Alcohol consumption and lifestyle factors in association

with pancreatic cancer in the EPIC study

Sabine Naudin

To cite this version:

Sabine Naudin. Alcohol consumption and lifestyle factors in association with pancreatic cancer in the
EPIC study. Health. Université de Lyon, 2019. English. �NNT : 2019LYSE1133�. �tel-02434882�

HAL Id: tel-02434882

https://tel.archives-ouvertes.fr/tel-02434882
Submitted on 10 Jan 2020

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci-
entific research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
N° d’ordre NNT : 2019LYSE1133

THESE de DOCTORAT DE L’UNIVERSITE DE LYON

opérée au sein de
l’Université Claude Bernard Lyon 1

Ecole Doctorale N°205

Ecole Doctorale Interdisciplinaire Science-Santé (EDISS)

Spécialité de doctorat : Epidémiologie, Santé Publique, Recherche sur les services de santé
Discipline : Epidémiologie

Soutenue publiquement le 20/09/2019, par :

Sabine NAUDIN

Le rôle de la consommation d’alcool et du mode de vie sur

le risque de cancer du pancréas dans l’étude EPIC

Devant le jury composé de :

KAMPMAN Ellen, PhD - Rapporteure

Professeure, Université de Wageningen, Division of Human Nutrition, Wageningen, Pays-Bas
WEIJENBERG Martje Petronella, PhD - Rapporteure
Professeure, Université de Maastricht, Department of Epidemiology, Maastricht, Pays-Bas
BOTTERI Edoardo, PhD - Rapporteur
Enseignant-Chercheur, Registre de Cancers de Norvège, Department of Bowel Cancer
Screening, Oslo, Norvège
FERVERS Béatrice – Examinatrice (Présidente du jury)
Professeure associée, Université Lyon 1,
Directrice, Départment « Cancer et Environnement », Unité INSERM UA8 « Radiations :
Défense, Santé, Environnement », Centre Léon Bérard, Lyon, France
LEITZMANN Michael - Examinateur
Professeur, Université de Regensburg, Department of Epidemiology and Preventive
Medicine, Regensburg, Allemagne
FERRARI Pietro - Directeur de thèse
Chercheur, Chef de Group, Centre International de Recherche sur le Cancer , Nutritional
Methodology and Biostatistics Group, Lyon, France
BRENNAN Paul - Co-directeur de thèse
Chercheur, Chef de Section, Centre International de Recherche sur le Cancer, Genetics
Section, Lyon, France
Alcohol consumption and lifestyle factors in association
with pancreatic cancer in the EPIC study
 Institute of thesis preparation

International Agency for Research on Cancer (IARC)

World Health Organization (WHO)

150, cours Albert Thomas

69372 Lyon Cedex 08, France

Director and Supervisor

Pietro FERRARI, PhD, Group Head

Nutritional Methodology and Biostatistics Group (NMB)

Nutrition and Metabolism Section (NME)

Co-director

Paul BRENNAN, PhD, Section Head

Genetics and Epidemiology Group (GEP)

Genetics Section (GEN)

PhD funding

Allocation doctorale de Recherche 2016

Région Auvergne-Rhône-Alpes
 ABSTRACT

Commonly diagnosed at late stage, pancreatic cancer (PC) is a highly fatal cancer with limited
opportunities for early detection and effective treatment. The identification of modifiable risk
factors may offer relevant scientific evidence for PC prevention. This doctoral research
program investigated PC etiology through a comprehensive examination of the role of alcohol
consumption and other lifestyle determinants in the occurrence of PC within the European
Prospective Investigation into Cancer and nutrition (EPIC) study, a multi-center cohort
involving more than 500,000 participants from 10 European countries.

Alcohol intake was evaluated with respect to the risk of PC. The role of different alcoholic
beverages and potential effect modification by smoking habits on PC risk were also
examined. Findings from this evaluation provided epidemiological evidence that large intakes
of alcohol were associated with an increased risk of PC.

The association between the healthy lifestyle index, a score combining information on
smoking history, alcohol intake, diet, obesity, and physical activity and the risk of PC was
examined. To quantify the impact of modifying several lifestyle factors, population
attributable fractions were estimated assuming counterfactual scenarios whereby study
participants hypothetically moved towards healthier behaviors. Adherence to healthy lifestyle
habits was strongly inversely related to PC.

This comprehensive evaluation provides informative insights on the etiology of PC and

supports the development and implementation of public health guidelines to promote
individuals’ adoption of healthy lifestyle habits for PC prevention.

Keywords: Pancreatic cancer, lifestyle factors, alcohol consumption, smoking, healthy

lifestyle index, EPIC, population attributable fractions.

1
2
 RESUME

Souvent diagnostiqué à un stade avancé, le cancer du pancréas (CP) est un cancer

particulièrement létal pour lequel il n’y a, à ce jour, que très peu de possibilités de traitement
et de diagnostic anticipé. L’identification de facteurs de risque modifiables pourrait fournir
des données épidémiologiques nécessaires au soutien de la mise en place de mesures
préventives. Cette thèse a pour but d’étudier l’étiologie du CP en évaluant le rôle de la
consommation d’alcool et du mode de vie dans l’étude prospective européenne sur le cancer
et la nutrition (EPIC), cohorte multicentrique de plus de 500,000 sujets provenant de 10 pays
européens.

La consommation d’alcool a d’abord été évaluée en regard du risque de CP. Les rôles des
différents types d’alcools ainsi que du tabagisme dans la relation entre l’alcool et le CP ont été
examinés. Cette évaluation a montré qu’une consommation d’alcool élevée était associée à
une augmentation du risque de CP.

Ensuite, la relation entre le Healthy Lifestyle Index, un indicateur combinant le passé

tabagique, la consommation d’alcool, l’alimentation, l’anthropométrie et l’activité physique,
et le risque de CP a été étudiée. Pour quantifier l’impact de l’amélioration de ces facteurs, des
fractions de CP attribuables ont été estimées en considérant des scénarios hypothétiques où les
participants adopteraient des modes de vie plus sains. L’adhérence à des habitudes saines était
fortement et inversement associée au risque de CP.

Ces travaux ont apporté des connaissances informatives sur l’étiologie du CP, et soutiennent
le développement de mesures de santé publique promouvant la prévention du CP par
l’adoption de modes de vie sains.

Mots clefs : cancer du pancréas, facteurs du mode de vie, consommation d’alcool, tabagisme,
healthy lifestyle index, EPIC, fractions de population attribuables.

3
4
 RESUME SUBTANTIEL

Souvent diagnostiqué à un stade avancé, le cancer du pancréas (CP) est un cancer

particulièrement létal pour lequel il n’y a, à ce jour, que très peu de possibilités de traitement
et de diagnostic anticipé. Les études épidémiologiques menées jusqu’à présent ont permis
d’établir que le tabagisme, la surcharge pondérale et le diabète de type II sont les principaux
facteurs de risque du CP. Des études récentes ont également suggéré qu’une consommation
considérable d’alcool pourrait être associée au CP. Les habitudes alimentaires et l’activité
physique, quant à eux, serraient inversement associés, mais les données épidémiologiques
actuelles ne convergent pas. L’identification de facteurs de risque modifiables pourrait fournir
des données épidémiologiques nécessaires au soutien de la mise en place de mesures
préventives. Cette thèse a pour but d’étudier l’étiologie du CP en évaluant le rôle de la
consommation d’alcool et du mode de vie dans l’étude prospective européenne sur le cancer
et la nutrition (EPIC), cohorte multicentrique de plus de 500,000 sujets provenant de 10 pays
européens.

La consommation d’alcool a d’abord été évaluée en regard du risque de CP. Bien que l’alcool
soit un carcinogène reconnu pour le foie, le colon-rectum, le sein et les voies aéro-digestives
supérieures, les données concernant le pancréas ont été jusque-là limitées par la faible
incidence du CP, notamment pour dissocier le rôle de l’alcool de celui du tabagisme. Dans
cette étude, les rôles de la consommation d’alcool au cours de la vie, des différents types
d’alcools ainsi que celui du tabagisme dans la relation entre l’alcool et le CP ont été examinés.
Dans la cohorte EPIC, 1,283 cas de CP (dont 57% de femmes) ont été rapportés au cours de
15 années de suivi. Les quantités d’alcool consommées au cours de la vie et pendant l’année
précédant le recrutement ont été estimées par un questionnaire de fréquences alimentaires et
un questionnaire portant sur le mode de vie. Des modèles de Cox à risques proportionnels
avec l’âge comme échelle de temps ont permis d’estimer des hazard ratios (HR) et leur
intervalle de confiance à 95% (IC95%). La consommation d’alcool était positivement
associée au risque de CP chez les hommes. Les associations mesurées étaient principalement
observées pour les niveaux d’alcool les plus élevés, avec des HRs comparant les grands
consommateurs (>60g/jour) à la catégorie de référence (0.1-4.9 g/jour) de 1.77 (IC95% : 1.06,
2.95) pour la consommation d’alcool au cours de la vie, et de 1.63 (1.16, 2.29) pour la
consommation d’alcool au recrutement. Les consommations de bière (>40 g/jour) et de

5
liqueurs/spiritueux (>10 g/jour) ont montré des HRs de 1.58 (1.07, 2.34) et de 1.41 (1.03,
1.94) respectivement, par rapport à la catégorie de référence (0.1-2.9 g/jour). Chez les
femmes, les HRs n’étaient pas statistiquement significatifs. La relation entre l’alcool et le
risque de CP n’était pas modifiée par le statut tabagique. Cette évaluation a donc montré
qu’une consommation d’alcool élevée était associée à une augmentation du risque de CP.

Bien que les habitudes liées au mode de vie aient tendance à être dépendantes les unes des
autres, les données épidémiologiques concernant l’association combinée de ces différents
facteurs et leur relation avec l’incidence du CP restent à ce jour limitées, particulièrement en
Europe. Dans la deuxième partie de ce travail de thèse, le Healthy Lifestyle Index (HLI), un
score combinant des informations sur le passé tabagique, la consommation d’alcool,
l’alimentation, l’anthropométrie et l’activité physique, a été étudié en relation avec le risque
de CP. Deux versions de ce score ont été analysées : le HLIIMC incluant l’indice de masse
corporelle (IMC) comme mesure de l’adiposité globale et le HLIRTH incluant le rapport taille-
hanche (RTH) comme mesure de l’adiposité centrale. Des modèles de Cox à risques
proportionnels avec l’âge comme échelle de temps ont permis d’estimer les HR et leur IC95%
pour une augmentation d’un écart type correspondant à 3 unités de HLI. Comme la relation
observée pouvait être le résultat de l’action d’un seul des facteurs du HLI, leur rôle dans la
relation entre le HLI et le CP a été évalué au cours d’analyses de sensibilité excluant à tour de
rôle chaque composante du score. Pour quantifier l’impact de l’amélioration de ces facteurs,
des fractions de population attribuables (FPA) ont été estimées en considérant des scénarios
hypothétiques où les participants adopteraient des modes de vie plus sains. Les HRs étaient de
0.84 (0.79, 0.89) et de 0.77 (0.72, 0.82), pour un écart type de HLIIMC et de HLIRTH,
respectivement. L’exclusion du tabagisme du HLIRTH a montré un HR de 0.88 (0.82, 0.94). La
FPA estimée pour une amélioration de la consommation d’alcool, de la qualité de
l’alimentation, du RTH et du niveau d’activité physique était de 14% (6%, 21%) et de 19%
(11%, 26%) si l’amélioration du tabagisme était considérée. L’adhérence à des habitudes
saines était fortement et inversement associée au risque de CP au-delà du rôle bénéfique de
l’évitement du tabagisme.

Les travaux inscrits dans ce projet de thèse ont apporté des connaissances informatives sur
l’étiologie du CP, et soutiennent le développement de mesures de santé publique promouvant
sa prévention par l’adoption de modes de vie sains. Les recherches sur les facteurs de risque
du CP seront poursuivies en s’appuyant sur des projets de plus grande envergure exploitant
des données d’études poolées. Une collaboration entre le CIRC et l’école de santé publique de

6
Harvard est actuellement en cours. Ce projet a pour but d’étudier les cancers pour lesquels le
niveau de preuve épidémiologique concernant leur relation avec la consommation d’alcool est
limité, le CP inclus, en regroupant à l’échelle internationale les données de 36 études
prospectives. Il permettra notamment d’évaluer la relation parmi les non-fumeurs et les profils
de consommation d’alcool plus spécifiques, en vue d’étudier des hypothèses
épidémiologiques jusque-là rarement examinées du fait de la faible occurrence du CP.

7
8
 ACKNOWLEDGMENTS

I would like to express my sincere gratitude to all the persons who have been involved all
along these years of my PhD at the International Agency for Research on Cancer (IARC).

Especially, I would like express my grateful acknowledgments to the following persons:

To Pietro Ferrari, PhD, Head of the Nutritional Methodology and Biostatistics Group, who
was my PhD supervisor and director, and offered me the incredible opportunity to spend these
3 epic years of PhD under his supervision. Thank you for what you taught me, for your
patience and for sharing with me your scientific values.

To Paul Brennan, PhD, Head of the Genetics Section, who was my PhD co-supervisor and
who brought his valuable guidance to this PhD program.

To Marc Gunter, PhD, Head of the Nutrition and Metabolism Section, who hosted me during
these 3 years of PhD in his section.

To Anouk Berger, Head of the Education and Training Group, for her support, her kindness,
as well as her positive and constructive advice all along these years.

To my colleagues and collaborators from the Nutritional Methodology and Biostatistics

Group on this work, Drs Vivian Viallon, Heinz Freizling, Flavie Perrier, Nada Assi,
Kuanrong Li, and Cristian Ricci, as well as to Ms Carine Biessy and Mr Bertrand Hemon,
who brought their brilliant contribution thanks to their precious advice.

To Drs Agnès Fournier (IARC) and Christine Lasset (Centre Léon Berard) for their
contribution and guidance during my PhD reviews.

To Drs Eric J Duell, Elisabeth Weiderpass, Bas Bueno-de-Mesquita, Alexandra Neiters,

Delphine Casabonne, Marta Solans, Fatemeh Saberi Hosnijeh, Roel Vermeulen, and all the
other members of the pancreatic cancer and lymphoma EPIC working groups, for sharing
their constructive recommendations and making me benefited from their expertise.

To Drs Ellen Kampman, Matty P Weijenberg and Edoardo Botteri for accepting to be the
reviewers of my thesis manuscript, and to Drs Béatrice Fervers and Michael Leitzmann to
take part as jury members to my thesis defence.

To my amazing colleagues, Drs Hannah Lennon, Nada Assi, Mathilde His, Flavie Perrier, and
Mr Sémi Zouiouich for their daily support, precious advice and participation to the review of

9
this thesis manuscript as well as my other colleagues, Drs Hwayoung Noh, Ana-Lucia Mayen
Chacon, Laura Trisjburg, Marta Pittavino, Ms Ghazaleh Dashti, Ms Anne-Sophie Navionis
and Ms Karina Zaluski and of course all my NMB and NME colleagues. All of them made me
proud to work by their side every single day during this thesis. I learned a lot from all of them.

To my family and friends Mr Clément Scala, Ms Elsa Pichon, Ms Emilie Naudin, Ms Sylvie
Goulet-Naudin, and Mr Eric Naudin for following and supporting me day by day in this
unexpected adventure.

10
 LIST OF SCIENTIFIC WORKS
Published

1. S Naudin, K Li, T Jaouen, N Assi, C Kyrø, A Tjønneland, K Overvad, MC Boutron-

Ruault, V Rebours, AL Vedie, H Boeing, R Kaaks, V Katzke, C Bamia, A Naska, A
Trichopoulou, F Berrino, G Tagliabue, D Palli, S Panico, R Tumino, C Sacerdote, PH Peeters,
HB Bueno-de-Mesquita, E Weiderpass, IT Gram, G Skeie, MD Chirlaque, M Rodrıguez-
Barranco, A Barricarte, JR Quiros, M Dorronsoro, I Johansson, M Sund, H Sternby, KE
Bradbury, N Wareham, E Riboli, M Gunter, P Brennan, EJ Duell and P Ferrari. Lifetime
and baseline alcohol intakes and risk of pancreatic cancer in the European Prospective
Investigation into Cancer and Nutrition study. Int J Cancer (2018). doi:10.1002/ijc.31367.

2. M Solans, Y Benavente, M Saez, A Agudo, S Naudin, FS Hosnijeh, H Noh, H

Freisling, P Ferrari, C Besson, Y Mahamat-Saleh, MC Boutron-Ruault, T Kühn, R Kaaks, H
Boeing, C Lasheras, M Rodríguez-Barranco, P Amiano, JM Huerta, A Barricarte, JA
Schmidt, P Vineis, E Riboli, A Trichopoulou, C Bamia, E Peppa, G Masala, C Agnoli, R
Tumino, C Sacerdote, S Panico, G Skeie, E Weiderpass, M Jerkeman, U Ericson, F Späth,
LM Nilsson, CC Dahm, K Overvad, AK Bolvig, A Tjønneland, S de Sanjose, G Buckland, R
Vermeulen, A Nieters and D Casabonne. Adherence to the Mediterranean diet and lymphoma
risk in the European Prospective Investigation into Cancer and Nutrition. Int J Cancer (2018).
doi:10.1002/ijc.32091.

3. M Solans, Y Benavente, M Saez, A Agudo, P Jakszyn, S Naudin, FS Hosnijeh, M

Gunter, I Huybrechts, P Ferrari, C Besson, Y Mahamat-Saleh, MC Boutron-Ruault, T Kühn,
R Kaaks, H Boeing, C Lasheras, MJ Sánchez, P Amiano, MD Chirlaque, E Ardanaz, JA
Schmidt, P Vineis, E Riboli, A Trichopoulou, A Karakatsani, E Valanou, G Masala, C
Agnoli, R Tumino, C Sacerdote, A Mattiello, G Skeie, E Weiderpass, M Jerkeman, J Alves
Dias, F Späth, LM Nilsson, CC Dahm, K Overvad, KE Nielsen Petersen, A Tjønneland, S de
Sanjose, R Vermeulen, A Nieters, and D Casabonne. Inflammatory potential of diet and risk
of lymphoma in the European Prospective Investigation into Cancer and Nutrition. Eur J Nutr
(2019). doi:10.1007/s00394-019-01947-0.

4. M Obón-Santacana, L Luján-Barroso, H Freisling, S Naudin, MC Boutron-Ruault, FR

Mancini, V Rebours, T Kühn, V Katzke, H Boeing, A Tjønneland, A Olsen, K Overvad, C

11
Lasheras, M Rodríguez-Barranco, P Amiano, C Santiuste, E Ardanaz, KT Khaw, NJ
Wareham, JA Schmidt, D Aune, A Trichopoulou, P Thriskos, E Peppa, G Masala, S Grioni, R
Tumino, S Panico, HB Bueno-de-Mesquita, V Sciannameo, R Vermeulen, E Sonestedt, M
Sund, E Weiderpass, G Skeie, CA González, E Riboli, and EJ Duell. Consumption of nuts
and seeds and pancreatic ductal adenocarcinoma risk in the European Prospective
Investigation into cancer and Nutrition. Int J Cancer (2019). doi:10.1002/ijc.32415.

5. L Cirera, JM Huerta, MD Chirlaque, K Overvad, M Lindström, S Regnér, A

Tjønneland, MC Boutron-Ruault, V Rebours, G Fagherazzi, VA Katzke, H Boeing, E Peppa,
A Trichopoulou, E Valanou, D Palli, S Grioni, S Panico, R Tumino, F Ricceri, C van Gils, R
Vermeulen, G Skeie, T Braaten, E Weiderpass, S Merino, MJ Sánchez, N Larrañaga, E
Ardanaz, M Sund, K Khaw, TJ Key, M Jenab, S Naudin, N Murphy, D Aune, H Ward, E
Riboli, HB Bueno-de-Mesquita, C Navarro, and EJ Duell. Socioeconomic effect of education
on pancreatic cancer risk in Western Europe: an update on the EPIC cohorts Study. Cancer
Epidemiol Biomarkers Prev (2019). doi:10.1158/1055-9965.EPI-18-1153.

6. S Naudin, V Viallon, D Hashim, H Freisling, M Jenab, E Weiderpass, F Perrier, F

McKenzie, HB Bueno-de-Mesquita, A Olsen, A Tjønneland, CC Dahm, K Overvad, FR
Mancini, V Rebours, MC Boutron-Ruault, V Katzke, R Kaaks, M Bergmann, H Boeing, E
Peppa, A Karakatsani, A Trichopoulou, VM Pala, G Masala, S Panico, R Tumino, C
Sacerdote, AM May, CH van Gils, C Rylander, KB Borch, MD Chirlaque López, MJ
Sánchez, E Ardanaz, JR Quirós, P Amiano Exezarreta, M Sund, I Drake, S Regnér, RC
Travis, N Wareham, D Aune, E Riboli, MJ Gunter, EJ Duell, P Brennan, and P Ferrari.
Healthy lifestyle and the risk of pancreatic cancer in the EPIC study. Eur J Epidemiol (2019).
doi:10.1007/s10654-019-00559-6.

