Thesis Submitted in Fulfilment of the Requirements for the Degree of Doctor of Philosophy

Home , Cretan cuisine, Ikaria Study, ProVeg Nederland

Understanding the ‘Traditional’ Mediterranean Cuisine: Relationship to Cognitive Health and Advances in Measurement of Adherence

Sue Radd-Vagenas

A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

Physical Activity, Lifestyle, Ageing and Wellbeing Research Group Faculty of Health Sciences The University of Sydney 2019

“The doctor of the future will no longer treat the human frame with drugs but rather will cure and prevent disease with nutrition.”

– Thomas Edison (1847-1931)

CANDIDATE’S CERTIFICATE

I, Sue Radd-Vagenas, hereby declare that the work contained in this thesis is my own and does not contain any material submitted for the award of another degree at any institution or university.

I, Sue Radd-Vagenas, further confirm I was the principal researcher of all the work contained in this thesis, including work published with co-authors acknowledged within the text.

I, Sue Radd-Vagenas, understand that if I am awarded a higher degree for my thesis entitled “Understanding the ‘Traditional’ Mediterranean Cuisine: Relationship to Cognitive Health and Advances in Measurement of Adherence” being submitted herewith for examination, the thesis will be lodged in the University Library and will be available for immediate use. I agree that the University Librarian (or in the case of the department, the Head of Department) may supply a photocopy or microform of the thesis to an individual for research or study or to a library.

Sue Radd-Vagenas

25th February 2019

iii SUPERVISOR’S CERTIFICATE

As supervisors for Sue Radd-Vagenas’ doctoral work, we certify that the thesis entitled “Understanding the ‘Traditional’ Mediterranean Cuisine: Relationship to Cognitive Health and Advances in Measurement of Adherence” is in a form ready for examination.

Professor Victoria M Flood

Faculty of Health Sciences

The University of Sydney 25th February 2019

Professor Maria A Fiatarone Singh

Faculty of Health Sciences

The University of Sydney 25th February 2019

ACKNOWLEDGEMENTS

This thesis would not have been possible without the unwavering support and encouragement of my husband, Nicholas Constantine Vagenas. Thanks ‘agapi mou’ for believing in me and the annual trips to Greece so I could immerse myself in Mediterranean culture. Also, for sharing your ‘traditional’ Greek recipe ideas, passed down from your mother. I found it delicious and enlightening to also be able to take the science into the kitchen.

Secondly, I’d like to acknowledge my Croatian family. My father, for instilling a strong work ethic in me and always pushing me to reach greater heights. My mother, for her soothing advice and practical support when I was stressed and needed it most.

I am indebted to Professors Vicki Flood and Maria Fiatarone Singh, my distinguished supervisors. You brought diverse and complementary insights and feedback during my candidature as well as challenging and inspiring me to complete a PhD – the first in my family!

To librarians Elaine Tam and Kanchana Ekanayake: thank you for sharing your tips for database searches and to master EndNote. You helped boost my confidence.

I also thank the co-authors mentioned in this thesis. Your collaboration and comments that emanated from various disciplines enabled me to write well thought out papers, which were published in leading nutrition journals during my candidature. This was a dream come true.

It takes special people to volunteer for research projects. I am very grateful to all the participants from The Study of Mental and Resistance Training (SMART) and the

Maintain Your Brain (MYB) trial for completing dietary surveys and keeping food records. Without you, none of this would have been possible.

Thank you also to my wonderful staff at the Nutrition and Wellbeing Clinic for keeping my private practice afloat while I focussed on my doctorate. Your ability to mostly work independently has been another blessing.

Finally, to friends with advanced technical skills and an eye for detail, who freely gave of their time to assist me, I give thanks. Robert Clark for help in Adobe with figures for my systematic review and Rebecca Lister for checking on my formatting and layout within Word and teaching me tricks to improve my thesis presentation. You have contributed to making this thesis more beautiful and enjoyable to read.

DEDICATION

In loving memory of my dearest grandmother, Nevenka Lazic (1923-2016), who inspired me in all things related to good food and nutrition, ever since I could remember holding on to her apron strings in her Croatian Mediterranean kitchen.

vii

ABSTRACT

The Mediterranean diet is a popular construct and already promoted by dietary guidelines as a healthy way to eat. It is also the most widely studied dietary pattern within nutrition science, having been associated with multiple health benefits for chronic disease. One of the least well researched areas, however, relates to brain health. Finding ways to prevent, slow or reverse cognitive decline has been identified as a priority for society and governments due to increasing rates of dementia within an ageing global population. Therefore, determining the effect of the Mediterranean diet on cognition and brain morphology and function, particularly for at-risk groups, is considered of paramount importance. However, there are inconsistencies in the definition of the Mediterranean diet, and these have impeded comparisons across trials.

The terminology for the Mediterranean diet came into existence after the Seven Countries Study conducted in the late 1950s and early 1960s. It therefore referred to the ‘traditional’ Mediterranean cuisine during that time period. Yet over the last 50 years the term ‘Mediterranean diet’ has been loosely used, often without specificity to the archetypal diet. Hence, Mediterranean diet interventions and adherence tools have variably represented the ‘traditional’ dietary pattern. In addition, no adherence tool exists that has been specifically developed and validated for online and in-person use to assess adherence to the ‘traditional’ Mediterranean dietary pattern and aspects of cuisine within at-risk Western cohorts, including individuals diagnosed with mild cognitive impairment (MCI).

To address these gaps, this thesis had four specific aims. First, to empirically describe the ‘traditional’ Mediterranean cuisine, improving its understanding within the literature. Second, to systematically investigate and synthesise findings from existing clinical trials using a Mediterranean diet intervention for cognition and brain morphology or function outcomes, and inform future research requirements. A particular focus of this review was

viii on the quality of interventions used and how they related to the ‘traditional’ dietary pattern, in terms of types and quantities of prescribed foods. Third, to develop a new diet index tool to assess, more broadly, adherence to the ‘traditional’ Mediterranean dietary pattern and aspects of cuisine for use within Western cohorts. Fourth, to test the reliability and validity of this tool for online administration to healthy Australians [based on n=84 from the Maintain Your Brain (MYB) validity study cohort] and in-person use among individuals from the same country diagnosed with MCI [based on n=68 from the Study of Mental and Resistance Training (SMART) cohort]. The goal of the discussion chapter was to sum up the thesis findings and discuss their implications, and delineate remaining research gaps to progress the field.

For the first aim, a narrative review was conducted investigating the evolution of the Mediterranean diet and cuisine from ancient to modern time periods (see Chapter Two). This review identified some additional concepts from the ‘traditional’ cuisine, which have previously been poorly captured, overlooked or not fully examined, but which may be potentially important and influence health outcomes. It also found that popular Mediterranean diet index tools rarely assess such elements. To address the second aim, a pre-defined protocol was developed and the first systematic review exclusively of randomised controlled trials was conducted on the effect of a Mediterranean diet on cognition and brain morphology and function (see Chapter Three). This review established that previous interventions varied considerably and did not necessarily represent the ‘traditional’ diet and cuisine. Overall, the results showed no benefit for a Mediterranean diet intervention, with mostly non-significant and small effect sizes for cognition, and no evidence for brain morphology or connectivity outcomes across the limited experimental evidence. However, in the most robustly-designed trial, which also included a more ‘traditional’ intervention, there was evidence for attenuation of cognitive decline over four years of ageing compared to a lower fat diet, with significant effect sizes for three domain composite scores representing Memory, Frontal and Global function. This finding persisted with the recent corrections within the original paper for a sub- cohort of this study. Hence, promising clinical evidence exists to support that a more ‘traditional’ Mediterranean diet may slow progression of cognitive decline in older adults with cardiovascular risk factors. To achieve the third aim, a new Mediterranean diet index

ix tool was developed (see Chapter Four). This was informed by earlier findings described in Chapter Two and the reported limitations of existing tools, which may be reducing their ability to consistently predict health outcomes. The tool, named Mediterranean Diet and Culinary Index (MediCul), is in the form of a survey. Scoring is from 0 to 100, with higher scores indicating greater adherence to the ‘traditional’ Mediterranean dietary pattern. MediCul includes both elements and cut-off points based on the ‘traditional’ Mediterranean diet and cuisine. It takes approximately 20 minutes to complete, and is therefore less onerous than a typical food frequency questionnaire, but more comprehensive than a short screener. Importantly, MediCul assesses exposure to discretionary foods that contribute significantly to energy intakes within Western countries. It also enables indirect derivation of a score for the popular Spanish Mediterranean Diet Adherence Screener (MEDAS), increasing its potential future utility. The fourth aim was met in Chapters Five and Six, which have reported on the reliability and validity testing of MediCul for use online and in-person with middle-aged and older Australians, including individuals diagnosed with MCI. Taken together, these Chapters are in agreement that MediCul is a highly reliable tool, based on intra-class correlation coefficients (ICC) (ICC=0.86, 95% CI: 0.789, 0.910 and ICC=0.93, 95% CI: 0.884, 0.954, respectively, p<0.0001). MediCul also has moderate validity according to Bland- Altman plots with the tool overestimating the MediCul score by 6% in both studies compared to a food record reference method. MediCul can therefore be used in future research to assess adherence to the ‘traditional’ Mediterranean dietary pattern and aspects of cuisine in middle-aged and older Australians who are at risk of cognitive decline.

While various nutrients, foods and plant-based dietary patterns may benefit health and slow cognitive decline, most evidence exists to support the Mediterranean diet. This diet has been identified as being of particular significance for older people and an ageing population, such as in Australia, since the confirmed benefits for the Mediterranean diet are relevant to key areas of morbidity and survival. Unfortunately, most Australians consume a poor quality Western diet according to the Australian Institute of Health and Welfare, and Western diets are associated with increased risk of cognitive decline and dementia. The Westernised nature of Australian diets was supported by the cohorts studied within this thesis. These cohorts had moderately low MediCul scores [MYB

x cohort: 55.2/100.0 (11.6); SMART cohort: 54.6/100.0 (13.0)] and derived MEDAS scores [MYB cohort: 6.5/14.0 (1.7); SMART cohort: 6.1/14.0 (2.2)], indicating poor adherence with a Mediterranean dietary pattern. Yet such individuals may benefit most by adopting more of a Mediterranean diet, due to their existing risk factors and/or age.

As there is no pharmacologic prevention or cure for dementia, robust trials are urgently required to test whether a healthy diet and lifestyle can slow the rate and severity of cognitive decline, both within normal ageing and for neurodegenerative disease. Trials are warranted specifically on the ‘traditional’ Mediterranean diet as an intervention, which may be combined with other lifestyle modalities. Future trials should fully describe their interventions so they can be replicated. They should also use a validated diet index tool to assess adherence to the ‘traditional’ Mediterranean dietary pattern.

In conclusion, promotion of lifestyle interventions to slow progression of cognitive decline in at-risk groups is critical. Additional well-designed trials on the relationship of the ‘traditional’ Mediterranean diet with cognitive health and other outcomes are warranted. This thesis has improved understanding of the ‘traditional’ Mediterranean cuisine, underpinning the development of a new diet index tool called MediCul to assess adherence to this pattern and provide greater discernment with other plant-based dietary patterns. MediCul has been tested in two Western cohorts at increased dementia risk, for online and in-person administration, and shown to be suitable for use. This new tool is now available for download from the Supplementary Materials in the British Journal of Nutrition where it was published, for researchers and educators globally. Availability of validated tools such as MediCul may advance understanding of the role of diet in healthy brain ageing.

TABLE OF CONTENTS

Candidate’s certificate…………………………………………………………………..iii Supervisor’s certificate…………………………………………………………….……iv Acknowledgements……………………………………………………………………...v Dedication…………………………………………………………………………...…vii Abstract…………………………………………………………………………..……viii Table of contents…………………………………………………………………..…...xii Outline of thesis……………………………………………………………………...xviii Dissemination of research……………………………………………………………...xx

CHAPTER ONE: Introduction……………………………………………………………………………..1 1.1 Popularity and promotion of the Mediterranean diet………………………………..2 1.1.1 Protection by UNESCO………………………………...…………………....2 1.1.2 Opportunity for healthier ageing…….……………...……………………...... 3 1.2 Health benefits for chronic disease.………………………………………………...3 1.2.1 Cognition, brain morphology and function..………………………………....5 1.2.1.1 Burden of dementia………………………………...………………….6 1.2.1.2 Limits of existing medical treatment…………………………………..7 1.2.1.3 Dementia risk factors………………………………………………….8 1.3 Measurement of Mediterranean diet adherence…..………………………………...9 1.3.1 Options for dietary assessment methods…………………………………....10 1.3.2 Statistical analysis for reliability and validity……………………………....11 1.3.3 Limitations of existing Mediterranean diet index tools…..………………...14 1.3.4 Discernment for the ‘traditional’ diet………...……….…………………….15 1.4 Gaps in research…………………………………………………………...... 15 1.5 Thesis aims.....………………………………………………………………….....16 1.6 References……………………………………………………………………...... 18

xii CHAPTER TWO: Evolution of Mediterranean diets and cuisine: Concepts and definitions…………37 Preface………………………………………………………………………………...38 Authorship statement……………………………………………………………….…40 Abstract……………………………………………………………………………..…41 Introduction…………………………………………………………………………...41 Methods……………………………………………………………………………….41 Results………………………………………………………………………………...42 Discussion………………………………………………………………………….....48 References………………………………………………………………………….…51

CHAPTER THREE: Effect of the Mediterranean diet on cognition and brain morphology and function: A systematic review of randomised controlled trials………………...……………....56 Preface………………………………………………………………………………...57 Authorship statement……………………………………………………………….…60 3.1 Abstract………………………………………………………………………....…64 3.2 Introduction…………………………………………………………………….…66 3.3 Methods………………………………………………………………………...... 67 3.3.1 Criteria for study inclusion...... 67 3.3.2 Search strategy...... 68 3.3.3 Study selection and data extraction...... 69 3.3.4 Data synthesis and analysis...... 70 3.3.5 Quality assessment...... 71 3.4 Results…………………………………………………………………………...... 72 3.4.1 Study selection…………………………………………………………...…72 3.4.2 Study characteristics……………………………………………………..…72 3.4.2.1 Study design and participants……………………………………..…72 3.4.2.2 Intervention………………………………………………………..…73 3.4.3 Cognitive outcomes………………………………………………………...77 3.4.3.1 Overview……………………………………………………….....…77 3.4.3.2 Global cognition…………………………………………………..…78 3.4.3.3 Attention…………………………………………………………..…79 3.4.3.4 Working memory…………………………………………….…...…79

xiii 3.4.3.5 Processing speed………………………………………………...... …80 3.4.3.6 Verbal and visual memory…………………………………………...80 3.4.3.7 Visuo-spatial…………………………………………………………81 3.4.3.8 Language…………………………………………………………..…81 3.4.3.9 Executive function…………………………………………….…...... 82 3.4.3.10 Motor control…………………………………………………...... 82 3.4.4 Brain morphology/function outcomes……………………………………...82 3.4.5 Incident MCI and dementia……………………………………………...... 83 3.5 Discussion………………………………………………………………………....84 3.5.1 Key findings……………………………………………………………...... 84 3.5.2 Mechanisms………………………………………………………………...85 3.5.3 Implications………………………………………………………………...86 3.5.4 Strengths and limitations…………………………………………………...86 3.5.5 Future research…………………………………………………………..….87 3.5.6 Conclusion…………………………………………………………....….…87 3.6 References………………………………………………………………………....88 Supplementary material………………………………………………………...... …114

CHAPTER FOUR: Development of the Mediterranean Diet and Culinary Index (MediCul) tool….....211 4.1 Introduction……………………………………………………………………....212 4.1.1 Rationale for use of diet index tools……………………………………….212 4.1.2 Mediterranean diet index tools…………………………………………….213 4.1.3 Limitations of existing tools……………………………………………….214 4.1.4 Ideal characteristics for a new tool………………………………………...216 4.2 Method…………………………………………………………………………...217 4.2.1 Identification of ‘traditional’ elements……………….…………………....218 4.2.2 Tool naming……………………………………………………………….218 4.2.3 Tool development………………………………………………………....218 4.2.4 Tool piloting………………………………………………………………220 4.2.5 Tool cut-off points………………………………………………………...221 4.2.6 Tool weighting………………………………………………………….....221 4.2.7 Tool scoring……………………………………………………………….222 4.2.8 Tool adaptation for online use…………………………………………...... 223

xiv 4.2.9 Scoring for MEDAS...... 224 4.3 Results…………………………………………………………………………...224 4.4 Discussion………………………………………………………………………..225 4.4.1 Strengths…………………………………………………………………..225 4.4.1.1 Original data sources used for cut-off points………………………..225 4.4.1.2 Novel Mediterranean elements assessed……………………………226 4.4.1.3 Exposure to Western foods assessed………………………………..226 4.4.1.4 Foods associated with cognition assessed…………….………….....226 4.4.1.5 Zero alcohol intake not penalised……………………………………226 4.4.1.6 Practical serves and images…………………………………………227 4.4.1.7 More comprehensive than a screener…………………………….....227 4.4.1.8 Complex computations not required…..…………………………....227 4.4.1.9 MEDAS score can be derived………………………………………227 4.4.2 Limitations………………………………………………………………...228 4.4.2.1 Limited dietary assessment period………………………………….228 4.4.2.2 Healthiness of ‘traditional’ diet assumed…………………………...228 4.4.2.3 Combined quantity and frequency questions used…………….……229 4.4.2.4 Less common Western foods not assessed………………………….229 4.4.2.5 Food preparation methods uncertain if not primary cook…………...229 4.4.2.6 Food avoidance due to intolerance/allergy affects scoring……….…229 4.4.2.7 Sex and energy-adjusted cut-off points unavailable..……………….229 4.4.2.8 Not designed to calculate nutrient intakes.……...………………….230 4.4.2.9 Timing of food intake not assessed…………………………………230 4.5 Conclusion……………………………………………………………………….231 4.6 References……………………………………………………………….……….232

CHAPTER FIVE: Validity of the Mediterranean diet and culinary index (MediCul) for online assessment of adherence to the ‘traditional’ diet and aspects of cuisine in older adults……………………………………………………………………………….…244 Preface…………………………………………………………………………….…245 Authorship statement………………………………………………………………...247 Abstract…………………………………………………………………………....…248 Introduction………………………………………………………………………..…249

xv Materials and methods………………………………………………..………….…..250 Results…………………………………………………………………………….….253 Discussion……………………………………………………………………….…...260 References………………………………………………………………………..…..261 Supplementary material………………………………………………………….…..265

CHAPTER SIX: Reliability and validity of the Mediterranean diet and culinary index (MediCul) tool in an older population with mild cognitive impairment…………………………....273 Preface…………………………………………………………………………...... 274 Authorship statement……………………………………………………………...…276 6.1 Abstract……………………………………………………………………….….280 6.2 Introduction………………………………………………………………...…….281 6.3 Methods……………………………………………………………………….….283 6.3.1 Participants…………………………………………………………...... …283 6.3.2 Data administration and collection………………………………………...284 6.3.3 Tool development……………………………………………………...... 285 6.3.4 Nutritional analysis………………………………………………………..286 6.3.5 Statistical analysis………………………………………………………....286 6.4 Results…………………………………………………………………………....289 6.4.1 Reliability…………………………………………………………….…....289 6.4.2 Validity……………………………………………………………………290 6.5 Discussion……………………………………………………………………...... 291 6.5.1 Limitations………………………………………………………………...293 6.5.2 Strengths…………………………………………………………………..295 6.5.3 Conclusions…………………………………………………………….….295 6.6 References…………………………………………………………………….….297 Supplementary material…………………………………………………………...... 322

CHAPTER SEVEN: Discussion and conclusions……………………………………………………….….380 7.1 Overview and outcomes of thesis.…………...…………………………………...381 7.1.1 Improved understanding of ‘traditional’ Mediterranean cuisine……….….381 7.1.2 Evidence for cognitive benefits warranting additional trials……………....382

xvi 7.1.3 Development of a new empirically-based adherence tool……………….…383 7.1.4 Tool validation within two cohorts for online and in-person use…………384 7.2 Limitations and strengths………………………………………………………...387 7.3 Implications……………………………………………………………………....391 7.3.1 Future research…………………………………………………………..…391 7.3.2 Public health and clinical significance………………………………….…392 7.4 Conclusions……………………………………………………………………....394 7.5 References…………………………………………………………………….….396

APPENDIX:…...……………………………………………………………………...405 List of appendices…………………………………………………………………....406 Appendix One: Conference posters and proceedings………………………………...408 Appendix Two: Additional works not forming part of this thesis………………….…414 Appendix Three: PROSPERO protocol for systematic review……………………....433 Appendix Four: Corrections within PREDIMED for cognitive outcomes…………...440 Appendix Five: Comparison of selected Mediterranean diet index tools………….…451 Appendix Six: Mediterranean Diet Adherence Screener (MEDAS)………………....493 Appendix Seven: Screen shots of sample MediCul questions formatted for online administration………………………………………………………………………..497 Appendix Eight: Ethics letters……………………………………………………….500 Appendix Nine: Research Food Diary app instructions……………………………...504 Appendix Ten: Dietitian assisted food record checklist……………………………...524 Appendix Eleven: Front sheet for validation and reliability testing of MediCul……..527 Appendix Twelve: 3-day paper food record template………………………………..529 Appendix Thirteen: FAQ on completing the 3-day food record…………………...... 539 Appendix Fourteen: Challis Bequest Society Lunch………………………………....542

xvii OUTLINE OF THESIS

This thesis is organised into seven chapters, with material submitted and published in peer-reviewed journals during the time of candidature.

Chapter One is an introduction to the thesis. It provides on overview of the significant issues identified within the thesis, the research gaps and specific aims to be addressed within the remaining chapters.

Chapter Two presents a literature review describing the evolution of Mediterranean diets and cuisine. It defines the ‘traditional’ diet using empirical data and identifies key concepts from this time period, frequently lacking in existing Mediterranean diet adherence tools. This work has been published in the Asia Pacific Journal of Clinical Nutrition.

Chapter Three presents a systematic review of randomised controlled trials on the effect of the Mediterranean diet on cognition and brain morphology. It includes a focus on how previous interventions have reflected the ‘traditional’ dietary pattern and makes detailed recommendations to assist future research. This work has been published in the American Journal of Clinical Nutrition.

Chapter Four details the development of a new diet index tool to assess adherence to the ‘traditional’ Mediterranean dietary pattern and aspects of cuisine within Western populations. This tool has overcome common limitations of existing tools and incorporates additional features to increase its potential future utility.

xviii Chapter Five presents the findings for the reliability and validity testing of the new index tool when administered online to healthy middle aged and older Australians. This work has been published in the journal Nutrients.

Chapter Six presents the findings for the reliability and validity testing of the new index tool when administered in-person to older Australians diagnosed with mild cognitive impairment, who are known to be at increased dementia risk. This work has been published in the British Journal of Nutrition.

Chapter Seven discusses the key findings from Chapters Two to Six and strengths and limitations of the thesis. It delineates remaining gaps in the evidence base and highlights implications of the contributed knowledge from this thesis for future research, public health and clinical practice.

Each chapter includes its own reference list and any Supplementary Materials published for that chapter.

xix

DISSEMINATION OF RESEARCH

Parts of the work presented in this thesis have been published in peer-reviewed journals and presented at conferences.

Research findings have also been disseminated to scientists, health professionals and the general public through various channels.

Peer-reviewed papers

Chapter Two:

• Radd-Vagenas, S., Kouris-Blazos, A., Fiatarone Singh, M., & Flood, V. M. (2017). Evolution of Mediterranean diets and cuisine: Concepts and definitions. Asia Pacific Journal of Clinical Nutrition, 26(5), 749-763. doi:10.6133/apjcn.082016.06

Chapter Three:

• Radd-Vagenas, S., Duffy, S. L., Naismith, S. L., Brew, B. J., Flood, V. M., & Fiatarone Singh, M. A. (2018). Effect of the Mediterranean diet on cognition and brain morphology and function: A systematic review of randomized controlled trials. American Journal of Clinical Nutrition, 107(3), 389-404. doi:10.1093/ajcn/nqx070

Chapter Five:

• Radd-Vagenas, S., Fiatarone Singh, M., Daniel, K., Noble, Y., Jain, N., O’Leary, F., . . . Flood, V. (2018). Validity of the Mediterranean Diet and Culinary Index (MediCul) for online assessment of adherence to the ‘traditional’ diet and aspects of cuisine in older adults. Nutrients, 10(12), 1913. doi:10.3390/nu10121913

Chapter Six:

• Radd-Vagenas, S., Fiatarone Singh, M. A., Inskip, M., Mavros, Y., Gates, N., Wilson, G. C., . . . Flood, V. M. (2018). Reliability and validity of a Mediterranean Diet and Culinary Index (MediCul) tool in an older population with mild cognitive impairment. British Journal of Nutrition, 120(10), 1189-1200. doi:10.1017/S0007114518002428

Conference presentations - podium

• Radd, S., Duffy, S., Naismith, S., Brew, B., Flood, V., & Fiatarone Singh, M. Effect of the Mediterranean diet on cognition and brain morphology/function: A systematic review and meta-analysis of RCTs. 10th Asia Pacific Conference on Clinical Nutrition, Adelaide, Australia., 26-29th November 2017. • Radd-Vagenas, S., Daniel, K., Noble, Y., Fiatarone Singh, M., & Flood, V. Reliability of new index tool to assess adherence to Mediterranean diet among people with mild cognitive impairment. Dietitians Association of Australia 35th National Conference, Sydney, Australia, 17-19th May 2018. • Radd-Vagenas, S., Fiatarone Singh, M. A., Daniel, K., Noble, Y., O’Leary, F., Mavros, Y., Brodaty, H., Flood, V. M. Reliability and validity of MediCul (Mediterranean Diet & Culinary Index) in older Australian adults. 42nd Nutrition Society of Australia (Inc.) Annual Scientific Meeting, Canberra, Australia, 27- 30th November 2018.

(See proceedings in Appendix One).

xxi

Conference presentations - poster

• Radd-Vagenas, S., Kouris-Blazos, A., Fiatarone Singh, M., Flood, V. What is the traditional Mediterranean diet? Joint Annual Scientific Meeting of the Nutrition Society of New Zealand and the Nutrition Society of Australia, Wellington, New Zealand, 1-4th December 2015. • Radd-Vagenas, S., Duffy, S. L., Naismith, S. L., Brew, B. J., Flood, V. M., & Fiatarone Singh, M. A. (2018). Could a plant-based diet affect cognition? Findings from a systematic review of randomized controlled trials on the Mediterranean diet. 7th International Congress on Vegetarian Nutrition, Loma Linda University, Loma Linda, USA, 26-28th February 2018.

(See posters in Appendix One).

Seminars – scientific community

• The Mediterranean diet and your mind: How does it affect memory and thinking performance?, Martin Thompson Seminar, Faculty of Health Sciences, The University of Sydney, Lidcombe, Australia, 21st June 2017. • The Mediterranean diet & health, webinar for Graduate Diploma in Lifestyle Medicine, Avondale College, 19th September 2017. • The Mediterranean diet and the brain: Where does the evidence stand?, speaker prior to Annual General Meeting of the Nutrition Society of Australia, Sydney group, The University of Sydney, Sydney, Australia, 14th March 2018.

Seminars – general public

• Eating the Mediterranean way & healthy ageing at Kick Start Healthy Ageing, organised by Aged Care Psychiatry Service, The Prince of Wales Hospital, Kingsford, The Juniors, Kingsford, Australia, 9th November 2016. • Benefits of the Mediterranean diet at Alive & Kicking elderly group, Summer Hill Community Centre, Summer Hill, Australia, 22nd November 2016.

xxii

• Healthy brain ageing, Sir Moses Montefiore Jewish Home Silver Ribbon Tea Club, Randwick, Australia, 22nd March 2017. • Could the Mediterranean diet affect dementia? Raising the Bar organised by The University of Sydney, PS40, Sydney, Australia, 25th October 2017. • Hot nutrition issues – The future of food as medicine, Loma Linda University, USA, 24th February 2018. • Panellist at Sydney Health Forum on Food as medicine, organised by The University of Sydney, Sydney, Australia, 23rd May 2018. • Panellist at Wolper Wellbeing Program on Healthy ageing organised by Wolper Jewish Hospital, Event Cinemas, Bondi Junction, Australia, 4th July 2018. • Menu review and provision of eating guidelines for the Mediterranean diet, Challis Bequest Society Lunch, Division of Alumni and Development, The University of Sydney, Sydney, Australia, 1st November 2018 (see Appendix Fourteen).

Awards

During their candidature, the candidate published a related consumer book, for which she was awarded Winner for ‘Best Health and Nutrition Book’ in the world at Gourmand World Cookbook Awards, Yantai, China, 26-29th May 2017:

- Food as Medicine: Cooking for Your Best Health by Sue Radd, Signs Publishing Company, Australia, October 2016. [ISBN 978 1 925044 62 1].

xxiii

Fifty five Mediterranean recipes from Food as Medicine: Cooking for Your Best Health are being used in Maintain Your Brain (MYB), an online, randomised trial funded by the National Health and Medical Research Council. Sixteen Mediterranean video recipes and tips were also produced by the candidate for use in the MYB trial. These can be publicly viewed as indicated below:

- Roasted red capsicum, walnut and pomegranate dip: https://www.youtube.com/watch?v=CIW-6BIKE2k - Village-style vegetable stew: https://www.youtube.com/watch?v=dmVdhmTKUCs - Black eyed bean salad with lemon and shallots: https://www.youtube.com/watch?v=dR-RZMfcW0o - Succulent eggplant and tomato bake: https://www.youtube.com/watch?v=eGB8tph1G4Y - Warm fava bean salad: https://www.youtube.com/watch?v=yQLWK7W14fY - Lentil olive and semi dried tomato pasta sauce: https://www.youtube.com/watch?v=Y0iRZs4heOc - Greek-style pea stew: https://www.youtube.com/watch?v=Pkq8KZ-B4LI - Wilted endive with lemon and olive oil: https://www.youtube.com/watch?v=kRxT6SsytCQ - Lemony chickpea soup: https://www.youtube.com/watch?v=Cl6OObRjfck - Spiced chickpeas with silverbeet and lemon: https://www.youtube.com/watch?v=2-0hWB2NOqA - Cannellini bean and carrot soup: https://www.youtube.com/watch?v=tbg_XVzAIfs - Shaved savoy cabbage with lemon dressing: https://www.youtube.com/watch?v=zpn4q-KAMS0 - Fresh dates stuffed with almonds: https://www.youtube.com/watch?v=_c-P9uTSPYQ

xxiv

- How to prepare dried oregano: https://www.youtube.com/watch?v=fua1pB4QXP4 - How to de-seed a pomegranate: https://www.youtube.com/watch?v=oiZLT8TrhFo - Bread with extra virgin olive oil and oregano: https://www.youtube.com/watch?v=JmNSKxBGoVQ

Additional works by the author not forming part of thesis

Peer-reviewed papers

• Heffernan, M., Andrews, G., Fiatarone Singh, M. A., Valenzuela, M., Anstey, K. J., Maeder, A., McNeil, J., Jorm, L., Lautenschlager, N., Sachdev, P., Ginige, A., Hobbs, M., Boulamatsis, C., Chau, T., Cobiac, L., Cox, K., Daniel, K., Flood, V. M., Guerrero, Y., Gunn, J., Jain, N., Kochan, N. A., Lampit, A., Mavros, Y., Meiklejohn, J., Noble, Y., O’Leary, F., Radd-Vagenas, S., & Brodaty, H. (2018). Maintain Your Brain: Protocol of a 3-year randomized controlled trial of a personalized multi-modal digital health intervention to prevent cognitive decline among community dwelling 55 to 77 year olds. Journal of Alzheimer's Disease, Advance online publication. doi:10.3233/JAD-180572 (see Appendix Two).

Published abstracts

• Flood, V., Russell, J., & Radd, S. (2015). Trends in legume consumption among ethnically diverse adults in a longitudinal cohort study in Australia. The FASEB Journal, 29(1) (abstract 381.4). (See Appendix Two).

xxv

CHAPTER ONE

Introduction

1 1.1 Popularity and promotion of the Mediterranean diet

The Mediterranean diet has become increasingly popular in both the scientific and lay literature. It was ranked the ‘best diet’ for 2019 by a panel of experts (US News and World Report, 2019) and is already being promoted by health authorities around the world. For example, the Australian and US dietary guidelines recommend the Mediterranean diet as a healthy eating pattern (National Health and Medical Research Council, 2013; US Department of Health and Human Services and US Department of Agriculture, 2015) as does the UK National Health Service (NHS, 2017). Various Mediterranean diet guidelines and pyramids exist for use in nutrition education (Bach-Faig et al., 2011; Fidanza & Alberti, 2005; Ministry of Health and Welfare: Supreme Scientific Health Council, 1999; Oldways, 2008; Willett et al., 1995). In Australia, the Royal Australian College of General Practitioners recommends a Mediterranean diet for primary and secondary prevention of cardiovascular events (National Health and Medical Research Council level 1 evidence) as part of their online Handbook of Non-Drug Interventions (HANDI) (The Royal Australian College of General Practitioners, 2014). Most recently, a Mediterranean-like diet is being promoted by the World Health Organization (WHO) for risk reduction of cognitive decline and dementia (World Health Organization, 2019).

1.1.1 Protection by UNESCO

This gastronomic eating pattern with intense colours and flavours is also the only diet in the world protected by UNESCO (United Nations Educational Scientific and Cultural Organization, 2010) to safeguard its ‘traditional’ aspects, which are increasingly being abandoned in Mediterranean countries (De Lorenzo et al., 2001; Hatzis et al., 2013). In 2010, UNESCO first recognised the Mediterranean diet as an ‘Intangible Cultural Heritage’ for safeguarding within Spain, Greece, Italy and Morocco. In 2013, this was extended to also include the countries of Croatia, Cyprus and Portugal (United Nations Educational Scientific and Cultural Organization, 2013). Further, this diet is the most extensively researched diet worldwide (Scarmeas, Anastasiou, & Yannakoulia, 2018). Interestingly, a large number of studies on the Mediterranean diet, investigating various outcomes, have been conducted outside the Mediterranean region (Bonaccio et al., 2018),

2 and span countries ranging from Sweden (Sjögren et al., 2010) and China (Woo et al., 2001) to the US (Fung et al., 2009) and Australia (Itsiopoulos et al., 2011).

1.1.2 Opportunity for healthier ageing

Adopting the Mediterranean diet could benefit wellbeing and promote healthier ageing by reducing the risk of, and better managing, chronic disease especially among middle- aged and older adults where health problems become evident and progressively increase. Improved metabolic health with ageing may also be effective for brain health (Wahl et al., 2019). Healthy ageing is a priority area for governments as there is rapid global population growth (Roser & Ortiz-Ospina, 2018) with shifts towards larger proportions of older adults, particularly in Western countries (Bongaarts, 2009). The United Nations recently projected that people aged 60 and older will outnumber children in 2047 (United Nations Department of Economic and Social Affairs Population Division, 2013). Hence, ageing is not just a biological process but a public policy issue with economic and societal impacts (Kennedy et al., 2014). To help promote healthy ageing, including for the brain, the UK National Institute for Health and Care Excellence guidelines on mid-life approaches to prevent or delay onset of dementia, disability and frailty (National Institute for Health and Care Excellence, 2015) recommend a diet mainly based on vegetables, fruits, beans and pulses, wholegrains and fish. These food groups represent key elements of healthy plant-based diets, including those rooted in ‘traditional’ culture, such as the Mediterranean diet.

1.2 Health benefits for chronic disease

Although the Mediterranean diet was not widely recognised until the 1990s (de Lorgeril, 2013) mounting evidence now exists that eating the Mediterranean way may provide a range of benefits to target risk factors, as well as existing chronic disease, thereby promoting healthier ageing. However, this evidence is primarily, but not exclusively, from observational studies with the highest level of evidence i.e., clinical, available for cardiovascular disease (Martínez-González, Gea, & Ruiz-Canela, 2019). The areas of research relevant to chronic disease prevention and management include the following:

3 • body weight/obesity (Buckland, Bach, & Serra‐Majem, 2008; Estruch et al., 2016; Martínez-González et al., 2012; Panagiotakos, Chrysohoou, Pitsavos, & Stefanadis, 2006); • cancer (Li et al., 2018; Sofi, Cesari, Abbate, Gensini, & Casini, 2008; Sofi, Macchi, Abbate, Gensini, & Casini, 2014; Toledo et al., 2015); • cardiovascular disease (CVD) (Davis et al., 2017; de Lorgeril et al., 1994; Estruch et al., 2018; Grosso et al., 2017); • dementia and mild cognitive impairment (Alonso et al., 2009; Psaltopoulou et al., 2013; Singh et al., 2014; Sofi et al., 2008); • gallstones (Barré, Gusto, Cadeau, Carbonnel, & Boutron-Ruault, 2017); • hip fracture (Benetou et al., 2013; Benetou et al., 2018); • kidney disease (Huang et al., 2013; Khatri et al., 2014); • macular degeneration (Hogg et al., 2016; Merle, Silver, Rosner, & Seddon, 2015); • mental health, including quality of life (Bonaccio et al., 2013; Govindaraju, Sahle, McCaffrey, McNeil, & Owen, 2018; Henríquez Sánchez et al., 2012) and depression (Jacka et al., 2017; Lassale et al., 2018; Parletta et al., 2017); • metabolic syndrome (Kastorini et al., 2011); • mortality risk (Sofi et al., 2014; Trichopoulou, Costacou, Bamia, & Trichopoulos, 2003), including in the elderly (Bonaccio et al., 2018; Kouris-Blazos & Itsiopoulos, 2014); • non-alcoholic fatty liver disease (Cueto-Galán et al., 2017; Kontogianni et al., 2014; Papamiltiadous et al., 2016; Ryan et al., 2013); • osteoarthritis (Veronese et al., 2017; Veronese et al., 2018); • Parkinson’s disease (PD) (Psaltopoulou et al., 2013; Sofi et al., 2008) and prodromal PD (Maraki et al., 2019); • physical function (Struijk, Guallar-Castillón, Rodríguez-Artalejo, & López- García, 2018) and frailty (Kojima, Avgerinou, Iliffe, & Walters, 2018); • telomere length (Boccardi et al., 2013; Crous-Bou et al., 2014); • the microbiome (De Filippis et al., 2016; Shively et al., 2018); • type 2 diabetes (Davies et al., 2018; Esposito et al., 2015; Itsiopoulos et al., 2011; Salas-Salvadó et al., 2014).

4 Various metabolic and molecular mechanisms have been proposed whereby particular nutrients, foods, the dietary pattern and cuisine aspects (e.g., food combinations, fasting, infrequent snacking) of the Mediterranean diet, as well as the synergies afforded (Jacobs & Tapsell, 2013), may influence multiple chronic diseases. These include lipid lowering effects; reduced oxidative stress, chronic systemic inflammation and platelet aggregation; as well as modification of hormones, growth factors, and nutrient sensing pathways via the microbiome (Tosti, Bertozzi, & Fontana, 2017), which are discussed more fully in Chapter Three.

1.2.1 Cognition, brain morphology and function

Most recently, interest has increased in healthy brain ageing and research is focussing on how the Mediterranean diet may influence cognition (Valls-Pedret et al., 2015), as well as brain morphology and function (Luciano et al., 2017; Pelletier et al., 2015) (see Chapter Three). This is imperative since pathological changes in the brain such as amyloid plaques and neurofibrillary tangles start appearing several decades before the onset of any clinical symptoms (Sperling et al., 2011). Interestingly, the ageing process and dementia share common characteristics at the cellular, molecular and system levels (Xia, Jiang, McDermott, & Han, 2018). A spectrum has therefore been proposed that links normal brain ageing and dementia, as the hallmarks of ageing are also present in dementia e.g., reduced DNA repair, abnormal neuronal activity and neuroinflammation (Wahl et al., 2019). However, while minor cognitive changes are common during normal brain ageing (Fletcher et al., 2018), dementia is progressive and influences function across various domains e.g., multiple domains of cognition, neuropsychiatric symptoms, activities of daily living (Livingston et al., 2017) and the degree of neuropathology, rather than specific pathological features, often differentiates normal brain ageing from dementia (Anderton, 1997). Successful interventions for cognitive decline may therefore be beneficial to both ends of this spectrum – normal brain ageing and dementia. In one longitudinal brain imaging study of cognitively normal people aged 30 to 60 years, higher adherence to a Mediterranean diet was associated with less beta-amyloid and a slower decline in brain metabolism over a three-year period. The authors of this study estimated

5 that high Mediterranean diet adherence may provide 1.5 to 3.5 years of protection against Alzheimer’s Disease (AD), the most common type of dementia (Berti et al., 2018).

According to the Academy of Nutrition and Dietetics, “the total diet or overall pattern of food eaten is the most important focus of healthy eating” (Freeland-Graves & Nitzke, 2013). Although the study of individual nutrients remains relevant (Wahl et al., 2018), nutrition science is moving away from a reductionist model (Fardet & Rock, 2014; Rutjes et al., 2018). The greatest evidence for cognitive benefits exists for dietary patterns that represent a ‘whole of diet approach’, which can also capture synergies between dietary elements. There is currently more evidence to support a Mediterranean dietary pattern than any other, such as patterns based on the Dietary Approaches to Stop Hypertension (DASH), Mediterranean-Dietary Approaches to Stop Hypertension (MIND) or the Nordic diet (Scarmeas et al., 2018). Hence, the Mediterranean diet is being increasingly studied, including within Australia (Hardman, Kennedy, Macpherson, Scholey, & Pipingas, 2015; McMaster et al., 2018; Wade et al., 2017) in an attempt to find ways to prevent or slow cognitive decline due to the significant burden of dementia (Livingston et al., 2017).

1.2.1.1 Burden of dementia

Dementia has been identified as the greatest challenge globally for health and social care in the 21st century (Livingston et al., 2017), with one new case being diagnosed every 3 seconds and a projected prevalence of 75 million by 2030 (Prince et al., 2015). In Australia, dementia is the second cause of death overall (Australian Institute of Health and Welfare, 2012) but the leading cause in women (Australian Institute of Health and Welfare, 2018). It is also the second most feared disease after cancer, among the general public (Pfizer, 2011). In terms of disability burden, which affects quality of life, dementia is the foremost contributor among Australians aged 65 years and older (Australian Institute of Health and Welfare, 2012). Hence, effective preventive measures and interventions are urgently required.

6 1.2.1.2 Limits of existing medical treatment

Despite the alarming trends and likely future healthcare costs, there is currently no medication that can successfully modify the course of cognitive decline towards dementia (Cooper, Li, Lyketsos, & Livingston, 2013). Development of dementia drugs has proven to be very difficult, with few candidates and frequent failures (Cummings, Morstorf, & Zhong, 2014). For example, during the 2002 to 2012 period, there was a very high attrition rate for compounds tested in drug trials, with only 0.4% success rate (99.6% failure rate) for progression to regulatory review (Cummings et al., 2014). This rate of success is among the lowest found in any therapeutic area. The success rate for development of cancer drugs, for example, is 19% (Tufts Center for the Study of Drug Development, 2013). Due to disappointing pharmaceutical research results in the dementia field, one major company recently announced they are pulling out of neuroscience research (Hawkes, 2018).

At best, existing medications may temporarily improve memory or behaviour by stabilising cognition and neuropsychiatric symptoms such as agitation, sleep and apathy, in a minority of individuals. However, their use is also associated with undesirable side effects (Buckley & Salpeter, 2015) and these medications do not improve long-term prognosis or survival (Casey, Antimisiaris, & O’Brien, 2010).

Therefore, finding new ways to prevent, slow and potentially reverse cognitive decline towards dementia is a high priority for society and governments (Australian Institute of Health and Welfare, 2012). For AD alone, if interventions could delay or slow its progression by just one year there would be 9 million fewer cases in 2050, according to estimates (Brookmeyer, Johnson, Ziegler-Graham, & Arrighi, 2007). Yet current approximations suggest there are nearly eight times as many people who have pre-clinical AD (amyloidosis, neurodegeneration or both without overt symptoms) as those with clinical AD and mild cognitive impairment (MCI) (Brookmeyer, Abdalla, Kawas, & Corrada, 2018). Thus, dementia statistics under-represent the extent of the problem and are akin to the ‘tip of an iceberg’, due to the known trajectory of the underlying disease pathology.

7 1.2.1.3 Dementia risk factors

Lifestyle has been identified as a modifiable risk factor for cognitive decline and dementia. Healthy lifestyle indicators, such as good quality sleep (Sexton, Storsve, Walhovd, Johansen-Berg, & Fjell, 2014) and a clear purpose in life (Ryff, Heller, Schaefer, van Reekum, & Davidson, 2016), have positive effects on brain health and ageing. Sedentary behaviour on the otherhand, which includes both screen time and total sitting time, is associated with increased mortality risk (Stamatakis et al., 2015) and lower cognitive performance over the lifespan (Falck, Davis, & Liu-Ambrose, 2017). A healthy diet contributes significantly to a healthy lifestyle and is already known to reduce the risk of stroke (Rutten-Jacobs et al., 2018), a condition associated with doubling the risk of dementia (Kuźma et al., 2018). Lifestyle interventions are appealing as potential prevention modalities for dementia as they can influence risk factors (Peters et al., 2019) and medical conditions through multiple pathways. Genes are thought to play a much smaller role (Livingston et al., 2017).

Presently, nine specific risk factors amenable to lifestyle change are estimated to collectively contribute to 35% of dementia cases (Livingston et al., 2017) (see Table 1.1). It is known that diet can significantly modulate at least four of these risk factors, namely hypertension, obesity, depression and diabetes. Urgent, well-designed, clinical trials are therefore required to test the effect of healthy diets on cognition, brain morphology and function, including the ‘traditional’ Mediterranean diet.

However, there is inconsistency in the definition of the Mediterranean diet (see Chapter Two) influencing the quality of the interventions used in clinical trials (Villani, Sultana, Doecke, & Mantzioris, 2018). Consequently, experts in the field have recognised that “the very definition of a Mediterranean dietary pattern is not a minor issue” (Martínez- González & Sánchez-Villegas, 2004). Further, there is substantial heterogeneity in the tools and scoring cut-off points used to measure adherence to a Mediterranean dietary pattern, and these tools do not necessarily reflect the ‘traditional’ cuisine (see Chapter Four). This is impeding comparison of data across studies (Galbete, Schwingshackl, Schwedhelm, Boeing, & Schulze, 2018).

8 Table 1.1. Modifiable risk factors for dementia, which in combination contribute to approximately 35% of dementia cases (Livingston et al., 2017)

Risk factor

1 No education beyond age 11-12 years or primary school only

2 mid-life hypertension

3 mid-life obesity

4 hearing loss

5 late-life depression

6 diabetes

7 physical inactivity

8 smoking

9 social isolation

1.3 Measurement of Mediterranean diet adherence

The ‘traditional’ Mediterranean diet within the literature represents the diets consumed in Southern Italy and Crete, Greece, in the late 1950s and early 1960s at the time of the Seven Countries Study (SCS) (see Chapter Two). Indeed, the ‘Mediterranean diet’ terminology in common use today was introduced as a result of the SCS (Hatzis, Sifaki- Pistolla, & Kafatos, 2015). Yet no previous studies have fully examined the effects of this ‘traditional’ Mediterranean diet (Trichopoulou et al., 2014), except for an Australian trial of people with type 2 diabetes, which utilised a fully reconstructed ‘traditional’ Cretan Mediterranean diet (Itsiopoulos et al., 2011). This archetypal ‘traditional’ diet may share similarities and also vary somewhat with diets from other Mediterranean countries during traditional and other times. Accurate measurement of adherence to the ‘traditional’ Mediterranean dietary pattern, including aspects of cuisine, would help progress the field.

9 1.3.1 Options for dietary assessment methods

Various options exist to assess adherence to a given diet or dietary pattern. The best method will depend on the purpose of study and degree of accuracy and precision required (Tapsell, Flood, Probst, Charlton, & Williams, 2013) and international guides are available to facilitate appropriate choice (NHS National Institute for Health Research and Medical Research Council, n.d.; NIH National Cancer Institute, n.d.). While food frequency questionnaires (FFQs) retrospectively assess usual diet (Thompson & Subar, 2017), they can be burdensome for participants since they typically include between 100 and 350 elements (Cade, Thompson, Burley, & Warm, 2002; Calfas, Zabinski, & Rupp, 2000). They also usually lack cuisine details, such as cooking methods and snacking habits. Food records (FR) prospectively assess actual intake (estimated or weighed) and can be used to determine dietary patterns but their data relates to a limited number of days, which may not represent usual diet depending on the days selected, chosen meals while recording and recording skills of participants (Thompson & Subar, 2017). Food records are also known to increase both participant and researcher burden and they may introduce optimistic biases (Miles & Scaife, 2003) and social desirability bias (Hebert, Clemow, Pbert, Ockene, & Ockene, 1995). Hence, FRs are utilised more in clinical trials rather than observational studies, and within cohorts that are motivated and literate. The 24-hour dietary recall method may be less burdensome for participants but it also only provides a snapshot of food intake from the previous day. Therefore, multiple recalls are usually undertaken and this method is mostly used within research to assess the diet of a population (Thompson & Subar, 2017).

In recent times, Short Question Surveys have become a popular way to derive an index or score for a Mediterranean dietary pattern (see Chapter Four). These tools are relatively simple to use and also have reduced burden for the researcher in terms of cost and analysis. They are developed using an a priori approach, with the included dietary elements usually informed by historical food intake data using a theoretical framework (Burggraf, Teuber, Brosig, & Meier, 2018). However, there are limitations of dietary patterns determined a priori, due to the subjectivity in selecting the elements for the index tool and defining their cut-off points. In contrast, dietary patterns can also be derived a

10 posteriori, using data from existing FFQs, FRs or 24-hour dietary recalls. In this instance, they are elicited through statistical methods that aggregate dietary elements into factors or clusters to reveal common underlying patterns e.g., Principal Component Analysis (PCA), K-Means (Newby & Tucker, 2004). However, these type of dietary patterns are often not reproducible since they are derived for a specific population (Burggraf et al., 2018). They are also not immune from inherent subjectivity as the methods used and the treatment of variables can vary across studies affecting the number and types of patterns derived (Newby & Tucker, 2004). Therefore, for future studies of the ‘traditional’ Mediterranean dietary pattern and aspects of cuisine, the use of a priori scores derived directly from diet index tools may offer various advantages and better reflect the ‘traditional’ Mediterranean diet (D'Alessandro & De Pergola, 2015). For example, when dietary indexes first started to gain momentum in research, an analysis of the famous Nurses’ Health Study found that the composite diet index score more strongly predicted risk of coronary heart disease in women than the major dietary patterns derived from factor analysis (Stampfer, Hu, Manson, Rimm, & Willett, 2000).

1.3.2 Statistical analysis for reliability and validity

Mediterranean diet index tools should be tested for reliability and validity. Reliability refers to the extent a tool provides the same/similar score when administered under similar circumstances whereas validity refers to whether the tool accurately measures what it is supposed to measure (Gleason, Harris, Sheean, Boushey, & Bruemmer, 2010). While different measures of reliability exist, test-retest reliability (within the same participant on two occasions) appears particularly important, and such analyses are currently lacking for most existing Mediterranean diet index tools (Zaragoza-Martí, Cabañero-Martínez, Hurtado-Sánchez, Laguna-Pérez, & Ferrer-Cascales, 2018). In the past, correlation coefficients, such as Pearson’s correlation coefficient used for continuous variables, were most frequently employed to describe the strength of the relationship between repeated measures from the same tool. However, since a high correlation does not necessarily mean good agreement, the Bland-Altman method was proposed more than 30 years ago as a more appropriate measure (Bland & Altman, 1986; Peat, 2001). Paired t-tests can be utilised to describe agreement between mean scores of two measures derived from the

11 same tool (Atkinson & Nevill, 1998; Margetts & Nelson, 1997). The intraclass correlation coefficient (ICC) is another useful test for reliability of a continuous score because it reflects both the degree of correlation and agreement between two measures (Koo & Li, 2016). Cohen’s kappa coefficient can be a reliability indicator for comparing percent agreement between categorical variables, such as food groups, derived from a diet index score, as it summarises cross classification with repeated measures (Peat, 2001).

With respect to the validity of a tool, different constructs have been tested in research, such as absolute validity (also known as criterion validity), content validity, convergent validity, discriminant validity, external validity, face validity, internal validity, predictive validity and relative validity (also known as concurrent validity) (Gleason et al., 2010). While the highest form is absolute validity, absolute or objective measures are lacking for many research areas, especially nutrition and total diets or dietary patterns (Margetts & Nelson, 1997; Streiner, Norman, & Cairney, 2015). Therefore, it is important that concurrent validity is established for a Mediterranean diet index tool, especially against a dietary reference method that is generally accepted as less biased and more accurate e.g., food record (Thompson & Subar, 2017). Convergent validity and discriminant validity are subtypes of concurrent validity that measure expected agreement or discrimination between a tool and other methods of assessment (Gleason et al., 2010). Face validity and content validity measure whether the tool captures the concept it is intending to measure and if it captures all the dimensions present in the concept, respectively. However, these are both subjective methods. The internal validity of a tool refers to how well the tool captures the intended concept among the population within which it was tested. External validity relates to whether the tool is valid for use across broader populations (Gleason et al., 2010) and this is simply described in text rather than measured using different statistical methods (Peat, 2001). Predictive validity refers to whether the tool can be associated with health and other outcomes in a predictive capacity (Panagiotakos, Pitsavos, Arvaniti, & Stefanadis, 2007).

12 For validity testing of diet index tools, which are used in medical and health sciences research, the Bland-Altman method mentioned above is generally preferred as it enables a graphical presentation of how well two quantitative methods agree with each other (Bland & Altman, 1999; Margetts & Nelson, 1997). The difference in the scores derived from the two methods is plotted against the means of the scores, with limits of agreement (LOA) shown as two standard deviations above, and below, the mean difference. The mean difference tells the researcher if the new tool is tending to under or over-report the score compared to the reference method and is referred to as the bias. The LOA show where 95% of the values for the mean difference in scores would be expected to fall. Linear regression is also used to indicate the direction of the bias and whether this is constant across the mean scores of the two measures. However, while the Bland-Altman plot quantifies the difference between two methods it does not provide a single statistic to confirm or deny a tool is valid, requiring interpretation by the researcher as to whether acceptable agreement exists. Importantly, it can identify whether any disagreement appears to be systematic (e.g., all values of one tool higher or lower than another), or non- systematic (e.g., degree of disagreement increases or decreases according to the absolute value of the construct in question). Indirect validity of a diet index tool can also be supported by analysis of expected Mediterranean nutrient trends (from the reference method) across quantiles of the tool score, using parametric or non-parametric tests, as required e.g., one-way ANOVA and Kruskal-Wallis, respectively.

Finally, it is important to stress that all dietary assessment methods are subject to measurement error (Kant, 2004) and the use of various statistical tests can strengthen the reliability and validity findings for a given tool. To date, few Mediterranean diet index tools have fulfilled quality criteria for both test-retest reliability and concurrent validity (Zaragoza-Martí et al., 2018), especially against a dietary reference method (see Chapter Four and Appendix Five). The available tools also have additional limitations with respect to the measurement of adherence to the actual ‘traditional’ Mediterranean dietary pattern.

13 1.3.3 Limitations of existing Mediterranean diet index tools

Heterogeneity exists among the currently available Mediterranean diet index tools with respect to the elements assessed, cut-off points used and the points awarded for scoring (see Chapter Four), limiting data synthesis and meta-analyses (Li et al., 2018). Most studies reporting health outcomes have used the Mediterranean Diet Score (MDS), which was originally developed and revised for a Greek population (Trichopoulou et al., 1995; Trichopoulou et al., 2003). While the elements in this tool were based on the ‘traditional’ diet, the cut-off points for scoring related to medians of this Mediterranean population. Medians in other populations, especially from Western countries, may not reflect ‘traditional’ Mediterranean intakes.

Furthermore, most available tools do not attempt to measure various aspects of cuisine, some of which are uniquely Mediterranean and may also contribute to health outcomes. These aspects include characterising foods (e.g., extra virgin olive oil (EVOO)), food combinations (e.g., sofrito, a tomato, onion/garlic and EVOO based sauce used as a base for many savoury dishes) and drinks (i.e., pure water); the lower temperature, moist, cooking methods traditionally used; and various habits related to eating (e.g., infrequent snacking occasions).

It is also of considerable concern that relatively few Mediterranean diet index tools (Hebestreit et al., 2017; Schroder et al., 2011; Sotos-Prieto et al., 2015) have been validated against a reference dietary method, particularly one that is not limited by the same recall bias as the tool being tested, such as a food record mentioned above. In addition, most Mediterranean diet scores reported have been indirectly derived from a food frequency questionnaire (FFQ) containing the required information for scoring, which also may or may not have been validated (see Chapter Four). However, repeatability coefficients are higher for a Mediterranean diet score calculated directly by administering the index tool itself (Bountziouka et al., 2012). Hence, existing tools have various limitations and do not necessarily measure adherence to the ‘traditional’ Mediterranean dietary pattern and aspects of cuisine.

14 1.3.4 Discernment for the ‘traditional’ diet

The ‘traditional’ Mediterranean dietary pattern should be discernible from other dietary patterns. However, most reported Mediterranean diet scores may simply be reflecting adherence to a more prudent, plant-based dietary pattern or measuring adherence to different types of Mediterranean diets, including modernised ones. For example, in an analysis of data from the Nurses’ Health Study II, increased adherence to three healthful dietary patterns, including the Mediterranean diet and the DASH diet, has been associated with a lower risk of hearing loss (Curhan, Wang, Eavey, Stampfer, & Curhan, 2018). Similarly, in another prospective study, high adherence to either the Mediterranean diet or MIND diet (a hybrid of the Mediterranean and DASH diets) was associated with reduced risk of incident AD (Morris et al., 2015). While overall diet quality may be most important, and explain the association of multiple healthy dietary patterns with improved health outcomes, greater discernment is required for the ‘traditional’ Mediterranean dietary pattern if studies are to claim this as their intervention or the dietary pattern followed.

Therefore, future research studying the effects of the ‘traditional’ Mediterranean diet requires a robust tool that includes both elements and scoring cut-off points determined a priori, based on empirical data. Such a tool should also be validated for use in Western populations, including consideration of the method for tool administration.

1.4 Gaps in research

With respect to the intent of this thesis, there is limited clinical research considering the effect of a Mediterranean diet on cognition or brain morphology and function. While systematic reviews on observational studies are plentiful, prior to research conducted in this thesis there was no systematic review focussed exclusively on clinical trials, which also included a focus on the quality of interventions and how these relate to the ‘traditional’ dietary pattern and aspects of cuisine (see Chapter Three).

15 Further, no publicly available tool has been specifically developed for, nor validated within, individuals at higher risk of cognitive decline, such as those diagnosed with MCI (see Chapter Six). Mild cognitive impairment may be a window of opportunity for interventions that might benefit the lives of carers, families and sufferers, since approximately 12% of people with MCI convert to AD per year, compared to an annual conversion rate of 1-2% in the general population (Petersen et al., 1999).

Finally, no Mediterranean diet adherence tool has been validated for use in an Australian population (representing a Western cohort), for online (see Chapter Five) and in-person (see Chapter Six) administration, which could broaden its utility for observational and clinical trials with potential cost savings to research budgets, as well as allowing it to be used in national and international population-based studies and surveys.

Addressing these gaps is a research priority as the ‘traditional’ Mediterranean diet holds promise to promote health (including brain health) and simultaneously reduce the risk of multiple chronic diseases, which increase with ageing. It is also a sustainable eating pattern, which is an important consideration for feeding the planet in the future (Dernini & Berry, 2015; Willett et al., 2019).

1.5 Thesis aims

The health benefits of the Mediterranean diet continue to be established, especially in the area of cognitive health. However, the existing definitions and measurement tools miss important ‘traditional’ cuisine elements, which could be incorporated into valid and reliable instruments for use in future clinical and observational studies. Therefore, it was hypothesised that a new measurement tool of Mediterranean diet and cuisine could be developed, drawing on the literature and dietary assessment methodology, and that the new tool would be able to be validated and tested for reliability among populations at risk of cognitive decline, suggesting a potential use for future clinical and observational studies.

16 Hence, the specific aims of this thesis were to:

1. Define the ‘traditional’ Mediterranean diet and cuisine from empirical data (see Chapter Two); 2. Investigate the relationship in clinical trials between a Mediterranean diet and cognition and brain morphology and function with a particular focus on the quality of interventions used and how they relate to the ‘traditional’ dietary pattern (see Chapter Three); 3. Develop a new Mediterranean diet index tool to assess adherence to the ‘traditional’ dietary pattern and aspects of cuisine, suitable for use within Western populations (see Chapter Four); 4. Test the reliability and validity of the new Mediterranean diet index tool, a) administered online to middle aged and older people with risk factors for cognitive decline, but no current diagnosis of dementia or neurodegenerative disease (see Chapter Five), and b) administered in-person to older individuals diagnosed with MCI (see Chapter Six).

The goal of the final chapter of this thesis (see Chapter Seven) was to sum up the thesis findings and discuss their implications, and to delineate remaining research gaps.

17 1.6 References

Alonso, A., Jacobs, D. R., Menotti, A., Nissinen, A., Dontas, A., Kafatos, A., & Kromhout, D. (2009). Cardiovascular risk factors and dementia mortality: 40 years of follow-up in the Seven Countries Study. Journal of the Neurological Sciences, 280(1), 79-83. doi:10.1016/j.jns.2009.02.004 Anderton, B. H. (1997). Changes in the ageing brain in health and disease. Philosophical Transactions of the Royal Society B: Biological Sciences, 352(1363), 1781-1792. doi:10.1098/rstb.1997.0162 Atkinson, G., & Nevill, A. M. (1998). Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Medicine, 26(4), 217-238. doi:10.2165/00007256-199826040-00002 Australian Institute of Health and Welfare. (2012). Dementia in Australia. Retrieved 21st December 2018 from https://www.aihw.gov.au/reports/dementia/dementia- in-australia/data Australian Institute of Health and Welfare. (2018). Deaths in Australia. Retrieved 2nd October 2018 from https://www.aihw.gov.au/reports/life-expectancy- death/deaths-in-australia/contents/leading-causes-of-death Bach-Faig, A., Berry, E. M., Lairon, D., Reguant, J., Trichopoulou, A., Dernini, S., . . . Mediterranean Diet Foundation Expert Group. (2011). Mediterranean diet pyramid today: Science and cultural updates. Public Health Nutrition, 14(12A), 2274-2284. doi:10.1017/S1368980011002515 Barré, A., Gusto, G., Cadeau, C., Carbonnel, F., & Boutron-Ruault, M.-C. (2017). Diet and risk of cholecystectomy: A prospective study based on the French E3N cohort. The American Journal of Gastroenterology, 112(9), 1448-1456. doi:10.1038/ajg.2017.216 Benetou, V., Orfanos, P., Pettersson-Kymmer, U., Bergstrom, U., Svensson, O., Johansson, I., . . . Trichopoulou, A. (2013). Mediterranean diet and incidence of hip fractures in a European cohort. Osteoporosis International, 24(5), 1587- 1598. doi:10.1007/s00198-012-2187-3

18 Benetou, V., Orfanos, P., Feskanich, D., Michaëlsson, K., Pettersson-Kymmer, U., Byberg, L., . . . Trichopoulou, A. (2018). Mediterranean diet and hip fracture incidence among older adults: The CHANCES project. Osteoporosis International, 29(7), 1591-1599. doi:10.1007/s00198-018-4517-6 Berti, V., Walters, M., Sterling, J., Quinn, C. G., Logue, M., Andrews, R., . . . Mosconi, L. (2018). Mediterranean diet and 3-year Alzheimer brain biomarker changes in middle-aged adults. Neurology, 90(20), e1789-e1798. doi:10.1212/WNL.0000000000005527 Bland, J. M., & Altman, D. G. (1986). Statistical methods for assessing agreement between two methods of clinical measurement. The Lancet, 327(8476), 307-310. doi:10.1016/S0140-6736(86)90837-8 Bland, J. M., & Altman, D. G. (1999). Measuring agreement in method comparison studies. Statistical Methods in Medical Research, 8(2), 135-160. doi:10.1177/096228029900800204 Boccardi, V., Esposito, A., Rizzo, M. R., Marfella, R., Barbieri, M., & Paolisso, G. (2013). Mediterranean diet, telomere maintenance and health status among elderly. PLoS ONE, 8(4), e62781. doi:10.1371/journal.pone.0062781 Bonaccio, M., Di Castelnuovo, A., Bonanni, A., Costanzo, S., De Lucia, F., Pounis, G., . . . Iacoviello, L. (2013). Adherence to a Mediterranean diet is associated with a better health-related quality of life: A possible role of high dietary antioxidant content. BMJ Open, 3(8). doi:10.1136/bmjopen-2013-003003 Bonaccio, M., Di Castelnuovo, A., Costanzo, S., Gialluisi, A., Persichillo, M., Cerletti, C., . . . Iacoviello, L. (2018). Mediterranean diet and mortality in the elderly: A prospective cohort study and a meta-analysis. The British Journal of Nutrition, 120(8), 841-854. doi:10.1017/S0007114518002179 Bongaarts, J. (2009). Human population growth and the demographic transition. Philosophical Transactions of the Royal Society B: Biological Sciences, 364(1532), 2985-2990. doi:10.1098/rstb.2009.0137 Bountziouka, V., Bathrellou, E., Zazpe, I., Ezquer, L., Martínez-González, M.-A., & Panagiotakos, D. B. (2012). Repeatability of food frequency assessment tools in relation to the number of items and response categories included. Food and Nutrition Bulletin, 33(4), 288-295. doi:10.1177/156482651203300409

19 Brookmeyer, R., Johnson, E., Ziegler-Graham, K., & Arrighi, H. M. (2007). Forecasting the global burden of Alzheimer's disease. Alzheimer's & Dementia, 3(3), 186- 191. doi:10.1016/j.jalz.2007.04.381 Brookmeyer, R., Abdalla, N., Kawas, C. H., & Corrada, M. M. (2018). Forecasting the prevalence of preclinical and clinical Alzheimer's disease in the United States. Alzheimer's & Dementia, 14(2), 121-129. doi:10.1016/j.jalz.2017.10.009 Buckland, G., Bach, A., & Serra‐Majem, L. (2008). Obesity and the Mediterranean diet: A systematic review of observational and intervention studies. Obesity Reviews, 9(6), 582-593. doi:10.1111/j.1467-789X.2008.00503.x Buckley, J. S., & Salpeter, S. R. (2015). A risk-benefit assessment of dementia medications: Systematic review of the evidence. Drugs and Aging, 32(6), 453- 467. doi:10.1007/s40266-015-0266-9 Burggraf, C., Teuber, R., Brosig, S., & Meier, T. (2018). Review of a priori dietary quality indices in relation to their construction criteria. Nutrition Reviews, 76(10), 747-764. doi:10.1093/nutrit/nuy027 Cade, J., Thompson, R., Burley, V., & Warm, D. (2002). Development, validation and utilisation of food-frequency questionnaires – A review. Public Health Nutrition, 5(4), 567-587. doi:10.1079/PHN2001318 Calfas, K. J., Zabinski, M. F., & Rupp, J. (2000). Practical nutrition assessment in primary care settings. American Journal of Preventive Medicine, 18(4), 289-299. doi:10.1016/S0749-3797(00)00116-1 Casey, D. A., Antimisiaris, D., & O’Brien, J. (2010). Drugs for Alzheimer’s disease: Are they effective? Pharmacy and Therapeutics. Retrieved 22nd December 2018 from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2873716/ Cooper, C., Li, R., Lyketsos, C., & Livingston, G. (2013). Treatment for mild cognitive impairment: Systematic review. British Journal of Psychiatry, 203(4), 255-264. doi:10.1192/bjp.bp.113.127811 Crous-Bou, M., Fung, T. T., Prescott, J., Julin, B., Du, M., Sun, Q., . . . De Vivo, I. (2014). Mediterranean diet and telomere length in Nurses’ Health Study: Population based cohort study. British Medical Journal, 349, g6674. doi:10.1136/bmj.g6674

20 Cueto-Galán, R., Barón, F., Valdivielso, P., Pintó, X., Corbella, E., Gómez-Gracia, E., & Wärnberg, J. (2017). Changes in fatty liver index after consuming a Mediterranean diet: 6-year follow-up of the PREDIMED-Malaga trial. Medicina Clínica (English Edition), 148(10), 435-443. doi:10.1016/j.medcle.2017.04.030 Cummings, J. L., Morstorf, T., & Zhong, K. (2014). Alzheimer's disease drug- development pipeline: Few candidates, frequent failures. Alzheimer's Research and Therapy, 6(4), 37-37. doi:10.1186/alzrt269 Curhan, S. G., Wang, M., Eavey, R. D., Stampfer, M. J., & Curhan, G. C. (2018). Adherence to healthful dietary patterns is associated with lower risk of hearing loss in women. The Journal of Nutrition, 148(6), 944-951. doi:10.1093/jn/nxy058 D'Alessandro, A., & De Pergola, G. (2015). Mediterranean diet and cardiovascular disease: A critical evaluation of a priori dietary indexes. Nutrients, 7(9), 7863- 7888. doi:10.3390/nu7095367 Davies, M. J., D'Alessio, D. A., Fradkin, J., Kernan, W. N., Mathieu, C., Mingrone, G., . . . Buse, J. B. (2018). Management of hyperglycemia in type 2 diabetes, 2018. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetes Care, 41(12), 2669-2701. doi:10.2337/dci18-0033 Davis, C. R., Hodgson, J. M., Woodman, R., Bryan, J., Wilson, C., & Murphy, K. J. (2017). A Mediterranean diet lowers blood pressure and improves endothelial function: Results from the MedLey randomized intervention trial. American Journal of Clinical Nutrition, 105(6), 1305-1313. doi:10.3945/ajcn.116.146803 De Filippis, F., Pellegrini, N., Vannini, L., Jeffery, I. B., La Storia, A., Laghi, L., . . . Ercolini, D. (2016). High-level adherence to a Mediterranean diet beneficially impacts the gut microbiota and associated metabolome. Gut, 65(11), 1812-1821. doi:10.1136/gutjnl-2015-309957 De Lorenzo, A., Alberti, A., Andreoli, A., Iacopino, L., Serrano, P., & Perriello, G. (2001). Food habits in a southern Italian town (Nicotera) in 1960 and 1996: Still a reference Italian Mediterranean diet? Diabetes Nutrition and Metabolism, 14(3), 121-125.

