities
I
-- as a result of any emotional problems (such as feeling depressed or anxious)? (Mark one response on each line.)
' Cut down the amount of tiiTl..tl you £>pent on work or other acttvltles Yt:is No
-
- b) Accomglished less than you would like
Didn't do work o other activities as as usual
Yes
Yes
No
No
- 48 During the past 4 weeks, to what extent has your physical health or have emotional problems interfered with your
-
- normal social activities with family, friends, neighbors, or groups?
- Not at all Slightly Moderately Quite a bit Extremely
- 49. How much bodily pain have you had during the past 4 weeks? 9
- None Very mtld Mild Moderate Severe Very severt
- 50 During the past 4 weeks, how much did bodily pain interfere with your normal work (including both work outside the 0
- home and housework)?

- Not at all A little bit Moderately Quite a bit Extremely
- 51. During the past 4 weeks, have you had any of the following problems with your work or other regular daily activities s
- as a result of your physical health? (Mark one response on each line.) e
-- , Cut down the amount o" time you spent on work or other ac 1vities
b ) Accomplished less than you would like
Yes
Yes
No
No PLEASE b
-- c) Were limited in the kind of work or other activities

d) Had difficulty performing the work or other activtties (for example, it took extra effort)
Yes
Yes
'\lo
No
CONTINUE ON
PAGES
PAGii: 6
- I
• In general, would you say your health is: Very Good Good Fair
53. Do you have an unreasonable fear of being in eo closed spaces such as stores, elevators, etc.7
Poor
"- I
.53 -
I
Ott ,. o"" Mf' r
5 . Do you find yourself worrying about getting some incurable illness? Often Sometimes Never
-- II
55. Are you scared of heights?
5 . Do you feel panicky in crowds?
·ery
A' "•metimes
Not at all
Never 56 -
-
57. Do you worry unduly when relatives are late coming home?
58. Do you feel more relaxed indoors? Sometimes
Yes No
Not particularly 58 -
-
59. Do you dislike going out alone? YP-s No 59 -
------
60. Do you feel uneasy traveling on buses or trains even when they are not crowded?
Vt:.ry ' li tl.. ,[ dt c I
--
- - - - -
61. If you are retired or stopped working due to illness/injury, at what age were you last in paid employment? 61 -
Still workmg • Age 2'5 • 29 JO 34 35 39 Age 40 or older
--
62. If you have been employed within the past 2 years, the following questions relate to your current or most recent job:
Not employed •n last 2 'ea• "
--
Please choose the answer which best describes the degree to w hich
you agree or disagree with each of the following statements. Drsagree Disagree Agree
Strongly
Agree
--
My JOI"\ rcquir < [lw I L ... n .. .aw --
My job involves a lot of rep etitive wo rk
My job requires me to be creat1ve
--
My job allows me to make a lot of decisions on my own
My 'ob requires a high level of skill
--
On my job, I have very little freedom to decide how I do my work
I get to do a variety of different things on my job
--
I have a lot of say abollt what happens on my job
I have an opportunity to develop my own special abilities
--
My job requires workmg very fast
My job requires working very hard
--
My job requires lots of physical effort
I am not asked to do an excessive amount of work
-- -
I have enough time to get the job done
My job security is good
--
I am free from conflicting demands that others make
People I work with are competent in doing their jobs
--
People I work with take a personal mterest in me
People I work with are friendly
--
People I work with are helpful in getting the job done --
D1sagree Disagree Agree
Strongly
Agree
Not
Applicable
--
My supervisor is concerned about th e w elfare of those under her
My supervisor pays attentton to what I am saymg _
--
My supervisor is helpful in getting the job done
My supervisor is successful in getting people to work together
--
b How long have you worked in the job you described above?
< 6 months 6 months - 11 month s 1 - 2 years 3- 4 years 5 - 9 years 10 or more years
--
c) How many hours per week do you work, on average, in your job?
15h Jr<; 15 20 21 40 41 · 60 61 80
--
63. ------------
How many hours per week do you spend in housework (including cooking, cleaning, shopping for food, doing laundry S3 -
and dishes, doing repairs, paying bills, making arrangements and caring for children)?
0 - 19 hours 20 39 110 .:>9 60 7c 80 100 hours
--
64. Thinking of all the things that are done in your household, what percentage do you personally do?
0 2'1 p • ct. t ?6 1Q 10 0 61 74 '5 99 100 percent -
65. Is there any one special person you know that you feel very close to; someone you feel you can share confidences and
feelings with 7 -
Ye"' . _ . How often do you see or talk with this person? X v -
No n-.ily Weekly Morthly Sever I 0nce/year nr lASS L • -
THANK You! Please return the questionnaire in the

enclosed postage-paid envelope to:
Walter Willett, M.D.
Nurses' Health Study II
677 Huntington Avenue
Boston, MA 02115
--
• Copyright® 1993. President and F llow: of H.. 'Varci College. Ill Rights Reserved We.. 1dwld .... Reflex\:11 by NCS EM·15117.... J21 A2403 Printed In U.S.A.