Ongoing publication / Article in preparation

1. S Naudin, M Solans, FS Hosnijeh, A Nieters, R Vermeulen, D Casabonne, P

Brennan, and P Ferrari. Healthy lifestyle and the risk of lymphoma in the European
Prospective Investigation into Cancer and Nutrition study (Under Revision) Int J Cancer
(2019).

12
2. FS Hosnijeh, M Kolijn, D Casabonne, A Nieters, M Solans, S Naudin, P Ferrari, JD
Mckay, E Weiderpass, C Besson, V Perduca, FR Mancini, G Masala, V Krogh, F Ricceri, JM
Huerta, D Petrova, N Sala, A Trichopoulou, A Karakatsani, C La Vecchia, R Kaaks, F
Canzian, D Aune, H Boeing, MB Schulze, A Perez-Cornago, AW Langerak, V van der
Velden, & R Vermeulen. Mediating effect of soluble B-cell activation immune markers on the
association between anthropometric and lifestyle factors and lymphoma development.
Haematologica (Submitted) (2019).

3. S Yammine, I Huybrechts, C Biessy, LDossus, EK Aglago, S Naudin, P Ferrari, E

Weiderpass, A Tjønneland, L Hansen, A Trichopoulou, A Karakatsani, C La Vecchia, B
Vassiliki, B, I Torhild Gram, MJ Sánchez, MJ Gunter & V Chajès. Dietary and circulating
fatty acids and ovarian cancer risk in the European Prospective Investigation into Cancer and
Nutrition study. Ann Oncol (Submitted) (2019).

4. FS Hosnijeh, D Casabonne, A Nieters, M Solans, S Naudin, S, P Ferrari, JD Mckay

& R Vermeulen. Association between anthropometry, lifestyle factors and future risk of B cell
lymphoma: an exposome wide analysis. (In preparation) (2019).

Poster abstract

1. S Naudin, C Biessy, F McKenzie, EJ Duell, P Ferrari, and P Brennan. Risk of

pancreatic cancer and non-Hodgkin lymphoma associated with healthy lifestyle behaviors in
the EPIC study. Rev DÉpidémiologie Santé Publique (2018).doi:10.1016/j.respe.2018.05.180.

Disclaimer for IARC authors

Where authors are identified as personnel of the International Agency for Research on Cancer
/ World Health Organization, the authors alone are responsible for the views expressed in this
article and they do not necessarily represent the decisions, policy or views of the International
Agency for Research on Cancer / World Health Organization.

13
14
 TABLE OF CONTENTS
Abstract ................................................................................................................................... 1
Résumé ................................................................................................................................... 3
Résumé subtantiel ................................................................................................................... 5
Acknowledgments .................................................................................................................. 9
List of scientific works ......................................................................................................... 11
Table of contents .................................................................................................................. 15
List of figures and tables ...................................................................................................... 17
Abbreviations........................................................................................................................ 18
Chapter I: Epidemiology of pancreatic cancer ................................................................... 20
1. Physiopathology of pancreatic cancer ........................................................................... 22
2. Descriptive epidemiology of pancreatic cancer ............................................................ 24
2.1. A rare and fatal cancer worldwide ..................................................................... 24
2.2. Increasing trends of pancreatic cancer ............................................................... 27
2.3. A rare cancer with a reduced screening effectiveness ........................................ 28
3. Etiology of pancreatic cancer ........................................................................................ 29
3.1. Lifestyle and environmental risk factors ............................................................ 29
3.1.1. Smoking ...................................................................................................... 29
3.1.2. Body fatness ................................................................................................ 30
3.1.3. Physical activity .......................................................................................... 31
3.1.4. Diet .............................................................................................................. 31
3.1.5. Alcohol ........................................................................................................ 32
3.2. Morbid conditions .............................................................................................. 34
3.2.1. Diabetes ....................................................................................................... 34
3.2.2. Pancreatitis .................................................................................................. 34
3.2.3. Infectious diseases ...................................................................................... 35
3.3. Genetic factors .................................................................................................... 36
4. Challenges of pancreatic cancer for epidemiological studies ....................................... 38
5. Thesis objectives ........................................................................................................... 38
Chapter II: The EPIC study .................................................................................................. 40
1. Context and objectives .................................................................................................. 42
2. Study population............................................................................................................ 42
3. Ascertainment of outcomes ........................................................................................... 43

15
 4. Diet and lifestyle assessment......................................................................................... 44
5. Biobank, blood samples collection ................................................................................ 44
Chapter III: Alcohol consumption and the risk of pancreatic cancer ............................. 46
1. Background ................................................................................................................... 48
2. Objectives ...................................................................................................................... 48
3. Methods ......................................................................................................................... 49
4. Results ........................................................................................................................... 50
5. Conclusions ................................................................................................................... 52
6. Scientific article ............................................................................................................. 52
Chapter IV: Healthy lifestyle index and the risk of pancreatic cancer ............................. 66
1. Background ................................................................................................................... 68
2. Objectives ...................................................................................................................... 68
3. Methods ......................................................................................................................... 69
4. Results ........................................................................................................................... 70
5. Conclusions ................................................................................................................... 71
6. Scientific article ............................................................................................................. 71
Chapter V: Additional epidemiological research evaluations............................................ 84
1. The Pooling Project on Alcohol and Cancer ................................................................. 86
1.1. Background and significance ............................................................................. 86
1.2. Objectives ........................................................................................................... 87
1.3. Study design ....................................................................................................... 87
1.4. Data centralization and harmonization ............................................................... 88
2. Lifestyle factors in association with the risk of lymphoma........................................... 92
2.1. Background ........................................................................................................ 92
2.2. Methods .............................................................................................................. 93
2.3. Results ................................................................................................................ 93
2.4. Conclusions ........................................................................................................ 93
2.5. Short communication ......................................................................................... 94
Chapter VI: General conclusion ......................................................................................... 120
1. Main findings .............................................................................................................. 122
2. Perspectives ................................................................................................................. 123
References ............................................................................................................................. 126
Annexes ................................................................................................................................. 142
Annex 1............................................................................................................................... 144

16
 LIST OF FIGURES AND TABLES
Figures

Figure 1. Anatomy and histology of the pancreas (Source: Encyclopedia Britannica, Inc.) .. 22

Figure 2. Model for pancreatic carcinogenesis displaying progression from normal cells to
PanIN precursor low grade (PanIN-1 and -2) and high grade (PanIN-3) lesions, invasive
cancer, and metastatic pancreatic cancer (Source: Chhoda et al., The American Journal of
Pathology, 2019). .............................................................................................................. 23

Figure 3. Estimated age-standardized incidence and mortality rates (per 100,000 person-
years) of pancreatic cancer by World areas in 2018 (both sexes and all ages), ranked by
incidence rate. The five firt and the five last world areas are displayed (Source:
GLOBOCAN 2018, IARC/WHO) .................................................................................... 25

Figure 4. Estimated age-standardized incidence rates of pancreatic cancer in the World in

2018 (for both sexes and all ages) (Source: GLOBOCAN 2018, IARC/WHO) .............. 25

Figure 5. Sex-specific incidence and mortality rates of pancreatic cancer by World area
(Source: GLOBOCAN 2018, IARC/WHO) ..................................................................... 26

Figure 6. Trends in recorded per capita alcohol consumption by adults in selected countries in
the World (Source: Ezzati and Riboli, The New England Journal of Medicine, 2013) ... 33

Figure 7. Countries and centers involved in the EPIC study (Source: IARC/WHO) ............. 42

Figure 8. Hazard Ratio (HR) functions (and 95% CI) describing the curvilinear dose–
response relationship between baseline and lifetime alcohol intake (g/day) and PC risk,
according to the frequency of cases in all participants. .................................................... 51

Figure 9. Sensitivity analysis using external information from the scientific literature on
history of chronic pancreatitis (Z), in the relationship between baseline alcohol intake (X)
and the risk of pancreatic cancer (D). ............................................................................... 52

Figure 10. Scoring system combining the 5 lifestyle factors into the Heathy Lifestyle Index
based on the waist-to-hip ratio (HLIWHR). 1 For the HLI based on the BMI (HLIBMI), sex-
specific waist-to-hip ratio quintiles were replaced by categories of BMI at baseline using
the following cut-offs: (4) 22–23.9 kg.m-2, (3) 24–25.9 kg.m-2, (2) <22 kg.m-2, (1) 26–
29.9 kg.m-2, and (0) >30 kg.m-2 ......................................................................................... 69

Tables

Table 1. Descriptive characteristics, alcohol consumption and smoking status of PPAC

participating cohorts. ......................................................................................................... 90

Table 2. Criteria for cancer outcomes harmonization in the PPAC project. ........................... 91

17
 ABBREVIATIONS
BCL: mature B-cell lymphoma

BMI: body mass index

CI: confidence interval

CLL/SLL: chronic lymphocytic leukemia and small lymphocytic leukemia

DLBCL: diffuse large B-cell lymphoma

EPIC: European prospective investigation into cancer and nutrition

FL: follicular lymphoma

HBV: hepatitis B virus

HCV: hepatitis C virus

HDI: human development index

HL: Hodgkin lymphoma

HLI: healthy lifestyle index

HR: hazard ratio

MT/NK: mature T and natural killer-cell lymphoma

NHL: non-Hodgkin lymphoma

OR: odds ratio

PanIN: pancreatic intraepithelial neoplasms

PC: pancreatic cancer

PCN/MM: plasma cell neoplasm and multiple myeloma

RR: risks ratio

SD: standard deviation

UADT: upper aero-digestive tract

WC: waist circumference

WCRF/AICR: World Cancer Research Fund / American Institute for Cancer Research

WHR: waist-to-hip ratio

18
19
 CHAPTER I:
EPIDEMIOLOGY
OF PANCREATIC CANCER

20
21
1. Physiopathology of pancreatic cancer

The pancreas is a gland of the digestive system involved in both endocrine and exocrine
metabolic activities (1,2). The endocrine function allows the regulation of the glucose
homeostasis through hormones secretion (i.e. glucagon and insulin). These hormones are
secreted by particular cell clusters called the islets of Langerhans which represent 2% of the
overall pancreatic mass. The exocrine function of the pancreas includes the secretion of
digestive fluids by acinar cells representing 85% of pancreatic mass. These produced
pancreatic fluids are drained to the duodenum through different pancreatic ducts (1) (Figure
1). Common disorders related to the pancreas are pancreatitis, pancreatic cysts, diabetes and
pancreatic cancer (PC).

Figure 1. Anatomy and histology of the pancreas (Source: Encyclopedia Britannica, Inc.)

PC can be categorized into exocrine and endocrine types depending on the function of the cell
from which it arises. Exocrine tumors represent around 95% of PC (3), of which more than
80% of these exocrine PC are adenocarcinomas and nearly all of them arise from the cells
lining the pancreatic ducts, thus named pancreatic ductal adenocarcinomas (PDAC). Rarer
types of exocrine PC tumors include cystic tumors and acinar cells tumors, however, only a
small proportion of cystic tumors are cancerous. Acinar cells tumors present a slower growing

22
pace than other pancreatic tumors and generally occur at younger age. Both these rarer types
of exocrine tumor have a better prognosis compared with other exocrine tumors (4).

A growing body of evidence has helped to establish that invasive adenocarcinoma develops
from well-defined progression of morphologic entities originating from noninvasive lesions.
The most common precursor lesion leading to invasive adenocarcinoma is a microscopic
lesion called the pancreatic intraepithelial neoplasia (PanIN) (5). PanIN lesions originate from
pancreatic duct cells and progress from PanIN-1 to PanIN-3 morphology, finally transitioning
to invasive carcinoma when the cancerous ductal cells move to the pancreatic parenchyma.
The transitions between the different PanIN stages are driven by accumulating genetic and
epigenetic alterations. Telomere shortening and kRAS mutations appear to be early events in
transformation from normal epithelium to PanIN-1, while mutations in TP53, BRCA2 and
SMAD4 mutations are late events in the progression from PanIN-1 to PanIN-3 (Figure 2) (6).
Eighty percent of these mutations occurring in the development of pancreatic adenocarcinoma
appear to be sporadic (7). Less frequently, PC adenocarcinoma arise from macroscopic cystic
precursor lesions including intraductal papillary mucinous neoplasm (IPMN) and mucinous
cystic neoplasm (MCN) (7–11). Usually diagnosed at early stages within younger
populations, IPMN and MCN benefit from a better prognosis compared with PanIN (12).

Figure 2. Model for pancreatic carcinogenesis displaying progression from normal cells to
PanIN precursor low grade (PanIN-1 and -2) and high grade (PanIN-3) lesions, invasive
cancer, and metastatic pancreatic cancer (Source: Chhoda et al., The American Journal of
Pathology, 2019).

23
Endocrine or neuroendocrine pancreatic tumors are uncommon and account for only 5% of
PC, with incidence rates below 1 case per 100,000 person-years (PY). Most of these tumors
are malignant, but slow-growing and have a better prognosis than exocrine tumors (13).

PC are mostly diagnosed at an advanced stage and more than 80% of cases have already
spread to a regional or distant stage (14). Early stages of PC tumor development are
asymptomatic, and when symptoms occur, they are usually non-specific to PC and can be
attributable to a broad number of diseases, including gastric ulcers, choledochal cysts,
pancreatitis as well as other gastro intestinal cancers (3,15). To date, treatment approaches in
patients diagnosed with PC largely consist in palliative care which allows a few months to be
saved only, even in countries with the most advanced cancer care protocols (7,16).

2. Descriptive epidemiology of pancreatic cancer

2.1. A rare and fatal cancer worldwide

PC is the 15th cause of cancer worldwide, with an estimated 458,918 new cases diagnosed in
2018, accounting for 2.5% of all cancers. This is a rare cancer with an age-standardized
incidence rate of 4.8 per 100,000 PY (17). Ranked as the 9th deadliest cancer worldwide, in
2018 PC accounted for 4.5% cancer-related deaths with a total number of 432,242 deaths and
an age-standardized mortality rate of 4.4 per 100,000 PY (17). With a mortality-incidence
ratio close to one, PC is one of the most fatal cancers, and it has been recently ranked as the
3rd leading cause of cancer-related death after lung and colorectal cancers in North America
and Western Europe populations (17).

The occurrence of PC differs according to World areas and in 2018 the highest incidence rates
were observed in Western Europe (8.3 per 100,000 PY), North America (7.6 per 100,000 PY)
and Central-Eastern Europe (7.5 per 100,000 PY), while the lowest rates were found in South
Central Asia (1.0 per 100,000 PY), Eastern Africa (1.4 per 100,000 PY) and South-Eastern
Asia (2.1 per 100,000 PY) (Figure 3) (17). Countries with the highest incidence of PC are
Hungary (10.8 per 100,000 PY) and Uruguay (10.7 per 100,000 PY), while the lowest
incidence rates were found in Guinea (0.35 per 100,000 PY) and Malawi (0.40 per 100,000
PY) (Figure 4) (17). Mortality rates follow almost the same pattern as incidence rates across
world areas with the highest rates in Western Europe (7.6 per 100,000 PY) and Central-
Eastern Europe (7.3 per 100,000 PY), while the lowest rates were found in South Central Asia
(1.0 per 100,000 PY) and Eastern Africa (1.4 per 100,000 PY) (Figure 3) (17). As for

24
incidence rates, mortality rates are about 1.8-time higher in men than in women (Figure 5)
(17).

Figure 3. Estimated age-standardized incidence and mortality rates (per 100,000 person-
years) of pancreatic cancer by World areas in 2018 (both sexes and all ages), ranked by
incidence rate. The five firt and the five last world areas are displayed (Source: GLOBOCAN
2018, IARC/WHO)

Figure 4. Estimated age-standardized incidence rates of pancreatic cancer in the World in

2018 (for both sexes and all ages) (Source: GLOBOCAN 2018, IARC/WHO)

25
It was reported that countries with high level of human development index (HDI) have the
largest incidence rates of PC (18). Although genetic and lifestyle profiles may account for the
observed geographic disparities, access to diagnostic tools and health care modalities may
well influence the chance of identifying PC in these populations. Differences may also be
attributed to the quality of cancer registries, specifically in terms of coverage, completeness
and accuracy (15).

PC occurrence also slightly differs by gender with a higher incidence rate in men (5.5 per
100,000) than in women (4.0 per 100,000) (17,19). Sex-specific incidence rates by world area
are presented in the Figure 5. The hypothesis suggesting that reproductive factors could
influence PC risk by gender have been studied during the last decade (20,21), and differences
in PC rates between men and women are believed to be due to differences in lifestyle patterns,
including smoking and obesity, or undiscovered genetic factors (22).

Figure 5. Sex-specific incidence and mortality rates of pancreatic cancer by World area
(Source: GLOBOCAN 2018, IARC/WHO)

The risk of developing PC increases with age. Almost 90% of cases are diagnosed after 55
years of age and the highest incidence is reported in individuals aged between 60 and 80
years, with a median age at diagnosis of 71 years (15,19,23). In the American population,

26
with a mean age at diagnosis of 65 and 67 in men and women, respectively, the age at
diagnosis is steadily increasing with a continuous trend towards the years of age 80 (24,25).
Whereas, in India, a country with a pyramidal age structure, the incidence of PC starts to rise
in the 5th decade, with a peak at age 70 (22,25).

Survival rates of PC are still very low, even in countries with high HDI (15,17). In 2015, the
Eurocare-5 Working Group examined data from 29 European countries, and reported a 1-year
survival rate of 26% and a 5-year survival rates of 7% (26)The United-Kingdom had one of
the lowest 5-year survival rates in Europe with a rate of 3% in 2011 (27). In US, the survival
rate was not considerably higher with a rate of 8% in 2018 (25). It has been shown that the 5-
year survival rate can be better when pancreatic tumors are diagnosed at less advanced stage
of progression with survival rate of 29% in localized tumors, while in tumors spread at
regional and distant levels the survival rates decreased considerably to 11% and 3%,
respectively. Other tumor’s characteristics, as well as age at diagnosis, gender, treatment and
heath care system modalities, preexisting health conditions and lifestyles factors affect PC
survival (15,27).

2.2. Increasing trends of pancreatic cancer

In past decades, PC incidence gradually increased worldwide (17), particularly in North

America, Western Europe, and Oceania for both sexes (19,28). In the last 20 years, the
greatest increase in incidence occurred in France between 2003-2012 with an average annual
percentage change of 4.4% and 5.4% in men and women, respectively (28), while in the US
population, PC incidence has gradually increased over the last 40 years, with an annual
percentage change of 1.6% between 2000 and 2014 (19,29). Trends of different histologic
subtypes of pancreatic tumor, extracted from the SEER database, showed that incidence rose
most rapidly during 1992-2013 for adenocarcinomas (the largest histologic subtype),
compared with other less common PC subtypes (19).

As a consequence of these increasing trends in incidence, in contrast to most cancers (14),

mortality from PC has globally steadily increased over the last 50 years for both men and
women, especially in Europe, Japan, Brazil, Korea, and US (30,31). Overall no substantial
changes in PC survival rates have occurred over (15,26). Minor improvement in mortality
rates have been observed during the last decade in some high income countries from Western

27
Europe, North America and in Japan (29,30), and overall PC prognosis has remained largely
unchanged worldwide (15,27,30).

Future temporal trends of PC over the period 2018-2040 were evaluated in GLOBOCAN
2018 (32). It is expected that PC incidence and mortality will follow an increasing trend, with
78% more cases and 80% more deaths from PC by 2040 in the World, notably among
European and American populations, in 2018. The number of deaths by PC are expected to be
25% higher than the number of deaths due to breast cancer in the European Union by 2025
(33). In the US population, it is possible that PC could surpass colorectal cancers and become
the second leading cause of cancer-related death after lung cancer by 2030 (34).

In absence of major improvement in survival rates, the rise in PC incidence and mortality has
become a major public health concern (15). Improved diagnostic and death certification of the
disease might partially explain the increase in PC incidence worldwide (31). Trends of PC
might partly reflect patterns of tobacco consumption throughout the world and it is likely that
a level off in incidence and mortality trends observed in high income countries between 1970
and 1990 might be related to implementation of tobacco control policies (19,30). In the most
recent decades, the increase in obesity and diabetes prevalence may also partly explain the
stable or slowly rising PC rates in the Western World (15,19,28,33).