21 de Lorgeril, M., Renaud, S., Salen, P., Monjaud, I., Mamelle, N., Martin, J. L., . . . Delaye, J. (1994). Mediterranean alpha-linolenic acid-rich diet in secondary prevention of coronary heart disease. Lancet, 343(8911), 1454-1459. doi:10.1016/S0140-6736(94)92580-1 de Lorgeril, M. (2013). Mediterranean diet and cardiovascular disease: historical perspective and latest evidence. Current Atherosclerosis Reports, 15(12), Art. No.:370. doi:10.1007/s11883-013-0370-4 Dernini, S., & Berry, E. M. (2015). Mediterranean Diet: From a healthy diet to a sustainable dietary pattern. Frontiers in Nutrition, 2, 15. doi:10.3389/fnut.2015.00015 Esposito, K., Maiorino, M. I., Bellastella, G., Chiodini, P., Panagiotakos, D., & Giugliano, D. (2015). A journey into a Mediterranean diet and type 2 diabetes: A systematic review with meta-analyses. BMJ Open, 5(8), e008222. doi:10.1136/bmjopen-2015-008222 Estruch, R., Martínez-González, M. A., Corella, D., Salas-Salvadó, J., Fitó, M., Chiva- Blanch, G., . . . Ros, E. (2016). Effect of a high-fat Mediterranean diet on bodyweight and waist circumference: A prespecified secondary outcomes analysis of the PREDIMED randomised controlled trial. The Lancet Diabetes & Endocrinology, 4(8), 666-676. doi:10.1016/S2213-8587(16)30085-7 Estruch, R., Ros, E., Salas-Salvado, J., Covas, M. I., Corella, D., Aros, F., . . . PREDIMED Study Investigators. (2018). Primary prevention of cardiovascular disease with a Mediterranean diet supplemented with extra-virgin olive oil or nuts. New England Journal of Medicine, 378(25), e34. doi:10.1056/NEJMoa1800389 Falck, R. S., Davis, J. C., & Liu-Ambrose, T. (2017). What is the association between sedentary behaviour and cognitive function? A systematic review. British Journal of Sports Medicine, 51(10), 800-811. doi:10.1136/bjsports-2015-095551 Fardet, A., & Rock, E. (2014). Toward a new philosophy of preventive nutrition: From a reductionist to a holistic paradigm to improve nutritional recommendations. Advances in Nutrition, 5(4), 430-446. doi:10.3945/an.114.006122 Fidanza, F., & Alberti, A. (2005). The healthy Italian Mediterranean diet temple food guide. Nutrition Today, 40(2), 71-78. doi:10.1097/00017285-200503000-00005

22 Fletcher, E., Gavett, B., Harvey, D., Farias, S. T., Olichney, J., Beckett, L., . . . Mungas, D. (2018). Brain volume change and cognitive trajectories in aging Neuropsychology, 32(4), 436-449. doi:10.1037/neu0000447 Freeland-Graves, J. H., & Nitzke, S. (2013). Position of the Academy of Nutrition and Dietetics: Total diet approach to healthy eating. Journal of the Academy of Nutrition and Dietetics, 113(2), 307-317. doi:10.1016/j.jand.2012.12.013 Fung, T., Rexrode, K., Mantzoros, C., Manson, J., Willett, W., & Hu, F. (2009). Mediterranean diet and incidence and mortality of coronary heart disease and stroke in women. The FASEB Journal, 23(S1). Galbete, C., Schwingshackl, L., Schwedhelm, C., Boeing, H., & Schulze, M. B. (2018). Evaluating Mediterranean diet and risk of chronic disease in cohort studies: An umbrella review of meta-analyses. European Journal of Epidemiology, 33(10), 909-931. doi:10.1007/s10654-018-0427-3 Gleason, P. M., Harris, J., Sheean, P. M., Boushey, C. J., & Bruemmer, B. (2010). Publishing nutrition research: Validity, reliability, and diagnostic test assessment in nutrition-related research. Journal of the American Dietetic Association, 110, 409-419. doi:10.1016/j.jada.2009.11.022 Govindaraju, T., Sahle, B. W., McCaffrey, T. A., McNeil, J. J., & Owen, A. J. (2018). Dietary patterns and quality of life in older adults: A systematic review. Nutrients, 10(8). doi:10.3390/nu10080971 Grosso, G., Marventano, S., Yang, J., Micek, A., Pajak, A., Scalfi, L., . . . Kales, S. N. (2017). A comprehensive meta-analysis on evidence of Mediterranean diet and cardiovascular disease: Are individual components equal? Critical Reviews in Food Science and Nutrition, 57(15), 3218-3232. doi:10.1080/10408398.2015.1107021 Hardman, R. J., Kennedy, G., Macpherson, H., Scholey, A. B., & Pipingas, A. (2015). A randomised controlled trial investigating the effects of Mediterranean diet and aerobic exercise on cognition in cognitively healthy older people living independently within aged care facilities: The Lifestyle Intervention in Independent Living Aged Care (LIILAC) study protocol [ACTRN12614001133628]. Nutrition Journal, 14, 53. doi:10.1186/s12937-015- 0042-z

23 Hatzis, C. M., Papandreou, C., Patelarou, E., Vardavas, C. I., Kimioni, E., Sifaki- Pistolla, D., . . . Kafatos, A. G. (2013). A 50-year follow-up of the Seven Countries Study: Prevalence of cardiovascular risk factors, food and nutrient intakes among Cretans. Hormones, 12(3), 379-385. doi:10.1007/BF03401303 Hatzis, C. M., Sifaki-Pistolla, D., & Kafatos, A. G. (2015). History of the Cretan cohort of the Seven Countries Study. Hormones. Retrieved 21st December 2018 from http://www.hormones.gr/5/search.html?s=a+50-year+follow- up+of+the+Seven+countries+study&flds=all&do=search&rpp=20&x=0&y=0 Hawkes, N. (2018). Pfizer abandons research into Alzheimer’s and Parkinson’s diseases. BMJ (Clinical Research Edition), 360, k122. doi:10.1136/bmj.k122 Hebert, J. R., Clemow, L., Pbert, L., Ockene, I. S., & Ockene, J. K. (1995). Social desirability bias in dietary self-report may compromise the validity of dietary intake measures. International Journal of Epidemiology, 24(2), 389-398. doi:10.1093/ije/24.2.389 Hebestreit, K., Yahiaoui-Doktor, M., Engel, C., Vetter, W., Siniatchkin, M., Erickson, N., . . . Bischoff, S. C. (2017). Validation of the German version of the Mediterranean Diet Adherence Screener (MEDAS) questionnaire. BMC Cancer, 17, 341. doi:10.1186/s12885-017-3337-y Henríquez Sánchez, P., Ruano, C., de Irala, J., Ruiz-Canela, M., Martínez-González, M. A., & Sánchez-Villegas, A. (2012). Adherence to the Mediterranean diet and quality of life in the SUN Project. European Journal of Clinical Nutrition, 66, 360-368. doi:10.1038/ejcn.2011.146 Hogg, R. E., Woodside, J. V., McGrath, A., Young, I. S., Vioque, J. L., Chakravarthy, U., . . . Fletcher, A. E. (2016). Mediterranean Diet Score and its association with age-related macular degeneration. Ophthalmology, 124(1), 82-89. doi:10.1016/j.ophtha.2016.09.019 Huang, X., Jiménez-Moleón, J. J., Lindholm, B., Cederholm, T., Ärnlöv, J., Risérus, U., . . . Carrero, J. J. (2013). Mediterranean diet, kidney function, and mortality in men with CKD. Clinical Journal of the American Society of Nephrology, 8(9), 1548-1555. doi:10.2215/CJN.01780213

24 Itsiopoulos, C., Brazionis, L., Kaimakamis, M., Cameron, M., Best, J. D., O’Dea, K., & Rowley, K. (2011). Can the Mediterranean diet lower HbA1c in type 2 diabetes? Results from a randomized cross-over study. Nutrition, Metabolism and Cardiovascular Diseases, 21(9), 740-747. doi:10.1016/j.numecd.2010.03.005 Jacka, F. N., O'Neil, A., Opie, R., Itsiopoulos, C., Cotton, S., Mohebbi, M., . . . Berk, M. (2017). A randomised controlled trial of dietary improvement for adults with major depression (the 'SMILES' trial). BMC Medicine, 15(1), 23. doi:10.1186/s12916-017-0791-y Jacobs, D. R., & Tapsell, L. C. (2013). Food synergy: The key to a healthy diet. Proceedings of the Nutrition Society, 72(2), 200-206. doi:10.1017/S0029665112003011 Kant, A. K. (2004). Dietary patterns and health outcomes. Journal of the American Dietetic Association, 104(4), 615-635. doi:10.1016/j.jada.2004.01.010 Kastorini, C.-M., Milionis, H. J., Esposito, K., Giugliano, D., Goudevenos, J. A., & Panagiotakos, D. B. (2011). The effect of Mediterranean diet on metabolic syndrome and its components: A meta-analysis of 50 studies and 534,906 individuals. Journal of the American College of Cardiology, 57(11), 1299-1313. doi:10.1016/j.jacc.2010.09.073 Kennedy, B. K., Berger, S. L., Brunet, A., Campisi, J., Cuervo, A. M., Epel, E. S., . . . Sierra, F. (2014). Geroscience: Linking aging to chronic disease. Cell, 159(4), 709-713. doi:10.1016/j.cell.2014.10.039 Khatri, M., Moon, Y. P., Scarmeas, N., Gu, Y., Gardener, H., Cheung, K., . . . Elkind, M. S. (2014). The association between a Mediterranean-style diet and kidney function in the Northern Manhattan Study cohort. Clinical Journal of the American Society of Nephrology, 9(11), 1868-1875. doi:10.2215/CJN.01080114 Kojima, G., Avgerinou, C., Iliffe, S., & Walters, K. (2018). Adherence to Mediterranean diet reduces incident frailty risk: Systematic review and meta‐ analysis. Journal of the American Geriatrics Society, 66(4), 783-788. doi:10.1111/jgs.15251

25 Kontogianni, M. D., Tileli, N., Margariti, A., Georgoulis, M., Deutsch, M., Tiniakos, D., . . . Papatheodoridis, G. (2014). Adherence to the Mediterranean diet is associated with the severity of non-alcoholic fatty liver disease. Clinical Nutrition, 33(4), 678-683. doi:10.1016/j.clnu.2013.08.014 Koo, T. K., & Li, M. Y. (2016). A guideline of selecting and reporting intraclass correlation coefficients for reliability research. Journal of Chiropractic Medicine, 15(2), 155-163. doi:10.1016/j.jcm.2016.02.012 Kouris-Blazos, A., & Itsiopoulos, C. (2014). Low all-cause mortality despite high cardiovascular risk in elderly Greek-born Australians: Attenuating potential of diet? Asia Pacific Journal of Clinical Nutrition, 23(4), 532-544. doi:10.6133/apjcn.2014.23.4.16 Kuźma, E., Lourida, I., Moore, S. F., Levine, D. A., Ukoumunne, O. C., & Llewellyn, D. J. (2018). Stroke and dementia risk: A systematic review and meta-analysis. Alzheimer's & Dementia, 14(11), 1416-1426. doi:10.1016/j.jalz.2018.06.3061 Lassale, C., Batty, G. D., Baghdadli, A., Jacka, F., Sánchez-Villegas, A., Kivimäki, M., & Akbaraly, T. (2018). Healthy dietary indices and risk of depressive outcomes: A systematic review and meta-analysis of observational studies. Molecular Psychiatry, 1-22. doi:10.1038/s41380-018-0237-8 Li, Y., Hu, B.-Q., Wu, X.-J., Qi, X.-W., Jiang, J., Cui, X., . . . Yang, X.-H. (2018). Adherence to Mediterranean diet and the risk of breast cancer: A meta-analysis. Translational Cancer Research. Retrieved 21st December 2018 from http://tcr.amegroups.com/article/view/24830 Livingston, G., Sommerlad, A., Orgeta, V., Costafreda, S. G., Huntley, J., Ames, D., . . . Mukadam, N. (2017). Dementia prevention, intervention, and care. The Lancet, 390(10113), 2673-2734. doi:10.1016/S0140-6736(17)31363-6 Luciano, M., Corley, J., Cox, S. R., Valdés Hernández, M. C., Craig, L. C. A., Dickie, D. A., . . . Deary, I. J. (2017). Mediterranean-type diet and brain structural change from 73 to 76 years in a Scottish cohort. Neurology, 88(5), 449-455. doi:10.1212/WNL.0000000000003559

26 Maraki, M. I., Yannakoulia, M., Stamelou, M., Stefanis, L., Xiromerisiou, G., Kosmidis, M. H., . . . Scarmeas, N. (2019). Mediterranean diet adherence is related to reduced probability of prodromal Parkinson's disease. Movement Disorders, 34(1), 48-57. doi:10.1002/mds.27489 Margetts, B. M., & Nelson, M. D. (1997). Design concepts in nutritional epidemiology (2nd ed.). New York, USA: Oxford University Press. Martínez-González, M. A., & Sánchez-Villegas, A. (2004). The emerging role of Mediterranean diets in cardiovascular epidemiology: Monounsaturated fats, olive oil, red wine or the whole pattern? European Journal of Epidemiology. Retrieved 21st December 2018 from https://www.jstor.org/stable/3582542?seq=1#metadata_info_tab_contents Martínez-González, M. A., García-Arellano, A., Toledo, E., Salas-Salvado, J., Buil- Cosiales, P., Corella, D., . . . Gómez-Gracia, E. (2012). A 14-item Mediterranean diet assessment tool and obesity indexes among high-risk subjects: The PREDIMED trial. PLoS ONE, 7(8), e43134. doi:10.1371/journal.pone.0043134 Martínez-González, M. A., Gea, A., & Ruiz-Canela, M. (2019). The Mediterranean diet and cardiovascular health: A critical review. 124, 779-798. doi:10.1161/CIRCRESAHA.118.313348 McMaster, M., Kim, S., Clare, L., Torres, S. J., D'Este, C., & Anstey, K. J. (2018). Body, Brain, Life for Cognitive Decline (BBL-CD): Protocol for a multidomain dementia risk reduction randomized controlled trial for subjective cognitive decline and mild cognitive impairment. Clinical Interventions in Aging, 13, 2397-2406. doi:10.2147/CIA.S182046 Merle, B. M. J., Silver, R. E., Rosner, B., & Seddon, J. M. (2015). Adherence to a Mediterranean diet, genetic susceptibility, and progression to advanced macular degeneration: A prospective cohort study. American Journal of Clinical Nutrition, 102(5), 1196-1206. doi:10.3945/ajcn.115.111047 Miles, S., & Scaife, V. (2003). Optimistic bias and food. Nutrition Research Reviews, 16(1), 3-19. doi:10.1079/NRR200249

27 Ministry of Health and Welfare: Supreme Scientific Health Council. (1999). Dietary guidelines for adults in Greece. Archives of Hellenic Medicine. Retrieved 19th December 2018 from http://www.mednet.gr/archives/1999-5/pdf/516.pdf Morris, M. C., Tangney, C. C., Wang, Y., Sacks, F. M., Bennett, D. A., & Aggarwal, N. T. (2015). MIND diet associated with reduced incidence of Alzheimer's disease. Alzheimer's & Dementia, 11(9), 1007-1014. doi:10.1016/j.jalz.2014.11.009 National Health and Medical Research Council. (2013). Australian dietary guidelines. Canberra, Australia: National Health and Medical Research Council. National Institute for Health and Care Excellence. (2015). Dementia, disability and frailty in later life - mid-life approaches to delay or prevent onset. Retrieved 3rd October 2018 from https://www.nice.org.uk/guidance/ng16 Newby, P. K., & Tucker, K. L. (2004). Empirically derived eating patterns using factor or cluster analysis: A review. Nutrition Reviews, 62(5), 177-203. doi:10.1111/j.1753-4887.2004.tb00040.x NHS. (2017). What is a Mediterranean diet? Retrieved 21st December 2018 from https://www.nhs.uk/live-well/eat-well/what-is-a-mediterranean-diet/ NHS National Institute for Health Research and Medical Research Council. (n.d.). DAPA measurement toolkit. Retrieved 9th June 2019 from https://www.dapa- toolkit.mrc.ac.uk/ NIH National Cancer Institute. (n.d.). Dietary assessment primer. Retrieved 9th June 2019 from https://dietassessmentprimer.cancer.gov/ Oldways. (2008). Mediterranean diet pyramid. Retrieved 18th May 2018 from http://oldwayspt.org/resources/heritage-pyramids/mediterranean- pyramid/overview Panagiotakos, D. B., Chrysohoou, C., Pitsavos, C., & Stefanadis, C. (2006). Association between the prevalence of obesity and adherence to the Mediterranean diet: The ATTICA study. Nutrition, 22(5), 449-456. doi:10.1016/j.nut.2005.11.004 Panagiotakos, D. B., Pitsavos, C., Arvaniti, F., & Stefanadis, C. (2007). Adherence to the Mediterranean food pattern predicts the prevalence of hypertension, hypercholesterolemia, diabetes and obesity, among healthy adults; the accuracy of the MedDietScore. Preventive Medicine, 44(4), 335-340. doi:10.1016/j.ypmed.2006.12.009

28 Papamiltiadous, E. S., Roberts, S. K., Nicoll, A. J., Ryan, M. C., Itsiopoulos, C., Salim, A., & Tierney, A. C. (2016). A randomised controlled trial of a Mediterranean Dietary Intervention for Adults with Non Alcoholic Fatty Liver Disease (MEDINA): Study protocol. BMC Gastroenterology, 16(1), 14. doi:10.1186/s12876-016-0426-3 Parletta, N., Zarnowiecki, D., Cho, J., Wilson, A., Bogomolova, S., Villani, A., . . . O'Dea, K. (2017). A Mediterranean-style dietary intervention supplemented with fish oil improves diet quality and mental health in people with depression: A randomized controlled trial (HELFIMED). Nutritional Neuroscience, 1-14. doi:10.1080/1028415X.2017.1411320 Peat, J. (2001). Health science research: A handbook of quantitative methods (1st ed.). Crows Nest, Australia: Allen & Unwin. Pelletier, A., Barul, C., Féart, C., Helmer, C., Bernard, C., Periot, O., . . . Samieri, C. (2015). Mediterranean diet and preserved brain structural connectivity in older subjects. Alzheimer's & Dementia, 11(9), 1023-1031. doi:10.1016/j.jalz.2015.06.1888 Peters, R., Booth, A., Rockwood, K., Peters, J., D’este, C., & Anstey, K. J. (2019). Combining modifiable risk factors and risk of dementia: A systematic review and meta-analysis. BMJ Open, 9(1), e022846. doi:10.1136/bmjopen-2018- 022846 Petersen, R. C., Smith, G. E., Waring, S. C., Ivnik, R. J., Tangalos, E. G., & Kokmen, E. (1999). Mild cognitive impairment: Clinical characterization and outcome. Archives of Neurology, 56(3), 303-308. doi:10.1001/archneur.56.3.303 Pfizer. (2011). Health report: Dementia is everybody's business, Issue 45. Retrieved 2nd October 2018 from https://www.dementia.org.au/publications/pfizer-health- reports Prince, M., Wimo, A., Guerchet, M., Ali, G.-C., Wu, Y.-T., & Prina, M. (2015). World Alzheimer report 2015, The global impact of dementia: An analysis of prevalence, incidence, cost and trends. Retrieved 23rd November 2018 from https://www.alz.co.uk/research/world-report-2015

29 Psaltopoulou, T., Sergentanis, T. N., Panagiotakos, D. B., Sergentanis, I. N., Kosti, R., & Scarmeas, N. (2013). Mediterranean diet, stroke, cognitive impairment, and depression: A meta-analysis. Annals of Neurology, 74(4), 580-591. doi:10.1002/ana.23944 Roser, M., & Ortiz-Ospina, E. (2018). World population growth. Retrieved 10th October 2018 from https://ourworldindata.org/world-population-growth Rutjes, A. W. S., Denton, D. A., Di Nisio, M., Chong, L. Y., Abraham, R. P., Al‐Assaf, A. S., . . . McCleery, J. (2018). Vitamin and mineral supplementation for maintaining cognitive function in cognitively healthy people in mid and late life. Cochrane Database of Systematic Reviews(12), Art. No.: CD011906. doi:10.1002/14651858.CD011906.pub2 Rutten-Jacobs, L. C., Larsson, S. C., Malik, R., Rannikmäe, K., Sudlow, C. L., Dichgans, M., . . . Traylor, M. (2018). Genetic risk, incident stroke, and the benefits of adhering to a healthy lifestyle: Cohort study of 306 473 UK Biobank participants. BMJ (Clinical Research Edition), 363, k4168. doi:10.1136/bmj.k4168 Ryan, M. C., Itsiopoulos, C., Thodis, T., Ward, G., Trost, N., Hofferberth, S., . . . Wilson, A. M. (2013). The Mediterranean diet improves hepatic steatosis and insulin sensitivity in individuals with non-alcoholic fatty liver disease. Journal of Hepatology, 59(1), 138-143. doi:10.1016/j.jhep.2013.02.012 Ryff, C. D., Heller, A. S., Schaefer, S. M., van Reekum, C., & Davidson, R. J. (2016). Purposeful engagement, healthy aging, and the brain. Current Behavioral Neuroscience Reports, 3(4), 318-327. doi:10.1007/s40473-016-0096-z Salas-Salvadó, J., Bulló, M., Estruch, R., Ros, E., Covas, M.-I., Ibarrola-Jurado, N., . . . Martínez-González, M. A. (2014). Prevention of diabetes with Mediterranean diets: A subgroup analysis of a randomized trial. Annals of Internal Medicine, 160(1), 1-10. doi:10.7326/M13-1725 Scarmeas, N., Anastasiou, C. A., & Yannakoulia, M. (2018). Nutrition and prevention of cognitive impairment. The Lancet Neurology, 17(11), 1006-1015. doi:10.1016/S1474-4422(18)30338-7

30 Schroder, H., Fito, M., Estruch, R., Martinez-Gonzalez, M. A., Corella, D., Salas- Salvado, J., . . . Covas, M. I. (2011). A short screener is valid for assessing Mediterranean diet adherence among older Spanish men and women. Journal of Nutrition, 141(6), 1140-1145. doi:10.3945/jn.110.135566 Sexton, C. E., Storsve, A. B., Walhovd, K. B., Johansen-Berg, H., & Fjell, A. M. (2014). Poor sleep quality is associated with increased cortical atrophy in community-dwelling adults. Neurology, 83(11), 967-973. doi:10.1212/WNL.0000000000000774 Shively, C. A., Register, T. C., Appt, S. E., Clarkson, T. B., Uberseder, B., Clear, K. Y. J., . . . Cook, K. L. (2018). Consumption of Mediterranean versus western diet leads to distinct mammary gland microbiome populations. Cell Reports, 25(1), 47-56.e43. doi:10.1016/j.celrep.2018.08.078 Singh, B., Parsaik, A. K., Mielke, M. M., Erwin, P. J., Knopman, D. S., Petersen, R. C., & Roberts, R. O. (2014). Association of Mediterranean diet with mild cognitive impairment and Alzheimer's disease: A systematic review and meta-analysis. Journal of Alzheimer's Disease, 39(2), 271-282. doi:10.3233/Jad-130830 Sjögren, P., Becker, W., Warensjö, E., Olsson, E., Byberg, L., Gustafsson, I.-B., . . . Cederholm, T. (2010). Mediterranean and carbohydrate-restricted diets and mortality among elderly men: A cohort study in Sweden. American Journal of Clinical Nutrition, 92(4), 967-974. doi:10.3945/ajcn.2010.29345 Sofi, F., Cesari, F., Abbate, R., Gensini, G. F., & Casini, A. (2008). Adherence to Mediterranean diet and health status: Meta-analysis. BMJ (Clinical Research Edition), 337, a1344. doi:10.1136/bmj.a1344 Sofi, F., Macchi, C., Abbate, R., Gensini, G. F., & Casini, A. (2014). Mediterranean diet and health status: An updated meta-analysis and a proposal for a literature-based adherence score. Public Health Nutrition, 17(12), 2769-2782. doi:10.1017/S1368980013003169 Sotos-Prieto, M., Santos-Beneit, G., Bodega, P., Pocock, S., Mattei, J., & Penalvo, J. L. (2015). Validation of a questionnaire to measure overall Mediterranean lifestyle habits for research application: The MEDiterranean LIFEstyle index (MEDLIFE). Nutricion Hospitalaria, 32(3), 1153-1163. doi:10.3305/nh.2015.32.3.9387

31 Sperling, R. A., Aisen, P. S., Beckett, L. A., Bennett, D. A., Craft, S., Fagan, A. M., . . . Phelps, C. H. (2011). Toward defining the preclinical stages of Alzheimer's disease: Recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimer's and Dementia, 7(3), 280-292. doi:10.1016/j.jalz.2011.03.003 Stamatakis, E., Rogers, K., Ding, D., Berrigan, D., Chau, J., Hamer, M., & Bauman, A. (2015). All-cause mortality effects of replacing sedentary time with physical activity and sleeping using an isotemporal substitution model: A prospective study of 201,129 mid-aged and older adults. The International Journal of Behavioral Nutrition and Physical Activity, 12(121). doi:10.1186/s12966-015- 0280-7 Stampfer, M. J., Hu, F. B., Manson, J. E., Rimm, E. B., & Willett, W. C. (2000). Primary prevention of coronary heart disease in women through diet and lifestyle. The New England Journal of Medicine, 343(1), 16-22. doi:10.1056/NEJM200007063430103 Streiner, D. L., Norman, G. R., & Cairney, J. (2015). Health measurement scales: A practical guide to their development and use (5th ed.). Oxford, United Kingdom: Oxford University Press. Struijk, E. A., Guallar-Castillón, P., Rodríguez-Artalejo, F., & López-García, E. (2018). Mediterranean dietary patterns and impaired physical function in older adults. The Journals of Gerontology. Series A, 73(3), 333-339. doi:10.1093/gerona/glw208 Tapsell, L., Flood, V., Probst, Y., Charlton, K., & Williams, P. (2013). Nutrition tools: Dietary assessment, food databases and dietary modelling. In L. Tapsell (Ed.), Food, nutrition and health (pp. 282-320). South Melbourne, Australia: Oxford University Press. The Royal Australian College of General Practitioners. (2014). HANDI: Mediterranean diet: Reducing cardiovascular risk. Retrieved 19th June 2017 from http://www.racgp.org.au/your- practice/guidelines/handi/interventions/nutrition/the-mediterranean-diet-for- reducing-cardiovascular-disease-risk/

32 Thompson, F. E., & Subar, A. F. (2017). Chapter 1 - Dietary assessment methodology. In A. M. Coulston, C. J. Boushey, M. G. Ferruzzi, & L. M. Delahanty (Eds.), Nutrition in the prevention and treatment of disease (4th ed., pp. 5-48). London, UK: Academic Press. Toledo, E., Salas-Salvadó, J., Donat-Vargas, C., Buil-Cosiales, P., Estruch, R., Ros, E., . . . Martínez-González, M. A. (2015). Mediterranean diet and invasive breast cancer risk among women at high cardiovascular risk in the PREDIMED trial: Randomized clinical trial. JAMA Internal Medicine, 175(11), 1752-1760. doi:10.1001/jamainternmed.2015.4838 Tosti, V., Bertozzi, B., & Fontana, L. (2017). Health benefits of the Mediterranean diet: Metabolic and molecular mechanisms. The Journals of Gerontology: Series A, 73(3), 318-326. doi:10.1093/gerona/glx227 Trichopoulou, A., Kouris-Blazos, A., Wahlqvist, M. L., Gnardellis, C., Lagiou, P., Polychronopoulos, E., . . . Trichopoulos, D. (1995). Diet and overall survival in elderly people. British Medical Journal, 311(7018), 1457-1460. doi:10.1136/bmj.311.7018.1457 Trichopoulou, A., Costacou, T., Bamia, C., & Trichopoulos, D. (2003). Adherence to a Mediterranean diet and survival in a Greek population. New England Journal of Medicine, 348(26), 2599-2608. doi:10.1056/nejmoa025039 Trichopoulou, A., Martinez-Gonzalez, M. A., Tong, T. Y. N., Forouhi, N. G., Khandelwal, S., Prabhakaran, D., . . . de Lorgeril, M. (2014). Definitions and potential health benefits of the Mediterranean diet: Views from experts around the world. BMC Medicine, 12(112). doi:10.1186/1741-7015-12-112 Tufts Center for the Study of Drug Development. (2013). Clinical success rates for new cancer drugs double while more enter testing. Retrieved 16th December 2018 from https://csdd.tufts.edu/impact-reports/ United Nations Department of Economic and Social Affairs Population Division. (2013). World population ageing 2013. Retrieved 21st December 2018 from http://www.un.org/en/development/desa/population/publications/ageing/WorldP opulationAgeingReport2013.shtml

33 United Nations Educational Scientific and Cultural Organization. (2010). Convention for the safeguarding of the intangible cultural heritage. Retrieved 21st December 2018 from www.unisob.na.it/ateneo/c002/unesco_nomination_file_rl2010.pdf United Nations Educational Scientific and Cultural Organization. (2013). Decision of the Intergovernmental Committee: Mediterranean diet (8.COM 8.10). Retrieved 30th January 2019 from https://ich.unesco.org/en/decisions/8.COM/8.10 US Department of Health and Human Services and US Department of Agriculture. (2015). 2015-2020 Dietary guidelines for Americans, 8th Edition. Washington DC: US Government Printing Office. US News and World Report. (2019). The best diets overall. Retrieved 8th January 2019 https://health.usnews.com/wellness/food/slideshows/best-diets- overall?onepage Valls-Pedret, C., Sala-Vila, A., Serra-Mir, M., Corella, D., de la Torre, R., Martínez- González, M. Á., . . . Salas-Salvadó, J. (2015). Mediterranean diet and age- related cognitive decline: A randomized clinical trial. JAMA Internal Medicine, 175(7), 1094-1103. doi:10.1001/jamainternmed.2015.1668 Veronese, N., Stubbs, B., Noale, M., Solmi, M., Luchini, C., Smith, T. O., . . . Maggi, S. (2017). Adherence to a Mediterranean diet is associated with lower prevalence of osteoarthritis: Data from the osteoarthritis initiative. Clinical Nutrition, 36(6), 1609-1614. doi:10.1016/j.clnu.2016.09.035 Veronese, N., La Tegola, L., Crepaldi, G., Maggi, S., Rogoli, D., & Guglielmi, G. (2018). The association between the Mediterranean diet and magnetic resonance parameters for knee osteoarthritis: Data from the Osteoarthritis Initiative. Clinical Rheumatology, 37(8), 2187-2193. doi:10.1007/s10067-018-4075-5 Villani, A., Sultana, J., Doecke, J., & Mantzioris, E. (2018). Differences in the interpretation of a modernized Mediterranean diet prescribed in intervention studies for the management of type 2 diabetes: How closely does this align with a traditional Mediterranean diet? European Journal of Nutrition, 1-12. doi:10.1007/s00394-018-1757-3

34 Wade, A., Davis, C., Dyer, K., Hodgson, J., Woodman, R., Keage, H., & Murphy, K. (2017). A Mediterranean diet to improve cardiovascular and cognitive health: Protocol for a randomised controlled intervention study. Nutrients, 9(2), 145- 145. doi:10.3390/nu9020145 Wahl, D., Solon-Biet, S. M., Wang, Q.-P., Wali, J. A., Pulpitel, T., Clark, X., . . . Le Couteur, D. G. (2018). Comparing the effects of low-protein and high- carbohydrate diets and caloric restriction on brain aging in mice. Cell Reports, 25(8), 2234-2243.e2236. doi:10.1016/j.celrep.2018.10.070 Wahl, D., Solon-Biet, S. M., Cogger, V. C., Fontana, L., Simpson, S. J., Le Couteur, D. G., & Ribeiro, R. V. (2019). Aging, lifestyle and dementia. Neurobiology of Disease, 130, 104481. doi:10.1016/j.nbd.2019.104481 Willett, W., Rockström, J., Loken, B., Springmann, M., Lang, T., Vermeulen, S., . . . Murray, C. J. L. (2019). Food in the Anthropocene: the EAT–Lancet Commission on healthy diets from sustainable food systems. The Lancet, 393(10170), 447-492. doi:10.1016/S0140-6736(18)31788-4 Willett, W. C., Sacks, F., Trichopoulou, A., Drescher, G., Ferro-Luzzi, A., Helsing, E., & Trichopoulos, D. (1995). Mediterranean diet pyramid: A cultural model for healthy eating. American Journal of Clinical Nutrition, 61(6 Suppl), 1402S- 1406S. doi:10.1093/ajcn/61.6.1402S Woo, J., Woo, K., Leung, S., Chook, P., Liu, B., Ip, R., . . . Celermajer, D. (2001). The Mediterranean score of dietary habits in Chinese populations in four different geographical areas. European Journal of Clinical Nutrition, 55(3), 215-220. doi:10.1038/sj.ejcn.1601150 World Health Organization. (2019). Risk reduction of cognitive decline and dementia: WHO guidelines. Retrieved 7th June 2019 from https://www.who.int/mental_health/neurology/dementia/guidelines_risk_reducti on/en/ Xia, X., Jiang, Q., McDermott, J., & Han, J. D. J. (2018). Aging and Alzheimer’s disease: Comparison and associations from molecular to system level. Ageing Cell, 17, e12802. doi:10.1111/acel.12802

35 Zaragoza-Martí, A., Cabañero-Martínez, M. J., Hurtado-Sánchez, J. A., Laguna-Pérez, A., & Ferrer-Cascales, R. (2018). Evaluation of Mediterranean diet adherence scores: A systematic review. BMJ Open, 8(2), e019033. doi:10.1136/bmjopen- 2017-019033

36 CHAPTER TWO

Evolution of Mediterranean diets and cuisine:

Concepts and definitions

37 PREFACE

The usefulness of studies on the Mediterranean diet is dependent on a clear understanding of its definition. However, there is confusion with various definitions used in interventions, pyramids and diet index tools, which do not all reflect the archetypal diet. This is impeding the quality, consistency and synthesis across trials. An improved understanding of the concepts and cuisine is important to assess how well previous interventions have reflected the ‘traditional’ dietary pattern, and to inform future research design and education initiatives. It is also required to adapt existing, or develop new, diet index tools to better capture this unique dietary pattern.

The aim of this chapter was to improve understanding of the ‘traditional’ Mediterranean cuisine and related eating habits and to identify any additional elements previously omitted or poorly studied, which may impact on health outcomes. To this end, a narrative review was selected as the preferred methodology, since it provides a broad scope to capture historical data and consolidate previous work.

This chapter was published in the Asia Pacific Journal of Clinical Nutrition. It includes the peer-reviewed version of the final form of the following article:

The published article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions.

38 Part of the work produced in this chapter has been presented at the following conference:

39 AUTHORSHIP STATEMENT

The co-authors of the paper: ‘Evolution of Mediterranean diets and cuisine: Concepts and definitions’ confirm that Sue Radd-Vagenas has made the following contributions:

• Conception and design of the research • Collection and extraction of data • Analysis and interpretation of the findings • Writing of the original draft • Critical appraisal of the content and revisions

As the primary supervisor for the candidature upon which this thesis is based, I can confirm that the above authorship attribution statement is true and correct.

Professor Victoria M Flood

Faculty of Health Sciences

The University of Sydney

25th February 2019

40 Asia Pac J Clin Nutr 2017;26(5):749-763 749 Review Article

Evolution of Mediterranean diets and cuisine: concepts and definitions

1 2 Sue Radd-Vagenas Grad Dip Diet , Antigone Kouris-Blazos PhD , Maria Fiatarone Singh 1 1,3,4 MD , Victoria M Flood PhD

1Faculty of Health Sciences, the University of Sydney, NSW, Australia 2Department of Rehabilitation, Nutrition and Sport, School of Allied Health, La Trobe University, Bundoora, Victoria, Australia 3St Vincent’s Health Network, Sydney, NSW, Australia 4Western Sydney Local Health District, NSW, Australia

Background and Objectives: The Mediterranean diet has been demonstrated to provide a range of health benefits in observational and clinical trials and adopted by various dietary guidelines. However, a broad range of definitions exist impeding synthesis across trials. This review aims to provide a historical description of Mediterrane- an diets, from the ancient to the modern, to inform future educational and diet index tool development representing the ‘traditional’ Mediterranean diet. Methods and Study Design: Nine databases were searched from inception to July 2015 to identify papers defining the Mediterranean diet. The definition accepted by the United Na- tions Educational, Scientific and Cultural Organization (UNESCO) was also reviewed. Results: The ‘traditional’ Mediterranean diet is described as high in unprocessed plant foods (grains, vegetables, fruits, legumes, nuts/seeds and extra virgin olive oil), moderate in fish/shellfish and wine and low in meat, dairy, eggs, animal fats and discretionary foods. Additional elements relating to cuisine and eating habits identified in this review include frequent intake of home cooked meals; use of moist, lower temperature, cooking methods; eating main meals in company; reduced snacking occasions; fasting practice; ownership of a vegetable garden; use of traditional foods and combinations; and napping after the midday meal. Conclusions: Scope exists for future tools to incorporate additional elements of the ‘traditional’ Mediterranean diet to improve the quality, consistency, and synthesis of ongoing research on the Mediterranean diet.

Key Words: Mediterranean diet, plant based diet, traditional foods, diet pattern, diet index

INTRODUCTION outcomes, to inform development of future educational The Mediterranean diet has been associated with multiple and diet index tools. health benefits and increased survival.1-7 However, this diet has proven difficult to define in the literature as it is METHODS surprisingly complex and varies between countries and We conducted a review of Mediterranean diet definitions time periods. Variations exist in definitions provided by as described by various literature and characterised by educational materials such as Mediterranean diet pyra- diet index tools. Nine databases were searched from in- mids and dietary guidelines used by educators and health ception to July 2015 to identify papers for a related sys- professionals when dealing with the general public.8-11 tematic review on the Mediterranean diet and cognitive Disparities also occur in the criteria, cut offs and scoring functioning and brain morphology or function. These systems selected by common Mediterranean diet pattern were MEDLINE, Premedline, AMED, All EBM review index tools or scores, which researchers use to assess ad- databases including Cochrane Database of Systematic herence to this diet pattern in observational studies and Reviews (DSR), ACP Journal Club, Database of Ab- clinical trials.5,12-16 stracts of Reviews of Effects (DARE), Cochrane Central This review provides a brief description of the ancient Mediterranean diet, focusses mostly on what is commonly regarded in the scientific literature as the ‘traditional’ Corresponding Author: Prof Victoria M Flood, Faculty of Health Sciences, The University of Sydney, 75 East Street, Lid- Mediterranean diet and, mentions some variations to this combe NSW 2141, Australia. with modern Mediterranean diets. It summarises key ‘tra- Tel: +61 2 9351 9001; Fax: +61 2 9351 9838 ditional’ Mediterranean diet foods and dishes. It also Email: [email protected]; identifies additional elements relating to ‘traditional’ cui- [email protected] sine and eating habits that have not been consistently or Manuscript received 25 February 2016. Initial review completed collectively assessed in the majority of research studies to 25 April 2016. Revision accepted 24 June 2016. date, yet may have an influence on wellbeing and health doi: 10.6133/apjcn.082016.06 41 750 S Radd-Vagenas, A Kouris-Blazos, M Fiatarone Singh and VM Flood

Register of Controlled Trials (CCTR), NHS Economic amount of meat with a strong preference for fish and sea- Evaluation Database (CLEED), Cochrane Methodology food.17 Register (CLMCR), Health Technology Assessment One significant yet underappreciated influence on the (CLHTA), CINAHL, EMBASE, PubMed, PsychINFO, historical path of the Mediterranean diet is the contribu- Web of Science. Key search terms included Mediterrane- tion made by Arabic/Islamic culture. As an example, at an diet, Mediterranean dietary pattern, Mediterranean the time when the Moors occupied much of Spain, the type diet, Cretan diet, MedDiet, MeDi and MeDiet. In Christian (Catholic) diet was based on meat, wine and addition grey literature was searched and additional use- grains with poor dietary variety. However, in Islamic ful historical references were identified from reference food culture meat comprised only a small part of the diet, lists. Details of the search can be found in our PROS- whereas a variety of vegetables were plentiful.23 Islamic PERO protocol registered at http://www.crd.york.ac.uk/ culture helped to transform the Mediterranean culinary PROSPERO/as CRD42015024088. model by introducing many new varieties of foods, such In the search we identified more than 30 Mediterranean as eggplant, spinach, citrus, pomegranate, almonds and diet index tools. We chose to review seven of these to spices, as well as providing preparation methods for reci- illustrate the diversity of tools in existence that measure pes.17 Grains, such as wheat, barley and rice (and espe- adherence to a Mediterranean diet. The tools selected cially bread), were considered important by Muslims and were either widely cited and/or included a greater number many legumes were classified as having medicinal prop- of elements to assess the intake of Mediterranean foods, erties.23 aspects of cuisine and broader eating habits. A further noteworthy influence on the Mediterranean diet was the discovery of the Americas. Christopher Co- RESULTS lumbus’s voyages (first journey was in 1492) resulted in Our main findings are presented within the following novel foods from the New World, such as tomato, potato, framework: a) a historical time line of Mediterranean capsicum, corn, various legumes and chilli, later becom- diets (ancient, ‘traditional’, modern), with details of char- ing commonplace within this diet.17,24 For example, while acterising foods and cooking methods used within the the tomato is now generally considered the food most ‘traditional’ Mediterranean diet b) additional culinary and synonymous with Mediterranean cuisine, the historical psycho-social elements of the ‘traditional’ Mediterranean account testifies this was not the case in the ancient diet.17 diet and c) relevance of existing diet index tools to the ‘traditional’ Mediterranean diet. ‘Traditional’ Mediterranean diet When speaking of the ‘traditional’ Mediterranean diet in Ancient Mediterranean diet the scientific literature, most researchers tend to refer to a The Mediterranean diet has its origins in antiquity, from plant based dietary pattern common to the olive growing the foods and dietary patterns of countries surrounding areas of the Mediterranean region as described in the late the Mediterranean basin considered by historians as the 1950s and early 1960s on the island of Crete, Greece and “cradle of civilization”.17 This region was a melting pot of rural villages in Southern Italy such as Nicotera, which influences from ancient civilizations such as the Minoan were part of the famous Seven Countries Study.25-27 (7000 BC-2000 BC) and Phoenician (1200 BC – 332 BC) ‘Traditional’ is defined by the European Union Eu- to various Greek periods, e.g., Classical Greece (479 BC- roFIR project as referring to practices or specifications 323 BC) and the Roman Empire (31 BC – 476 AD).18-20 It established prior to the Second World War, meaning be- was also exposed to Arabic/Islamic culture and food dis- fore the modern era of factory farming and mass scale coveries from the New World. The diet in the Mediterra- food production.28 As Trichopoulou et al point out, ‘tradi- nean is therefore not a single construct but has progres- tional’ reflects a period “when most population groups sively evolved over time and lands. Specific food compo- still applied simple, time honoured approaches” to grow nents have been more or less emphasised according to the most of the food they personally consumed.29 It also re- historical period, culture, religion, agricultural production fers to a time before the globalisation of ultra-processed and the economy of the time. In addition, climatic condi- foods. Although sounding like the distant past, Keys30 tions, poverty and hardship have contributed to shaping argued the ‘traditional’ Mediterranean diet definition was the Mediterranean diet rather than intellectual insight or a relatively new invention rather than what was consumed wisdom, which tends to be the method used in modern by natives of the Mediterranean region in ancient times. times to construct popular diets.21 The ‘traditional’ Mediterranean diet can be defined as A study of the writings of the Greek philosopher Plato being high in unprocessed plant foods (cereals (grains), (5th to 4th Century BC) provides valuable insight into vegetables, fruits, legumes, nuts/seeds and extra virgin some of the early dietary customs of the Mediterranean olive oil); moderate in fish/shellfish (until recent times, region. For example, Plato stated that a healthy diet con- fish consumption was a function of proximity to the sea) sists of cereals, legumes, fruit, milk, honey, fish and wa- and wine (usually consumed only with meals and if ac- ter. He cautioned that meat, confectionary and wine ceptable by religious belief); and low in meat and dairy should only be consumed in moderate quantities.22 products, eggs, animal fats and discretionary foods.31 In the Roman tradition, which was based on the ancient When the ratio of plant versus animal foods is considered, Greek model, bread, wine and oil were the triad charac- the ‘traditional’ Mediterranean diet presents as a semi teristic of the Mediterranean diet. Additional foods in- vegetarian diet, abundant in unrefined plant foods with a cluded some sheep’s cheese, vegetables (chicory, leeks, notable absence of ultra-processed foods, such as extrud- lettuce, mallow and mushrooms) and only a small ed snacks, sugary foods and drinks, fast foods and refined 42 Historical review Med diets and cuisine 751 fats and oils, now commonplace in modern day diets. meal including soups, cooked dishes, salads and even fruit.5 These breads were based on stone ground whole- A plant based diet meal flour and made with sourdough culture, and would In his personal reflections, the principal investigator of therefore have been lower in Glycaemic Index (GI) than the Seven Countries Study, Ancel Keys,30 stated the most modern breads consumed.31 Rusks (paximadia) “heart of this diet is mainly vegetarian and differs from made from barley were staple foods in Crete. These American and Northern European diets in that it is much would presumably have elicited a particularly low gly- lower in meat and dairy products and uses fruit for des- caemic response since barley is rich in viscous dietary sert”. Keys described “pasta in many forms, leaves sprin- fibre and the GI of barley is low compared with other kled with olive oil, all kinds of vegetables in season, and grains.33,34 In addition, rusks were always softened with often cheese, all finished off with fruit, and frequently olive oil prior to consumption further lowering their gly- washed down with wine”. Further, when comparing mod- caemic response. Whole grains, including any foods made ern interpretations of the Mediterranean diet he criticised from unrefined or minimally processed grains such as Italian restaurants in the US and England as serving a whole wheat and barley (intact, cracked, rolled or flour travesty of the healthy ‘traditional’ Mediterranean diet as products), also provided good amounts of dietary fibre “everything has to be loaded with butter or margarine and and unique phytonutrients such as lignans, phytates and ground meat”. Also, he rebuked them because “serving phenolic acids that complemented those found in fruit and only fruit for dessert is not common” whereas “ice cream vegetables.35 or pie is customary”.30 Despite the fact that the dishes traditionally consumed Dishes based on vegetables, legumes and extra virgin by the Greeks were not the same as that consumed by the olive oil Spaniards, Lebanese or Moroccan, the Cretan diet circa Large amounts of salads, soups and cooked vegetable 1960’s remains the preferred prototype in the literature dishes were traditionally consumed. These were based on for the ‘traditional’ Mediterranean diet.21,32 See Table1 seasonal vegetables and legumes from the region (e.g., for a descriptive list of foods consumed by Cretan men in lentils, chickpeas, broad beans, white beans, lupins), the 1960’s. doused with copious amounts of extra virgin olive oil leading to composite dishes with mixtures of antioxi- Bread as a staple food dants.33 Examples include stuffed vegetables (tomato, Local food choices available for the preparation of ‘tradi- capsicum, eggplant and zucchini), white beans with fen- tional’ Mediterranean meals were influenced by the sea- nel, chickpeas with spinach and broad beans with wild sons, which can help promote dietary diversity over the artichokes. Lemon juice or vinegar, herbs and on- year or be a vulnerability in the case of bad weather. Sea- ion/garlic were also frequently used to season salads or sons determined the cuisine, helped to impart varied fla- legume soups before consumption, as well as in food vours and increase phytonutrient intakes when abundant preparation, such as when baking ‘lemon potatoes’ or harvests were available. Unlike recent times where low cooking goat or lamb. Modern research suggests addition carbohydrate diets have been increasingly promoted in of condiments such as lemon juice/vinegar may help low- the popular and social media, it was common to eat large er the GI of a meal and reduce the formation of advanced quantities of unrefined bread in many forms with each glycation end products (AGE’s), generally low in raw or

Table 1. Description of the food intake of Cretan men in the 1960’s†

Food Description Quantity/frequency Bread Made with stoneground wholemeal flour, sourdough At least 6 slices per day Fruit Seasonal e.g. grapes, figs, apples, melons At least 2 pieces per day Vegetables Seasonal; regular inclusion of wild greens and dishes cooked in a rich tomato sauce with onions/garlic, herbs and olive oil Meat Mainly sheep or goat; usually boiled or casseroled Once per week or less Fish/seafood Any type available, depending on proximity to the sea e.g. sardines, Once per week or less cod, herring, octopus Legumes Lentils, chickpeas, broad beans etc. in place of meat 2-3 times per week Nuts All types available e.g. walnuts, almonds, chestnuts At least 3 times per week Olives All types available Daily Eggs Free range Average of 3 per week Cheese/yoghurt Made from sheep/goats milk 2-3 times per week or more Milk From sheep/goats milk; full cream Reserved for children Olive oil Extra virgin More than 4 tablespoons daily Wine Red wine or retsina 100-200 mL per day Tea Sage, mountain tea (sideritis or ironwort), lemon verbena, dittany, Daily chamomile, mint Coffee Made from grounds and boiled Daily, if available Herbs Oregano, parsley, celery leaf, dill etc. Daily Spices Cinnamon, cumin, cloves, saffron etc. Several times per week

†Adapted from Kouris-Blazos et al and used with permission.32

43 752 S Radd-Vagenas, A Kouris-Blazos, M Fiatarone Singh and VM Flood boiled vegetables but formed in significant amounts including the preparation of rich tomato pasta sauces when frying potatoes with added fat at high temperature where, together with the extra virgin olive oil, they in- e.g. French fries.36,37 creased the content of flavonoids.49,50 The regular use of While Greece still claims the highest per capita con- herbs/spices in cooking, even in small quantities, further sumption of olive oil as a country, Cretans still consume contributed to a higher dietary intake of polyphenols.21,51 the largest amounts in the world.38,39 A survey carried out For example, dried oregano, which is one of the richest by the American Rockefeller Foundation in 1948 in Crete sources of polyphenols among herbs, was regularly added indicated that olive oil contributed heavily to the daily to almost all dishes, along with some salt, pepper, lemon caloric intake and “food seemed literally to be ‘swim- and extra virgin olive oil.52 Cultivated for centuries in the ming’ in oil”.40 Recently, Ramirez-Anaya Jdel et al Mediterranean, and used also as medicine, oregano has showed that frying vegetables in extra virgin olive oil can more recently been shown to contain several antioxidants increase their overall polyphenol content by virtue of ab- with anti-inflammatory, blood glucose and lipid lowering sorption of the polyphenols from the oil.41 In Greece and potential although clinical trials are lacking.53 Other Southern Italy, cooking vegetables in olive oil was com- popular herbs used traditionally include dill, mint, rose- monplace. Fresh dark green leaves and/or boiled wild mary and wild fennel, although their usage depended on (often bitter) varieties such as endive, chicory, sow thistle, the region. Recent analyses of traditional Greek foods and amaranth and purslane drizzled with olive oil and lemon dishes determined that the addition of capers in a sauce to juice were an important part of daily cuisine when in sea- the humble Santorini dish called fava (made from split son. Vine leaves stuffed with rice and spices were also peas cooked with onion) further enriched it with flavonol, consumed. The bitter cooking water that remained after flavone and flavan-3-ol.29 A simple green bean dish boiling greens was usually drunk by the cook as it was (known as fasolakia) was found to contain the highest believed to be highly nutritious. More than 80 different content of flavonoids in an analysis of a weekly Mediter- edible varieties of wild greens (weeds) have been identi- ranean menu due to the addition of onion and parsley in fied on Crete.33 Wild greens and green pies prepared with cooking.21 Indeed, the Greek version of the Mediterrane- a combination of leaves and herbs were found to contain an diet used for this analysis was found to be high in fla- exceptionally high levels of antioxidants, including a vonoids compared with diets in Northern European coun- wide range of flavonoids.27 Dark leafy greens are also tries.21 Interestingly, Cretan cuisine incorporated a wider particularly rich in the carotenoid lutein, which can be range of herbs and spices due to its extensive history of preferentially taken up by the brain and macula and pro- invasions from the Byzantines and Arabs to the Venetians vides antioxidant and anti-inflammatory effects.42 Further, and Ottoman Turks prior to Crete becoming unified with in addition to beetroot, which was cooked (both bulb and and influenced by mainland Greece. For example, herbs leaves) and eaten cold as a salad dressed with olive oil and spices often associated with Asian cuisine, such as and lemon or vinegar, dark green leafy vegetables were a cumin, coriander and saffron, were used. This may help major contributor to the high nitrate content of the ‘tradi- explain why Cretan cuisine was even more beneficial tional’ Mediterranean diet, which promotes vasodilation than the cuisine of mainland Greece at that time period.33 of blood vessels and lowers blood pressure.43-45 Emerging Trichopoulou29 pointed out that the nutritional secrets of research has also shown that leafy greens contain a sugar ‘traditional’ Mediterranean foods and dishes were not called sulfoquinovose that can beneficially improve the immediately obvious as they appeared to relate to their gut microbiota.46 phytonutrients. This is not surprising since dietary analy- The type of olive oil used traditionally for all cooking sis using standard food composition tables found in most and eating purposes was extra virgin. There are two ma- apps and software programs omits the diverse range of jor reasons why extra virgin olive oil might be preferen- phytonutrients existing in plant foods (as there is incom- tial for health compared with refined olive or vegetable plete compositional data), so that the analytical focus re- oil extracted with solvents. First, it contains a high pro- mains on macro and micronutrients. portion of monounsaturated to polyunsaturated fatty acids, which makes it more resistant to lipid peroxidation.47 Se- Nuts and honey to make nutritive sweets cond, ‘extra virgin’ means that the olives were mechani- Nuts and seeds were eaten by the handful and used to cally pressed to squeeze out their juice, preserving their make nutritive sweets together with honey e.g. pastelli polyphenol and tocopherol content. Such phytonutrients (made from sesame seeds and honey) or grape juice syr- further contribute to oxidative stability and may provide up/must (petimezi) and occasionally dried fruit, since the anti-inflammatory benefits.48 It was not customary to use of table sugar was very low in the ‘traditional’ diet. spread any butter or margarine on bread in the Mediterra- Spoon sweets were made by cooking fruit, vegetables or nean. Twice baked Cretan rusks, which could survive green walnuts in their shell with sugar to create a syrupy long journeys by shepherds were softened before eating sweet. However, these sweets were considered only an by drizzling with extra virgin olive oil and fresh tomato occasional treat, provided as one spoonful followed by juices.33 Olives themselves, a fermented food with live water, and mostly served to guests as a gesture of hospi- cultures, were also regularly consumed as a condiment to tality. Key Mediterranean nuts include almonds, hazel- meals. nuts, pine nuts, pistachios, and walnuts.54 Walnuts provide a particularly rich source of alpha linolenic acid and Traditional seasonings the highest level of phenolic compounds compared to Onion and garlic were generously used, both raw as part other nuts.55 Honey has been used traditionally as medi- of daily salads and as the basis of most cooked dishes, cine and recent research suggests it can provide antioxi- 44 Historical review Med diets and cuisine 753 dant, anti-inflammatory and anti-microbial effects benefi- that fermented foods in the ‘traditional’ Mediterranean cial for treating certain conditions.56,57 diet, including yoghurt, cheese, trahana (fermented wheat and milk), olives and sourdough bread, may contribute to Beverages: water as the main drink a healthy microbiome and that, at least, some of the bene- In addition to water being consumed as the daily drink, fits obtained from Mediterranean diets might be operating moderate amounts of herbal tea, such as from sage and through the microbiome.32 various mountain herbs, and/or boiled Greek coffee were consumed as part of the ‘traditional’ Mediterranean diet, Meat in very small amounts although there is a paucity of data on actual quantities.58 The small amount of meat consumed in the prototype It was the job of children to pick herbs for making tea, Cretan diet usually came from small animals that grazed while the coffee beans were purchased according to on pasture, (e.g., goats and sheep), rather than cows lot household budgets which, for most, were limited. More fed or lot finished with grains, as is the case with most recent research has confirmed these beverages supply beef produced today in Western countries.65 These ani- various antioxidants, which may provide health bene- mals provided an additional source of omega-3 fatty acids fits.59,60 While coffee consumption may potentially pro- to the diet, as did snails which were popular during reli- vide both positive and negative health effects, Siasos et al gious fasting and throughout the year.66 As compared to reported boiled Greek coffee contained only a moderate modern diets where meat takes centre stage on the plate amount of caffeine while it is rich in antioxidants, such as and dishes are named accordingly further elevating the polyphenols, and magnesium.61,62 In the ‘traditional’ diet, status of meat, legumes were the pillars of the ‘tradition- herbal teas and coffee were also consumed with lower al’ Mediterranean diet. The Southern Greeks had quite a quantities of honey or sugar compared with amounts of puritanical attitude towards meat, viewing regular meat added sugars contained per serve in modern sugar sweet- intake as being greedy.33 ened beverages. Further, the cooking methods used to prepare meats were usually lower temperature and higher moisture e.g. Some dairy from sheep and goats boiling, stewing with the frequent addition of herbs and Dairy was traditionally sourced from sheep and goats lemon retarding the formation of AGE’s, polyaromatic (occasionally buffalo), which provide A2 beta-casein, as hydrocarbons (PAH’s) and/or heterocyclic amines found in breast milk.63 In Western society however, most (HCA’s), which have all been associated with chronic milk is sourced from cows that exist in mixed herds sup- disease.37,67 plying a combination of A1 and A2 beta-casein. While While traditional dishes varied in Mediterranean coun- further research is required, some studies suggest A2 be- tries, modulated by ethnic background, agricultural pro- ta-casein may be better tolerated by certain individuals duction, economy and religious beliefs, what all Mediter- and may be associated with a reduced risk of some chron- ranean countries shared in common was the unrefined ic diseases compared with A1 beta-casein.63,64 In the Med- plant based dietary pattern typical of this time period.68 iterranean, full cream sheep or goats milk was reserved See Table 2 for a sample one day ‘traditional’ Mediterra- for drinking by children, whereas yoghurt (usually nean (Greek) summer menu. strained, which decreases lactose and increases the protein content) and soft white fermented cheese was pre- Whole diet pattern more important than individual ferred by adults. In Greece, for example, fetta cheese foods (with live cultures) was regularly added to salads and to With regards to the question of which food/s in a Medi- finish the preparation of vegetable stews.5 On Crete, myz- terranean diet are most important, current data suggest ithra (a mild whey cheese that hardens with age and can there is no specific food or component that is as benefi- be grated) was most commonly produced, used and cial as the whole diet pattern and, possibly, the broader shipped around the Mediterranean as the cheeses of Crete traditional cuisine and lifestyle habits which may be sig- have been famous since antiquity.33 It has been suggested nificant elements. This is because simply changing one

Table 2. Sample one day ‘traditional’ Mediterranean (Greek) summer menu

Eating occasion Foods Breakfast Sourdough bread topped with grated fresh tomato and raw onion, crumbled feta, dried oregano and drizzled with extra virgin olive oil Herbal tea or boiled Greek coffee

Lunch (main meal) Oven baked giant Lima beans in tomato sauce + boiled potato and endive drizzled with olive oil and lemon + sourdough bread Small glass red wine or retsina

Dinner (lighter meal) Vegetable stew with eggplant, green beans, zucchini, potato, tomato, onion, garlic and parsley + sourdough bread OR village salad with barley rusks, tomato, cucumber, onion, olives and oregano drizzled with extra virgin olive oil OR Greek yoghurt topped with walnuts and drizzled with honey