Tesis RomanosNanclares21

Uploaded by

Copyright:

Available Formats

Tesis RomanosNanclares21

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tesis RomanosNanclares21

Uploaded by

Copyright:

Available Formats

School of Medicine

DIETARY FACTORS AND BREAST CANCER

ANDREA ROMANOS NANCLARES

Pamplona, a 29 de marzo de 2021

Dña. ESTEFANÍA A. TOLEDO ATUCHA, PROFESORA TITULAR DEL

D. ALFREDO GEA SÁNCHEZ, PROFESOR CONTRATADO DOCTOR DEL

Que Dña. ANDREA ROMANOS NANCLARES ha realizado el estudio “dietary

Dra. Estefanía Toledo Atucha Dr. Alfredo Gea Sánchez

I extend my sincerest appreciation to my supervisor, Dr. Estefania Toledo, for her

I am grateful to my supervisor, Dr. Alfredo Gea, whose patience, serenity and

The Department of Preventive Medicine has been my workplace since

Despite the extraordinary circumstances as a visiting grad student engulfed in

I would like to express my sincere gratitude to the Spanish Association Against

De todos ellos, los factores relacionados con la dieta se consideran de gran

En este contexto, el objetivo global de la presente tesis doctoral fue valorar

El proyecto SUN es una cohorte prospectiva, dinámica y multipropósito iniciada

En el Proyecto SUN, una adhesión moderada a un patrón provegetariano, donde

Existen alimentos que son ampliamente consumidos dentro de patrones

En línea con lo anterior, se debería de poner más énfasis en la calidad de los

Los polifenoles son responsables de algunas de las propiedades de los

Por último, numerosos estudios experimentales y epidemiológicos han puesto de

Figure 1. Breast anatomy. (Taken from Harbeck et al.4).

Arterial and Venous Supply

Carcinogenesis is a multifaceted process that involves numerous etiologies

At the molecular level, there is a parallelism between normal cell development

Figure 3. Microenvironmental alterations during tumor progression. Schematic view of

Typically, cancer is influenced by genetic and epigenetic alterations which allow

The most predominant signaling pathways implicated in the normal development of

1. ER signaling and ER-Positive breast cancer

2. HER2 signaling and HER2-Positive breast cancer

Human epidermal growth factor receptor (HER)s 1 to 4 comprise a family of tyrosine

3. Canonical Wnt/β-catenin signaling in breast cancer

4. Other signaling pathways

• Notch signaling: Notch signaling is implicated in the pathogenesis of breast cancer

• Breast tumor kinase (BRK): BRK is a non-receptor tyrosine kinase overexpressed

• PI3K/AKT/mTOR pathway: oncogenic mutation of PI3K is generally known in

3. Breast Cancer Subtypes

Types of breast cancer based on pathology, invasiveness and prevalence

• Non-invasive (or in situ) breast cancer

• Invasive or infiltrating breast cancer

1. Invasive Ductal Carcinoma (IDC)

Figure 5. Invasive Ductal Carcinoma subtypes47–49.

• Metastatic breast cancer

Inflammatory breast cancer (IBC)

Male breast carcinomas

Paget disease of the breast

Angiosarcoma of the breast

Box 1. Less common types of breast cancer47–49.

Molecular or intrinsic subtypes of breast cancer

Classical immunohistochemistry (IHC) markers such as ER, PR and HER2, along

Intrinsic subtype Molecular Signatures

Luminal A ER+, PR±, HER2-, Low Ki67

Figure 7. Summary of breast cancer subtypes57,58.

Clinical staging and survival rates of breast cancer

Table 1. Anatomic stage groups of breast cancer.

Breast cancer epidemiology pattern differences across countries are further

Over 355,000 women in the European Union were estimated to be diagnosed

A risk factor is defined as anything that affects an individual’s chance of getting a

5.1. Non-modifiable risk factors