2.3. A rare cancer with a reduced screening effectiveness

The lack of opportunity of effective treatments for PC is mostly due to late diagnoses. PC
early detection, prior to invasion, is to date the primary goal of PC screening to improve the
chance for patients to undergo curative chemotherapy and surgery rather than palliative
treatments (6). However, despite its aggressiveness and high degree of lethality, PC screening
is not a standard practice in the general population with the present technologies of diagnosis
(6,35). Current screening strategies have been so far limited to the clinical follow-up of
patients already identified as high risk individuals based on family history of PC or
underlying genetic syndrome, who represent 7% and 3% of PC cases, respectively (12). As a
result, ongoing approaches target only a small part of the PC burden. In addition, the success
of screening resides in noninvasive detection of the smallest lesions, but no gold standard
technic for PC diagnosis has been yet identified, and expensive imaging techniques combined
with highly invasive endoscopic examination have been so far the most reliable methods of
diagnosis and staging of the disease (15).

28
Recently, advances in the understanding of the carcinogenic pathways driving progression of
precursor lesions have shown promising results (35). Progression from PanIN-3 lesion to
malignant adenocarcinoma would occur in about 12 years (24,36,37), allowing a large time
window for early detection. The use of blood biomarker for early detection is a fast
developing new field of research, but current candidate biomarkers are non-specific to PC and
at present can only be used to confirm a diagnosis, predict a prognosis, or evaluate a
recurrence after resection (12,38). Additionally, knowledge is so far limited to biopsy
biomarkers collected in high risk populations undergoing invasive medical examination, and
biomarkers collected through non-invasive technique in general population are yet to be
defined (6,15).

In absence of reliable screening strategies in the general population, primary prevention

aiming at improving awareness on the management of modifiable lifestyle, and thus the
quantification of the associations of these modifiable risk factors, is of utmost importance.

3. Etiology of pancreatic cancer

Over the last decades, the identification of risk factors causally related to the occurrence of
pancreatic tumors has been a main research priority. Although the knowledge about the
pathogenesis of pancreatic tumors has greatly improved, PC etiology remains poorly
understood and only a few risk factors have been consistently causally related to PC
occurrence.

3.1. Lifestyle and environmental risk factors

3.1.1. Smoking

Smoking is one of the major risk factors for PC. In 2004, the International Agency for
Research on Cancer (IARC) monographs recognized the causal relationship, with a strong
positive association (39). It has been estimated that tobacco smoking account for 11 to 32 %
of the total number of PC cases (40). Compared with non-smokers, current smokers have been
estimated to have a 74% (95%CI: 61%, 87%) higher risk of PC, while weaker increases in
risk have been observed in former smokers with a 20% (11%, 29%) higher risk, in a meta-
analysis based on 82 independent studies (41). In a more recent meta-analysis of 42 studies,
where authors reported similar strength of association, the risk of PC was positively

29
associated with the duration of smoking, the number of cigarettes smoked per day and also
with passive smoking, while it was inversely related with time since smoking cessation (42).
Cigarette smoking has been hypothesized to generate circulating carcinogens that cause
pancreatic inflammation and mutations in proto-oncogenes (KRAS) and tumor suppressor
genes (TP53). Furthermore, smoking appears to interact with genetic predispositions,
including hereditary pancreatitis and familial PC, other strong risk factors for PC (43).

3.1.2. Body fatness

It has been suggested that the increase in obesity prevalence, as well as metabolism related
diseases such as diabetes, could be partly responsible for the positive trends of PC, especially
in Western countries (44).

As a result, the relationship between excess body fatness and the risk of PC has been widely
explored in cohort studies and several meta-analyses, which consistently have shown positive
associations. Evidence from epidemiological studies was evaluated and considered as
convincing by the WCRF/AICR expert panel and the association has been even more recently
confirmed by the IARC monograph program (45,46). The body mass index (BMI), reflecting
overall adiposity, is the most common measure evaluated in studies in relation to PC risk.
Consistently in men and women, a continuous 5 kg/m² increase in BMI has been associated
with a 10% to 18% increase in PC risk (46–48). Further, the specific role of central adiposity
was evaluated through measures of both waist circumference (WC) and waist-to-hip ratio
(WHR) in several studies, showing that a 10 cm increase in WC or a 0.1 increase in WHR
were associated with 7% and 9% higher risks of PC, respectively (46,49–52). It has been
suggested that the underlying mechanisms for the association with body fatness are driven by
chronic inflammation due to remodeling of adipose tissue and ectopic lipid accumulation.
Excess lipid storage in adipose tissue can lead to infiltration of lipids within muscle, liver and
pancreatic tissue, and cause impairment of lipid processing and clearance within these tissues
(53,54). Therefore, lipid intermediates impair the function of cells and increase the release of
cytokines, which foster inflammation as well as insulin resistance (45). Recent findings from
6 European cohorts showed that 42% of the relationship between BMI and PC risk was
mediated by insulin resistance in men and women collectively, with a larger proportion
mediated in women (91%) than men (20%) (55). It has been also suggested that the
relationship between obesity and PC risk could be additionally mediated by adiponectin, an

30
inflammatory factors (56), especially among non-smokers (57), possibly suggesting common
mechanism between obesity and tobacco smoking.

The role of adult attained height in the occurrence of PC was also reported by the
WCRF/AICR panel with probable evidence for a modest but significant positive association.
Adult attained height might be a marker of genetic, environmental, hormonal and nutritional
growth factors during growth (46).

3.1.3. Physical activity

Physical activity is believed to prevent PC by regulating body weight and decreasing insulin
resistance (58), DNA damage (59), and chronic inflammation (60). Since practice of physical
activity is a relatively easily modifiable lifestyle factor, it may offer possibilities for primary
prevention of PC. Nonetheless, evidence for an association between physical activity and the
risk of PC has been considered as limited and inconclusive in 2012 by the WCRF/AICR
experts’ panel. The summary RR of high versus low total physical activity was of 0.74
(95%CI: 0.55, 1.00) when based on a small number of studies (46,61). The statistical power in
this study to assess consistency of results regarding the role of physical activity over time was
insufficient, as well as the observed variation of the association by smoking status or level of
obesity. A recent meta-analysis including 30 individual studies showed a weak but
statistically significant inverse association for high versus low level of physical activity (RR:
0.93, 95%CI: 0.88, 0.98), with strongest estimate for consistent level of physical activity over
time (RR: 0.86, 95%CI: 0.76, 0.96, as compared with irregular past physical activity levels)
(62). No evidence for variation by smoking status or BMI was found (62), while a recent
evaluation in Asian population from the Shanghai Women and Men Health Studies showed a
significant inverse association in men only (63). Despite rising recent epidemiological
evidence and strong mechanistic hypotheses, the causal association between physical activity
and PC risk is still not strongly supported by epidemiological evidence. Results need to be
confirmed through interventional studies (62).

3.1.4. Diet

Despite the direct influence of diet energy imbalance on adiposity and evidence for the
relationship between obesity and PC, epidemiological studies examining the association
between any specific dietary patterns or food group and PC has been so far considered as

31
inconsistent and limited by the WCRF/AICR experts’ panel in 2012. However, since 2012
several individual studies have provided evidence of positive associations with meat intakes
(64), saturated fatty acids and industrial trans fatty acids (65), while whole grains (66), n-3
polyunsaturated fatty acids (65) and low-fat dietary patterns (67) were inversely related to PC
risk. These results were confirmed in a meta-analysis on the relation between dietary patterns
and the risk of PC with a data driven approach, were associated with reduced risk of PC,
where dietary patterns were characterized by high consumption of animal foods and
associated nutrients, starch, or other typical Western type foods elevated PC risk, while
dietary patterns rich in plant-based foods, whole grains, white meat, fiber, and associated
antioxidants (68). A series of a priori diet scores have been also implemented to reflect
adherence to specific dietary patterns, i.e. Mediterranean diet score, low inflammatory
potential diet score, adherence to diet guidelines, and overall did not find significant inverse
associations with PC (69–71). Overall, evidence of a relationship between diet and the risk of
PC are so far inconsistent, perhaps due to several methodological limitations, including the
lack of data from prospective studies and standardized methods to assess diet exposure.

3.1.5. Alcohol

The consumption of alcohol is one of the top-10 risk factors contributing to the worldwide
burden of disease. Alcohol consumption is responsible for around 2.7 million annual deaths
and 3.9% of the global burden of disease. Over the past four decades, per capita alcohol
consumption has doubled in the United Kingdom and Denmark; during the same period, it has
decreased by about one half in traditional wine-producing and wine-drinking countries such
as Italy and France, and levels of consumption in these countries have converged. Alcohol
consumption has increased steadily in Japan, China, and many other countries in Asia, where
it was previously low, while trends have been constant in the United States (Figure 6) (72).

There are several mechanisms through which alcohol is believed to promote cancer.
Conversion of ethanol in acetaldehyde has a direct carcinogenic effect on the cells throughout
the digestive tube where conversion occurs (73,74). Ethanol promotes production of highly
reactive oxygen species, which can damage DNA, can alter DNA methylation, and has
hormonal effects, including increasing estradiol levels (75). Furthermore, ethanol also
facilitates uptake of other carcinogens (e.g. from tobacco smoking), thus also increasing the
risk of tobacco-induced cancers. It may also propel already existing cancers through
immunosuppression, and angiogenesis, and decrease the effect of chemotherapeutics (76).
32
Figure 6. Trends in recorded per capita alcohol consumption by adults in selected countries in
the World (Source: Ezzati and Riboli, The New England Journal of Medicine, 2013)

In 2012, the Monograph program at the International Agency for Research on Cancer (IARC)
reviewed the available epidemiological evidence on the possible association between
alcoholic beverage consumption and the risk of cancer at 27 anatomical sites, and reported
that cancers of the upper digestive, liver, colorectum and breast are causally related to the
consumption of alcoholic beverages. However, evidence regarding the risk of PC was
considered as limited (77,78). Consecutively, the WCRF/AICR panel of experts also reviewed
evidence for an association between alcohol drinking and PC as limited or suggestive as
major part of evidence issued from case-control studies (46). In a more recent large meta-
analysis of 18 cohort studies, it was estimated that the association was restricted to heavy
alcohol drinkers of more than 50 grams of ethanol per day (g/day) with an increased risk of
PC by 19% when compared with nondrinkers or occasional drinkers (RR=1.19, 95%CI: 1.11,
1.28) (79). Similar strengths of associations were observed more recently in a meta-analysis
by Wang et al. with a 15% increasing risk of PC among individuals who drunk more than 24
g/day (RR= 1.15, 95 CI: 1.06, 1.25) when compared with light drinkers of 0 to 12 g/day (80).
These findings strengthen evidence of the associations observed for heavy alcohol use in prior
pooled analyses (81–84). However, these previous studies had relatively limited statistical
power to examine specific alcoholic beverages and effect modification by smoking status.
Given that alcohol abuse is a cause of chronic pancreatitis, it has been proposed that excessive
alcoholic intake could increase the risk of PC indirectly through this pathway (43).

33
 3.2. Morbid conditions

3.2.1. Diabetes

The relationship between both type I and II diabetes and the risk of developing PC has been
extensively evaluated over past decades and showed strong positive associations (85,86). PC
burden attributable to diabetes has been evaluated to represent 10 to 16% of PC cases.
Approximatively 80% of patients present an impaired glucose tolerance or a recent onset of
type 2 diabetes at the moment of PC diagnosis (87). Indeed, due to the major role of the
pancreas into glucose homeostasis, altered glucose metabolism may be a result of a
progressive loss in pancreatic functions and reflect the tumor invasion. Therefore, type-2
diabetes can be both a cause and a consequence of PC and experience a bidirectional
association (88).

More recent works have therefore focused on the evaluation of the duration between diabetes
onset and PC diagnosis to account for potential reverse causation (89–91). Evidence
suggested that overall long-term diabetes was associated with at 50% increased risk of PC
(40), with PC risk decreasing with the duration of diabetes. A meta-analysis of 88 studies
showed that RR for a duration of diabetes <1 year was 6.69 (95% CI: 3.80, 11.78), for a
duration <5 years 1.86 (1.56, 2.21), for a duration <10 years 1.72 (1.47, 2.00), and after 10
years RR was 1.36 (1.19, 1.55) (90). Another study showed that 30% excess risk persists for
more than two decades after a diagnosis of diabetes (91).

The underlying mechanisms potentially explaining this association combines a tumorigenic

role of hyperglycemia, the mitogenic effect of hyperinsulinemia and chronic inflammation
related to fat infiltration of the pancreas. This could be enhanced in the context of obesity
related diabetes due to insulin resistance (89).

3.2.2. Pancreatitis

Pancreatitis is a condition characterized by inflammation of the pancreatic tissue. While acute

pancreatitis occurs punctually under the influence of heavy alcohol drinking, gallstones
formation, medication, or infection, chronic pancreatitis is a progressive and long standing
inflammation of the pancreatic tissue leading to irreversible fibrosis and loss of pancreatic
cells’ functions overtime (92). This chronic inflammation has been postulated to facilitate
pancreatic cell progression to cancer (92). As for diabetes, the relation between pancreatitis
and PC is bidirectional (88,93).

34
Both acute and chronic pancreatitis have been evaluated in association with PC risk in a
pooled-analysis of 10 case-control studies from the international Pancreatic Cancer Case-
Control Consortium (PanC4). Pancreatitis diagnosed more than 2 years before PC diagnosis
was associated with PC risk, with odds ratio (OR) equal to 2.71 (95% CI: 1.96–3.74), and OR
was 13.56 (8.72–21.90) for pancreatitis occurring within less than 2 years before PC
occurrence. Population attributable fraction was also estimated at 1.3% (95% CI: 0.61%,
2.07%), suggesting that a relatively small proportion of PC could be avoided if pancreatitis
could be prevented (94). Evaluations focusing on chronic pancreatitis showed stronger
associations. A meta-analysis by Raimondi and colleagues suggested a strong association
between chronic pancreatitis and PC with pooled RR of 13.3 (95%CI: 6.1, 28.9) (95).
Nonetheless, chronic pancreatitis is a rare disease with an incidence rate ranging from 2 to 14
cases per 100,000 PY. Additionally, less than 5% of PC cases occur among individuals with
chronic pancreatitis (96). As heavy alcohol drinking is also a major cause of chronic
pancreatitis, it has been suggested that alcohol consumption might be related to PC risk with
chronic pancreatitis being the underlying mechanism between the two of them (7).

3.2.3. Infectious diseases

The associations between infections and occurrence of PC have been also investigated,
including both bacterial and viral.

Gastric colonization with Helicobacter pylori has been related with greater risk of PC. In a
recent meta-analysis of 8 case-control studies, a pooled estimate showed a positive significant
relationship between Helicobacter pylori infection and PC risk with a OR equal to 1.45 (95%
CI: 1.09, 1.92). In a multi-center hospital-based American case-control study, the association
has been examined with respect to Helicobacter pylori virulence protein phenotype CagA and
showed significant increased risk of PC in relation to CagA-negative Helicobacter pylori
seropositivity (OR: 1.68, 95% CI: 1.07, 2.66), but no association was observed for CagA
seropositivity (OR: 0.77, 95% CI: 0.52, 1.16) (97). The association was more particularly
restricted to individuals with non–O blood type with an OR equal to 2.78 (1.49, 5.20). It has
been suggested that the presence of A or B blood group antigens might influence the
properties of Helicobacter pylori binding (97). Although Helicobacter pylori prevalence varies
widely across continent and countries, population attributable fraction of PC was estimated as
great as 4-25% (40). Mechanisms involved in this relationship are still unknown.

35
Since the last decade, some studies have raised attention to the association between chronic
infections such as hepatitis B and C viruses (HBV and HCV) and PC occurrence (15,87,98).
An increasing amount of evidence from meta-analyses has indicated a positive association
between active HBV and HCV infections and PC (99–102). However, so far evidence from
observational studies are scarce, and require cautious interpretation regarding study design
and definitions of infection status and PC cases (101).

3.3. Genetic factors

Several hereditary and genetic factors have been identified as possible cause of PC. PC
occurring in people who do not carry a high risk genetic mutation, referred to as sporadic PC,
represent 90% of cases. Only a small proportion (approximatively 5-10%) is thought to be
explained by inherited forms like familial aggregations and high risk genetic syndromes
(103), which includes hereditary pancreatitis, hereditary genetic syndromes and familial
history of PC. The risk of PC related to such familial or genetic predisposition is considerably
greater than in the general population and varies depending on the gene affected and
concurrent lifestyle factors.

Hereditary pancreatitis is a rare inflammatory disorder characterized by recurrent acute

pancreatitis with a typical onset during childhood. It is caused by gain-of-function mutations
in the cationic trypsinogen gene (PRSSI) (104,105). PRSS1 mutations result in the over-
activation of the trypsin, a digestive enzyme produced by the pancreas. Approximately 70%
of pancreatitis result from the mutations R122H and N291 in the PRSS1 gene (105). At least
half of the patients diagnosed with hereditary pancreatitis will develop chronic pancreatitis
(106), and PC generally occurs about 30 years after the onset of chronic pancreatitis as a
result of a prolonged trypsin related chronic inflammation. Compared with the general
population, individuals with PRSSI mutations have been estimated to have 50 to 87 higher
risk of PC than non-mutation carriers (107,108), while the risk of PC has been estimated to be
between 54 and 154 higher in smokers who are also mutation carriers when compared to non-
smokers who are also non-mutation carriers (109).

The risk of PC may be higher in subjects with hereditary genetic syndromes for who germ-
line mutations have been identified. Mutations occurring in tumor suppressor genes like
STK11 gene (Peutz-Jeghers syndrome or PJS), CDKN2A gene (familial atypical malignant
melanoma syndrome or FAMMM), BRCA1 and BRCA2 genes (hereditary breast and ovarian

36
cancer syndrome or HBOC) and APC gene (familial adenomatous polyposis or FAP
syndrome) have been related to increase in risk ratio of PC ranging from 4 for HBCO
syndrome to 132 in PJS as compared to the general population (104,105,110). Mutations in
the mismatch repair genes MSH2 or MLH1 resulting in hereditary nonpolyposis colorectal
cancer (HNPCC or Lynch syndrome) were associated with a cumulative risk of PC 9-time
higher in older individuals. Mutation on the PALLD gene, encoding for control of cells shape
and motility, was also recently related to site-specific familial PC, a familial syndrome with
risk ratio ranging from 18-57 in comparison to general population. Additionally, KRAS, TP53
and SMAD4 mutations have been suspected to occur in precursor lesion of sporadic forms of
PC. Overall, whatever PC be sporadic or related to inherited syndromes, it has been estimated
that around 1,000 genetic mutations occur during the pancreatic tumorogenesis (111).

In the absence of known hereditary cancer syndrome related to a specific gene mutation,
family history of PC has been established as a major risk factor of PC (104). Familial PC
cases, defined as individuals with at least two first-degree relatives (parent, sibling or child)
diagnosed with PC, is estimated to affect about 5-10% of PC cases (15). In a pooled analysies
using data from nested case-control studies from the PanScan consortium, individual with a
family history of PC with one first degree relative diagnosed with PC had an 80% increased
risk of pancreatic adenocarcinoma (RR: 1.76, 95%CI: 1.19, 2.61) compared with individuals
with no reported family history (112).

The ABO blood group is also a genetically defined factor, which has been related to PC risk
studies over the last decades. Recent meta-analyses and pooled prospective studies showed
that the risk of developing PC was positively associated with non-O blood groups with 30–
40% higher risk of PC compared to individuals with the O blood group (40,113,114). In the
PanScan consortium study, Wolpin and colleagues combined data from 12 prospective
cohorts and showed that in comparison with subjects of O blood group, patients with blood
group A, AB, and B had higher risk of PC with ORs equal to 1.38 (95%CI: 1.18, 1.62), 1.47
(1.07,2.02), and 1.53 (1.21, 1.92), respectively (114). The mechanism by which the ABO
blood group is involved in the occurrence of PC is not known, although it has been suggested
that the alteration of the ABO glycosyltransferase activity, an enzyme which might be
involved in PC carcinogenesis, mainly by affecting cell proliferation, tumor invasion and
metastatic spread (115).

37
4. Challenges of pancreatic cancer for epidemiological studies

PC etiology has been extensively investigated. Although genetic factors are well identified,
they offer the potential to understand only a small proportion of PC cases occurring
worldwide. Besides, epidemiological investigations have led to the identification of a limited
number of modifiable lifestyle factors for which there are strong evidence for causal
relationships with PC occurrence, including tobacco smoking and body fatness. Evaluations
of the role of alcohol consumption, level of physical activity or diet have been so far limited
or inconclusive and offer limited room for the implementation of primary prevention policies.