Snacks Water Fresh fruit in season Nuts Dried fruit e.g. figs, raisins

45 754 S Radd-Vagenas, A Kouris-Blazos, M Fiatarone Singh and VM Flood food or adding a few traditional dietary components may meat intake was provided in PREDIMED (less than one not completely override the detrimental effects of a West- serve daily) compared with intakes in the ‘traditional’ ern diet.69 Mediterranean diet (a few times per month). Unlike other However, some studies have attempted to identify the studies, PREDIMED did include cuisine based advice relative importance of individual foods within predictive with the recommendation to consume meals with sofrito index tools. When considering the Mediterranean Diet (a rich sauce made with tomato and onion, often includ- Pattern Score (MDPS) used for data from the Food Habits ing garlic and aromatic herbs, slowly simmered in olive in Later Life (FHILL) study, which included five ethnic oil) at least two times per week. cohorts, Darmadi-Blackberry et al found that a higher Therefore, strictly speaking, neither the Lyon Diet legume intake was the most predictive dietary factor for Heart Study nor PREDIMED used a purely ‘traditional’ longevity, with a 7-8% reduction in mortality risk for Mediterranean diet as their intervention, making interpre- every 20 g increase in daily legume intake.70 In the Greek tation and comparison of the results more complex.73 It is European Prospective Investigation into Cancer and Nu- interesting, however, that both trials were stopped early trition (EPIC) where a higher adherence to the Mediterra- as interim analyses indicated the Mediterranean diet nean diet was associated with a significant reduction in groups had a significantly reduced risk of disease com- total mortality, the dominant components of the Mediter- pared with controls. ranean diet pattern index that predicted lower mortality Unfortunately, the diets currently consumed in Mediter- were moderate alcohol intake, low intake of meat and ranean countries (including the region where the proto- meat products and high intake of vegetables, fruits and type ‘traditional’ Mediterranean diet was described) are nuts, olive oil and legumes.71 A meta-analysis on the progressively being modernised and influenced by West- Mediterranean diet and cardiovascular disease (CVD) ern dietary habits due to rapid economic development and revealed that the protective effects of the diet pattern spe- high urbanisation rates and there is decreasing adherence cifically related to CVD as an outcome, and assessed us- to a ‘traditional’ Mediterranean diet pattern.78 While the ing various index tools, could mostly be attributed to ol- influence of American fast food culture can easily be ob- ive oil, fruit, vegetables and legumes.72 served, some of the move towards a greater use of animal products in modern Greek diets and a reduced use of Modern Mediterranean diet herbs and spices has been more subtle and can be traced While it is believed that Ancel Keys first coined the term back to earlier culinary instruction by the famous Chef ‘Mediterranean diet’, the concept of the Mediterranean Tselementes (1878-1958). After spending considerable diet was not widely recognised in the scientific literature time abroad at prestigious eating establishments, Tsele- until the 1990s with the publication of the Lyon Diet mentes returned to Greece in the early 20th century to Heart Study. This clinical trial showed a 50% reduction in open a cooking school and publish his first cookbook. His risk of a second heart attack or cardiovascular complica- cooking methods reduced the emphasis on herbs and tions, fewer cancers and a significant decrease in all- spices and, instead, added butter and cream to foods so cause mortality among those receiving the intervention that recipes conformed more to classic French cuisine and diet.73-75 Yet despite being a type of Mediterranean diet, appealed to the cosmopolitan upper class.79 While his the intervention used in the Lyon Diet Heart Study was influence was not apparent in Crete in the 1960s (the ma- not ‘traditional’ and included some variations. For exam- jor location and period that defined the “traditional” Med- ple, the intervention was not high in total fat and the sub- iterranean diet), it prevails today in most modern Greek jects were provided with an additional margarine made kitchens in the form of more bland dishes with rich cream from rapeseed oil to increase their alpha linolenic acid sauces i.e. bechamel, such as moussaka and pasticcio, intake, since at the time of the study the exact dietary which are mistakenly considered traditional. Hence, sources of alpha linolenic acid in the ‘traditional’ Medi- modern Mediterranean diets have been shaped by external terranean (Greek) diet were unknown.73 and internal forces. More recently, the Mediterranean diet gained acclaim Ironically, Tourlouki et al have suggested the ‘tradi- after the publication of the PREDIMED trial from Spain. tional’ Mediterranean diet is being forsaken at the fastest This was a primary prevention trial of subjects at high pace possibly within the countries of its origin, based on risk of CVD placed on one of two versions of a Mediter- data from the MEDIS (Mediterranean Islands Study) of ranean diet compared with a slightly lower fat control diet. people aged over 65 years from 12 islands in the Mediter- The findings were a 30% reduction in CVD risk, primari- ranean, and research conducted on primary school chil- ly due to decreased stroke risk, with either of the Mediter- dren in Malta.80,81 In the MEDIS study of ‘elders’, Crete ranean diets.76 Also, a reduction in the risk of type 2 dia- was recently found to have the highest frequency of fast betes and peripheral artery disease was reported, as well food and sweets consumption.81 as attenuation of decline or slight improvement in some aspects of cognitive function over time, compared with Additional culinary and psycho-social elements of the the control diet.7,77 PREDIMED is a further example of a ‘traditional’ Mediterranean diet modernised Mediterranean diet since the design of the In 2010 the United Nations Educational, Scientific and study meant that one intervention group was supplement- Cultural Organization (UNESCO) recognised the Medi- ed with polyphenol rich extra virgin olive oil while the terranean diet as an Intangible Cultural Heritage. Indeed, other received supplements of mixed nuts to encourage a to date, it is the only diet in the world protected by higher intake of these key foods in the respective groups, UNESCO in an effort to safeguard the traditional aspects compared with controls. Also, a higher allowance for of the diet, which are at risk of being abolished. As the 46 Historical review Med diets and cuisine 755 word ‘diet’ comes from the Greek word ‘diaita’, meaning meals may drive obesity as eating frequency has been a way of life or lifestyle, the recognised Mediterranean shown to account for twice the variance in energy intake diet is much more than just a diet pattern that can be de- compared with portion size.97 Fifth, in the Orthodox, scribed by consideration of its foods and/or nutrients, Catholic and Islamic religions some form of ‘fasting’ or although a quantitative definition may have particular food restriction was traditionally practised. For example, uses in research.82 The Candidature Dossier submitted to those compliant with the Orthodox Church recommenda- UNESCO describes the Mediterranean diet as “a social tions abstain from eating most foods of animal origin eve- practice based on the set of skills, knowledge, practices ry Wednesday and Friday as well as for three other signif- and traditions ranging from the landscape to the cuisine, icant periods leading to religious holidays. This means which in the Mediterranean basin concern the crops, har- they virtually follow a total plant based diet (vegan) for vesting, fishing, conservation, processing, preparation and, up to 200 days, or more than half of the year.98 Modern particularly, consumption”.83 Hence, domestic cooking science has confirmed fasting or abstaining from meat, using distinct methods, as well as eating and other social chicken, fish, dairy and eggs may have a favourable im- habits that revolve around the foods consumed, also play pact on the microbiome and result in less production of an integral role in the ‘traditional’ Mediterranean diet. potentially harmful bacterial metabolites and inflamma- Yet these elements have generally escaped the attention tion.32,99-102 Further, high fibre vegetable based diets of researchers, despite the likelihood that they may inter- whether they be vegan, vegetarian or Mediterranean are act with the type and quantity of foods consumed and associated with increased levels of faecal short chain fatty possibly exert independent health effects. While evidence acids, which confer a protective effect against various is currently lacking within the context of a Mediterranean chronic diseases.103 Fasting practice may also contribute diet to support the importance of such elements, emerging to moderation in overall food intake and a more frugal research within the wider scientific literature suggests diet, which is less commonly observed in Western society. these elements could be a significant contributing factor Sixth, Mediterranean people traditionally grew most of to the value of the ‘traditional’ Mediterranean diet. their own produce, a practice which has been suggested In the Mediterranean, food was traditionally cooked at to impact consumption levels of vegetables. The Australi- home or prepared by family and friends rather than by an arm of the MEDIS found having a home garden is pos- commercial and industrial operators. This is important itively correlated with a high Mediterranean diet pattern because recent data on Western populations has suggested index score.104 Seventh, certain traditional foods and more home cooking is associated with higher diet quality culinary combinations were consumed regularly in the as indicated by increased consumption of fruit, vegetables ‘traditional’ Mediterranean diet, such as Greek yoghurt and salads.84,85 Because of their dominant role within the and olives (both fermented foods), tomato salsa/sofrito, home in the past, Mediterranean women were the custo- wild dark green leafy vegetables drizzled with lemon dians of their culinary culture and self-production of juice and olive oil and vegetables and/or legumes cooked food.86 Second, the traditional slow cooking methods with herbs and olive oil, discussed earlier. Such foods and used in the Mediterranean included high moisture and combinations lend themselves to providing a greater di- lower temperature, e.g., boiling or stewing. These condi- versity and/or higher concentration of phytonutrients tions are now known to reduce the formation of various compared with adding a slice of tasty cheese, squirt of noxious chemicals in foods such as AGE’s, mentioned ketchup, a side salad without dressing (or low fat dressing) above, and acrylamide produced by faster modern cook- or a portion of steamed vegetables to the dinner plate, ing methods using intense and dry heat, associated with which tends to be more common in Western society. chronic disease.37,87-91 It is also now appreciated that Eighth, taking a short nap following the main midday preparation methods where vegetables are soft cooked so meal is well known to be part of the ‘traditional’ Mediter- that cell walls are broken down increase bioavailability of ranean lifestyle,105 which also included a considerable phytonutrients, further aided by the presence of olive oil amount of physical activity. A daily siesta has been asso- since many phytonutrients, e.g. carotenoids, are fat solu- ciated with reduced coronary mortality in a Greek popula- ble. Third, main meals in the Mediterranean were usual- tion,106 with a recent study on midday napping finding an ly consumed by sitting around the table (rather than in association with reduced blood pressure.107 However, front of a screen or while driving), where food can be while the additional elements identified in this review, savoured and eaten more mindfully in a convivial envi- were an integral part of the ‘traditional’ Mediterranean ronment with the company of family and friends.92 The lifestyle, further research is required to confirm their con- Greek historian Plutarch stated “we do not sit at table to tribution to reducing chronic disease risk within the con- eat, but to eat together”.83 This method of eating provides text of that lifestyle. Finally, despite the ‘traditional’ social connectedness and a sense of community, now rec- Mediterranean diet being used as a valuable reference ognised to be important for health.93 Indeed, frequent point for health within the scientific literature, capturing socialisation is a hallmark of Mediterranean traditions such a diet and trying to preserve and use it in different renowned for their longevity.94 Fourth, frequent snacking contexts may be fraught with difficulties and may have between meals was not common in the ‘traditional’ Medi- some unintended consequences. For example, gender terranean diet. In contrast, 96% of Australians say they roles, time availability and daily routines are changing regularly consume snack foods and Australians are now across the world, which may impact on the ability to snacking four times as much as they did a decade ago regularly cook at home, tend a garden or take a short nap according to market research.95,96 Recent data from four after the midday meal, even in Mediterranean countries. US surveys suggests excessive grazing between main 47 756 S Radd-Vagenas, A Kouris-Blazos, M Fiatarone Singh and VM Flood

Relevance of existing diet index tools to the ‘traditional’ fruit and nuts or legumes, nuts and seeds, whereas others Mediterranean diet incorporate measures for discretionary foods not part of Existing Mediterranean diet pattern index tools have been the ‘traditional’ Mediterranean diet, such as sugary bev- used in epidemiology and clinical trials to assess whole erages and sweets, in order to capture the potentially neg- dietary patterns and measure the combined effects of food ative effects of such items now common in Westernised components on disease outcomes, overcoming limitations diets. The intake of discretionary foods may be of particu- of studies on single nutrients or foods.108,109 However, lar interest when assessing non-Mediterranean popula- there have been disparities in their ability to predict simi- tions, and the degree of acculturation by Mediterranean lar risk reductions for chronic disease. These tools, de- populations. veloped a priori, have mostly focussed their attention on While an exhaustive comparison of Mediterranean diet foods and, in some cases, nutrients and they generally do index tools is beyond the scope of this review, the ele- not attempt to measure the broader elements of the ‘tradi- ments from seven commonly used tools out of more than tional’ Mediterranean diet and cuisine discussed in this 30 existing in the literature are provided (see Table 3). A review.5,12-14,16 Most index tools, with the exception of comparison of these example indices reveals that all in- MEDLIFE (see below), are also not based on the defini- clude some measure of vegetables, meat and meat prod- tion of the ‘traditional’ Mediterranean diet.110 ucts and fruit (even if as a composite group with nuts). The first Mediterranean diet pattern index tool (MDPS) Most include a measure of fish/seafood, dairy, legumes, was developed in 1995 by Trichopoulou et al.5 It exam- alcohol and olive oil intake. Few include a measure of ined the association between defined Mediterranean food wholegrain cereals, sofrito, poultry/white meat, eggs, patterns and overall survival using data collected for the vegetable oil, butter, margarine, cream, sugary beverages FHILL study. Scoring for the MDPS ranged from 0-8 and and sweets while none include a measure of vegetable focussed on broad food groups derived from the tradi- variety, dark green leafy vegetables, onions/garlic/leeks/ tional Cretan diet as described in the Seven Countries shallots, olives, fermented dairy, ‘extra virgin’ olive oil, Study but it did not reflect cooking methods or other as- lemon/vinegar use in food preparation, herbs and spices, pects of cuisine. Nevertheless this pioneering tool, and water, herbal tea and coffee, despite the latter being part the study it was based on, facilitated the work of re- of the ‘traditional’ Mediterranean diet and cuisine. searchers around the world to explore additional benefits Therefore, it has been suggested that while existing of a Mediterranean diet, resulting in rapid growth of pub- Mediterranean diet pattern index tools may be a helpful lications and development of multiple other index tools way of describing a Mediterranean diet, such indices may with variations.111 have led to an over simplification and confusion about the The variations in index tools appear primarily due to ‘traditional’ Mediterranean diet.73 There are also currently differences in purpose and how these tools were con- no Mediterranean diet indices that capture the broader structed. While the majority of Mediterranean diet pat- elements of ‘traditional’ cuisine and consumption habits tern tools seemingly aim to assess overall adherence to a discussed in this review. The most holistic example iden- Mediterranean diet pattern, specific elements included in tified is MEDLIFE, a 28 item index based on criteria some tools would suggest greater relevance to a particular from the new Spanish Mediterranean diet pyramid. This health outcome or population being studied. For example, includes measures for diet, physical activity patterns and Ciccarone et al who studied peripheral artery disease in social interaction but lacks the same elements as the edu- Italian patients with type 2 diabetes, included a frequency cational pyramid, discussed above, upon which it is based. measure of raw vegetable, carrot, butter and cream intake, 8,114 presumably to enable some proxy for carotenoid and saturated fat intake, which are related to diabetes complica- DISCUSSION tions.112 The Mediterranean Diet Adherence Screener Emulating and assessing the ‘traditional’ Mediterranean (MEDAS) used in PREDIMED on the other hand seems diet can be a challenge. The Mediterranean Diet Founda- to have been largely designed to enable a rapid assess- tion from Spain in conjunction with experts in the field, ment of compliance to a Mediterranean diet pattern so recently updated the ‘traditional’ Mediterranean diet pyr- feedback can be provided to participants during dietary amid in areas that have evolved with modernisation and interventions as part of a clinical trial. Unfortunately, to incorporate cultural and lifestyle elements.8 However, MEDAS also lacks a measure for frequency of intake of some of the updated dietary recommendations may be home cooked meals, the cooking methods used, socialisa- controversial, such as the higher allowance for red and tion at meals, snacking occasions throughout the day, processed meat, which exceeds amounts consumed tradi- fasting practice, availability of a home vegetable garden tionally in the prototype diet.26 Also, despite considering and napping after the main meal, which are characteristics a wide range of Mediterranean lifestyle elements, this of the ‘traditional’ Mediterranean diet.15,113 educational tool does not emphasise the potentially im- Indices also vary according to the number and actual portant role of traditional cooking methods such as stew- dietary elements assessed, whether quantity or frequency ing rather than barbecuing meat, unique traditional foods of the elements is being measured, cut-offs used for scor- such as fermented dairy products, fasting practice, and ing (absolute or a scale, and determined a priori or based ownership of a home vegetable garden, which may con- on gender specific mean intakes of the population as- tribute to the health outcomes provided by the ‘tradition- sessed, which may not reflect true adherence to the ‘tradi- al’ Mediterranean diet. tional’ Mediterranean diet) and the computation method. The literature supports that the ‘traditional’ Mediterra- Some index tools include composite food groups such as nean diet is much more comprehensive than just a collec- 48 Historical review Med diets and cuisine 757

Table 3. Elements measured in selected Mediterranean diet index tools†

Trichopoulou et al Trichopoulou et al Ciccarone et al Goulet et al Panagiotakos et al Schroder et al Sotos-Prieto et al 1995,5 2003,16 2003,112 2003,13 2007,109 2011,15 2014,114 Greece (MDPS) Greece Italy Canada Greece (MedDietScore) Spain (MEDAS) Spain (MEDLIFE) Type of index Mediterranean diet Mediterranean diet Mediterranean diet Mediterranean diet Mediterranean diet Mediterranean diet Mediterranean diet pattern pattern pattern pattern pattern pattern + cuisine pattern + lifestyle Number of elements assessed 8 9 15 11 11 14 28 Vegetables X X X X X X X Vegetables raw - - X - - - - Vegetables variety ------Dark green leafy vegetables ------Sofrito - - - - - X - Carrots - - X - - - - Potatoes - - - - X - X Onions/garlic/leeks/shallots ------Fruit - - X X X X X Fruits & nuts composite X X - - - - - Nuts - - - - - X - Olives ------Nuts & olives composite ------X Legumes X X - - X X X Legumes, nuts & seeds compo- - - - X - - - site Cereals X X - - - - X Cereals wholegrain - - - X X - - Wholegrain preference ------X Dairy X X - X X - - Dairy low fat ------X Dairy fermented e.g., yoghurt, ------fetta Cheese (all types) - - X - - - - Meat & meat products X X X X X X X Processed meat - - X - - - X Poultry/white meat - - - X X - X White meat preference - - - - - X - Fish/seafood - X X X X X X Eggs - - X X - - X Olive oil extra virgin ------Olive oil - - X X X X X Olive oil principal fat - - - - - X - Vegetable oil - - X - - - -

†For the sake of brevity, the lifestyle and non-food elements of cooking or eating behaviour discussed in this review are not represented in this table as none of the existing tools assessed these except for a measure of snacking behaviour in MEDLIFE, which also includes additional lifestyle elements. For Goulet et al a maximum of 1 point from refined grain products counted towards wholegrain products; a maximum of 1 point 13 from vegetable juice counted towards vegetables; a maximum of 1 point from fruit juice counted towards fruit; a maximum of 1 point from canola oil or olive oil margarine counted towards olive oil.

49 758 S Radd-Vagenas, A Kouris-Blazos, M Fiatarone Singh and VM Flood

Table 3. Elements measured in selected Mediterranean diet index tools† (cont.)

Trichopoulou et al Trichopoulou et al Ciccarone et al Goulet et al Panagiotakos et al Schroder et al Sotos-Prieto et al 1995,5 2003,16 2003,112 2003,13 2007,109 2011,15 2014,114 Greece (MDPS) Greece Italy Canada Greece (MedDietScore) Spain (MEDAS) Spain (MEDLIFE) Type of index Mediterranean diet Mediterranean diet Mediterranean diet Mediterranean diet Mediterranean diet Mediterranean diet Mediterranean diet pattern pattern pattern pattern pattern pattern + cuisine pattern + lifestyle Butter - - X - - - - Margarine - - X - - - - Cream - - X - - - - Butter, margarine, cream compo- - - X - - X - site MUFA: saturated fat ratio X X - - - - - Sugary beverages - - - - - X - Sugary beverages limited intake ------X Sweets - - - X - X X Processed snacks ------X Salt limit at meals ------X Lemon/vinegar in food prepara------tion Herbs & spices ------Herbs, spices & onion/garlic ------X garnish Water ------Herbal teas ------Water or herbal teas ------X Coffee ------Alcohol X X X - X - - Wine - - - - - X X Snacking behaviour ------X

†For the sake of brevity, the lifestyle and non-food elements of cooking or eating behaviour discussed in this review are not represented in this table as none of the existing tools assessed these except for a measure of snacking behaviour in MEDLIFE, which also includes additional lifestyle elements. For Goulet et al a maximum of 1 point from refined grain products counted towards wholegrain products; a maximum of 1 point from vegetable juice counted towards vegetables; a maximum of 1 point from fruit juice counted towards fruit; a maximum of 1 point from canola oil or olive oil margarine counted towards olive oil.13

50 Historical review Med diets and cuisine 759 tion of foods. It includes a cuisine with unique recipes Foundation provided a multi-disciplinary grant to support publi- and other psycho-social factors that may also impact on cation of this manuscript. food and nutrient intake, although a broader discussion of social science parameters is outside the scope of this re- REFERENCES view. Taking these factors into consideration, it has been 1. Gotsis E, Anagnostis P, Mariolis A, Vlachou A, Katsiki N, Karagiannis A. Health benefits of the Mediterranean diet: an proposed that integrating the concept of the Mediterrane- update of research over the last 5 years. Angiology. 2015;66: an diet with the concept of the Mediterranean way of 115 304-18. doi: 10.1177/0003319714532169. cooking should become a priority for the future. Can- 2. Sofi F, Abbate R, Gensini GF, Casini A. Accruing evidence 116 non suggests nutrition scientists should pay attention to on benefits of adherence to the Mediterranean diet on health: the history, tradition and culture of foods as they do to an updated systematic review and meta-analysis. Am J Clin biochemistry, food chemistry and other emerging and Nutr. 2010;92:1189-96. doi: 10.3945/ajcn.2010.29673. relevant scientific areas e.g., the microbiome and nutri- 3. Sofi F, Casini A. Mediterranean diet and non-alcoholic fatty genomics. liver disease: new therapeutic option around the corner? When designing new tools or updating existing ones to World J Gastroenterol. 2014;20:7339-46. doi: 10.3748/wjg. assess or promote adherence to the ‘traditional’ Mediter- v20.i23.7339. ranean diet, researchers may wish to consider a more ho- 4. Sofi F, Cesari F, Abbate R, Gensini GF, Casini A. Adherence to Mediterranean diet and health status: meta- listic definition as endorsed by UNESCO and include analysis. BMJ. 2008;337:a1344. doi: 10.1136/bmj.a1344. additional measures (either directly or by proxy), while 5. Trichopoulou A, Kouris-Blazos A, Wahlqvist ML, giving consideration to regional food cultures where the Gnardellis C, Lagiou P, Polychronopoulos E et al. Diet and 117 tool is intended for use. These measures include: overall survival in elderly people. BMJ. 1995;311:1457-60. 1. Frequency of intake of home cooked meals 6. Trichopoulou A, Orfanos P, Norat T, Bueno-de-Mesquita B, 2. Style of cooking methods used Ocké MC, Peeters PHM et al. Modified Mediterranean diet 3. Frequency of eating main meals in company vs. alone and survival: EPIC-elderly prospective cohort study. BMJ. 4. Number of snacking occasions outside main meals 2005;330:991. doi: 10.1136/bmj.38415.644155.8f. 5. Regular fasting of any type 7. Valls-Pedret C, Sala-Vila A, Serra-Mir M, Corella D, de la 6. Availability of a home vegetable garden Torre R, Martínez-González MÁ et al. Mediterranean diet 7. Use of traditional foods (or their nutritional counter- and age-related cognitive decline: a randomized clinical trial. JAMA Intern Med. 2015;175:1094-103. doi: 10.1001/jama parts) and combinations within dishes internmed.2015.1668. 8. Napping after the main midday meal 8. Bach-Faig A, Berry EM, Lairon D, Reguant J, Trichopoulou Tools that incorporate traditional culinary practices and A, Dernini S et al. Mediterranean diet pyramid today. eating habits from the ‘traditional’ Mediterranean diet Science and cultural updates. Public Health Nutr. 2011;14: may have increased usefulness for researchers, health 2274-84. doi: 10.1017/S136898001100 2515. professionals, school educators, chefs and the general 9. Ministry of Health and Welfare: Supreme Scientific Health public.113,118,119 Council. Dietary guidelines for adults in Greece. Arch Hellen Med. 1999;16:516-24. Conclusion 10. Oldways. Mediterranean diet pyramid 2008. 2015/12/18 Plant based diets appear best for the prevention of chronic [cited 2016/02/24]; Available from: http://oldwayspt.org/ disease,67,120 and a more sustainable way of eating for the resources/heritage-pyramids/mediterranean-pyramid/overvie w. future.121 The Mediterranean diet shares many food simi- 11. Willett WC, Sacks F, Trichopoulou A, Drescher G, Ferro- larities with modern vegetarian diets, but may find greater Luzzi A, Helsing E et al. Mediterranean diet pyramid: a acceptance due to its high palatability and allowance for a cultural model for healthy eating. Am J Clin Nutr. 1995;61: 122,123 low intake of animal foods. The Mediterranean diet 1402S-6S. is recommended for use by dietary guidelines in Australia 12. Alberti-Fidanza A, Fidanza F, Chiuchiu MP, Verducci G, and the US.124,125 Increased compliance with a Mediterra- Fruttini D. Dietary studies on two rural Italian population nean diet pattern, even in non-Mediterranean countries, groups of the Seven Countries Study. 3. Trend of food and has been associated with protection against chronic dis- nutrient intake from 1960 to 1991. Eur J Clin Nutr. 1999;53: ease and a decreased risk of premature mortality.29,73,126 854-60. doi: 10.1038/sj.ejcn.1600865. As the ‘traditional’ Mediterranean diet includes additional 13. Goulet J, Nadeau G, Lemieux S. Effect of a nutritional culinary and consumption habits for a ‘way of life’ un- intervention promoting the Mediterranean food pattern on plasma lipids, lipoproteins and body weight in healthy common in modern society, future research should assess French-Canadian women. Atherosclerosis. 2003;170:115-24. the relevant contribution of these elements to the effec- doi: 10.1016/s0021-9150(03)00243-0. tiveness of the Mediterranean diet and give consideration 14. Panagiotakos DB, Milias GA, Pitsavos C, Stefanadis C. to their inclusion in educational and index tools for re- MedDietScore: A computer program that evaluates the search, clinical practice and public health promotion. adherence to the Mediterranean dietary pattern and its relation to cardiovascular disease risk. Comput Methods ACKNOWLEDGEMENTS Programs Biomed. 2006;83:73-7. doi: 10.1016/j.cmpb.2006. We thank St Vincent’s Clinic Foundation, Sydney, Australia for 05.003. funding of a multidisciplinary grant. 15. Schroder H, Fito M, Estruch R, Martinez-Gonzalez MA, Corella D, Salas-Salvado J et al. A short screener is valid for AUTHOR DISCLOSURES assessing Mediterranean diet adherence among older On behalf of all authors, the corresponding author declares there Spanish men and women. J Nutr. 2011;141:1140-5. doi: 10. is no conflict of interest for this review. St Vincent’s Clinic 3945/jn.110.135566.

51 760 S Radd-Vagenas, A Kouris-Blazos, M Fiatarone Singh and VM Flood

16. Trichopoulou A, Costacou T, Bamia C, Trichopoulos D. 37. Uribarri J, Woodruff S, Goodman S, Cai W, Chen X, Pyzik Adherence to a Mediterranean diet and survival in a Greek R et al. Advanced glycation end products in foods and a population. N Engl J Med. 2003;348:2599-608. doi: 10. practical guide to their reduction in the diet. J Am Diet 1056/nejmoa025039. Assoc. 2010;110:911-6.e12. doi: 10.1016/j.jada.2010.03.01 17. Altomare R, Cacciabaudo F, Damiano G, Palumbo VD, 8. Gioviale MC, Bellavia M et al. The Mediterranean diet: a 38. International Olive Council. Market newsletter No. 91: history of health. Iran J Publ Health. 2013;42:449-57. trends in world olive oil consumption. 2016/02/24 [cited 18. Bispham E, Harrison TJ, Sparkes BA. The Edinburgh 2016/02/24]; Available from: http://www.internationalolive companion to ancient Greece and Rome. Edinburgh: oil.org/news/view/682-year-2015-news/578-market-newslett Edinburgh University Press; 2006. er-february-2015. 19. Boardman J, Edwards IES, Sollberger E, Hammond NGL. 39. Karpadakis M. World (extra virgin) olive oil consumption. The Cambridge ancient history. Cambridge: Cambridge The Chania Post. Feb-March, 2015:24. University Press; 1992. 40. Allbaugh LG. Crete: a case study of an underdeveloped area. 20. Mellersh HEL, Williams N. Chronology of world history. New Jersey: Princeton University Press; 1953. Santa Barbara, California: ABC-CLIO; 1999. 41. Ramirez-Anaya Jdel P, Samaniego-Sanchez C, Castaneda- 21. Vasilopoulou E, Georga K, Joergensen M, Naska A, Saucedo MC, Villalon-Mir M, de la Serrana HL. Phenols Trichopoulou A. The antioxidant properties of Greek foods and the antioxidant capacity of Mediterranean vegetables and the flavonoid content of the Mediterranean menu. Curr prepared with extra virgin olive oil using different domestic Med Chem Immunol Endocr Metab Agents. 2005;5:33-45. cooking techniques. Food Chem. 2015;188:430-8. doi: 10. doi: 10.2174/1568013053005508. 1016/j.foodchem.2015.04.124. 22. Skiadas P, Lascaratos J. Dietetics in ancient Greek 42. Johnson EJ. Role of lutein and zeaxanthin in visual and philosophy: Plato's concepts of healthy diet. Eur J Clin Nutr. cognitive function throughout the lifespan. Nutr Rev. 2014; 2001;55:532-7. doi: 10.1038/sj.ejcn.1601179. 72:605-12. doi: 10.1111/nure.12133. 23. Salas-Salvadó J, Huetos-Solano MD, García-Lorda P, Bulló 43. Ashworth A, Mitchell K, Blackwell JR, Vanhatalo A, Jones M. Diet and dietetics in al-Andalus. Br J Nutr. 2006;96: AM. High-nitrate vegetable diet increases plasma nitrate and S100-S04. doi: 10.1079/bjn20061710. nitrite concentrations and reduces blood pressure in healthy 24. Mariani-Costantini A, Ligabue G. Did Columbus also open women. Public Health Nutr. 2015;18:2669-78. doi: 10.1017/ the exploration of the modern diet? Nutr Rev. 1992;50:313- S1368980015000038. 9. doi: 10.1111/j.1753-4887.1992.tb07717.x. 44. Lidder S, Webb AJ. Vascular effects of dietary nitrate (as 25. Alberti-Fidanza A, Fidanza F. Mediterranean adequacy found in green leafy vegetables and beetroot) via the index of Italian diets. Public Health Nutr. 2004;7:937-41. nitrate‐‐ nitrite nitric oxide pathway. Br J Clin Pharmacol. doi: 10.1079/phn2004557. 2013;75:677-96. doi: 10.1111/j.1365-2125.2012.04420.x. 26. Kromhout D, Keys A, Aravanis C, Buzina R, Fidanza F, 45. Su Q, Rowley K, Itsiopoulos C, O'Dea K. Identification and Giampaoli S et al. Food consumption patterns in the 1960s quantitation of major carotenoids in selected components of in seven countries. Am J Clin Nutr. 1989;49:889-94. the Mediterranean diet: green leafy vegetables, figs and 27. Trichopoulou A. Mediterranean diet, traditional foods, and olive oil. Eur J Clin Nutr. 2002;56:1149-54. doi: 10.1038/sj. health: evidence from the Greek EPIC cohort. Food Nutr ejcn.1601472. Bull. 2007;28:236-40. 46. Speciale G, Jin Y, Davies GJ, Williams SJ, Goddard-Borger 28. Weichselbaum E, Benelam B, Soares Costa H. Traditional ED. YihQ is a sulfoquinovosidase that cleaves foods in Europe: Synthesis report No. 6. 2015/12/18 [cited sulfoquinovosyl diacylglyceride sulfolipids. Nat Chem Biol. 2016/02/24]; Available from: http://www.eurofir.org/?page_ 2016;12:215. doi: 10.1038/nchembio.2023. id=10. 47. Velasco J, Dobarganes C. Oxidative stability of virgin olive 29. Trichopoulou A. Diversity v. globalization: traditional foods oil. Eur J Lipid Sci Tech. 2002;104:661-76. doi: 10.1002/ at the epicentre. Public Health Nutr. 2012;15:951-4. doi: 10. 1438-9312(200210)104:9/10<661::AID-EJLT661>3.0.CO;2 1017/s1368980012000304. -D. 30. Keys A. Mediterranean diet and public health: personal 48. Aiello A, Guccione GD, Accardi G, Caruso C. What olive reflections. Am J Clin Nutr. 1995;61:1321S-3S. oil for healthy ageing? Maturitas. 2015;80:117-8. doi: 10. 31. Fidanza F, Alberti A. The healthy Italian Mediterranean diet 1016/j.maturitas.2014.10.016. temple food guide. Nutr Today. 2005;40:71-8. doi: 10.1097/ 49. Slimestad R, Fossen T, Vågen IM. Onions: a source of 00017285-200503000-00005. unique dietary flavonoids. J Agric Food Chem. 2007;55: 32. Kouris-Blazos A, Itsiopoulos C. Low all-cause mortality 10067-80. doi: 10.1021/jf0712503. despite high cardiovascular risk in elderly Greek-born 50. Virruso C, Accardi G, Colonna-Romano G, Candore G, Australians: attenuating potential of diet? Asia Pac J Clin Vasto S, Caruso C. Nutraceutical properties of extra-virgin Nutr. 2014;23:532-44. doi: 10.6133/apjcn.2014.23.4.16. olive oil: a natural remedy for age-related disease? 33. Farr Louis D. Feasting & fasting in Crete. Athens, Greece: Rejuvenation Res. 2014;17:217-20. doi: 10.1089/rej.2013.15 Kedros Publications; 2001. 32. 34. The University of Sydney. Glycemic index website. 51. Trichopoulou A, Vasilopoulou E, Georga K, Soukara S, 2015/12/18 [cited 2016/02/24]; Available from: http://www. Dilis V. Traditional foods: why and how to sustain them. glycemicindex.com/foodSearch.php. Trends Food Sci Technol. 2006;17:498-504. doi: 10.1016/j. 35. Adom KK, Sorrells ME, Liu RH. Phytochemical profiles tifs.2006.03.005. and antioxidant activity of wheat varieties. J Agric Food 52. Opara EI, Chohan M. Culinary herbs and spices: their Chem. 2003;51:7825-34. doi: 10.1021/jf030404l. bioactive properties, the contribution of polyphenols and the 36. Ostman E, Granfeldt Y, Persson L, Bjorck I. Vinegar challenges in deducing their true health benefits. Int J Mol supplementation lowers glucose and insulin responses and Sci. 2014;15:19183-202. doi: 10.3390/ijms151019183. increases satiety after a bread meal in healthy subjects. Eur J 53. Singletary K. Oregano: Overview of the literature on health Clin Nutr. 2005;59:983-8. doi: 10.1038/sj.ejcn.1602197. benefits. Nutr Today. 2010;45:129-38. doi: 10.1097/NT.0b 013e3181dec789.

52 Historical review Med diets and cuisine 761

54. Ros E. Contribution of nuts to the Mediterranean diet. In: important dietary predictor of survival in older people of Preedy V, Watson R, editors. The Mediterranean diet: an different ethnicities. Asia Pac J Clin Nutr. 2004;13:217-20. evidence-based approach. London: Academic Press; 2015. 71. Trichopoulou A, Bamia C, Trichopoulos D. Anatomy of pp. 175-84. health effects of Mediterranean diet: Greek EPIC 55. Souza RGM, Gomes AC, Naves MMV, Mota JF. Nuts and prospective cohort study. BMJ. 2009;338:b2337. doi: 10. legume seeds for cardiovascular risk reduction: scientific 1136/bmj.b2337. evidence and mechanisms of action. Nutr Rev. 2015;73:335- 72. Grosso G, Marventano S, Yang J, Micek A, Pajak A, Scalfi 47. doi: 10.1093/nutrit/nuu008. L et al. A comprehensive meta-analysis on evidence of 56. Cohen HA, Rozen J, Kristal H, Laks Y, Berkovitch M, Uziel Mediterranean diet and cardiovascular disease: are Y et al. Effect of honey on nocturnal cough and sleep quality: individual components equal? Crit Rev Food Sci a double-blind, randomized, placebo-controlled study. Nutr. 2017;57:3218-32. doi: 10.1080/10408398.2015.11070 Pediatrics. 2012;130:465-71. doi: 10.1542/peds.2011-3075. 21. 57. Molan PC. Re-introducing honey in the management of 73. de Lorgeril M. Mediterranean diet and cardiovascular wounds and ulcers - theory and practice. Ostomy Wound disease: historical perspective and latest evidence. Curr Manage. 2002;48:28-40. Atheroscler Rep. 2013;15:370. doi: 10.1007/s11883-013- 58. Matalas A-L. Disparities within traditional Mediterranean 0370-4. food patterns: an historical approach of the Greek diet. Int J 74. de Lorgeril M, Renaud S, Salen P, Monjaud I, Mamelle N, Food Sci Nutr. 2006;57:529-36. doi: 10.1080/096374806010 Martin JL et al. Mediterranean alpha-linolenic acid-rich diet 41037. in secondary prevention of coronary heart disease. Lancet. 59. Niseteo T, Komes D, Belščak-Cvitanović A, Horžić D, 1994;343:1454-9. doi: 10.1016/S0140-6736(94)92580-1. Budeč M. Bioactive composition and antioxidant potential 75. de Lorgeril M, Salen P, Martin J-L, Monjaud I, Delaye J, of different commonly consumed coffee brews affected by Mamelle N. Mediterranean diet, traditional risk factors, and their preparation technique and milk addition. Food Chem. the rate of cardiovascular complications after myocardial 2012;134:1870-77. doi: 10.1016/j.foodchem.2012.03.095. infarction: final report of the Lyon diet heart study. 60. Samanidou V, Tsagiannidis A, Sarakatsianos I. Simultane- Circulation. 1999;99:779-85. doi: 10.1161/01.cir.99.6.779. ous determination of polyphenols and major purine alkaloids 76. Estruch R, Ros E, Salas-Salvado J, Covas MI, Corella D, in Greek Sideritis species, herbal extracts, green tea, black Aros F et al. Primary prevention of cardiovascular disease tea, and coffee by high-performance liquid chromatography- with a Mediterranean diet. N Engl J Med. 2013;368:1279-90. diode array detection. J Sep Sci. 2012;35:608-15. doi: 10. doi: 10.1056/NEJMoa1200303. 1002/jssc. 201100894. 77. Martínez-González MA, Salas-Salvadó J, Estruch R, Corella 61. Cano-Marquina A, Tarín JJ, Cano A. The impact of coffee D, Fitó M, Ros E. Benefits of the Mediterranean diet: on health. Maturitas. 2013;75:7-21. doi: 10.1016/j.maturitas. insights from the PREDIMED study. Prog Cardiovasc Dis. 2013.02.002. 2015;58:50-60. doi: 10.1016/j.pcad.2015.04.003. 62. Siasos G, Oikonomou E, Chrysohoou C, Tousoulis D, 78. Tyrovolas S, Polychronopoulos E, Bountziouka V, Panagiotakos D, Zaromitidou M et al. Consumption of a Zeimbekis A, Tsiligiani I, Papoutsou S et al. Level of boiled Greek type of coffee is associated with improved adherence to the Mediterranean diet among elderly endothelial function: The Ikaria Study. Vasc Med. 2013;18: individuals living in Mediterranean Islands: nutritional 55-62. doi: 10.1177/1358863x13480258. report from the MEDIS study. Ecol Food Nutr. 2009;48:76- 63. Pal S, Woodford K, Kukuljan S, Ho S. Milk intolerance, 87. doi: 10.1080/03670240802577390. beta-casein and lactose. Nutrients. 2015;7:7285-97. doi: 10. 79. Kremezi A. Nikolas Tselementes. In: Walker H, ed. Cooks 3390/nu7095339. & other people: proceedings of the Oxford symposium of 64. Bell SJ, Grochoski GT, Clarke AJ. Health implications of food and cookery 1995. England: Prospect Books 1996. pp. milk containing β-casein with the A2 genetic variant. Crit 162-9. Rev Food Sci Nutr. 2006;46:93-100. doi: 10.1080/10408390 80. Piscopo S. Socio-ecological factors influencing food choices 591001144. and behaviors of Maltese primary schoolchildren [PhD 65. Wahlqvist ML, Kouris-Blazos A, Trichopoulou A, thesis]. England: University of Birmingham; 2004. Polychronopoulos E. The wisdom of the Greek cuisine and 81. Tourlouki E, Matalas A-L, Bountziouka V, Tyrovolas S, way of life: comparison of the food and health beliefs of Zeimbekis A, Gotsis E et al. Are current dietary habits in elderly Greeks in Greece and Australia. Age Nutr. 1991;2: Mediterranean Islands a reflection of the past? Results from 163-73. the MEDIS Study. Ecol Food Nutr. 2013;52:371-86. doi: 66. Simopoulos AP. The Mediterranean diets: What is so special 10.1080/03670244.2012.707431. about the diet of Greece? The scientific evidence. J Nutr. 82. Davis C, Bryan J, Hodgson J, Murphy K. Definition of the 2001;131:3065S-73S. Mediterranean diet: a literature review. Nutrients. 2015;7: 67. World Cancer Research Fund/American Institute for Cancer 9139-53. doi: 10.3390/nu7115459. Research. Food, nutrition, physical activity, and the 83. UN Educational Scientific and Cultural Organization. prevention of cancer: a global perspective. Washington DC: Convention for the safeguarding of the intangible cultural AICR; 2007. heritage. 2015/12/18 [cited 2015/12/18]; Available from: 68. Iriti M, Vitalini S. Health-promoting effects of traditional www.unisob.na.it/ateneo/c002/ unesco_nomination_file_rl2 mediterranean diets: a review. Pol J Food Nutr Sci. 2012;62: 010.pdf. 71-6. doi: 10.2478/v10222-011-0047-z. 84. Crawford D, Ball K, Mishra G, Salmon J, Timperio A. 69. Widmer RJ, Flammer AJ, Lerman LO, Lerman A. The Which food-related behaviours are associated with healthier Mediterranean diet, its components, and cardiovascular intakes of fruits and vegetables among women? Public disease. Am J Med. 2015;128:229-38. doi: 10.1016/j. Health Nutr. 2007;10:256-65. doi: 10.1017/S13689800072 amjmed.2014.10.014. 46798. 70. Darmadi-Blackberry I, Wahlqvist ML, Kouris-Blazos A, 85. Monsivais P, Aggarwal A, Drewnowski A. Time spent on Steen B, Lukito W, Horie Y et al. Legumes: the most home food preparation and indicators of healthy eating. Am

53 762 S Radd-Vagenas, A Kouris-Blazos, M Fiatarone Singh and VM Flood

J Prev Med. 2014;47:796-802. doi: 10.1016/j.amepre.2014. 102. Tuohy KM, Fava F, Viola R. ‘The way to a man's heart is 07.033. through his gut microbiota’ - dietary pro- and prebiotics for 86. Pes GM, Tolu F, Dore MP, Sechi GP, Errigo A, Canelada A the management of cardiovascular risk. Proc Nutr Soc. 2014; et al. Male longevity in Sardinia, a review of historical 73:172-85. doi: 10.1017/S0029665113003911. sources supporting a causal link with dietary factors. Eur J 103. De Filippis F, Pellegrini N, Vannini L, Jeffery IB, La Storia Clin Nutr. 2015;69:411-8. doi: 10.1038/ejcn.2014.230. A, Laghi L et al. High-level adherence to a Mediterranean 87. Birlouez-Aragon I, Saavedra G, Tessier FJ, Galinier A, Ait- diet beneficially impacts the gut microbiota and associated Ameur L, Lacoste F et al. A diet based on high-heat-treated metabolome. Gut. 2016;65:1812-21. doi: 10.1136/gutjnl- foods promotes risk factors for diabetes mellitus and 2015-309957. cardiovascular diseases. Am J Clin Nutr. 2010;91:1220-6. 104. Darmos-Thodis T. Depressive symptoms and the prevalence doi: 10.3945/ajcn.2009.28737. of cardiovascular risk factors among Greek Australians: pre- 88. Dybing E, Farmer PB, Andersen M, Fennell TR, Lalljie liminary findings from the Mediterranean Islands (MEDIS- SPD, Müller DJG et al. Human exposure and internal dose Australia) study. 2015/12/18 [cited 2015/12/18]; Available assessments of acrylamide in food. Food Chem Toxicol. from: www.atiner.gr/papers/PSY2013-0470.pdf. 2005;43:365-410. doi: 10.1016/j.fct.2004.11.004. 105. Naska A, Oikonomou E, Trichopoulou A, Psaltopoulou T, 89. Negrean M, Stirban A, Stratmann B, Gawlowski T, Trichopoulos D. Siesta in healthy adults and coronary mor- Horstmann T, Götting C et al. Effects of low-and high- tality in the general population. Arch Intern Med. 2007;167: advanced glycation endproduct meals on macro-and 296-301. doi: 10.1001/archinte.167.3.296. microvascular endothelial function and oxidative stress in 106. Trichopoulos D, Tzonou A, Christopoulos C, Havatzoglou S, patients with type 2 diabetes mellitus. Am J Clin Nutr. 2007; Trichopoulou A. Siesta and risk of coronary heart disease. 85:1236-43. Stress Med.1988;4:143-8. doi: 10.1002/smi.2460040306. 90. Riboldi BP, Vinhas AM, Moreira JD. Risks of dietary 107. Kallistratos M, Poulimenos L, Karamanou A, Kouremenos acrylamide exposure: a systematic review. Food Chem. N, Koukouzeli A, Zacharopoulou I et al. Association of 2014;157:310-22. doi: 10.1016/j.foodchem.2014.02.046. mid-day naps occurrence and duration with blood pressure 91. Vlassara H, Striker GE. Advanced glycation endproducts in levels in hypertensive patients. A prospective observational diabetes and diabetic complications. Endocrinol Metab Clin study. J Hypertens.2015;33:e277. doi: 10.1097/01.hjh.0000 North Am. 2013;42:697-719. doi: 10.1016/j.ecl.2013.07.005 468209.39013.ca. 92. Sukkar SG. Mediterranean diet? no, thanks: Mediterranean 108. Oliveira A, Lopes C, Rodriguez-Artalejo F. Adherence to lifestyle! Med J Nutr Metab. 2011;4:79-81. doi: 10.1007/s12 the Southern European Atlantic Diet and occurrence of non- 349-011-0070-y. fatal acute myocardial infarction. Am J Clin Nutr. 2010;92: 93. Serra-Majem L, Bach-Faig A, Raido-Quintana B. 211-7. doi: 10.3945/ajcn.2009.29075. Nutritional and cultural aspects of the Mediterranean diet. 109. Panagiotakos DB, Pitsavos C, Arvaniti F, Stefanadis C. Int J Vitam Nutr Res. 2012;82:157-62. doi: 10.1024/0300- Adherence to the Mediterranean food pattern predicts the 9831/a000106. prevalence of hypertension, hypercholesterolemia, diabetes 94. Panagiotakos DB, Chrysohoou C, Siasos G, Zisimos K, and obesity, among healthy adults; the accuracy of the Skoumas J, Pitsavos C et al. Sociodemographic and lifestyle MedDietScore. Prev Med. 2007;44:335-40. doi: 10.1016/j. statistics of oldest old people (>80 years) living in Ikaria ypmed.2006.12.009. island: the Ikaria Study. Cardiol Res Pract. 2011;2011:1-7. 110. D'Alessandro A, De Pergola G. Mediterranean diet and doi: 10.4061/2011/679187. cardiovascular disease: a critical evaluation of a priori die- 95. Lamisere J. Mindful snacking on the increase. Food tary indexes. Nutrients. 2015;7:7863-88. doi: 10.3390/nu709 Australia. 2015;67:22-3. 5367. 96. Langley S. Australia's snacking habits revealed, Nielsen. 111. Food habits in later life: a cross-cultural study [computer 2015/12/18 [cited 2015/12/18]; Available from: program]. Tokyo: United Nations University Press/Asia Pac http://ausfoodnews.com.au/2014/11/19/australias-snacking- J Clin Nutr; 1995. habits-revealed-nielsen.html?utm_source=feedburner&utm 112. Ciccarone E, Di Castelnuovo A, Salcuni M, Siani A, Giacco _medium=email&utm_campaign=Feed:+AustralianFoodNe A, Donati MB et al. A high-score Mediterranean dietary pat- ws+(Australian+Food+News). tern is associated with a reduced risk of peripheral arterial 97. Duffey KJ, Popkin BM. Energy density, portion size, and disease in Italian patients with type 2 diabetes. J Thromb eating occasions: contributions to increased energy intake in Haemost. 2003;1:1744-52. doi: 10.1046/j.1538-7836.2003. the United States, 1977-2006. PLoS Med. 2011;8:e1001050. 00323.x. doi: 10.1371/journal. pmed.1001050. 113. Radd-Vagenas S, Kouris-Blazos A, Fiatarone Singh M, 98. Sarri KO, Linardakis MK, Bervanaki FN, Tzanakis NE, Flood VM. What is the traditional Mediterranean diet? Kafatos AG. Greek Orthodox fasting rituals: a hidden JNIM. 2015;4:45. doi: 10.1016/j.jnim.2015.12.320. characteristic of the Mediterranean diet of Crete. Br J Nutr. 114. Sotos-Prieto M, Moreno-Franco B, Ordovás JM, León M, 2004;92:277-84. doi: 10.1079/BJN20041197. Casasnovas JA, Peñalvo JL. Design and development of an 99. Cotillard A, Kennedy SP, Kong LC, Prifti E, Pons N, Le instrument to measure overall lifestyle habits for epidemio- Chatelier E et al. Dietary intervention impact on gut logical research: the Mediterranean Lifestyle (MEDLIFE) microbial gene richness. Nature. 2013;500:585-8. doi: 10. index. Public Health Nutr. 2014;18:959-67. doi: 10.1017/S 1038/nature12480. 1368980014001360. 100. Koeth RA, Wang Z, Levison BS, Buffa JA, Org E, Sheehy 115. Ghiselli A, Damicis A, Giacosa A. The antioxidant potential BT et al. Intestinal microbiota metabolism of L-carnitine, a of the Mediterranean diet. Eur J Cancer Prev. 1997;6:S15-9. nutrient in red meat, promotes atherosclerosis. Nat Med. doi: 10.1097/00008469-199703001-00004. 2013;19:576-85. doi: 10.1038/nm.3145. 116. Cannon G. Out of the Christmas box. Public Health Nutr. 101. Tang WH, Wang Z, Levison BS, Koeth RA, Britt EB, Fu X 2004;7:987-90. doi: 10.1079/PHN2004684. et al. Intestinal microbial metabolism of phosphatidylcho- 117. Wahlqvist ML, Lee MS. Regional food culture and devel- line and cardiovascular risk. N Engl J Med. 2013;368:1575- opment. Asia Pac J Clin Nutr. 2007;16(Suppl 1):2-7. 84. doi: 10.1056/ NEJMoa1109400.

54 Historical review Med diets and cuisine 763

118. De Lorenzo A, Alberti A, Andreoli A, Iacopino L, Serrano P, Engl J Med. 2013;369:676-7. doi: 10.1056/NEJMc1306659. Perriello G. Food habits in a southern Italian town (Nicotera) 123. Ros E, Martinez-Gonzalez MA, Estruch R, Salas-Salvado J, in 1960 and 1996: still a reference Italian Mediterranean diet? Fito M, Martinez JA et al. Mediterranean diet and cardio- Diabetes Nutr Metab. 2001;14:121-5. vascular health: Teachings of the PREDIMED study. Adv 119. Helsing E. Traditional diets and disease patterns of the Nutr. 2014;5:330S-6S. doi: 10.3945/an.113.005389. Mediterranean, circa 1960. Am J Clin Nutr. 1995;61:1329S- 124. National Health and Medical Research Council. Australian 37S. dietary guidelines. Canberra: National Health and Medical 120. Fardet A, Boirie Y. Associations between food and beverage Research Council; 2013. groups and major diet-related chronic diseases: an exhaus- 125. US Department of Health and Human Services and US tive review of pooled/meta-analyses and systematic reviews. Department of Agriculture. 2015-2010 Dietary guidelines Nutr Rev. 2014;72:741-62. doi: 10.1111/nure.12153. for Americans, 8th Edition. Washington DC: US Govern- 121. Tilman D, Clark M. Global diets link environmental sus- ment Printing Office; 2015. tainability and human health. Nature. 2014;515:518-22. doi: 126. Woo J, Woo K, Leung S, Chook P, Liu B, Ip R et al. The 10.1038/nature13959. Mediterranean score of dietary habits in Chinese popula- 122. Estruch R, Ros E, Martinez-Gonzalez MA. Mediterranean tions in four different geographical areas. Eur J Clin Nutr. diet for primary prevention of cardiovascular disease. N 2001;55:215-20. doi: 10.1038/sj.ejcn.1601150.

55 CHAPTER THREE

Effect of the Mediterranean diet on cognition and

brain morphology and function:

A systematic review of randomised controlled trials

56 PREFACE

Dementia is a major health problem worldwide that has no effective medical treatment to prevent, delay or modify its course. Observational studies show promising associations between increased adherence to the Mediterranean diet and a slower rate of cognitive decline and brain atrophy with ageing, as well as a reduced risk of mild cognitive impairment (MCI), dementia and conversion of MCI to dementia. However, data from randomised controlled trials are limited. As there is urgency to identify successful interventions that may promote healthy brain ageing, an understanding of the existing literature and the gaps in previous clinical trials is required to progress the field.

The aim of this chapter was to conduct the first systematic review exclusively focussed on randomised controlled trials (RCTs), on the effect of a Mediterranean diet intervention on cognition and brain morphology and function. An additional focus was to determine the quality of the interventions prescribed, including their relation to the ‘traditional’ Mediterranean dietary pattern. A systematic review was selected as the preferred methodology since it provides a rigorous approach for a specific and quantitative focus, which can be reproduced by others.

This is a pre-copyedited, author-produced version of an article accepted for publication in the American Journal of Clinical Nutrition following peer-review. English (Australia) spelling has been retained for the purposes of this thesis. The version of record [Radd- Vagenas, S., Duffy, S. L., Naismith, S. L., Brew, B. J., Flood, V. M., & Fiatarone Singh, M. A. (2018). Effect of the Mediterranean diet on cognition and brain morphology and function: A systematic review of randomized controlled trials. American Journal of Clinical Nutrition, 107(3), 389-404.] is available online at: https://academic.oup.com/ajcn/article/107/3/389/4939347?searchresult=1 or https://doi.org/10.1093/ajcn/nqx070

57 The chapter includes 13 published supplementary materials:

- Sample search strategy for PsychINFO - Quality rating of Mediterranean diet interventions template - Study design – randomised controlled trials (RCTs) with pre/post-test comparisons - Study design – PREDIMED RCT with only post-test comparisons - Participant characteristics - Intervention characteristics – means of implementation of the experimental Mediterranean and control condition - Dietary compliance outcomes - Intervention characteristics – dietary elements/nutritional behaviours prescribed for experimental Mediterranean condition - Intervention characteristics – dietary elements/nutritional behaviours prescribed for control/comparative condition - Quality review of five included studies according to modified PEDro scale - Quality review of five included studies for Mediterranean diet interventions - Cognitive function outcomes - Brain morphology or function outcomes

Shortly prior to finalising this thesis, and since publication of this chapter as a paper in the American Journal of Clinical Nutrition, the authors of one of the included publications (Valls-Pedret et al., 2015) corrected some values within their original paper for a subset of the Prevencion con Dieta Mediterranea (PREDIMED) cohort. They did this as it was discovered that 5/344 of their participants were randomised incorrectly to the same intervention as previously randomised relatives residing in the same household. Their corrected values for individual cognitive test scores were minor (see Appendix Four). However, the two neuropsychological tests (i.e., Rey Auditory Verbal Learning Test (RAVLT) total learning and the Color Trail Test part 2) that were significantly improved in the experimental Mediterranean diet plus extra virgin olive oil arm became non- significant. Importantly, however, there were no appreciable differences in the three composite cognitive domain change scores, which remained significant. Hence, the

58 conclusion in this paper remained the same, i.e., that a Mediterranean diet supplemented with extra virgin olive oil or nuts was associated with improved composite measures of cognitive function over four years of ageing, as compared to a healthy lower fat diet.

Further information related to this chapter is presented in Appendices Three and Four and includes: the PROSPERO protocol and the corrections within PREDIMED for cognitive outcomes.

Part of the work produced in this chapter has been presented at the following conference:

x Radd, S., Duffy, S., Naismith, S., Brew, B., Flood, V., & Fiatarone Singh, M. Effect of the Mediterranean diet on cognition and brain morphology/function: A systematic review and meta-analysis of RCTs. 10th Asia Pacific Conference on Clinical Nutrition, Adelaide, Australia., 26-29th November 2017. x Radd-Vagenas, S., Duffy, S. L., Naismith, S. L., Brew, B. J., Flood, V. M., & Fiatarone Singh, M. A. (2018). Could a plant-based diet affect cognition? Findings from a systematic review of randomized controlled trials on the Mediterranean diet. 7th International Congress on Vegetarian Nutrition, Loma Linda University, Loma Linda, USA, 26-28th February 2018.

59 AUTHORSHIP STATEMENT

The co-authors of the paper: ‘Effect of the Mediterranean diet on cognition and brain morphology and function: A systematic review of randomised controlled trials’ confirm that Sue Radd-Vagenas has made the following contributions:

x Conception and design of the research x Collection and extraction of data x Analysis and interpretation of the findings x Writing of the original draft x Critical appraisal of the content and revisions

As the primary supervisor for the candidature upon which this thesis is based, I can confirm that the above authorship attribution statement is true and correct.

Professor Victoria M Flood

Faculty of Health Sciences

The University of Sydney

25th February 2019

60 Effect of the Mediterranean diet on cognition and brain morphology and function: A systematic review of randomised controlled trials

Sue Radd-Vagenas1, Shantel L Duffy2,3, Sharon L Naismith2,4, Bruce J Brew5,6, Victoria M Flood1,7, Maria A Fiatarone Singh1,8

1Faculty of Health Sciences, The University of Sydney, NSW, Australia;

2Healthy Brain Ageing Program; Brain and Mind Centre, The University of Sydney, NSW, Australia;

3Charles Perkins Centre & Woolcock Institute of Medical Research; Central Clinical School; Faculty of Medicine; The University of Sydney, NSW, Australia;

4Charles Perkins Centre; School of Psychology; Faculty of Science, The University of Sydney, NSW, Australia;

5Faculty of Medicine, University of New South Wales and Department of Neurology, NSW, Australia;

6St Vincent’s Hospital and Peter Duncan Neurosciences Unit for Applied Medical Research, NSW, Australia;

7Western Sydney Local Health District, Westmead Hospital, Westmead, NSW, Australia;

8Hebrew SeniorLife and Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, MA, USA.

Corresponding author

Prof Victoria M Flood

Faculty of Health Sciences

The University of Sydney

75 East Street, Lidcombe NSW 2141, Australia

Tel: +61 2 9351 9001

Email: [email protected]

61 Abbreviations

AD Alzheimer’s Disease

BDNF Brain-derived Neurotrophic Factor

BRIDGE Building Research in Diet and CoGnition study

CDT Clock Drawing Test

CLEED NHS Economic Evaluation Database

CLMCR Cochrane Methodology Register

CVD Cardiovascular disease

DARE Database of Abstracts of Reviews of Effects

DBT Digit Backward Test

DSB Digit Span Backward

DSF Digit Span Forward

EBM Evidence Based Medicine

ELF Excluded Letter Fluency

ES Effect size

EVOO Extra virgin olive oil

FINGER Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability trial

ILF Initial Letter Fluency

ITT Intention to treat

MCI Mild Cognitive Impairment

MedLey Cognitive function and cardiovascular health in the elderly study

MET Metabolic equivalent of task

MMSE Mini-Mental State Examination

PREDIMED Prevención con Dieta Mediterránea

62 PRISMA Preferred Reporting Items for Systematic Reviews and Meta-Analyses

RAVLT Rey Auditory Verbal Learning Test

RCT Randomised controlled trial

ROCF Rey-Osterrieth Complex Figure

Acknowledgements

We would like to thank Elaine Tam for technical advice with database searches, Robert Clark for assistance with figures and Yorgi Mavros, Mark Halaki, Rob Heard and Matthew Hollings for help with statistical analyses.

Sources of support

We thank St Vincent’s Clinic Foundation for funding of a multidisciplinary grant that supported publication of this manuscript. This work was done in partial fulfilment of the requirements for the PhD degree of SRV at The University of Sydney.

Authorship

The author’s contributions were as follows – SRV, VMF and MAFS conceived the study; SRV conducted the literature search; SRV and SLD performed data extraction and statistical analysis; SRV wrote the paper; MAFS, VMF, SLD, SLN and BB provided critical review and analysis; MAFS had primary responsibility for final content.

Conflict of interest

No authors had a conflict of interest to disclose, financially or otherwise.

3.1 Abstract

Background:

Observational studies of the Mediterranean diet suggest cognitive benefits, potentially reducing dementia risk.

Objectives:

To perform the first published review to our knowledge of randomised controlled trials (RCTs) investigating Mediterranean diet effects on cognition or brain morphology/function, with an additional focus on intervention diet quality and its relation to ‘traditional’ Mediterranean dietary patterns.

Design:

We searched nine databases from inception to July 2017 to identify RCTs which tested a Mediterranean diet compared to an alternate diet for cognitive or brain morphology/function outcomes.

Results:

Analyses were based on 66 cognitive tests and one brain function outcome from five included studies (n=1888 participants). The prescribed Mediterranean diets varied considerably between studies, particularly with regards to quantitative food advice. Only 8/66 (12.1%) of individual cognitive outcomes at trial level significantly favoured a Mediterranean diet for cognitive performance, with effect sizes (ESs) ranging from small (0.32) to large (1.66); whereas two outcomes favoured control. Data limitations precluded a meta-analysis. In eight domain composite cognitive scores from two studies, the three (Memory, Frontal and Global function) from PREDIMED (Prevención con Dieta Mediterránea) were significant, with ESs ranging from 0.39 to 1.29. A post-test comparison at a second PREDIMED site found the Mediterranean diet modulates the effect of several genotypes associated with dementia risk for some cognitive outcomes, with mixed results. Finally, the risk of low plasma Brain-derived Neurotrophic Factor (BDNF) was reduced by 78% (OR=0.22; 95% CI: 0.05, 0.90) in those on a Mediterranean diet compared to control at 3 years in this trial. There was no benefit of the Mediterranean diet for incident cognitive impairment or dementia.

Conclusions:

Five RCTs of the Mediterranean diet and cognition have been published to date. The data are mostly non-significant, with small ESs. However, the significant improvements in cognitive domain composites in the most robustly-designed study warrant additional research.

3.2 Introduction

Dementia is a major health problem worldwide, with one new case diagnosed every three seconds and a projected global prevalence of 75 million by 2030 (1). An estimated 5.4 million Americans were affected by the most common type — Alzheimer’s Disease (AD) — in 2016 with more than 96% aged 65 and over (2). Demographic shifts thus suggest that dementia incidence will double by 2050 (3), increasing the burden of disability, and reducing the quality of life for sufferers, carers and families (4, 5). Currently there is no medical treatment to prevent, delay, or modify the course of dementia (6). Therefore, attention has turned to the control of vascular and other risk factors, including hypertension and diabetes, as well as the promotion of lifestyle behaviours such as diet and exercise, which are associated with the rate of cognitive decline and dementia risk (7-11).

With respect to human nutrition, particular nutrients, foods and dietary patterns have been associated with cognitive function and brain morphology/function (12-21). Among these, the Mediterranean dietary pattern has generated recent interest based on observational data suggesting it promotes slower cognitive decline, lower risk of dementia (especially AD), reduced conversion of Mild Cognitive Impairment (MCI) to AD and better overall survival in those affected (22-25). A Mediterranean diet has been defined as a plant-based dietary pattern common to the olive growing areas of the Mediterranean region as described in the late 1950s and early 1960s in the Seven Countries Study (26, 27). The ‘traditional’ pattern was high in unprocessed plant foods (grains, vegetables, fruits, legumes, nuts/seeds and extra virgin olive oil), moderate in fish/shellfish and red wine and low in meat, dairy, eggs, animal fats and discretionary foods. It also included additional cuisine elements, such as use of moist, lower temperature cooking methods, reduced snacking occasions and fasting practice (28).

There are now at least 13 systematic reviews of a Mediterranean diet or dietary pattern considering cognition (18, 19, 25, 29-38). However, these are primarily reviews of observational studies, with only three reviews including 1 to 3 randomised controlled trials (RCTs) (18, 31, 33), and none including a pre-defined protocol to identify RCTs

with cognition and brain morphology/function outcomes. Notably, the experimental evidence linking a Mediterranean diet to cognition or brain morphology/function is limited and has not been comprehensively reviewed. Therefore, our aim was to perform the first systematic review and meta-analysis of all published RCTs investigating the effects of a Mediterranean diet compared to an alternate diet, on any aspect of cognition or brain morphology/function, with additional focus on the quality of the intervention and how it is related to the ‘traditional’ Mediterranean dietary pattern.

3.3 Methods

We used a protocol prospectively designed and registered with PROSPERO (registration number: CRD42011100485 http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42015024088) and followed the statement and general principles of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (39).

3.3.1 Criteria for study inclusion

Studies were included if they met all the following criteria:

(1) RCT design or post-test comparison within an RCT;

(2) full length article published in peer-reviewed journal in any language;

(3) only human participants of any age;

(4) intervention included a Mediterranean diet (with/without supplied food); for the purpose of this review, all studies that described their intervention as a Mediterranean diet or Mediterranean-style dietary pattern were included;

(5) comparator included an alternate or usual diet;

(6) outcomes included at least one measure of cognition, or brain morphology/function;

(7) outcomes were measured at two time points or compared at one time point during follow-up between the intervention and comparator groups.

Studies were excluded if:

(1) acute/single exposures were tested and duration of study period was less than 10 days, at which time point it could be expected that physiological and cognitive outcomes of repeated/chronic exposures might be observed;

(2) single foods/nutrients or a partial Mediterranean diet was tested (e.g. low fat, low carbohydrate, low calorie);

(3) additional non-dietary modalities (e.g., exercise) were not equally prescribed to intervention and comparator groups.

3.3.2 Search strategy

The following nine electronic databases were searched from inception to July 2015 with a final update in December 2017: MEDLINE, PreMedline, AMED, all EBM review databases including Cochrane Database of Systematic Reviews (DSR), ACP Journal Club, Database of Abstracts of Reviews of Effects (DARE), Cochrane Central Register of Controlled Trials (CCTR), NHS Economic Evaluation Database (CLEED), Cochrane Methodology Register (CLMCR), Health Technology Assessment (CLHTA), CINAHL, EMBASE, PubMed, PsycINFO, Web of Science. Between the comprehensive searches, the lead author reviewed reference lists of eligible trials and review articles, hand searched relevant websites and journals and was in contact with known experts in the field to identify additional studies.

A search strategy was developed for each database. This included a combination of 'Intervention' AND 'Outcome' terms (Figure 3.1) and did not include 'Population', 'Comparator' or 'Study design' terms. We searched for ‘search terms’ in all fields and no language or date restrictions were applied to the search strategy in order to maximize search sensitivity. A sample full search strategy is provided for the PsycINFO database (Supplementary Material 3.1).

3.3.3 Study selection and data extraction

One reviewer (SRV) conducted the search, removed duplicates then screened papers by title and abstract according to eligibility criteria. Papers retrieved for full evaluation were appraised by two reviewers (SRV and SLD) who extracted data into pre-designed, piloted tables. Any disagreements were resolved by consensus with two additional reviewers (VMF and MAFS).

Multiple fields were extracted as outlined in our protocol, based on the following categories: (1) study design, (2) participant characteristics, (3) intervention characteristics, (4) dietary compliance outcomes, (5) cognitive function outcomes, (6) brain morphology or function outcomes, (7) study quality, (8) funding source. Cognitive outcomes were categorized under their respective domains according to Strauss, et al. (40), except for Rey-Osterrieth Complex Figure (ROCF) copy scores which we categorized as visuo-spatial.

Additional outcomes relevant to cognition or brain morphology/function were searched for in all publications reporting on the same study. Authors were contacted to clarify the definition of ‘meat’ (41, 42) and confirm for two publications (43, 44) whether study participants and cognitive evaluations overlapped since they were derived from the same node of PREDIMED. If data were presented in figures (45), a customized MATLAB R2015b script (MathWorks, Natick, Massachusetts, USA) was used to extract data. Unadjusted means and differences were extracted from papers unless these were unavailable, in which case fully adjusted values were extracted and footnoted. Where the standard deviation (SD) was unavailable, this was calculated from reported confidence intervals (CI):

ଽହΨ ஼ூ ିଽହΨ ஼ூ ܵܦ = ξܰ × ೠ೛೛೐ೝ ೗೚ೢ೐ೝ. ଶ×௧ ௩௔௟௨௘

When required, change data were calculated as follow-up minus baseline data, or follow- up data were imputed using baseline plus change data.

69 3.3.4 Data synthesis and analysis

Between-group effect sizes (ESs) for various tests within cognitive domains were calculated by subtracting the mean change in the control group from the mean change in the experimental group and dividing by the pooled SD at baseline:

௫ҧ ି௫ҧ (ܧܵ = భ మ). ௦

Individual ESs were then corrected for small sample size (Hedges’ g ES) (46) and 95% CIs calculated. For post-test comparisons, Cohen’s d ESs were calculated by subtracting follow-up data in the control group from follow-up data in the experimental group and dividing by the pooled post SD. We used Coe’s calculator (http://www.cem.org/effect- size-calculator) to generate data from parallel trials in extraction tables. For the cross- over trial (42) where only change scores were available, we used the F test calculator from Campbell’s Collaboration (https://www.campbellcollaboration.org/escalc/html/EffectSizeCalculator-SMD5.php) to calculate ESs and CIs and then imputed pooled baseline SD (ܵܦ = ܯܦ/ܧܵ).

Cochrane Review Manager Software, RevMan Version 5.3 (The Cochrane Collaboration, Copenhagen) was used to create forest plots when baseline and follow-up data were available. A single primary outcome for a given test (e.g., time rather than errors) was selected for plotting. The final outcome measure was used where more than two time points were provided. The ESs were graphed and described according to Cohen’s LQWHUSUHWDWLRQRIµWULYLDO¶ µVPDOO¶ WR µPRGHUDWH¶ to <0.80) DQGµODUJH¶ (47). For graphical representation of ESs only, signs were reversed in some cases (e.g., processing speed) so that a positive ES represented an improvement with Mediterranean diet. We did not plot cognitive composites (e.g., Executive Domain) since categorization of cognitive tests within domains varied across studies. However, where z scores were available for composites we used these to calculate and discuss ESs.

Where two Mediterranean interventions were compared against one control arm, the control number of participants (n) was divided by two, consistent with the recommendations of the Cochrane Handbook (48, 49). For the cross-over trial we divided

the n of both groups by two because the authors did not conduct a paired sample analysis (50). Statistical heterogeneity was tested using the Chi2 test (significance level: 0.1) and I2 statistic (0% to 40%: might not be important; 30% to 60%: may represent moderate heterogeneity; 50% to 90%: may represent substantial heterogeneity; 75% to 100%: considerable heterogeneity) (51). A random effects model was used as studies included in the analysis were assumed to be a random sample of all possible studies that met the inclusion criteria for the review (52).

We provided a narrative summary of the standardised mean differences and did not pool data within cognitive domains wherever there was: a) a high I2 E DFRPELQDWLRQ of parallel and cross-over trial data without a paired sample analysis (50), or c) outcomes for a given domain were not available from at least three different trials. We considered the use of robust variance estimation and multilevel meta-analysis for the cross-over trial group, but the number of studies was insufficient (53).