PC is a rare cancer which has been mostly investigated in case-control studies, widely subject
to recall bias and selection bias. Potential true causal relationship between lifestyle risk
factors and PC, if they exist, are likely to be weak. Therefore, the most favorable context to
investigate such relationship should allow a large-scale prospective examination and have a
long follow-up of participants to have a sufficiently large number of cases to detect an
association.

5. Thesis objectives

The main objective of this PhD program was to investigate the etiology of PC by
strengthening epidemiological evidence on the role of alcohol consumption and more
generally the role of a combination of lifestyle factors in relation to PC incidence. For this
purpose, evidence from the ongoing multicenter prospective cohort study, the European
Prospective Investigation into Cancer and nutrition (EPIC) study, was used.

38
39
 CHAPTER II:
THE EPIC STUDY

40
41
1. Context and objectives

The European Prospective Investigation on Cancer and Nutrition (EPIC) study was initiated
in the 1990’s at the International Agency for Research on Cancer (IARC), based on the
limited success in identifying links between diet and cancer through experimental and case-
control studies, and the necessity to investigate prospectively at a large scale the occurrence of
cancer in relation to diet and lifestyle (116). The project started with a series of pilot studies to
evaluate the methodology for the establishment of a very large prospective cohort with the
collection and storage of blood samples and the feasibility of recruiting participants (117).

The principal goal of the EPIC study was to investigate the etiology and the incidence of
cancers, as well as many kinds of chronic diseases, in relation to diet, lifestyle and
environmental factors taking advantage of the contrast in cancer rates and dietary habits
between centers and countries and of the large overall size of the study.

2. Study population

From 1992 to 2000, 521,324 participants were recruited in 23 centers across 10 European
countries including Spain, Italy, France, the United Kingdom, Greece, Germany, Netherlands,
Denmark, Sweden, and Norway (Figure 7), and have been followed-up for more than 15
years. To date, the EPIC study represents one of the largest ongoing multicenter prospective
studies worldwide.

Figure 7. Countries and centers involved in the EPIC study (Source: IARC/WHO)

42
EPIC participants were recruited in the general population, of which 70% are women, aged
from 35 to 70 years. Exceptions were the French cohort (school and university employees),
the Spanish and Italian centers (blood donors), Utrecht and Florence centers (breast cancer
screening participants), and Oxford (vegetarians and ‘health conscious’ participants). In
France, Norway, Utrecht and Naples women only were recruited. Study participants provided
informed consent before completing questionnaires at baseline. At baseline, all of the cohort
members signed an informed consent form. Approval for this study was obtained from the
relevant ethical review boards of the participating institutions (116).

3. Ascertainment of outcomes

Cancer cases were identified during follow-up based on population cancer registries in
Denmark, Italy, Netherlands, Norway, Spain, Sweden, and the United Kingdom, and on a
combination of methods, including health insurance records, contacts with cancer and
pathology registries, and active follow-up of EPIC participants and their next of kin in France,
Germany, and Greece. Mortality data were collected from, either the cancer or mortality
registries at the regional or national level.

The most recent vital status and cancer diagnosis are known for 98.4% of all EPIC subjects,
while 1.6% of participants emigrated, withdrew or were lost to follow-up. The current follow-
up period ended as follows: December 2009 in Varese and Murcia, December 2010 in
Florence, Ragusa, Turin, Asturias, Bilthoven and Utrecht, December 2011 in Granada,
Navarra, San Sebastian and Cambridge, December 2012 in Oxford, Umeå, Norway and
Denmark, and December 2013 in Malmö. The end of follow-up was considered as the last
known contact with participants in France (June 2008), Heidelberg and Potsdam (December
2009), and Naples (December 2010) and Greece (December 2012).

In the present work, cases of PC were identified as primary incident tumor of the pancreas.
They were coded according to the International Classification of Diseases (10th edition), and
included all invasive PC (C25.0–C25.3, C25.7–C25.9). Because of they have different
prognosis, endocrine and neuroendocrine tumors of the pancreas (C25.4), as well as carcinoid
tumors (8240/3) and lymphoma (9591/3) were censored at date of diagnosis (n=54). As result,
there was 1,283 PC cases identified within a median follow-up of 15 years in the EPIC study.
Microscopically confirmed PC represented 83% of the cases (n=928) based on histology of
the primary tumor or metastases, cytology or autopsy reports.

43
4. Diet and lifestyle assessment

Information on habitual diet and alcohol intakes over the year preceding recruitment was
assessed at recruitment by validated centre-/country- specific dietary questionnaires designed
to capture local-dietary habits with high compliance (116,117). To calibrate dietary
measurements, a standardized computer-assisted 24-hour dietary recall was implemented at a
center level on stratified random samples of the participants, for a total of 36,900 subjects and
used as a reference method to correct for systematic between-centre over- or underestimations
in the baseline dietary assessments (118).

Anthropometric measurements were performed (self-reported in France, Norway and the UK

Oxford centre) (119,120) to collect information on height, weight, waist circumference and
hip circumference. Occupational and physical activities, education, prevalent chronic
conditions (hypertension, diabetes, cardio-vascular diseases), lifetime history of tobacco
smoking, lifetime alcohol intake and menstrual and reproductive history (including the use of
exogenous hormones for contraception and post-menopausal replacement therapy) were
collected through lifestyle questionnaires.

Follow-up measures of diet and lifestyle exposures have been collected in about 70% of the
centers. The centralization and harmonization are currently ongoing at IARC and data should
be available in 2020.

5. Biobank, blood samples collection

Blood samples were collected at baseline from 387 889 individuals (88%). Plasma, serum,
leukocytes, and erythrocytes have been separated and aliquoted for long-term storage in liquid
nitrogen at IARC and mirrored at EPIC collaborating centers. Overall, the EPIC
biorepositories host more than 9 million aliquots, constituting one of the largest biobanks in
the world for biochemical and genetic investigations on cancer and other chronic diseases.
that can integrate questionnaire data on lifestyle and diet, biomarkers of diet and of
endogenous metabolism (e.g. hormones and growth factors) and genetic polymorphisms.

44
45
 CHAPTER III:
ALCOHOL CONSUMPTION AND
THE RISK OF PANCREATIC CANCER

46
47
1. Background

PC is one of the most fatal cancers worldwide (17,26). Its etiology has been extensively
researched, leading to the identification of tobacco smoking, body fatness, type-II diabetes
mellitus and chronic pancreatitis as major risk factors (23,40,46,98). Although alcohol
consumption has been recognized as a type 1 carcinogen by IARC Monograph, the evidence
for an association between alcohol use and the risk of PC has been considered as suggestive
and limited by international experts panels (46,77). Studies examining the association of PC
risk with smoking status, type of alcoholic beverages and drinking history have been limited
by the small number of cases in prospective studies as PC remains a particularly rare cancer
(17).

Previous prospective studies yielded evidence for an excess risk of PC limited to high levels
of alcohol consumption (81,121–123). Recent meta analyses have shown that alcohol intake
increased the risk of PC by at least 15% in drinkers consuming more than 25 g/day compared
with light drinkers with alcohol intake below 12g/day (79,80). The vast majority of these
investigations has primarily focused on baseline alcohol intake, though two early analyses
from the EPIC study indicated that neither baseline nor cumulative lifetime alcohol intake
were related to PC risk (124,125). In addition, heavy drinking is a risk factor for chronic
pancreatitis, a condition posited to be involved in the underlying mechanisms of PC (126). It
has been hypothesized that the relationship between alcohol and PC could be entirely
explained by occurrence of chronic pancreatitis (127).

In this study, the association between alcohol intake and the risk of PC was comprehensively
investigated using information from the EPIC study and involving a larger number of incident
PC cases than earlier evaluations.

2. Objectives

- To investigate the role of past alcohol consumption examining lifetime alcohol intake
in relation to PC risk.
- To evaluate the potential differences in the relationship according to the type of
alcoholic beverages consumed.
- To investigate the role of smoking habits in the relationship between alcohol
consumption and PC risk, especially in non-smokers.

48
 - To evaluate the potential role of chronic pancreatitis in the relationship between
alcohol intake and PC risk.

3. Methods

This study was carried out using information from the EPIC cohort (128). After 15 years of
follow-up, 1,283 incident PC cases were diagnosed from 521,324 cancer-free participants.

Baseline alcohol intake was assessed through dietary questionnaire (116,117) and computed
from the number of glasses of beer and/or cider, wine, sweet liquors and/or distilled spirits
and fortified wines drunk per day or week during the 12 months preceding recruitment.
Lifetime alcohol consumption was measured as the number of glasses from the different types
of beverages consumed per week at 20, 30, 40 and 50 years of age. It was assessed from a
lifestyle questionnaire and was available for 76% of study participants corresponding to 966
PC cases over 363,265 participants.

Cox proportional hazard models with age as primary time variable were used to estimate PC
hazard ratios (HR) and 95% confidence intervals (CI). Separate models were run by gender to
account for the behavioural differences in alcohol intake between men and women. Baseline
and lifetime alcohol intakes were modeled as continuous variables, with HR estimates for a 12
g/day increase in alcohol intake (corresponding to a standard drink of beer, spirits/liquors, or
wine) and in categories (<0.1 g/day, 0.1–4.9, 5–14.9, 15–29.9, 30–59.9 and >60), using the
group of light drinkers (0.1-4.9 g/day) as reference. Dose–response analyses were performed
for baseline and lifetime alcohol intake in men and in all participants together. Potential
departures from linearity in the associations were examined by fitting restricted cubic spline
models (129). Effect modification in the relationship between alcohol and PC risk by smoking
status (never, current smokers) was evaluated by comparing models with and without
interaction terms.

In the absence of information on chronic pancreatitis in EPIC, a sensitivity analysis was

carried out to account for the potential role of chronic pancreatitis (Z) between baseline
alcohol intake (X) and risk of PC (D) using information from the literature (130). In EPIC,
HR estimate for the risk of PC in baseline heavy drinkers (>60g/day) compared to light
drinkers (0.1-4.9g/day) was equal to 1.64 (95% CI: 1.22, 2.21) in men and women combined.
PC HRs for heavy drinkers vs. light drinkers considering information on chronic pancreatitis
(HRDX) were estimated assuming values from the literature for pancreatitis prevalence among

49
moderate drinkers ranging from 0.005 to 0.02 (131), relative risk estimates of chronic
pancreatitis in relation to alcohol intake greater than 25 g/day compared to the never drinkers
(RRZX) ranging from 2 to 6 (131,132), and relative risk estimates of PC associated with
chronic pancreatitis (RRDZ) ranging from 1.5 to 15 (94,95,126).

4. Results

In men, baseline alcohol intake was significantly associated with PC risk, with a 5% (95% CI:
1%, 9%) increased PC risk for every increment of 12 g/day of alcohol. HR for men heavy
drinkers with more than 60 g/day of alcohol was 1.63 (95%CI: 1.16, 2.29). No evidence for an
association between alcohol and PC risk was found in women. Similar observations were
found for lifetime alcohol intake. In men, a 12 g/day increase of lifetime alcohol intake was
related to a 6% (2%, 10%) increased PC risk, with a HR for heavy drinkers (>60 g/day) equal
to 1.77 (1.06, 2.95) when compared to the reference category. No statistically significant
association between lifetime alcohol intake and PC risk was found in women. In evaluations
by types of alcoholic beverages, beer consumption was positively associated with a 9% (2%,
15%) and a 22% (3%, 44%) increase in PC risk for 12 g/day in men and women, respectively.
In men, spirits/liquors were associated with a 17% (4%, 32%) higher PC risk for a 12 g/day
increase, while no association was observed in women. Wine intake was not associated with
PC risk irrespective of sex. Similar results were observed for lifetime alcohol intake from the
different beverages and PC risk.

Evaluation of the dose response relationship between baseline and lifetime alcohol intake and
PC risk in men suggested a linear shaped association, without evidence for departure from
linearity either for baseline (pnon-linearity=0.83) or lifetime intakes (pnon-linearity=0.57). Similar
results were found when examining all participants, with p-values for non-linearity equal to
0.90 and 0.75 for baseline and lifetime intakes, respectively (Figure 8).

Although alcohol intake was not associated with PC risk among never smokers (HR per 12
g/day increase: 1.06; 95% CI: 0.98, 1.15), current smokers had a HR equal to 1.05 (1.00,
1.11). There was no evidence for heterogeneity between alcohol and PC risk with respect to
smoking statuses (pheterog= 0.84).

50
Figure 8. Hazard Ratio (HR) functions (and 95% CI) describing the curvilinear dose–
response relationship between baseline and lifetime alcohol intake (g/day) and PC risk,
according to the frequency of cases in all participants.

The sensitivity analysis using information from the litterature about history of chronic
pancreatitis indicated that the HR estimate comparing baseline heavy drinkers (>60 g/day) vs.
light drinkers (0.1-4.9 g/day) was marginally attenuated when considering realistic scenarios
with a prevalence of chronic pancreatitis in moderate drinkers of 0.01, with a relative risk
estimate for the relation between alcohol and chronic pancreatitis as large as 4, and a PC
relative risk associated with chronic pancreatitis not exceeding 10. Larger attenuations of HR
estimates were observed for more extreme scenarios showing weaker positive HRDX
estimated (Figure 9). Findings suggested that alcohol intake may exert its carcinogenic role
only partially through chronic pancreatitis. Studies with individual data on history of
pancreatitis may allow for a more advanced mechanistic approach to estimate the proportion
of alcohol-PC association mediated by pancreatitis.

51
Figure 9. Sensitivity analysis using external information from the scientific literature on
history of chronic pancreatitis (Z), in the relationship between baseline alcohol intake (X) and
the risk of pancreatic cancer (D).

5. Conclusions

In summary, our study showed a moderate but statistically significant increase in PC risk with
high alcohol intake, either baseline or lifetime, and particularly for beer and spirits/ liquors.
These findings provide epidemiologic evidence for the role of alcohol consumption as a
potential carcinogen of the pancreas.

6. Scientific article

“Lifetime and baseline alcohol intakes and risk of pancreatic cancer in the European
Prospective Investigation into Cancer and Nutrition study”, Naudin S et al., International
Journal of Cancer, Published, March 2018.

Roles of the PhD student: 1st author, participated to the analytical strategy, conducted all
analyses, wrote the manuscript, revised and submitted it.

52
53
54
55
56
57
58
59
60
61
62
63
64
65
 CHAPTER IV:
HEALTHY LIFESTYLE INDEX AND
THE RISK OF PANCREATIC CANCER

66
67
1. Background

Commonly diagnosed at late stages, PC is a highly fatal cancer with limited opportunities for
effective treatment (17,35). In the current absence of available screening tools, the
identification of modifiable risk factors may offer opportunities for PC primary prevention.
Examinations of lifestyle risk factors with respect to PC incidence in prospective studies have
led to the identification of smoking, body fatness, adult attained height, type-2 diabetes, and
heavy alcohol drinking as positive risk factors, while diet and physical activity have been
inconsistently associated with PC risk (40,46). Although lifestyle habits often cluster together
(133), there is limited evidence regarding the joint association of different lifestyle factors on
PC incidence, especially among European populations.

A multi-component score termed the Healthy Lifestyle Index (HLI), combining information
on smoking, alcohol intake, dietary habits, body mass index (BMI), and physical activity has
been previously related to several cancers sites in the European Prospective Investigation into
Cancer and Nutrition (EPIC) study (134–137). Within the American Association of Retired
Persons (AARP) cohort a strong inverse association was observed between the HLI and PC
risk (138). So far, evidence for an inverse association in European population has been
limited to an evaluation from an Italian case-control study (139).

2. Objectives

- To examine the relationship between the HLI and PC risk within the EPIC study using
two versions of the score: (i) with BMI to reflect overall adiposity and (ii) with waist-
to-hip ratio to reflect central adiposity.

- To investigate the marginal role of each single lifestyle factors of the HLI score in
association with PC risk, particularly for smoking.

- To estimate population attributable fractions based on counterfactual and realistic

scenari where study participants adopted healthier behaviors.

- To examine the pattern of the association between the HLI and PC incidence over
age.

68
3. Methods

This study was carried out using information from the EPIC cohort. After 15 years of follow-
up, 1,283 incident PC cases were diagnosed from 521,324 cancer-free participants. The HLI
score was generated from the combination of information on smoking history, alcohol intake,
a dietary score, BMI, and physical activity, with larger values indicating adherence to
healthier habits (137). Since previous evidence on PC etiology identified waist-to-hip ratio
(WHR), an indicator of central adiposity, as a PC risk factor (50,140), a HLI based on the
WHR (HLIWHR) was computed replacing BMI categories with sex-specific WHR quintiles
(Figure 10). Subjects with missing information on smoking status (n=15,684) and physical
activity (n=65,054) were excluded. For analyses with HLIWHR, subjects with missing WHR
were also excluded (n=45,105). The final study population included 400,577 participants of
which 1,113 PC cases in analyses with HLIBMI, and 1,075 cases in analyses with HLIWHR.

Figure 10. Scoring system combining the 5 lifestyle factors into the Heathy Lifestyle Index
based on the waist-to-hip ratio (HLIWHR).
1
For the HLI based on the BMI (HLIBMI), sex-specific waist-to-hip ratio quintiles were
replaced by categories of BMI at baseline using the following cut-offs: (4) 22–23.9 kg.m-2,
(3) 24–25.9 kg.m-2, (2) <22 kg.m-2, (1) 26–29.9 kg.m-2, and (0) >30 kg.m-2

69
Cox proportional hazard models with age as primary time variable were used to estimate PC
hazard ratios (HR) and their 95% confidence intervals (CI). The HLIBMI and HLIWHR were, in
turn, modeled as continuous variables computing HR estimates for a one-standard deviation
(1-SD) increase in the HLI score corresponding to approximatively 3 units, and in categories
(0-4, 5-9, 10-14, 15-20), using the group 5-9 as reference. All models were stratified by study
center, sex and age at recruitment. Models were adjusted for education level, diabetes status,
energy intake from non-alcohol sources, and height was undertaken, and additionally for BMI
in models with HLIWHR.

Sensitivity analyses were carried out by excluding, in turn, each factor from the HLI scores to
identify which factors are driving the HLI association with PC risk. The excluded component
was used as a confounder in the model.

Population attributable fractions (PAF) were estimated as the reduction in PC incidence that
would occur (i) if study participants shifted to the adjacent healthier category of HLIWHR, and
(ii) assuming a counterfactual scenario whereby men adopted women’s lifestyle habits (141).
Confidence intervals were obtained using bootstrap sampling (142).

The pattern of HR for a 1-SD increase of HLIWHR by age was examined using a flexible
parametric survival model on the cumulative hazard scale. Restricted cubic splines with 5
internal knots were used to model the baseline hazard using attained age as the time scale and
a time-varying coefficient on HLIWHR (143).

4. Results

HR estimates for PC risk given for a one-standard deviation increment of HLIBMI and
HLIWHR were equal to 0.84 (95% CI: 0.79, 0.89; ptrend=4.3e-09) and 0.77 (0.72, 0.82;
ptrend=1.7e-15), respectively. Exclusions of, in turn, smoking, alcohol and WHR from HLIWHR
resulted in HRs of 0.88 (0.82, 0.94; ptrend=4.9e-04), 0.79 (0.74, 0.84; ptrend=7.0e-13), and 0.79
(0.74, 0.85; ptrend=3.2e-11), respectively. PAF based on counterfactual scenari where
participants adopted healthier behaviors indicated that controlling obesity, alcohol
consumption, having a healthy diet and improving physical activity could prevent no less
than 14% (95% CI: 6%, 21%) of PC cases, and up to 19% (11%, 26%) if quitting smoking
was also considered. Interestingly, 13% (9%, 16%) of PC cases would be prevented if men
adopted lifestyle habits of women. Examination of PC HR estimates pattern for a 1-SD of
HLIWHR showed weaker associations at older ages.

70
5. Conclusions

In this large European prospective study, healthy lifestyle habits expressed in a HLI score
were strongly inversely related to the risk of PC. Adherence to healthy behaviors
corresponding to a three-point increase in the score was associated with a 16% (95%CI: 11%,
21%) lower PC risk for a score that included BMI, and 23% (18%, 28%) lower PC risk for a
score including WHR. Our findings provided evidence that adherence to healthy lifestyle
habits could be an effective primary prevention strategy to control the incidence of PC.

6. Scientific article

“The Healthy lifestyle and the risk of pancreatic cancer in the EPIC study”, Naudin et al.,
European Journal of Epidemiology, Published, September 2019.

Roles of the PhD student: 1st author, participated to the analytical strategy, conducted all
analyses, wrote the manuscript, revised and submitted it.