3.3.5 Quality assessment

Two reviewers (SRV and SLD) independently assessed the quality of included studies according to the Physiotherapy Evidence Database (PEDro) scale, designed to assess quality of RCTs for major sources of bias (54). However, we modified PEDro criteria as follows: a) for studies with multiple publications, a criterion was deemed to be met if information was retrievable from any publication; b) for post-test comparisons, the criterion for baseline similarity with important prognostic indicators was taken to be satisfied if commonly accepted risk factors for cognitive outcomes (e.g., age) were similar across groups at baseline.

As there is inconsistency in the definition of the Mediterranean diet in the literature, seven additional quality indicators were developed, including 19 unique elements describing the ‘traditional’ Mediterranean dietary pattern and cuisine (28). However, these were not included in scoring as they have not been validated (See Supplementary Material 3.2 for quality rating of Mediterranean diet interventions template). For each of these 19

elements, we rated whether studies: a) prescribed the element, b) specified a quantity, and c) met our minimum criterion for a ‘traditional’ pattern. Discrepancies in rating were resolved by consensus with a third reviewer (VMF).

3.4 Results

3.4.1 Study selection

The initial keyword search identified 6385 titles. Following exclusions based on title or abstract and additional hand searching, 30 full text articles were retrieved for evaluation. One additional RCT was identified after the initial search. Among the 31 selected articles, 22 were excluded based on eligibility criteria: not RCT design (n=12), not Mediterranean diet (n=6), cognitive/brain outcomes not reported (n=3), and secondary intervention modality (exercise) not provided to both groups (n=1). The remaining nine articles were included for review, representing five RCTs. The largest of these (n=1527), PREDIMED (Prevención con Dieta Mediterránea), was a multi-site study that included some cognitive outcomes from two sites. One site (Navarra) published four papers (43, 44, 55, 56) and correspondence with the authors confirmed that the 2013 publications did not include overlap of participants or outcomes, so all were included. The PREDIMED publication from Barcelona (57) also did not overlap with the Navarra node participants.

3.4.2 Study characteristics

3.4.2.1 Study design and participants

All RCTs were of parallel design, except for one cross-over trial (42) (Supplementary Material 3.3). PREDIMED also provided post-test comparisons (43, 44, 55, 56) (Supplementary Material 3.4). Across the five RCTs, the duration of interventions ranged from 10 days to 6.5 years. These community dwelling cohorts (Table 3.1) ranged in size from 24 to 1497 (median 166), were 65% female (range 52-100%), with mean age 65.24

years (SD=6.15) (range of individual study mean ages 21.1 to 72.1). Only one trial (PREDIMED) was conducted in a Mediterranean region/cultural cohort.

Three studies (42, 58) including one PREDIMED sub-study (57) reported normal cognitive status at baseline. Comorbidities were specifically excluded at screening by two studies (42, 58) and reported at baseline by another two (45, 57), with the most prevalent being hypertension, diabetes and hyperlipidaemia. Relevant medication use was poorly reported, with only one PREDIMED sub-study describing medications that were most commonly used for hypertension, hyperlipidaemia and diabetes (57). Only PREDIMED, in publications for three of its sub-studies, provided data for one or more genotypes related to dementia risk (43, 44, 56) (Supplementary Material 3.5).

Four studies (42, 45, 57, 58) reported BMI at baseline, with average overweight status of 27.78 kg/m2 (SD=3.82). Only PREDIMED reported habitual physical activity and educational level at baseline in four of its publications and adjusted for these covariates (43, 44, 56, 57) (Supplementary Material 3.5). Reported physical activity level was very high in this Spanish study, with up to approximately four times the upper recommended MET minutes per week advised for good health to Americans (https://health.gov/paguidelines). Educational level was reported, but not adjusted for, in the Mediterranean Diet for Cognitive Function and Cardiovascular Health in the elderly study (MedLey) (58). The remaining trials did not report measurement or adjustment for these factors.

We attempted to extract habitual sedentary time, size of family unit, whether participant was the primary cook, socioeconomic status/financial level and ethnicity but these were not reported by any study.

3.4.2.2 Intervention

Behavioural strategies utilised to implement diet. All studies provided individual dietary counselling sessions to the intervention group, ranging widely from one session (41, 42)

to eight mixed individual and group sessions (45), to both individual and group sessions every three months on average, resulting in 12-26 sessions of each type, depending on sub-study length (57). Telephone calls or texts for additional supervision were not reported by any studies. Logging of dietary intake was reported by four studies with two using daily food records (41, 42), one a daily checklist (58) and another (45) a 7-day food record but only at baseline and 12 weeks (Supplementary Material 3.6). The use of dietary biomarkers to check on dietary compliance was reported by two studies, with PREDIMED measuring urinary hydroxytyrosol and plasma alpha linolenic acid (markers of specific food intake, namely olive oil and nuts, respectively) and MedLey measuring erythrocyte fatty acids, plasma carotenoids and urinary metabolites (markers of the dietary pattern, namely monounsaturated to saturated fatty acid ratio, fruit and vegetable and mineral intake, respectively) (Supplementary Material 3.7).

Dietary elements. The prescribed Mediterranean diet varied considerably between the five studies, particularly with regards to quantitative advice (Table 3.2). All studies provided some advice on the intake of fruit, vegetables, fish/seafood but only PREDIMED and MedLey gave quantitative guidance for these foods (57-59). Four studies advised on the use of fats other than olive oil (41, 42, 45, 57). Four (41, 42, 57- 59) also gave recommendations for the consumption of nuts, dairy, sweets and sweetened drinks, but amounts for these foods were variably specified by 2 to 4 of these studies. Both qualitative and quantitative advice was provided by three studies for the intake of wine (42, 57-59) and by four for red/processed meats (41, 42, 57-59). However, only two studies provided this level of comprehensive guidance for consumption of olive oil, legumes and eggs (57-59). Four out of five studies provided some recommendation on poultry (41, 42, 57-59) but only three of these specified how much to consume (41, 42, 58, 59). PREDIMED was the only study to promote use of sofrito (a rich sauce made with tomato and onion simmered in olive oil, widely consumed in the Mediterranean) and traditional cooking methods. Despite grain foods, such as bread, playing a major role in the Mediterranean diet, only two studies (41, 58, 59) provided any guidance on their intake, with MedLey specifying quantities (58, 59), as shown in Table 3.2. Thus, overall, most of the experimental diets differed in substantially important ways from the elements and quantities of ‘traditional’ Mediterranean diets (28) (Supplementary Material 3.8).

Control group intervention. The effect of a Mediterranean diet was compared to either a waiting list, usual diet or a low fat control group, depending on the trial (Supplementary Material 3.9). A low fat diet was prescribed to controls in PREDIMED based on American Heart Association guidelines (57). Three studies (41, 42, 58) from Australia recommended unspecified ‘usual diet’. One UK study (45) included a waiting list arm, in which participants were not discouraged from making dietary changes to their usual diet, reporting that some participants elected to do so. A dietitian administered the control intervention for only two studies (57, 58). Three of the five studies (41, 42, 45) required controls to keep a food record during the intervention period.

Study quality. The overall quality of the included studies according to the modified PEDro tool was moderate with four out of five scoring 6 or above (range: 4-8/10). The most common limitations were in the areas of subject and therapist blinding, which is unavoidable in dietary trials. Three out of five studies (41, 42, 45) did not meet criteria for assessor blinding and intention to treat analysis (ITT), despite PEDro being less rigorous for ITT compared to CONSORT (60), potentially inflating reported quality.

With respect to our additional criteria for the intervention itself, only the MedLey study met the criterion for participant perception of diet burden or benefit, relevant for translating research findings into the public domain. Only PREDIMED and MedLey had their diets designed and administered by a dietitian; specified minimum amounts of Mediterranean foods to be consumed by participants; or recommended a diet that met at least 8/19 of our minimum criterion for the ‘traditional’ Mediterranean diet. Further, just two studies reported on diet tolerance (45, 57). However, all studies reported on frequency and setting for their diet instruction and assessment of dietary compliance. Figure 3.2 illustrates the percentage of included studies fulfilling the combined quality criteria. (Supplementary Material 3.10 and 3.11 show details of quality rating).

When considering the quality of interventions as they relate specifically to the ‘traditional’ Mediterranean dietary pattern and cuisine, the majority (4/5) (41, 42, 57-59) met our minimum criterion for sweetened drinks; only three studies (41, 42, 58, 59)

achieved this for red/processed meats and sweets; two studies for olive oil, nuts, fruit, vegetables and fish/seafood (57-59); two for poultry (41, 42); one study for other fats, legumes and sofrito (57) and one for grain foods (58, 59) and wine (42). No studies fulfilled our minimum criteria defining the traditionally low intake of dairy products, adequate water intake, use of moist cooking methods and eating main meals in company as shown in Table 3.2.

Further limitations not captured with the above tools, which could have introduced bias or limited robustness of the interventions include: a) use of a waiting list as the control group (45), where participants were not discouraged from making dietary or lifestyle changes; b) no wash out period for cross-over design (42); and c) outcomes tested only in those participants who arrived at scheduled visits for post-test comparisons (43). In addition, the MedLey study provided recommendations for dietary serves at four levels of energy intake ranging from 6400-9600 kJ per day (59). While the optimal or absolute minimum amounts of key Mediterranean foods for cognition and brain function are not yet known these, and traditional levels of intake, may not have been reached at the lower energy levels for some foods. For example, for energy level 4 (6400 kJ), MedLey recommended 1-2 tablespoons of extra virgin olive oil per day, compared to at least 4 tablespoons reported to have been consumed in the Cretan Mediterranean diet in the 1960s (61).

Study adherence and tolerance. Only two out of five studies (42, 45) reported participant attendance at intervention sessions, ranging from 86-100%, respectively. Dropout rates for the RCTs varied from a median of 13.1% (range 0-19.5%) in the experimental groups to 10.7% (range 0-33.1%) in the control groups, increasing with trial duration. Adverse events were poorly reported, with only 1 out of 5 studies (45) reporting that no participants experienced adverse effects from the intervention. PREDIMED authors have indicated that such data were collected (62), but this has not yet been published, to our knowledge, for the sub-cohorts in the included studies.

3.4.3 Cognitive outcomes

3.4.3.1 Overview

The five included studies encompassed 1888 participants and reported on 66 cognitive test outcomes. Effect sizes were able to be calculated from reported data or authors’ information in 63/66 (95.5%). We extracted mean cognitive scores for all studies and outcomes and report these in Supplementary Material 3.12. We plotted the most indicative outcomes for those tests, which had both baseline and follow-up data to illustrate the scope of current research (Figures 3.3-3.5). Outcomes and original Cohen’s d ESs from trials where cognition was measured at only one time point are not plotted and can be viewed in Supplementary Material 3.12. However, these are included in the text summary of overall ESs within each domain. Overall, only 8/66 (12.1%) of the individual cognitive outcomes were statistically significant at the trial level in favour of a Mediterranean diet. Two additional cognitive outcomes (3.0%) significantly favoured controls, however these were from studies which either had a ‘waiting list’ control group, which may have adopted some elements of the experimental diet (45) or did not include a washout where the design was cross-over (42). Two studies reported composite domain z scores (57, 58) (Supplementary Material 3.12), although component tests were not the same. For PREDIMED (57) the three composites included the following tests: 1) Memory (Rey Auditory Verbal Learning Test (RAVLT) and Verbal Paired Associates from the Wechsler Memory Scale); 2) Frontal (Digit Backward Test from the Wechsler Adult Intelligence Scale (DBT) and Color Trail Test (parts 1 and 2); 3) Global (all individual tests within study, including the Mini-Mental State Examination). For MedLey, the five composites and their included tests were: 1) Executive (Stroop (interference control), Initial Letter Fluency (ILF), Excluded Letter Fluency (ELF) and Tower of London (TOL)); 2) Memory (RAVLT, Digit Span Forward (DSF); Digit Span Backward (DSB) and Letter-Number Sequencing (LNS)); 3) Speed of Processing (Symbol Search and Coding); 4) Visual-Spatial Memory (The Benton Visual Retention Test (BVRT)); 5) Global (all individual tests within study). The three composites from PREDIMED (Memory, Frontal, Global) were significantly improved with the intervention, with ESs ranging from small (0.39) to large (1.29), to different degrees by the EVOO or Nuts arm

of the trial, as described below. By contrast, the MedLey study composites were not improved by the Mediterranean diet. No results were able to be pooled for a meta-analysis due to pre-defined limitations with regards to heterogeneity and number of studies available (see above).

3.4.3.2 Global cognition

Three tests from one study (PREDIMED) reported on global cognition, including one parallel (57) (Figure 3.3A) and two post-test (43, 44) (Supplementary Material 3.12) comparisons with ESs ranging from 0.15 to 0.63, one being significant. While PREDIMED used two experimental groups, only the Mediterranean Nuts group performed significantly better than control with a moderate ES (ES=0.63; 0.26, 0.99) for MMSE (44) at the 6.5 year time point in the Navarra sub-study site (Supplementary Material 3.12). Limited outcomes from this single study precluded data pooling for this domain.

In addition to the single tests noted above, two studies (57, 58) provided composite global cognition z scores as one of their primary cognitive outcomes. In the PREDIMED sub- study from Barcelona (57) with participants selected for cardiovascular risk, there were large and significant ESs for Global Cognition composite for both Mediterranean diet types: ES=1.29 (0.65, 1.92) for Mediterranean EVOO and ES=1.12 (0.44, 1.81) for Mediterranean Nuts at 4.1 years. The MedLey study (58) of 24 weeks duration with healthy participants, by contrast, did not report benefits for the Mediterranean diet on its global (Total Cognitive Function) composite score. Overall, there is evidence of large, significant global cognitive benefits for the Mediterranean diet in the most robust trial.

A post-test comparison from the PREDIMED Navarra node (56), at 6.5 years, reported for the first time on Mediterranean diet-gene interactions for cognition, pooling the EVOO and Nuts Mediterranean groups for this analysis. These investigations were undertaken as genome-wide association studies have identified several single nucleotide polymorphisms that increase the risk of late-onset AD (63, 64). A significant interaction

was found between the CLU-rs11136000 gene and the Mediterranean diet, with significantly improved performance (ES=0.45; 0.12, 0.79) on the MMSE in the Mediterranean diet groups only for carriers of the harmful T allele (which occurs at a minimum in 40% of the population). By contrast, there was a small significant benefit (ES=0.40; 0.09, 0.70) for MMSE scores in the Mediterranean diet group for homozygotes with both protective GG alleles for CR1-rs3818361 compared to other genotypes. Finally, MMSE scores were significantly better in the Mediterranean group compared to controls regardless of ApoE4 status (ApoE4 non-carriers: p=0.007; ApoE4 carriers: p=0.037). Data were not provided to allow comparison of ESs relative to ApoE status, however.

3.4.3.3 Attention

Seven tests from three studies reported on attention, including four parallel (41, 45) and three cross-over (42) comparisons (Figure 3.3B). The ESs ranged widely IURPí15.95 to 0.63, and no significant benefits were attributed to the Mediterranean diet. Unexpectedly, the control group from one 12-week study (45) performed significantly better on one test (sustained attention task (SAT)) (45). Non-robust design features may explain these results as the control arm was a waiting list where participants were not discouraged from making dietary changes, and some elected to do so, suggesting potential contamination. With limited outcomes and heterogeneity between studies (I2=97%) precluding pooling, definitive conclusions about this domain are not possible.

3.4.3.4 Working memory

Thirteen tests from four studies reported on working memory, including eight parallel (41, 57, 58) and three cross-over (42) comparisons (Figure 3.3C) as well as two post-test (44) comparisons (Supplementary Material 3.12). These ESs ranged widely IURPí1.49 to 0.74 with three being significant. There was a moderate ES benefit for Mediterranean Nuts on the DSB test (ES=0.74; 0.08, 1.40) (57) (Figure 3.3C) and a small ES benefit for Mediterranean EVOO on DSF (ES=0.47; 0.11, 0.83) post-test comparison (44) (Supplementary Material 3.12) in the PREDIMED study. By contrast, the control group performed significantly better on 3-back correct responses test in the cross-over study (42), although interpretation is limited by this study design as noted above. Heterogeneity

across studies was moderate (I2=46%) but the number of RCTs was insufficient to warrant a multilevel meta-analysis to control for the correlation between subjects in cross-over trials and multiple outcomes of each study. Overall, no consistent evidence of Mediterranean diet benefit was present.

3.4.3.5 Processing speed

Nine tests from five studies reported on processing speed, including six parallel (41, 45, 57, 58), two cross-over (42) and one post-test (44) comparison. The ESs ranged from í0.55 to 0.51, none of which were significant (Figure 3.4A). Heterogeneity across studies was low (I2=0%) but the number of studies was insufficient to warrant a multilevel meta- analysis to control for the correlation between participants in cross-over trials and multiple outcomes of each study. A composite z score for Processing Speed from MedLey was also non-significant (58). Thus, there is no evidence for benefits of a Mediterranean diet intervention on processing speed in the existing literature.

3.4.3.6 Verbal and visual memory

Nineteen tests from five studies reported on verbal and visual memory, including 10 parallel (41, 45, 57, 58) and four cross-over (42) comparisons (Figure 3.4B) and five post- test (44) comparisons (Supplementary Material 3.12). The ESs ranged from í0.55 to 1.66, with only one large ES in the cross-over trial (42) without a wash out period favouring the Mediterranean diet for the immediate word recall-correct responses test (ES=1.66; 0.67, 2.66). Valls-Pedret, et al. (57) also reported in their paper that performance on RAVLT total learning test by the Mediterranean EVOO group was significantly different to control (p<0.05), which may be likely due to appropriate statistical adjustment which we were unable to make. Heterogeneity across studies was moderate (I2=46%) but the number of studies was insufficient to warrant a multilevel meta-analysis. There was also a small but significant ES for the composite z scores for Memory domain in PREDIMED for the Mediterranean Nuts group (ES=0.39; 0.05, 0.73) (57) but not in MedLey. Overall, there appears to be limited and inconsistent evidence of benefit of the Mediterranean diet for verbal and visual memory.

3.4.3.7 Visuo-spatial

Three tests from one study (44) reported on visuo-spatial domain based on post-test comparisons after 6.5 years, with ESs ranging from 0.00 to 0.32 (Supplementary Material 3.12). At this time point, only the Mediterranean EVOO group performed significantly better on the Clock Drawing Test (CDT) with a small ES (ES=0.32; 0.04, 0.60). The limited outcomes precluded data pooling. A composite of z scores was only available for Visual-Spatial Memory domain from MedLey (58) but this was not significant. Overall, for visuo-spatial function, there is limited evidence of benefit for a Mediterranean diet.

With respect to genetic predictors of performance on the CDT, in post-test analyses from PREDIMED (56), individuals with the harmful T allele for CLU-rs11136000 did significantly better on the Mediterranean diet compared to controls, with a small ES (ES=0.42; 0.09, 0.75). By contrast, for both CR1-rs3818361 and PICALM-rs3851179, those with protective alleles receiving the Mediterranean diet significantly outperformed control (ES=0.33; 0.03, 0.64 and ES=0.40; 0.03, 0.76, respectively). Cognitive performance on the CDT was only better than control for ApoE4 non-carriers in the Mediterranean EVOO group (p<0.001). Notably, these gene interaction patterns were the same across both the MMSE and CDT tests for the CLU-rs11136000 and CR1-rs3818361 genes, where those with harmful and beneficial genotypes, respectively, responded better to the Mediterranean diet. However, while APOE4 carrier status had no effect on MMSE performance, non-carriers (beneficial genotype with lower risk of dementia) responded better on the CDT of visuo-spatial function.

3.4.3.8 Language

Six tests from two studies reported on language, including three parallel (57, 58) (Figure 3.5A) and three post-test (44) (Supplementary Material 3.12) comparisons. The ESs UDQJHGIURPí0.48 to 0.45. Only one test (phonemic verbal fluency-letter F, A, S) (44) at trial level favoured Mediterranean EVOO with a small ES at 6.5 years follow-up (ES=0.45; 0.09, 0.81) (Supplementary Material 3.12). Heterogeneity across studies was low (I2=0%) but the number of studies was insufficient to pool data for a meta-analysis.

Thus, there is minimal evidence for a benefit of the Mediterranean diet on language in current literature.

3.4.3.9 Executive function

Five tests from two studies reported on executive function, including three parallel (57, 58) (Figure 3.5B) and two post-test (44) (Supplementary Material 3.12) comparisons. Overall, the ESs for this domain ranged from í0.16 to 0.75 with only one being significant. Specifically, the Mediterranean EVOO group performed better than control on the Color Trail Test 2 (CTT 2) at 4.1 years in PREDIMED with a moderate ES (ES=0.75; 0.14, 1.35) (Figure 3.5B). Heterogeneity across studies was low (I2=29%) but the number of studies was insufficient to pool data for a meta-analysis. PREDIMED (57) also reported a composite of z scores for executive function (described as Frontal Cognition), finding a moderate significant ES for Mediterranean Nuts (ES=0.72; 0.06, 1.38) and large significant ES for Mediterranean EVOO (ES=1.02; 0.40, 1.61). By contrast, the ES for the Executive Function composite from MedLey was not significant (58). Thus, overall, there is some evidence for moderate to large benefit in this domain, although the data supporting this are limited to one study.

3.4.3.10 Motor control

No studies reported on outcomes within this domain.

3.4.4 Brain morphology/function outcomes

Only one post-test comparison from PREDIMED Navarra (55) (n=243) reported on a plasma neuroplasticity biomarker (BDNF), which promotes connectivity between neurons, at 3 years in a randomly selected subsample from the original cohort (Supplementary Material 3.13). Overall, there were no differences between groups in plasma BDNF levels. However, the Mediterranean Nuts group had a 78% decreased risk of plasma BDNF levels below the 10th percentile (OR=0.22; 95% CI: 0.05, 0.90) at this time point. When analyses were stratified according to prevalence of different diseases at baseline there were no significant differences between groups with our calculated ESs

UDQJLQJ IURP í0.17 to 0.11. However, the authors of this PREDIMED comparison reported that significantly higher BDNF levels (p<0.05) were found in those with depression at baseline assigned to the Mediterranean Nuts diet, compared to control, for whom baseline BDNF levels were also available (n=37), presumably due to appropriate statistical adjustments, which we were unable to make. Conclusions are limited as plasma BDNF was not measured at baseline for most participants, which could have introduced bias.

3.4.5 Incident MCI and dementia

Three PREDIMED post-test comparisons (43, 44, 57) included incidence rates for MCI and one for dementia (43) at 4.1 or 6.5 years but these were not significantly different between groups. For the smaller cohort from the Navarra node (44) incident MCI was 7.80%-11.80% in the two Mediterranean diet groups vs. 19.30% in the control group whereas a sub-study of a larger cohort from this node (43) reported incident MCI as 5.13%-5.40% vs. 6.53%, respectively. The latter sub-study however may include measurement error, as data were ascertained from medical records rather than actual assessment, which could underestimate cases. In the post-test comparison from Barcelona (57) incident MCI was 7.1%-13.4% vs. 12.6% (adjusted p=<0.28) in Mediterranean diet and control groups, respectively. Incident dementia was reported from the Navarra node as 1.70%-3.42% in the Mediterranean diet groups vs. 4.83% in the control (43) but, for the reason outlined above, may also underestimate cases, although both MCI and dementia outcomes in PREDIMED were blindly assessed, so this bias should have affected both groups equally. Thus, there is no experimental evidence at this time to corroborate observational data (30) showing an association between followers of a Mediterranean dietary pattern and incident dementia.

3.5 Discussion

3.5.1 Key findings

This is the first systematic review to our knowledge to evaluate the effect of a Mediterranean diet on cognitive function or brain morphology/function exclusively in RCTs. Convincing evidence for a significant beneficial effect of the Mediterranean diet on cognitive function or brain morphology/function across all available studies and outcomes is lacking, and there was a lack of dietary uniformity and trial design precluding meta-analysis. However, at the individual trial level, there were significant ESs, ranging from 0.32 to 1.66, in favour of a Mediterranean diet for eight test outcomes related to global cognition, working memory, verbal and visual memory, visuo-spatial, language and executive function domains.

Importantly, significant ESs were found for composite z scores for Global Cognition, Memory and Frontal Cognition domains within PREDIMED (57), ranging from 0.39 to 1.29. The size of these composite effects may be clinically relevant, as the controls z scores declined by íWRí0.38 over the trial, whereas the Mediterranean diet groups either remained stable or improved by í0.05 to 0.23 z scores (Supplementary Material 3.12). This suggests the Mediterranean diet may attenuate cognitive decline, or improve cognition, despite four years of aging.

The success of the PREDIMED sub-study from Barcelona should be interpreted in light of its design features and cohort selection. Robust design features include diet design and administration by dietitians, prescription and provision of actual quantities of targeted foods, consistency with ‘traditional’ Mediterranean dietary patterns, long duration (4.1 years) and measurement of cognition using gold-standard neuropsychological tests at baseline and follow-up. It is likely that the cognitive benefits seen in PREDIMED may represent a conservative estimate of potential efficacy, since the two Mediterranean diets were compared to a healthy low fat diet, which is an intervention in itself. Additionally,

PREDIMED was conducted in highly active people living within a Mediterranean culture, so the external generalisability of their results would need to be proven.

There is insufficient evidence relating genotype to Mediterranean diet benefits at this time, as mixed results were obtained for four genes only studied in PREDIMED. No studies reported brain morphology outcomes, and only one study found a potential benefit for BDNF in a subset of participants. Thus, suggestions of protective effects on brain morphology from observational studies (15, 65-68) could not be corroborated. Incident MCI and dementia rates did not differ significantly, although no study was powered for these as primary outcomes and numbers of cases were low.

To put our findings in perspective, the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER), which used a multi-domain intervention including diet (non-Mediterranean), exercise, cognitive training and vascular risk monitoring over two years reported an ES of 0.13 for overall cognitive performance (69). Meta-analyses of RCTs on the effect of aerobic exercise (70) and computerized cognitive training (71) on cognitive performance have reported ESs from 0.12 to 0.16 and 0.08 to 0.31, respectively. Thus, the most robust trial of the Mediterranean diet to date compares favourably to these other interventions.

Potential reasons for inconsistent efficacy of a Mediterranean diet across RCTs compared to more homogeneous benefits seen in previously reviewed observational studies include: a) short intervention time to see significant change in physically and cognitively healthy cohorts in four of the five studies, b) use of varied Mediterranean diet interventions which may not have provided an adequate ‘dose’ of foods important for cognition, and c) lack of consistent and sensitive methodologies to measure cognitive change.

3.5.2 Mechanisms

Potential mechanisms for a protective effect of the Mediterranean diet for dementia (30) as well as cardiovascular disease (CVD) and its risk factors (72), type 2 diabetes (73, 74),

breast cancer (75), macular degeneration (76), osteoarthritis (77), and fatty liver (78), have been extensively reviewed. These include anti-inflammatory, antioxidant and microbiome effects, which are also associated with multiple vascular risk factors (16, 79). The Mediterranean dietary pattern could potentially contribute to better health and cognition by lowering the glycaemic load (80) and advanced glycation end products (AGEs) (81); and increasing dietary fibre, omega-3, polyphenols (16), arginine (82), nitrate (83) and melatonin (84), which have been suggested to influence mechanistic pathways. Unfortunately, none of the trials reviewed reported vascular risk factor profiles/biomarkers in relation to cognitive health outcomes, so it is unclear if lack of risk factor improvement explains the minimal cognitive benefits observed across the majority of outcomes reported. Given that these risk factor profile data are available in other reports from PREDIMED (85-87), future exploration of these relationships may be possible in these or future trials.

3.5.3 Implications

In the absence of treatment or cure for dementia, reducing known risk factors takes on added importance (2). PREDIMED, the largest RCT testing a Mediterranean diet supplemented with EVOO or Nuts, showed significant risk reduction of primary CVD by 30% after a median follow-up of 4.8 years (72), mainly due to decreased stroke risk. As stroke and vascular factors, particularly in mid-life (88), are thought to drive cognitive decline with age, such a finding is also of relevance to cognition and brain morphology/function. However, despite the CVD evidence being used as a basis for recommending this dietary pattern (89-93) it is premature to specifically advise the Mediterranean diet for cognition due to the limited empirical evidence at this time. PREDIMED PLUS (94) and the BRIDGE trial (95), which also include energy restriction and physical activity elements will provide new data on cognitive benefits.

3.5.4 Strengths and limitations

We used a robust and sensitive search strategy with no language or date limitations. Another strength of our review is that our authors included experienced dietitians

86 specialising in the Mediterranean diet, who undertook an analysis of the dietary components and interventions used, which is a novel aspect ofr our pape and addresses a recent criticism on methodologic quality of systematic reviews on the Mediterranean diet (96). Further, we attempted to contact authors to clarify poor reporting and increase the accuracy of our reporting. However, only one author was responsible for initial study selection. Unpublished data were not searched for and we included RCTs with outcomes measured only at follow-up or of short duration (10 days).

3.5.5 Future research

Future RCTs should: a) include larger participant numbers and test various cognitive and brain morphology/function outcomes, be longer than 1.5 years in duration (97) and have a parallel design to avoid potential residual brain structural or connectivity changes in cross-over studies (98), b) target at-risk groups, c) characterise relevant genotypes and health status/lifestyle factors related to cognition, d) include baseline assessment of Mediterranean diet adherence and apply exclusion where this is high; define minimum quantities of recommended foods; provide culinary instructions by dietitians, including cooking; assess potential impact of chrononutrition, e.g., meal size/frequency and timing; use novel technologies to increase compliance; use dietary compliance biomarkers including development of a Mediterranean food metabolome (99), e) utilise a usual diet as comparator (100), f) undertake standardised neuropsychological assessment at baseline and at least one follow-up time point, g) include biomarkers and neuroimaging.

3.5.6 Conclusion

There is currently inconsistent empirical evidence for significant benefits of the Mediterranean diet for cognitive function in a small body of literature. No empirical data have been reported linking this diet to brain morphology or connectivity. However, significant and clinically meaningful effect sizes were found for cognitive composites in the largest and most robust trial, indicating promising scope for future well-designed trials.

3.6 References

1. Prince M, Wimo A, Guerchet M, Ali G-C, Wu Y-T, Prina M. World Alzheimer report 2015, The global impact of dementia: An analysis of prevalence, incidence, cost and trends. Internet: https://www.alz.co.uk/research/world- report-2015 (accessed 23rd November 2018). 2. Alzheimer's Association. 2016 Alzheimer's disease facts and figures. Alzheimers Dement 2016;12:459-509. 3. Hebert LE, Beckett LA, Scherr PA, Evans DA. Annual incidence of Alzheimer disease in the United States projected to the years 2000 through 2050. Alzheimer Dis Assoc Disord 2001;15:169-73. 4. Harwood RH, Sayer AA, Hirschfeld M. Current and future worldwide prevalence of dependency, its relationship to total population, and dependency ratios. Bull World Health Organ 2004;82:251-58. 5. Australian Institute of Health and Welfare. Dementia in Australia. Internet: https://www.aihw.gov.au/reports/dementia/dementia-in-australia/data (accessed 21st December 2018). 6. Buckley JS, Salpeter SR. A risk-benefit assessment of dementia medications: Systematic review of the evidence. Drugs Aging 2015;32:453-67. 7. Gu YA, Nieves JW, Stern Y, Luchsinger JA, Scarmeas N. Food combination and Alzheimer disease risk: A protective diet. Arch Neurol 2010;67:699-706. 8. Barberger-Gateau P, Raffaitin C, Letenneur L, Berr C, Tzourio C, Dartigues J-F, Alpérovitch A. Dietary patterns and risk of dementia: The Three-City cohort study. Neurology 2007;69:1921-30. 9. Blondell SJ, Hammersley-Mather R, Veerman JL. Does physical activity prevent cognitive decline and dementia?: A systematic review and meta-analysis of longitudinal studies. BMC Public Health 2014;14:510. 10. Smyth A, Dehghan M, O'Donnell M, Anderson C, Teo K, Gao P, Sleight P, Dagenais G, Probstfield JL, Mente A, et al. Healthy eating and reduced risk of cognitive decline: A cohort from 40 countries. Neurology 2015;84:2258-65. 11. Shakersain B, Santoni G, Larsson SC, Faxen-Irving G, Fastbom J, Fratiglioni L, Xu WL. Prudent diet may attenuate the adverse effects of Western diet on cognitive decline. Alzheimers Dement 2016;12:100-09.

12. Ashby-Mitchell K, Peeters A, Anstey KJ. Role of dietary pattern analysis in determining cognitive status in elderly Australian adults. Nutrients 2015;7:1052- 67. 13. Jacka FN, Cherbuin N, Anstey KJ, Sachdev P, Butterworth P. Western diet is associated with a smaller hippocampus: A longitudinal investigation. BMC Med 2015;13:215. 14. Kanoski SE, Davidson TL. Western diet consumption and cognitive impairment: Links to hippocampal dysfunction and obesity. Physiol Behav 2011;103:59-68. 15. Luciano M, Corley J, Cox SR, Valdés Hernández MC, Craig LCA, Dickie DA, Karama S, McNeill GM, Bastin ME, Wardlaw JM, et al. Mediterranean-type diet and brain structural change from 73 to 76 years in a Scottish cohort. Neurology 2017;88:449-55. 16. Maruszak A, Pilarski A, Murphy T, Branch N, Thuret S. Hippocampal neurogenesis in Alzheimer's disease: Is there a role for dietary modulation? J Alzheimers Dis 2014;38:11-38. 17. Morris MC, Tangney CC, Wang Y, Sacks FM, Barnes LL, Bennett DA, Aggarwal NT. MIND diet slows cognitive decline with aging. Alzheimers Dement 2015;11:1015-22. 18. Petersson SD, Philippou E. Mediterranean diet, cognitive function, and dementia: A systematic review of the evidence. Adv Nutr 2016;7:889-904. 19. Singh B, Parsaik AK, Mielke MM, Erwin PJ, Knopman DS, Petersen RC, Roberts RO. Association of Mediterranean diet with mild cognitive impairment and Alzheimer's disease: A systematic review and meta-analysis. J Alzheimers Dis 2014;39:271-82. 20. Solfrizzi V, Frisardi V, Seripa D, Logroscino G, Imbimbo B, D'Onofrio G, Addante F, Sancarlo D, Cascavilla L, Pilotto A, et al. Mediterranean diet in predementia and dementia syndromes. Curr Alzheimer Res 2011;8:520-42. 21. Tangney CC, Li H, Wang Y, Barnes L, Schneider JA, Bennett DA, Morris MC. Relation of DASH- and Mediterranean-like dietary patterns to cognitive decline in older persons. Neurology 2014;83:1410-16. 22. Scarmeas N, Stern Y, Mayeux R, Luchsinger JA. Mediterranean diet, Alzheimer disease, and vascular mediation. Arch Neurol 2006;63:1709-17.

23. Scarmeas N, Stern Y, Mayeux R, Manly JJ, Schupf N, Luchsinger JA. Mediterranean diet and mild cognitive impairment. Arch Neurol 2009;66:216- 25. 24. Tangney CC, Kwasny MJ, Li H, Wilson RS, Evans DA, Morris MC. Adherence to a Mediterranean-type dietary pattern and cognitive decline in a community population. Am J Clin Nutr 2011;93:601-07. 25. Psaltopoulou T, Sergentanis TN, Panagiotakos DB, Sergentanis IN, Kosti R, Scarmeas N. Mediterranean diet, stroke, cognitive impairment, and depression: A meta-analysis. Ann Neurol 2013;74:580-91. 26. Fidanza F, Alberti A, Fruttini D. The Nicotera diet: The reference Italian Mediterranean diet. World Rev Nutr Diet 2005;95:115. 27. Kromhout D, Keys A, Aravanis C, Buzina R, Fidanza F, Giampaoli S, Jansen A, Menotti A, Nedeljkovic S, Pekkarinen M. Food consumption patterns in the 1960s in seven countries. Am J Clin Nutr 1989;49:889-94. 28. Radd-Vagenas S, Kouris-Blazos A, Fiatarone Singh M, Flood VM. Evolution of Mediterranean diets and cuisine: Concepts and definitions. Asia Pac J Clin Nutr 2017;26:749-63. 29. Aridi YS, Walker JL, Wright OR. The Association between the Mediterranean dietary pattern and cognitive health: A systematic review. Nutrients 2017;9:674. 30. Cao L, Tan L, Wang H-F, Jiang T, Zhu X-C, Lu H, Tan M-S, Yu J-T. Dietary patterns and risk of dementia: A systematic review and meta-analysis of cohort studies. Mol Neurobiol 2016;53:6144-54. 31. Hardman RJ, Kennedy G, Macpherson H, Scholey AB, Pipingas A. Adherence to a Mediterranean-style diet and effects on cognition in adults: A qualitative evaluation and systematic review of longitudinal and prospective trials. Front Nutr 2016;3:22. 32. Knight A, Bryan J, Murphy K. The Mediterranean diet and age-related cognitive functioning: A systematic review of study findings and neuropsychological assessment methodology. Nutr Neurosci 2017;20:449-68. 33. Lourida I, Soni M, Thompson-Coon J, Purandare N, Lang IA, Ukoumunne OC, Llewellyn DJ. Mediterranean diet, cognitive function, and dementia: A systematic review. Epidemiology 2013;24:479-89.

34. Opie RS, Ralston RA, Walker KZ. Adherence to a Mediterranean-style diet can slow the rate of cognitive decline and decrease the risk of dementia: A systematic review. Nutr Diet 2013;70:206-17. 35. Sofi F, Abbate R, Gensini GF, Casini A. Accruing evidence on benefits of adherence to the Mediterranean diet on health: An updated systematic review and meta-analysis. Am J Clin Nutr 2010;92:1189-96. 36. Sofi F, Cesari F, Abbate R, Gensini GF, Casini A. Adherence to Mediterranean diet and health status: Meta-analysis. BMJ 2008;337:a1344. 37. Sofi F, Macchi C, Abbate R, Gensini GF, Casini A. Mediterranean diet and health status: An updated meta-analysis and a proposal for a literature-based adherence score. Public Health Nutr 2014;17:2769-82. 38. Wu L, Sun DL. Adherence to Mediterranean diet and risk of developing cognitive disorders: An updated systematic review and meta-analysis of prospective cohort studies. Sci Rep 2017;7:41317. 39. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. Ann Intern Med 2009;151:264-69. 40. Strauss E, Sherman EMS, Spreen O. A compendium of neuropsychological tests: Administration, norms, and commentary. 3rd ed. Oxford, New York: Oxford University Press, 2006. 41. McMillan L, Owen L, Kras M, Scholey A. Behavioural effects of a 10-day Mediterranean diet. Results from a pilot study evaluating mood and cognitive performance. Appetite 2011;56:143-47. 42. Lee J, Pase M, Pipingas A, Raubenheimer J, Thurgood M, Villalon L, Macpherson H, Gibbs A, Scholey A. Switching to a 10-day Mediterranean-style diet improves mood and cardiovascular function in a controlled crossover study. Nutrition 2015;31:647-52. 43. Martinez-Lapiscina EH, Clavero P, Toledo E, Estruch R, Salas-Salvado J, Julian BS, Sanchez-Tainta A, Ros E, Valls-Pedret C, Martinez-Gonzalez MA. Mediterranean diet improves cognition: The PREDIMED-NAVARRA randomised trial. J Neurol Neurosurg Psychiatry 2013;84:1318-25.

44. Martinez-Lapiscina EH, Clavero P, Toledo E, San Julian B, Sanchez-Tainta A, Corella D, Lamuela-Raventos R, Martinez J, Martinez-Gonzalez M. Virgin olive oil supplementation and long-term cognition: the PREDIMED-NAVARRA randomized, trial. J Nutr Health Aging 2013;17:544-52. 45. Wardle J, Rogers P, Judd P, Taylor MA, Rapoport L, Green M, Nicholson Perry K. Randomized trial of the effects of cholesterol-lowering dietary treatment on psychological function. Am J Med 2000;108:547-53. 46. Hedges LV. Distribution theory for glass's estimator of effect size and related estimators. J Educ Behav Stat 1981;6:107-28. 47. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates, 1988. 48. Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions Version 5.1.0 [updated March 2011]. Internet: http://handbook-5- 1.cochrane.org/ (accessed 19th June 2017). 49. Senn SJ. Overstating the evidence: Double counting in meta-analysis and related problems. BMC Med Res Methodol 2009;9:10. 50. Li T, Yu T, Hawkins BS, Dickersin K. Design, analysis, and reporting of crossover trials for inclusion in a meta-analysis. PLoS ONE 2015;10:e0133023. 51. Shamseer L, Moher D, Clarke M, Ghersi D, Liberati A, Petticrew M, Shekelle P, Stewart LA. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015: Elaboration and explanation. BMJ 2015;349:g7647. 52. Borenstein M, Hedges LV, Higgins JPT, Rothstein HT. Introduction to meta- analysis. Chichester, UK: John Wiley & Sons, 2009. 53. Moeyaert M, Ugille M, Natasha Beretvas S, Ferron J, Bunuan R, Van den Noortgate W. Methods for dealing with multiple outcomes in meta-analysis: A comparison between averaging effect sizes, robust variance estimation and multilevel meta-analysis. Int J Soc Res Methodol 2017;20:559-72. 54. Maher CG, Sherrington C, Herbert RD, Moseley AM, Elkins M. Reliability of the PEDro scale for rating quality of randomized controlled trials. Phys Ther 2003;83:713-21.

55. Sanchez-Villegas A, Galbete C, Martinez-Gonzalez MA, Martinez JA, Razquin C, Salas-Salvado J, Estruch R, Buil-Cosiales P, Marti A. The effect of the Mediterranean diet on plasma brain-derived neurotrophic factor (BDNF) levels: The PREDIMED-NAVARRA randomized trial. Nutr Neurosci 2011;14:195- 201. 56. Martinez-Lapiscina EH, Galbete C, Corella D, Toledo E, Buil-Cosiales P, Salas- Salvado J, Ros E, Martinez-Gonzalez MA. Genotype patterns at CLU, CR1, PICALM and APOE, cognition and Mediterranean diet: The PREDIMED- NAVARRA trial. Genes Nutr 2014;9:393. 57. Valls-Pedret C, Sala-Vila A, Serra-Mir M, Corella D, de la Torre R, Martínez- González MÁ, Martínez-Lapiscina EH, Fitó M, Pérez-Heras A, Salas-Salvadó J. Mediterranean diet and age-related cognitive decline: A randomized clinical trial. JAMA Intern Med 2015;175:1094-103. 58. Knight A, Bryan J, Wilson C, Hodgson JM, Davis CR, Murphy KJ. The Mediterranean diet and cognitive function among healthy older adults in a 6- month randomised controlled trial: the MedLey study. Nutrients 2016;8:579. 59. Davis CR, Bryan J, Hodgson JM, Wilson C, Dhillon V, Murphy KJ. A randomised controlled intervention trial evaluating the efficacy of an Australianised Mediterranean diet compared to the habitual Australian diet on cognitive function, psychological wellbeing and cardiovascular health in healthy older adults (MedLey study): Protocol paper. BMC Nutrition 2015;1:35. 60. Schulz KF, Altman DG, Moher D, for the Consort Group. CONSORT 2010 statement: Updated guidelines for reporting parallel group randomized trials. Ann Intern Med 2010;152:726-32. 61. Kouris-Blazos A, Itsiopoulos C. Low all-cause mortality despite high cardiovascular risk in elderly Greek-born Australians: Attenuating potential of diet? Asia Pac J Clin Nutr 2014;23:532-44. 62. Martínez-Gonzalez MA, Corella D, Salas-Salvadó J, Ros E, Covas MI, Fiol M, Wärnberg J, Arós F, Ruíz-Gutiérrez V, Lamuela-Raventós RM. Cohort profile: Design and methods of the PREDIMED study. Int J Epidemiol 2012;41:377-85.

63. Harold D, Abraham R, Hollingworth P, Sims R, Gerrish A, Hamshere ML, Pahwa JS, Moskvina V, Dowzell K, Williams A, et al. Genome-wide association study identifies variants at CLU and PICALM associated with Alzheimer's disease. Nat Genet 2009;41:1088-156. 64. Lambert JC, Heath S, Even G, Campion D, Sleegers K, Hiltunen M, Combarros O, Zelenika D, Bullido MJ, Tavernier B, et al. Genome-wide association study identifies variants at CLU and CR1 associated with Alzheimer's disease. Nat Genet 2009;41:1094-99. 65. Gu Y, Brickman AM, Stern Y, Habeck CG, Razlighi QR, Luchsinger JA, Manly JJ, Schupf N, Mayeux R, Scarmeas N. Mediterranean diet and brain structure in a multiethnic elderly cohort. Neurology 2015;85:1774-51. 66. Merrill DA, Siddarth P, Raji CA, Emerson ND, Rueda F, Ercoli LM, Miller KJ, Lavretsky H, Harris LM, Burggren AC, et al. Modifiable risk factors and brain positron emission tomography measures of amyloid and tau in nondemented adults with memory complaints. Am J Geriatr Psychiatry 2016;24:729-37. 67. Staubo SC, Aakre JA, Vemuri P, Syrjanen JA, Mielke MM, Geda YE, Kremers WK, Machulda MM, Knopman DS, Petersen RC, et al. Mediterranean diet, micronutrients and macronutrients, and MRI measures of cortical thickness. Alzheimers Dement 2017;13:168-77. 68. Titova OE, Ax E, Brooks SJ, Sjogren P, Cederholm T, Kilander L, Kullberg J, Larsson EM, Johansson L, Ahlstrom H, et al. Mediterranean diet habits in older individuals: Associations with cognitive functioning and brain volumes. Exp Gerontol 2013;48:1443-8. 69. Ngandu T, Lehtisalo J, Solomon A, Levalahti E, Ahtiluoto S, Antikainen R, Backman L, Hanninen T, Jula A, Laatikainen T, et al. A 2 year multidomain intervention of diet, exercise, cognitive training, and vascular risk monitoring versus control to prevent cognitive decline in at-risk elderly people (FINGER): A randomised controlled trial. Lancet 2015;385:2255-63. 70. Smith PJ, Blumenthal JA, Hoffman BM, Cooper H, Strauman TA, Welsh- Bohmer K, Browndyke JN, Sherwood A. Aerobic exercise and neurocognitive performance: A meta-analytic review of randomized controlled trials. Psychosom Med 2010;72:239-52.

71. Lampit A, Hallock H, Valenzuela M. Computerized cognitive training in cognitively healthy older adults: A systematic review and meta-analysis of effect modifier. PLoS Med 2014;11:e1001756. 72. Estruch R, Ros E, Salas-Salvado J, Covas MI, Corella D, Aros F, Gomez-Gracia E, Ruiz-Gutierrez V, Fiol M, Lapetra J, et al. Primary prevention of cardiovascular disease with a Mediterranean diet. N Engl J Med 2013;368:1279- 90. 73. Diaz-Lopez A, Babio N, Martinez-Gonzalez MA, Corella D, Amor AJ, Fito M, Estruch R, Aros F, Gomez-Gracia E, Fiol M, et al. Mediterranean diet, retinopathy, nephropathy, and microvascular diabetes complications: A post hoc analysis of a randomized trial. Diabetes Care 2015;38:2134-41. 74. Itsiopoulos C, Brazionis L, Kaimakamis M, Cameron M, Best JD, O’Dea K, Rowley K. Can the Mediterranean diet lower HbA1c in type 2 diabetes? Results from a randomized cross-over study. Nutr Metab Cardiovasc Dis 2011;21:740- 47. 75. Toledo E, Salas-Salvadó J, Donat-Vargas C, Buil-Cosiales P, Estruch R, Ros E, Corella D, Fitó M, Hu FB, Arós F, et al. Mediterranean diet and invasive breast cancer risk among women at high cardiovascular risk in the PREDIMED trial: Randomized clinical trial. JAMA Intern Med 2015;175:1752-60. 76. Merle BMJ, Silver RE, Rosner B, Seddon JM. Adherence to a Mediterranean diet, genetic susceptibility, and progression to advanced macular degeneration: A prospective cohort study. Am J Clin Nutr 2015;102:1196-206. 77. Veronese N, Stubbs B, Noale M, Solmi M, Luchini C, Maggi S. Adherence to the Mediterranean diet is associated with better quality of life: Data from the Osteoarthritis Initiative. Am J Clin Nutr 2016;104:1403-09. 78. Kontogianni MD, Tileli N, Margariti A, Georgoulis M, Deutsch M, Tiniakos D, Fragopoulou E, Zafiropoulou R, Manios Y, Papatheodoridis G. Adherence to the Mediterranean diet is associated with the severity of non-alcoholic fatty liver disease. Clin Nutr 2014;33:678-83.

79. Caracciolo B, Xu WL, Collins S, Fratiglioni L. Cognitive decline, dietary factors and gut-brain interactions. Mech Ageing Dev 2014;136:59-69. 80. Rodríguez-Rejón AI, Castro-Quezada I, Ruano-Rodríguez C, Ruiz-López MD, Sánchez-Villegas A, Toledo E, Artacho R, Estruch R, Salas-Salvadó J, Covas MI, et al. Effect of a Mediterranean diet intervention on dietary glycemic load and dietary glycemic index: The PREDIMED study. J Nutr Metab 2014;2014:1- 10. 81. Rodriguez JM, Balich LL, Concha MJ, Mizon C, Barnett DB, Acevedo GB, Birn SH, Jaime TJ, Henriquez S, Uribarri J, et al. Reduction of serum advanced glycation end-products with a low calorie Mediterranean diet. Nutr Hosp 2015;31:2511-17. 82. Barbour JA, Howe PRC, Buckley JD, Bryan J, Coates AM. Nut consumption for vascular health and cognitive function. Nutr Res Rev 2014;27:131-58. 83. Lidder S, Webb AJ. Vascular effects of dietary nitrate (as found in green leafy vegetables and beetroot) via the nitrate-nitrite-nitric oxide pathway. Br J Clin Pharmacol 2013;75:677-96. 84. Iriti M, Varoni EM. Melatonin in Mediterranean diet, a new perspective. J Sci Food Agric 2015;95:2355-59. 85. Martínez-González MA, Salas-Salvadó J, Estruch R, Corella D, Fitó M, Ros E. Benefits of the Mediterranean diet: Insights from the PREDIMED study. Prog Cardiovasc Dis 2015;58:50-60. 86. Medina-Remón A, Casas R, Tressserra-Rimbau A, Ros E, Martínez-González MA, Fitó M, Corella D, Salas-Salvadó J, Lamuela-Raventos RM, Estruch R, et al. Polyphenol intake from a Mediterranean diet decreases inflammatory biomarkers related to atherosclerosis: A substudy of the PREDIMED trial. Br J Clin Pharmacol 2017;83:114-28. 87. Storniolo CE, Casillas R, Bullo M, Castaner O, Ros E, Saez GT, Toledo E, Estruch R, Ruiz-Gutierrez V, Fito M, et al. A Mediterranean diet supplemented with extra virgin olive oil or nuts improves endothelial markers involved in blood pressure control in hypertensive women. Eur J Nutr 2017;56:89-97.

88. Bendlin BB, Carlsson CM, Gleason CE, Johnson SC, Sodhi A, Gallagher CL, Puglielli L, Engelman CD, Ries ML, Xu G, et al. Midlife predictors of Alzheimer's disease. Maturitas 2010;65:131-37. 89. National Health and Medical Research Council. Australian dietary guidelines. Canberra, Australia: National Health and Medical Research Council, 2013. 90. US Department of Health and Human Services and US Department of Agriculture. 2015-2020 Dietary guidelines for Americans, 8th Edition. Washington DC: US Government Printing Office, 2015. 91. Sarris J, Logan AC, Akbaraly TN, Amminger GP, Balanza-Martinez V, Freeman MP, Hibbeln J, Matsuoka Y, Mischoulon D, Mizoue T, et al. Nutritional medicine as mainstream in psychiatry. Lancet Psychiatry 2015;2:271-74. 92. Barnard ND, Bush AI, Ceccarelli A, Cooper J, de Jager CA, Erickson KI, Fraser G, Kesler S, Levin SM, Lucey B, et al. Dietary and lifestyle guidelines for the prevention of Alzheimer's disease. Neurobiol Aging 2014;35:S74-S78. 93. The Royal Australian College of General Practitioners. HANDI: Mediterranean diet: reducing cardiovascular risk. Internet: http://www.racgp.org.au/your- practice/guidelines/handi/interventions/nutrition/the-mediterranean-diet-for- reducing-cardiovascular-disease-risk/ (accessed 19th June 2017). 94. European Commission. Long-term effects of an energy-restricted Mediterranean diet on mortality and cardiovascular disease: The PREDIMED PLUS Study. Internet: https://cordis.europa.eu/project/rcn/188509/factsheet/en (accessed 21st December 2018). 95. Tussing-Humphreys L, Lamar M, Blumenthal JA, Babyak M, Fantuzzi G, Blumstein L, Schiffer L, Fitzgibbon ML. Building research in diet and cognition: The BRIDGE randomized controlled trial. Contemp Clin Trials 2017;59:87-97. 96. Huedo-Medina TB, Garcia M, Bihuniak JD, Kenny A, Kerstetter J. Methodologic quality of meta-analyses and systematic reviews on the Mediterranean diet and cardiovascular disease outcomes: A review. Am J Clin Nutr 2016;103:841-50.

97. Vauzour D, Camprubi-Robles M, Miquel-Kergoat S, Andres-Lacueva C, Banati D, Barberger-Gateau P, Bowman GL, Caberlotto L, Clarke R, Hogervorst E, et al. Nutrition for the ageing brain: Towards evidence for an optimal diet. Ageing Res Rev 2017;35:222-40. 98. Rebok GW, Ball K, Guey LT, Jones RN, Kim HY, King JW, Marsiske M, Morris JN, Tennstedt SL, Unverzagt FW, et al. Ten-year effects of the ACTIVE cognitive training trial on cognition and everyday functioning in older adults. J Am Geriatr Soc 2014;62:16-24. 99. Scalbert A, Brennan L, Manach C, Andres-Lacueva C, Dragsted LO, Draper J, Rappaport SM, van der Hooft JJ, Wishart DS. The food metabolome: A window over dietary exposure. Am J Clin Nutr 2014;99:1286-308. 100. National Heart Lung and Blood Institute. The National Heart Lung and Blood Institute Workshop: "Toward testing the effects of a Mediterranean dietary pattern on cardiovascular and other diseases in the United States". Internet: https://www.nhlbi.nih.gov/events/2016/national-heart-lung-and-blood-institute- workshop-toward-testing-effects-mediterranean (accessed 17th June 2017).

List of tables and figures

Table 3.1. Study characteristics at baseline

Table 3.2. Dietary elements prescribed for the Mediterranean diet

Figure 3.1. PRISMA flow chart of study screening and selection process

Figure 3.2. Percentage of five included studies fulfilling quality criteria

Figure 3.3. Effect of Mediterranean diet on global cognition, attention and working memory

Figure 3.4. Effect of Mediterranean diet on processing speed and verbal and visual memory

Figure 3.5. Effect of Mediterranean diet on language and executive function

Table 3.1. Study characteristics at baseline1

Study Author (yr), Study design n (% F) Mean age Cognitive status Health status Comparative country & duration (y) (SD) condition/s to Mediterranean diet 1 Wardle (2000), Parallel, 12 wk 176 (52) 53.01 (10.1) Not specified at Selected for Waiting list Britain (45) baseline CV risk: total (Exp low fat cholesterol diet also mM compared to this) 2 McMillan (2011), Parallel, 10 d 25 (100) 21.1 (3.3) Not specified at Healthy Usual diet Australia (41) baseline

3 PREDIMED sub-studies (2011-2015) Sanchez-Villegas Post-test 243 (51) 67.85 (6.0) Not specified at Selected for Low fat diet (2011), Spain (55) comparison at baseline CV risk: type 2 3y '0RU major CV risk factors 3 Martinez- Post-test 285 (55) 67.34 (5.7) Not specified at Selected for Low fat diet Lapiscina (2013), comparison at baseline CV risk: type 2 Spain (43) 6.5 y '0RU major CV risk factors

100 Table 3.1. - cont.

Study Author (yr), Study design n (% F) Mean age Cognitive status Health status Comparative country & duration (y) (SD) condition/s to Mediterranean diet 3 Martinez- Post-test 522 (55) 67.38 (5.7) Not specified at Selected for Low fat diet Lapiscina (2013), comparison at baseline CV risk: type 2 Spain (44) 6.5 y '0RU major CV risk factors 3 Martinez- Post-test 522 (56) 67.00 (6.0) Not specified at Selected for Low fat diet Lapiscina (2014), comparison at baseline CV risk: type 2 Spain (56) 6.5 y '0RU major CV risk factors 3 Valls-Pedret Parallel, 4.1 y 447 (51) 66.73 (5.5) Normal (MCI Selected for Low fat diet (2015), Spain (57) excluded at baseline CV risk: type 2 with '0RU neuropsychological major CV risk testing) factors 4 Lee (2015), Cross-over, 24 (100) 25.6 (5.1) Normal (neurological Healthy Usual diet Australia (42) 10 d & psychiatric disorders excluded at baseline)

101 Table 3.1. - cont.

Study Author (yr), Study design n (% F) Mean age Cognitive status Health status Comparative country & duration (y) (SD) condition/s to Mediterranean diet 5 Knight (2016), 2-cohort 166 (gender 72.1 (5.0)2 Normal (MCI & Healthy Usual diet Australia (58) parallel, 24 wk for dementia excluded at completers: baseline with 53) neuropsychological testing) 1 CV, cardiovascular; DM, diabetes mellitus; Exp, experimental; MCI, Mild Cognitive Impairment; SD, standard deviation; wk, week; y, years; yr, year; %, percent.

2 Mean age and SD for completers only.

102

Table 3.2. Dietary elements prescribed for the Mediterranean diet1

Dietary Minimum Criterion2 Wardle McMillan PREDIMED Lee (2015) MedLey Element (2000) (45) (2011) (41) Study* (42) Study (2016) (58, 59) 3 Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Olive oil 2 TBSP (40 mL)/d or to be used as XXX XXX principal fat/oil in diet Other fats4 Max allowed: 10 g/d (about 2 tsp) X X5 X6 XXXX 7 X Nuts 3 serves/wk (1 serve = 30 g) X XXXX XXX Legumes 2 serves/wk (1 serve = 150 g or 1 cup XXX XX cooked or canned legumes) Vegetables 2 serves at 2 meals/d or 4 serves/d X X XXXX XXX total (1 serve = ½ cup cooked (75 g) or 1 cup raw) Sofrito 2 times/wk rich tomato sauce with XXX onion/garlic simmered in EVOO Fruit 2 serves/d (1 serve = 1 medium or 2 X X XXXX XXX small pieces or 1 cup diced fruit); excludes FJ

103 Table 3.2. - cont.

Dietary Minimum Criterion2 Wardle McMillan PREDIMED Lee (2015) MedLey Element (2000) (45) (2011) (41) Study* (42) Study (2016) (58, 59) 3 Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Grain foods 1-2 serves/meal (or 3-6 serves/d) of X XXX any, preferably wholegrain; (1 serve = 1 slice bread or ½ cup cooked pasta, rice, barley, bulgur etc.) Fish or 2 serves/wk unless vegetarian (1 X X XXXX XXX shellfish serve = 1 small fillet or 1 small can or 200 g shellfish) Poultry Max allowed: 2 serves/wk (1 serve = X 8 X 8 X 8 X X 8 X 8 X 8 XX 80 g cooked or 100 g raw) Red meat & <2 serves/wk red meat & <1 X 8 X 8 X 8 XX XXXXXX processed serve/wk processed meat (1 serve = meat 100-150 g) Eggs Max allowed: 4/wk X X 9 XX Dairy Max allowed: 2 serves/d (1 serve = ½ XXXXXX products cup (120 g) ricotta/cottage, 50 g fetta, ¾ cup (200 g) yogurt, 2 slices (40 g) cheese, 1 cup (250 mL) milk) Sweets Max allowed: 2 times/wk XXXXX XXXXXX

104 Table 3.2. - cont.

Dietary Minimum Criterion2 Wardle McMillan PREDIMED Lee (2015) MedLey Element (2000) (45) (2011) (41) Study* (42) Study (2016) (58, 59) 3 Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Mentioned Qty Given Met Min. Sweetened Max allowed: <1 serve/d; (1 serve = XXXXXXXXXXXX drinks 1 cup (250 mL)) Wine Max allowed: 14 std drinks/wk (1 std XX XXXXX drink = 100 mL) Water 5 cups/d Moist )RUPDLQPHDOVZN X11 cooking methods10 Main meals )RUPDLQPHDOVZN eaten in company 1 EVOO, extra virgin olive oil; FJ, fruit juice; Max, maximum; Mentioned, element mentioned in study; Met Min., minimum criterion for traditional diet was met; Min, minimum; Qty. Given, quantity was specified by authors; std, standard; TBSP, tablespoon; tsp, teaspoon; wk, week.

2 Minimum criterion for traditional Mediterranean diet as defined in PROSPERO protocol registration no.: 42015024088.

3 Provided four levels of energy intake ranging from 6400-9600 kJ/d with adjustments made for serves from food groups. Comparison was made with level 2 (8600 kJ/d) recommendations as participants were mostly older men. Advised on discretionary foods as one group.

105

4 Other fats include butter, margarine, cream, refined vegetable oils etc.

5 Paper doesn’t provide quantities but recommended reduce total fat to 30% of energy and swap most saturated fats with monounsaturated fats.

6 Paper doesn’t provide quantities but recommended ‘healthy fat (providing essential fatty acids)’.

7 While paper describes recommendation to abstain from butter and margarine, use of other oils is not specified.

8 Correspondence with authors confirmed both red meat and poultry were excluded from eating plan and only fish was permitted.

9 While allowance is not in accordance with minimum criterion the protocol provides quantitative guidance for eggs, i.e. unlimited.

10 Moist cooking methods include boiling, stewing, steaming etc.

11 Recommended traditional cooking methods are assumed to be moist cooking methods.

* Includes (43, 44, 55-57)

106 Outcome Terms brain-derived neurotrophic factor, CT scan*, executive function, gray matter, grey matter, lumbar puncture, magnetic resonance imaging, mild cognitive impairment, Intervention Terms neurological test*, neuropsychological test*, post- &UHWDQGLHW0HGLWHUUDQHDQGLHW mortem brain examination, problem solving, reaction 0HGLWHUUDQHDQGLHWDU\SDWWHUQ time, spinal puncture, spinal tap, white matter, 0HGLWHUUDQHDQW\SHGLHW Alzheimer, BDNF, brain, cerebral, cerebrovascular 0HG'LHW0H'L0H'LHW cognition, cognitive, cortical, dement*, fMRS, intellect*, intelligen*, IQ, memory, mental, MRI, neural, neuro*, neurologic, neurotransmitter*, thinking, tomography, visuo-spatial.

Search results n = 6385

Duplicates removed n = 1752 After removal of duplicates n = 4633 Excluded on basis of title or abstract n = 4603 Retrieved for evaluation n=30 Added from hand searching n=1

References fully assessed n=31 Excluded on basis of eligibility criteria n=22

Included articles n=9 (representing 5 trials)

FLJXUH 1 35,60$ IORZ FKDUW RI VWXG\ VFUHHQLQJ DQG VHOHFWLRQ SURFHVV LQGLFDWHV DOO SRVVLEOH ZRUG H[WHQVLRQV XVHG ZKHQ VHDUFKLQJ GDWDEDVHV %'1) %UDLQGHULYHG 1HXURWURSKLF )DFWRU &7 FRPSXWHG WRPRJUDSK\ I056 IXQFWLRQDO PDJQHWLF UHVRQDQFH RI WKH EUDLQ ,4 LQWHOOLJHQFH TXRWLHQW05,PDJQHWLFUHVRQDQFHLPDJLQJ

107 Eligibility criteria specified 80% Random allocation 100% Concealed allocation 80% Baseline similarity 80% Subject blinding 0% Therapist blinding 0% Assessors blinding 40% Key outcomes for >85% initial subjects 78% ITT analysis 40% Between group statistics 100% Point & variability measures 100% Diet designed & administered by dietitian 40% Minimum Med diet foods specified 40% Diet meets at least 8 defined minimum criterion 40% Diet tolerance reported 40% Frequency & setting for diet instruction reported 100% Diet compliance assessed 100% Perception of diet burden or benefit reported 20% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

FLJXUH 23HUFHQWDJHRIILYHLQFOXGHGVWXGLHVIXOILOOLQJTXDOLW\FULWHULD1RWH±FULWHULRQIRU'NH\RXWFRPHV! LQLWLDO VXEMHFWV' LV RXW RI QLQH EHFDXVH LW GLIIHUHG EHWZHHQ SXEOLFDWLRQV IRU D VLQJOH VWXG\ ,77, LQWHQWLRQ WR WUHDW DQDO\VLV0HG, 0HGLWHUUDQHDQ

108 A Global cognition (95% CI lower, upper) (95% CI lower, upper) (57) (57)

B Attention (95% CI lower, upper) (95% CI lower, upper) (45) (41) (41) (41) (42) (42) (42)

C Working memory (95% CI lower, upper) (95% CI lower, upper) (41) (41) (41) (57) (57) (57) (57) (42) (42) (42) (58) (58) (58)

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

110 A Processing speed (95% CI lower, upper) (95% CI lower, upper) (41) (41) (57) (57) (42) (42) (58) (58)

B Verbal and visual memory (95% CI lower, upper) (95% CI lower, upper) (41) (41) (41) (41) (57) (57) (57) (57) (57) (57) (42) (42) (42) (42) (58) (58)

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

112 A Language (95% CI lower, upper) (95% CI lower, upper) (57) (57) (58) (58)

B Executive function (95% CI lower, upper) (95% CI lower, upper) (57) (57) (58) (58)

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

113 SUPPLEMENTARY MATERIAL

114 Supplementary Material 3.1. - Sample search strategy for PsycINFO

Database: PsycINFO <1806 to July Week 1 2017>

Search Strategy:

------1 "cretan diet".af. (1) 2 "Mediterranean diet".af. (1442) 3 "Mediterranean dietary pattern* ".af. (215) 4 "Mediterranean type diet".af. (43) 5 MedDiet.af. (8) 6 MeDi.af. (1337) 7 MeDiet.af. (1) 8 1 or 2 or 3 or 4 or 5 or 6 or 7 (2835) 9 "brain-derived neurotrophic factor".af. (21890) 10 "CT scan* ".af. (5030) 11 "executive function".af. (44784) 12 "gray matter".af. (28935) 13 "grey matter".af. (12160) 14 "lumbar puncture".af. (1656) 15 "magnetic resonance imaging".af. (122765) 16 "mild cognitive impairment".af. (32190) 17 "neurological test* ".af. (367) 18 "neuropsychological test* ".af. (73150) 19 "post-mortem brain examination".af. (6) 20 "problem solving".af. (120925) 21 "reaction time".af. (91822) 22 "spinal puncture".af. (373) 23 "spinal tap".af. (70) 24 "white matter".af. (47367) 25 Alzheimer.af. (70303) 26 BDNF.af. (25684)

115 Supplementary Material 3.1. - cont.

27 brain.af. (786482) 28 cerebral.af. (255093) 29 cerebrovascular.af. (36294) 30 cognition.af. (492421) 31 cognitive.af. (1003953) 32 cortical.af. (232573) 33 "dement*".af. (150743) 34 fMRS.af. (39) 35 "intellect*".af. (179777) 36 "intelligen*".af. (283243) 37 IQ.af. (55067) 38 memory.af. (491901) 39 mental.af. (1048152) 40 MRI.af. (106385) 41 neural.af. (392217) 42 "neuro*".af. (1233624) 43 neurologic.af. (36100) 44 "neurotransmitter*".af. (52637) 45 thinking.af. (219110) 46 tomography.af. (62312) 47 visuo-spatial.af. (9637) 48 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46 or 47 (2537049) 49 8 and 48 (2259) (1806-July 2015) 50 limit 49 to yr="2015 -Current" (574)

***************************

116 Supplementary Material 3.2. - Quality rating of Mediterranean diet interventions template

Yes (Y)orno (N)rating.

Note – this tool has not been validated.

Quality Indicator Study 1 Study 2 Study 3 Study 4 Study 5 Study 6 Study 7

1. Diet designed and administered by dietitian 2. Minimum amounts specified to participants for identified Mediterranean foods 3. Prescribed diet meets at least 8 out of 19 defined minimum criterion* 4. Diet tolerance reported 5. Frequency and setting for diet instruction reported 6. Diet compliance assessed 7. Perception of diet burden or benefit reported

117 Supplementary Material 3.2. - cont.

Definitions for quality indicators

Criterion 1 This criterion is satisfied if the report specifically mentions that the Mediterranean diet was both designed by a dietitian and participants were instructed by a dietitian on how to follow this diet. Criterion 2 Specific daily/weekly quantities are stated for most of the Mediterranean foods recommended. This criterion is not met if only general guidance is provided to include more of certain Mediterranean foods or less of non- Mediterranean foods without instruction on minimum amounts required for consumption of the Mediterranean foods. Criterion 3 When at least 40% of the minimum criterion* for 19 Mediterranean foods/cuisine aspects are met by the experimental diet. Criterion 4 The study reports any intolerances to specific Mediterranean diet foods or the prescribed Mediterranean diet as a whole. Criterion 5 The frequency or number of dietary instruction sessions provided during the study and the setting for these sessions (e.g., individual or group) is reported. Criterion 6 Dietary compliance to the Mediterranean diet has been assessed in some way, at any time point, during the trial. This may include food records, FFQ and indices. Criterion 7 The study collected data from participants on their perception of burden or benefit for the Mediterranean diet. Note - the original report may cross reference additional protocol papers where criterion is reported. Criterion may also be satisfied by citing additional publications from the same study and cohort.

*Minimum criterion for 19 Mediterranean foods/cuisine can be found in Table 3.2 of main paper (1).

FFQ, food frequency questionnaire; %, percent.

118 Supplementary Material 3.3. - Study design - randomised controlled trials (RCTs) with pre/post-test comparisons

Author Country Study name Study Study sizea (n) Study Duration of Health Primary Funding source (yr) (city) years design intervention status of study participants outcome Wardle Britain N/R N/R Exp Med: 61 Parallel 12 wks Selected for Effect of Biotechnology (2000) (2) (London & Exp LF: 59 CV risk: Tot- cholesterol and Biological Southeast Con: 56 cholesterol lowering Sciences England) >5.2 mM on mood Research (198 mg/dL) & Council cognition (BBSRC), UK McMillan Australia N/R N/R Exp: 12 Parallel 10 d Healthy Mood & N/R (2011) (3) (Melbourne) Con: 13 cognition Valls- Spain PREDIMED, 2003 to Exp Med Parallel 4.1 yrs Selected for Cognition University of Pedret (Barcelona Barcelona 2009 EVOO: 155 CV risk: type Navarra, (2015) (4) North sub- North sub- Exp Med Nuts: '0RU Instituto de study) study 147 major CV Salud Carlos III, Con: 145 risk factors Centros de Investigacio´n Biome´dica En Red) )LVLRSDWRORJÕD de la Obesidad y Nutricio´n (CIBERobn), donations by food companies (Patrimonio Comunal Olivarero,

119 Supplementary Material 3.3. - cont.

Author Country Study name Study Study sizea (n) Study Duration of Health Primary Funding source (yr) (city) years design intervention status of study participants outcome California Walnut Commission, Borges, La Morella Nuts, Hojiblanca) Lee (2015) Australia N/R N/R 24 Crossover 10 d Healthy Mood, Nil (5) (Melbourne) cognition & CV outcomes Knight Australia MedLey 2013 to Exp: 85 2-cohort 24 wks Healthy Cognition National Health (2016) (6) (Adelaide) 2015 Con: 81 parallel Medical Research Council; University of South Australia Postgraduate Award a study n reported as randomised except for McMillan (2011) (3) who report baseline n for each group after n=2 excluded/dropped out from originally randomised n=27 as not specified whether both of these were from Med group.

Con, control group; CV, cardiovascular; d, days; DM, diabetes mellitus; Exp, experimental group; LF, low fat intervention; Med, Mediterranean intervention; n, number; N/R, outcome measured but not reported; PREDIMED, Prevención con Dieta Mediterránea, translated to PREvention with MEDiterranean Diet; wks, weeks; yr, year; yrs, years.

120 Supplementary Material 3.4. - Study design – PREDIMED RCT with only post-test comparisons

Author Country Study name Sub-study size (n) Follow Health status of Primary Funding source (yr) (city) up time participants study point outcome Sanchez- Spain PREDIMED, Exp Med EVOO: 91 3 yrs Selected for CV Plasma Department of Health of the Villegas (Navarra NAVARRA Exp Med Nuts: 75 risk: type 2 DM BDNF Navarra government, Linea (2011) (7) region sub- sub-study Con: 77 RUPDMRU&9 Especial of the University of study) risk factors Navarra & Instituto de Salud Randomly selected Carlos III 23% of cohort Martinez- Spain PREDIMED, Exp Med EVOO: 95 6.5 yrs Selected for CV Cognition Official agency for funding Lapiscina (Navarra NAVARRA Exp Med Nuts: 95 risk: type 2 DM biomedical (2013) (8) region sub- sub-study Con: 95 RUPDMRU&9 research of the Spanish study) risk factors Government, Thematic Randomly selected Network 27% of cohort, of Nutrition & Cardiovascular which 95% agreed to disease, Thematic participate Network PREDIMED, FEDER (FondoEuropeo de Desarrollo Regional), CIBERobn (Virtual Center for Biomedical Research on Obesity and Nutrition) & Government of Navarra Martinez- Spain PREDIMED, Exp Med EVOO: 224 6.5 yrs Selected for CV Global Official agency for funding Lapiscina (Navarra NAVARRA Exp Med Nuts: 166 risk: type 2 DM cognition biomedical (2013) (9) region sub- sub-study Con: 132 RUPDMRU&9 research of the Spanish study) risk factors Government, Thematic Network

121 Supplementary Material 3.4. - cont.

Author Country Study name Sub-study size (n) Follow Health status of Primary Funding source (yr) (city) up time participants study point outcome Non-randomly Nutrition & Cardiovascular selected 49.5% of disease, Thematic cohort Network PREDIMED, CIBERobn (Virtual Center for Biomedical Research on Obesity and Nutrition) & Government of Navarra Martinez- Spain PREDIMED, Exp Med EVOO: 224 6.5 yrs Selected for CV Effect of Official agency for funding Lapiscina (Navarra NAVARRA Exp Med Nuts: 166 risk: type 2 DM Med diet on biomedical research of the (2014) region sub- sub-study Con: 132 RUPDMRU&9 cognition Spanish Government, (10) study) risk factors across Thematic Network Nutrition Non-randomly different & Cardiovascular selected 49.5% of gene disease, Thematic Network cohort variants PREDIMED, FEDER (Fondo Europeo de Desarrollo Regional), CIBERobn (Virtual Center for Biomedical Research on Obesity and Nutrition) & Government of Navarra BDNF, brain derived neurotrophic factor; Con, control group; CV, cardiovascular; DM, diabetes mellitus; Exp, experimental group; Med, Mediterranean intervention; n, number; PREDIMED, Prevencion con Dieta Mediterranea, translated to PREvention with MEDiterranean Diet; yr, year; yrs, years; %, percent.

122 Supplementary Material 3.5. - Participant characteristics

Author Age Sex Residential Cognitive Relevant Relevant BMI Habitual Educational (yr) (y) mean (% setting status chronic medications or (kg/m2) physical level mean (SD), female) diseasesa (%) supplements (SD) activity (SD), yrs range (%) mean (SD), MET- min/d Wardle Exp Exp Free living Not Mildly or N/R Exp Med: N/A N/A (2000) (2) Med: 54 Med: specified moderately 29 (6) (11) 56 at baseline raised serum Exp LF: 27 Exp LF: Exp cholesterol (5) 52 (11) LF: 42 (100) Con: 27 (4) Con: 53 Con: 57 (8) McMillan 21.1 100 Free living Not N/R N/R N/R N/A N/A (2011) (3) (3.26), specified 19-30 at baseline PREDIMED publications (2011-2015) Sanchez- Exp Med Exp Free living Not Exp Med N/R Exp Med N/R N/R Villegas EVOO: Med specified EVOO: EVOO: (2011) (7) 68.1 EVOO: at baseline Hypertension 29.7 (3.6) (6.1) 53.8 (83.5); Exp Med Exp Med Exp diabetes Nuts: 29.1 Nuts: Med (25.3); (2.7) 67.4 Nuts: hypercholester Con: 28.5 (5.7) 48.0 olaemia (79.1); (3.4) Con: Con: depression 68.0 51.9 (15.4) (6.1)

123 Supplementary Material 3.5. - cont.

Con: Hypertension (84.4); diabetes (31.2); hypercholester olaemia (66.2); depression (11.7) Martinez- Exp Med Exp Free living Not Exp Med N/R Exp Med Exp Med Exp Med Lapiscina EVOO: Med specified EVOO: EVOO: EVOO: 302 EVOO: 8.87 (2013)b (8) 67.18 EVOO: at baseline Hypertension 28.68 (3.6) (219) (2.08) (5.61) 58.2 (74.7); diabetes

124 Supplementary Material 3.5. - cont.

Con: Hypertension (79.8); diabetes (32.6); dyslipidaemia (66.3); ApoE4

125 Supplementary Material 3.5. - cont.

Author Age Sex Residential Cognitive Relevant Relevant BMI Habitual Educational (yr) (y) mean (% setting status chronic medications or (kg/m2) physical level mean (SD), female) diseasesa (%) supplements (SD) activity (SD), yrs range (%) mean (SD), MET- min/d genotype (15.7) Martinez- Exp Med Exp Free living Not Exp Med N/R Exp Med Exp Med Exp Med Lapiscina EVOO: Med specified EVOO: EVOO: EVOO: 282 EVOO: 8.50 (2013)c (9) 67.35 EVOO: at baseline Hypertension 29.30 (3.4) (200) (2.79) (5.65) 54.5 (77.7); Exp Med Exp Med Exp Med Exp Med Exp diabetes Nuts: 28.96 Nuts: 280 Nuts: 8.45 Nuts: Med (37.1); (3.1) (194) (3.00) 67.30 Nuts: dyslipidaemia Con: 29.94 Con: 253 Con: 8.54 (5.77) 57.2 (70.5); family (3.4) (197) (3.38) Con: Con: Hx cognitive 67.55 54.5 decline (18.8); (5.54) ApoE4 genotype (12.5) Exp Med Nuts: Hypertension (83.1); diabetes (34.9); dyslipidaemia (69.3); family Hx cognitive decline (22.8);

126 Supplementary Material 3.5. - cont.

Con: Hypertension (81.8); diabetes (26.5); dyslipidaemia (67.4); family Hx cognitive decline (20.8); ApoE4 genotype (14.4) Martinez- Exp Med Exp Free living Not Exp Med N/R Exp Med Exp Med Exp Med Lapiscina EVOO: Med specified EVOO: EVOO: EVOO: 283 EVOO: 8.50 (2014)d 67 (6) EVOO: at baseline Hypertension 29.3 (3.4) (199) (2.8) (10) Exp Med 54 (77.3); Exp Med Exp Med Exp Med Nuts: 67 Exp diabetes Nuts: 28.9 Nuts: 279 Nuts: 8.40 (6) Med (37.7); (3.2) (196) (2.9) Con: 67 Nuts: dyslipidaemia Con: 29.0 Con: 252 Con: 8.50 (6) 58 (70.5); family (3.4) (198) (3.4)

127 Supplementary Material 3.5. - cont.

128 Supplementary Material 3.5. - cont.

Con: Hypertension (81.4); diabetes (27.1); dyslipidaemia (66.7); family Hx cognitive decline (14.7); ApoE4 genotype (15.6); CLU MAF (0.39); CR1 MAF (0.14); PICALM (0.34)

129 Supplementary Material 3.5. - cont.

Author Age Sex Residential Cognitive Relevant Relevant BMI Habitual Educational (yr) (y) mean (% setting status chronic medications or (kg/m2) physical level mean (SD), female) diseasesa (%) supplements (SD) activity (SD), yrs range (%) mean (SD), MET- min/d Valls- Exp Med Exp Free living Normal Exp Med Exp Med Exp Med Exp Med Exp Med Pedret EVOO: Med (MCI EVOO: EVOO: EVOO: EVOO: 541 EVOO: 6.8 (2015)e (4) 67.9 EVOO: excluded hypertension antihypertensive 28.5 (3.3) (358) (3.0) (5.4) 52.8 at (77.2%); (67.7%); Exp Med Exp Med Exp Med Exp Med Exp baseline) diabetes antidiabetic Nuts: 28.4 Nuts: 554 Nuts: 7.6 Nuts: Med (55.1%); (26.8%); (3.2) (339) (3.3) 66.7 Nuts: dyslipidaemia hypolipidaemic Con: 28.5 Con: 516 Con: 7.1 (5.3) 48.2 (75.6%) (52.8%); (3.3) (349) (2.8) Con: Con: anticholinergic 65.5 51.6 (15.7%) (5.8) Exp Med Nuts: Exp Med Nuts: hypertension antihypertensive (68.8%); (63.4%); diabetes antidiabetic (58.9%); (36.6%); dyslipidaemia hypolipidaemic (74.1%) (47.3%); anticholinergic (11.6%) Con: hypertension Con: (76.8%); antihypertensive diabetes (66.3%);

130 Supplementary Material 3.5. - cont.

Author Age Sex Residential Cognitive Relevant Relevant BMI Habitual Educational (yr) (y) mean (% setting status chronic medications or (kg/m2) physical level mean (SD), female) diseasesa (%) supplements (SD) activity (SD), yrs range (%) mean (SD), MET- min/d (50.5%); antidiabetic dyslipidaemia (37.9%); (71.6%) hypolipidaemic (49.5%); anticholinergic (15.8%) Lee 25.6 100 Free living Normal Excluded at N/A 23.01 (2.33) N/A N/A (2015) (5) (5.14), (neurologi screening 20-38 cal & psychiatric disorders excluded at baseline) Knight Exp: Exp: Free living Normal Excluded at N/R Exp: 26.5 N/A Exp: 19.7% (2016)f (6) 72.1 24.1 (MCI & screening (3.5) only (4.9) Con: dementia Con: 26.9 secondary Con: 29.2 excluded (4.1) Con: 21.2% 72.0 at only (5.0) baseline) secondary a Relevant chronic diseases refer to any diseases (or risk factors) shown in the literature to impair cognition or modify brain function/morphology or interfere with the ability to adopt and adhere to the dietary intervention.

131 Supplementary Material 3.5. - cont. b Baseline data for age, sex, relevant chronic diseases, BMI, habitual physical activity and education relates to participants analysed: Exp Med EVOO=91, Exp Med Nuts=88, Con=89. c Baseline data for age, sex, relevant chronic diseases, BMI, habitual physical activity and education relates to participants analysed: Exp Med EVOO=224, Exp Med Nuts=166, Con=132. d Baseline data for age, sex, relevant chronic diseases, BMI, habitual physical activity and education relates to n=510 in total: Exp Med EVOO=220, Exp Med Nuts=161, Con=129. e Baseline data for age, sex, relevant chronic diseases, BMI, habitual physical activity and education relates to participants analysed: Exp Med EVOO=127, Exp Med Nuts=112, Con=95. f Baseline data for age, sex, relevant chronic diseases, BMI and education relates to participants analysed: Exp Med=70, Con=67.

BMI, body mass index; CLU, CR1, PICALM & ApoE4, genotypes previously identified as related to cognitive decline; Con, control group; EVOO, extra virgin olive oil; Exp, experimental group; LF, low fat intervention; MAF, minimum allele frequency; Med, Mediterranean intervention; MET-min, minutes at a given metabolic equivalent level (units of energy expenditure in physical activity, 1 MET-min roughly equivalent to 1 kCal); MCI, mild cognitive impairment; n, number; neuropsych, neuropsychological; N/A, not available as not collected during study; N/R, outcome measured but not reported; PREDIMED, Prevencion con Dieta Mediterranea, translated to PREvention with MEDiterranean Diet; SD, standard deviation; yr, year; yrs, years; %, percent.

132 Supplementary Material 3.6. - Intervention characteristics: means of implementation of the experimental Mediterranean and control condition

Author (yr) Food provided Diet Diet Individual Group Telephone Logging Assessment developed by administered by counselling counselling calls, texts of dietary of tolerance dietitian dietitian & mode etc. intake of administration Experimental Wardle Y, high PUFA N/R Y, in person (also Y, 8 Y, 8 N Y, 7-d Y (2000) (2) spread & oil psychologist) sessions of sessions of food mixed mixed record at individual individual baseline & & group & group 12 wks McMillan N, $75 N/R N Y, 1 session N N Y, daily N (2011) (3) supermarket food voucher provided record PREDIMED publications (2011-2015) Sanchez- Y, Med EVOO: 1 Y Y, in person Y, 12 Y, 12 NNY Villegas liter EVOO/wk; sessions on sessions on (2011) (7) Med Nuts: 30g/d average average raw unprocessed (every 3 (every 3 nuts (15 g walnuts MThs) MThs) & 15 g almonds) Martinez- Y, Med EVOO: 1 Y Y, in person Y, 26 Y, 26 NNY Lapiscina liter EVOO/wk; sessions on sessions on (2013) (8) Med Nuts: 30g/d average average raw unprocessed (every 3 (every 3 nuts (15 g walnuts, MThs) MThs)

133 Supplementary Material 3.6. - cont.

Author (yr) Food provided Diet Diet Individual Group Telephone Logging Assessment developed by administered by counselling counselling calls, texts of dietary of tolerance dietitian dietitian & mode etc. intake of administration 7.5 g almonds & 7.5 g hazelnuts) Martinez- Y, Med EVOO: 1 Y Y, in person Y, 26 Y, 26 NNY Lapiscina liter EVOO/wk; sessions on sessions on (2013) (9) Med Nuts: 30g/d average average raw unprocessed (every 3 (every 3 nuts (15 g walnuts, MThs) MThs) 7.5 g almonds & 7.5 g hazelnuts) Martinez- Y, Med EVOO: 1 Y Y, in person Y, 26 Y, 26 NNY Lapiscina liter EVOO/wk; sessions on session on (2014) (10) Med Nuts: 30g/d average average raw unprocessed (every 3 (every 3 nuts (15 g walnuts, MThs) MThs) 7.5 g almonds & 7.5 g hazelnuts) Valls-Pedret Y, Med EVOO: 1 Y Y, in person Y, 16 Y, 16 NNY (2015) (4) liter EVOO/wk; sessions on sessions on Med Nuts: 30g/d average average raw unprocessed (every 3 (every 3 nuts (15 g walnuts, MThs) MThs) 7.5 g almonds & 7.5 g hazelnuts)

134 Supplementary Material 3.6. - cont.

Author (yr) Food provided Diet Diet Individual Group Telephone Logging Assessment developed by administered by counselling counselling calls, texts of dietary of tolerance dietitian dietitian & mode etc. intake of administration Lee (2015) N, $150 N/R N Y, 1 session N N Y, daily N (5) supermarket food voucher provided record Knight Y, canned Y Y, in person Y, 14 N N Y, daily N (2016) (6) legumes, yoghurt sessions food (natural/flavored checklist Greek), EVOO, canned tuna, unsalted nuts (walnuts, almonds, peanuts) Control Wardle N N N N N N Y, 7-d N (2000) (2) food record baseline only McMillan N, $75 N N N N N Y, daily N (2011) (3) supermarket food voucher provided record PREDIMED publications (2011-2015) Sanchez- N N Y, in person Y, 1 session N NNY Villegas (2011) (7)

135 Supplementary Material 3.6. - cont.

Author (yr) Food provided Diet Diet Individual Group Telephone Logging Assessment developed by administered by counselling counselling calls, texts of dietary of tolerance dietitian dietitian & mode etc. intake of administration Martinez- N N Y, in person Y, 1 session N NNY Lapiscina (2013) (8) Martinez- N N Y, in person Y, 1 session N NNY Lapiscina (2013) (9) Martinez- N N Y, in person Y, 1 session N NNY Lapiscina (2014) (10) Valls-Pedret N N Y, in person Y, 3 Y, 4 NNY (2015) (4) sessions sessions over first 3 during last yrs (1/yr) + yr only 4 sessions for last yr = 7 sessions total Lee (2015) N, $150 N N N N N Y, daily N (5) supermarket food voucher provided record Knight N, $10 N Y, in person Y, 14 NNNN (2016) (6) supermarket sessions voucher provided per fortnight

136 Supplementary Material 3.6. cont. d, days; EVOO, extra virgin olive oil; Med, Mediterranean intervention; MThs, months; MUFA, monounsaturated fatty acids; N, no; N/R, outcome measured but not reported; PREDIMED, Prevencion con Dieta Mediterranea, translated to PREvention with MEDiterranean Diet; PUFA, polyunsaturated fatty acids; wks, weeks; Y, yes; yr, year; years, years.

137 Supplementary Material 3.7. - Dietary compliance outcomes

Author (yr) Attendance Dietary log Consumed Diet Dietary Perceptions Dropout Reason for at completion free foods compliance biomarkers of burden rate dropout intervention (% cohort) provided assessment or benefit (%)a sessions (% (% cohort) tool (score) cohort) Experimental Wardle 86.9% N/R N/A 7-d food N/A N/A 13.1 Moving, taking (2000) (2) completed 12 record; score new job, no time wk follow up N/R to attend sessions & attended Tx sessions McMillan N/R 11/12 N/A 10-d food N/A N/A 3.7b 1 participant (2011) (3) (91.7%) record; 93% withdrew & 1 was of meals & excluded in 95% of analysis due to snacks failure to comply with eating plan PREDIMED N/R N/A N/R 137-item Y, Med EVOO: Y, yrly Med Med EVOO: 2 publicationsc FFQ; score increase in EVOO: deceased, 4 (2011-2015): N/R urinary 16.1 illness, 15 Sanchez- 14-item hydroxytyrosold Med declined to Villegas screener (mean increase Nuts: participate, 4 lost (2011) (7), score = 49.6 μg/L); 23.1 contact Martinez- Pre Med Y, Med Nuts: Med Nuts: 1 Lapiscina EVOO: sig. increase in deceased, 3 (2013) (8), 8.6/14 plasma ALAe illness, 29 Martinez- (SD:1.91) declined to

138 Supplementary Material 3.7. - cont.

139 Supplementary Material 3.7. - cont.

Author (yr) Attendance Dietary log Consumed Diet Dietary Perceptions Dropout Reason for at completion free foods compliance biomarkers of burden rate dropout intervention (% cohort) provided assessment or benefit (%)a sessions (% (% cohort) tool (score) cohort) food record carotenoids, commitments, 2 & FFQ at urinary did not comply baseline, 2 & metabolites with diet, 4 had 4 MThs; 92% (potassium, family/personal compliance sodium, issues, 3 had reported but magnesium, health reasons score N/R calcium) (unrelated to trial) Control Wardle 89.3% N/A N/A N/A N/A N/A 10.7 Moving, taking (2000) (2) completed 12 new job, no time wk follow up to attend sessions McMillan N/R 11/13 N/A 10-d food N/A N/A 3.7b 1 participant (2011) (3) (84.6%) record; 71% withdrew & 1 was of meals & excluded in 68% of analysis due to snacks not failure to comply adhered to with eating plan Med PREDIMED N/R N/A N/A 137-item Y, decrease in Y, yrly 33.1 1 deceased, 6 publicationsc FFQ; score urinary illness, 41 (2011-2015): N/R hydroxytyrosold declined to Sanchez- 14-item (mean decrease participate Villegas screener = 8.9 μg/L); (2011) (7), score Pre: Y, sig. increase Martinez- in plasma ALAe

140 Supplementary Material 3.7. - cont.

Author (yr) Attendance Dietary log Consumed Diet Dietary Perceptions Dropout Reason for at completion free foods compliance biomarkers of burden rate dropout intervention (% cohort) provided assessment or benefit (%)a sessions (% (% cohort) tool (score) cohort) Lapiscina 8.7/14 (mean increase (2013) (8), (SD:2.15) = 0.03%) Martinez- Post: 9.1/14 Lapiscina Change: (2013) (9), 0.40/14 Martinez- (SD:2.27) Lapiscina (2014) (10), Valls-Pedret (2015) (4) Lee (2015) 100% 100% N/A 10-d food N/A N/A 0 No drop outs (5) record; 76% compliance to control Knight N/R N/R N/A 3-d weighed Erythrocyte N/A 17.3 9 withdrew before (2016) (6) food record fatty acids, commencement, 2 & FFQ at plasma had other baseline, 2 & carotenoids, commitments, 2 4MThs; urinary had score N/R metabolites family/personal (potassium, issues, 1 had sodium, health reasons magnesium, (unrelated to trial) calcium)

141 Supplementary Material 3.7. - cont. a Dropout rate includes those randomised who withdrew, were deceased, didn’t complete the study and therefore lost to analysis. It does not include those excluded for analysis due to low adherence. b Percent refers to overall dropout rate as paper does not specify which group drop outs relate to. c Diet compliance assessment tool, dietary biomarkers, dropout rate and reason for drop out relate to Valls-Pedret (2015) (4). d Hydroxytyrosol measured in n=65 (Med EVOO=20; Med Nuts=21; Con=24). e ALA measured in n=75 (Med EVOO=24; Med Nuts=28; Con=23).

ALA, alpha linolenic acid; d, day; EVOO, extra virgin olive oil; FFQ, food frequency questionnaire; Med, Mediterranean intervention; MThs, months; N/A, not available as not collected during study; N/R, outcome measured but not reported; PREDIMED, Prevencion con Dieta Mediterranea, translated to PREvention with MEDiterranean Diet; SD, standard deviation; sig., significant; Tx, treatment; wk, week; Y, yes; yr, year; yrly, yearly; %, percent.

142 Supplementary Material 3.8. - Intervention characteristics: dietary elements/nutritional behaviours prescribed for experimental Mediterranean condition

Dietary Minimum Wardle (2000) McMillan PREDIMED Lee (2015) (5) Knight (2016)b Element criteriona (2) (2011) (3) publications (2011- (6) 2015): Sanchez- Villegas (2011) (7), Martinez-Lapiscina (2013) (8), Martinez- Lapiscina (2013) (9), Martinez-Lapiscina (2014) (10), Valls- Pedret (2015) (4) Olive oil 2 TBSP (40 1ĻWRWDOIDWWR N; combine <7%63GZLWK N; focus on foods Y; 2-3 TBSP/d ml)/d or to be 30% E healthy fat with ‘abundant use for providing source used as principal protein & CHO cooking & dressing of of healthy fats fat/oil in diet dishes’ Other fats Max. allowed: 1ĻWRWDOIDWWR N Y; ‘eliminate or limit’ ~Y; no butter or N (e.g., butter, 10 g/d (about 2 30% E & swap with <1 serve/d spread margarine margarine, tsp) most SFA with fats cream, refined MUFA vegetable oils) Nuts 3 serves/wk N 1ĹQXWVVHHGV <VHUYHVZNQXWVRU 1ĹQXWV <VHUYHVZN 1 serve = 30 g seeds (30 g/serve) (35 g/serve) Legumes 2 serves/wk NN<VHUYHVZN N 1VHUYHVZN 1 serve = 150 g (75 g/serve) (1 cup cooked or canned legumes)

143 Supplementary Material 3.8. - cont.

144 Supplementary Material 3.8. - cont.

145 Supplementary Material 3.8. - cont.

Dietary Minimum Wardle (2000) McMillan PREDIMED Lee (2015) (5) Knight (2016)b Element criteriona (2) (2011) (3) publications (2011- (6) 2015): Sanchez- Villegas (2011) (7), Martinez-Lapiscina (2013) (8), Martinez- Lapiscina (2013) (9), Martinez-Lapiscina (2014) (10), Valls- Pedret (2015) (4) Poultry Max allowed: 2 N Y N; preference for white Y; abstain from N serves/wk meat, e.g., poultry white meat 1 serve = 100- without skin or rabbit 150 g Red meat & <2 serves/wk N Y N; <1 serve/d with Y; abstain from <Verve/wk processed meat red meat & <1 preference for white all red meats red meat (100 serve/wk meat, e.g., poultry g/serve); 2 processed meat without skin or rabbit serves/wk 1 serve = 100- processed meat 150 g (50 g/serve) Eggs Max allowed: 4 N N N; permitted ad lib N N /wk Dairy products Max allowed: 2 N 1DGYLVHGWRĹ N; LF cheese permitted 1DGYLVHGWRĹ N serves/d LF dairy ad lib LF dairy 1 serve = ½ cup (120 g) ricotta/cottage cheese, 50 g fetta, ¾ cup (200 g) yoghurt,

146 Supplementary Material 3.8. - cont.

147 Supplementary Material 3.8. - cont.

Dietary Minimum Wardle (2000) McMillan PREDIMED Lee (2015) (5) Knight (2016)b Element criteriona (2) (2011) (3) publications (2011- (6) 2015): Sanchez- Villegas (2011) (7), Martinez-Lapiscina (2013) (8), Martinez- Lapiscina (2013) (9), Martinez-Lapiscina (2014) (10), Valls- Pedret (2015) (4) Moist cooking For PDLQ N N N; recommended NN methods used, meals/wk ‘traditional’ cooking e.g., boiling, methods stewing, steaming Main meals )RUPDLQ NNN N N eaten in meals/wk company a Our set minimum criterion to describe compliance to the traditional Mediterranean diet. b MedLey study provided recommendations for serves for four levels of energy intake ranging from 6400-9600 kJ/d. Comparison was made with level 2 (8600 kJ/d) recommendations as participants were mostly older men. MedLey study also advised on discretionary foods as one group. ad lib, ad libitum; CHO, carbohydrate; d, day; E, energy; EVOO, extra virgin olive oil; FJ, fruit juice; kJ, kilojoules; LF, low fat; Max., maximum; MUFA, monounsaturated fatty acids; N, no; N/R, outcome measured but not reported; PREDIMED, Prevencion con Dieta Mediterranea, translated to PREvention with MEDiterranean Diet; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids; TBSP, tablespoon; tsp, teaspoon; veg, vegetables; wk, week; Y, yes; ~Y, partially meets specified criteria; %, percent.

148 Supplementary Material 3.9. - Intervention characteristics: dietary elements/nutritional behaviours prescribed for control/comparative condition

Author (yr) Type of diet Macronutrient Micronutrient Other recommendations recommendations recommendations Wardle (2000) (2) Waiting list (Tx offered end of Nil Nil Not discouraged from making period) dietary changes and some elected to do so McMillan (2011) (3) Usual diet Nil Nil Nil

PREDIMED publications LF based on guidelines from AHA Nil Nil Nil; no energy restriction (2011-2015): Sanchez- National Cholesterol Education Villegas (2011) (7), Program Adult Treatment Panel III Martinez-Lapiscina (2013) (8), Martinez- Lapiscina (2013) (9), Martinez-Lapiscina (2014) (10), Valls-Pedret (2015) (4) Lee (2015) (5) Usual diet Nil Nil Nil

Knight (2016) (6) Usual diet Nil Nil Maintain current diet, physical activity and dietary supplements unless instructed otherwise by medical professional AHA, American Heart Association; LF, low fat intervention; PREDIMED, Prevencion con Dieta Mediterranea, translated to PREvention with MEDiterranean Diet; Tx, treatment; yr, year.

149 Supplementary Material 3.10. - Quality review of five included studies according to modified PEDro scale

Quality Indicator Wardle McMillan PREDIMED Study Lee MedLey (2000) (2) (2011) (3) (2015) (5) Study Sanchez- Martinez- Martinez- Martinez- Valls- Knight Villegas Lapiscina Lapiscina Lapiscina Pedret (2016) (6) (2011) (7) (2013) (8) (2013) (9) (2014) (10) (2015) (4)

1. Eligibility criteria 101111111 specified 2. Random allocation 1 1 1 1 1 1 1 1 1 3. Concealed allocation 1 1 1 1 1 1 1 0 1 4. Baseline similarity with 111111101 important prognostic indicators 5. Blinding of all subjects 0 0 0 0 0 0 0 0 0 6. Blinding of all therapists 0 0 0 0 0 0 0 0 0 7. Blinding of all assessors 0 0 1 1 1 1 1 0 1 8. Measures of key 111111010 outcomes (>85% of initial subjects) 9. Intention to treat analysis 0 0 1 1 1 1 1 0 1 10. Between group statistics 1 1 1 1 1 1 1 1 1 11. Point measures & 111111111 measures of variability Final score ( /10) 6 6 8 8 8 8 7 4 7 Quality indicator met: 0=no; 1=yes. N.B. number 1 indicator is not included in PEDro scoring. It measures external validity or generalisability of the trial and is retained so the Delphi list is complete.

150 Supplementary Material 3.11. - Quality review of five included studies for Mediterranean diet intervention

Quality Indicator Wardle (2000)(2) McMillan (2011) PREDIMED Lee MedLey Study (3) Study Sanchez- (2015) (5) Knight (2016) (6), Villegas (2011) Davis (2015) (11), (7), Martinez- Davis (2017) (12) Lapiscina (2013) (8), Martinez- Lapiscina (2013) (9), Martinez- Lapiscina (2014) (10), Valls-Pedret (2015) (4) 1. Diet designed and NNYNY administered by dietitian 2. Minimum amounts NNYNY specified to participants for identified Mediterranean foods 3. Prescribed diet meets at NNY NY least 8 out of 19 defined minimum criterion* 4. Diet tolerance reported YNYNN 5. Frequency and setting YYYYY for diet instruction reported 6. Diet compliance YYYYY assessed

151 Supplementary Material 3.11. - cont.

152 Supplementary Material 3.11. - cont.

Definitions for quality indicators

Criterion 1 This criterion is satisfied if the report specifically mentions that the Mediterranean diet was both designed by a dietitian and participants were instructed by a dietitian on how to follow this diet. Criterion 2 Specific daily/weekly quantities are stated for most of the Mediterranean foods recommended. This criterion is not met if only general guidance is provided to include more of certain Mediterranean foods or less of non- Mediterranean foods without instruction on minimum amounts required for consumption of the Mediterranean foods. Criterion 3 When at least 40% of the minimum criterion* for 19 Mediterranean foods/cuisine aspects are met by the experimental diet. Criterion 4 The study reports any intolerances to specific Mediterranean diet foods or the prescribed Mediterranean diet as a whole. Criterion 5 The frequency or number of dietary instruction sessions provided during the study and the setting for these sessions (e.g., individual or group) is reported. Criterion 6 Dietary compliance to the Mediterranean diet has been assessed in some way, at any time point, during the trial. This may include food records, FFQ and indices. Criterion 7 The study collected data from participants on their perception of burden or benefit for the Mediterranean diet. *Minimum criterion for 19 Mediterranean foods/cuisine can be found in Table 3.2 of main paper (1).

FFQ, food frequency questionnaire; %, percent.

153 Supplementary Material 3.12. - Cognitive function outcomes

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Global cognition Wardle N/A N/A N/A N/A N/A N/A N/A N/A (2000) (2) McMillan N/A N/A N/A N/A N/A N/A N/A N/A (2011) (3) PREDIMED publications (2011-2015) Sanchez- N/A N/A N/A N/A N/A N/A N/A N/A Villegas (2011) (7) Martinez- MMSEc 6.5 yrs Pre Med Pre: N/A Med EVOO: Med EVOO: 0.33 0.130 N Lapiscina EVOO: N/A 0.66 (-0.06, (-0.03, 0.69) (2013) (8) Pre Med 1.38) Nuts: N/A Med Nuts: 1.35 Med Nuts: 0.63 Y Post Med Post: 27.48 (0.57, 2.13) (0.26, 0.99) EVOO: 28.14 (1.97) (2.02) Post Med Nuts: 28.83 (2.22)

Change Med Change: N/A EVOO: N/A

154 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Change Med Nuts: N/A

Incident MCI Exp Med 17 (19.30%) N/A Exp Med EVOO: 0.024 N cases (%)d,e EVOO: 7 -0.57 (-5.44, 4.30) (7.80%) Exp Med Nuts: - Exp Med 0.32 (-4.65, 4.01) 0.171 N Nuts: 10 (11.80%) Martinez- MMSEf 6.5 yrs Pre Med Pre: N/A Med EVOO: Med EVOO: 0.27 0.030 N Lapiscina EVOO: N/A 0.60 (-0.01, (-0.01, 0.54) (2013) (9) Pre Med 1.21) Nuts: N/A Med Nuts: 0.56 Med Nuts: 0.25 (- N Post Med Post: 27.40 (-0.07, 1.19) 0.03, 0.54) EVOO: 28.00 (2.38) (2.18) Post Med Nuts: 27.96 (2.12)

Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A

155 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Incident MCI Exp Med 23 (6.53%) N/A Med EVOO: - N/R N cases (%)e,g EVOO: 18 0.14 (-6.71, 6.43) (5.13%) Exp Med Med Nuts: -0.11 Nuts: 19 (-6.59, 6.37) N/R N (5.40%)

Incident Exp Med 17 (4.83%) N/A Med EVOO: - N/R N dementia EVOO: 12 0.20 (-7.99, 7.60) cases (%)e,g (3.42%) Exp Med Med Nuts: -0.59 Nuts: 6 (-10.35, 9.17) N/R N (1.70%) Martinez- MMSEh,i,j 6.5 yrs CLU CLU 0.97 (0.45, 1.49) 0.45 (0.12, 0.79) <0.001 Y Lapiscina CT/TT: 28.20 CT/TT: 27.23 (2014) (10) (2.15) (2.02)

CLU CLU 0.18 (-0.52, 0.09 (-0.35, 0.53) 0.612 N CC: 27.74 CC: 27.56 0.88) (2.00) (2.01)

CR1 CR1 0.76 (0.32, 1.20) 0.40 (0.09, 0.70) 0.001 Y GG: 28.12 GG: 27.37 (1.87) (1.89)

156 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value CR1 CR1 0.15 (-0.83, 0.06 (-0.46, 0.59) 0.762 N GA/AA: GA/AA: 1.13) 27.66 (2.30) 27.51 (2.30)

PICALM PICALM 0.51 (0.01, 1.00) 0.29 (-0.08, 0.65) 0.046 N CT/TT: 28.11 CT/TT: 27.61 (1.73) (1.75)

PICALM PICALM 0.76 (0.10, 1.43) 0.32 (-0.06, 0.71) 0.024 N CC: 27.89 CC: 27.18 (2.20) (2.21)

ApoE4 non- ApoE4 non- 0.56 (0.15, 0.97) N/A 0.007 N/A carriers: N/R carriers: N/R

ApoE4 ApoE4 1.61 (0.10, 3.13) N/A 0.037 N/A carriers: N/R carriers: N/R Valls-Pedret MMSEk,l Baseline, Pre Med Pre: 28.38 Med EVOO: Med EVOO: 0.32 Overall: N (2015) (4) as close to EVOO: 28.01 (1.30) 0.42 (-0.01, (-0.01, 0.66) 0.15u study (1.28) 0.85) termination date as Pre Med Med Nuts: 0.19 Med Nuts: 0.15 (- N possible Nuts: 28.11 (-0.25, 0.63) 0.19, 0.49) (average (1.28) 4.1 yrs) Post Med Post: 28.12 EVOO: 28.17

157 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Post Med Nuts: 28.04

Change Med Change: - EVOO: 0.16 0.26 (1.55) (1.59)

Change Med Nuts: -0.07 (1.58)

Global Pre Med Pre: -0.07 Med EVOO: Med EVOO: 1.29 <0.01 Y cognition EVOO: -0.07 (0.32) 0.43 (0.23, 0.63) (0.65, 1.92) compositel,m (0.33) Pre Med Med Nuts: 0.33 Med Nuts: 1.12 Nuts: -0.05 (0.14, 0.52) (0.44, 1.81) <0.25 Yx (0.27) Overall: Post Med Post: -0.45 0.005u EVOO: -0.02 Post Med Nuts: -0.10

Change Med Change: - EVOO: 0.05 0.38 (0.52) (0.51)

158 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Change Med Nuts: -0.05 (0.55)

Incident MCI Exp Med 12 (12.6%) N/A Med EVOO: 0.28 N cases (%)e,n EVOO: 17 0.04 (-4.51, 4.58) (13.4%) Exp Nuts: 8 Med Nuts: N (7.1%) -0.35 (-5.67, 4.97) Lee (2015) N/A N/A N/A N/A N/A N/A N/A N/A (5) Knight Total age- Baseline, 3 Pre: N/R Pre: N/R 0.10v (2016)t (6) related MThs, 6 cognitive MThs 3MThs: 3MThs: -1.58 (-3.41, -0.29 (-0.63, 0.05) N function Post: -0.76 Post: 0.82 0.25) score (5.37) (5.43) composite 6MThs: 6MThs: -0.77 (-2.52, -0.15 (-0.48, 0.19) N Post: -0.37 Post: 0.40 0.98) (5.14) (5.23) Attention Wardle Sustained Baseline, 6 Pre: 34.86 Pre: Overall: (2000) (2) attention task wks, 12 (0.62) 34.31 (0.54) <0.001w (SAT) wks adapted from 6 wks: 6 wks: 6 wks: 6 wks: Y Baker Post: Post: -2.08 (-2.31, - -3.54 (-4.16, - 32.81 (0.50) 34.34 (0.55) 1.85) 2.92)

159 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value vigilance Change: - Change: 0.03 paradigma,b 2.05

12 wks: 12 wks: 12 wks: 12 wks: Y Post: Post: -9.32 (-9.55, - -15.86 (-18.06, - 33.80 (0.48) 42.57 (0.47) 9.09) 13.66) Change: - Change: 8.26 1.06 McMillan Serial 3 RT Baseline, Pre: 4094 Pre: 3983 222.00 (-734.25, 0.19 (-0.60, 0.97) 0.34 N (2011) (3) (ms) 10 d (1010) (1273) 1178.25) Post: 3762 Post: 3429 (1467) (1103) Change: - Change: - 332.00 554.00

Serial 3 Pre: 88.01 Pre: 91.95 5.72 (-4.12, 0.47 (-0.33, 1.26) 0.27 N accuracy (%) (12.12) (11.67) 15.56) Post: 91.46 Post: 89.68 (9.18) (13.23) Change: 3.45 Change: - 2.27

Serial 7 RT Pre: 7813 Pre: 5981 -1668.00 (- -0.42 (-1.22, 0.37) 0.09 N (ms) (5043) (2105) 4818.68, Post: 6700 Post: 6536 1482.68) (2854) (2689)

160 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Change: - Change: 1113.00 555.00

Serial 7 Pre: 80.15 Pre: 89.96 9.62 (-4.21, 0.56 (-0.24, 1.36) 0.29 N accuracy (%) (20.04) (12.90) 23.45) Post: 83.58 Post: 83.77 (10.16) (23.86) Change: 3.43 Change: - 6.19

RVIP RT Pre: 467 (56) Pre: 471 (57) -21 (-67.81, -0.36 (-1.15, 0.43) 0.35 N (ms) Post: 453 Post: 478 25.81) (37) (61) Change: -14 Change: 7.0

RVIP Pre: 39.96 Pre: 46.32 4.26 (-16.27, 0.17 (-0.62, 0.95) 0.58 N accuracy (24.05) (25.44) 24.79) (100%) Post: 54.55 Post: 56.65 (20.69) (20.24) Change: Change: 14.59 10.33 PREDIMED publications (2011-2015) Sanchez- N/A N/A N/A N/A N/A N/A N/A N/A Villegas (2011) (7)

161 Supplementary Material 3.12. - cont.

Lee (2015) Serial 3 – Baseline, Pre: N/R Pre: N/R -3.01 -0.69 (-1.85, 0.48) 0.25 N (5) correct 11 d Post: N/R Post: N/R responseso,p Change: 0.86 Change: 3.87 (7.69) (5.84)

Serial 3 RT Pre: N/R Pre: N/R 153.41 0.95 (-0.25, 2.14) 0.61 N (ms)o,p Post: N/R Post: N/R Change: - Change: - 348.35 501.76 (948.92) (595.79)

Serial 3 – Pre: N/R Pre: N/R -2.13 -0.59 (-1.75, 0.56) 0.31 N total Post: N/R Post: N/R responseso,p Change: 2.17 Change: 4.3 (5.48) (5.34)

162 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Serial 7 – Pre: N/R Pre: N/R 0.26 0.65 (-0.51, 1.81) 0.72 N correct Post: N/R Post: N/R responseso,p Change: 1.48 Change: 1.22 (4.96) (4.11)

Serial 7 RT Pre: N/R Pre: N/R -900.13 -0.74 (-1.91, 0.43) 0.21 N (ms)o,p Post: N/R Post: N/R Change: - Change: 707.96 192.17 (2374.76) (1515.71)

Serial 7 – Pre: N/R Pre: N/R 1.44 0.57 (-0.58, 1.73) 0.82 N total Post: N/R Post: N/R responseso,p Change: 0.87 Change: - (4.61) 0.57 (3.49)

RVIP RT Pre: N/R Pre: N/R 6.54 0.57 (-0.58, 1.73) 0.48 N (ms)o,p Post: N/R Post: N/R Change: - Change: - 5.38 (72.50) 11.92 (89.59)

RVIP Pre: N/R Pre: N/R -4.34 -0.57 (-1.73, 0.58) 0.39 N accuracy Post: N/R Post: N/R (100%)o,p Change: 4.08 Change: 8.42 (14.23) (16.52)

163 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value RVIP – false Pre: N/R Pre: N/R -0.22 -0.57 (-1.73, 0.58) 0.97 N alarm Post: N/R Post: N/R responseso,p Change: - Change: - 0.74 (6.70) 0.52 (5.81) Knight N/A N/A N/A N/A N/A N/A N/A N/A (2016) (6) Working memory Wardle N/A N/A N/A N/A N/A N/A N/A N/A (2000) (2) McMillan Corsi blocks Baseline, Pre: 3109 Pre: 3088 -1257 (-1762.77, -1.99 (-2.95, - <0.001 Y (2011) (3) RT (ms) 10 d (683) (536) -751.23) 1.03) Post: 2689 Post: 3925 (557) (1129) Change: -420 Change: 837

Corsi blocks Pre: 5.57 Pre: 5.77 (87) 0.45 (-51.6, 0.01 (-0.78, 0.79) 0.28 N score (1.24) Post: 6.00 52.5) Post: 6.25 (92) (98) Change: 0.23 Change: 0.68

N-Back RT Pre: 990 Pre: 1045 339 (-44.55, 0.71 (-0.10, 1.52) 0.11 N (ms) (346) (549) 722.55) Post: 1025 Post: 741 (746) (439) Change: 35 Change: -304

164 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value N-Back Pre: 86.85 Pre: 85.64 2.74 (-3.45, 0.35 (-0.44, 1.15) 0.38 N accuracy (%) (7.55) (7.40) 8.93) Post: 86.85 Post: 82.90 (9.96) (11.63) Change: 0 Change: - 2.74

Numeric WM Pre: 782 Pre: 860 (16) 137 (71.01, 1.66 (0.75, 2.57) 0.04 Y RT (ms) (114) Post: 738 202.99) Post: 797 (116) (191) Change: -122 Change: 15

Numeric WM Pre: 88.98 Pre: 96.33 10.54 (-4.50, 0.56 (-0.24, 1.36) 0.21 N RT accuracy (25.99) (3.64) 25.58) (%) Post: 95.58 Post: 92.39 (4.59) (13.39) Change: 6.60 Change: - 3.94 PREDIMED publications (2011-2015) Sanchez- N/A N/A N/A N/A N/A N/A N/A N/A Villegas (2011) (7) Martinez- Digit, 6.5 yrs Pre Med Pre: N/A Med EVOO: Med EVOO: 0.47 0.01 Y Lapiscina forwardc EVOO: N/A 0.90 (0.21, 1.59) (0.11, 0.83) (2013) (8) Pre Med Nuts: N/A

165 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Post Med Post: 6.90 Med Nuts: 0.50 Med Nuts: 0.28 (- N EVOO: 7.80 (1.64) (-0.14, 1.14) 0.08, 0.64) (2.02) Post Med Nuts: 7.40 (1.84)

Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A

Digit, Pre Med Pre: N/A Med EVOO: Med EVOO: 0.13 0.49 N backwardc EVOO: N/A 0.24 (-0.41, (-0.23, 0.49) Pre Med 0.89) Nuts: N/A Med Nuts: 0.31 Med Nuts: 0.17 (- N Post Med Post: 4.20 (-0.35, 0.97) 0.19, 0.53) EVOO: 4.44 (1.38) (1.99) Post Med Nuts: 4.51 (2.01)

Change Med Change: N/A EVOO: N/A

166 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Change Med Nuts: N/A Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2013) (9) Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2014) (10) Valls-Pedret Digit span Baseline, Pre Med Pre: 5.23 Med EVOO: Med EVOO: 0.20 Overall: N (2015) (4) forwardl,m as close to EVOO: 5.33 (0.91) 0.19 (-0.38, (-0.39, 0.79) 0.28u study (0.95) 0.76) termination Pre Med date as Nuts: 5.20 Med Nuts: 0.44 Med Nuts: 0.49 (- N possible (0.86) (-0.14, 1.02) 0.16, 1.14) (average 4.1 yrs) Post Med Post: 5.15 EVOO: 5.44 Post Med Nuts: 5.56

Change Med Change: - EVOO: 0.11 0.08 (0.87) (0.86) Change Med Nuts: 0.36 (0.90)

167 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Digit span Pre Med Pre: 3.95 Med EVOO: Med EVOO: 0.58 Overall: N backwardl,m EVOO: 3.76 (0.79) 0.49 (-0.01, (-0.02, 1.19) 0.14u (0.84) 0.99) Pre Med Nuts: 3.83 Med Nuts: 0.58 Med Nuts: 0.74 Y (0.75) (0.07, 1.09) (0.08, 1.40)

Post Med Post: 3.71 EVOO: 4.01 Post Med Nuts: 4.17

Change Med Change: - EVOO: 0.25 0.24 (1.08) (1.09) Change Med Nuts: 0.34 (1.13) Lee (2015) Corsi blocks Baseline, Pre: N/R Pre: N/R 925.26 1.18 (-0.05, 2.40) 0.054 N (5) RT (ms)o,p 11 d Post: N/R Post: N/R Change: Change: - 603.26 322.00 (1254.19) (1368.18)

Corsi blocks Pre: N/R Pre: N/R 0.15 0.57 (-0.58, 1.73) 0.70 N spano,p Post: N/R Post: N/R

168 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Change: 0.28 Change: 0.13 (1.10) (0.84)

3-Back Pre: N/R Pre: N/R -251.27 -1.03 (-2.28, 0.23) 0.1 N reaction - Post: N/R Post: N/R time (ms)p,q Change: - Change: - 374.77 123.50 (499.25) (426.18)

3-Back - Pre: N/R Pre: N/R -6.46 -1.55 (-2.90, 0.21) 0.02 N correct Post: N/R Post: N/R responsesp,q Change: - Change: 4.14 2.32 (5.80) (7.27)

Numeric WM Pre: N/R Pre: N/R -10.47 -0.57 (-1.73, 0.58) 0.75 N RT (ms)o,p Post: N/R Post: N/R Change: - Change: - 85.96 75.49 (59.80) (181.35)

Numeric WM Pre: N/R Pre: N/R 2.03 0.57 (-0.58, 1.73) 0.87 N accuracyo,p Post: N/R Post: N/R Change: 1.11 Change: - (5.54) 0.92 (6.36) Knight Digit span Baseline, 3 Pre: 11.2 Pre: 10.9 0.20u (2016)t (6) forward MThs, 6 (2.0) (2.1) MThs

169 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value 3MThs: 3MThs: -0.56 (-1.25, -0.27 (-0.61, 0.06) N Post: 11.18 Post: 11.44 0.13) (1.99) (2.03) Change: - Change: 0.54 0.02

6MThs: 6MThs: 0.00 (-0.69, 0.00 (-0.33, 0.33) N Post: 11.04 Post: 10.74 0.69) (2.26) (2.30) Change: - Change: - 0.16 0.16

Digit span Pre: 9.2 (2.2) Pre: 9.3 (2.1) 0.84u backward

3MThs: 3MThs: 0.21 (-0.52, 0.10 (-0.24, 0.43) N Post: 9.49 Post: 9.38 0.94) (1.85) (1.89) Change: 0.29 Change: 0.08

6MThs: 6MThs: 0.00 (-0.73, 0.00 (-0.33, 0.33) N Post: 9.41 Post: 9.51 0.73) (2.26) (2.28) Change: 0.21 Change: 0.21

170 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Letter- Pre: 20.5 Pre: 20.9 0.85u number (2.5) (2.8) sequencing (LNS) 3MThs: 3MThs: -0.12 (-1.02, -0.05 (-0.38, 0.29) N Post: 20.40 Post: 20.92 0.78) (2.47) (2.52) Change: - Change: 0.02 0.10

6MThs: 6MThs: 0.20 (-0.70, 0.08 (-0.26, 0.41) N Post: 20.68 Post: 20.88 1.10) (3.15) (3.24) Change: 0.18 Change: - 0.02 Processing speed Wardle Choice Baseline, 6 N/R N/R N/A N/A 6 wks & N (2000) (2) reaction time wks, 12 12 wks: (modification wks >0.05u of Erikson & Erikson task) McMillan Simple RT Baseline, Pre: 277 (59) Pre: 289.04 -8.63 (-56.21, -0.15 (-0.93, 0.64) 0.66 N (2011) (3) (ms) 10 d Post: 253.33 (56) 38.95) (32) Post: 274 Change: - (46) 23.67 Change: - 15.04

171 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Choice RT Pre: 392 (86) Pre: 377 (45) -18 (-74.13, -0.26 (-1.04, 0.53) 0.30 N (ms) Post: 375 Post: 378 38.13) (37) (57) Change: - Change: 1.00 17.00

Choice RT Pre: 97.22 Pre: 96.16 -3.01 (-5.61, - -0.93 (-1.75, - 0.13 Y accuracy (%) (2.24) (3.79) 0.41) 0.10) Post: 94.20 Post: 96.15 (4.19) (3.79) Change: - Change: - 3.02 0.01 PREDIMED publications (2011-2015) Sanchez- N/A N/A N/A N/A N/A N/A N/A N/A Villegas (2011) (7) Martinez- TMT-Ac 6.5 yrs Pre Med Pre: N/A Med EVOO: - Med EVOO: - 0.41 N Lapiscina EVOO: N/A 5.29 (-17.08, 0.16 (-0.52, 0.20) (2013) (8) Pre Med 6.50) Nuts: N/A Med Nuts: 0.78 Med Nuts: 0.02 (- N Post Med Post: 70.79 (-11.31, 12.87) 0.34, 0.38) EVOO: 65.50 (34.13) (31.98) Post Med Nuts: 71.57 (32.94)

172 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2013) (9) Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2014) (10) Valls-Pedret Color Trail Baseline, Pre Med Pre: 57.00 Med EVOO: - Med EVOO: - Overall: N (2015) (4) test part 1l,m as close to EVOO: 62.60 (19.78) 10.30 (-22.36, 0.51 (-1.11, 0.09) 0.045u study (19.99) 1.76) termination Pre Med date as Nuts: 61.15 Med Nuts: -2.05 Med Nuts: -0.11 N possible (18.39) (-14.55, 10.45) (-0.75, 0.53) (average 4.1 yrs) Post Med Post: 61.53 EVOO: 56.83 Post Med Nuts: 63.63

Change Med Change: 4.53 EVOO: -5.77 (17.78) (17.38)

173 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Change Med Nuts: 2.48 (18.84) Lee (2015) Simple RT Baseline, Pre: N/R Pre: N/R 5.24 0.57 (-0.58, 1.73) 0.78 N (5) (ms)o,p 11 d Post: N/R Post: N/R Change: 8.79 Change: 3.55 (21.88) (64.84)

Choice RT Pre: N/R Pre: N/R 5.49 0.57 (-0.58, 1.73) 0.53 N (ms)o,p Post: N/R Post: N/R Change: 0.47 Change: - (28.61) 5.02 (23.57)

Choice RT - Pre: N/R Pre: N/R 1.04 0.57 (-0.58, 1.73) 0.35 N correct Post: N/R Post: N/R responseso,p Change: 0.78 Change: - (1.68) 0.26 (2.98) Knight Symbol Baseline, 3 Pre: 19.3 Pre: 19.5 0.78u (2016)t (6) search MThs, 6 (4.1) (4.5) MThs 3MThs: 3MThs: -0.54 (-1.99, -0.12 (-0.46, 0.21) N Post: 21.39 Post: 22.13 0.91) (5.07) (5.12) Change: 2.09 Change: 2.63

6MThs: 6MThs: -0.54 (-1.99, -0.12 (-0.46, 0.21) N 0.91)

174 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Post: 24.83 Post: 25.57 (- (5.20) 12.26) Change: 5.53 Change: 6.07

Coding Pre: 40.7 Pre: 42.8 0.92u (9.3) (10.8)

3MThs: 3MThs: -0.65 (-4.05, -0.06 (-0.40, 0.27) N Post: 46.07 Post: 48.82 2.75) (9.88) (9.94) Change: 5.37 Change: 6.02

6MThs: 6MThs: -0.29 (-3.69, -0.03 (-0.36, 0.31) N Post: 52.34 Post: 54.73 3.11) (9.90) (9.96) Change: Change: 11.64 11.93

Processing Pre: N/R Pre: N/R 0.81u speed composite 3MThs: 3MThs: -0.41 (-0.96, -0.25 (-0.59, 0.09) N Post: -0.20 Post: 0.21 0.14) (1.61) (1.64)

175 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value 6MThs: 6MThs: -0.37 (-0.91, -0.23 (-0.57, 0.10) N Post: -0.18 Post: 0.19 0.17) (1.57) (1.60) Verbal & visual memory Wardle Verbal Baseline, 6 N/R N/R N/A N/A 6 wks & N/A (2000) (2) immediate wks, 12 12 wks: free recall wks >0.05u McMillan Immediate Baseline, Pre: 7.67 Pre: 7.67 0.38 (-1.32, 0.18 (-0.61, 0.97) 0.13 N (2011) (3) word recall 10 d (2.00) (2.10) 2.08) Post: 8.92 Post: 8.54 (2.39) (2.18) Change: 1.25 Change: 0.87

Delayed word Pre: 6.33 Pre: 7.00 1.65 (-0.2, 3.5) 0.72 (-0.09, 1.53) 0.17 N recall (2.38) (2.08) Post: 7.83 Post: 6.85 (2.36) (2.64) Change: 1.5 Change: - 0.15

Word recog Pre: 81.95 Pre: 88.72 3.59 (-5.22, 0.33 (-0.46, 1.12) 0.31 N accuracy (%) (14.03) (6.02) 12.40) Post: 85.28 Post: 88.46 (8.46) (7.28) Change: 3.33 Change: - 0.26

176 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Word recog Pre: 894 Pre: 945 162 (1.32, 0.81 (-0.01, 1.62) 0.04 N RT (ms) (150) (227) 322.68) Post: 865 Post: 754 (210) (152) Change: -29 Change: -191

Delayed Pre: 88.75 Pre: 90.00 0.22 (-7.32, 0.02 (-0.76, 0.81) 0.95 N picture recog (7.65) (10.26) 7.76) (%) Post: 86.66 Post: 87.69 (7.92) (12.35) Change: - Change: - 2.09 2.31

Delayed Pre: 742 Pre: 752 (97) 143 (2.72, 0.82 (0.00, 1.63) 0.16 Y picture recog (223) Post: 719 283.28) RT Post: 852 (104) (199) Change: -33 Change: 110 PREDIMED publications (2011-2015) Sanchez- N/A N/A N/A N/A N/A N/A N/A N/A Villegas (2011) (7) Martinez- RAVLT, 6.5 yrs Pre Med Pre: N/A Med EVOO: Med EVOO: 0.25 0.27 N Lapiscina immediate EVOO: N/A 2.28 (-0.99, (-0.11, 0.61) (2013) (8) recallc Pre Med 5.55) Nuts: N/A N

177 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Post Med Post: 30.62 Med Nuts: 0.66 Med Nuts: 0.06 (- EVOO: 32.90 (9.54) (-3.06, 4.38) 0.30, 0.42) (8.83) Post Med Nuts: 31.28 (10.60)

Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A

RAVLT, Pre Med Pre: N/A Med EVOO: Med EVOO: 0.21 0.25 N delayed EVOO: N/A 0.60 (-0.40, (-0.14, 0.57) recallc Pre Med 1.60) Nuts: N/A Med Nuts: -0.04 Med Nuts: -0.01 N Post Med Post: 5.21 (-1.13, 1.05) (-0.37, 0.35) EVOO: 5.81 (2.97) (2.69) Post Med Nuts: 5.17 (3.02)

Change Med Change: N/A EVOO: N/A

178 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Change Med Nuts: N/A

Verbal paired Pre Med Pre: N/A Med EVOO: Med EVOO: 0.17 0.05 N associatesc EVOO: N/A 0.62 (-0.73, (-0.19, 0.52) Pre Med 1.97) Nuts: N/A Med Nuts: -0.76 Med Nuts: -0.20 N Post Med Post: 12.81 (-2.13, 0.61) (-0.56, 0.16) EVOO: 13.43 (3.80) (3.70) Post Med Nuts: 12.05 (3.75)

Change Med EVOO: N/A Change: N/A Change Med Nuts: N/A

ROCF, Pre Med Pre: N/A Med EVOO: Med EVOO: 0.33 0.01 N immediatec EVOO: N/A 1.93 (-0.18, (-0.03, 0.69) Pre Med 4.04) Nuts: N/A N

179 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Post Med Post: 11.95 Med Nuts: -1.01 Med Nuts: -0.17 EVOO: 13.88 (3.70) (-3.11, 1.09) (-0.53, 0.19) (6.67) Post Med Nuts: 10.94 (6.61)

Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A

ROCF, Pre Med Pre: N/A Med EVOO: Med EVOO: 0.30 0.03 N delayedc EVOO: N/A 2.14 (-0.42, (-0.06, 0.66) Pre Med 4.70) Nuts: N/A Med Nuts: -0.30 Med Nuts: -0.05 N Post Med Post: 11.13 (-2.63, 2.03) (-0.41, 0.31) EVOO: 13.27 (6.15) (7.54) Post Med Nuts: 10.83 (6.56)

Change Med Change: N/A EVOO: N/A

180 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Change Med Nuts: N/A Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2013) (9) Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2014) (10) Valls-Pedret RAVLT, total Baseline, Pre Med Pre: 39.24 Med EVOO: Med EVOO: 0.30 <0.05 N (2015) (4) learningk,l as close to EVOO: 39.31 (7.90) 2.40 (-0.25, (-0.03, 0.64) study (7.92) 5.05) termination date as Pre Med Med Nuts: 2.16 Med Nuts: 0.27 (- >0.05 N possible Nuts: 39.46 (-0.53, 4.85) 0.07, 0.61) (average (7.90) Overall: 4.1 yrs) 0.04u Post Med Post: 41.34 EVOO: 43.81 Post Med Nuts: 43.72

Change Med Change: 2.10 EVOO: 4.50 (7.19) (7.20) Change Med Nuts: 4.26 (7.18)

181 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value RAVLT, Pre Med Pre: 6.37 Med EVOO: Med EVOO: 0.18 Overall: N delayed EVOO: 6.61 (2.92) 0.52 (-0.47, (-0.16, 0.51) 0.06u recallk,l (2.96) 1.51) Pre Med Nuts: 6.48 Med Nuts: 0.85 Med Nuts: 0.29 (- N (2.94) (-0.15, 1.85) 0.05, 0.63)

Post Med Post: 7.32 EVOO: 8.08 Post Med Nuts: 8.28

Change Med Change: 0.95 EVOO: 1.47 (2.55) (2.56) Change Med Nuts: 1.80 (2.56)

Paired Pre Med Pre: 14.89 Med EVOO: Med EVOO: 0.11 Overall: N associatesk,l EVOO: 15.25 (3.19) 0.34 (-0.73, (-0.23, 0.44) 0.56u (3.22) 1.41) Pre Med Nuts: 15.25 Med Nuts: 0.50 Med Nuts: 0.16 (- N (3.20) (-0.59, 1.59) 0.18, 0.49)

182 Supplementary Material 3.12. - cont.

Change Med Change: - EVOO: 0.25 0.09 (3.29) (3.30) Change Med Nuts: 0.41 (3.31)

Memory Pre Med Pre: -0.04 Med EVOO: Med EVOO: 0.33 <0.25 N compositek,l EVOO: 0.02 (0.22) 0.21 (0.00, 0.42) (-0.01, 0.66) (0.74) Pre Med Med Nuts: 0.26 Med Nuts: 0.39 >0.05 Y Nuts: 0.013 (0.04, 0.48) (0.05, 0.73) (0.77) Overall: 0.04u Post Med Post: -0.21 EVOO: 0.06 Post Med Nuts: 0.10

Change Med Change: - EVOO: 0.04 0.17 (0.76)

183 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value (0.77) Change Med Nuts: 0.09 (0.75) Lee (2015) Immediate Baseline, Pre: N/R Pre: N/R 1.78 1.73 (0.34, 3.11) 0.01 Y (5) word recall – 11 d Post: N/R Post: N/R correct Change: 1.14 Change: - responsesp,q,s (2.01) 0.64 (1.97)

Immediate Pre: N/R Pre: N/R -0.68 -1.33 (-2.63, - 0.04 Y word recall – Post: N/R Post: N/R 0.02) incorrect Change: - Change: 0.36 responsesp,q,s 0.32 (0.95) (0.79)

Delayed word Pre: N/R Pre: N/R 1.15 0.85 (-0.38, 2.09) 0.17 N recall – Post: N/R Post: N/R correct Change: 0.67 Change: - responsesp,r,s (2.46) 0.48 (2.42)

Delayed word Pre: N/R Pre: N/R -1.04 -1.29 (-2.59, 0.01) 0.046 N recall – Post: N/R Post: N/R incorrect Change: - Change: 0.71 responsesp,r,s 0.33 (1.15) (1.38)

184 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Word recog Pre: N/R Pre: N/R -8.73 -0.57 (-1.73, 0.58) 0.78 N accuracy Post: N/R Post: N/R (%)o,p Change: - Change: - 8.87 (9.11) 0.14 (9.51)

Word recog Pre: N/R Pre: N/R -15.7 -0.57 (-1.73, 0.58) 0.69 N RT (ms)o,p Post: N/R Post: N/R Change: - Change: - 49.09 33.39 (93.63) (123.81)

Picture recog Pre: N/R Pre: N/R 2.38 0.72 (-0.44, 1.89) 0.23 N – correct Post: N/R Post: N/R responseso,p Change: - Change: - 0.14 (5.17) 2.52 (6.96)

Picture recog Pre: N/R Pre: N/R 112.18 0.57 (-0.58, 1.73) 0.36 N RT (ms)o,p Post: N/R Post: N/R Change: - Change: - 5.04 (131.30) 117.22 (530.47) Knight Benton Baseline, 3 Pre: 6.2 (1.5) Pre: 6.1 (1.3) 0.55u (2016)t (6) Visual MThs, 6 Retention MThs 3MThs: 3MThs: -0.34 (-0.82, -0.24 (-0.58, 0.10) N Test (BVRT) Post: 5.21 Post: 5.45 0.14) (1.66) (1.68)