71
72
73
74
75
76
77
78
79
80
81
82
83
 CHAPTER V:
ADDITIONAL EPIDEMIOLOGICAL
RESEARCH EVALUATIONS

84
85
1. The Pooling Project on Alcohol and Cancer

As part of this PhD program, I had the opportunity to take part to the Pooling Project on
Alcohol and Cancer (PPAC), a large ongoing study aiming at the evaluation of the association
between alcohol consumption and cancer sites for which the relationship with alcohol is not
established within prospective studies participating to the National Cancer Institute (NCI)
Cohort Consortium. This project was initiated as a collaborative project coordinated by IARC
(Principal investigators (PI): Drs. Pietro Ferrari and Paul Brennan) and the Harvard School of
Public Health (Co-PI: Dr. Stephanie Smith-Warner). In 2016, the project was successfully
granted by the National Institute on Alcohol Abuse and Alcoholism (NI-AAA). The following
part will provide an overview of the ongoing work with this project and what was
accomplished as part of the PhD program.

1.1. Background and significance

In 2012, the IARC Monograph program defined alcohol consumption as causally related to
the incidence of several cancer sites including cancers of the upper aero-digestive tract
(UADT), liver, colorectal and female breast cancers, as confirmed by the WCRF/AICR
experts’ panels in 2012 (77,144). Whereas the association of these specific cancer sites with
alcohol use has been consistently observed, the evidence of the link with other cancers,
primarily prostate and pancreatic, but also kidney, thyroid cancer and non-Hodgkin
lymphoma (NHL) was either inconsistent or sparse (79). Despite the importance in human
carcinogenesis, several research questions on alcohol and cancer remain far to be answered.

The relationship between alcohol use and cancer sites with an established link with alcohol
was recently confirmed in the EPIC study with respect to cancer incidence (145) and overall
and cancer-specific mortality (146), both for baseline and lifetime alcohol uses (147).
Lifetime alcohol use was also significantly associated with mortality from the ensemble of
non-alcohol related cancer sites (147), suggesting a potential role of alcohol for cancer sites
other than those evaluated by the IARC Monograph program (77). It has been observed more
recently in pooled studies and meta-analyses that alcohol use was positively associated with
risk of PC (84) and prostate cancer (79). Also there is suggestive evidence for an inverse
association between alcohol use and the risk of cancer of the kidney (148), thyroid (79), and
NHL (149,150). However, scientific evidence on these associations is still very inconsistent
and unclear regarding plausible biological mechanisms involved.

86
The evaluation of the relationship between alcohol and cancer presents several analytical
challenges. First, if a relationship between alcohol and any of these cancer sites existed, it
would likely be weak; thus, requiring a large sample size to have sufficient statistical power to
detect associations. Second, many of the cancer sites investigated in the PPAC are fairly rare
cancers, in particular for PC and NHL (in women) and thyroid cancer (in men). Any
epidemiological investigation would require information from large study population, with a
long follow-up in order to develop a sufficiently large number of cancer cases. This scenario
is likely to materialize through a collaborative effort to pool data from ongoing large
prospective epidemiological investigations, as in the PPAC consortium.

1.2. Objectives

The global aim of the PPAC project is to conduct a pooled analysis of 36 prospective studies
on the role of alcohol use and the risk of cancer sites for which the evidence is suggestive or
limited, namely: prostate, pancreas, UADT, kidney, thyroid, and NHL. The evaluation will
particularly focus on the following specific objectives:

- To evaluate the associations between alcohol consumption and cancers of the prostate,
pancreas, kidney and UADT in non-smokers

- To evaluate associations of lifetime alcohol use (assessed in 7 studies) with risk of

advanced and localized prostate, pancreatic, kidney, and thyroid cancer, and NHL

- To evaluate specific patterns of alcohol use, including binge drinking (assessed in 11

studies), in relation to risk of advanced and localized prostate, pancreas, kidney, and
thyroid cancer, and NHL.

This project will produce solid evidence on the role of alcohol on the aforementioned cancer
sites for which the relationship with alcohol is not established, thus allowing an update of the
burden of cancer attributable to alcohol use in different populations across the world.

1.3. Study design

Individual-level data from 36 cohort studies spanning 4 continents (American, Europe, Asia
and Oceania) including over 2.7 million women and men were centralized and will be
evaluated in the coming year. The Table 1 provides the list of the studies involved in the
project. Information about alcohol intake (including past consumption, intake at recruitment,

87
and intake during follow-up), dietary intake and lifestyle covariates is collected through food
frequency questionnaires, lifestyle questionnaires and questionnaire during follow-up within
each individual study is centralized at Harvard. The follow-up of participants will start at the
date of completion of the baseline questionnaire for the exposure of interest and last until the
date of death, date of lost-to-follow-up, or administrative end date, whichever occurred first.
Cancer outcomes considered in this project are primary invasive cancers of the prostate, the
pancreas, the kidney, the thyroid, UADT cancers and NHL. At the beginning of the study it
was expected to reach over 161,000 incident cases with 4,800 thyroid cancer; 9,900 kidney
cancer, and more than 10,000 each for pancreatic, UADT and prostate cancer, and NHL.

1.4. Data centralization and harmonization

Centralization and harmonization of primary participant level data originating from the 36
studies involved in the project was initiated in 2017, and are still ongoing. These steps
benefits from expertise and data already collected by Harvard collaborators few years ago as
part of a previous long-standing international consortium called the Pooling Project of
Prospective Studies of Diet and Cancer (DCPP) which included 26 cohorts.

Data centralization started with the implementation of a list of variables that were requested to
all participating studies of the PPAC. For this purpose, I developed the dictionary providing
definitions for variables on mortality, end of follow-up, cancer outcomes, covariates collected
at baseline and during follow-up (including lifestyle, dietary and medical history), and on
alcohol exposure (including information about alcohol intakes at baseline, during lifetime and
during follow-up) (Annex 1). The principal investigators of each cohort were contacted to
require a list of variables in August 2017. The 26 cohorts already involved in the DCPP
received a shorten list focusing on variables on alcohol consumption and on update of cancer
outcomes. Data were centralized by Harvard scientists, and were remotely accessible from
IARC using a secured server.

A second part of my contribution was to harmonize information on of cancer outcomes across

participating cohorts. The task started during the summer 2018 based on data already
centralized at that time. Under the supervision of an IARC pathologist, cancer site-specific
criteria of inclusion and exclusion were defined. Cancer sites considered in this project were
cancer of the prostate (C61), the pancreas (C25.0, C25.1, C25.2, C25.3, C25.7, C25.8, C25.9),
the kidney (C64), the thyroid (C73) and UADT (C00, C01, C02, C03, C04, C05, C06, C09,

88
C10, C11, C13, C14, C15.0, C15.1, C15.3, C15.8, C15.9) based on the International
Classification of Diseases (ICD) 10th edition (Table 2). Criteria were refined at the
histological and morphological level using the ICD-Oncology (ICD-O) 3rd edition coding
system. The exhaustive list of histological and morphological type is provided in Table 2. As
the classification for NHL outcomes has changed several time over the last decades due to
their biologically and clinically changing definition (151,152), the participating cohorts might
have used different versions of the ICD-O classification to define the histology of their NHL
cases. As a result, I produced a separate table summarizing for each subtype of lymphoma the
corresponding histology codes providing equivalence between the different IDC-O versions
(Annex 1).

The harmonization of the data collected on alcohol intake is currently ongoing (Table 1).
Next steps of the project will be to analyze participant-level data from each cohort using
uniform criteria across studies and control for key covariates (e.g., smoking habits) to provide
the most complete and reliable information for the evaluation of these associations between
alcohol use and the targeted cancer- sites.

89
Table 1. Descriptive characteristics, alcohol consumption and smoking status of PPAC participating cohorts.

90
Table 2. Criteria for cancer outcomes harmonization in the PPAC project.

91
2. Lifestyle factors in association with the risk of lymphoma

While the various steps of the PPAC were implemented, including data centralization and
harmonization, interest emerged on the examination within the EPIC study of the relationship
between alcohol and risk of several cancer sites that were the target of the PPAC. The
association between alcohol intake and lymphoma had not been investigated in EPIC yet,
mainly due the limited number of lymphoma cases, particularly by subtypes. Statistical
analyses were therefore undertaken as part of this PhD program. Results showed weak
associations for baseline alcohol intake, with HR for non-Hodgkin lymphoma (NHL) and for
Hodgkin lymphoma (HL) equal to 1.00 (95% CI: 0.98, 1.03) and 0.85 (0.74, 0.99),
respectively, for a 12 g/day increase in alcohol intake. Similar risk patterns were observed for
lifetime alcohol intake, overall and by lymphoma subtypes. The investigation was also
extended to a larger list of lifestyle factors, through the HLI score, and the risk of lymphoma
in EPIC.

2.1. Background

Lymphoma is a heterogeneous group of malignancies occurring in the lymphatic system with

biologically and clinically diverse definitions depending on immune cells involved in the
carcinogenic process. Traditionally, they are grouped in Hodgkin (HL) and non-Hodgkin
lymphoma (NHL). NHL is divided into two major subtypes, mature T and natural killer-cell
lymphoma (MT/NK) and mature B-cell lymphoma (BCL), with BCL subdivided by diffuse
large B-cell lymphoma (DLBCL), follicular lymphoma (FL), chronic lymphocytic leukemia
and small lymphocytic leukemia (CLL/SLL), and plasma cell neoplasm and multiple
myeloma (PCN/MM) (152). In 2018, lymphoma accounted for 3.2% of incident cancers
worldwide, and their incidence steadily increased over the past decades, particularly in high-
incomes countries (17,153). Additionally, a large part of the increase in incidence is still
poorly understood and the role of lifestyle exposure in lymphomagenesis remains particularly
unclear (154). Although individual lifestyle factors have been related to the risk of lymphoma
(149,155), there is limited evidence regarding the joint association of lifestyle factors on the
risk of lymphoma. In this work, the association between adherence to healthy lifestyles and
risks of HL and NHL was evaluated in a large European prospective study.

92
 2.2. Methods

Within the European Prospective investigation into Cancer and Nutrition (EPIC) study 2,659
incident lymphoma cases, including 120 HL and 2,446 NHL, were diagnosed from 400,622
cancer-free participants after 14 years of follow-up. The healthy lifestyle index (HLI) score
combined information on the history of smoking, alcohol intake, diet, the level of physical
activity and BMI, with large values of HLI expressing adherence to healthy behaviors. Cox
proportional hazards models with age as primary time variable were used to estimate
lymphoma hazard ratios (HR) and their 95% confidence interval (CI). Models were stratified
by sex, country and age at recruitment, and adjusted for education level, height, and energy
intake from non-alcohol sources. The HLI was modeled in continuous for a 1-SD increase in
the HLI (around 3 units) and in quartiles using the second quartile as the reference category.
Sensitivity analyses were conducted by excluding, in turn, each lifestyle factor from the HLI
score.

2.3. Results

Our results showed an inverse association between the HLI and the risk of HL, with HR for a
1-SD increment in the score equal to 0.81 (95%CI: 0.68, 0.97). Sensitivity analyses showed
that the association was mainly driven by smoking, and marginally by diet and BMI.

Overall lymphoma and NHL were not associated with HLI, with HRs for a 1-SD increment
equal to 0.99 (0.96, 1.03) and 1.00 (0.96, 1.04), respectively.

There was no association in evaluations by NHL subtypes. HR estimates comparing the fourth
to the second quartile were equal to 0.93 (95%CI: 0.81, 1.06; ptrend=0.40) for BCL,1.31 (1.00,
1.74; ptrend=0.25) for DLBCL, 1.01 (0.90, 1.13; ptrend=0.62) for FL, 1.02 (0.93, 1.12;
ptrend=0.62) for CLL/SLL, and 0.98 (0.90, 1.06; ptrend=0.61) for PCN/MM. Although, the EPIC
study benefits from a large sample size, this evaluation was possibly underpowered to detect
likely weak associations in analysis by subtypes of NHL.

2.4. Conclusions

In the EPIC study, adherence to healthy lifestyle was not associated with overall lymphoma or
NHL risks. An inverse association was observed for HL, although it was largely attributable
to smoking. Overall, these findings suggested a limited role of lifestyle factors in the etiology

93
of lymphoma subtypes. However, this study is one of the first attempts to investigate the risk
of lymphoma with respect to healthy lifestyle habits, and results should be replicated in other
studies.

2.5. Short communication

“Healthy lifestyle and the risk of lymphoma in the EPIC study”, Naudin et al., Under revision,
International Journal of Cancer, November 2019.

Roles of the PhD student: 1st author, participated to the analytical strategy, conducted all
analyses, wrote the manuscript, and revised it.

94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
 CHAPTER VI:
GENERAL CONCLUSION

120
121
1. Main findings

Commonly diagnosed at late stage, PC is a highly fatal cancer with limited opportunities for
early detection and effective treatment. In this context, the identification of modifiable risk
factors may contribute to PC prevention. This doctoral research program intended to enhance
the understanding of PC etiology, focusing on the role of alcohol consumption and more
globally on lifestyle determinants.

Our overall findings showed a moderate but significant positive association with high alcohol
intake. Evaluation of the dose-response relationship suggested a linear alcohol- PC
association. Consistent results were observed using information on alcohol intake at baseline
and alcohol consumed during the life course. Examination by type of alcoholic beverages
showed positive relationships with beer and spirits/liquors, and a weaker association with
wine. Although the relationship was weaker in never than in current smokers and no
association was observed in women, there was no statistically significant evidence for
heterogeneity by smoking status or by gender, potentially pointing to limitations regarding
subgroup analyses in the evaluation of rare disease. A sensitivity analysis accounting for the
role of chronic pancreatitis showed evidence for a moderate mediating role of chronic
pancreatitis. If alcohol is causally related to PC risk, it is likely to only partly involve
pancreatitis mechanism.

The second evaluation undertaken during this PhD resulted in strong evidence for an inverse
relationship between healthy lifestyle habits and PC occurrence, using a score combining
information on smoking history, alcohol intake, obesity, diet and physical activity. The
evaluation suggested a marginally stronger association with a score considering information
on central adiposity rather than overall adiposity. Our results pointed to associations being
marginally stronger at younger ages. Estimated population attributable fractions suggested
that controlling obesity, alcohol consumption, diet quality and increasing the level of physical
activity could prevent 14% of PC cases, and up to 19% if quitting smoking was also
considered. Interestingly, 13% of PC cases could be prevented if men adopted lifestyle habits
of women. Overall, although a large proportion of PC cases is still unexplained, these findings
provide evidence that realistic changes in common lifestyle habits could sizably decrease the
risk of PC.

122
2. Perspectives

With respect to the work undertaken during this PhD program, future research aiming at
improving the understanding of PC etiology through investigation of alcohol consumption
and lifestyle factors should consider the following challenges.

While our findings strengthen the epidemiological evidence for the relationship between
alcohol use and PC risk, future research steps should focus on the clarification of the
association among specific populations. In the current work, the evaluation was hampered by
the low number of cases in nonsmokers and women, who are also less likely to be heavy
drinkers. As PC remains of low incidence, it is likely that the collective effort of prospective
investigations will be required to provide a sufficiently large number of cancer cases to
conduct an accurate investigation of the association among these subgroups. Future
evaluations should benefit from the strong framework provided by consortia of prospective
studies within the NCI cohort consortium, and in particular the Pooling Project on Alcohol
and Cancer.

Since extreme alcohol intake is a cause of chronic pancreatitis, it has been suggested that
alcohol intake could increase PC risk through this pathway. Findings from our first
evaluation, based on a sensitivity analysis using information on the link between pancreatitis
and PC risk extracted from the scientific literature, suggested that the relationship between
alcohol and PC could be moderately attenuated. Although there is a strong association
between individuals with chronic pancreatitis and the risk of PC compared with the general
population (95), the proportion of PC arising from patients with chronic pancreatitis is very
low (131). The use of mediation analysis in studies where the information on the history of
pancreatitis was collected could clarify to what extent the alcohol-PC association is mediated
by chronic pancreatitis and provide a better understanding about whether the carcinogenic
role of alcohol on PC is dependent from the pancreatitis mechanism.

Our second evaluation provided strong evidence for an inverse relationship between healthy
lifestyle habits and the risk of PC. Notably, we estimated that 13% of PC cases could have
been avoided if men adopted similar lifestyle habits as women. This proportion partially
explains the difference in PC incidence in men (Europe ASR=5.5 per 100,000 PY) compared
with women (Europe ASR=4.0 per 100,000 PY), although other elements are to be considered
to explain the remaining difference. Our study used a healthy lifestyle score implemented on
the basis of a priori general knowledge on the relationship between lifestyle factors and risk
of chronic diseases, and was not specific to PC. Yet, it has the advantage of producing
123
evidence comparable across different cancer sites. Given the difference in smoking habits in
men and women, the use of weights to combine the various components of the score may
have provided results more specific to PC. Also, the score could have included other factors
related to cancer, e.g. information on infectious diseases and morbid conditions (type-2
diabetes, hypertension), which might have contributed to explain more accurately the gender
difference. Lastly, as pancreatic carcinogenesis has a genetic origin, information on genetic
profiles could also have provided key elements to elucidate the gap between men and women.

Our findings on lifestyle habits were based on European populations and yielded similar
evidence on the strong relationship between lifestyle habits and PC as in an American
investigation in the NIH-AARP cohort (138). This evidence needs to be confirmed in other
populations, possibly characterized by diverse lifestyle habits in different geographical
contexts.

Socio-economic status is believed to play a major role in the occurrence of PC (156,157) and
some of the components of the HLI may act as mediators of this association (158). The
relationships observed in our study between healthy lifestyle habits and the risk of PC could
possibly provide a partial overview of the causal pathway between socio-economic status and
PC risk. Additionally, other lifestyle factors not included in the HLI could be on the same
pathway. As a result, socio-economic status is an important confounder of the HLI-PC
association, and our analyses were consistently adjusted for participants’ level of education,
although the use of education as a proxy for socio-economic status may have produced
residual confounding.

Our studies considered lifestyle and dietary information collected at recruitment among adult
participants. However, within 15 years of follow-up, individuals may have modified their
lifestyle habits, possibly leading to different findings. Thanks to the collaboration between the
EPIC participating centers and IARC, data on lifestyle and dietary factors collected during
follow-up will soon be available. Examination of serial questionnaire information collected on
alcohol intake and lifestyle factors will enable future studies to consider potential changes in
alcohol intake and lifestyle habits in relation to the risk of PC, and further complete the work
initiated during this PhD.

Furthermore, information on potential mechanisms related to the role of alcohol and a

combination of lifestyle factors on pancreatic carcinogenesis are far from being identified.
Metabolomics data could be instrumental for the investigation of pathological processes and

124
could notably enhance the understanding of the mechanisms of alcohol–related pancreatic
carcinogenesis (159).

This comprehensive evaluation provides informative insights on the role of alcohol

consumption and healthy lifestyle habits in the etiology of PC. It strongly supports the
development of effective primary prevention strategy and the implementation of public health
guidelines to promote the adoption of healthy lifestyle habits for the sake of PC prevention.

125
 REFERENCES

126
1. Longnecker DS. Anatomy and Histology of the Pancreas. Pancreapedia Exocrine
Pancreas Knowl Base [Internet]. 2014 Mar 20 [cited 2019 Apr 11]; Available from:
/reviews/anatomy-and-histology-of-pancreas

2. Pancreas | anatomy [Internet]. Encyclopedia Britannica. [cited 2019 Apr 11]. Available
from: https://www.britannica.com/science/pancreas

3. Hans-Olov Adami, David Hunter, Dimitrios Trichopoulos. Textbook of Cancer

Epidemiology [Internet]. Second edition. Oxford University Press; 2008 [cited 2018 Jul
17]. Available from:
http://www.oxfordscholarship.com/view/10.1093/acprof:oso/9780195311174.001.0001/ac
prof-9780195311174

4. Cancer Research UK. Types of pancreatic cancer | Cancer Research UK [Internet]. Cancer
Research UK. 2019 [cited 2019 May 5]. Available from:
https://www.cancerresearchuk.org/about-cancer/pancreatic-cancer/stages-types-
grades/types

5. Brosens LAA, Hackeng WM, Offerhaus GJ, Hruban RH, Wood LD. Pancreatic
adenocarcinoma pathology: changing “landscape.” J Gastrointest Oncol [Internet]. 2015
Aug [cited 2019 Apr 11];6(4):358–74. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4502163/

6. Chhoda A, Lu L, Clerkin BM, Risch H, Farrell JJ. Current Approaches to Pancreatic

Cancer Screening. Am J Pathol [Internet]. 2019 Jan 1 [cited 2019 May 5];189(1):22–35.
Available from: http://www.sciencedirect.com/science/article/pii/S0002944018302529

7. McGuigan A, Kelly P, Turkington RC, Jones C, Coleman HG, McCain RS. Pancreatic
cancer: A review of clinical diagnosis, epidemiology, treatment and outcomes. World J
Gastroenterol [Internet]. 2018 Nov 21 [cited 2019 Apr 10];24(43):4846–61. Available
from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6250924/

8. Hruban RH, Takaori K, Klimstra DS, Adsay NV, Albores-Saavedra J, Biankin AV, et al.
An illustrated consensus on the classification of pancreatic intraepithelial neoplasia and
intraductal papillary mucinous neoplasms. Am J Surg Pathol. 2004 Aug;28(8):977–87.

9. Maitra A, Adsay NV, Argani P, Iacobuzio-Donahue C, De Marzo A, Cameron JL, et al.

Multicomponent analysis of the pancreatic adenocarcinoma progression model using a
pancreatic intraepithelial neoplasia tissue microarray. Mod Pathol Off J U S Can Acad
Pathol Inc. 2003 Sep;16(9):902–12.

10. Feldmann G, Beaty R, Hruban RH, Maitra A. Molecular genetics of pancreatic

intraepithelial neoplasia. J Hepatobiliary Pancreat Surg [Internet]. 2007 [cited 2019 Apr
10];14(3):224–32. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2666331/

11. Bosman F.T. CF, Hruban R.H. TND. WHO Classification of Tumours of the Digestive
System. 4th edition. Lyon: International Agency for Research on Cancer (IARC); 2010.

12. Chari ST, Kelly K, Hollingsworth MA, Thayer SP, Ahlquist DA, Andersen DK, et al.
Early Detection of Sporadic Pancreatic Cancer. Pancreas [Internet]. 2015 Jul [cited 2019

127
 May 5];44(5):693–712. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4467589/

13. Ricardo V. Lloyd, Robert Y. Osamura, Günter Klöppel, Juan Rosai. WHO Classification
of Tumours of Endocrine organs. 4th edition. Lyon, France: International Agency for
Research on Cancer; 2017.

14. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2017. CA Cancer J Clin [Internet]. 2017
[cited 2019 Apr 10];67(1):7–30. Available from:
http://onlinelibrary.wiley.com/doi/abs/10.3322/caac.21387

15. Rawla P, Sunkara T, Gaduputi V. Epidemiology of Pancreatic Cancer: Global Trends,

Etiology and Risk Factors. World J Oncol [Internet]. 2019 Feb [cited 2019 Apr
10];10(1):10–27. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6396775/

16. Conroy T, Desseigne F, Ychou M, Bouché O, Guimbaud R, Bécouarn Y, et al.

FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med. 2011
May 12;364(19):1817–25.

17. 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 [Internet]. 2018 Nov 1 [cited 2019 Apr 4];68(6):394–424.
Available from: http://onlinelibrary.wiley.com/doi/full/10.3322/caac.21492

18. Wong MCS, Jiang JY, Liang M, Fang Y, Yeung MS, Sung JJY. Global temporal patterns
of pancreatic cancer and association with socioeconomic development. Sci Rep [Internet].
2017 Jun 9 [cited 2019 Apr 10];7. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5466634/

19. Gordon-Dseagu VL, Devesa SS, Goggins M, Stolzenberg-Solomon R. Pancreatic cancer

incidence trends: evidence from the Surveillance, Epidemiology and End Results (SEER)
population-based data. Int J Epidemiol [Internet]. 2018 Apr [cited 2019 Apr
10];47(2):427–39. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5913617/

20. Wahi MM, Shah N, Schrock CE, Rosemurgy AS, Goldin SB. Reproductive factors and
risk of pancreatic cancer in women: a review of the literature. Ann Epidemiol. 2009
Feb;19(2):103–11.

21. Lujan-Barroso L, Zhang W, Olson SH, Gao Y-T, Yu H, Baghurst PA, et al. Menstrual
and Reproductive Factors, Hormone Use, and Risk of Pancreatic Cancer: Analysis From
the International Pancreatic Cancer Case-Control Consortium (PanC4). Pancreas.
2016;45(10):1401–10.

22. Midha S, Chawla S, Garg PK. Modifiable and non-modifiable risk factors for pancreatic
cancer: A review. Cancer Lett [Internet]. 2016 Oct 10 [cited 2019 Apr 10];381(1):269–77.
Available from: http://www.sciencedirect.com/science/article/pii/S0304383516304396

23. Ilic M, Ilic I. Epidemiology of pancreatic cancer. World J Gastroenterol [Internet]. 2016
Nov 28 [cited 2017 Sep 11];22(44):9694–705. Available from:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC5124974/

128
24. Yu J, Blackford AL, Molin M dal, Wolfgang CL, Goggins M. Time to progression of
pancreatic ductal adenocarcinoma from low-to-high tumour stages. Gut [Internet]. 2015
Nov 1 [cited 2019 Apr 17];64(11):1783–9. Available from:
https://gut.bmj.com/content/64/11/1783

25. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, et al. Cancer

incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN
2012. Int J Cancer. 2015 Mar 1;136(5):E359-386.

26. Lepage C, Capocaccia R, Hackl M, Lemmens V, Molina E, Pierannunzio D, et al.

Survival in patients with primary liver cancer, gallbladder and extrahepatic biliary tract
cancer and pancreatic cancer in Europe 1999–2007: Results of EUROCARE-5. Eur J
Cancer [Internet]. 2015 Oct 1 [cited 2017 Sep 6];51(15):2169–78. Available from:
http://www.sciencedirect.com/science/article/pii/S0959804915007145

27. Cancer Reseach UK. Pancreatic cancer survival statistics [Internet]. Cancer Research UK.
2015 [cited 2019 Apr 28]. Available from: https://www.cancerresearchuk.org/health-
professional/cancer-statistics/statistics-by-cancer-type/pancreatic-cancer/survival

28. Luo G, Zhang Y, Guo P, Ji H, Xiao Y, Li K. Global Patterns and Trends in Pancreatic
Cancer Incidence: Age, Period, and Birth Cohort Analysis. Pancreas. 2019;48(2):199–
208.

29. Wu W, He X, Yang L, Wang Q, Bian X, Ye J, et al. Rising trends in pancreatic cancer

incidence and mortality in 2000–2014. Clin Epidemiol [Internet]. 2018 Jul 9 [cited 2019
Apr 10];10:789–97. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6042490/

30. Lucas AL, Malvezzi M, Carioli G, Negri E, La Vecchia C, Boffetta P, et al. Global
Trends in Pancreatic Cancer Mortality From 1980 Through 2013 and Predictions for
2017. Clin Gastroenterol Hepatol Off Clin Pract J Am Gastroenterol Assoc.
2016;14(10):1452-1462.e4.

31. Fernandez E, La Vecchia C, Porta M, Negri E, Lucchini F, Levi F. Trends in pancreatic

cancer mortality in Europe, 1955-1989. Int J Cancer. 1994 Jun 15;57(6):786–92.

32. Ferlay J, Ervik M, Lam F, Colombet M, Mery L, Pineros M, et al. Global Cancer
Observatory: Cancer Tomorrow [Internet]. 2018. Available from:
https://gco.iarc.fr/tomorrow

33. Ferlay J, Partensky C, Bray F. More deaths from pancreatic cancer than breast cancer in
the EU by 2017. Acta Oncol Stockh Swed. 2016 Oct;55(9–10):1158–60.

34. Rahib L, Smith BD, Aizenberg R, Rosenzweig AB, Fleshman JM, Matrisian LM.
Projecting cancer incidence and deaths to 2030: the unexpected burden of thyroid, liver,
and pancreas cancers in the United States. Cancer Res. 2014 Jun 1;74(11):2913–21.

35. Lindquist CM, Miller FH, Hammond NA, Nikolaidis P. Pancreatic cancer screening.
Abdom Radiol N Y. 2018;43(2):264–72.

36. Peters MLB, Eckel A, Mueller PP, Tramontano AC, Weaver DT, Lietz A, et al.
Progression to pancreatic ductal adenocarcinoma from pancreatic intraepithelial

129
 neoplasia: Results of a simulation model. Pancreatology [Internet]. 2018 Dec 1 [cited
2019 Apr 23];18(8):928–34. Available from:
http://www.sciencedirect.com/science/article/pii/S1424390318306367

37. Yachida S, Jones S, Bozic I, Antal T, Leary R, Fu B, et al. Distant metastasis occurs late
during the genetic evolution of pancreatic cancer. Nature. 2010 Oct 28;467(7319):1114–7.

38. Khan MA, Azim S, Zubair H, Bhardwaj A, Patel GK, Khushman M, et al. Molecular
Drivers of Pancreatic Cancer Pathogenesis: Looking Inward to Move Forward. Int J Mol
Sci [Internet]. 2017 Apr 6 [cited 2019 Apr 10];18(4). Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5412363/

39. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Tobacco
smoke and involuntary smoking. IARC Monogr Eval Carcinog Risks Hum. 2004;83:1–
1438.

40. Maisonneuve P, Lowenfels AB. Risk factors for pancreatic cancer: a summary review of
meta-analytical studies. Int J Epidemiol. 2015 Feb;44(1):186–98.

41. Iodice S, Gandini S, Maisonneuve P, Lowenfels AB. Tobacco and the risk of pancreatic
cancer: a review and meta-analysis. Langenbecks Arch Surg. 2008 Jul;393(4):535–45.

42. Zou L, Zhong R, Shen N, Chen W, Zhu B, Ke J, et al. Non-linear dose-response

relationship between cigarette smoking and pancreatic cancer risk: evidence from a meta-
analysis of 42 observational studies. Eur J Cancer Oxf Engl 1990. 2014 Jan;50(1):193–
203.

43. Duell EJ. Epidemiology and potential mechanisms of tobacco smoking and heavy alcohol
consumption in pancreatic cancer. Mol Carcinog [Internet]. 2012 Jan 1 [cited 2017 Jun
16];51(1):40–52. Available from: http://onlinelibrary.wiley.com.docelec.univ-
lyon1.fr/doi/10.1002/mc.20786/abstract

44. Ma J, Jemal A. The rise and fall of cancer mortality in the USA: why does pancreatic
cancer not follow the trend? Future Oncol [Internet]. 2013 Jul 1 [cited 2019 May
21];9(7):917–9. Available from: https://www.futuremedicine.com/doi/10.2217/fon.13.76

45. IARC. Absence of Excess Body Fatness / IARC Working Group on the Evaluation of
Cancer-Preventive Interventions [Internet]. Vol. 16. IARC Handbooks of Cancer
Prevention; 2018. 658 p. Available from: http://publications.iarc.fr/570

46. World Cancer Research Fund/American Institute for Cancer Research. Continuous
Update Project Expert Report 2018. Diet, nutrition, physical activity and pancreatic
cancer. www.dietandcancerreport.org; 2018.

47. Stolzenberg-Solomon RZ, Schairer C, Moore S, Hollenbeck A, Silverman DT. Lifetime

adiposity and risk of pancreatic cancer in the NIH-AARP Diet and Health Study
cohort123. Am J Clin Nutr [Internet]. 2013 Oct [cited 2018 Mar 23];98(4):1057–65.
Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3778860/

48. Genkinger JM, Spiegelman D, Anderson KE, Bernstein L, van den Brandt PA, Calle EE,
et al. A POOLED ANALYSIS OF 14 COHORT STUDIES OF ANTHROPOMETRIC
FACTORS AND PANCREATIC CANCER RISK. Int J Cancer J Int Cancer [Internet].

130
 2011 Oct 1 [cited 2018 Oct 18];129(7):1708–17. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3073156/

49. Stolzenberg-Solomon RZ, Adams K, Leitzmann M, Schairer C, Michaud DS, Hollenbeck

A, et al. Adiposity, physical activity, and pancreatic cancer in the National Institutes of
Health-AARP Diet and Health Cohort. Am J Epidemiol. 2008 Mar 1;167(5):586–97.

50. González AB de, Spencer EA, Bueno-de-Mesquita HB, Roddam A, Stolzenberg-Solomon

R, Halkjær J, et al. Anthropometry, Physical Activity, and the Risk of Pancreatic Cancer
in the European Prospective Investigation into Cancer and Nutrition. Cancer Epidemiol
Prev Biomark [Internet]. 2006 May 1 [cited 2018 Mar 27];15(5):879–85. Available from:
http://cebp.aacrjournals.org/content/15/5/879

51. Luo J, Iwasaki M, Inoue M, Sasazuki S, Otani T, Ye W, et al. Body mass index, physical
activity and the risk of pancreatic cancer in relation to smoking status and history of
diabetes: a large-scale population-based cohort study in Japan–The JPHC study. Cancer
Causes Control [Internet]. 2007 Aug 1 [cited 2018 Apr 26];18(6):603–12. Available from:
https://link.springer.com/article/10.1007/s10552-007-9002-z

52. Genkinger JM, Kitahara CM, Bernstein L, Berrington de Gonzalez A, Brotzman M, Elena
JW, et al. Central adiposity, obesity during early adulthood, and pancreatic cancer
mortality in a pooled analysis of cohort studies. Ann Oncol [Internet]. 2015 Nov [cited
2018 Mar 23];26(11):2257–66. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4621029/

53. Suganami T, Tanaka M, Ogawa Y. Adipose tissue inflammation and ectopic lipid
accumulation. Endocr J. 2012;59(10):849–57.

54. Henry SL, Bensley JG, Wood-Bradley RJ, Cullen-McEwen LA, Bertram JF, Armitage
JA. White adipocytes: more than just fat depots. Int J Biochem Cell Biol. 2012
Mar;44(3):435–40.

55. Fritz J, Bjørge T, Nagel G, Manjer J, Engeland A, Häggström C, et al. The triglyceride-
glucose index as a measure of insulin resistance and risk of obesity-related cancers. Int J
Epidemiol. 2019 Mar 3;

56. Bao Y, Giovannucci EL, Kraft P, Qian ZR, Wu C, Ogino S, et al. Inflammatory plasma
markers and pancreatic cancer risk: a prospective study of five U.S. cohorts. Cancer
Epidemiol Biomark Prev Publ Am Assoc Cancer Res Cosponsored Am Soc Prev Oncol.
2013 May;22(5):855–61.

57. Grote VA, Kaaks R, Nieters A, Tjønneland A, Halkjær J, Overvad K, et al. Inflammation
marker and risk of pancreatic cancer: a nested case-control study within the EPIC cohort.
Br J Cancer. 2012 May 22;106(11):1866–74.

58. Mann S, Beedie C, Balducci S, Zanuso S, Allgrove J, Bertiato F, et al. Changes in insulin
sensitivity in response to different modalities of exercise: a review of the evidence.
Diabetes Metab Res Rev. 2014 May;30(4):257–68.

59. Cash SW, Beresford SA, Vaughan TL, Heagerty PJ, Bernstein L, White E, et al. Recent
physical activity in relation to DNA damage and repair using the comet assay. J Phys Act
Health. 2014 May;11(4):770–6.

131
60. Herder C, Peltonen M, Koenig W, Sütfels K, Lindström J, Martin S, et al. Anti-
inflammatory effect of lifestyle changes in the Finnish Diabetes Prevention Study.
Diabetologia. 2009 Mar;52(3):433–42.

61. Norat T, Aune D, Vieira AR, Chan D, Rosenblatt DN, Vieira R. The Associations
between Food, Nutrition and Physical Activity and the Risk of Pancreatic Cancer.
2011;470.

62. Behrens G, Jochem C, Schmid D, Keimling M, Ricci C, Leitzmann MF. Physical activity
and risk of pancreatic cancer: a systematic review and meta-analysis. Eur J Epidemiol.
2015 Apr;30(4):279–98.

63. Wu L, Zheng W, Xiang Y-B, Gao Y-T, Li H-L, Cai H, et al. Physical Activity and
Pancreatic Cancer Risk among Urban Chinese: Results from Two Prospective Cohort
Studies. Cancer Epidemiol Prev Biomark [Internet]. 2018 Apr 1 [cited 2018 Apr
4];27(4):479–87. Available from: http://cebp.aacrjournals.org/content/27/4/479

64. Taunk P, Hecht E, Stolzenberg-Solomon R. Are meat and heme iron intake associated
with pancreatic cancer? Results from the NIH-AARP Diet and Health Cohort. Int J
Cancer J Int Cancer [Internet]. 2016 May 1 [cited 2018 Mar 27];138(9):2172–89.
Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4764390/

65. Matejcic M, Lesueur F, Biessy C, Renault AL, Mebirouk N, Yammine S, et al.

Circulating plasma phospholipid fatty acids and risk of pancreatic cancer in a large
European cohort. Int J Cancer. 2018 15;143(10):2437–48.

66. Lei Q, Zheng H, Bi J, Wang X, Jiang T, Gao X, et al. Whole Grain Intake Reduces
Pancreatic Cancer Risk. Medicine (Baltimore) [Internet]. 2016 Mar 7 [cited 2019 May
26];95(9). Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4782845/

67. Jiao L, Chen L, White DL, Tinker L, Chlebowski RT, Van Horn LV, et al. Low-fat
Dietary Pattern and Pancreatic Cancer Risk in the Women’s Health Initiative Dietary
Modification Randomized Controlled Trial. JNCI J Natl Cancer Inst [Internet]. 2017 Aug
18 [cited 2019 May 26];110(1):49–56. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6059156/

68. Zheng J, Guinter MA, Merchant AT, Wirth MD, Zhang J, Stolzenberg-Solomon RZ, et al.
Dietary patterns and risk of pancreatic cancer: a systematic review. Nutr Rev [Internet].
2017 Nov [cited 2019 May 26];75(11):883–908. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5914454/

69. Molina-Montes E, Sánchez M-J, Buckland G, Bueno-de-Mesquita HB, Weiderpass E,

Amiano P, et al. Mediterranean diet and risk of pancreatic cancer in the European
Prospective Investigation into Cancer and Nutrition cohort. Br J Cancer [Internet]. 2017
Mar 14 [cited 2017 Jun 5];116(6):811–20. Available from:
https://www.nature.com/bjc/journal/v116/n6/full/bjc201714a.html

70. Zheng Jiali, Merchant Anwar T., Wirth Michael D., Zhang Jiajia, Antwi Samuel O.,
Shoaibi Azza, et al. Inflammatory potential of diet and risk of pancreatic cancer in the
Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. Int J Cancer
[Internet]. 2018 Jan 22 [cited 2018 Mar 22];0(0). Available from: https://onlinelibrary-
wiley-com.docelec.univ-lyon1.fr/doi/abs/10.1002/ijc.31271
132
71. Zheng J, Wirth MD, Merchant AT, Zhang J, Shivappa N, Stolzenberg-Solomon RZ, et al.
Inflammatory potential of diet, inflammation-related lifestyle factors and risk of
pancreatic cancer: Results from the NIH-AARP Diet and Health Study. Cancer Epidemiol
Biomark Prev Publ Am Assoc Cancer Res Cosponsored Am Soc Prev Oncol. 2019 Apr
30;

72. Ezzati M, Riboli E. Behavioral and dietary risk factors for noncommunicable diseases. N
Engl J Med. 2013 Sep 5;369(10):954–64.

73. Seitz HK, Stickel F. Molecular mechanisms of alcohol-mediated carcinogenesis. Nat Rev
Cancer [Internet]. 2007 Aug [cited 2019 May 20];7(8):599–612. Available from:
http://www.nature.com/articles/nrc2191

74. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Alcohol
consumption and ethyl carbamate. IARC Monogr Eval Carcinog Risks Hum. 2010;96:3–
1383.

75. Romieu I, Scoccianti C, Chajès V, de Batlle J, Biessy C, Dossus L, et al. Alcohol intake
and breast cancer in the European prospective investigation into cancer and nutrition. Int J
Cancer. 2015 Oct 15;137(8):1921–30.

76. Seitz HK, Becker P. Alcohol metabolism and cancer risk. Alcohol Res Health J Natl Inst
Alcohol Abuse Alcohol. 2007;30(1):38–41, 44–7.

77. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Personal
habits and indoor combustions. Volume 100 E. A review of human carcinogens. IARC
Monogr Eval Carcinog Risks Hum World Health Organ Int Agency Res Cancer.
2012;100(Pt E):1–538.

78. Cogliano VJ, Baan R, Straif K, Grosse Y, Lauby-Secretan B, El Ghissassi F, et al.

Preventable exposures associated with human cancers. J Natl Cancer Inst. 2011 Dec
21;103(24):1827–39.