185 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Change: - Change: - 0.99 0.65

6MThs: 6MThs: -0.32 (-0.80, -0.23 (-0.56, 0.11) N Post: 6.07 Post: 6.29 0.16) (1.78) (1.80) Change: - Change: 0.19 0.13

RAVLT Pre: 76.7 Pre: 75.7 0.24u (total score) (14.9) (13.8)

3MThs: 3MThs: -0.18 (-5.04, -0.01 (-0.35, 0.32) N Post: 79.70 Post: 78.88 4.68) (14.66) (14.86) Change: 3.0 Change: 3.18

6MThs: 6MThs: -2.36 (-7.22, -0.16 (-0.50, 0.17) N Post: 86.18 Post: 87.54 2.50) (13.73) (15.44) Change: 9.48 Change: 11.84

Memory Pre: N/R Pre: N/R 0.44u composite

186 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value 3MThs: 3MThs: -0.45 (-1.29, -0.18 (-0.52, 0.15) N Post: -0.22 Post: 0.23 0.39) (2.45) (2.50)

6MThs: 6MThs: -0.10 (-0.97, -0.04 (-0.37, 0.30) N Post: -0.05 Post: 0.05 0.77) (2.56) (2.58) Visuo-spatial Wardle N/A N/A N/A N/A N/A N/A N/A N/A (2000) (2) McMillan N/A N/A N/A N/A N/A N/A N/A N/A (2011) (3) PREDIMED publications (2011-2015) Sanchez- N/A N/A N/A N/A N/A N/A N/A N/A Villegas (2011) (7) Martinez- Clock 6.5 yrs Pre Med Pre: N/A Med EVOO: Med EVOO: 0.28 0.160 N Lapiscina Drawing Test EVOO: N/A 0.46 (-0.13, (-0.08, 0.64) (2013) (8) (CDT)c Pre Med 1.05) Nuts: N/A Med Nuts: 0.13 Med Nuts: 0.08 (- N Post Med Post: 5.07 (-0.48, 0.74) 0.28, 0.44) EVOO: 5.53 (1.66) (1.61) Post Med Nuts: 5.20 (1.68)

187 Supplementary Material 3.12. - cont.

ROCF, copyc Pre Med Pre: N/A Med EVOO: Med EVOO: 0.29 0.11 N EVOO: N/A 1.74 (-0.43, (-0.07, 0.65) Pre Med 3.91) Nuts: N/A Med Nuts: 0.00 Med Nuts: 0.00 (- N Post Med (-2.45, 2.45) 0.36, 0.36) EVOO: 29.87 Post: 28.13 (5.98) (6.08) Post Med Nuts: 28.13 (7.06)

Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A Martinez- Clock 6.5 yrs Pre Med Pre: N/A Med EVOO: Med EVOO: 0.32 0.020 Y Lapiscina Drawing Test EVOO: N/A 0.50 (0.07, 0.93) (0.04, 0.60) (2013) (9) (CDT)f Pre Med Nuts: N/A Med Nuts: 0.32 Med Nuts: 0.19 (- N (-0.16, 0.80) 0.09, 0.48)

188 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Post Med Post: 4.95 EVOO: 5.45 (1.66) (1.53) Post Med Nuts: 5.27 (1.66)

Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A Martinez- Clock 6.5 yrs CLU CLU 0.60 (0.24, 0.96) 0.42 (0.09, 0.75) 0.001 Y Lapiscina Drawing Test CT/TT: 5.50 CT/TT: 4.90 (2014) (10) (CDT)h,i,j (1.42) (1.40)

CLU CLU 0.40 (-0.11, 0.28 (-0.16, 0.72) 0.126 N CC: 5.24 CC: 4.83 0.92) (1.47) (1.36)

CR1 CR1 0.46 (0.13, 0.79) 0.33 (0.03, 0.64) 0.006 Y GG: 5.43 GG: 4.96 (1.40) (1.41)

CR1 CR1 0.31 (-0.30, 0.22 (-0.31, 0.74) 0.762 N GA/AA: 5.24 GA/AA: 4.92 0.93) (1.47) (1.46)

189 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value PICALM PICALM 0.59 (0.18, 1.00) 0.40 (0.03, 0.76) 0.005 Y CT/TT: 5.36 CT/TT: 4.78 (1.46) (1.46)

PICALM PICALM 0.27 (-0.13, 0.16 (-0.22, 0.55) 0.186 N CC: 5.41 CC: 5.19 0.68) (1.35) (1.35)

ApoE4 non- ApoE4 non- 0.55 (0.25, 0.85) N/A <0.001 N/A carriers: N/R carriers: N/R

ApoE4 ApoE4 0.33 (-0.60, N/A 0.477 N/A carriers: N/R carriers: N/R 1.27) Valls-Pedret N/A N/A N/A N/A N/A N/A N/A N/A (2015) (4)

Lee (2015) N/A N/A N/A N/A N/A N/A N/A N/A (5) Knight Visual-spatial Baseline, 3 Pre: N/R Pre: N/R 0.10u (2016) (6) score MThs, 6 composite MThs 3MThs: 3MThs: -0.14 (-0.48, -0.14 (-0.48, 0.20) N Post: -0.07 Post: 0.07 0.20) (1.03) (0.96)

6MThs: 6MThs: -0.12 (-0.44, -0.12 (-0.46, 0.21) N Post: -0.06 Post: 0.06 0.20) (0.94) (0.98)

190 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value

Language Wardle N/A N/A N/A N/A N/A N/A N/A N/A (2000) (2) McMillan N/A N/A N/A N/A N/A N/A N/A N/A (2011) (3) PREDIMED publications (2011-2015) Sanchez- N/A N/A N/A N/A N/A N/A N/A N/A Villegas (2011) (7) Martinez- Semantic 6.5 yrs Pre Med Pre: N/A Med EVOO: Med EVOO: 0.22 0.24 N Lapiscina Verbal EVOO: N/A 0.89 (-0.57, (-0.14, 0.58) (2013) (8) Fluency- Pre Med 2.35) Animalsc Nuts: N/A Med Nuts: 0.01 Med Nuts: 0.00 (- N Post Med Post: 12.16 (-1.50, 1.52) 0.36, 0.36) EVOO: 13.05 (3.96) (4.11) Post Med Nuts: 12.17 (4.25)

Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A

191 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Phonemic Pre Med Pre: N/A Med EVOO: Med EVOO: 0.45 0.01 Y Verbal EVOO: N/A 4.37 (0.91, 7.83) (0.09, 0.81) Fluency- Pre Med FASc Nuts: N/A Med Nuts: 2.05 Med Nuts: 0.21 (- (-1.53, 5.63) 0.15, 0.57) N Post Med Post: 21.29 EVOO: 25.66 (8.78) (9.99) Post Med Nuts: 23.34 (10.38)

Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A

Boston Pre Med Pre: N/A Med EVOO: Med EVOO: 0.01 0.18 N Naming Testc EVOO: N/A 0.08 (-2.53, (-0.35, 0.37) Pre Med 2.69) Nuts: N/A Med Nuts: -1.76 Med Nuts: -0.23 N Post Med Post: 47.19 (-4.50, 0.98) (-0.59, 0.13) EVOO: 47.27 (6.10) (7.73)

192 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Post Med Nuts: 45.43 (8.21)

Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2013) (9) Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2014) (10) Valls-Pedret Verbal Baseline, Pre Med Pre: 19.02 Med EVOO: Med EVOO: 0.06 Overall: N (2015) (4) fluencyl,m as close to EVOO: 18.44 (3.99) 0.23 (-2.18, (-0.53, 0.65) 0.13u study (3.99) 2.64) termination Pre Med date as Nuts: 20.33 Med Nuts: -1.81 Med Nuts: -0.48 N possible (3.55) (-4.27, 0.65) (-1.13, 0.17) (average 4.1 yrs) Post Med Post: 19.32 EVOO: 18.97 Post Med Nuts: 18.82

193 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Change Med Change: 0.30 EVOO: 0.53 (4.20) (4.15) Change Med Nuts: -1.51 (3.80) Lee (2015) N/A N/A N/A N/A N/A N/A N/A N/A (5) Knight Initial letter Baseline, 3 Pre: 24.7 Pre: 25.4 0.34u (2016)t (6) fluency (ILF) MThs, 6 (8.5) (8.5) MThs 3MThs: 3MThs: -2.03 (-4.90, -0.24 (-0.57, 0.10) N Post: 24.37 Post: 27.10 0.84) (7.63) (7.75) Change: - Change: 1.70 0.33

6MThs: 6MThs: -0.95 (-3.82, -0.11 (-0.45, 0.22) N Post: 25.60 Post: 27.25 1.92) (8.03) (8.14) Change: 0.90 Change: 1.85

Excluded Pre: 23.3 Pre: 21.9 0.18u letter fluency (9.4) (7.8) (ELF) 3MThs: 3MThs: -0.32 (-0.66, 0.02) N

194 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Post: 22.39 Post: 23.77 -2.78 (-5.71, (7.21) (7.32) 0.15) Change: - Change: 1.87 0.91

6MThs: 6MThs: -2.32 (-5.25, -0.27 (-0.60, 0.07) N Post: 24.26 Post: 25.18 0.61) (8.89) (9.04) Change: 0.96 Change: 3.28 Executive function Wardle N/A N/A N/A N/A N/A N/A N/A N/A (2000) (2) McMillan N/A N/A N/A N/A N/A N/A N/A N/A (2011) (3) PREDIMED publications (2011-2015) Sanchez- N/A N/A N/A N/A N/A N/A N/A N/A Villegas (2011) (7) Martinez- Similaritiesc 6.5 yrs Pre Med Pre: N/A Med EVOO: Med EVOO: 0.15 0.58 N Lapiscina EVOO: N/A 0.79 (-1.04, (-0.20, 0.51) (2013) (8) Pre Med 2.62) Nuts: N/A Med Nuts: 0.53 Med Nuts: 0.10 (- N Post Med Post: 10.11 (-1.37, 2.43) 0.26, 0.46) EVOO: 10.90 (4.37) (5.40)

195 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Post Med Nuts: 10.64 (5.64)

Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A

TMT-Bc Pre Med Pre: N/A Med EVOO: - Med EVOO: - 0.21 N EVOO: N/A 20.61 (-67.25, 0.16 (-0.52, 0.20) Pre Med 26.03) Nuts: N/A Med Nuts: 12.22 Med Nuts: 0.10 (- N Post Med Post: 220.19 (-31.03, 55.47) 0.26, 0.46) EVOO: (117.18) 199.58 (134.95) Post Med Nuts: 232.41 (120.35)

Change Med Change: N/A EVOO: N/A Change Med Nuts: N/A

196 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2013) (9) Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2014) (10) Valls-Pedret Color Trail Baseline, Pre Med Pre: 129.99 Med EVOO: - Med EVOO: - <0.05 Y (2015) (4) test part 2l,m as close to EVOO: (41.79) 31.90 (-57.42, - 0.75 (-1.35, -0.14) study 136.55 6.38) termination (42.31) date as Pre Med Med Nuts: - Med Nuts: -0.32 >0.05 N possible Nuts: 131.59 13.33 (-40.17, (-0.97, 0.32) (average (39.88) 13.51) Overall: 4.1 yrs) 0.045u Post Med Post: 167.55 EVOO: 142.21 Post Med Nuts: 155.82

Change Med Change: EVOO: 5.66 37.56 (52.00) (50.34) Change Med Nuts: 24.23 (55.40)

197 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Frontal Pre Med Pre: 0.12 Med EVOO: Med EVOO: 1.02 <0.01 Y cognition EVOO: -0.06 (0.31) 0.56 (0.23, 0.89) (0.40, 1.64) compositel,m (0.60) Pre Med Med Nuts: 0.36 Med Nuts: 0.72 <0.25 Yx Nuts: 0.04 (0.04, 0.68) (0.06, 1.38) (0.57) Overall: 0.003u Post Med Post: -0.21 EVOO: 0.17 Post Med Nuts: 0.07

Change Med Change: - EVOO: 0.23 0.33 (0.64) (0.63) Change Med Nuts: 0.03 (0.68) Lee (2015) N/A N/A N/A N/A N/A N/A N/A N/A (5) Knight Stroop test Baseline, 3 Pre: 2.5 (0.5) Pre: 2.5 (0.6) 0.82u (2016)t (6) (interference MThs, 6 control) MThs 3MThs: 3MThs: 0.00 (-0.19, 0.00 (-0.33, 0.33) N Post: 2.33 Post: 2.33 0.19) (0.52) (0.53) Change: - Change: - 0.17 0.17

198 Supplementary Material 3.12. - cont.

Tower of Pre: 15.4 Pre: 15.9 0.67u London (3.4) (3.3) (TOL) 3MThs: 3MThs: 0.18 (-0.95, 0.05 (-0.28, 0.39) N Post: 16.20 Post: 16.52 1.31) (3.06) (3.12) Change: 0.80 Change: 0.62

6MThs: 6MThs: 0.48 (-0.65, 0.14 (-0.19, 0.48) N Post: 17.49 Post: 17.51 1.61) (3.21) (3.24) Change: 2.09 Change: 1.61

Executive Pre: N/R Pre: N/R 0.11u function composite 3MThs: 3MThs: -1.10 (-2.48, -0.27 (-0.60, 0.07) N Post: -0.28 Post: 0.82 0.28) (2.08) (5.43)

199 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value 6MThs: 6MThs: -0.39 (-1.07, -0.19 (-0.53, 0.14) N Post: -0.09 Post: 0.30 0.29) (1.89) (2.11) Motor control Wardle Tapping Baseline, 6 N/R N/R N/A N/A 6 wks & N (2000) (2) speed wks, 12 12 wks: wks >0.05u McMillan N/A N/A N/A N/A N/A N/A N/A N/A (2011) (3) PREDIMED publications (2011-2015) Sanchez- N/A N/A N/A N/A N/A N/A N/A N/A Villegas (2011) (7) Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2013) (8) Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2013) (9) Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2014) (10) Valls-Pedret N/A N/A N/A N/A N/A N/A N/A N/A (2015) (4) Lee (2015) N/A N/A N/A N/A N/A N/A N/A N/A (5)

200 Supplementary Material 3.12. - cont.

Cognitive Follow up Exp mean Control Between group Between group Between ES function test time points (SD) mean (SD) MD (95% CIs) ES (95% CIs) group p- significant value Knight N/A N/A N/A N/A N/A N/A N/A N/A (2016) (6) We used random-effect models. Standardised mean differences were calculated by subtracting the mean change in the control group from the mean change in the Med diet group and dividing by the pooled SD at baseline. Data is presented as standardised mean difference and 95% CI. For PREDIMED studies with two Med Exp groups (4, 7-10) we divided Con by two to calculate ES. For post-test comparisons, ESs were calculated by subtracting the Con post-test from the Exp post-test and dividing by the pooled post SD (7-10). Values for PREDIMED and MedLey composites (4, 6) relate to z scores. a All measures extracted from Figure 2 in Wardle (2000) (2) using Matlab and number of correct responses across 10 trials summed at each time point. b Outcomes reported for n=155 in total (Med n=53; Con n=50; LF n=52). c Outcomes reported for n=268 in total (Med EVOO n=91; Med Nuts n=88; Con n=89). d Outcomes available for n=263 in total (Med EVOO n=90; Med Nuts n=85; Con n=88). e ES and CI calculated from incident percentage using Campbell Collaboration calculator: https://www.campbellcollaboration.org/effect-size- calculato.html f Outcomes available for n=522 in total (Med EVOO n=224; Med Nuts n=166; Con n=132). g Incident MCI and dementia obtained from review of medical records for total n=1055 as randomised.

201 Supplementary Material 3.12. - cont. h For this test the intervention group gathers two Med diet groups, i.e., Med Diet EVOO and Med Diet Nuts merged to evaluate effect of Med as whole dietary pattern. i Multivariable adjusted means and differences extracted from paper as unadjusted N/R. j Outcomes for genotypes reported for variable n’s. CLU CT/TT total=309 (Med=226; Con=83); CLU CC total=198 (Med=153; Con=45); CR1 GG total=375 (Med=278; Con=97); CR1 GA/AA total=133 (Med=101; Con=32); PICALM CT/TT total=248 (Med=179; Con=69); PICALM CC total=256 (Med=196; Con=60). k Outcomes reported for n=334 in total (Med EVOO n=127; Med Nuts n=112; Con n=95). l Multivariable adjusted means and differences extracted as reported in main paper (although unadjusted values available in author suppl. tables). m Outcomes reported for n=96 in total (Med EVOO n=41; Med Nuts n=25; Con n=30). n Incident MCI cases reported for completers n=334. No dementia cases documented.

0 Change score reported for n=23 in cross-over RCT. p ES and CI calculated from F test and n divided by two, as cross-over trial, using Campbell Collaboration calculator: https://www.campbellcollaboration.org/escalc/html/EffectSizeCalculator-SMD5.php q Change score reported for n=22 in cross-over RCT. r Change score reported for n=21 in cross-over RCT. s Missing cases reported but number unclear. t Outcomes reported for n=137 completers (Med=70; Con=67).

202 Supplementary Material 3.12. - cont. u p value by analysis of covariance. v Interaction effect from mixed factorial repeated measures ANCOVA. w p value by analysis of covariance of three groups. However, post hoc tests indicated Con did better than either Med or LF. x The p value is discordant with the CI. We note the ANCOVA used a small n with a very large number of co-variates and may therefore have suffered from reduced statistical power.

CIs, confidence intervals; CLU, CR1, PICALM & ApoE4, genotypes previously identified as related to cognitive decline; Con, control group; d, days; ES, effect size; EVOO, extra virgin olive oil; Exp, experimental group; FAS, letters used in phonemic verbal fluency test; MCI, mild cognitive impairment; Med, Mediterranean intervention; MD, mean difference; MMSE, mini mental state examination; MThs, months; N, no; N/A, not available as not collected during study; N/R, outcome measured but not reported; PREDIMED, Prevencion con Dieta Mediterranea, translated to PREvention with MEDiterranean Diet; RAVLT, Rey auditory verbal learning test; recog, recognition; ROCF, Rey–Osterrieth complex figure; RT, reaction time; RVIP, rapid visual information processing; SAT, sustained attention task; SD, standard deviation; suppl., supplementary; TMT-A, trail making test part A; TMT-B, trail making test part B; wks, weeks; WM, working memory; Y, yes; yrs, years; %, percent.

203 Supplementary Material 3.13. - Brain morphology or function outcomes

Author Brain Follow up Exp mean (SD) Control mean Between Between Between ES (yr) markers time (SD) group MD group ES group p- significant points (95% CIs) (95% CIs) value Functional outcomes Wardle N/A N/A N/A N/A N/A N/A N/A N/A (2000) (2) McMillan N/A N/A N/A N/A N/A N/A N/A N/A (2011) (3) PREDIMED publications (2011-2015) Sanchez- Plasma Baseline Pre Med EVOO: Pre: N/R 0.68 Villegas BDNF only for N/R (2011) (7) participant Pre Med Nuts: N/R s with prevalent Post Med EVOO: Post: 23.37 (33.66) Med EVOO: Med EVOO: N depression, 24.66 (38.32) 1.29 (-12.72, 0.03 (-0.34, 3 yrs for Post Med Nuts: 15.30) 0.41) entire 24.87 (34.21) cohort Med Nuts: Med Nuts: N Change Med Change: N/A 1.50 (-11.81, 0.04 (-0.34, EVOO: N/R 14.81) 0.43) Change Med Nuts: N/R

Risk of Pre Med EVOO: Pre: N/R 0.97 plasma N/R BDNF below Pre Med Nuts: N/R

204 Supplementary Material 3.13. - cont.

Author Brain Follow up Exp mean (SD) Control mean Between Between Between ES (yr) markers time (SD) group MD group ES group p- significant points (95% CIs) (95% CIs) value 10th percentile Post Med EVOO: Post: N/R N/A Med EVOO: 0.04 N N/R OR=1.02 Post Med Nuts: N/R (0.38, 2.76)

Change Med Med Nuts: Y EVOO: N/R OR=0.22 Change Med Nuts: (0.05, 0.90) N/R

Plasma Pre Med EVOO: Pre: N/R 0.74 BDNF when N/R hypertension Pre Med Nuts: N/R prevalent at baselinea Post Med EVOO: Post: 25.77 (30.61) Med EVOO: Med EVOO: N 27.20 (34.94) 1.43 (-12.50, 0.04 (-0.37, Post Med Nuts: 15.36) 0.45) 26.05 (31.56) Med Nuts: Med Nuts: N Change Med Change: N/R 0.28 (-12.98, 0.01 (-0.41, EVOO: N/R 13.54) 0.43) Change Med Nuts: N/R

205 Supplementary Material 3.13. - cont.

Author Brain Follow up Exp mean (SD) Control mean Between Between Between ES (yr) markers time (SD) group MD group ES group p- significant points (95% CIs) (95% CIs) value Plasma Pre Med EVOO: Pre: N/R 0.51 BDNF when N/R diabetes Pre Med Nuts: N/R prevalent at baselineb Post Med EVOO: Post: 30.54 (22.63) Med EVOO: Med EVOO: N 26.89 (20.71) -3.65 (-19.13, -0.17 (-0.87, Post Med Nuts: 11.83) 0.53) 28.34 (25.25) Change Med Change: N/R Med Nuts: - Med Nuts: - N EVOO: N/R 2.20 (-19.00, 0.09 (-0.75, Change Med Nuts: 14.60) 0.58) N/R

Plasma Pre Med EVOO: Pre: N/R 0.98 BDNF when N/R hypercholest Pre Med Nuts: N/R erolaemia N prevalent at Post Med EVOO: Post: 23.52 (33.07) Med EVOO: Med EVOO: baselinec 23.31 (40.96) -0.21 (-17.95, -0.01 (-0.45, Post Med Nuts: 17.53) 0.44) 23.81 (29.60) N Med Nuts: Med Nuts: Change Med Change: N/R 0.29 (-15.51, 0.01 (-0.49, EVOO: N/R 16.09) 0.51) Change Med Nuts: N/R

206 Supplementary Material 3.13. - cont.

Author Brain Follow up Exp mean (SD) Control mean Between Between Between ES (yr) markers time (SD) group MD group ES group p- significant points (95% CIs) (95% CIs) value

Plasma Pre Med EVOO: Pre: N/R 0.007 BDNF when N/R depression Pre Med Nuts: N/R prevalent at baselined Post Med EVOO: Post: 22.72 (9.68) Med EVOO: Med EVOO: N 24.16 (13.80) 1.44 (-12.79, 0.11 (-0.92, Post Med Nuts: 15.67) 1.13) 32.37 (11.59) Med Nuts: Med Nuts: N Change Med Change: N/R 9.65 (-2.63, 0.83 (-0.23, EVOO: N/R 21.93) 1.88) Change Med Nuts: N/R Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2013) (8) Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2013) (9) Martinez- N/A N/A N/A N/A N/A N/A N/A N/A Lapiscina (2014) (10)

207 Supplementary Material 3.13. - cont.

Author Brain Follow up Exp mean (SD) Control mean Between Between Between ES (yr) markers time (SD) group MD group ES group p- significant points (95% CIs) (95% CIs) value Valls- N/A N/A N/A N/A N/A N/A N/A N/A Pedret (2015) (4) Lee N/A N/A N/A N/A N/A N/A N/A N/A (2015) (5) Knight N/A N/A N/A N/A N/A N/A N/A N/A (2016) (6) We used random-effect models. Standardised mean differences were calculated by subtracting the mean change in the control group from the mean change in the Med diet group and dividing by the pooled SD at baseline. Data is presented as standardised mean difference and 95% CI. a Outcome available for n=206 of total cohort. b Outcome available for n=79 of total cohort. c Outcome available for n=161 of total cohort. d Outcome available for n=37 of total cohort.

BDNF, brain derived neurotrophic factor; CIs, confidence intervals; ES, effect size; EVOO, extra virgin olive oil; Med, Mediterranean intervention; MD, mean difference; N, no; N/A, not available as not collected during study; N/R, outcome measured but not reported; OR, odds ratio; PREDIMED, Prevencion con Dieta Mediterranea, translated to PREvention with MEDiterranean Diet; SD, standard deviation; Y, yes; yr, year; %, percent.

208 References for Supplementary Materials

1. Radd-Vagenas S, Duffy SL, Naismith SL, Brew BJ, Flood VM, Fiatarone Singh MA. Effect of the Mediterranean diet on cognition and brain morphology and function: A systematic review of randomized controlled trials. Am J Clin Nutr 2018;107:389-404.

2. Wardle J, Rogers P, Judd P, Taylor MA, Rapoport L, Green M, Nicholson Perry K. Randomized trial of the effects of cholesterol-lowering dietary treatment on psychological function. Am J Med 2000;108:547-53.

3. McMillan L, Owen L, Kras M, Scholey A. Behavioural effects of a 10-day Mediterranean diet. Results from a pilot study evaluating mood and cognitive performance. Appetite 2011;56:143-47.

4. Valls-Pedret C, Sala-Vila A, Serra-Mir M, Corella D, de la Torre R, Martínez- González MÁ, Martínez-Lapiscina EH, Fitó M, Pérez-Heras A, Salas-Salvadó J. Mediterranean diet and age-related cognitive decline: a randomized clinical trial. JAMA Intern Med 2015;175:1094-103.

5. Lee J, Pase M, Pipingas A, Raubenheimer J, Thurgood M, Villalon L, Macpherson H, Gibbs A, Scholey A. Switching to a 10-day Mediterranean-style diet improves mood and cardiovascular function in a controlled crossover study. Nutrition 2015;31:647-52.

6. Knight A, Bryan J, Wilson C, Hodgson JM, Davis CR, Murphy KJ. The Mediterranean diet and cognitive function among healthy older adults in a 6- month randomised controlled trial: the MedLey study. Nutrients 2016;8:579.

7. Sanchez-Villegas A, Galbete C, Martinez-Gonzalez MA, Martinez JA, Razquin C, Salas-Salvado J, Estruch R, Buil-Cosiales P, Marti A. The effect of the Mediterranean diet on plasma brain-derived neurotrophic factor (BDNF) levels: The PREDIMED-NAVARRA randomized trial. Nutr Neurosci 2011;14:195-201.

209 8. Martinez-Lapiscina EH, Clavero P, Toledo E, San Julian B, Sanchez-Tainta A, Corella D, Lamuela-Raventos R, Martinez J, Martinez-Gonzalez M. Virgin olive oil supplementation and long-term cognition: the PREDIMED-NAVARRA randomized, trial. J Nutr Health Aging 2013;17:544-52.

9. Martinez-Lapiscina EH, Clavero P, Toledo E, Estruch R, Salas-Salvado J, Julian BS, Sanchez-Tainta A, Ros E, Valls-Pedret C, Martinez-Gonzalez MA. Mediterranean diet improves cognition: The PREDIMED-NAVARRA randomised trial. J Neurol Neurosurg Psychiatry 2013;84:1318-25.

10. Martinez-Lapiscina EH, Galbete C, Corella D, Toledo E, Buil-Cosiales P, Salas- Salvado J, Ros E, Martinez-Gonzalez MA. Genotype patterns at CLU, CR1, PICALM and APOE, cognition and Mediterranean diet: the PREDIMED- NAVARRA trial. Genes Nutr 2014;9:393.

11. Davis CR, Bryan J, Hodgson JM, Wilson C, Dhillon V, Murphy KJ. A randomised controlled intervention trial evaluating the efficacy of an Australianised Mediterranean diet compared to the habitual Australian diet on cognitive function, psychological wellbeing and cardiovascular health in healthy older adults (MedLey study): Protocol paper. BMC Nutrition 2015;1:35.

12. Davis C, Hodgson J, Bryan J, Garg M, Woodman R, Murphy K. Older Australians can achieve high adherence to the Mediterranean diet during a 6 month randomised intervention; results from the Medley study. Nutrients 2017;9:534.

210 CHAPTER FOUR

Development of the Mediterranean Diet and Culinary

Index (MediCul) tool

211 4.1 Introduction

Existing Mediterranean diet index tools have various limitations. For example, Chapters Two and Three within this thesis identified that previous Mediterranean diet interventions and index tools used to assess adherence to the dietary pattern did not necessarily represent the ‘traditional’ diet. In order to conduct well-designed trials in the future to test the effect of the ‘traditional’ Mediterranean diet on brain health, as well as other outcomes, a new tool is required.

The aim of this chapter was to describe the development of a new Mediterranean diet index tool to assess adherence to the ‘traditional’ dietary pattern and aspects of cuisine, which can be used within Western populations including among individuals at higher risk of cognitive decline.

4.1.1 Rationale for use of diet index tools

Diet index tools are widely used in research as they allow for a combined measure of dietary factors (e.g., foods, food groups, habits related to eating) that can be conveniently summarised in a single score (Bach et al., 2006; Hernández Ruiz et al., 2015). Index tools were originally developed to assess overall diet quality in relation to dietary guidelines, such as with the Diet Quality Index (DQI) (Patterson, Haines, & Popkin, 1994), Healthy Eating Index (HEI) (Kennedy, Ohls, Carlson, & Fleming, 1995), Alternative Healthy Eating Index (AHEI) (McCullough et al., 2002) and, more recently, the Australian Recommended Food Score (ARFS) (Collins et al., 2015), Healthy Eating Index for Australian Adults (HEIFA-2013) (Roy, Hebden, Rangan, & Allman-Farinelli, 2016) and Dietary Guideline Index (DGI) (McNaughton, Ball, Crawford, & Mishra, 2008) now adapted for use also with food records (Ward, Coates, & Hill, 2019). Latter indexes, such as those based on the Dietary Approaches to Stop Hypertension (DASH) (Kwan et al., 2013), Mediterranean-Dietary Approaches to Stop Hypertension (MIND) (Morris et al., 2015a), Nordic (Hillesund, Bere, Haugen, & Øverby, 2014) and the popular Mediterranean diet (Trichopoulou et al., 1995) aimed to assess adherence to a particular dietary pattern and relate the scores to various outcomes. The Dietary Inflammatory Index

212 (DII) aims to capture the inflammatory potential across any eating pattern and may therefore be relevant to multiple diets, including the Mediterranean diet (Shivappa, Steck, Hurley, Hussey, & Hébert, 2014).

4.1.2 Mediterranean diet index tools

Since 1995 at least 30 original Mediterranean diet index tools and their variants have been developed and used, many of which remain unnamed (see Appendix Five). Adherence scores from these tools have been associated with survival/mortality (Bonaccio et al., 2018; Sofi, Macchi, Abbate, Gensini, & Casini, 2014; Trichopoulou, Costacou, Bamia, & Trichopoulos, 2003); multiple chronic diseases (see Chapter One) risk ;factors such as inflammation, lipids, blood pressure, coagulation markers (Panagiotakos, Pitsavos, & Stefanadis, 2006), body weight/waist circumference (Estruch et al., 2016); the microbiome (De Filippis et al., 2016); diet quality (Mariscal-Arcas et al., 2007); and the degree of adherence to a Mediterranean diet (Trichopoulou et al., 1995). Surprisingly, most index tools to date have studied adherence to a Mediterranean dietary pattern within countries outside the Mediterranean region, such as the US, Canada, Northern Europe, Australia, Middle East and China.

The most widely used Mediterranean diet index tool (Hoffman & Gerber, 2013; Zaragoza-Martí, Cabañero-Martínez, Hurtado-Sánchez, Laguna-Pérez, & Ferrer- Cascales, 2018), frequently modified by researchers to better suit various populations and cohorts, is the Mediterranean Diet Score (MDS). This has also been called the Mediterranean Diet Pattern Score (MDPS) (Kouris-Blazos & Itsiopoulos, 2014). The MDS was originally developed, and later updated, by Trichopoulou (Trichopoulou et al., 1995; Trichopoulou et al., 2003) to include fish. More recently, the Mediterranean Diet Score (MedDietScore) (Panagiotakos, Milias, Pitsavos, & Stefanadis, 2006) and the Mediterranean Diet Adherence Screener (MEDAS) (see Appendix Six) (Schroder et al., 2011) have become increasingly popular. MEDAS has also been adapted for use in other populations, such as with the development of the Mediterranean Eating Pattern for Americans (MEPA) screener, which incorporates selected elements protective for brain health (Cerwinske, Rasmussen, Lipson, Volgman, & Tangney, 2017).

213 While most Mediterranean diet index tools aim to assess adherence to a Mediterranean dietary pattern among the general adult population, it is worth noting that some tools focus on specific stages of the life cycle, cohorts with particular conditions, or even specific meal occasions. Examples include the Mediterranean Diet Score for Pregnancy (MDS-P) (Mariscal-Arcas et al., 2009), KIDMED (Serra-Majem et al., 2004), one unnamed Mediterranean diet index tool used in relation to peripheral artery disease risk in people with type 2 diabetes (Ciccarone et al., 2003) and the Breakfast Quality Index (BQI) (Monteagudo et al., 2013).

The Mediterranean Lifestyle (MEDLIFE) index (Sotos-Prieto et al., 2014; Sotos-Prieto et al., 2015), which is based on the Spanish Mediterranean diet pyramid (Bach-Faig et al., 2011), was devised to more broadly assess adherence to the Mediterranean lifestyle as it includes physical activity, rest, social interaction and conviviality elements. This tool also assesses some unique Mediterranean dietary habits, e.g., infrequent snacking, which have rarely been captured within the Mediterranean context. Most recently, the Total Lifestyle Index (TLI) was constructed to measure adherence to the overall Mediterranean lifestyle pattern, including the Mediterranean diet, physical activity, sleep and daily living activities with social/intellectual aspects (Anastasiou et al., 2018). While the whole Mediterranean lifestyle may be of greatest importance for health, this has not yet been well studied, and is beyond the scope of this thesis.

4.1.3 Limitations of existing tools

Existing Mediterranean diet index tools and/or resulting scores show variation (Milà- Villarroel et al., 2011) based on:

a) their construction (a priori versus a posteriori; see Chapter One);

b) the number and type of elements included (typical range: 6-28 main elements);

c) cut-off points used for scoring;

d) total score (typical range: 0-130), which can also have variable weighting applied to the elements.

214 Scores for tools constructed a priori may better reflect the ‘traditional’ Mediterranean diet as the dietary elements are informed by historical intake data from Mediterranean populations at that time period (D'Alessandro & De Pergola, 2015) as compared to scores derived using data-driven models such as Principal Component Analysis (PCA).

However, while the original MDS index (Trichopoulou et al., 1995; Trichopoulou et al., 2003) was based on elements developed a priori, the cut-off points used to generate scoring were arbitrarily based on gender- and population-specific median intakes within Greece at that time. The use of such cut-off points may have been appropriate within a Mediterranean population which has maintained a traditional diet until the 1990s (Trichopoulou et al., 1995). However, this approach has been identified as a limitation when used in other countries or cohorts since such cut-off points will differ according to population characteristics (Martinez-Gonzalez, Hershey, Zazpe, & Trichopoulou, 2017; Sofi et al., 2014). For example, in some cohorts within the studies evaluated in a recent review of Mediterranean diet indexes, median intakes of meat and dairy were very high (D'Alessandro & De Pergola, 2018). Nevertheless, the use of medians, means and tertiles, derived from food frequency questionnaires has been common in research when deriving a score, even though these cut-off points may not necessarily represent levels related to ‘traditional’ Mediterranean intakes. Indeed, such cut-off points may simply delineate participants with more desirable versus undesirable intakes of the elements assessed.

Mediterranean diet index tools have also been criticised for lacking critical elements that characterise the ‘traditional’ Mediterranean diet and cuisine (D'Alessandro & De Pergola, 2015; Hoffman & Gerber, 2013) (see Chapter Two). For example, while a monounsaturated fatty acid (MUFA) to saturated fatty acid ratio may be appropriate for use within the Mediterranean region where MUFA predominantly comes from the high intake of extra virgin olive oil (EVOO), in Australia the major source of MUFA in the diet reported is meat (Baghurst, 1999), which has different and usually opposing health effects, compared to olive oil. More recently, meat products were also found to be the highest contributors to dietary MUFA among older Australians from the Blue Mountains Eye Study (BMES) (Flood, Webb, Rochtchina, Kelly, & Mitchell, 2007), and for adults

215 aged 19 years and older, when compared to other food groups, based on the Australian Health Survey 2011-12 (Australian Bureau of Statistics, 2014). However, very recent data are lacking and may be different due to the rise in popularity of olive oil. In addition, not all existing tools adequately assess the contribution of modern, ultra-processed foods to the diet, which is important if they are intended for use within Western populations. Further, to derive scores for some tools involves complex computations, so the tools cannot be easily used in clinical practice.

From a reliability and validity viewpoint, a recent systematic review of Mediterranean diet index tools found that few fulfilled quality criteria, especially with regards to test- retest reliability and concurrent validity against another diet reference method (Zaragoza- Martí et al., 2018). For example, relatively few tools have been compared for performance against a 24-hour food recall (Benitez-Arciniega et al., 2011), food record (Ciccarone et al., 2003), food frequency questionnaire (FFQ) (Schroder et al., 2011) or dietary biomarkers (Gerber et al., 2000; Panagiotakos et al., 2009). Tool validation against another diet reference method, particularly one that is not limited by the same recall bias, is now considered benchmark practice to determine whether a tool accurately measures what it intends to measure. Also, many existing index tools are not in the form of a dietary survey so a score cannot be directly derived, sometimes requiring assumptions to be made when deriving the score from a FFQ. Finally, to the author’s knowledge, no Mediterranean diet index tool has been specifically developed for, and validated within, the Australian context including among individuals diagnosed with mild cognitive impairment (MCI), which is a precursor to dementia (Petersen, 2004). This is an important research gap, given the emerging evidence that a Mediterranean dietary pattern may reduce rates of cognitive decline among high risk populations (Psaltopoulou et al., 2013; Scarmeas et al., 2009; Tangney et al., 2011).

4.1.4 Ideal characteristics for a new tool

Patterson et al. (1994) who developed the original Diet Quality Index (DQI) (not for a Mediterranean diet), identified that “the most serious limitation to research in nutrition epidemiology has been the lack of accurate and practical methods to measure diet”. This

216 concern is relevant also for clinical studies. However, ‘gold standard’ assessment methods such as a FR can be burdensome for the participant, as well as expensive and labour-intensive for the researcher (Collins, Watson, & Burrows, 2010). The use of diet index tools may overcome some of the limitations of using a FR. However, individuals may still misrepresent their diets in line with their optimistic biases about food benefits/risks (Miles & Scaife, 2003) and social desirability bias (Hebert, Clemow, Pbert, Ockene, & Ockene, 1995).

Nevertheless, diet index tools can capture the multi-dimensional nature of individual diets (Waijers, Feskens, & Ocké, 2007) as well as the cumulative impact of the elements measured (Golley et al., 2012) with reduced participant burden. An ideal diet index tool would: a) address known limitations of existing index tools, b) assess the whole of diet, c) be relatively simple to use, and d) be tested for reliability and validity in target populations, including against a dietary reference method that is not limited by the same recall bias.

Therefore, the aim of this Chapter was to describe the development of a new diet index tool to assess adherence to the ‘traditional’ Mediterranean dietary pattern and aspects of cuisine which:

a) is suitable for use in the Australian context; b) lends itself to various administrations, including online and in-person; c) can be used among adults with mild cognitive impairment (MCI); d) can be used to also indirectly derive a MEDAS score; e) has potential utility for research and clinical practice, without complex mathematical computations.

4.2 Method

The new Mediterranean diet index tool, which is in the form of a survey, was developed using guidelines (Passmore, Dobbie, Parchman, & Tysinger, 2002) and general recommendations (Cade, Thompson, Burley, & Warm, 2002) for survey construction and refined at multiple stages, as outlined below.

217 4.2.1 Identification of ‘traditional’ elements

A literature review was first conducted (see Chapter Two) and the ‘traditional’ Mediterranean cuisine was described from original data sources to improve current understanding. This was used to check against elements already included in existing Mediterranean tools from various countries (see Appendix Five) and to identify any gaps or potentially important elements that should also be included in the new tool to assess adherence to the ‘traditional’ diet. Further, consideration was given to definitions of the ‘traditional’ Mediterranean diet from various educational Mediterranean diet pyramids (Bach-Faig et al., 2011; Oldways, 2008; Willett et al., 1995) to inform the final selection of elements for the new index tool using professional judgement.

4.2.2 Tool naming

To reduce confusion in future with other Mediterranean tools, which can sometimes be difficult to discern, a name was devised. The new diet index tool was called the Mediterranean Diet and Culinary Index (MediCul) in an attempt to link together concepts of the Mediterranean diet, cuisine/culinary matters and a ‘food as medicine’ philosophy.

4.2.3 Tool development

It was intended that the MediCul tool approximate a short diet survey (defined as <50 items) (Calfas, Zabinski, & Rupp, 2000). Hence, short questions were developed using simple language, appropriate for an Australian and Western context. Expert dietetic knowledge was employed to minimise the use of leading questions since people tend to over-report what they perceive as good foods and under-report bad foods.

Both frequency and serve questions were developed to reduce participant burden, as frequency questions tend to be easier to answer. Frequency data can also explain much of the variation in dietary intake or be a greater contributor to variation than serve size for most foods (Collins et al., 2014; Thompson & Subar, 2017). Questions were worded in a systematic way. For example, “How often do you usually…” was stated for frequency

218 questions, and consistent response options were offered for “times per day, times per week, times per month and I don’t eat…”. For serve questions, “How many serves of X do you usually eat..?” was stated with relevant serve size examples being provided, as well as consistent response options.

Care was taken to ensure the final tool addressed gaps identified in existing tools. In particular, questions were devised, and checked, to ensure the tool assesses:

1. intake of unique Mediterranean foods and cuisine aspects. For example, typical food combinations such as sofrito; high moisture, lower heat cooking methods; growing of own vegetables; wholegrain/low glycaemic index (GI) bread preference (as a proxy for wholegrains); wine in small amounts and only with meals; meals frequently home cooked; fasting practice; napping after the midday meal; 2. discretionary foods, as these are frequently consumed in Western countries and make up approximately 35% of the energy intake in Australian diets (Australian Bureau of Statistics, 2014), e.g., hot chips, takeaway and sugary drinks; 3. foods consumed in the ‘traditional’ Mediterranean diet, which have more recently been specifically associated with cognitive function, e.g., dark green leafy vegetables (Morris et al., 2015a; Morris et al., 2015b), nuts (Barbour, Howe, Buckley, Bryan, & Coates, 2014; O'Brien et al., 2014) and EVOO (Valls-Pedret et al., 2015).

To be able to derive a MEDAS score, the 14 MEDAS questions optimised for the English language (or their substance) were included within the MediCul tool. Guidelines were developed to convert participant responses, if necessary, to the relevant frequency for MEDAS scoring. This was considered valuable due to the importance in research of the Prevencion con Dieta Mediterranea (PREDIMED) randomised controlled trial, which also includes cognitive outcomes (Valls-Pedret et al., 2015), and the recent popularity of the MEDAS tool, also in studies outside of Spain, such as the AUSMED Heart Trial (Mayr et al., 2018; Mayr, Tierney, Kucianski, Thomas, & Itsiopoulos, 2019).

219 The sequence of questions within MediCul was determined following common sense principles such as, assessing the total intake from a food group prior to assessing individual elements of that food group and grouping related questions together.

4.2.4 Tool piloting

A 40-item draft MediCul tool was pilot-tested (Passmore et al., 2002) among five healthy consumers from the general public and three participants from an existing cohort of older individuals with MCI (Singh et al., 2014) for readability, comprehension and timing. The Flesch Reading Ease and Flesch-Kincaid Grade Level were computed with Microsoft Word and employed to assess readability and comprehension as these are well established methods and have been previously used to assess consent forms (Priestley, Campbell, Valentine, Denison, & Buller, 1992) and patient information leaflets (Williamson & Martin, 2010). These indexes check the number of syllables per word (a measure of word difficulty) and sentence length (indicates syntactic complexity). Responses were sought to the following questions: length of time for completion of MediCul, any unclear questions, recommendations for tool improvement and perceived benefits of having a dietitian present to clarify questions. Feedback on MediCul was also obtained from several scientific experts on the Mediterranean diet before the MediCul tool was refined and finalised.

The revision process resulted in evolutionary changes to MediCul instructions and some questions to improve their clarity and sensitivity. Additional questions were added (increasing the items from 40 to 50) to ensure the ‘traditional’ Mediterranean dietary pattern and aspects of cuisine could be adequately assessed. For example, a separate question was introduced for onion/garlic exposure, which is lacking in existing tools, yet the allium family of vegetables is consumed almost daily in the Mediterranean in cooked and/or raw forms. Images were added for less well-known foods such as dark green leafy vegetables, legumes, nuts/seeds and standard alcoholic drink quantities. See Chapter Six, Supplementary Material 6.2, for the final 50-item MediCul tool.

220 4.2.5 Tool cut-off points

Absolute cut-off points for scoring were determined based on ‘traditional’ intakes. For some questions, multiple cut-off points were allocated to provide greater discernment, e.g., intake of nut serves. Maximum points were awarded for the highest level of adherence compared to ‘traditional’ intakes or exposures, and minimum points to the lowest.

Cut-off points were informed by considering a combination of information sources: the traditional Cretan diet described in the Seven Countries Study (SCS), which gave rise to the ‘Mediterranean diet’ terminology and is considered the archetypal Mediterranean diet (Hatzis et al., 2013; Kromhout et al., 1989); popular Mediterranean diet index tools, such as MEDAS, which was itself informed by dose-response relationships between each food item and risk of myocardial infarction (Martinez-Gonzalez, Fernandez-Jarne, Serrano- Martinez, Wright, & Gomez-Gracia, 2004; Schroder et al., 2011); Mediterranean diet pyramids such as the original pyramid by Willett et al. (1995), Oldways pyramid (Oldways, 2008), Spanish Mediterranean Diet Foundation pyramid (Bach-Faig et al., 2011) and National Health and Medical Research Council Australian dietary guidelines and alcohol guidelines (National Health and Medical Research Council, 2001, 2013) for limits on alcohol. Although considered, the current dietary guidelines from the Greek Ministry of Health and Welfare (Ministry of Health and Welfare: Supreme Scientific Health Council, 1999) were not used as a basis for cut-off points, except for legumes, as these do not all reflect ‘traditional’ intakes. For example, these guidelines recommend red/white meat products 4-5 times per week, whereas ‘traditional’ intakes of such animal products were considerably lower. See Chapter Six, Supplementary Material 6.1, for the cut-off points used for each MediCul question.

4.2.6 Tool weighting

The weighting of points for various elements was based on the significance of the component to the ‘traditional’ Mediterranean cuisine in terms of quantity and frequency of intake and guided by weighting used in existing tools, particularly MEDAS, which has

221 been associated with health outcomes including cognitive function (Valls-Pedret et al., 2015). While some tools provide equal weighting for each element, assuming all contribute equally to outcomes, this approach was avoided as it has been identified as a potential weakness which may affect a tools accuracy and predictive ability for future health outcomes (Panagiotakos et al., 2009).

With regards to weighting for specific food groups, the MEDAS tool awards the highest weighting to foods with high fat content, i.e., EVOO and nuts, followed by a lower but equal weighting to vegetables, animal protein and discretionary foods, with the lowest weighting given to all other foods. This hierarchy was also used when weighting the scoring for MediCul food groups. Overall, MediCul awards the highest weighting to healthy Mediterranean foods. The remainder of MediCul points are distributed to unhealthy Western foods followed by cuisine elements, which have been less studied in relation to health outcomes.

4.2.7 Tool scoring

MediCul was designed to be scored in three steps:

1. convert raw responses for all questions to the relevant frequency required for scoring. See Table 4.1 for conversions; 2. score each item based on its cut-off point/s; 3. sum the scores for all items to derive a total MediCul score.

The scoring process can be done manually or automated for convenience using software such as Microsoft Excel. See Chapter Six, Supplementary Material 6.1, for the maximum score possible for each MediCul question.

222 Table 4.1. Conversions prior to scoring

Convert from Convert to Calculation

Serves or frequencies per Serves or frequencies per day Divide by 30 month

Serves or frequencies per Serves or frequencies per week Divide by 30 then month multiply by 7

Serves or frequencies per week Serves or frequencies per day Divide by 7

Serves or frequencies per day Serves or frequencies per week Multiply by 7

Days per month Days per year Divide by 30 then multiply by 365

4.2.8 Tool adaptation for online use

In order for MediCul to be administered online (e.g., offered to participants via the SurveyMonkey platform as used in Chapter Five) the following adaptations were made:

1. overall instructions were slightly modified to reflect online use, e.g., Next/Previous buttons, use of drop down menus, etc.; 2. the number of questions per page was limited for ease of viewing on screen and to minimise required scrolling; 3. drop down menus were added to offer frequency/quantity options; 4. some questions were slightly re-worded to better fit the response options in drop down menus.

See Appendix Seven for screen shots of sample MediCul questions formatted for online administration.

223 4.2.9 Scoring for MEDAS

The MEDAS score, used in PREDIMED, can be indirectly derived from responses provided to MediCul using a similar 3-step process for tool scoring discussed above (see Chapter Five, Supplementary Material Table S1).

4.3 Results

A new 50-item Mediterranean diet index tool named MediCul was developed a priori to capture the ‘traditional’ dietary pattern and certain aspects of cuisine. It measures some unique elements, not assessed by previous Mediterranean diet index tools. For example, the use of a home garden to grow vegetables, dark green leafy vegetable consumption, and the intake of olives and other traditional fermented foods. It also assesses exposure to discretionary foods, enabling it to be suitable for use within Western populations where such foods contribute significantly to energy intakes.

MediCul includes 17 main food groups/elements: olive oil, vegetables, fruit, nuts, wholegrains, legumes, fish/shellfish, eggs, dairy products, white meat, red/processed meats, sweets and sugary drinks, takeaway, water, alcohol, coffee, cuisine. Nine of these elements cover desirable features of a ‘traditional’ Mediterranean diet and aspects of cuisine, and six cover undesirable elements of a Western diet.

MediCul is scored out of 100, with a higher score representing increased adherence to the ‘traditional’ dietary pattern and aspects of cuisine. However, it is presently difficult to categorise the level of adherence, e.g., low, moderate, high, without further testing the tool in association with health outcomes. This is beyond the scope of this thesis, but is planned for future research.

MediCul takes approximately 20 minutes to complete and is relatively simple to use. It has good readability as confirmed by the Flesch Reading Ease = 76.4 and Flesch-Kincaid Grade Level = 5.4, where text with scores of at least 60-70, and at least 7-8, respectively,

224 is considered to be well written. MediCul is designed to be suitable for online and in- person administration.

4.4 Discussion

It is commonly assumed that all Mediterranean diet index tools assess adherence to the ‘traditional’ diet. This is a misconception since such tools, with both elements and absolute cut-off points determined a priori and based on traditional intakes have been reported far less in the literature (see Appendix Five).

MediCul is a newly developed diet index tool that has avoided the identified risks of moving towards a modernised Mediterranean model, far away from the ‘traditional’ one originally described in the SCS (Ciancarelli, Di Massimo, De Amicis, & Ciancarelli, 2017). During the development of MediCul some previous criticisms of Mediterranean diet research were also addressed as MediCul gathers information on cooking styles and certain food combinations, as well as various cuisine practices. Therefore, MediCul may assist in future research by enabling greater discernment between the ‘traditional’ Mediterranean dietary pattern and other plant-based dietary patterns. However, the strengths and limitations of MediCul should be considered.

4.4.1 Strengths

4.4.1.1 Original data sources used for cut-off points

MediCul uses cut-off points based on ‘traditional’ intakes (circa 1950s to early 1960s in Crete, Greece, and Southern Italy), which have been associated with reduced risks of chronic diseases now prevalent in Western society, including dementia (Hatzis, Sifaki- Pistolla, & Kafatos, 2015).

225 4.4.1.2 Novel Mediterranean elements assessed

Aspects of ‘traditional’ cuisine or habits related to eating, not captured by other index tools, are measured by MediCul. These include use of lemon/vinegar in food preparation; frequent home cooked main meals; exposure to moist, lower temperature cooking methods; eating main meals in company; fasting practice and napping after the midday meal.

4.4.1.3 Exposure to Western foods assessed

MediCul includes a range of questions that assess exposures to commonly consumed Western foods that have been associated with health risks, e.g., sugary drinks and takeaways. Although such foods were not part of the ‘traditional’ Mediterranean diet, Westernisation of dietary habits is occurring globally, including within Mediterranean countries (D'Alessandro & De Pergola, 2018).

4.4.1.4 Foods associated with cognition assessed

MediCul includes elements that have been related to cognitive function, e.g., dark green leafy vegetables (Mewborn, Terry, Renzi-Hammond, Hammond, & Miller, 2017), water (Lieberman, 2007), coffee (Santos, Costa, Santos, Vaz-Carneiro, & Lunet, 2010) and fasting practice (Brandhorst et al., 2015). Some of these elements are not measured by other tools.

4.4.1.5 Zero alcohol intake not penalised

Unlike most existing Mediterranean diet index tools, MediCul does not penalise participants if they avoid alcohol, for two of its alcohol questions. This is based on the following rationale: a) not all women consumed alcohol in the ‘traditional diet’, b) even small to moderate, but regular amounts of alcohol are now appreciated to increase cancer risk (Winstanley et al., 2011) with previously reported health benefits now believed to have been overstated due to confounding (Knott, Coombs, Stamatakis, & Biddulph, 2015) and, c) alcohol contributes to adverse brain morphology and function outcomes, particularly at high levels (Guest, Grant, Mori, & Croft, 2014; Topiwala et al., 2017;

226 Verbaten, 2009). While some observational studies show an association for low exposures to alcohol and reduced risk of cognitive decline and dementia, especially in older adults, systematic reviews have been criticised for not categorising former drinkers separately from lifetime abstainers in their analyses (Ilomaki, Jokanovic, Tan, & Lonnroos, 2015). Hence, MediCul is similar to HEIFA-2013 (Roy et al., 2016), a non- Mediterranean index tool based on the updated Australian Dietary Guidelines (National Health and Medical Research Council, 2013), which awards maximum points for no or low alcohol intakes.

4.4.1.6 Practical serves and images used

For foods where quantity is assessed, MediCul uses practical serve sizes informed partly by MEDAS serves, as well as the commonly consumed portions in Australia (Zheng et al., 2016) and the Australian Guide to Healthy Eating (AGTHE), as appropriate. The tool also includes images for some less well-known foods in Western populations.

4.4.1.7 More comprehensive than a screener

While short screener tools may be useful to elicit behaviour change within an intervention, MediCul provides a comprehensive assessment of adherence to the ‘traditional’ dietary pattern and aspects of cuisine with less participant burden than an extensive FFQ.

4.4.1.8 Complex computations not required

A MediCul score can be derived without further statistical analyses. It can be determined manually or calculation can be automated using software, such as Microsoft Excel.

4.4.1.9 MEDAS score can be derived

Responses to MediCul questions allow for indirect derivation of a MEDAS score, which is ideal for comparison to other studies that have used, or may use this screener in future.

227 4.4.2 Limitations

4.4.2.1 Limited dietary assessment period

Inquiring about the last six months could be a limitation of MediCul, as this period doesn’t cover the impact of all seasons. Also, it is in contrast to many FFQs which typically ask about periods of 12 months. However, the time period of dietary assessment for MediCul could be adjusted to suit research needs. Also, the shorter time frame of dietary assessment selected for MediCul was deliberate as it may be difficult to assess diet over extended periods, especially among individuals with MCI. It is known that recall errors increase with time (Livingstone, Robson, & Wallace, 2004), increasing the chance of recall bias. This possibility was flagged by comments made when piloting the tool, where some individuals with MCI reported they could only consider the last month when answering questions about their diet. MediCul does, however, ask about cooking methods related to both warmer and cooler weather, as these can vary substantially.

4.4.2.2 Healthiness of ‘traditional’ diet assumed

The SCS gave rise to the Mediterranean diet terminology, as the ‘traditional’ Cretan diet consumed during the time of the study has been associated with the lowest coronary heart disease, cancer and dementia mortality, compared to the diets from 15 other cohorts (Hatzis et al., 2013; Hatzis et al., 2015). While the Mediterranean diet is the most extensively studied dietary pattern to date, with similar high quality evidence not available for any other dietary patterns (Martinez-Gonzalez et al., 2017), it is well-known that changes have occurred to what is now consumed in the Mediterranean. For example, there has been an increase in consumption of animal products and ultra processed foods. Since such foods have been associated with chronic disease, it was decided that MediCul should be based on the ‘traditional’ cuisine (see Chapter Two), which may represent a healthier time point in consumption of the Mediterranean diet.

228 4.4.2.3 Combined quantity and frequency questions used

A few questions within MediCul include the simultaneous assessment of both quantity and frequency variables, e.g., serves of vegetables consumed. Such questions tend to be more difficult to answer and may reduce accuracy (Passmore et al., 2002). However, these questions were based on standard questions within use in Australian dietary surveys, so this was unavoidable.

4.4.2.4 Less common Western foods not assessed

MediCul does not assess foods that are used less commonly within the general Western population, but which may contribute significantly to the diets of individuals. For example, there is no response category for coconut oil, coconut yoghurt, probiotic shots, margarine containing olive oil, dairy blend or paleo bread. However, no dietary assessment tool can measure everything, which is why study purpose is an important consideration in tool selection.

4.4.2.5 Food preparation methods uncertain if not primary cook

Food preparation methods may be unknown by respondents who are not the primary cook within their household. However, this is a limitation across all tools where it is not compulsory for the partner/carer to be present during an assessment.

4.4.2.6 Food avoidance due to intolerance/allergy affects scoring

Some individuals may have a food intolerance/allergy to particular foods/nutrients, which could influence their MediCul score. For example, a nut allergy or fructan intolerance. However, this is unavoidable, if the intake of certain ‘traditional’ foods is to be assessed.

4.4.2.7 Sex and energy-adjusted cut-off points unavailable

Due to the paucity of data for traditional intakes according to gender and energy intake level, MediCul uses non-gender specific cut-off points, as does MEDAS and the various Mediterranean diet pyramids.

229 4.4.2.8 Not designed to calculate nutrient intakes

Few diet index tools are used to estimate the intake of nutrients or phytochemicals (Hernández Ruiz et al., 2015) and MediCul is no exception. For example, while MediCul includes a short question to assess intake of total coffee cups, this cannot be used to accurately calculate caffeine intake since coffee cup sizes vary. Also, the type of coffee beans used and method of preparation can impact the caffeine content of coffee. Hence, multiple questions are required to accurately assess caffeine intake. This is beyond the scope of MediCul and best done using specialised tools, such as the caffeine food frequency questionnaire (C-FFQ) recently developed for the Australian population (Watson, Kohler, Banks, & Coates, 2017).

4.4.2.9 Timing of food intake not assessed

Traditionally, most eating within the Mediterranean occurred during the day with the main meal being a cooked lunch followed by a light supper (Hoffman & Gerber, 2013). Emerging research suggests late night eating, common in Western countries, disrupts the circadian rhythm and is associated with multiple chronic diseases (Grant et al., 2017; St- Onge et al., 2017). While the timing of food intake over a 24-hour period may also be an important consideration for health, this is not assessed by MediCul due to the complexity of chrononutrition, which also considers meal size/frequency and timing.

Finally, although MediCul provides the opportunity to comprehensively assess adherence to a ‘traditional’ Mediterranean dietary pattern and aspects of cuisine, it should be noted that previous studies using various Mediterranean diet index tools, which did not include measures of cuisine, nor Mediterranean specific cut-off points, have still found benefits for a Mediterranean-style dietary pattern. These include a reduced risk of mortality and chronic disease (Kouris-Blazos et al., 1999; Trichopoulou et al., 1995). However, higher scores from some of these tools, particularly when applied to Western populations, may simply be reflecting increased adherence to a prudent plant-based dietary pattern, rather than being discerning for the ‘traditional’ Mediterranean diet.

230 4.5 Conclusion

A new 50-item diet index tool, MediCul, has been developed to assess adherence to the ‘traditional’ Mediterranean dietary pattern and aspects of cuisine. MediCul is intended for online and in-person use within a Western context, including among adults with risk factors for cognitive decline and those diagnosed with MCI. Validation of the MediCul tool against a dietary reference method, within two such cohorts, is reported in Chapters Five and Six.