79. Bagnardi V, Rota M, Botteri E, Tramacere I, Islami F, Fedirko V, et al. Alcohol

consumption and site-specific cancer risk: a comprehensive dose–response meta-analysis.
Br J Cancer [Internet]. 2015 Feb 3 [cited 2018 Mar 19];112(3):580–93. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4453639/

80. Wang Y-T, Gou Y-W, Jin W-W, Xiao M, Fang H-Y. Association between alcohol intake
and the risk of pancreatic cancer: a dose–response meta-analysis of cohort studies. BMC
Cancer [Internet]. 2016 Mar 12 [cited 2018 Mar 19];16. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4788838/

81. Tramacere I, Scotti L, Jenab M, Bagnardi V, Bellocco R, Rota M, et al. Alcohol drinking
and pancreatic cancer risk: a meta-analysis of the dose-risk relation. Int J Cancer. 2010
Mar 15;126(6):1474–86.

82. Lucenteforte E, La Vecchia C, Silverman D, Petersen GM, Bracci PM, Ji BT, et al.
Alcohol consumption and pancreatic cancer: a pooled analysis in the International
Pancreatic Cancer Case–Control Consortium (PanC4). Ann Oncol [Internet]. 2012 Feb
[cited 2018 Apr 26];23(2):374–82. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3265544/

133
83. Genkinger JM, Spiegelman D, Anderson KE, Bergkvist L, Bernstein L, van den Brandt
PA, et al. Alcohol intake and pancreatic cancer risk: a pooled analysis of fourteen cohort
studies. Cancer Epidemiol Biomark Prev Publ Am Assoc Cancer Res Cosponsored Am
Soc Prev Oncol. 2009 Mar;18(3):765–76.

84. Michaud DS, Vrieling A, Jiao L, Mendelsohn JB, Steplowski E, Lynch SM, et al. Alcohol
intake and pancreatic cancer: a pooled analysis from the pancreatic cancer cohort
consortium (PanScan). Cancer Causes Control [Internet]. 2010 Aug 1 [cited 2018 Apr
26];21(8):1213–25. Available from: https://link.springer.com/article/10.1007/s10552-010-
9548-z

85. Huxley R, Ansary-Moghaddam A, Berrington de González A, Barzi F, Woodward M.

Type-II diabetes and pancreatic cancer: a meta-analysis of 36 studies. Br J Cancer. 2005
Jun 6;92(11):2076–83.

86. Stevens RJ, Roddam AW, Beral V. Pancreatic cancer in type 1 and young-onset diabetes:
systematic review and meta-analysis. Br J Cancer. 2007 Feb 12;96(3):507–9.

87. Muñoz AR, Chakravarthy D, Gong J, Halff GA, Ghosh R, Kumar AP. Pancreatic cancer:
Current status and Challenges. Curr Pharmacol Rep [Internet]. 2017 Dec [cited 2019 Apr
10];3(6):396–408. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5795623/

88. Risch HA. Diabetes and Pancreatic Cancer: Both Cause and Effect. J Natl Cancer Inst.
2019 Jan 1;111(1):1–2.

89. Song S, Wang B, Zhang X, Hao L, Hu X, Li Z, et al. Long-Term Diabetes Mellitus Is

Associated with an Increased Risk of Pancreatic Cancer: A Meta-Analysis. PloS One.
2015;10(7):e0134321.

90. Batabyal P, Vander Hoorn S, Christophi C, Nikfarjam M. Association of diabetes mellitus

and pancreatic adenocarcinoma: a meta-analysis of 88 studies. Ann Surg Oncol. 2014
Jul;21(7):2453–62.

91. Bosetti C, Rosato V, Li D, Silverman D, Petersen GM, Bracci PM, et al. Diabetes,
antidiabetic medications, and pancreatic cancer risk: an analysis from the International
Pancreatic Cancer Case-Control Consortium. Ann Oncol Off J Eur Soc Med Oncol. 2014
Oct;25(10):2065–72.

92. Kolodecik T, Shugrue C, Ashat M, Thrower EC. Risk factors for pancreatic cancer:
underlying mechanisms and potential targets. Front Physiol [Internet]. 2014 Jan 16 [cited
2017 Jun 30];4. Available from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3893685/

93. Mizuno S, Nakai Y, Isayama H, Yanai A, Takahara N, Miyabayashi K, et al. Risk factors
and early signs of pancreatic cancer in diabetes: screening strategy based on diabetes
onset age. J Gastroenterol. 2013 Feb;48(2):238–46.

94. Duell EJ, Lucenteforte E, Olson SH, Bracci PM, Li D, Risch HA, et al. Pancreatitis and
pancreatic cancer risk: a pooled analysis in the International Pancreatic Cancer Case-
Control Consortium (PanC4). Ann Oncol Off J Eur Soc Med Oncol ESMO. 2012
Nov;23(11):2964–70.

134
95. Raimondi S, Lowenfels AB, Morselli-Labate AM, Maisonneuve P, Pezzilli R. Pancreatic
cancer in chronic pancreatitis; aetiology, incidence, and early detection. Best Pract Res
Clin Gastroenterol. 2010 Jun;24(3):349–58.

96. Yadav D, Lowenfels AB. The Epidemiology of Pancreatitis and Pancreatic Cancer.
Gastroenterology [Internet]. 2013 Jun [cited 2016 Nov 15];144(6):1252–61. Available
from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3662544/

97. Risch HA, Yu H, Lu L, Kidd MS. ABO blood group, Helicobacter pylori seropositivity,
and risk of pancreatic cancer: a case-control study. J Natl Cancer Inst. 2010 Apr
7;102(7):502–5.

98. Barone E, Corrado A, Gemignani F, Landi S. Environmental risk factors for pancreatic
cancer: an update. Arch Toxicol. 2016 Nov;90(11):2617–42.

99. Wang Y, Yang S, Song F, Cao S, Yin X, Xie J, et al. Hepatitis B virus status and the risk
of pancreatic cancer: a meta-analysis. Eur J Cancer Prev [Internet]. 2013 Jul 1 [cited 2019
Jun 20];22(4):328–34. Available from: insights.ovid.com

100. Luo G, Hao N-B, Hu C-J, Yong X, Lü M-H, Cheng B-J, et al. HBV infection
increases the risk of pancreatic cancer: a meta-analysis. Cancer Causes Control CCC.
2013 Mar;24(3):529–37.

101. Fiorino S, Lorenzini S, Masetti M, Deleonardi G, Grondona AG, Silvestri T, et al.

Hepatitis B and C virus infections as possible risk factor for pancreatic adenocarcinoma.
Med Hypotheses. 2012 Nov;79(5):678–97.

102. Xu J-H, Fu J-J, Wang X-L, Zhu J-Y, Ye X-H, Chen S-D. Hepatitis B or C viral
infection and risk of pancreatic cancer: a meta-analysis of observational studies. World J
Gastroenterol. 2013 Jul 14;19(26):4234–41.

103. Hruban RH, Canto M, Goggins M, Schulick R, Klein AP. Update on Familial
Pancreatic Cancer. Adv Surg [Internet]. 2010 [cited 2019 May 5];44:293–311. Available
from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2966038/

104. Flores K, Dinh K, Rouleau E, Whalen G, Wassef W, LaFemina J. Identification of

genetic risk for pancreatic adenocarcinoma. Cancer Genet. 2015 Nov;208(11):559–63.

105. Greer JB, Whitcomb DC, Brand RE. Genetic Predisposition to Pancreatic Cancer: A
Brief Review. Am J Gastroenterol [Internet]. 2007 Nov [cited 2018 Apr 5];102(11):2564–
9. Available from: https://www.nature.com/articles/ajg2007495

106. Lowenfels AB, Maisonneuve P, Lankisch PG. Chronic pancreatitis and other risk
factors for pancreatic cancer. Gastroenterol Clin North Am. 1999 Sep;28(3):673–85, x.

107. Chen F, Roberts NJ, Klein AP. Inherited pancreatic cancer. Chin Clin Oncol. 2017
Dec;6(6):58.

108. Del Chiaro M, Segersvärd R, Lohr M, Verbeke C. Early detection and prevention of
pancreatic cancer: is it really possible today? World J Gastroenterol. 2014 Sep
14;20(34):12118–31.

135
109. Lowenfels AB, Maisonneuve P, Whitcomb DC, Lerch MM, DiMagno EP. Cigarette
smoking as a risk factor for pancreatic cancer in patients with hereditary pancreatitis.
JAMA. 2001 Jul 11;286(2):169–70.

110. Amundadottir LT. Pancreatic Cancer Genetics. Int J Biol Sci. 2016;12(3):314–25.

111. Bailey P, Chang DK, Nones K, Johns AL, Patch A-M, Gingras M-C, et al. Genomic
analyses identify molecular subtypes of pancreatic cancer. Nature [Internet]. 2016 Mar
[cited 2019 Jun 17];531(7592):47–52. Available from:
https://www.nature.com/articles/nature16965

112. Jacobs EJ, Chanock SJ, Fuchs CS, Lacroix A, McWilliams RR, Steplowski E, et al.
Family history of cancer and risk of pancreatic cancer: a pooled analysis from the
Pancreatic Cancer Cohort Consortium (PanScan). Int J Cancer. 2010 Sep 1;127(6):1421–
8.

113. Risch HA, Lu L, Wang J, Zhang W, Ni Q, Gao Y-T, et al. ABO blood group and risk
of pancreatic cancer: a study in Shanghai and meta-analysis. Am J Epidemiol. 2013 Jun
15;177(12):1326–37.

114. Wolpin BM, Kraft P, Gross M, Helzlsouer K, Bueno-de-Mesquita HB, Steplowski E,

et al. Pancreatic cancer risk and ABO blood group alleles: results from the pancreatic
cancer cohort consortium. Cancer Res. 2010 Feb 1;70(3):1015–23.

115. Hakomori S. Antigen structure and genetic basis of histo-blood groups A, B and O:
their changes associated with human cancer. Biochim Biophys Acta. 1999 Dec
6;1473(1):247–66.

116. Riboli E, Hunt KJ, Slimani N, Ferrari P, Norat T, Fahey M, et al. European
Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data
collection. Public Health Nutr. 2002 Dec;5(6B):1113–24.

117. Kaaks R, Slimani N, Riboli E. Pilot phase studies on the accuracy of dietary intake
measurements in the EPIC project: overall evaluation of results. European Prospective
Investigation into Cancer and Nutrition. Int J Epidemiol. 1997;26 Suppl 1:S26-36.

118. Kaaks R, Riboli E. Validation and calibration of dietary intake measurements in the
EPIC project: methodological considerations. European Prospective Investigation into
Cancer and Nutrition. Int J Epidemiol. 1997;26 Suppl 1:S15-25.

119. Spencer EA, Appleby PN, Davey GK, Key TJ. Validity of self-reported height and
weight in 4808 EPIC-Oxford participants. Public Health Nutr. 2002 Aug;5(4):561–5.

120. Spencer EA, Roddam AW, Key TJ. Accuracy of self-reported waist and hip
measurements in 4492 EPIC-Oxford participants. Public Health Nutr. 2004 Sep;7(6):723–
7.

121. Heinen MM, Verhage BAJ, Ambergen TAW, Goldbohm RA, van den Brandt PA.
Alcohol consumption and risk of pancreatic cancer in the Netherlands cohort study. Am J
Epidemiol. 2009 May 15;169(10):1233–42.

136
122. Jiao L, Silverman DT, Schairer C, Thiébaut ACM, Hollenbeck AR, Leitzmann MF, et
al. Alcohol use and risk of pancreatic cancer: the NIH-AARP Diet and Health Study. Am
J Epidemiol. 2009 May 1;169(9):1043–51.

123. Gapstur SM, Jacobs EJ, Deka A, McCullough ML, Patel AV, Thun MJ. Association
of Alcohol Intake With Pancreatic Cancer Mortality in Never Smokers. Arch Intern Med
[Internet]. 2011 Mar 14 [cited 2018 Mar 19];171(5):444–51. Available from:
https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/226816

124. Bueno de Mesquita HB, Maisonneuve P, Moerman CJ, Runia S, Boyle P. Lifetime
consumption of alcoholic beverages, tea and coffee and exocrine carcinoma of the
pancreas: a population-based case-control study in The Netherlands. Int J Cancer. 1992
Feb 20;50(4):514–22.

125. Rohrmann S, Linseisen J, Vrieling A, Boffetta P, Stolzenberg-Solomon RZ,

Lowenfels AB, et al. Ethanol intake and the risk of pancreatic cancer in the European
Prospective Investigation into Cancer and Nutrition (EPIC). Cancer Causes Control CCC.
2009 Jul;20(5):785–94.

126. Dhar P, Kalghatgi S, Saraf V. Pancreatic Cancer in Chronic Pancreatitis. Indian J Surg
Oncol [Internet]. 2015 Mar [cited 2017 Jun 30];6(1):57–62. Available from:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4412870/

127. Raimondi S, Maisonneuve P, Lowenfels AB. Epidemiology of pancreatic cancer: an

overview. Nat Rev Gastroenterol Hepatol. 2009 Dec;6(12):699–708.

128. Riboli E, Kaaks R. The EPIC Project: rationale and study design. European
Prospective Investigation into Cancer and Nutrition. Int J Epidemiol. 1997;26 Suppl 1:S6-
14.

129. Durrleman S, Simon R. Flexible regression models with cubic splines. Stat Med. 1989
May;8(5):551–61.

130. Greenland S. Basic Methods for Sensitivity Analysis of Biases. Int J Epidemiol
[Internet]. 1996 Dec 1 [cited 2017 Jan 27];25(6):1107–16. Available from:
https://academic.oup.com/ije/article/25/6/1107/672425/Basic-Methods-for-Sensitivity-
Analysis-of-Biases

131. Setiawan VW, Pandol SJ, Porcel J, Wilkens LR, Le Marchand L, Pike MC, et al.
Prospective Study of Alcohol Drinking, Smoking, and Pancreatitis: The Multiethnic
Cohort. Pancreas. 2016;45(6):819–25.

132. Samokhvalov AV, Rehm J, Roerecke M. Alcohol Consumption as a Risk Factor for
Acute and Chronic Pancreatitis: A Systematic Review and a Series of Meta-analyses.
EBioMedicine. 2015 Dec;2(12):1996–2002.

133. Schuit AJ, van Loon AJM, Tijhuis M, Ocké M. Clustering of lifestyle risk factors in a
general adult population. Prev Med. 2002 Sep;35(3):219–24.

134. Aleksandrova K, Pischon T, Jenab M, Bueno-de-Mesquita HB, Fedirko V, Norat T, et

al. Combined impact of healthy lifestyle factors on colorectal cancer: a large European
cohort study. BMC Med. 2014 Oct 10;12:168.

137
135. McKenzie F, Ferrari P, Freisling H, Chajès V, Rinaldi S, de Batlle J, et al. Healthy
lifestyle and risk of breast cancer among postmenopausal women in the European
Prospective Investigation into Cancer and Nutrition cohort study. Int J Cancer. 2015 Jun
1;136(11):2640–8.

136. Buckland G, Travier N, Huerta J m., Bueno-de-Mesquita HB, Siersema P d., Skeie G,
et al. Healthy lifestyle index and risk of gastric adenocarcinoma in the EPIC cohort study.
Int J Cancer [Internet]. 2015 Aug 1 [cited 2017 Aug 22];137(3):598–606. Available from:
http://onlinelibrary.wiley.com.docelec.univ-lyon1.fr/doi/10.1002/ijc.29411/abstract

137. McKenzie F, Biessy C, Ferrari P, Freisling H, Rinaldi S, Chajès V, et al. Healthy

Lifestyle and Risk of Cancer in the European Prospective Investigation Into Cancer and
Nutrition Cohort Study. Medicine (Baltimore). 2016 Apr 22;95(16).

138. Jiao L, Mitrou PN, Reedy J, Graubard BI, Hollenbeck AR, Schatzkin A, et al. A
Combined Healthy Lifestyle Score and Risk of Pancreatic Cancer - A Large Cohort
Study. Arch Intern Med [Internet]. 2009 Apr 27 [cited 2017 Dec 15];169(8):764–70.
Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3498842/

139. Lucas AL, Bravi F, Boffetta P, Polesel J, Serraino D, La Vecchia C, et al. Adherence
to World Cancer Research Fund/American Institute for Cancer Research
recommendations and pancreatic cancer risk. Cancer Epidemiol. 2016 Feb;40:15–21.

140. Janssen I, Heymsfield SB, Allison DB, Kotler DP, Ross R. Body mass index and waist
circumference independently contribute to the prediction of nonabdominal, abdominal
subcutaneous, and visceral fat. Am J Clin Nutr. 2002 Apr;75(4):683–8.

141. Barendregt JJ, Veerman JL. Categorical versus continuous risk factors and the
calculation of potential impact fractions. J Epidemiol Community Health [Internet]. 2010
Mar 1 [cited 2018 Mar 12];64(3):209–12. Available from:
http://jech.bmj.com/content/64/3/209

142. Efron B, Tibshirani R. Bootstrap Methods for Standard Errors, Confidence Intervals,
and Other Measures of Statistical Accuracy. Stat Sci [Internet]. 1986 [cited 2018 Sep
13];1(1):54–75. Available from: https://www.jstor.org/stable/2245500

143. Lambert PC, Royston P. Further Development of Flexible Parametric Models for
Survival Analysis. Stata J. 2009;Volume 9 Number 2.:21.

144. WCRF/AICR. World Cancer Research Fund/American Institute for Cancer Research.
Continuous Update Project Expert Report 2018. Alcoholic drinks and the risk of cancer.
[Internet]. WCRF/AICR; 2018. Available from: dietandcancerreport.org

145. Schütze M, Boeing H, Pischon T, Rehm J, Kehoe T, Gmel G, et al. Alcohol

attributable burden of incidence of cancer in eight European countries based on results
from prospective cohort study. BMJ [Internet]. 2011 Apr 7 [cited 2017 Jun
30];342:d1584. Available from: http://www.bmj.com/content/342/bmj.d1584

146. Bergmann MM, Rehm J, Klipstein-Grobusch K, Boeing H, Schütze M, Drogan D, et

al. The association of pattern of lifetime alcohol use and cause of death in the European
prospective investigation into cancer and nutrition (EPIC) study. Int J Epidemiol. 2013
Dec;42(6):1772–90.

138
147. Ferrari P, Licaj I, Muller DC, Kragh Andersen P, Johansson M, Boeing H, et al.
Lifetime alcohol use and overall and cause-specific mortality in the European Prospective
Investigation into Cancer and nutrition (EPIC) study. BMJ Open [Internet]. 2014 Jul 2
[cited 2017 Mar 6];4(7). Available from:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4091394/

148. Ljungberg B, Campbell SC, Choi HY, Cho HY, Jacqmin D, Lee JE, et al. The
epidemiology of renal cell carcinoma. Eur Urol. 2011 Oct;60(4):615–21.

149. Morton LM, Slager SL, Cerhan JR, Wang SS, Vajdic CM, Skibola CF, et al. Etiologic
Heterogeneity Among Non-Hodgkin Lymphoma Subtypes: The InterLymph Non-
Hodgkin Lymphoma Subtypes Project. JNCI Monogr [Internet]. 2014 Aug 1 [cited 2017
Jun 9];2014(48):130–44. Available from:
https://academic.oup.com/jncimono/article/2014/48/130/942634/Etiologic-Heterogeneity-
Among-Non-Hodgkin-Lymphoma

150. Tramacere I, Pelucchi C, Bonifazi M, Bagnardi V, Rota M, Bellocco R, et al. Alcohol

drinking and non-Hodgkin lymphoma risk: a systematic review and a meta-analysis. Ann
Oncol [Internet]. 2012 Nov 1 [cited 2017 Jun 5];23(11):2791–8. Available from:
https://academic.oup.com/annonc/article/23/11/2791/233973/Alcohol-drinking-and-non-
Hodgkin-lymphoma-risk-a

151. Swerdlow S, Campo E, Harris N, Jaffe E, Pileri S, Stein H, et al. WHO Classification
of Tumours of Haematopoietic and Lymphoid Tissues [Internet]. 2017 [cited 2018 Jul 9].
Available from: http://publications.iarc.fr/Book-And-Report-Series/Who-Iarc-
Classification-Of-Tumours/Who-Classification-Of-Tumours-Of-Haematopoietic-And-
Lymphoid-Tissues-2017

152. Turner JJ, Morton LM, Linet MS, Clarke CA, Kadin ME, Vajdic CM, et al.
InterLymph hierarchical classification of lymphoid neoplasms for epidemiologic research
based on the WHO classification (2008): update and future directions. Blood [Internet].
2010 Nov 18 [cited 2017 May 24];116(20):e90–8. Available from:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2993636/

153. Chihara D, Ito H, Matsuda T, Shibata A, Katsumi A, Nakamura S, et al. Differences in

incidence and trends of haematological malignancies in Japan and the United States. Br J
Haematol. 2014 Feb;164(4):536–45.