231 4.6 References

Anastasiou, C., Yannakoulia, M., Kontogianni, M., Kosmidis, M., Mamalaki, E., Dardiotis, E., . . . Scarmeas, N. (2018). Mediterranean lifestyle in relation to cognitive health: Results from the HELIAD study. Nutrients, 10(10), 1557. doi:10.3390/nu10101557 Australian Bureau of Statistics. (2014). 4364.0.55.007 - Australian Health Survey: Nutrition first results - Foods and nutrients, 2011-12. Canberra, Australia: FSANZ. Bach-Faig, A., Berry, E. M., Lairon, D., Reguant, J., Trichopoulou, A., Dernini, S., . . . Mediterranean Diet Foundation Expert Group. (2011). Mediterranean diet pyramid today: Science and cultural updates. Public Health Nutrition, 14(12A), 2274-2284. doi:10.1017/S1368980011002515 Bach, A., Serra-Majem, L., Carrasco, J. L., Roman, B., Ngo, J., Bertomeu, I., & Obrador, B. (2006). The use of indexes evaluating the adherence to the Mediterranean diet in epidemiological studies: A review. Public Health Nutrition, 9(1a), 132-146. doi:10.1079/PHN2005936 Baghurst, K. (1999). Red meat consumption in Australia: Intakes, contributions to nutrient intake and associated dietary patterns. European Journal of Cancer Prevention, 8(3), 185-191. doi:10.1097/00008469-199906000-00005 Barbour, J. A., Howe, P. R. C., Buckley, J. D., Bryan, J., & Coates, A. M. (2014). Nut consumption for vascular health and cognitive function. Nutrition Research Reviews, 27(1), 131-158. doi:10.1017/S0954422414000079 Benitez-Arciniega, A. A., Mendez, M. A., Baena-Diez, J. M., Rovira Martori, M. A., Soler, C., Marrugat, J., . . . Schroder, H. (2011). Concurrent and construct validity of Mediterranean diet scores as assessed by an FFQ. Public Health Nutrition, 14(11), 2015-2021. doi:10.1017/S1368980011001212 Bonaccio, M., Di Castelnuovo, A., Costanzo, S., Gialluisi, A., Persichillo, M., Cerletti, C., . . . Iacoviello, L. (2018). Mediterranean diet and mortality in the elderly: A prospective cohort study and a meta-analysis. The British Journal of Nutrition, 120(8), 841-854. doi:10.1017/S0007114518002179

232 Brandhorst, S., Choi, In Y., Wei, M., Cheng, Chia W., Sedrakyan, S., Navarrete, G., . . . Longo, V. (2015). A periodic diet that mimics fasting promotes multi-system regeneration, enhanced cognitive performance, and healthspan. Cell Metabolism, 22(1), 86-99. doi:10.1016/j.cmet.2015.05.012 Cade, J., Thompson, R., Burley, V., & Warm, D. (2002). Development, validation and utilisation of food-frequency questionnaires – A review. Public Health Nutrition, 5(4), 567-587. doi:10.1079/PHN2001318 Calfas, K. J., Zabinski, M. F., & Rupp, J. (2000). Practical nutrition assessment in primary care settings. American Journal of Preventive Medicine, 18(4), 289-299. doi:10.1016/S0749-3797(00)00116-1 Cerwinske, L. A., Rasmussen, H. E., Lipson, S., Volgman, A. S., & Tangney, C. C. (2017). Evaluation of a dietary screener: The Mediterranean Eating Pattern for Americans tool. Journal of Human Nutrition and Dietetics, 30(5), 596-603. doi:10.1111/jhn.12451 Ciancarelli, M. G. T., Di Massimo, C., De Amicis, D., & Ciancarelli, I. (2017). Mediterranean diet and health promotion: Evidence and current concerns. Medical Research Archives. Retrieved 23rd December 2018 from https://journals.ke-i.org/index.php/mra/article/view/1385 Ciccarone, E., Di Castelnuovo, A., Salcuni, M., Siani, A., Giacco, A., Donati, M. B., . . . Gendiabe, I. (2003). A high-score Mediterranean dietary pattern is associated with a reduced risk of peripheral arterial disease in Italian patients with Type 2 diabetes. Journal of Thrombosis and Haemostasis, 1(8), 1744-1752. doi:10.1046/j.1538-7836.2003.00323.x Collins, C. E., Watson, J., & Burrows, T. (2010). Measuring dietary intake in children and adolescents in the context of overweight and obesity. International Journal of Obesity, 34(7), 1103-1115. doi:10.1038/ijo.2009.241 Collins, C. E., Boggess, M. M., Watson, J. F., Guest, M., Duncanson, K., Pezdirc, K., . . . Burrows, T. L. (2014). Reproducibility and comparative validity of a food frequency questionnaire for Australian adults. Clinical Nutrition, 33(5), 906- 914. doi:10.1016/j.clnu.2013.09.015

233 Collins, C. E., Burrows, T. L., Rollo, M. E., Boggess, M. M., Watson, J. F., Guest, M., . . . Hutchesson, M., J. (2015). The comparative validity and reproducibility of a diet quality index for adults: The Australian Recommended Food Score. Nutrients, 7(2), 785-798. doi:10.3390/nu7020785 D'Alessandro, A., & De Pergola, G. (2015). Mediterranean diet and cardiovascular disease: A critical evaluation of a priori dietary indexes. Nutrients, 7(9), 7863- 7888. doi:10.3390/nu7095367 D'Alessandro, A., & De Pergola, G. (2018). The Mediterranean Diet: Its definition and evaluation of a priori dietary indexes in primary cardiovascular prevention. International Journal of Food Sciences and Nutrition, 69(6), 647-659. doi:10.1080/09637486.2017.1417978 De Filippis, F., Pellegrini, N., Vannini, L., Jeffery, I. B., La Storia, A., Laghi, L., . . . Ercolini, D. (2016). High-level adherence to a Mediterranean diet beneficially impacts the gut microbiota and associated metabolome. Gut, 65(11), 1812-1821. doi:10.1136/gutjnl-2015-309957 Estruch, R., Martínez-González, M. A., Corella, D., Salas-Salvadó, J., Fitó, M., Chiva- Blanch, G., . . . Ros, E. (2016). Effect of a high-fat Mediterranean diet on bodyweight and waist circumference: A prespecified secondary outcomes analysis of the PREDIMED randomised controlled trial. The Lancet Diabetes & Endocrinology, 4(8), 666-676. doi:10.1016/S2213-8587(16)30085-7 Flood, V. M., Webb, K. L., Rochtchina, E., Kelly, B., & Mitchell, P. (2007). Fatty acid intakes and food sources in a population of older Australians. Asia Pacific Journal of Clinical Nutrition, 16(2), 322-330. Gerber, M. J., Scali, J. D., Michaud, A., Durand, M. D., Astre, C. M., Dallongeville, J., & Romon, M. M. (2000). Profiles of a healthful diet and its relationship to biomarkers in a population sample from Mediterranean southern France. Journal of the American Dietetic Association, 100(10), 1164-1171. doi:10.1016/S0002- 8223(00)00340-0 Golley, R. K., Smithers, L. G., Mittinty, M. N., Brazionis, L., Emmett, P., Northstone, K., . . . Lynch, J. W. (2012). An index measuring adherence to complementary feeding guidelines has convergent validity as a measure of infant diet quality. Journal of Nutrition, 142(5), 901-908. doi:10.3945/jn.111.154971

234 Grant, C. L., Coates, A. M., Dorrian, J., Kennaway, D. J., Wittert, G. A., Heilbronn, L. K., . . . Banks, S. (2017). Timing of food intake during simulated night shift impacts glucose metabolism: A controlled study. Chronobiology International, 34(8), 1003-1013. doi:10.1080/07420528.2017.1335318 Guest, J., Grant, R., Mori, T. A., & Croft, K. D. (2014). Changes in oxidative damage, inflammation and [NAD(H)] with age in cerebrospinal fluid. PLoS ONE, 9(1). doi:10.1371/journal.pone.0085335 Hatzis, C. M., Papandreou, C., Patelarou, E., Vardavas, C. I., Kimioni, E., Sifaki- Pistolla, D., . . . Kafatos, A. G. (2013). A 50-year follow-up of the Seven Countries Study: Prevalence of cardiovascular risk factors, food and nutrient intakes among Cretans. Hormones, 12(3), 379-385. doi:10.1007/BF03401303 Hatzis, C. M., Sifaki-Pistolla, D., & Kafatos, A. G. (2015). History of the Cretan cohort of the Seven Countries Study. Hormones. Retrieved 21st December 2018 from http://www.hormones.gr/5/search.html?s=a+50-year+follow- up+of+the+Seven+countries+study&flds=all&do=search&rpp=20&x=0&y=0 Hebert, J. R., Clemow, L., Pbert, L., Ockene, I. S., & Ockene, J. K. (1995). Social desirability bias in dietary self-report may compromise the validity of dietary intake measures. International Journal of Epidemiology, 24(2), 389-398. doi:10.1093/ije/24.2.389 Hernández Ruiz, A., García-Villanova, B., Guerra Hernández, E. J., Amiano, P., Azpiri, M., & Molina Montes, E. (2015). Description of indexes based on the adherence to the Mediterranean dietary pattern: A review. Nutricion Hospitalaria, 32(5), 1872-1884. doi:10.3305/nh.2015.32.5.9629 Hillesund, E. R., Bere, E., Haugen, M., & Øverby, N. C. (2014). Development of a New Nordic Diet score and its association with gestational weight gain and fetal growth - A study performed in the Norwegian Mother and Child Cohort Study (MoBa). Public Health Nutrition, 17(9), 1909-1918. doi:10.1017/S1368980014000421 Hoffman, R., & Gerber, M. (2013). Evaluating and adapting the Mediterranean diet for non‐Mediterranean populations: A critical appraisal. Nutrition Reviews, 71(9), 573-584. doi:10.1111/nure.12040

235 Ilomaki, J., Jokanovic, N., Tan, E. C. K., & Lonnroos, E. (2015). Alcohol consumption, dementia and cognitive decline: An overview of systematic reviews. Current Clinical Pharmacology, 10(3), 204-212. doi:10.2174/157488471003150820145539 Kennedy, E. T., Ohls, J., Carlson, S., & Fleming, K. (1995). The Healthy Eating Index: Design and applications. Journal of the American Dietetic Association, 95(10), 1103-1108. doi:10.1016/S0002-8223(95)00300-2 Knott, C. S., Coombs, N., Stamatakis, E., & Biddulph, J. P. (2015). All cause mortality and the case for age specific alcohol consumption guidelines: Pooled analyses of up to 10 population based cohorts. BMJ (Clinical Research Edition), 350, h384. doi:10.1136/bmj.h384 Kouris-Blazos, A., Gnardellis, C., Wahlqvist, M. L., Trichopoulos, D., Lukito, W., & Trichopoulou, A. (1999). Are the advantages of the Mediterranean diet transferable to other populations? A cohort study in Melbourne, Australia. British Journal of Nutrition, 82(01), 57-61. doi:10.1017/S0007114599001129 Kouris-Blazos, A., & Itsiopoulos, C. (2014). Low all-cause mortality despite high cardiovascular risk in elderly Greek-born Australians: Attenuating potential of diet? Asia Pacific Journal of Clinical Nutrition, 23(4), 532-544. doi:10.6133/apjcn.2014.23.4.16 Kromhout, D., Keys, A., Aravanis, C., Buzina, R., Fidanza, F., Giampaoli, S., . . . Pekkarinen, M. (1989). Food consumption patterns in the 1960s in seven countries. American Journal of Clinical Nutrition, 49(5), 889-894. doi:10.1093/ajcn/49.5.889 Kwan, M. W.-M., Wong, M. C.-S., Wang, H. H.-X., Liu, K. Q.-L., Lee, C. L.-S., Yan, B. P.-Y., . . . Griffiths, S. M. (2013). Compliance with the Dietary Approaches to Stop Hypertension (DASH) diet: A systematic review. PLoS ONE, 8(10), e78412. doi:10.1371/journal.pone.0078412 Lieberman, H. R. (2007). Hydration and cognition: A critical review and recommendations for future research. Journal of the American College of Nutrition, 26(Supplement 5), 555S-561S. doi:10.1080/07315724.2007.10719658

236 Livingstone, M. B. E., Robson, P. J., & Wallace, J. M. W. (2004). Issues in dietary intake assessment of children and adolescents. British Journal of Nutrition, 92(Suppl2), S213-S222. doi:10.1079/BJN20041169 Mariscal-Arcas, M., Romaguera, D., Rivas, A., Feriche, B., Pons, A., Tur, J. A., & Olea-Serrano, F. (2007). Diet quality of young people in southern Spain evaluated by a Mediterranean adaptation of the Diet Quality Index-International (DQI-I). British Journal of Nutrition, 98(6), 1267-1273. doi:10.1017/S0007114507781424 Mariscal-Arcas, M., Rivas, A., Monteagudo, C., Granada, A., Cerrillo, I., & Olea- Serrano, F. (2009). Proposal of a Mediterranean diet index for pregnant women. British Journal of Nutrition, 102(5), 744-749. doi:10.1017/S0007114509274769 Martinez-Gonzalez, M. A., Fernandez-Jarne, E., Serrano-Martinez, M., Wright, M., & Gomez-Gracia, E. (2004). Development of a short dietary intake questionnaire for the quantitative estimation of adherence to a cardioprotective Mediterranean diet. European Journal of Clinical Nutrition, 58(11), 1550-1552. doi:10.1038/sj.ejcn.1602004 Martinez-Gonzalez, M. A., Hershey, M. S., Zazpe, I., & Trichopoulou, A. (2017). Transferability of the Mediterranean diet to non-Mediterranean countries: What is and what is not the Mediterranean diet. Nutrients, 9(11), 1226. doi:10.3390/nu9111226 Mayr, H. L., Thomas, C. J., Tierney, A. C., Kucianski, T., George, E. S., Ruiz-Canela, M., . . . Itsiopoulos, C. (2018). Randomization to 6-month Mediterranean diet compared with a low-fat diet leads to improvement in Dietary Inflammatory Index scores in patients with coronary heart disease: The AUSMED Heart Trial. Nutrition Research, 55, 94-107. doi:10.1016/j.nutres.2018.04.006 Mayr, H. L., Tierney, A. C., Kucianski, T., Thomas, C. J., & Itsiopoulos, C. (2019). Australian patients with coronary heart disease achieve high adherence to 6- month Mediterranean diet intervention: Preliminary results of the AUSMED Heart Trial. Nutrition, 61, 21-31. doi:10.1016/j.nut.2018.10.027

237 McCullough, M. L., Feskanich, D., Stampfer, M. J., Giovannucci, E. L., Rimm, E. B., Hu, F. B., . . . Willett, W. C. (2002). Diet quality and major chronic disease risk in men and women: Moving toward improved dietary guidance. American Journal of Clinical Nutrition, 76(6), 1261-1271. doi:10.1093/ajcn/76.6.1261 McNaughton, S. A., Ball, K., Crawford, D., & Mishra, G. D. (2008). An index of diet and eating patterns is a valid measure of diet quality in an Australian population. The Journal of Nutrition, 138(1), 86-93. doi:10.1093/jn/138.1.86 Mewborn, C. M., Terry, D. P., Renzi-Hammond, L. M., Hammond, B. R., & Miller, L. S. (2017). Relation of retinal and serum lutein and zeaxanthin to white matter integrity in older adults: A diffusion tensor imaging study. Archives of Clinical Neuropsychology, 33(7), 1-14. doi:10.1093/acn/acx109 Milà-Villarroel, R., Bach-Faig, A., Puig, J., Puchal, A., Farran, A., Serra-Majem, L., & Carrasco, J. L. (2011). Comparison and evaluation of the reliability of indexes of adherence to the Mediterranean diet. Public Health Nutrition, 14(12A), 2338- 2345. doi:10.1017/S1368980011002606 Miles, S., & Scaife, V. (2003). Optimistic bias and food. Nutrition Research Reviews, 16(1), 3-19. doi:10.1079/NRR200249 Ministry of Health and Welfare: Supreme Scientific Health Council. (1999). Dietary guidelines for adults in Greece. Archives of Hellenic Medicine. Retrieved 19th December 2018 from http://www.mednet.gr/archives/1999-5/pdf/516.pdf Monteagudo, C., Palacín-Arce, A., Bibiloni, M. D. M., Pons, A., Tur, J. A., Olea- Serrano, F., & Mariscal-Arcas, M. (2013). Proposal for a Breakfast Quality Index (BQI) for children and adolescents. Public Health Nutrition, 16(4), 639- 644. doi:10.1017/S1368980012003175 Morris, M. C., Tangney, C. C., Wang, Y., Sacks, F. M., Barnes, L. L., Bennett, D. A., & Aggarwal, N. T. (2015a). MIND diet slows cognitive decline with aging. Alzheimer's & Dementia, 11(9), 1015-1022. doi:10.1016/j.jalz.2015.04.011 Morris, M. C., Tangney, C. C., Wang, Y., Sacks, F. M., Bennett, D. A., & Aggarwal, N. T. (2015b). MIND diet associated with reduced incidence of Alzheimer's disease. Alzheimer's & Dementia, 11(9), 1007-1014. doi:10.1016/j.jalz.2014.11.009

238 National Health and Medical Research Council. (2001). Australian alcohol guidelines: Health risks and benefits. Canberra, Australia: Commonwealth of Australia. National Health and Medical Research Council. (2013). Australian dietary guidelines. Canberra, Australia: National Health and Medical Research Council. O'Brien, J., Okereke, O., Devore, E., Rosner, B., Breteler, M., & Grodstein, F. (2014). Long-term intake of nuts in relation to cognitive function in older women. Journal of Nutrition Health & Aging, 18(5), 496-502. doi:10.1007/s12603-014- 0014-6 Oldways. (2008). Mediterranean diet pyramid. Retrieved 18th May 2018 from http://oldwayspt.org/resources/heritage-pyramids/mediterranean- pyramid/overview Panagiotakos, D., Kalogeropoulos, N., Pitsavos, C., Roussinou, G., Palliou, K., Chrysohoou, C., & Stefanadis, C. (2009). Validation of the MedDietScore via the determination of plasma fatty acids. International Journal of Food Sciences and Nutrition, 60(Suppl5), 168-180. doi:10.1080/09637480902810338 Panagiotakos, D. B., Milias, G. A., Pitsavos, C., & Stefanadis, C. (2006). MedDietScore: A computer program that evaluates the adherence to the Mediterranean dietary pattern and its relation to cardiovascular disease risk. Computer Methods and Programs in Biomedicine, 83(1), 73-77. doi:10.1016/j.cmpb.2006.05.003 Panagiotakos, D. B., Pitsavos, C., & Stefanadis, C. (2006). Dietary patterns: A Mediterranean diet score and its relation to clinical and biological markers of cardiovascular disease risk. Nutrition, Metabolism and Cardiovascular Diseases, 16(8), 559-568. doi:10.1016/j.numecd.2005.08.006 Passmore, C., Dobbie, A. E., Parchman, M., & Tysinger, J. (2002). Guidelines for constructing a survey. Family Medicine. Retrieved 23rd December 2018 from https://www.stfm.org/FamilyMedicine/Vol34Issue4/Passmore281 Patterson, R. E., Haines, P. S., & Popkin, B. M. (1994). Diet quality index: Capturing a multidimensional behavior. Journal of the American Dietetic Association, 94(1), 57-64. doi:10.1016/0002-8223(94)92042-7 Petersen, R. C. (2004). Mild cognitive impairment as a diagnostic entity. Journal of Internal Medicine, 256(3), 183-194. doi:10.1111/j.1365-2796.2004.01388.x

239 Priestley, K. A., Campbell, C., Valentine, C. B., Denison, D. M., & Buller, N. P. (1992). Are patient consent forms for research protocols easy to read? BMJ (Clinical Research Edition), 305, 1263-1264. doi:10.1136/bmj.305.6864.1263 Psaltopoulou, T., Sergentanis, T. N., Panagiotakos, D. B., Sergentanis, I. N., Kosti, R., & Scarmeas, N. (2013). Mediterranean diet, stroke, cognitive impairment, and depression: A meta-analysis. Annals of Neurology, 74(4), 580-591. doi:10.1002/ana.23944 Roy, R., Hebden, L., Rangan, A., & Allman-Farinelli, M. (2016). The development, application, and validation of a Healthy Eating Index for Australian Adults (HEIFA—2013). Nutrition, 32(4), 432-440. doi:10.1016/j.nut.2015.10.006 Santos, C., Costa, J., Santos, J., Vaz-Carneiro, A., & Lunet, N. (2010). Caffeine intake and dementia: Systematic review and meta-analysis. Journal of Alzheimer's Disease, 20(1), S187-S204. doi:10.3233/JAD-2010-091387 Scarmeas, N., Stern, Y., Mayeux, R., Manly, J. J., Schupf, N., & Luchsinger, J. A. (2009). Mediterranean diet and mild cognitive impairment. Archives of Neurology, 66(2), 216-225. doi:10.1001/archneurol.2008.536 Schroder, H., Fito, M., Estruch, R., Martinez-Gonzalez, M. A., Corella, D., Salas- Salvado, J., . . . Covas, M. I. (2011). A short screener is valid for assessing Mediterranean diet adherence among older Spanish men and women. Journal of Nutrition, 141(6), 1140-1145. doi:10.3945/jn.110.135566 Serra-Majem, L., Ribas, L., Ngo, J., Ortega, R. M., García, A., Pérez-Rodrigo, C., & Aranceta, J. (2004). Food, youth and the Mediterranean diet in Spain. Development of KIDMED, Mediterranean Diet Quality Index in children and adolescents. Public Health Nutrition, 7(7), 931-935. doi:10.1079/PHN2004556 Shivappa, N., Steck, S. E., Hurley, T. G., Hussey, J. R., & Hébert, J. R. (2014). Designing and developing a literature-derived, population-based dietary inflammatory index. Public Health Nutrition, 17(8), 1689-1696. doi:10.1017/S1368980013002115

240 Singh, M. A. F., Gates, N., Saigal, N., Wilson, G. C., Meiklejohn, J., Brodaty, H., . . . Suo, C. (2014). The Study of Mental and Resistance Training (SMART) study— resistance training and/or cognitive training in mild cognitive impairment: A randomized, double-blind, double-sham controlled trial. Journal of the American Medical Directors Association, 15(12), 873-880. doi:10.1016/j.jamda.2014.09.010 Sofi, F., Macchi, C., Abbate, R., Gensini, G. F., & Casini, A. (2014). Mediterranean diet and health status: An updated meta-analysis and a proposal for a literature-based adherence score. Public Health Nutrition, 17(12), 2769-2782. doi:10.1017/S1368980013003169 Sotos-Prieto, M., Moreno-Franco, B., Ordovás, J. M., León, M., Casasnovas, J. A., & Peñalvo, J. L. (2014). Design and development of an instrument to measure overall lifestyle habits for epidemiological research: The Mediterranean lifestyle (MEDLIFE) index. Public Health Nutrition, 18(6), 959-967. doi:10.1017/S1368980014001360 Sotos-Prieto, M., Santos-Beneit, G., Bodega, P., Pocock, S., Mattei, J., & Penalvo, J. L. (2015). Validation of a questionnaire to measure overall Mediterranean lifestyle habits for research application: The MEDiterranean LIFEstyle index (MEDLIFE). Nutricion Hospitalaria, 32(3), 1153-1163. doi:10.3305/nh.2015.32.3.9387 St-Onge, M.-P., Ard, J., Baskin, M. L., Chiuve, S. E., Johnson, H. M., Kris-Etherton, P., & Varady, K. (2017). Meal timing and frequency: Implications for cardiovascular disease prevention: A scientific statement from the American Heart Association. Circulation, 135(9), e96-e121. doi:10.1161/CIR.0000000000000476 Tangney, C. C., Kwasny, M. J., Li, H., Wilson, R. S., Evans, D. A., & Morris, M. C. (2011). Adherence to a Mediterranean-type dietary pattern and cognitive decline in a community population. American Journal of Clinical Nutrition, 93(3), 601- 607. doi:10.3945/ajcn.110.007369

241 Thompson, F. E., & Subar, A. F. (2017). Chapter 1 - Dietary assessment methodology. In A. M. Coulston, C. J. Boushey, M. G. Ferruzzi, & L. M. Delahanty (Eds.), Nutrition in the prevention and treatment of disease (4th ed., pp. 5-48). London, UK: Academic Press. Topiwala, A., Allan, C. L., Valkanova, V., Zsoldos, E., Filippini, N., Sexton, C., . . . Ebmeier, K. P. (2017). Moderate alcohol consumption as risk factor for adverse brain outcomes and cognitive decline: longitudinal cohort study. BMJ (Clinical Research Edition), 357, j2353. doi:10.1136/bmj.j2353 Trichopoulou, A., Kouris-Blazos, A., Wahlqvist, M. L., Gnardellis, C., Lagiou, P., Polychronopoulos, E., . . . Trichopoulos, D. (1995). Diet and overall survival in elderly people. British Medical Journal, 311(7018), 1457-1460. doi:10.1136/bmj.311.7018.1457 Trichopoulou, A., Costacou, T., Bamia, C., & Trichopoulos, D. (2003). Adherence to a Mediterranean diet and survival in a Greek population. New England Journal of Medicine, 348(26), 2599-2608. doi:10.1056/nejmoa025039 Valls-Pedret, C., Sala-Vila, A., Serra-Mir, M., Corella, D., de la Torre, R., Martínez- González, M. Á., . . . Salas-Salvadó, J. (2015). Mediterranean diet and age- related cognitive decline: A randomized clinical trial. JAMA Internal Medicine, 175(7), 1094-1103. doi:10.1001/jamainternmed.2015.1668 Verbaten, M. N. (2009). Chronic effects of low to moderate alcohol consumption on structural and functional properties of the brain: Beneficial or not? Human Psychopharmacology, 24(3), 199-205. doi:10.1002/hup.1022 Waijers, P. M. C. M., Feskens, E. J. M., & Ocké, M. C. (2007). A critical review of predefined diet quality scores. British Journal of Nutrition, 97(2), 219-231. doi:10.1017/S0007114507250421 Ward, S. J., Coates, A. M., & Hill, A. M. (2019). Application of an Australian Dietary Guideline Index to weighed food records. Nutrients, 11, 1286. doi:10.3390/nu11061286 Watson, E. J., Kohler, M., Banks, S., & Coates, A. M. (2017). Validation and reproducibility of an Australian caffeine food frequency questionnaire. International Journal of Food Sciences and Nutrition, 68(5), 617-626. doi:10.1080/09637486.2016.1268102

242 Willett, W. C., Sacks, F., Trichopoulou, A., Drescher, G., Ferro-Luzzi, A., Helsing, E., & Trichopoulos, D. (1995). Mediterranean diet pyramid: A cultural model for healthy eating. American Journal of Clinical Nutrition, 61(6 Suppl), 1402S- 1406S. doi:10.1093/ajcn/61.6.1402S Williamson, J. M. L., & Martin, A. G. (2010). Analysis of patient information leaflets provided by a district general hospital by the Flesch and Flesch–Kincaid method. International Journal of Clinical Practice, 64(13), 1824-1831. doi:10.1111/j.1742-1241.2010.02408.x Winstanley, M. H., Pratt, I. S., Chapman, K., Griffin, H. J., Croager, E. J., Olver, I. N., . . . Slevin, T. J. (2011). Alcohol and cancer: A position statement from Cancer Council Australia. Medical Journal of Australia. Retrieved 23rd December 2018 from https://www.cancer.org.au/policy-and-advocacy/position- statements/alcohol-and-cancer/ Zaragoza-Martí, A., Cabañero-Martínez, M. J., Hurtado-Sánchez, J. A., Laguna-Pérez, A., & Ferrer-Cascales, R. (2018). Evaluation of Mediterranean diet adherence scores: A systematic review. BMJ Open, 8(2), e019033. doi:10.1136/bmjopen- 2017-019033 Zheng, M., Wu, J. H. Y., Louie, J. C. Y., Flood, V. M., Gill, T., Thomas, B., . . . Rangan, A. (2016). Typical food portion sizes consumed by Australian adults: Results from the 2011-12 Australian National Nutrition and Physical Activity Survey. Scientific Reports, 6, 19596. doi:10.1038/srep19596

243

CHAPTER FIVE

Validity of the Mediterranean Diet and Culinary Index

(MediCul) for online assessment of adherence to the

‘traditional’ diet and aspects of cuisine in older adults

244 PREFACE

In order for research tools to be used, they should first be validated. This is to ensure they are measuring what is intended, and can do so with reliability among the population of interest. While there are various types of validity of interest within research, it is now generally accepted that diet index tools should be tested for concurrent validity against a reference dietary method that is not limited by the same recall bias.

The aim of this chapter was to test the reliability and validity of the newly developed Mediterranean Diet and Culinary Index (MediCul) against a 3-day food record, when administered online to middle-aged and older Australian adults with risk factors for cognitive decline, but no current diagnosis of dementia or neurodegenerative disease, representing a Western population.

This chapter was published in the journal Nutrients:

The chapter includes two published supplementary materials:

- How to derive a MEDAS score from MediCul - Kappa statistics for percent agreement within the same category of 17 MediCul groups, between survey A and survey B

Further information related to this chapter is presented in Appendices Seven to Eleven and includes: screen shots of sample MediCul questions formatted for online

245 administration, ethics letter, Research Food Diary app instructions, dietitian assisted food record checklist, front sheet for validation and reliability testing of MediCul.

Part of the work produced in this chapter has been presented at the following conference:

• Radd-Vagenas, S., Fiatarone Singh, M. A., Daniel, K., Noble, Y., O’Leary, F., Mavros, Y., Brodaty, H., Flood, V. M. Reliability and validity of MediCul (Mediterranean Diet & Culinary Index) in older Australian adults. 42nd Nutrition Society of Australia (Inc.) Annual Scientific Meeting, Canberra, Australia, 27- 30th November 2018.

246 AUTHORSHIP STATEMENT

The co-authors of the paper: ‘Validity of the Mediterranean Diet and Culinary Index (MediCul) for online assessment of adherence to the ‘traditional’ diet and aspects of cuisine in older adults’ confirm that Sue Radd-Vagenas has made the following contributions:

As the primary supervisor for the candidature upon which this thesis is based, I can confirm that the above authorship attribution statement is true and correct.

Professor Victoria M Flood

Faculty of Health Sciences

The University of Sydney

25th February 2019

247 nutrients

Article Validity of the Mediterranean Diet and Culinary Index (MediCul) for Online Assessment of Adherence to the ‘Traditional’ Diet and Aspects of Cuisine in Older Adults

Sue Radd-Vagenas 1 , Maria A. Fiatarone Singh 1,2, Kenneth Daniel 1, Yian Noble 1, Nidhi Jain 1, Fiona O’Leary 3, Yorgi Mavros 1 , Megan Heffernan 4, Jacinda Meiklejohn 1, Yareni Guerrero 1, Tiffany Chau 4, Perminder S. Sachdev 4,5 , Henry Brodaty 4,5 and Victoria M. Flood 1,6,*

1 The University of Sydney, Physical Activity, Lifestyle, Ageing and Wellbeing Research Group, Faculty of Health Sciences, Lidcombe, NSW 2141, Australia; [email protected] (S.R.-V.); maria.fi[email protected] (M.A.F.S.); [email protected] (K.D.); [email protected] (Y.N.); [email protected] (N.J.); [email protected] (Y.M.); [email protected] (J.M.); [email protected] (Y.G.) 2 The University of Sydney, Sydney Medical School, Camperdown, NSW 2006, Australia and Hebrew SeniorLife and Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, MA 02111, USA 3 The University of Sydney, Nutrition and Dietetics Group, School of Life and Environmental Science, Faculty of Science and The Charles Perkins Centre, Camperdown, NSW 2006, Australia; fi[email protected] 4 Centre for Healthy Brain Ageing (CHeBA), School of Psychiatry, University of New South Wales, Randwick, NSW 2031, Australia; [email protected] (M.H.); [email protected] (T.C.); [email protected] (P.S.S.); [email protected] (H.B.) 5 Dementia Centre for Research Collaboration, University of New South Wales, Sydney, NSW 2052, Australia 6 Western Sydney Local Health District, Westmead Hospital, Westmead, NSW 2145, Australia * Correspondence: vicki.fl[email protected]; Tel.: +61-2-9351-9001

Received: 29 October 2018; Accepted: 28 November 2018; Published: 4 December 2018

Abstract: The Mediterranean diet is associated with multiple health benefits. Yet, no tool has been specifically developed to assess adherence to the ‘traditional’ Mediterranean diet and cuisine within a Western cohort, and validated for online use. We tested the reliability and validity of online administration of the Mediterranean Diet and Culinary Index (MediCul) among middle-aged and older adults. Participants were recruited in January–March 2017 from the 45 and Up Study, completing MediCul twice. Test-retest reliability was assessed using the paired t-test, intra-class correlation coefficient (ICC) and Bland-Altman plot. Validity was tested against a three-day food record (FR)-derived MediCul score using Bland-Altman and nutrient trends across the MediCul score tertiles. Participants (n = 84; 60% female; 65.4 years (SD = 5.9)), were overweight (BMI 26.1; SD = 4.0) with 1.7 (SD = 1.5) chronic illnesses/conditions. Sequential MediCul tool scores were 56.1/100.0 and 56.8/100.0, respectively (t = −1.019; p = 0.311). Reliability via ICC (ICC = 0.86, 95% CI: 0.789, 0.910, p < 0.0001) and Bland-Altman was good. In Bland-Altman validity analyses, the tool over-reported FR MediCul score by 5.6 points with no systematic bias ((y = 8.7 − 0.06*x) (95% CI: −0.278, 0.158, p = 0.584)). Nutrient trends were identified for MediCul consistent with expected Mediterranean patterns. Online MediCul administration demonstrated good reliability and moderate validity for assessing adherence to a ‘traditional’ Mediterranean pattern among older Australians.

Keywords: validity; reliability; repeatability; dietary assessment; index tool; score; Mediterranean diet; Mediterranean dietary pattern; traditional

Nutrients 2018, 10, 1913; doi:10.3390/nu10121913 www.mdpi.com/journal/nutrients 248 Nutrients 2018, 10, 1913 2 of 17

1. Introduction The Mediterranean diet is a focus of research interest worldwide, as it has been associated with multiple health benefits. These include a reduced risk of premature death, cardiovascular disease (CVD), metabolic syndrome, cancer, liver disease, type 2 diabetes, depression, cognitive decline with ageing, mild cognitive impairment (MCI) and dementia, in particular Alzheimer’s disease [1–11]. However, there has been inconsistency in the literature with regards to the definition used for the Mediterranean diet [12,13]. Also, variable or inadequately reported Mediterranean interventions have been used in clinical trials, which may not represent the ‘traditional’ diet [7]. The ‘traditional’ Mediterranean diet has been described as the dietary pattern existing in Greece and Southern Italy during the late 1950s and early 1960s [13]. Briefly, it pertains to a more frugal, plant-based pattern with higher intakes of unrefined foods, such as grains, vegetables, fruits, legumes and extra virgin olive oil, and lower levels of exposure to meats, dairy and modern discretionary foods. While the benefits ascribed to a Mediterranean diet parallel those of other plant-based diets, such as vegetarian diets and the Dietary Approaches to Stop Hypertension (DASH) diet [14,15], the ‘traditional’ Mediterranean diet also includes some unique cuisine aspects. For example, low advanced glycation end products (AGEs) cooking methods [16], such as boiling and stewing, and the use of acidic ingredients in cooking, such as lemon or vinegar; particular food combinations, e.g., onions sautéed in olive oil with tomato as the basis for many cooked dishes; regular inclusion of fermented foods, e.g., olives, feta cheese; limited snacking between meals; and water as the main beverage. Such elements may contribute to increased antioxidant bioavailability, reduced chronic systemic inflammation and positively influence the microbiome, but they have been rarely captured in existing tools [13]. Hence, previous Mediterranean diet index tools may not have been sufficiently nuanced to assess adherence to the ‘traditional’ Mediterranean diet and cuisine [17–19]. This may partly explain differences in reported health outcomes, associated with a Mediterranean diet score across various studies [12,20,21]. Some findings in the literature ascribed to the ‘traditional’ Mediterranean diet may simply represent increased adherence to healthier plant-based dietary patterns, which have also been associated with benefits, such as reduced risk of type 2 diabetes and mortality, outside the context of a particular diet [20,21]. Additional limitations of existing Mediterranean diet index tools, especially when used within a Western population, include the type and number of elements assessed, source of cut-off points used, score derivation methodology and lack of validation against a dietary reference method, which is not limited by the same recall biases [22]. For example, the number of components assessed in reported indexes has ranged from only six [23] to 28 in the MEDiterranean LIFEstyle index (MEDLIFE), which evaluates the Mediterranean diet and lifestyle more broadly [24]. Studies in Western countries have used sex-specific median values from their respective populations as cut-off points for the elements detailed in the first Mediterranean Diet Score (MDS) [11,25], originally developed for use in a Greek population. This is problematic, since the median intakes for certain foods, such as meat and dairy, are high in some Western cohorts [12]. Most Mediterranean diet adherence scores, including multiple variants of the MDS, the MedDietScore [26] and Mediterranean Style Dietary Pattern Score (MSDPS) [27], require derivation from a food frequency questionnaire (FFQ) rather than being calculated directly from responses to an administered index tool. Notable exceptions are MEDLIFE [24], the Mediterranean Diet Adherence Screener (MEDAS) [28] from the PREvencion conDIeta MEDiterranea (PREDIMED) study in Spain and the Mediterranean Eating Pattern for Americans (MEPA) screener adapted from MEDAS [29]. Only some Mediterranean diet index tools attempt to adjust for Western dietary habits or non-typical Mediterranean foods, such as discretionary foods [24,28–31], which were not part of the ‘traditional’ diet and may provide different health effects. Finally, in the era of e-Health, which opens the potential for wider tool administration including regional and remote areas, to the best of our knowledge, no Mediterranean diet index tool based on the ‘traditional’ diet and cuisine has been validated for online administration. Such a tool could assist with broad population-based and longitudinal studies.

249 Nutrients 2018, 10, 1913 3 of 17

We recently developed the Mediterranean Diet and Culinary Index (MediCul) tool, to overcome these common limitations. We have tested the performance of MediCul when administered in-person, to a cohort of older Australians with MCI [22]. As MediCul has the potential for wider use, the purpose of the present study was to test its reliability and validity when administered online to middle-aged and older people in the general population, living in the same Western country.

2. Materials and Methods

2.1. Participants We aimed to recruit a cross-sectional sample of 100 [32,33] community-dwelling participants living in Sydney from January to March 2017, from the Sax Institute’s 45 and Up Study [34], as part of a validation study for various tools to be used in the Maintain Your Brain (MYB) trial administered within an online platform [35]. The 45 and Up Study is a large prospective cohort in New South Wales (NSW), Australia (n = 267,153) of participants recruited from the Department of Human Services (formerly Medicare) enrolment database representing diverse socio-demographic, geographic and cultural backgrounds, investigating healthy ageing [34]. Participants initially joined the study by completing the baseline questionnaire (between January 2006 and December 2009) and giving their consent for follow-up and linkage of their information to routine health databases. Inclusion criteria for the MYB validity study were: (1) Previously enrolled in the 45 and Up Study and agreeable to further contact with email on record, (2) aged 55–75 years in 2016, (3) absence of dementia, Parkinson’s disease and multiple sclerosis, (4) home internet/computer access, (5) access to iPad or iPhone (for activity and diet logging), and (6) residing in postcode within 30 km radius of Lidcombe, NSW (site of clinic visit in Sydney). Exclusion criteria were: (1) Unable to use a computer, (2) life threatening condition, (3) inability to write/converse in English, (4) serious psychiatric condition, on antipsychotics or suicidal ideation, and (5) unstable medical condition. A potentially eligible subset of the 45 and Up Study cohort (n = 27,848) was identified by the Sax Institute, which administers the 45 and Up Study [34], based on above inclusion criteria except for home internet/computer access and access to iPad/iPhone, details which were unavailable to the Sax Institute. Based on an estimated recruitment rate of 30%, 300 invitations were initially sent by the Sax Institute to a random sample of participants. A further 303 invitations were sent later to achieve the target sample. Participation in the MYB validity study was voluntary and electronic informed consent was obtained before starting any data collection. This was followed by full online screening for study eligibility, using both the inclusion and exclusion criteria. Ethics approval for MYB was obtained from the University of New South Wales Human Research Ethics Committee (#16252) and NSW Population and Health Services Ethics Committee (#2016/03/636). Ethical conduct for the 45 and Up Study was approved by the University of New South Wales Human Research Ethics Committee (HREC).

2.2. Assessments The MYB validity study included four assessment time points relevant to the validation of the MediCul tool. MediCul was administered twice online utilizing SurveyMonkey® (www.surveymonkey. com), at least one week apart (survey A and survey B), with an invitation for a clinic visit in between. The clinic visit was used to measure: (a) Weight in light clothing, without shoes, using calibrated digital scales (AND HW-100K and Avery Berkel HL122), (b) height using a wall-mounted Harpenden stadiometer (Holtain Limited, Crymmych, Pembrokeshire, UK), (c) waist circumference taken at the umbilicus with a flexible tape measure (Grafco®, Graham-Field, Model # 17-1340-2), and (d) resting heart rate, taken manually, after a five-minute seated resting period. All anthropometric measurements were made in triplicate with the mean being used. Body mass index (BMI) was calculated by dividing weight (kg) by height2 (m). Data relating to demographics, self-reported medical conditions, number of medications taken and supplement use were also collected as part of a suite of surveys during the first

250 Nutrients 2018, 10, 1913 4 of 17

MediCul (survey A) administration. Composite medications were counted according to the number of active ingredients they contained. Prevalence of nutrition supplement use was based on any reported vitamin/mineral/botanical supplements, excluding probiotics. Vitamin B12 injections were counted as medication. At the end of the clinic visit, or via email if a clinic visit was not possible, participants were asked to keep a three-day food record (FR) as soon as they had completed MediCul survey B, selecting one weekend and two weekdays, which needed to be within a seven-day period, but did not need to be consecutive. Participants were instructed to use the ‘Research Food Diary’ app [36] as this may reduce participant and researcher burden. This app has previously been reported to have a high level of acceptance by adults compared to a 24-hour dietary recall [37] and mobile nutrition apps, in general, have been shown in healthy populations to be comparable to traditional dietary assessment methods [38]. The Research Food Diary app is based on Australian food composition data, and entries can be directly imported into FoodWorks 8 Professional Edition software: 8.0.3553 (Xyris Software Pty Ltd., Brisbane, Australia). We provided detailed downloadable instructions for the Research Food Diary app, including color screen shots. Some participants however, declined to use the app so a hard copy FR template was provided. The FR instructions provided to all participants advised them to log typical days (not to change dietary intake), record any recipes, enter foods as consumed throughout the day, and estimate serve sizes using household measures (spoons, cups) or weights (grams (g)), if preferred, using kitchen scales. Participants were advised to expect a call from a dietitian shortly after submitting their FR. A phone call was made to each participant within 24–48 hours of receiving their FR to query unusual quantities and conﬁrm entries were complete using a checklist for potentially forgotten foods, based on principles from the ASA24 Automated Self-Administered 24-hour Dietary Assessment Tool [39,40]. Where MediCul survey A or survey B was not completed within one week of administration, an email reminder was sent with a direct link to prompt participants to complete their surveys. If either survey still remained incomplete 1–2 weeks later, a second email reminder was sent. If FRs were not submitted within one week, participants also received a reminder email. If FRs were not submitted within two weeks, telephone reminders were made.

2.3. The MediCul Tool MediCul is a 50-item short dietary survey developed empirically to assess adherence to a ‘traditional’ Mediterranean dietary pattern and aspects of cuisine [13] within a Western population. It applies reverse scoring for high exposures to foods that were not part of the ‘traditional’ pattern, such as modern discretionary foods, which now account for approximately 35% of the total energy intake of Australians [41]. A score for the popular MEDAS screener can also be derived from MediCul (Supplementary Table S1). MediCul is scored between 0 and 100, with a higher score representing greater adherence, and takes approximately 20 minutes to complete. Details about MediCul development, elements, cut-off points, scoring and rationale have previously been reported as Supplementary Materials in the British Journal of Nutrition and this tool is freely available for download and use in research and education [22]. For the present validity study, MediCul was modiﬁed for online administration using SurveyMonkey with drop-down response options, logic and data limits applied to facilitate correct dietary reporting without a dietitian being present. These backend adjustments were tested for accurate performance by ﬁve researchers.

2.4. Dietary Analysis We used Excel (MS Ofﬁce Professional Plus 2013) to derive scores for both the MediCul and MEDAS tools [22]. The FRs were coded and entered into FoodWorks 8 and we selected AusBrands 2015 and AusFoods 2015 data sources which map to the AUSNUT 2011–2013 Food Standards Australia New Zealand (FSANZ) nutrient database for analysis by S.R.-V. As the food groups within FoodWorks

251 Nutrients 2018, 10, 1913 5 of 17 draw on the concept of the USDA Food Patterns Equivalents Database (FPED), which counts hot chips/fries/crisps and legumes in the vegetable group, we manually adjusted serves for the FoodWorks vegetable group to exclude these discretionary and legume foods. We also manually added some legume serves to the legume protein foods group, which were not counted by FoodWorks. We did not analyze the nutrient contribution from supplements as our focus was on validating a dietary pattern.

2.5. Statistical Analysis Statistical analyses were conducted using the Statistical Package for Social Sciences (SPSS) for Windows version 24 (SPSS, Inc., Chicago, IL, USA). Histograms, skewness, kurtosis and minimum and maximum values were used to examine distributions for normality and identify potential data entry errors for index scores, nutrients and foods groups. Means and standard deviations (SD) were calculated and reported for normally distributed nutrients whereas medians and the interquartile range was reported for non-normally distributed nutrients. The derived MEDAS score from survey A (n = 84) was compared to the baseline MEDAS score reported for participants in PREDIMED, being the largest randomized controlled trial (RCT) of a Mediterranean diet [42]. Further, we estimated the percent of our cohort that reached previously defined Mediterranean thresholds [22] for selected foods/’traditional’ aspects of cuisine, determined by their FRs. Test-retest reliability for MediCul was assessed using the paired t-test and intra-class correlation coefficient (ICC), which was classified as poor (<0.40), fair to good (≥0.4 and <0.75), and very good (≥0.75) [43]. A Bland-Altman plot was also used to determine the level of agreement between survey A and survey B time points, since a high correlation does not necessarily mean good agreement [44]. Kappa was utilized to check percent agreement within the same category [32] for 17 major food groups within MediCul and values were interpreted using cut-offs proposed by Landis and Koch [45]. We used Spearman’s correlation coefficient to assess whether the lag time between survey A and survey B was correlated with the difference in MediCul scores at the two time points. We also investigated whether selected categorical (i.e., sex, supplement use, primary grocery shopper, primary cook) or continuous (i.e., age, BMI) variables might be driving the improved performance for survey B, using Chi-square, linear regression and paired t-tests, as appropriate. Concurrent validity for MediCul was determined by examining the Bland-Altman plot of the difference in MediCul scores between the survey tool and the FR versus the mean of the scores from the two assessment methods, with limits of agreement (LOA) being ±2 standard deviations (SD) from the mean difference. Bias was assessed using linear regression analysis. A paired t-test was conducted for scores from the survey at time points A, B, and the mean of AB versus the FR, to check the mean differences and confidence intervals (CI) across these comparisons. We also indirectly validated MediCul by examining whether expected (Mediterranean) nutrient trends from the FR were associated with tertiles of the MediCul score. For positively or negatively skewed nutrients, we conducted a log 10 transformation, then re-checked their distribution for normality to determine whether parametric or non-parametric statistical tests should be applied. Normally distributed and normalized nutrients were compared across tertiles of the MediCul score derived from both survey A and the FR, using one-way ANOVA, while non-normally distributed nutrients were analyzed using the Kruskal-Wallis test. Variances were checked for equality using the Levene’s test; the Bonferroni post hoc test was applied for equal variances and the Games-Howell for unequal variances. First (linear)- and second (quadratic)-order polynomial contrasts were applied to test for nutrient trends across tertiles, as well as the line of best fit for normally distributed/normalized nutrients. The Jonckheere trend test was used for non-normally distributed nutrients. As the FR did not include details for three MediCul questions (i.e., growing of own vegetables, weekly main meals eaten alone and fasting frequency), the validation of MediCul was based on scoring out of 97 whereas reliability analysis was based on scoring out of 100.

252 Nutrients 2018, 10, 1913 6 of 17

Nutrients 2018, 10, x FOR PEER REVIEW 6 of 17 3. Results 3. Results We received 175 responses to the 603 invitations sent by the Sax Institute to potentially eligible participantsWe fromreceived the 175 45 responses and Up Study.to the 603 Of invitations these, 93 sent participants by the Sax wereInstitute eligible to potentially after full eligible screening (Figureparticipants1). A total from of the 84/93 45 and (90%) Up Study. completed Of these, the 93 ﬁrst participants administration were eligible of MediCul after full (survey screening A) and 65 participants(Figure 1). A elected total of to84/ make93 (90%) a cliniccompleted visit the for first physical administration measurements. of MediCul Seventy-six (survey A) participants and 65 participants elected to make a clinic visit for physical measurements. Seventy-six participants completed MediCul on two occasions. The ﬁnal number included for reliability testing was 74, completed MediCul on two occasions. The final number included for reliability testing was 74, since since two participants started their FR prior to completing MediCul survey B and were excluded. two participants started their FR prior to completing MediCul survey B and were excluded. Validity Validitytesting testing was wasconducted conducted with with71 participants 71 participants who completed who completed both survey both surveyA and the A andFR, thewith FR, 69% with of 69% of thesethese using using the the app app and and submitting submitting an an electronic electronic FR. FR. Data Data collection collection for for all participantsall participants was was completedcompleted by mid-May by mid-May 2017. 2017.

FigureFigure 1. Flow 1. Flow chart chart of participants. of participants. Abbreviations: Abbreviations: MediCul MediCul survey survey A, ﬁrst A, administration first administration of MediCul of tool;MediCul tool; survey MediCul B, second survey administration B, second administration of MediCul of MediCul tool; MS, tool; multiple MS, multiple sclerosis; sclerosis; n, number n, of participants;number of Sax, participants; the Sax Institute. Sax, the Sax Institute.

The majorityThe majority of the of recruited the recruited participants participants were were female, female, and reported and reported taking taking at least at one least nutritional one supplement,nutritional with supplement, about one-half with about being one-half the being primary the primary grocery grocery shopper shopper and and primary primary cook cook at at home home (Table 1). Participants ranged in age from 56 to 76 years, were overweight on average (BMI = (Table 1). Participants ranged in age from 56 to 76 years, were overweight on average (BMI = 26.1; 26.1; SD = 4.0), with approximately one-half having a waist circumference associated with high SD = 4.0), with approximately one-half having a waist circumference associated with high chronic chronic disease risk [46] (Table 1). They had 1.7 (SD = 1.5) chronic illnesses/conditions, on average, diseasewith risk the [ 46most] (Table prevalent1). They being had osteoarthritis, 1.7 (SD = 1.5)followed chronic by illnesses/conditions,hypercholesterolemia and on hypertension average, with. the mostThe prevalent median being number osteoarthritis, of medications followed taken was by 1.5 hypercholesterolemia (IQR: 0.0, 3.0). and hypertension. The median number ofThe medications mean MediCul taken score was from 1.5 survey (IQR: 0.0,A (n 3.0).= 84) was 55.2/100.0 (SD = 11.6; range: 27.5, 75.5). The Thederived mean mean MediCul MEDAS score score fromfor the survey same time A ( npoint= 84) was was 6.5/14.0 55.2/100.0 (SD = 1.7; (SD range: = 11.6; 2.0, range: 10.0), which 27.5, 75.5). The derivedwas approximately mean MEDAS two points score lower for the than same the baseline time point MEDAS was score 6.5/14.0 of 8.6 (SD(SD = 2) 1.7; [42 range:] reported 2.0, in 10.0), whichthe was Spanish approximately PREDIMED two population, points lower prior than to intervention the baseline with MEDAS a Mediterranean score of 8.6 diet (SD supplemented = 2) [42] reported in thewith Spanish extra virgin PREDIMED olive oil population, or nuts or a reduced prior to fat intervention control diet. with However, a Mediterranean it was one point diet higher supplemented than reported recently for a Western UK cohort [47]. Only 3/84 (4%) of our participants achieved a MEDAS with extra virgin olive oil or nuts or a reduced fat control diet. However, it was one point higher than score of ≥10, which has been classified as high adherence to a Mediterranean diet and associated with reported recently for a Western UK cohort [47]. Only 3/84 (4%) of our participants achieved a MEDAS score of ≥10, which has been classiﬁed as high adherence to a Mediterranean diet and associated with

253 Nutrients 2018, 10, 1913 7 of 17

Nutrients 2018, 10, x FOR PEER REVIEW 7 of 17 better health outcomes in PREDIMED (2, 48–50). All of these participants were in the highest tertile of the MediCulbetter health score outcomes from survey in PREDIMED A (cut-off (2, =48 59.0–50). for All tertile of these 3) whenparticipants compared were toin nutrientsthe highest from tertile the FR. Few participantsof the MediCul reached score from defined survey Mediterranean A (cut-off = 59.0 thresholds for tertile 3) for when selected compared foods/’traditional’ to nutrients from aspects the of cuisineFR. [ 22 Few], especially participants for reachedolive oil defined (0%), legumes Mediterranean (6%), sofrito, thresholds a sauce for made selected with foods/ tomato,’traditional’ onion/garlic and oliveaspects oil, of usedcuisine as [22 the], especially base for for many olive dishes oil (0%), (10%), legumes vegetables (6%), sofrito, (11%) a sauce and made fruit with (18%) tomato (Figure, 2). However,onion/garlic the majority and olive reported oil, used cooking as the base most for ofmany their dishes main (10%), meals vegetables at home (73%),(11%) and limiting fruit (18%) snacking occasions(Figure (79%), 2). However, limiting the sugary majority drinks reported (83%) cooking and consuming most of their raw main vegetables/salads meals at home (73%), most limiting days of the snacking occasions (79%), limiting sugary drinks (83%) and consuming raw vegetables/salads most week (86%). Almost half included an adequate intake of nuts (44%) and fish/shellfish (48%). days of the week (86%). Almost half included an adequate intake of nuts (44%) and fish/shellfish (48%). Table 1. Participant characteristics at baseline (n = 84).

TableDescriptive 1. ParticipantStatistics characteristics (Mean (SD) at orbaseline Proportion) (n = 84). Age (years)Descriptive statistics (mean (SD) or proportion)65.4 (5.9) FemalesAge (%) (years) 65.4 (5.9)60 BMI * (kg/mFemales2) (%) 26.160 (4.0) <18.5 (%)BMI * (kg/m2) 26.1 (4.0)1.5 18.5 to 24.9<18.5 (%) (%) 1.535.4 25.0 to 29.918.5 to (%) 24.9 (%) 35.447.7 ≥30 (%)25.0 to 29.9 (%) 47.715.4 Waist circumference≥30 (%) *—females (cm) 91.415.4 (12.4) ≥88 cmWaist (%) circumference *—females (cm) 91.4 (12.4)56.4 Waist circumference≥88 cm (%) *—males (cm) 100.256.4 (8.2) ≥102 cmWaist (%) circumference *—males (cm) 100.2 (8.2)46.2 Resting≥102 heart cm rate (%) * (beats per minute) 46.266 (8) PrimaryResting cook atheart home rate (%) * (beats per minute) 66 (8)49 PrimaryPrimary grocery cook shopper at home at (%) home (%) 49 51 SupplementPrimary use grocery (%) shopper at home (%) 51 63 NumberSupplement of chronic use illnesses/conditions (%) 63 1.7 (1.5) OsteoarthritisNumber (%) of chronic illnesses/conditions 1.7 (1.5)36 HypercholesterolemiaOsteoarthritis (%) (%) 36 30 HypertensionHypercholesterolemia (%) (%) 30 24 Gastro-esophagealHypertension reﬂux (%) disease (%) 24 16 Gastro-esophageal reflux disease (%) 16 Depression (%) 7 Depression (%) 7 Diabetes (%) 4 Diabetes (%) 4 Number of medications (median (IQR)) 1.5 (0.0, 3.0) Number of medications (median (IQR)) 1.5 (0.0, 3.0) Abbreviations:Abbreviations: IQR, IQR,interquartile interquartile range range representing representing the 25th the and 25th 75th and percentiles; 75th percentiles;n, number n, number of participants; of SD, standard deviation; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared).participants; * measured SD, at standard clinic visit deviation; (n = 65). BMI, body mass index (calculated as weight in kilograms divided by height in meters squared). * measured at clinic visit (n = 65).

Figure 2. Percent of participants who reached Mediterranean diet thresholds according to three-day Figure 2. food Percentrecords ( ofn = participants 71). Thresholds who defined reached by highest Mediterranean cut-offs used diet in thresholds MediCul tool according for the relevant to three-day foodquestions records ([n22=]. 71). Abbreviations: Thresholds n,deﬁned number byof participants; highest cut-offs d, day; used Prep, in preparation; MediCul tool Sofrito, for thea sauce relevant questions [22]. Abbreviations: n, number of participants; d, day; Prep, preparation; Sofrito, a sauce made with tomato, onion/garlic and olive oil, used as a base for savory dishes; TBSP, tablespoon; wk, week.

254 Nutrients 2018, 10, 1913 8 of 17

3.1. Reliability There was a considerable lag time between MediCul survey A and survey B for some participants, so the time between surveys was non-normally distributed (median = 16 days; IQR: 13, 20). As we did not have an a priori criterion for the length of time between surveys, we did not exclude any participants for reliability testing based on the time between surveys. MediCul B score was similar to MediCul A 56.8/100 and 56.1/100.0, respectively (t = −1.019; p = 0.311). The MediCul tool was highly reliable as assessed by ICC and 95% CI, (ICC = 0.86, 95% CI: 0.789, 0.910, p < 0.0001). Using the Bland-Altman test for repeated measures, the mean difference between the scores at the two time points was −0.69 with lower and upper LOA being −12.09 and 10.71, respectively. Seventy of the 74 participants (95%) fell within the LOA representing ±2 SD, with a reasonably even distribution of the mean difference in scores. There was no indication of bias according to the regression coefficient ((y= −2.95 + 0.04*x) (95% CI: −0.088, 0.168; p = 0.535)), further supporting the null hypothesis that the scores were equally variable at the two time points. The slightly higher mean score from survey B was unrelated to age, BMI, sex, supplement use, being the primary grocery shopper or primary cook. The lag time between repeated tool administration was unrelated to the difference between MediCul scores (Spearman’s p = −0.102; p = 0.385), suggesting that MediCul is a stable index for usual intakes across the time spans in this study. Percent agreement within the same category of the 17 main MediCul food groups for survey A and survey B, determined by Kappa, was ‘almost perfect’ for coffee; ‘substantial’ for fruit, wholegrains, fish/shellfish, eggs, dairy products, takeaway and water; ‘moderate’ for olive oil, nuts, legumes, white meat preference, red/processed meat, sweets and sugary drinks and alcohol; and ‘fair’ for vegetables and cuisine (Supplementary Table S2). No groups were found to have poor agreement within the same category between the two time points of MediCul tool administration.

3.2. Validity Analysis of the paired t-tests for MediCul scores from the survey administered at time points A, B and mean of AB versus the FR (n = 68) indicated a similar mean difference and 95% CI across the three comparisons, which were all statistically significant (p < 0.0001) (Table 2). We thus decided to proceed with survey A for the remaining validity testing, because this time point included most data points (n = 71) and represented first time MediCul use, which could be applied in future research. The Bland-Altman plot showed a mean difference of 5.6 between the MediCul score derived from survey A (54.6/97.0, SD = 11.2, range: 26.5, 73.5) and the FR (49.0/97.0, SD = 11.7, range: 25.0, 80.0), and the mean difference was lower among participants who provided a paper version of their FR compared to the electronic version (supplied through the app), although both were similarly checked by the dietitian (Table 2). The scores for 69/71 participants (97%) fell within or on the 95% LOA with lower and upper LOA being −13.0 and 24.2, respectively, with reasonably even distribution across mean scores. There was no significant linear trend for the fitted regression line ((y = 8.7 − 0.06*x) (95% CI: −0.278, 0.158, p = 0.584)), indicating no systematic bias between the two measurement methods (Figure 3).

Table 2. Mean difference from paired samples t-test for the Mediterranean Diet and Culinary Index (MediCul) scores from surveys A, B, mean AB versus three-day food record (n = 68).

Mean Difference 95% CI of the Difference p Value A versus FR * 5.32 3.03, 7.62 <0.0001 B versus FR 6.38 4.16, 8.59 <0.0001 Mean AB versus FR 5.85 3.70, 8.00 <0.0001 Abbreviations: A, ﬁrst administration of MediCul; B, second administration of MediCul; mean AB, mean of A and B MediCul administrations; n, number of participants; FR, three-day food record; CI, conﬁdence intervals. * When compared to Survey A with maximum data points (n = 71), participants using a paper FR (n = 22) had a mean difference of 2.6 points (95% CI: −2.36, 7.63; t = 1.10; p = 0.285), whereas those using the electronic FR (n = 49) had a mean difference of 6.9 points (95% CI: 4.50, 9.30; t = 5.77; p < 0.0001).

255 Nutrients 2018, 10, x FOR PEER REVIEW 9 of 17

using a paper FR (n = 22) had a mean difference of 2.6 points (95% CI: −2.36, 7.63; t = 1.10; p = 0.285), Nutrientswhereas2018, 10 those, 1913 using the electronic FR (n = 49) had a mean difference of 6.9 points (95% CI: 4.50, 9.30;9 of 17 t = 5.77; p < 0.0001).

Figure 3. Bland-AltmanBland-Altman plot of the the difference difference in the Mediterranean Diet and Culinary Index (MediCul) score between survey A (first (first administration of of MediCul tool) tool) and and three-day FR versus the mean MediCul score score of of the the two two methods methods (n (n = =71). 71). The The solid solid line line indicates indicates the themean mean difference, difference, whereas whereas the dottedthe dotted lines lines are the are limits the limits of agreement of agreement representing representing ±2 SD± from2 SD the from mean the difference mean difference within withinwhich 95%which of 95% the of differences the differences between between the methods the methods are expected are expected to fall. to fall. The The fitted fitted regression regression line line is ((y=8.7is ((y =−0.06*x) 8.7−0.06*x (95%) (95% CI: CI: −0.278,−0.278, 0.158, 0.158, p =p 0.584)),= 0.584)), and and indicates indicates no no significant significant linear linear trend. trend. Abbreviations: FR, food record.

We hypothesized that that a a higher higher intake intake of of certain certain nutrients nutrients (determined (determined by byFR FR analysis) analysis) would would be associatedbe associated with with the the higher higher tool tool (survey (survey A) A) and and FR-derived FR-derived MediCul MediCul scores, scores, categorized categorized as as tertiles. tertiles. This was true for: Polyunsaturated Polyunsaturated fatty fatty acids acids (PUFA) (PUFA) as as % % fat, fat, nn-3-3 long long chain chain (LC) (LC) PUFA PUFA (mg), (mg), eicosapentaenoic acid acid (EPA) (EPA) (mg), docosahexaenoic acid acid (DHA) (mg), and and ratio of unsaturated to saturated fatty fatty acids acids (SFA), (SFA), supporting supporting increased increased adherence adherence to a Mediterranean to a Mediterranean dietary pattern dietary (Table pattern3). (TableAdditionally, 3). Additionally, we hypothesized we hypothesized that a lower that intake a lowerof some intake other of nutrients some other would nutrients be associated would with be associatedthe higher MediCulwith the higher score tertiles MediCul from score both tertiles sources. from This both was sources. true for SFAThis aswas % fat,true alcohol for SFA (g) as and % fat the, alcoholsodium (g)to potassium and the sodium ratio, being to potassium consistent ratio, with thebeing plant consistent based, minimally with the plantprocessed, based, nature minimally of the ‘traditional’processed, nature Mediterranean of the ‘traditional’ diet, which Mediterranean is proportionally diet, lower which in is SFA proportionally and includes lower only in a moderateSFA and includesamountof only alcohol. a moderate For the sodium amount to ofpotassium alcohol. ratio, For thea significant sodium difference to potassium (p = 0.004) ratio, was a significant observed differencebetween MediCul (p = 0.004) score was tertiles observed 1 and between 3, derived MediCul from thescore FR. tertiles 1 and 3, derived from the FR. For some nutrients, the expected relationship was observed only with the MediCul score tertiles from the FR e.g., PUFA (g), alpha linolenic acid (ALA) (mg), ratio of monounsaturated fatty fatty acids acids (MUFA) toto SFA,SFA, dietary dietary fiber fiber (g), (g), vitamin vitamin E (mg),E (mg), potassium potassium (mg), (mg), magnesium magnesium (mg) (mg) and and selenium selenium (µg). (μg).For magnesium, For magnesium, not only not was only there was athere positive a positive linear trendlinear across trend tertiles across oftertiles the MediCul of the MediCul score from score the fromFR, but the a FR, significant but a significant difference difference was observed was observed between between tertiles 1 tertiles and 3 (1p and= 0.015) 3 (p = and 0.015) tertiles and 2tertiles and 3 2(p and= 0.016). 3 (p = 0.016). PUFA %PUFA fat, EPA% fat, and EPA DHA and areDHA nutrient are nutrient examples examples for which for which significant significant differences differences were werefound found between between tertiles tertiles 1 and 1 3 and of the 3 of MediCul the MediCul score, score, derived derived fromboth from the both tool the and tool FR and (Table FR (Table3). 3). In all cases linear trends were significant and there were no significant deviations from normality exceptIn all for cases retinol linear equivalents trends were (µg), significant beta-carotene and there (µg) were and ironno significant (mg), where deviations a quadratic from trendnormality was excepta better for fit retinol for these equivalents nutrients in(μg), relation beta-carotene to the MediCul (μg) and score iron tertiles (mg), fromwhere survey a quadratic A. trend was a betterThere fit for was these no nutrients relationship in relation between to tertiles the MediCul for either score of thetertiles MediCul from survey scores andA. intake of energy (kJ), proteinThere was (g), no total relationship fat (g), SFA between (g), MUFA tertiles (g), cholesterolfor either of (mg), the MediCul carbohydrate scores (g), and total intake sugars of energy (g) and (kJ),folate protein (µg). (g), total fat (g), SFA (g), MUFA (g), cholesterol (mg), carbohydrate (g), total sugars (g) and folate (μg).

256 Nutrients 2018, 10, 1913 10 of 17

Table 3. Nutrient intakes compared with tertiles of the Mediterranean Diet and Culinary Index (MediCul) score from FR and survey A (n = 71) *.

257 Nutrients 2018, 10, 1913 11 of 17

Table 3. Cont.

258 Nutrients 2018, 10, 1913 12 of 17

Table 3. Cont.

Source of the MediCul Nutrient Intake for Tertile 1 Nutrient Intake for Tertile 2 Nutrient Intake for Tertile 3 Comparison of Nutrient Nutrients from Food of the MediCul Score of the MediCul Score of the MediCul Score Test for Trend p Value, Record Score Tertile Cut-Off Intakes across Tertiles of Direction ↑ ↓ † Points Mean/median SD/IQR Mean/median SD/IQR Mean/median SD/IQR the MediCul Score p Value Retinol equivalents FR 1092 557 1193 568 1266 587 0.577 0.298 µ g/d Survey 1153 404 972 b 557 1454 637 0.012 0.016 FR 4567 3307 5619 3245 6047 3273 0.285 0.125 Beta carotene µg/d Survey 5043 2477 4167 b 3112 7196 3593 0.004 0.013 FR 2330 753 2021 750 1920 688 0.141 0.058 Sodium mg/d Survey 2298 836 2064 699 1901 652 0.190 0.071 FR 3513 813 3690 1057 4105 945 0.096 0.036 ↑ Potassium mg/d Survey 3691 860 3520 939 4124 1028 0.089 0.135 a Sodium:Potassium FR 0.7 0.3 0.6 0.2 0.5 0.2 0.006 0.001 ↓ ratio Survey 0.6 0.2 0.6 0.2 0.5 0.2 0.072 0.031 ↓ FR 390 a 349, 500 361 b 321, 466 470 440, 525 0.006 0.005 ↑ Magnesium mg/d Survey 418 333, 489 393 365, 469 466 365, 557 0.154 0.148 FR 12.1 3.3 11.7 3.4 13.2 3.5 0.288 0.254 Iron mg/d Survey 12.9 3.6 11.1 2.7 13.1 3.7 0.091 0.030 FR 12.1 4.0 10.7 3.2 11.9 3.1 0.336 0.868 Zinc mg/d Survey 11.8 4.2 11.1 3.1 11.8 3.1 0.721 0.960 FR 89 70, 109 91 76, 105 103 92, 134 0.035 0.017 ↑ Selenium µg/d Survey 96 81, 117 92 68, 109 96 83, 110 0.435 0.751 Abbreviations: n, number of participants; SD, standard deviation; IQR, interquartile range (representing the 25th and 75th percentiles); kJ, kilojoules; d, day; FR, food record; g, grams; SFA, saturated fatty acids; PUFA, polyunsaturated fatty acids; MUFA, monounsaturated fatty acids; n-3, omega 3; LC, long chain; mg, milligrams; ALA, alpha linolenic acid; EPA, eicosapentaenoic acid; DPA, docosapentaenoic acid; DHA, docosahexaenoic acid. * MediCul index score tertiles are derived from FR and survey A (first administration of MediCul). For survey A, the cut-off points for tertiles 2 and 3 were 50.0 and 59.0, respectively. Values are presented as mean (SD) for normally distributed data or median (IQR) for non-normally distributed data. These data were normalized by logarithmic transformation for use in ANOVA models with the exception of alcohol, selenium, EPA and DHA, which were unable to be normalized and therefore analyzed using the Kruskal-Wallis model. When ANOVA F ratio was significant, variances were checked for equality using the Levene’s test, and Bonferroni was applied for equal variances or Games-Howell post hoc t test for unequal variances. First (linear)- and second (quadratic)-order polynomial contrasts were applied to test for trends across tertiles, as well as the line of best fit for normally distributed/normalized nutrients. The Jonckheere trend test was used for non-normally distributed nutrients. In all cases linear trends were significant and there were no significant deviations from normality, except for retinol equivalents, beta carotene and iron when compared with tertiles from survey A, where the quadratic trend was positive and the most significant. Significant differences (p < 0.05) between tertiles denoted as: a = 1 vs. 3; b = 2 vs. 3. † Linear trend direction indicated as ↑ (increasing) or ↓ (decreasing). ‡ Naturally occurring food folates only.

259 Nutrients 2018, 10, 1913 13 of 17

4. Discussion Our study indicates that online administration of MediCul has good reliability and moderate validity, as an assessment of adherence to a ‘traditional’ Mediterranean dietary pattern and aspects of cuisine among the middle-aged and older individuals studied who live in a Western country. These findings are consistent with those reported for in-person administration among older individuals with MCI [22]. The participants included in our study were Australians aged 55–75 years, living in suburban Sydney. Only 4% had high adherence to a Mediterranean diet based on their derived MEDAS scores. Higher MEDAS scores have been related to positive health outcomes in PREDIMED [2,48–50]. Similarly, a very small proportion reached thresholds for ‘traditional’ Mediterranean foods or exposures. For example, 11% reported adequate vegetable intake, 6% had adequate legumes, and 49% limited their exposure to red meat, in accordance with ‘traditional’ levels. These participants represented a fairly homogenous group in regards to their dietary intakes, with few people achieving a high MediCul score. Such a cohort may be at higher chronic disease risk. Education on the Mediterranean diet, which is already recommended by dietary guidelines, may be prudent. Targets for ‘traditional’ Mediterranean foods and cuisine aspects can be viewed in previously published Supplementary Materials that describe cut-off points used in the MediCul tool [22]. For example, four tablespoons of extra virgin olive oil per day; one serve (30 g) of nuts, five or more days per week; inclusion of dark green leafy vegetables (e.g., spinach, kale, silverbeet), four or more times per week. Concurrent validity of MediCul has therefore been demonstrated within a cohort that is almost entirely Western in their eating habits. Further research is required to validate MediCul use among other population groups, including those with broader dietary habits (particularly at the high end), and in relation to biomarkers of dietary intake. The MYB trial, a three-year randomized controlled trial of a fully online multi-modal intervention targeting modifiable dementia risk factors in 6236 participants, will utilize MediCul at baseline and annual follow up. This will generate data on change in MediCul score over time and its association with chronic disease risk factors, which will provide external validity, since clinical end points were not available from this validation study [35]. Given the good reliability of the MediCul tool shown here, we anticipate that changes in the index score, in the context of the MYB trial and similar dietary interventions, would represent true dietary change, and could potentially serve as a marker of positive health outcomes related to the ‘traditional’ diet and cuisine.

Strengths and Limitations The MediCul tool assesses the overall dietary pattern within a Western context, capturing the multi-dimensional nature of individual diets. With 50 items being assessed, it provides a comprehensive score for adherence to a ‘traditional’ Mediterranean dietary pattern and aspects of cuisine. Many of these elements are lacking from most other index tools. MediCul also enables derivation of a MEDAS score for comparison with studies that have used this short screener. Importantly, it is relatively simple to use and compute its scoring and MediCul can be administered online and in-person [22], increasing its potential utility, feasibility in large cohorts and cost-effectiveness. Our study has some limitations. Data are currently lacking to support external validity for MediCul, which is of clinical relevance. Generalizability may also have been tempered by selection bias as our participants came from the 45 and Up Study, thus representing individuals who are better educated, more health aware and who may respond differently to dietary surveys compared to non-volunteers [32]. Our participants had less obesity and diabetes compared to the general Australian population for a similar age group [51,52]. MediCul was also administered within a suite of surveys, which may have caused fatigue, impacting on responses and scores. This was unavoidable as we wished to simulate the online testing experience of the subsequent MYB trial. Our testing, nevertheless, conﬁrmed good reliability and moderate validity. It is also well known that food records may under-report intake or include atypical days [53] and there were some limitations speciﬁcally

260 Nutrients 2018, 10, 1913 14 of 17 related to use of the Research Food Diary app. For example, certain recipes entered into the app by participants included some scanned foods that did not exist in the FoodWorks data source available to us, and needed to be re-entered by the dietitian, which was done by selecting nutritionally similar foods. If participants selected composite dishes within the app or reported a composite meal in a paper FR (e.g., stir fry chicken and vegetables), the dietitian checked all entries similarly during the phone call by probing for potentially missed foods and ingredients, such as oil. While the MediCul tool performed well overall, it performed better against a traditional (paper) FR, as compared to an electronic FR (using an app). However, it was not our intention to compare methods of keeping an FR, and we therefore did not exclude any participants who provided a paper FR as we did not have an a priori criterion to do so. In some participants, the approximately 20 point difference in MediCul score between the new tool and the FR, could be clinically meaningful. However, it has been suggested that a certain degree of disagreement is to be expected when instruments assessing usual diet are validated against food records, due to within-person variations that typically occur over the shorter reference period of an FR [32]. Also, the lower and upper LOA (13% and 25%, respectively) were similar to ﬁndings for MediCul from our validity study among individuals with MCI [22] and well within the 50% and 200% LOA suggested by Ambrosini et al. [54] to classify agreement between an FFQ and food record as being acceptable. Furthermore, the new MediCul tool appears to be closer to the reference method, as compared to MEDAS. MEDAS under- and overestimated scores by 43% and 53%, respectively, when validated in a Spanish population against an FFQ [28], or by 16% and 36% when validated within a UK cohort against an FR [47].