154. Bassig BA, Lan Q, Rothman N, Zhang Y, Zheng T. Current Understanding of

Lifestyle and Environmental Factors and Risk of Non-Hodgkin Lymphoma: An
Epidemiological Update. J Cancer Epidemiol [Internet]. 2012 [cited 2019 Feb 20];2012.
Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3447374/

155. World Cancer Research Fund / American Institute for Cancer Rsearch. Food,
nutrition, physical activity, and the prevention of cancer: a global perspective. [Internet].
2007 [cited 2016 May 12]. Available from:
http://www.aicr.org/assets/docs/pdf/reports/Second_Expert_Report.pdf

156. Bakens MJAM, Lemmens VEPP, de Hingh IHJT. Socio-economic status influences
the likelihood of undergoing surgical treatment for pancreatic cancer in the Netherlands.
HPB. 2017;19(5):443–8.

139
157. Bonaccio M, Castelnuovo AD, Costanzo S, Curtis AD, Persichillo M, Cerletti C, et al.
Socioeconomic trajectories across the life course and risk of total and cause-specific
mortality: prospective findings from the Moli-sani Study. J Epidemiol Community Health
[Internet]. 2019 Jun 1 [cited 2019 Jun 20];73(6):516–28. Available from:
https://jech.bmj.com/content/73/6/516

158. Akinyemiju T, Ogunsina K, Okwali M, Sakhuja S, Braithwaite D. Lifecourse

Socioeconomic Status and Cancer-Related Risk Factors: Analysis of the WHO study on
Global Ageing and Adult Health (SAGE). Int J Cancer [Internet]. 2017 Feb 15 [cited
2018 Oct 26];140(4):777–87. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5330617/

159. Jenab M, Slimani N, Bictash M, Ferrari P, Bingham SA. Biomarkers in nutritional

epidemiology: applications, needs and new horizons. Hum Genet [Internet]. 2009 Jun 1
[cited 2017 Jun 19];125(5–6):507–25. Available from:
https://link.springer.com/article/10.1007/s00439-009-0662-5

140
141
ANNEXES
 ANNEX 1
List of variables requested to cohorts participating to
the Pooling Pooling Project on Alcohol and Cancer as part of data centralization
 Variable List for Alcohol and Cancer Project

for New Studies in the Diet and Cancer Pooling Project

August, 2017

- Cohort name
- Study Center, if applicable
- Country
- Region of the country
- Participant ID code
- Personal history of cancer at baseline

Mortality and end of follow-up data

- Date of administrative end of follow-up

- Age at end of follow-up (years; please include decimals, if possible)
- Death status (yes, no)
- Date of death
- Age at death (years; please include decimals, if possible)
- ICD code for underlying cause of death
- ICD version
- Date of loss to follow-up
- Age lost to follow-up (years; please include decimals, if possible)

Cancer outcomes (only first primary, invasive incident tumors are of interest)

- Date of diagnosis of any cancer (if you cannot provide the date, please provide the year of cancer
diagnosis, if possible)
- Age at diagnosis of any cancer (years; please include decimals, if possible)
**This information is instrumental to censor study participants’ follow-up time in statistical analyses
on each cancer site detailed below.

Kidney cancer

Body-site ICD9Dx-CM codes ICD10Dx-CM codes

Kidney, except pelvis 189.0 C64.-

- Kidney cancer case status (yes, no)

- Date of diagnosis of kidney cancer (if you cannot provide the date, please provide the year of
kidney cancer diagnosis, if possible)
- Age at diagnosis of kidney cancer (years; please include decimals, if possible)
- Location of the tumor (ICD code)
- ICD version
- Stage of cancer (In situ, localized, metastatic, metastatic regional, metastatic distant, unknown)
- TNM classification
- Version of TNM classification

Annexe 1, page 146

- Behavior of the tumor (carcinoma in situ, malignant primary site, metastatic, benign)
- Morphology of the tumor
- Grade of the tumor (undifferentiated, poorly, moderately, well or not differentiated)

Pancreatic cancer

Body-site ICD9Dx-CM codes ICD10Dx-CM codes

Head of pancreas 157 C25.0

Body of pancreas 157.1 C25.1

Tail of pancreas 157.2 C25.2

Pancreatic duct 157.3 C25.3

Other specified sites of pancreas 157.8 C25.7

Other specified sites of pancreas 157.8 C25.8

Pancreas, part unspecified 157.9 C25.9

- Pancreatic cancer case status (yes, no)

- Date of diagnosis of pancreatic cancer (if you cannot provide the date, please provide the year of
pancreatic cancer diagnosis, if possible)
- Age at diagnosis of pancreatic cancer (years; please include decimals, if possible)
- Location of the tumor (ICD code)
- ICD version
- Stage of cancer (In situ, localized, metastatic, metastatic regional, metastatic distant, unknown)
- TNM classification
- Version of TNM classification
- Behavior of the tumor (carcinoma in situ, malignant primary site, metastatic, benign)
- Morphology of the tumor
- Grade of the tumor (undifferentiated, poorly, moderately, well or not differentiated)

Thyroid cancer

Body-site ICD9Dx-CM codes ICD10Dx-CM codes

Malignant neoplasm of thyroid gland 193 C73

- Thyroid cancer case status (yes, no)

- Date of diagnosis of thyroid cancer (if you cannot provide the date, please provide the year of
thyroid cancer diagnosis, if possible)
- Age at diagnosis of thyroid cancer (years; please include decimals, if possible)
- Location of the tumor (ICD code)
- ICD version
- Stage of cancer (In situ, localized, metastatic, metastatic regional, metastatic distant, unknown)
- TNM classification
- Version of TNM classification
- Behavior of the tumor (carcinoma in situ, malignant primary site, metastatic, benign)
- Morphology of the tumor
- Grade of the tumor (undifferentiated, poorly, moderately, well or not differentiated)

Annexe 1, page 147

- Methodology of diagnosis

Upper aero-digestive tract (UADT) cancer

Body-site ICD9Dx-CM codes ICD10Dx-CM codes

Malignant neoplasm of lip (inner aspect) 140.3, 140.4, 140.5 C00.3, C00.4, C00.5

C01.x, C02.0, C02.1, C02.2,

Malignant neoplasm of tongue 141 to 141.9
C02.3, C02.4, C02.8, C02.9

Malignant neoplasm of gum 143, 143.0, 143.1, 143.8, 143.9 C03, C03.0, C03.1, C03.9

C04, C04.0, C04.1, C04.8,

Malignant neoplasm of floor of mouth 144, 144.0, 144.1, 144.8, 144.9
C04.9

C05, C05.0, C05.1, C05.2,

Malignant neoplasm of other and unspecified
145 to 145.9 C05.8, C05.9, C06, C06.0,
parts of mouth
C06.1, C06.2, C06.8, C06.9

C10.0, C10.1, C10.2, C10.3,

Malignant neoplasm of oropharynx and tonsils 146 to 146.9 C10.4, C10.8, C10.9, C09.0,
C09.1, C09.8, C09.9

147, 147.0, 147.1, 147.2, 147.3, C11.0, C11.1, C11.2,C11.3,

Malignant neoplasm of nasopharynx
147.8, 147.9 C11.8, C11.9

148, 148.0, 148.1, 148.2, 148.3, C13.0, C13.1, C13.2, C13.8,

Malignant neoplasm of hypopharynx
148.8, 148.9 C13.9

Malignant neoplasm of other and ill-defined sites

149, 149.0, 149.1, 149.8, 149.9 C14, C14.0, C14.2, C14.8
within the lip oral cavity and pharynx

Malignant neoplasm of nasal cavities middle ear 160, 160.0, 160.1, 160.2, 160.3, C30.0, C30.1, C31.0, C31.1,
and accessory sinuses 160.4, 160.5, 160.8, 160.9 C31.2, C31.3, C31.9, C31.9

161, 161.0, 161.1, 161.2, 161.3, C32.0, C32.1, C32.2, C32.3,

Malignant neoplasm of larynx
161.8, 161.9 C32.8, C32.9

150.0, 150.1, 150.2, 150.3, 150.4, C15.0, C15.1, C15.2, C15.3,

Malignant neoplasm of oesophagus
150.5, 150.8, 150.9 C15.4, C15.5, C15.8, C15.9.

- UADT cancer case status (yes, no)

- Date of diagnosis of UADT cancer (if you cannot provide the date, please provide the year of
UADT cancer diagnosis, if possible)
- Age at diagnosis of UADT cancer (years; please include decimals, if possible)
- Upper aero digestive tract site (ICD code)
- ICD version
- Stage of cancer (In situ, localized, metastatic, metastatic regional, metastatic distant, unknown)
- TNM classification
- Version of TNM classification
- Behavior of the tumor (carcinoma in situ, malignant primary site, metastatic, benign)
- Morphology of the tumor

Annexe 1, page 148

- Grade of the tumor (undifferentiated, poorly, moderately, well or not differentiated)

Non-Hodgkin lymphoma (NHL)

Eligible codes, by coding scheme†

Site ICD9Dx-CM codes ICD10Dx-CM codes ICD-O-3 codes‡
Codes 202.0, 202.1, 202.2, 9590 to 9597,
Non-Hodgkin Code C96.Z.
202.4, 202.7, 202.8, 202.9, 207.8 9670 to 9729,
lymphoma (NHL) Codes and “child codes”
and 273.3. 9731 to 9738,
or beginning with C82, C83,
Codes and “child codes” 9761, 9764, 9766,
Lymphoid neoplasm C84, C85, C86, C88, C90
beginning with 200, 203 and 9811 to 9837,
(LN) and C91.
204. 9940, 9948, 9970
† For more detailed list that include the names of individual conditions, see Appendix table on InterLymph
Hierarchical Classification for Lymphoid Neoplasms and correspondences for ICD-O-3, ICD-9 and ICD-10
codes.
‡ All site codes are allowable; report site code and cell lineage (B, T, NK) if available.

- Non-Hodgkin lymphoma (NHL) case status (yes, no)

- Date of diagnosis of NHL (if you cannot provide the date, please provide the year of NHL
diagnosis, if possible)
- Age at diagnosis of NHL (years; please include decimals, if possible)
- Cell type (B, T, NK, unknown)
- Nodal or extranodal
o If extranodal, please specify site(s) and/or ICD-O-3 site (topography) code if possible
- Provisional subtype classification (Diffuse large B-cell, follicular, etc.)
- Classification system for subtype (WHO, REAL, Working formulation etc.)
- Histology type code (ICD-O-2 or -3 if possible) or diagnosis code (ICD-9 etc.)
- Classification used for histology or diagnosis, and version (ICD-O-2, ICD-O-3, ICD-9, etc.)
- Stage of cancer (In situ, localized, metastatic, metastatic regional, metastatic distant, unknown)
o Indications of extranodality (“E”), A/B symptoms and/or bulk (“X”) if available
- Staging method used (Ann Arbor or other [please specify])
- Grade of the tumor (undifferentiated, poorly, moderately, well or not differentiated) if
available/applicable
- International Prognostic Index (IPI), if available;
- For follicular lymphoma cases, in addition to above items,
o "FLIPI" if available
- For chronic lymphocytic leukemia (CLL) cases,
o Year of diagnosis of CLL
o Month and day of CLL diagnosis, if possible
o Age at diagnosis of CLL (years; please include decimals, if possible)
o B or T cell type, if known
o Histology type code (ICD-O-2 or -3 codes if possible; specify which type code is
provided) or diagnosis code (specify ICD-9, -10 etc.)
o Rai stage, if known
o Binet stage, if known

Prostate cancer

Body-site ICD9Dx-CM codes ICD10Dx-CM codes

Malignant neoplasm of prostate 185 C61

Annexe 1, page 149

- Prostate cancer case status (yes, no)
- Date of prostate cancer diagnosis (if you cannot provide the date, please provide the year of
prostate cancer diagnosis, if possible)
- Age at prostate cancer diagnosis (years; please include decimals, if possible)
- Location of the tumor (ICD code)
- ICD version
- Stage of disease and staging method used. We would prefer to receive separate data on
1) Primary tumor (T), Regional lymph nodes (N) and Distant metastases (M), and the version
of TNM classification used
2) or SEER extent of disease code (if available), the name of the manual(s), and the version
used
- Behavior of the tumor (carcinoma in situ, malignant primary site, metastatic, benign)
- Morphology of the tumor
- Grade of the tumor (undifferentiated, poorly, moderately, well or not differentiated)
- Gleason score
- Primary and secondary Gleason pattern scores separately, if available
- PSA at diagnosis, if available
- Diagnosed during a screening process (Yes, No)

Covariate data collected at baseline

General and lifestyle covariates

- Date the baseline questionnaire was returned

- Age at baseline questionnaire return (please include decimals if possible)
- Date of birth
- Gender
- Race
- Educational level
- Marital status
- Weight
- Height
- Waist circumference
- Hip circumference
- Body weight at age 18/20 years or young adulthood
- Physical activity:
Since this information has been asked in a variety of ways across studies, we would like to
create a physical activity variable with three categories: ‘sedentary’, ‘moderate activity’
and ‘vigorous activity’. If you have already created a variable for physical activity
reflecting these levels, we would like the derived variable and an explanation of how the
variable was derived, in addition to the raw data. Otherwise, please send us the data that
you currently have available for analyzing physical activity.

- Smoking:
For all smoking variables, separate information for cigarettes, pipes, and cigars would be
useful, if available

o Smoking status (never, former, current)

o Smoking pack-years
o Age at which smoking was initiated
Annexe 1, page 150
 o Age at which smoking ceased
o Time since quitting smoking
o Duration of smoking
o Amount smoked
o Passive smoking
- Non-steroidal anti-inflammatory drug use
- Personal history of hypertension
- Treatment for hypertension
- Age at menarche, women only
- Parity, women only
- Age at first birth, women only
- Menopausal status, women only
o Yes, No
o Age at menopause
o Type of menopause
- Hormone replacement therapy use (never, former, current), women only
- Oral contraceptive use (never, former, current), women only
- Latitude
- Family history of pancreatic cancer
- Family history of thyroid cancer
- Family history of kidney cancer
- Family history of lymphoma
- Family history of prostate cancer
- Family history of upper aero-digestive tract (UADT) cancer
- Personal history of pancreatitis
o Yes, No, Unknown
o Type of pancreatitis: Acute, Chronic, Other type, Unknown
- Diabetes status (Yes, No, Unknown)
- Type of diabetes
- Personal history of benign thyroid diseases
- Multivitamin use
o Yes, no
o Number of multivitamins per day
o Duration of use

Dietary covariates

- Date the baseline dietary questionnaire was returned

- Age at dietary questionnaire return, if different from age at baseline questionnaire return (please
include decimals if possible)
- Energy intake (Calories/day)
- Energy intake excluding alcohol drinking (Calories/day)
- Protein (g/day)
- Carbohydrates (g/day)
- Total fat (g/day)
- Saturated fat (g/day)
- Monounsaturated fat (g/day)
- Polyunsaturated fat (g/day)
- Cholesterol (mg/day)
- Dietary fiber (g/day)
- Vitamin B6 from food sources only (mg/day)
- Vitamin B6 from foods plus supplements (mg/day)
- Vitamin B12 from food sources only (ucg/day)
- Vitamin B12 from foods plus supplements (ucg/day)

Annexe 1, page 151

- Folate from foods (ucg/day)
- Folate from foods plus supplements (ucg/day)
- Methionine (g/day)
- Incomplete or not usable dietary questionnaire (yes, no); (we will exclude participants from
analyses involving dietary variables other than alcohol consumption)

Covariate data collected during follow-up

Please create separate variables for information coming from each follow-up cycle and mention for
each of them:

- Date of follow-up questionnaire

- Age at follow-up questionnaire return (please include decimals if possible)
- Duration covered by the follow-up questionnaire (1 year, 1 month, 1week, other).

General and lifestyle covariates

- Weight
- Height
- Waist circumference
- Hip circumference
- Physical activity:
Since this information has been asked in a variety of ways across studies, we would like to
create a physical activity variable with three categories: ‘sedentary’, ‘moderate activity’
and ‘vigorous activity’. If you have already created a variable for physical activity
reflecting these levels, we would like the derived variable and an explanation of how the
variable was derived, in addition to the raw data. Otherwise, please send us the data that
you currently have available for analyzing physical activity.

- Smoking:
For all smoking variables, separate information for cigarettes, pipes, and cigars would be
useful, if available

o Smoking status (never, former, current)

o Smoking pack-years
o Age at which smoking ceased
o Time since quitting smoking
o Duration of smoking
o Amount smoked
o Passive smoking
- Non-steroidal anti-inflammatory drug use
- Personal history of hypertension
- Treatment for hypertension
- Parity, women only
- Age at first birth, women only
- Menopausal status, women only
o Yes, No
o Age at menopause
o Type of menopause
- Hormone replacement therapy use (never, former, current), women only
- Oral contraceptive use (never, former, current), women only
- Family history of pancreatic cancer

Annexe 1, page 152

- Family history of thyroid cancer
- Family history of kidney cancer
- Family history of lymphoma
- Family history of prostate cancer
- Family history of upper aero-digestive tract (UADT) cancer
- Personal history of pancreatitis
o Yes, No, Unknown
o Type of pancreatitis: Acute, Chronic, Other type, Unknown
- Personal history of diabetes
- Diabetes status (Yes, No, Unknown)
- Type of diabetes
- Personal history of benign thyroid diseases
- Multivitamin use
o Yes, no
o Number of multivitamins per day
o Duration of use

Dietary covariates

- Date the follow-up dietary questionnaire was returned

- Age at follow-up dietary questionnaire return (please include decimals if possible)
- Energy intake (Calories/day)
- Energy intake excluding alcohol drinking (Calories/day)
- Protein (g/day)
- Carbohydrates (g/day)
- Total fat (g/day)
- Saturated fat (g/day)
- Monounsaturated fat (g/day)
- Polyunsaturated fat (g/day)
- Cholesterol (mg/day)
- Dietary fiber (g/day)
- Vitamin B6 from food sources only (mg/day)
- Vitamin B6 from foods plus supplements (mg/day)
- Vitamin B12 from food sources only (ucg/day)
- Vitamin B12 from foods plus supplements (ucg/day)
- Folate from foods (ucg/day)
- Folate from foods plus supplements (ucg/day)
- Methionine (g/day)
- Incomplete or not usable dietary questionnaire (yes, no); (we will exclude participants from
analyses involving dietary variables other than alcohol consumption)

Alcohol exposure data

Baseline alcohol information

- Baseline alcohol intake (g/d)

- Baseline alcohol intake from beer (g/d)
- Baseline alcohol intake from wine (g/d)
- Baseline alcohol intake from liquors/spirits (g/d)

Annexe 1, page 153

Lifetime alcohol information

Please provide details on the assessment method.

- Lifetime total alcohol intake (g/d)

- Did lifetime include baseline alcohol (yes, no)
- Lifetime alcohol intake from beer (g/d)
- Lifetime alcohol intake from wine (g/d)
- Lifetime alcohol intake from liquors/spirits (g/d)
- Alcohol intake (g/d) during different life periods if available, also providing the ages
corresponding to the different life periods
o Alcohol at age-i (g/d)
- Information on drinking patterns
o Number of days per week that alcohol was consumed (day(s))
o Number of alcohol drinks consumed on a day (when alcohol is consumed) (glass(es))
o How often a large number of drinks is consumed on a single day
- Information on binge drinking
Please, give details on the assessment method if other variables were used.

o Binge drinking (yes, no, unknown)

o Number of drink(s) usually consumed per drinking occasion
- Former drinking
o Age started drinking alcohol (year)
o Age stopped drinking alcohol (year)
- Former drinker of beer (yes, no)
o Age started drinking alcohol from beer (year)
o Age stopped drinking alcohol from beer (year)
- Former drinker of wine (yes, no)
o Age started drinking alcohol from wine (year)
o Age stopped drinking alcohol from wine (year)
- Former drinker of spirits and liquors (yes, no)
o Age started drinking alcohol from spirits and liquors (year)
o Age stopped drinking alcohol from spirits and liquors (year)

Alcohol intake during follow-up

Please create separate variables for each follow-up cycle and mention for each of them:

- year of follow-up questionnaire

- duration covered by the follow-up questionnaire (1 year, 1 month, 1week, other)
- Alcohol intake during follow-up (g/d)
- Alcohol intake from beer during follow-up (g/d)
- Alcohol intake from wine during follow-up (g/d)
- Alcohol intake from liquor during follow-up (g/d)

Appendix Table on the following pages

Annexe 1, page 154

Annexe 1, page 155
Annexe 1, page 156
Annexe 1, page 157
Annexe 1, page 158