5. Conclusions MediCul is a highly reliable and moderately valid Mediterranean diet index tool for online administration among health aware, better educated, middle-aged and older individuals living in a Western country, which can be used to assess adherence to a ‘traditional’ dietary pattern and aspects of cuisine relevant to many important health outcomes.

Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6643/10/12/1913/ s1, Table S1: How to derive a MEDAS score from MediCul and Table S2: Kappa statistics for percent agreement within the same category of 17 MediCul groups, between survey A and survey B. Author Contributions: Conceptualization, S.R.-V., M.A.F.S. and V.M.F.; Methodology, S.R.-V., M.A.F.S. and V.M.F.; Investigation, S.R.-V.; Formal Analysis, S.R.-V.; Supervision, M.A.F.S. and V.M.F.; Writing-Original Draft, S.R.-V.; Writing–Review and Editing, M.A.F.S., K.D., Y.N., N.J., F.O., Y.M., M.H., J.M., Y.G., T.C., P.S.S., H.B., and V.M.F.; Funding Acquisition, M.A.F.S., H.B. and P.S.S; All authors provided critical revision of the manuscript and approved the final version. Funding: The MYB study was funded by a NHMRC Dementia Research Team Grant (APP1095097). Acknowledgments: This research was completed using data collected through the 45 and Up Study (www. saxinstitute.org.au). The 45 and Up Study is managed by the Sax Institute in collaboration with major partner Cancer Council NSW; and partners: The National Heart Foundation of Australia (NSW Division); NSW Ministry of Health; NSW Government Family and Community Services—Ageing, Carers and the Disability Council NSW; and the Australian Red Cross Blood Service. We thank the many thousands of people participating in the 45 and Up Study. We also wish to thank Lyn Cobiac for feedback on the original MediCul tool and Deborah Black for statistical advice. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Esposito, K.; Maiorino, M.I.; Ceriello, A.; Giugliano, D. Prevention and control of type 2 diabetes by Mediterranean diet: A systematic review. Diabetes Res. Clin. Pract. 2010, 89, 97–102. [CrossRef][PubMed] 2. Estruch, R.; Ros, E.; Salas-Salvado, J.; Covas, M.I.; Corella, D.; Aros, F.; Gomez-Gracia, E.; Ruiz-Gutierrez, V.; Fiol, M.; Lapetra, J.; et al. Primary prevention of cardiovascular disease with a Mediterranean diet supplemented with extra-virgin olive oil or nuts. N. Engl. J. Med. 2018, 378, e34. [CrossRef][PubMed]

261 Nutrients 2018, 10, 1913 15 of 17

3. Jacka, F.N.; O’Neil, A.; Opie, R.; Itsiopoulos, C.; Cotton, S.; Mohebbi, M.; Castle, D.; Dash, S.; Mihalopoulos, C.; Chatterton, M.L.; et al. A randomised controlled trial of dietary improvement for adults with major depression (the ‘SMILES’ trial). BMC Med. 2017, 15, 23. [CrossRef][PubMed] 4. Kastorini, C.-M.; Milionis, H.J.; Esposito, K.; Giugliano, D.; Goudevenos, J.A.; Panagiotakos, D.B. The effect of Mediterranean diet on metabolic syndrome and its components: A meta-analysis of 50 studies and 534,906 individuals. J. Am. Coll. Cardiol. 2011, 57, 1299–1313. [PubMed] 5. Parletta, N.; Zarnowiecki, D.; Cho, J.; Wilson, A.; Bogomolova, S.; Villani, A.; Itsiopoulos, C.; Niyonsenga, T.; Blunden, S.; Meyer, B.; et al. A Mediterranean-style dietary intervention supplemented with fish oil improves diet quality and mental health in people with depression: A randomized controlled trial (HELFIMED). Nutr. Neurosci. 2017.[CrossRef][PubMed] 6. Psaltopoulou, T.; Sergentanis, T.N.; Panagiotakos, D.B.; Sergentanis, I.N.; Kosti, R.; Scarmeas, N. Mediterranean diet, stroke, cognitive impairment, and depression: A meta-analysis. Ann. Neurol. 2013, 74, 580–591. [CrossRef][PubMed] 7. Radd-Vagenas, S.; Duffy, S.L.; Naismith, S.L.; Brew, B.J.; Flood, V.M.; Fiatarone Singh, M.A. Effect of the Mediterranean diet on cognition and brain morphology and function: A systematic review of randomized controlled trials. Am. J. Clin. Nutr. 2018, 107, 389–404. [CrossRef] 8. Ryan, M.C.; Itsiopoulos, C.; Thodis, T.; Ward, G.; Trost, N.; Hofferberth, S.; O’Dea, K.; Desmond, P.V.; Johnson, N.A.; Wilson, A.M. The Mediterranean diet improves hepatic steatosis and insulin sensitivity in individuals with non-alcoholic fatty liver disease. J. Hepatol. 2013, 59, 138–143. [CrossRef] 9. Sofi, F.; Cesari, F.; Abbate, R.; Gensini, G.F.; Casini, A. Adherence to Mediterranean diet and health status: Meta-analysis. BMJ 2008, 337, a1344. [CrossRef] 10. Sofi, F.; Macchi, C.; Abbate, R.; Gensini, G.F.; Casini, A. Mediterranean diet and health status: An updated meta-analysis and a proposal for a literature-based adherence score. Public Health Nutr. 2014, 17, 2769–2782. [CrossRef] 11. Trichopoulou, A.; Costacou, T.; Bamia, C.; Trichopoulos, D. Adherence to a Mediterranean diet and survival in a Greek population. N. Engl. J. Med. 2003, 348, 2599–2608. [CrossRef][PubMed] 12. D’Alessandro, A.; De Pergola, G. The Mediterranean Diet: Its definition and evaluation of a priori dietary indexes in primary cardiovascular prevention. Int. J. Food Sci. Nutr. 2018, 69, 647–659. [CrossRef][PubMed] 13. Radd-Vagenas, S.; Kouris-Blazos, A.; Fiatarone Singh, M.; Flood, V.M. Evolution of Mediterranean diets and cuisine: Concepts and definitions. Asia Pac. J. Clin. Nutr. 2017, 26, 749–763. [PubMed] 14. Locke, A.; Schneiderhan, J.; Zick, S.M. Diets for health: Goals and guidelines. Am. Fam. Phys. 2018, 97, 721–728. [PubMed] 15. Martínez-González, M.A.; Sánchez-Tainta, A.; Corella, D.; Salas-Salvadó, J.; Ros, E.; Arós, F.; Gómez-Gracia, E.; Fiol, M.; Lamuela-Raventós, R.M.; Schröder, H.; et al. A provegetarian food pattern and reduction in total mortality in the Prevención con Dieta Mediterránea (PREDIMED) study. Am. J. Clin. Nutr. 2014, 100, 320S–328S. [CrossRef][PubMed] 16. Uribarri, J.; Woodruff, S.; Goodman, S.; Cai, W.; Chen, X.; Pyzik, R.; Yong, A.; Striker, G.E.; Vlassara, H. Advanced glycation end products in foods and a practical guide to their reduction in the diet. J. Am. Diet. Assoc. 2010, 110, 911–916. [CrossRef][PubMed] 17. D’Alessandro, A.; De Pergola, G. Mediterranean diet and cardiovascular disease: A critical evaluation of a priori dietary indexes. Nutrients 2015, 7, 7863–7888. [CrossRef] 18. Davis, C.; Bryan, J.; Hodgson, J.; Murphy, K. Definition of the Mediterranean diet: A literature review. Nutrients 2015, 7, 9139–9153. [CrossRef] 19. Hoffman, R.; Gerber, M. Evaluating and adapting the Mediterranean diet for non-Mediterranean populations: A critical appraisal. Nutr. Rev. 2013, 71, 573–584. [CrossRef] 20. Kim, H.; Caulfield, L.E.; Rebholz, C.M. Healthy plant-based diets are associated with lower risk of all-cause mortality in US adults. J. Nutr. 2018, 148, 624–631. [CrossRef] 21. Satija, A.; Bhupathiraju, S.N.; Rimm, E.B.; Spiegelman, D.; Chiuve, S.E.; Borgi, L.; Willett, W.C.; Manson, J.E.; Sun, Q.; Hu, F.B. Plant-based dietary patterns and incidence of type 2 diabetes in US men and women: Results from three prospective cohort studies. PLoS Med. 2016, 13, e1002039. [CrossRef][PubMed]

262 Nutrients 2018, 10, 1913 16 of 17

22. Radd-Vagenas, S.; Fiatarone Singh, M.A.; Inskip, M.; Mavros, Y.; Gates, N.; Wilson, G.C.; Jain, N.; Meiklejohn, J.; Brodaty, H.; Wen, W.; et al. Reliability and validity of a Mediterranean diet and culinary index (MediCul) tool in an older population with mild cognitive impairment. Br. J. Nutr. 2018, 120, 1189–1200. [CrossRef][PubMed] 23. Tognon, G.; Hebestreit, A.; Lanfer, A.; Moreno, L.A.; Pala, V.; Siani, A.; Tornaritis, M.; De Henauw, S.; Veidebaum, T.; Molnár, D.; et al. Mediterranean diet, overweight and body composition in children from eight European countries: Cross-sectional and prospective results from the IDEFICS study. Nutr. Metab. Cardiovasc. Dis. 2013, 24, 205–213. [CrossRef][PubMed] 24. Sotos-Prieto, M.; Santos-Beneit, G.; Bodega, P.; Pocock, S.; Mattei, J.; Penalvo, J.L. Validation of a questionnaire to measure overall Mediterranean lifestyle habits for research application: The MEDiterranean LIFEstyle index (MEDLIFE). Nutr. Hosp. 2015, 32, 1153–1163. [PubMed] 25. Trichopoulou, A.; Kouris-Blazos, A.; Wahlqvist, M.L.; Gnardellis, C.; Lagiou, P.; Polychronopoulos, E.; Vassilakou, T.; Lipworth, L.; Trichopoulos, D. Diet and overall survival in elderly people. BMJ 1995, 311, 1457–1460. [CrossRef][PubMed] 26. Panagiotakos, D.B.; Milias, G.A.; Pitsavos, C.; Stefanadis, C. MedDietScore: A computer program that evaluates the adherence to the Mediterranean dietary pattern and its relation to cardiovascular disease risk. Comput. Methods Programs Biomed. 2006, 83, 73–77. [CrossRef][PubMed] 27. Rumawas, M.E.; Dwyer, J.T.; Mckeown, N.M.; Meigs, J.B.; Rogers, G.; Jacques, P.F. The development of the Mediterranean-style dietary pattern score and its application to the American diet in the Framingham Offspring Cohort. J. Nutr. 2009, 139, 1150–1156. [CrossRef] 28. Schroder, H.; Fito, M.; Estruch, R.; Martinez-Gonzalez, M.A.; Corella, D.; Salas-Salvado, J.; Lamuela-Raventos, R.; Ros, E.; Salaverria, I.; Fiol, M.; et al. A short screener is valid for assessing Mediterranean diet adherence among older Spanish men and women. J. Nutr. 2011, 141, 1140–1145. [CrossRef] 29. Cerwinske, L.A.; Rasmussen, H.E.; Lipson, S.; Volgman, A.S.; Tangney, C.C. Evaluation of a dietary screener: The Mediterranean Eating Pattern for Americans tool. J. Hum. Nutr. Diet. 2017, 30, 596–603. [CrossRef] 30. Alberti-Fidanza, A.; Fidanza, F. Mediterranean adequacy index of Italian diets. Public Health Nutr. 2004, 7, 937–941. [CrossRef] 31. Vitale, M.; Racca, E.; Izzo, A.; Giacco, A.; Parente, E.; Riccardi, G.; Giacco, R. Adherence to the traditional Mediterranean diet in a population of South of Italy: Factors involved and proposal of an educational ﬁeld-based survey tool. Int. J. Food Sci. Nutr. 2018.[CrossRef][PubMed] 32. Cade, J.; Thompson, R.; Burley, V.; Warm, D. Development, validation and utilisation of food-frequency questionnaires-A review. Public Health Nutr. 2002, 5, 567–587. [CrossRef][PubMed] 33. Peat, J. Health Science Research: A Handbook of Quantitative Methods, 1st ed.; Allen & Unwin: Crows Nest, Australia, 2001. 34. 45 and Up Study Collaborators. Cohort Proﬁle: The 45 and Up Study. Int. J. Epidemiol. 2008, 37, 941–947. [CrossRef][PubMed] 35. Heffernan, M.; Andrews, G.; Fiatarone Singh, M.A.; Valenzuela, M.; Anstey, K.; Maeder, A.; McNeil, J.J.; Jorm, L.; Lautenschlager, N.; Sachdev, P.; et al. Maintain Your Brain: Protocol of a 3-year randomised controlled trial of an individualised multi-modal eHealth intervention to prevent cognitive decline amongst community dwelling 55 to 77 year olds. J. Alzheimers Dis. 2018, in press. [CrossRef][PubMed] 36. Apple App Store. Research Food Diary Xyris Software (Australia) Pty Ltd. Available online: https://itunes. apple.com/au/app/research-food-diary/id1043440131 (accessed on 30 September 2018). 37. Ambrosini, G.L.; Hurworth, M.; Giglia, R.; Trapp, G.; Strauss, P. Feasibility of a commercial smartphone application for dietary assessment in epidemiological research and comparison with 24-h dietary recalls. Nutr. J. 2018, 17, 5. [CrossRef][PubMed] 38. Lieffers, J.R.L.; Hanning, R.M. Dietary assessment and self-monitoring: With nutrition applications for mobile devices. Can. J. Diet. Pract. Res. 2012, 73, e253–e260. [CrossRef][PubMed] 39. Ingwersen, L.; Raper, N.; Anand, J.; Moshfegh, A. Validation study shows importance of probing for forgotten foods during a dietary recall (abstract). J. Am. Diet. Assoc. 2004, 104, 13. [CrossRef] 40. NIH National Cancer Institute. Automated Self-Administered 24-hour (ASA24) Dietary Assessment Tool. Available online: https://epi.grants.cancer.gov/asa24/respondent/methodology.html (accessed on 2 August 2018).

263 Nutrients 2018, 10, 1913 17 of 17

41. Australian Bureau of Statistics. 4364.0.55.007-Australian Health Survey: Nutrition First Results-Foods and Nutrients, 2011–2012; FSANZ: Canberra, Australia, 2014. 42. Martínez-González, M.A.; García-Arellano, A.; Toledo, E.; Salas-Salvado, J.; Buil-Cosiales, P.; Corella, D.; Covas, M.I.; Schröder, H.; Arós, F.; Gómez-Gracia, E. A 14-item Mediterranean diet assessment tool and obesity indexes among high-risk subjects: The PREDIMED trial. PLoS ONE 2012, 7, e43134. [CrossRef] 43. Rosner, B. Fundamentals of Biostatistics, 8th ed.; Cengage Learning: Boston, MA, USA, 2015. 44. Bland, J.M.; Altman, D.G. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986, 1, 307–310. [CrossRef] 45. Landis, J.R.; Koch, G.G. The measurement of observer agreement for categorical data. Biometrics 1977, 33, 159–174. [CrossRef] 46. National Health and Medical Research Council. Clinical Practice Guidelines for the Management of Overweight and Obesity in Adults, Adolescents and Children in Australia; National Health and Medical Research Council: Melbourne, Australia, 2013. 47. Papadaki, A.; Johnson, L.; Toumpakari, Z.; England, C.; Rai, M.; Toms, S.; Penfold, C.; Zazpe, I.; Martínez-Gonzalez, M.A.; Feder, G. Validation of the English version of the 14-item Mediterranean Diet Adherence Screener of the PREDIMED study, in people at high cardiovascular risk in the UK. Nutrients 2018, 10, 138. [CrossRef][PubMed] 48. Schröder, H.; Marrugat, J.; Vila, J.; Covas, M.I.; Elosua, R. Adherence to the traditional Mediterranean diet is inversely associated with body mass index and obesity in a Spanish population. J. Nutr. 2004, 134, 3355–3361. [CrossRef][PubMed] 49. Toledo, E.; Salas-Salvadó, J.; Donat-Vargas, C.; Buil-Cosiales, P.; Estruch, R.; Ros, E.; Corella, D.; Fitó, M.; Hu, F.B.; Arós, F.; et al. Mediterranean diet and invasive breast cancer risk among women at high cardiovascular risk in the PREDIMED trial: Randomized clinical trial. JAMA. Intern. Med. 2015, 175, 1752–1760. [CrossRef][PubMed] 50. Valls-Pedret, C.; Sala-Vila, A.; Serra-Mir, M.; Corella, D.; de la Torre, R.; Martínez-González, M.Á.; Martínez-Lapiscina, E.H.; Fitó, M.; Pérez-Heras, A.; Salas-Salvadó, J. Mediterranean diet and age-related cognitive decline: A randomized clinical trial. JAMA. Intern. Med. 2015, 175, 1094–1103. [CrossRef][PubMed] 51. Australian Bureau of Statistics. National Health Survey: First Results, 2014-15-Australia, ‘Table 8.3 Body Mass Index, Waist Circumference, Height and Weight’, Time Series Spreadsheet, Cat. No. 4364.0.55.001. Available online: http://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/4364.0.55.0012014-15?OpenDocument (accessed on 19 September 2018). 52. Australian Bureau of Statistics. National Health Survey: First Results, 2014-15-Australia, ‘Table 3.3 Long-Term Health Conditions’, Time Series Spreadsheet, Cat. No. 4364.0.55.0013. Available online: http: //www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/4364.0.55.0012014-15?OpenDocument (accessed on 19 September 2018). 53. Thompson, F.E.; Subar, A.F. Chapter 1-Dietary Assessment Methodology. In Nutrition in the Prevention and Treatment of Disease, 4th ed.; Coulston, A.M., Boushey, C.J., Ferruzzi, M.G., Delahanty, L.M., Eds.; Academic Press: London, UK, 2017; pp. 5–48. 54. Ambrosini, G.L.; Van Roosbroeck, S.A.H.; Mackerras, D.; Fritschi, L.; De Klerk, N.H.; Musk, A.W. The reliability of ten-year dietary recall: Implications for cancer research. J. Nutr. 2003, 133, 2663–2668. [CrossRef][PubMed]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

264 SUPPLEMENTARY MATERIAL

265 Supplementary Material Table S1. - How to derive a MEDAS score from MediCul

Aim: To derive a Mediterranean Diet Adherence Screener (MEDAS) score out of 14 from the responses provided to the 50-item Mediterranean Diet and Culinary Index (MediCul) tool. Instructions: 1. Check the participant responses to the relevant MediCul question numbers listed in column three below. 2. Convert these responses to the relevant frequency for scoring, where required*. 3. Follow the instructions in columns four and five below to score each MEDAS question either 1 or 0 (zero) points; if a response cannot be awarded 1 point, it is automatically awarded a 0 (zero) point as there are no intermediate points. 4. Add up the points for the 14 MEDAS questions to provide the total MEDAS score. Note – some MEDAS questions used within MediCul have been optimised for the English language.

1. MEDAS question [1] 2. MEDAS 3. Relevant 4. MediCul response category 5. MEDAS scoring criteria for 1 MediCul instructions point Question/s [2]

1. Do you use olive oil as Yes Q. 29 This MediCul question is only used to This MediCul main culinary fat? score for MEDAS, hence no question is only used adjustment is required to score for MEDAS, hence no adjustment is required 2. How much olive oil do ≥ 4 TBSP/d Q. 28 First check that response to Q. 28 is Apply MEDAS you consume in a given given per day or convert to per day* criteria (column 2) day (including oil used and score for frying, salads, out-of- house meals, etc.)?

266 Supplementary Material Table S1. - cont.

1. MEDAS question [1] 2. MEDAS 3. Relevant 4. MediCul response category 5. MEDAS scoring criteria for 1 MediCul instructions point Question/s [2]

3. How many vegetable ≥ 2 (≥ 1 Q. 1 First check that response to Q. 1 is Take total MediCul servings do you consume portion raw given per day or convert to per day* serves/d and per day? [1 serving: 200 or as a salad) multiply by 75 g g (consider side dishes as then divide by 200 g; half a serving)] Apply MEDAS criteria (column 2) and score 4. How many fruit units  3 serve/d Q. 12 and Q. 34 First check that responses given in Q. Apply MEDAS (including natural fruit 12 and Q. 34 are expressed as per day criteria (column 2) juices) do you consume or convert to per day*; and score per day? Finally, add together responses from Q. 12 and Q. 34 5. How many servings of < 1 serve/d Q. 13 and Q. 14 First check that responses given in Q. Apply MEDAS red meat, hamburger or 13 and Q. 14 are expressed as per day criteria (column 2) meat products (ham, or convert to per day*; and score sausage, etc.) do you Add together responses for Q. 13 and consume per day? 14

267 Supplementary Material Table S1. - cont.

1. MEDAS question [1] 2. MEDAS 3. Relevant 4. MediCul response category 5. MEDAS scoring criteria for 1 MediCul instructions point Question/s [2]

6. How many servings of < 1 serve/d Q. 26 and Q. 27 First check that responses given in Q. Divide total for Q. butter, margarine, or 26 and Q. 27 are expressed as per day 26 & Q. 27 by 2 to cream do you consume or convert to per day*; correct for MEDAS per day? (1 serving: 12 g) Add together daily serves for Q. 26 serve size; and Q. 27 Apply MEDAS criteria (column 2) and score 7. How many < 1/d Q. 33 First check that response given in Q. Apply MEDAS sweet/carbonated 33 is expressed as per day or convert criteria (column 2) beverages do you drink to per day* and score per day? 8. How much wine do ≥ 7 Q. 41c First check that response given in Q. Apply MEDAS you drink per week? glasses/wk 41c is expressed as per week or criteria (column 2) convert to per week* and score 9. How many servings of ≥ 3/wk Q. 19 First check that response given in Q. Apply MEDAS legumes do you consume 19 is expressed as per week or convert criteria (column 2) per week? (1 serving: to per week* and score 150 g)

268 Supplementary Material Table S1. - cont.

1. MEDAS question [1] 2. MEDAS 3. Relevant 4. MediCul response category 5. MEDAS scoring criteria for 1 MediCul instructions point Question/s [2]

10. How many servings ≥ 3/wk Q. 16 First check that response given in Q. Apply MEDAS of fish or shellfish do 16 is expressed as per week or convert criteria (column 2) you consume per week? to per week* and score (1 serving: 100-150 g of fish or 4-5 units or 200 g shellfish) 11. How many times per < 3/wk Q. 31, Q. 32a ASSUMPTION: serves/wk for Apply MEDAS week do you consume cakes/biscuits = times/wk; criteria (column 2) commercial sweets or and score First check that response given in Q. pastries (not homemade), 31 is expressed as per week or such as cakes, cookies, convert to per week*; biscuits or custard? Add Q. 31 response + Q.32a response (custard only) 12. How many servings ≥ 3/wk Q. 24 First check that response given in Q. Apply MEDAS of nuts (including 24 is expressed as per week or convert criteria (column 2) peanuts) do you consume to per week* and score per week? (1 serving: 30 g)

269 Supplementary Material Table S1. - cont.

1. MEDAS question [1] 2. MEDAS 3. Relevant 4. MediCul response category 5. MEDAS scoring criteria for 1 MediCul instructions point Question/s [2]

13. Do you preferentially Yes Q. 17 Assume Yes = if ‘chicken, turkey or Apply MEDAS consume chicken, turkey rabbit’ OR ‘I don’t eat chicken or criteria (column 2) or rabbit meat instead of meat’ options ticked; and score veal, pork, hamburger or Assume No = if ‘beef, pork, sausage? hamburgers or sausages’ ticked 14. How many times per ≥ 2/wk Q. 6 First check that response given in Q. 6 Apply MEDAS week do you consume is expressed as per week or convert to criteria (column 2) vegetables, pasta, rice or per week* and score other dishes seasoned with sofrito (sauce made with tomato and onion, leek or garlic and simmered with olive oil)? Q, question; TBSP, tablespoons; d, day; g, grams; wk, week. *As the MediCul questions may include daily, weekly and monthly response options, some conversions are required to ensure responses relate to the same time period as in the MEDAS questions:  To convert responses given for serves/frequency per month to serves/frequency per week: divide by 30 then multiply by 7.  To convert responses given for serves/frequency per week to serves/frequency per day: divide by 7.  To convert responses given for serves/frequency per month to serves/frequency per day: divide by 30.

270

References for Supplementary Material Table S1

1. Schroder, H.; Fito, M.; Estruch, R.; Martinez-Gonzalez, M.A.; Corella, D.; Salas-Salvado, J.; Lamuela-Raventos, R.; Ros, E.; Salaverria, I.; Fiol, M., et al. A short screener is valid for assessing Mediterranean diet adherence among older Spanish men and women. J. Nutr. 2011, 141, 1140-1145, doi:10.3945/jn.110.135566. 2. Radd-Vagenas, S.; Fiatarone Singh, M.A.; Inskip, M.; Mavros, Y.; Gates, N.; Wilson, G.C.; Jain, N.; Meiklejohn, J.; Brodaty, H.; Wen, W., et al. Reliability and validity of a Mediterranean diet and culinary index (MediCul) tool in an older population with mild cognitive impairment. Br. J. Nutr. 2018, 120, 1189-1200, doi:10.1017/S0007114518002428.

271 Supplementary Material Table S2. - Kappa statistics for percent agreement within the same category of 17 MediCul groups between survey A and survey B

MediCul group Percent (%) Kappa value Level of agreement [1] agreement within same category 1. Olive oil 59.5% 0.47 (p<0.0001) Moderate 2. Vegetables 32.4% 0.21 (p<0.0001) Fair 3. Fruit 81.1% 0.63 (p<0.0001) Substantial 4. Nuts 64.9% 0.50 (p<0.0001) Moderate 5. Wholegrains 93.2% 0.77 (p<0.0001) Substantial 6. Legumes 67.6% 0.49 (p<0.0001) Moderate 7. Fish/shellfish 82.4% 0.69 (p<0.0001) Substantial 8. Eggs 90.5% 0.74 (p<0.0001) Substantial 9. Dairy products 86.5% 0.63 (p<0.0001) Substantial 10. White meat 70.3% 0.50 (p<0.0001) Moderate preference 11. Red/processed meat 63.5% 0.53 (p<0.0001) Moderate 12. Sweets & sugary 59.5% 0.51 (p<0.0001) Moderate drinks 13. Takeaway 82.4% 0.64 (p<0.0001) Substantial 14. Water 78.4% 0.71 (p<0.0001) Substantial 15. Alcohol 66.2% 0.58 (p<0.0001) Moderate 16. Coffee 100.0% 1.00 (p<0.0001) Almost perfect 17. Cuisine 39.2% 0.23 (p<0.0001) Fair survey A, first administration of MediCul tool; survey B, second administration of MediCul tool.

Reference for interpretation of Kappa values 1. Landis, J.R.; Koch, G.G. The measurement of observer agreement for categorical data. Biometrics 1977, 33, 159-174.

272 CHAPTER SIX

Reliability and validity of the Mediterranean Diet and

Culinary Index (MediCul) tool in an older population

with mild cognitive impairment

273 PREFACE

Diet research tools should be tested for reliability and concurrent validity against a reference dietary method within all populations where they are intended to be used.

The aim of this chapter was to test the reliability and validity of the newly developed Mediterranean Diet and Culinary Index (MediCul) against a 3-day food record, when administered in-person to older Australian adults with mild cognitive impairment, representing a Western population at high risk of dementia.

This chapter was published in the British Journal of Nutrition by Cambridge University Press (© The Authors 2018):

x Radd-Vagenas, S., Fiatarone Singh, M. A., Inskip, M., Mavros, Y., Gates, N., Wilson, G. C., . . . Flood, V. M. (2018). Reliability and validity of a Mediterranean diet and culinary index (MediCul) tool in an older population with mild cognitive impairment. British Journal of Nutrition, 120(10), 1189-1200. doi:10.1017/S0007114518002428

The chapter includes three published supplementary materials:

- MediCul tool elements, cut-off points, scoring and rationale - MediCul index tool - Kappa statistics for percent agreement within the same category of 17 MediCul elements between survey A and B

274 Further information related to this chapter is presented in Appendices Eight, and Ten to Thirteen and includes: ethics letter, dietitian assisted food record checklist, front sheet for validation and reliability testing of MediCul, 3-day paper food record template and FAQ on completing the 3-day food record.

Part of the work produced in this chapter has been presented at the following conference:

x Radd-Vagenas, S., Daniel, K., Noble, Y., Fiatarone Singh, M., & Flood, V. Reliability of new index tool to assess adherence to Mediterranean diet among people with mild cognitive impairment. Dietitians Association of Australia 35th National Conference, Sydney, Australia, 17-19th May 2018.

275 AUTHORSHIP STATEMENT

The co-authors of the paper: ‘Reliability and validity of a Mediterranean Diet and Culinary Index (MediCul) tool in an older population with mild cognitive impairment’ confirm that Sue Radd-Vagenas has made the following contributions:

As the primary supervisor for the candidature upon which this thesis is based, I can confirm that the above authorship attribution statement is true and correct.

Professor Victoria M Flood

Faculty of Health Sciences

The University of Sydney

25th February 2019

276 Reliability and validity of the Mediterranean Diet and Culinary Index (MediCul) tool in an older population with mild cognitive impairment

Sue Radd-Vagenas1, Maria A. Fiatarone Singh1,2, Michael Inskip1, Yorgi Mavros1,Nicola Gates3, Guy C. Wilson1, Nidhi Jain1, Jacinda Meiklejohn1, Henry Brodaty3,4, Wei Wen3,5, Nalin Singh1, Bernhard T. Baune6, Chao Suo7, Michael K. Baker1,8, Nasim Foroughi9, Perminder S. Sachdev3,4, Michael Valenzuela10, Victoria M. Flood1,11

1Physical Activity, Lifestyle, Ageing and Wellbeing Research Group, Faculty of Health Sciences, The University of Sydney, Lidcombe, NSW 2141, Australia

2Hebrew SeniorLife and Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, MA 02111, USA

3Centre for Healthy Brain Ageing (CHeBA), School of Psychiatry, University of New South Wales, Sydney, NSW 2031, Australia

4Dementia Centre for Research Collaboration, University of New South Wales, Sydney, NSW 2052, Australia

5Neuropsychiatric Institute, Prince of Wales Hospital, Randwick, NSW 2031, Australia

6Department of Psychiatry, Adelaide Medical School, University of Adelaide, Adelaide, SA 5000, Australia

7Brain and Mental Health Hub, Monash Institute of Cognitive and Clinical Neuroscience, School of Psychological Science, Monash University, Clayton, VIC 3800, Australia

8School of Exercise Science, Australian Catholic University, Strathfield, NSW 2135, Australia

9Clinical and Rehabilitation Research Group, Faculty of Health Sciences, The University of Sydney, Lidcombe, NSW 2141, Australia

10Regenerative Neuroscience Group, Brain and Mind Centre and Sydney Medical School, The University of Sydney, Camperdown, NSW 2050, Australia

11Western Sydney Local Health District, Westmead Hospital, Westmead, NSW 2145, Australia

277 Corresponding author

Prof Victoria M Flood

Faculty of Health Sciences

The University of Sydney

75 East Street, Lidcombe NSW 2141, Australia

Email: [email protected]

Tel: +61 2 9351 9001

Fax: +61 2 9351 9838

Disclaimers

Nil to declare.

Acknowledgements

The authors thank Kenneth Daniel and Yian Noble for assistance with operationalising MediCul scoring. The authors are grateful to Fiona O’Leary and Lynne Cobiac for comments on a draft of the MediCul index tool. The authors also acknowledge Adela Yip for help with data entry and greatly appreciate the involvement of all SMART participants.

Financial support

The original SMART trial was funded by a National Health and Medical Research Council (NHMRC) of Australia Dementia Research Grant, project grant ID No. 512672 from 2008-2011. Additional funding for a research assistant position was provided by NHMRC Program Grant ID No. 568969, and the project was supported by the University of Sydney and University of New South Wales. M. V. was supported by a University of New South Wales Vice Chancellor’s Fellowship and consecutive NHMRC Clinical Career Development Fellowships. The original trial fulfilled a portion of the degree requirements for PhD for N. G. and C. S. This validity study did not receive financial

278 support from any organisation and fulfilled a portion of the degree requirements for PhD for S. R-V.

Authorship

The authors’ contributions are as follows: S. R-V. contributed to study design, data collection, data analyses, interpretation of findings and wrote the manuscript; S. R-V. in conjunction with V. M. F, prepared the first draft of the MediCul tool; V. M. F. and M. A. F. S. contributed to study design, statistical analyses and interpretation of findings. M. A. F. S. and all other authors with the exception of S. R-V. and V. M. F. contributed to the funding acquisition, study design and/or data collection/management and statistical analysis for the original SMART trial and the follow-up assessments of participants. All authors provided critical review and approved the final version of the manuscript.

Conflict of interest

The authors declare no personal or financial conflicts of interest.

279 6.1 Abstract

Dementia is a leading cause of morbidity and mortality without pharmacologic prevention or cure. Mounting evidence suggests adherence to a Mediterranean dietary pattern may slow cognitive decline, and is important to characterise in at-risk cohorts. Thus, we determined the reliability and validity of Mediterranean Diet and Culinary Index (MediCul), a new tool, among community-dwelling individuals with mild cognitive impairment (MCI). Sixty-eight participants (66% female) aged 75.9 years (6.6), from the Study of Mental and Resistance Training study MCI cohort completed the 50-item MediCul at two time points, followed by a 3-day food record (FR). MediCul test-retest reliability was assessed using intraclass correlation coefficients (ICC), Bland-Altman plots and Kappa agreement within 17 dietary element categories. Validity was assessed against the FR using the Bland-Altman method and nutrient trends across MediCul score tertiles. The mean MediCul score was 54.6/100.0 with few participants reaching thresholds for key Mediterranean foods. MediCul had very good test-retest reliability (ICC=0.93, 95 CI% 0.884, 0.954, p<0.0001) with fair-to-almost-perfect agreement for classifying elements within the same category. Validity was moderate with no systematic bias between methods of measurement, according to the regression coefficient \ í30+0.17*x) (95% CI íP=0.091). MediCul overestimated the mean FR score by 6%, with limits of agreement (LOA) being under- and overestimated by 11% and 23%, respectively. Nutrient trends were significantly associated with increased MediCul scoring, consistent with a Mediterranean pattern. MediCul provides reliable and moderately valid information about Mediterranean diet adherence among older individuals with MCI, with potential application in future studies assessing relationships between diet and cognitive function.

280 6.2 Introduction

The Mediterranean diet has been associated with many health benefits such as a reduced risk of cardiovascular disease (CVD) (Estruch et al., 2018; Gardener et al., 2011; Gates et al., 2011; Martinez-Gonzalez & Bes-Rastrollo, 2014; Tong, Wareham, Khaw, Imamura, & Forouhi, 2016), cancer (Couto et al., 2011; Toledo et al., 2015), type 2 diabetes (Esposito, Maiorino, Ceriello, & Giugliano, 2010; Salas-Salvadó et al., 2014) and neurodegenerative diseases (Psaltopoulou et al., 2013; Sofi, Macchi, Abbate, Gensini, & Casini, 2014). The latter includes a slower rate of cognitive decline with age (Tangney et al., 2011), reduced risk of dementia (particularly Alzheimer’s disease (AD)) (Scarmeas, Stern, Mayeux, & Luchsinger, 2006) and reduced risk of mild cognitive impairment (MCI) and conversion of MCI to AD (Scarmeas et al., 2009). These findings are important since dementia is now a leading cause of morbidity and mortality globally with no pharmacologic options available to prevent, slow or reverse its course (World Health Organization, 2017).

More than 30 Mediterranean diet indexes and their variations (Bach et al., 2006; D'Alessandro & De Pergola, 2015; Milà-Villarroel et al., 2011; Radd-Vagenas, Kouris- Blazos, Fiatarone Singh, & Flood, 2017; Ruiz et al., 2015) (including short screeners) (Cerwinske, Rasmussen, Lipson, Volgman, & Tangney, 2017; Schroder et al., 2011) have been reported in the literature for use in assessing adherence to a Mediterranean dietary pattern. Indexes are popular as they can assess overall dietary patterns, while reducing participant and researcher burden associated with more classical methods of dietary measurement such as long food frequency questionnaires (FFQs) and weighed food records (Thompson & Subar, 2017).

However, limitations exist with the currently available Mediterranean diet indexes. For example, while the original and widely used Mediterranean Diet Score (MDS), and its many iterations (Trichopoulou et al., 1995; Trichopoulou, Costacou, Bamia, & Trichopoulos, 2003), includes elements determined a priori, the cut-off points used for this tool vary between the populations studied as they are related to mean or median intakes, which may not reflect ‘traditional’ Mediterranean or optimal intakes. Also,

281 relatively few Mediterranean diet indexes have been validated directly against an alternate dietary assessment method (Hebestreit et al., 2017; Schroder et al., 2011; Sotos- Prieto et al., 2015), especially one not limited by the same recall biases. Further, most Mediterranean diet index scores reported in the literature have been derived indirectly from FFQs (which may or may not be validated), then used to look for associations with health outcomes. Direct validation of dietary tools is now appreciated to be important to reliably interpret results (Freedman et al., 2015).

Importantly, to our knowledge, no existing Mediterranean diet index tools have been validated for use among individuals at various stages of cognitive decline, such as MCI. MCI is considered a pre-dementia stage, defined by subjective concern and mild objective cognitive changes without significant changes in daily functioning related to cognition (Petersen, 2004). In terms of prevention, MCI has been identified as a potential window of opportunity for lifestyle or other interventions since approximately 12% of individuals with MCI convert to AD per year, compared to an annual conversion rate of 1-2% in the general population (Petersen et al., 1999). It is therefore important to be able to accurately measure adherence to a Mediterranean dietary pattern in older adults who may have already begun to manifest memory difficulties, or are diagnosed with MCI. A Mediterranean diet index tool, which has been validated in such at-risk populations, would be useful for future interventions investigating the potential of nutrition to slow progression of cognitive decline.

A short Spanish Mediterranean Diet Adherence Screener (MEDAS) used in the largest randomised controlled trial of the Mediterranean diet, that is, PREvencion con DIeta MEDiterranea (PREDIMED) in cognitively normal but high CVD risk participants, is the only tool, to our knowledge, which has been associated with clinically demonstrated cognitive benefits. In a sub-cohort of these PREDIMED participants, their MEDAS score (out of 14) increased over a four-year period by approximately two points from a baseline of 8.3 and 8.6 in the groups supplemented with extra virgin olive oil and nuts, respectively, to 10.7 and 10.5 (Radd-Vagenas et al., 2017; Valls-Pedret et al., 2015). Also, an increase in MEDAS scoring has been associated with other benefits, such as reduced

282 risk of obesity and breast cancer (Martínez-González et al., 2012; Toledo et al., 2015). However, it is unclear exactly what the cut-off points in MEDAS mean for cognitive and other health outcomes. Also, the score interpretations for what is considered low, medium and high adherence to the diet in relation to studied outcomes, vary for this tool (Hernandez-Galiot & Goni, 2017; Martínez-González et al., 2012).

In summary, deficiencies in existing Mediterranean diet index tools may reduce their ability to predict health outcomes and guide lifestyle interventions. In addition, no tools have been developed for, and tested in, a cohort with pre-existing cognitive impairment at higher risk of conversion to dementia. Our aim was to test the reliability and validity of a more comprehensive, newly constructed, Mediterranean diet index tool including elements and cut-off points based on the ‘traditional’ Mediterranean diet, within a cohort of older people with MCI living in a non-Mediterranean country, in order to facilitate clinical research in various at-risk populations.

6.3 Methods

6.3.1 Participants

We recruited a convenience sample of community-dwelling participants from Sydney, Australia, who fulfilled MCI criteria (Gates et al., 2011; Petersen et al., 1999) from an existing clinical trial cohort (Singh et al., 2014), The Study of Mental and Resistance Training (SMART). The flow of participants from the original SMART trial into this validity study can be seen in Figure 6.1. SMART participants had been diagnosed with MCI but without dementia, and 100 were randomised between 2008 and 2011 to resistance training and/or cognitive training for six months, with follow-up at 18 months and then annually, where possible, to confirm their ongoing cognitive and health status (Singh et al., 2014). All SMART participants, except those who were deceased, dropped out, uncontactable, involved in piloting the Mediterranean Diet and Culinary Index (MediCul) tool or known to have reverted to normal cognition or progressed to dementia were invited to participate in this validity study during one of their annual re-assessment

283 visits, which occurred, on average, 78 months from the time they were originally recruited to SMART with the diagnosis of MCI.

6.3.2 Data administration and collection

The new Mediterranean diet index tool named MediCul was administered at the University clinic site as a paper survey (survey), twice, one week apart (time points A, B), with a dietitian observing and available to clarify questions (S. R-V.). The dietitian also checked responses to ensure no question was missed. Immediately following survey B, participants were instructed to keep a 3-day food record (FR) on any two weekdays and one weekend day within a seven day period, representing usual intake. They were asked to specify brands of foods/drinks, preparation methods and recipes, as well as to use the supplied Australian standard household measures (i.e., metric cups, spoons, jug), to estimate quantities. Participants were not required to weigh foods, although some elected to do so. Returned FRs were queried with the participant by the dietitian for potentially missed food categories using a checklist.

Anthropometric data was collected at time point A using calibrated digital scales and a wall-mounted stadiometer at the clinic (in light clothing and no shoes) or portable scales and stadiometer (UC-321PBT (A&D Company Limited) and Seca 213, respectively), if a home visit was required. Body mass index (BMI) was calculated by dividing weight (kg) by height (m2).

Additional participant characteristics including education level, marital status, number of chronic diseases, cognitive and physical function scores were sourced from original or follow-up SMART data, selecting the closest time point available for the entire cohort in our validity study. On average, this was 78 months prior to the validity study for education and marital status, and 59 months prior for number of chronic diseases, Alzheimer’s Disease Assessment Scale-Cognitive subscale (ADAS-Cog), Katz Activities of Daily Living (Katz ADL) and Bayer Informant Activities of Daily Living (Bayer-IADL) scores. ADAS-Cog (Rosen, Mohs, & Davis, 1984) and Bayer-IADL (Hindmarch, Lehfeld, de

284 Jongh, & Erzigkeit, 1998) were used as primary outcomes in SMART for global cognitive function and functional independence, respectively (Singh et al., 2014).

The study was approved by the Ethics Review Committee (RPAH Zone) of the Sydney Local Health District (Protocol No. X08-0064 & HREC/08/RPAH/106).

6.3.3 Tool development

The 50-item MediCul index tool was developed empirically in the form of a short question survey (See Supplementary Material 6.1 for elements, cut-off points, scoring and rationale) to a) reflect a ‘traditional’ Mediterranean dietary pattern and certain aspects of cuisine not assessed by previous tools (Bach-Faig et al., 2011; Oldways, 2008; Radd- Vagenas et al., 2017; Willett et al., 1995), b) include 14 questions from the validated MEDAS screener optimised for the English language (Schroder et al., 2011), and c) incorporate discretionary foods, commonly consumed in Western populations (Australian Bureau of Statistics, 2014).

After conducting a literature review to identify important Mediterranean dietary elements and existing tools (Radd-Vagenas et al., 2017) draft questions were developed in consultation with Mediterranean diet and survey tool experts. The tool was pilot-tested with five healthy people from the general public aged 50-80 and three SMART participants with MCI for readability, ambiguity and completion timing, which is 20 minutes on average, before being finalised.

MediCul includes a blend of frequency and serve questions spanning 17 main elements and assesses their exposure over the past six months: olive oil, vegetables, fruit, nuts, wholegrains, legumes, fish/shellfish, eggs, dairy products, white meat, red/processed meats, sweets and sugary drinks, takeaway, water, alcohol, coffee and certain aspects of Mediterranean cuisine. In all, nine of these elements cover desirable features of the ‘traditional’ Mediterranean diet and four cover undesirable features of a Western diet.

285 MediCul is scored from 0-100, with a higher score representing increased adherence to a ‘traditional’ dietary pattern (Supplementary Material 6.2).

6.3.4 Nutritional analysis

Scoring for both the MediCul and MEDAS tools was operationalised using Excel (MS Office Professional Plus 2013). The FRs were coded and entered into FoodWorks 8 Professional Edition: 8.0.3553 (Xyris Software Pty Ltd, Brisbane, Australia) selecting AusBrands 2015 and AusFoods 2015 data sources which map to the AUSNUT 2011-13 Food Standards Australia New Zealand (FSANZ) nutrient database for analysis by S. R- V. Average intakes for food group outputs were adjusted manually, where required, as FoodWorks draws on the concept of USDA Food Patterns Equivalents Database (FPED), which is sometimes contrary to current nutrition guidelines that also consider diet quality (e.g., hot chips are counted in the vegetable group). Missing foods that have become popular in recent times (e.g., paleo bread), were entered into FoodWorks using data from nutrition panels on packaging, and by basing such foods on similar products. Nutrient intakes from supplements were not included as the aim was to test validity of MediCul based on foods alone (Margetts & Nelson, 1997).

6.3.5 Statistical analysis

The Statistical Package for Social Sciences (SPSS) for Windows version 24 (SPSS Inc.) was used for all analyses. We aimed to have a minimum of 50 participants as recommended by Peat (2001) for adequate assessment of repeatability and agreement.

The distribution of MediCul scores, nutrients and food groups was examined for plausibility with the aid of histograms and by considering minimum and maximum values, to identify potential data entry errors. We did not use cut-offs for potential outlying values in reported dietary intake as we were specifically testing the tool among participants with MCI and did not want to subjectively exclude clinically conceivable answers for possible cognitive influences in reporting. As people with MCI are known to be clinically and neurologically heterogeneous (Livingston et al., 2017), a range of

286 cognitive performances, and thereby accuracy of recall for dietary intake within such a cohort, were therefore able to be captured.

A comparison was made of MediCul and the derived MEDAS scores, relevant to cognitive outcomes reported in the literature (Valls-Pedret et al., 2015). In addition, we estimated the percent of MCI participants that reached Mediterranean diet thresholds for selected foods/aspects of ‘traditional’ cuisine presumed to be health promoting, and those rarely used traditionally but consumed at significant levels in Western populations and known to be harmful at high or frequent levels of exposure. This was done using cut-off points for the highest score for relevant questions, from the MediCul index tool (Supplementary Material 6.1).

Reliability for MediCul across the two administrations, one week apart, was assessed using the intraclass correlation coefficient (ICC) and classified as poor (<0.40), fair to good 0.4 and <0.75), and YHU\JRRG (Rosner, 2015). A Bland-Altman plot was used to assess the level of agreement between survey A and B time points, since a high correlation does not necessarily mean good agreement (Bland & Altman, 1986). Kappa was also used to check percentage agreement within the same category for the 17 dietary elements. The Kappa values were characterised as showing almost perfect agreement (0.81-1.00), substantial agreement (0.61-0.80), moderate agreement (0.41-0.60), fair agreement (0.21-0.40), slight agreement (0.00-0.20) and poor agreement (<0.00) (Landis & Koch, 1977).

Validity was assessed by comparing MediCul scores derived from survey A versus MediCul scores from the FR using the Bland-Altman method. The differences between the two methods were plotted against the means of the methods, with limits of agreement (LOA) as two standard deviations (SD) above, and below, the mean difference. Linear regression analysis was used to indicate the direction of bias and whether it was constant across mean scores.

287 A total of three questions from MediCul relating to growing own vegetables, main meal eaten alone and fasting frequency, were unable to be validated as the FR did not include these details. We chose not to score for napping (traditionally conducted immediately after lunch in the Mediterranean), hence this question was also not validated. Finally, the validation of the MediCul index tool was based on scoring out of 97 whereas the reliability analysis was out of 100.

Indirect validity was investigated by examining if MediCul scores were associated with expected trends in nutrient intakes extracted from the FR. Nutrient values from the FR were checked for normal distribution using graphical methods and skewness, and log 10 transformed where positively/negatively skewed, then re-checked for normality to inform the statistical tests to be used. Normally distributed or normalised nutrients from the FR were compared across tertiles of MediCul score derived from both survey A and the FR using parametric tests (one-way ANOVA), whereas non-normally distributed nutrients were analysed using non-parametric tests (Kruskal-Wallis). When the ANOVA F ratio was significant, variances were checked for equality, and Bonferroni’s test was applied for equal variances or the Games-Howell post hoc test was applied for unequal variances. In addition, first (linear)- and second (quadratic)-order polynomial contrasts were applied to test for nutrient trends across tertiles, as well as the line of best fit.

Means and SD were calculated from FR values for normally distributed nutrients: kilojoules, protein, fat, fat as percentage energy, saturated fatty acids (SFA), SFA as percentage energy, SFA as percentage fat, polyunsaturated fatty acids (PUFA) as percentage fat, monounsaturated fatty acids (MUFA), MUFA as percentage fat, cholesterol, carbohydrate, carbohydrate as percentage energy, sugars, water, dietary fibre, vitamin C, vitamin A, beta carotene, sodium, potassium, magnesium, iron and zinc. Medians and the interquartile range, representing tertiles 1 and 3 of the MediCul score, were calculated for non-normally distributed nutrients: protein as percentage energy, PUFA, n-3 long chain (LC) PUFAĮ-linolenic acid (ALA), EPA, docosapentaenoic acid, DHA, ratio of MUFA to SFA, ratio of total unsaturated fatty acids to SFA, vitamin E, vitamin B12, total folate, selenium and alcohol.

288 Linear regression analysis was undertaken to assess precision of the MediCul tool across ADAS-Cog scores, being an index of global cognition (n=67).

6.4 Results

We recruited 68 participants from the 100 originally randomised to the SMART trial (Figure 6.1). All were included in the reliability study and 65 participated in the validity study. A total of two participants did not complete the FR and one had an incomplete FR. The majority of the recruited participants were female (65%), married/de facto (56%) and the primary cook at home (71%). On average, they were aged 75.9 years (SD=6.6), overweight (BMI=27.3 kg/m2; SD=5.2), well educated (13 years; SD=4), had 2.8 chronic diseases (SD=1.6), good physical function and a confirmed MCI diagnosis based on the most recently available ADAS-Cog scores (Table 6.1).

The mean MediCul score for survey A was 54.6/100.0 (SD=13.0; range: 32.5, 85.5) with 4.4% of participants VFRULQJ.0. The mean derived MEDAS score for survey A was 6.1/14.0 (SD=2.2; range: 1.0, 11.0). All those with 0('$6VFRUHVRI.0 (Valls-Pedret et al., 2015) had a MediCul score of 81.5.

Based on their FR, few participants reached thresholds for Mediterranean foods considered to be protective for cognitive or vascular function, such as olive oil (2%), legumes (3%), fruit (14%) and water (15%) (Figure 6.2). Fewer than half had minimal exposure to potentially harmful foods, used at low levels in the ‘traditional’ diet, such as processed meat (43%) and red meat (46%). However, three-quarters of the participants did report they mostly cooked their main meals at home and 89% kept sugary drinks to a minimum.

6.4.1 Reliability

The reliability of the MediCul index tool, based on single measures and 95% confidence intervals of the ICC, was very good (ICC=0.93, 95 CI% 0.884, 0.954, p<0.0001),

289 indicating the total score was measured similarly at the two time points (A and B). The Bland-Altman test for repeated measures showed a mean difference between the two scores of -0.04 with a lower LOA of í9.7 and an upper LOA of 9.6. Sixty-six of the 68 (97%) participants fell within or on the lower and upper LOA with a fairly even distribution across mean scores. There was also no indication of bias according to the regression coefficient (y=í0.79+0.01*x) (95% CI: í0.082, 0.109; p=0.778), supporting the null hypothesis that the scores at two time points were equally variable.

Groups that performed well for percentage agreement within the same category at time points A and B were as follows: wholegrains and coffee (almost perfect agreement); fruit, nuts, fish/shellfish, eggs, white meat preference, water and alcohol (substantial agreement); olive oil, dairy products, red/processed meats, sweets and sugary drinks (moderate agreement). Groups with fair agreement within the same category were: vegetables, legumes, takeaway and cuisine (Landis & Koch, 1977) (Supplementary Material 6.3). No groups had poor agreement for the proportion within each category at the two time points.

6.4.2 Validity

We assessed paired t-tests for scores from the survey administered at time points A, B and the mean of AB versus the FR (n=65). This analysis indicated a very similar mean difference and CI across the three comparisons, which were all significant (p<0.0001) (Table 6.2). We therefore used survey A time point for the remaining validity testing, given that this represented first time MediCul use, as could be applied in future research.

Bland-Altman analysis showed a positive mean difference of 6.0 between the MediCul score derived from survey A (52.8/97.0, SD=12.4, range: 31.5, 83.5) and the FR (46.8/97.0, SD=10.8, range: 21.5, 72.0). Scores for all but one participant fell within or on the 95% limits of agreement ZLWKDORZHU/2$RIí10.7 and an upper LOA of 22.6. There was also no significant linear trend for WKHILWWHGUHJUHVVLRQOLQH \ í2.30+0.17*x)

290 (95% CI: í0.027, 0.358; p=0.091), indicating no systematic bias between the two methods of measurement (Figure 6.3).

Significant linear trends in relation to tertiles of MediCul score were identified for the following nutrients, consistent with what would be expected for a Mediterranean dietary pattern: total fat (grams and %), including PUFA and MUFA (grams), which increased across tertiles of MediCul score from both the survey and FR (Table 6.3). This also translated into highly significant trends for ratios of MUFA to SFA and total unsaturated fatty acids to SFA (p<0.0001). Further, dietary fibre, vitamin C, vitamin E and magnesium all increased with increasing tertiles of MediCul score from both methods. Conversely, a significant trend for the reduction in carbohydrate as percent of energy was observed across tertiles of MediCul score from both methods. Protein was either unrelated (grams) or significantly decreased (percentage energy) when comparing tertiles from the survey (p=0.041), consistent with the fact that a Mediterranean diet is not a high protein diet. In some instances, there was a trend for nutrients by tertiles of scores from the FR but not the survey (and vice versa), e.g., n-3 LC PUFA. There was no trend observed for total folate. In all cases, linear trends were significant, except for sodium when compared to tertiles from the FR and sugars when compared to tertiles from survey A, where the quadratic trend provided a better fit for the data.

MediCul precision was stable across a range of cognition IURPQRUPDO WR0&, - 12)) using ADAS-Cog as a measure of global cognitive performance.

6.5 Discussion

MediCul is a short survey index tool (takes 20 minutes to complete, on average), developed to assess adherence to a ‘traditional’ Mediterranean dietary pattern and certain aspects of cuisine, within a Western population. It also assesses some cultural (fasting) and lifestyle factors (gardening, napping) that can impact on dietary intake or are uniquely related to consumption of the main (midday) meal. On the basis of our analyses, MediCul has very good reliability and moderate validity relative to a FR, among older individuals

291 with MCI. To our knowledge, this is the first Mediterranean diet index tool to be validated in a group at higher risk of dementia. Our results cannot be generalised to younger or cognitively unimpaired individuals without further testing. Also, our participants were originally volunteers for a randomised clinical trial and well educated, which may have influenced our results. Although we chose not to exclude participants based on extreme energy intakes, these were nevertheless all within plausible limits (Willett, 1998).

The MediCul index tool overestimated the mean total score compared to its reference method by 6%, and this was similar to the findings for MEDAS in Spanish (5%) and German (9%) cohorts (Hebestreit et al., 2017; Schroder et al., 2011). However, it is well- known that questionnaires tend to overestimate intakes compared with FRs (Klipstein- Grobusch et al., 1998). Although there was a considerable range for LOA from the Bland- Altman method when comparing scores from the MediCul index tool versus the FR, and it is unknown if this may have clinical implications, no systematic bias was found across mean scores. Hence, while the new index tool may be under and overestimating the FR- derived MediCul score by 11% and 23%, respectively, the Spanish cohort MEDAS scores were under and overestimated by 43% and 53%, respectively, compared with FFQ estimates (Schroder et al., 2011). Further, the under- and overestimates for the MediCul index tool are of a similar range reported for an alternate diet quality index score (Huybrechts et al., 2010), and well within limits proposed by Ambrosini et al. (2003) who classified agreement between a FFQ and FR as being acceptable when LOA were between 50% and 200%.

MediCul captures wide elements of the Mediterranean dietary pattern as a continuous measure. The cut-off points used and nutrient patterns identified suggest that diet quality may be improving with an increased MediCul score. For example, with increasing tertiles of the MediCul score there is a significant increase in healthy fats (and ratios of MUFA or total unsaturated fats to SFA), as well as dietary fibre, vitamin C and vitamin E, whereas carbohydrate as percentage energy declines correspondingly, and protein remains the same or decreases slightly. These directions are as anticipated for a ‘traditional’ Mediterranean diet, and macronutrient levels in the third tertile approximate

292 a Mediterranean diet model proposed in Australia (George et al., 2018). For example, the macronutrient proportions in the third tertile of MediCul score from the index tool were: fat, 41% of energy; protein, 15% of energy; carbohydrate, 36% of energy; MUFA, 47% of total fat. No trend was observed for total folate, probably a result of fortification in the Australian food supply, making interpretation of folate intakes difficult without additional and specific questions to assess this nutrient.

In our cohort of MCI participants, few reached Mediterranean diet thresholds for adequate intake of certain protective foods, such as olive oil, legumes, fruit and water and under half met our criterion for high vegetable variety, adequate fish intake and limited red/processed meat intake (Figure 6.2). The mean scores from the MediCul index tool and the derived MEDAS were also moderately low: 54.6/100.0 (SD=13.0) and 6.1/14.0 (SD=2.2), respectively. As a 0HGL&XOVFRUHRI.5 was equivalent to a MEDAS score RI.0, a level associated with cognitive benefit in the PREDIMED trial (Valls-Pedret et al., 2015), it is of concern that only 3/68 (4.4%) of older participants with MCI included in our study scored in this range. These findings suggest that individuals with MCI living in Western countries, even those well-educated, may not be optimally protected by a Mediterranean dietary pattern, which is recommended for chronic disease prevention by US (US Department of Health and Human Services and US Department of Agriculture, 2015) and Australian (National Health and Medical Research Council, 2013) dietary guidelines and the National Health Service (NHS, 2017) in the UK. Future studies, however, are required to determine the direction of this relationship, as reverse causality is possible.

6.5.1 Limitations

Individual diets are complex and tend to vary over time, making measurement errors inevitable for all dietary methods (Willett, 2001). The best methods for assessing populations at risk of dementia are yet to be elucidated (Bowman et al., 2011). The MediCul index tool relies on self-reported data that could bias our results, especially given the cohort investigated. Yet there is limited research on cognitive status impact on the integrity of self-reported dietary data (Zuniga & McAuley, 2015). One small study,

293 including MCI participants of a similar age to our participants, found cognitive impairment may inflate reliability and decrease validity of a FFQ (Bowman et al., 2011), which may also be the case with our findings. Further, MediCul has not yet been validated against disease risk factors and health outcomes or using biochemical measures of food intake, as has been reported for MEDAS (Díez-Espino et al., 2011; Hebestreit et al., 2017; Sánchez-Taínta et al., 2008; Schröder, Marrugat, Vila, Covas, & Elosua, 2004), and there has generally been limited use of biomarkers to investigate the relationship between diet and cognitive function (Zuniga & McAuley, 2015). However, this type of validation may be most relevant for assessment of absolute nutrient intakes rather than an index for an overall dietary pattern. While most FFQs solicit information about intake over the past year (Thompson & Subar, 2017) the MediCul index tool asks participants about their last six months, which together with specific questions relating to cooking methods for both warmer and cooler weather, may address some seasonal variation. However, this time period may still be problematic for information retrieval among individuals with MCI, although it has been reported that if the information recalled is considered inadequate, respondents rely on general knowledge of what they routinely eat (Zuniga & McAuley, 2015). We also had a dietitian present, available to answer questions and check responses were complete, limiting conclusions about other types of administrations or if the tool is entirely self-administered. Finally, the primary measure of global cognition (ADAS- Cog), assessed at the same time point for the whole cohort in our validity study, was taken, on average, 59 months earlier and it is possible that some participants may have reverted to normal cognition, inflating our results.

To reduce participant burden we required only a 3-day food record using household measures, which is not ideal for foods that are not consumed daily. However, 3- to 4-day records appear acceptable as it has been reported that the validity of collected information decreases in the latter days of a 7-day record with recording periods of more than four days thought to be unsatisfactory owing to fatigue/disinterest, creating reactivity bias (Thompson & Subar, 2017). In common with most other indexes, no energy adjustment was made for age or sex, which is unavoidable with tools designed for easy use. The FoodWorks nutritional analysis program has some limitations, with missing foods and categorisation used for some food groups, however, we adjusted for this manually.

294 FoodWorks also contains Australian compositional data but this is unlikely to vary in ways that would influence reliability and validity of MediCul for use in other countries.

6.5.2 Strengths

Small scale indexes such as screeners may not capture extreme levels of intakes, leading to overestimation of associations with health outcomes (Panagiotakos, Pitsavos, & Stefanadis, 2006). More comprehensive surveys may also have higher validity (Martinez- Gonzalez et al., 2002), although a ceiling of validity may exist (Willett, 2001). MediCul may be likened more to the larger scale modified MedDietScore index tool, which has scoring from 0-130 (Panagiotakos et al., 2009), yet it is relatively quick to complete and compute scoring for, compared to a typical FFQ. MediCul also measures some unique aspects of Mediterranean cuisine such as high moisture, lower temperature cooking methods; frequent use of herbs and spices; and exposure to fermented foods like olives. Such elements have been recommended to improve calculation of Mediterranean diet scores (Hoffman & Gerber, 2013). In its development, MediCul considered various best practice guidelines for dietary assessment (Cade, Burley, Warm, Thompson, & Margetts, 2004) now advised by The DIETary Assessment Tool NETwork (Cade et al., 2017). The fact that a MEDAS score can also be derived from MediCul improves its utility so that comparisons with different studies, for various outcomes, can also be made.

6.5.3 Conclusions

Preventing or slowing cognitive decline may have a significant impact on the lives of individuals, families and carers, as well as future public health budgets. The Mediterranean diet is a promising lifestyle modality based on current evidence. Accurate measurement of adherence to the ‘traditional’ dietary pattern, among at-risk individuals or those with existing cognitive impairment is vital to progress the field. We found that MediCul is a reliable and moderately valid tool to assess adherence to a Mediterranean dietary pattern among individuals with MCI who are at higher risk of converting to dementia. In our cohort of older Australians with MCI, the mean MediCul score was moderately low, suggesting poor compliance to this dietary pattern. MediCul may be a

295 useful tool for future studies testing a Mediterranean diet intervention for various stages of cognitive decline, including MCI.

296 6.6 References

Ambrosini, G. L., Van Roosbroeck, S. A. H., Mackerras, D., Fritschi, L., De Klerk, N. H., & Musk, A. W. (2003). The reliability of ten-year dietary recall: Implications for cancer research. Journal of Nutrition, 133(8), 2663-2668. doi:10.1093/jn/133.8.2663 Australian Bureau of Statistics. (2014). 4364.0.55.007 - Australian Health Survey: Nutrition first results - Foods and nutrients, 2011-12. Canberra, Australia: FSANZ. Bach-Faig, A., Berry, E. M., Lairon, D., Reguant, J., Trichopoulou, A., Dernini, S., . . . Mediterranean Diet Foundation Expert Group. (2011). Mediterranean diet pyramid today: Science and cultural updates. Public Health Nutrition, 14(12A), 2274-2284. doi:10.1017/S1368980011002515 Bach, A., Serra-Majem, L., Carrasco, J. L., Roman, B., Ngo, J., Bertomeu, I., & Obrador, B. (2006). The use of indexes evaluating the adherence to the Mediterranean diet in epidemiological studies: A review. Public Health Nutrition, 9(1a), 132-146. doi:10.1079/PHN2005936 Bland, J. M., & Altman, D. G. (1986). Statistical methods for assessing agreement between two methods of clinical measurement. The Lancet, 327(8476), 307-310. doi:10.1016/S0140-6736(86)90837-8 Bowman, G. L., Shannon, J., Ho, E., Traber, M. G., Frei, B., Oken, B. S., . . . Quinn, J. F. (2011). Reliability and validity of food frequency questionnaire and nutrient biomarkers in elders with and without mild cognitive impairment. Alzheimer Disease and Associated Disorders, 25(1), 49-57. doi:10.1097/WAD.0b013e3181f333d6 Cade, J. E., Burley, V. J., Warm, D. L., Thompson, R. L., & Margetts, B. M. (2004). Food-frequency questionnaires: a review of their design, validation and utilisation. Nutrition Research Reviews, 17(1), 5-22. doi:10.1079/NRR200370 Cade, J. E., Warthon-Medina, M., Albar, S., Alwan, N. A., Ness, A., Roe, M., . . . Consortium, D. N. (2017). DIET@NET: Best Practice Guidelines for dietary assessment in health research. BMC Medicine, 15(1), 202. doi:10.1186/s12916- 017-0962-x

297 Cerwinske, L. A., Rasmussen, H. E., Lipson, S., Volgman, A. S., & Tangney, C. C. (2017). Evaluation of a dietary screener: The Mediterranean Eating Pattern for Americans tool. Journal of Human Nutrition and Dietetics, 30(5), 596-603. doi:10.1111/jhn.12451 Couto, E., Boffetta, P., Lagiou, P., Ferrari, P., Buckland, G., Overvad, K., . . . Kariologi. (2011). Mediterranean dietary pattern and cancer risk in the EPIC cohort. British Journal of Cancer, 104(9), 1493-1499. doi:10.1038/bjc.2011.106 D'Alessandro, A., & De Pergola, G. (2015). Mediterranean diet and cardiovascular disease: A critical evaluation of a priori dietary indexes. Nutrients, 7(9), 7863- 7888. doi:10.3390/nu7095367 Díez-Espino, J., Buil-Cosiales, P., Serrano-Martínez, M., Toledo, E., Salas-Salvadó, J., & Martínez-González, M. Á. (2011). Adherence to the Mediterranean diet in patients with type 2 diabetes mellitus and HbA1c level. Annals of Nutrition and Metabolism, 58(1), 74-78. doi:10.1159/000324718 Esposito, K., Maiorino, M. I., Ceriello, A., & Giugliano, D. (2010). Prevention and control of type 2 diabetes by Mediterranean diet: A systematic review. Diabetes Research and Clinical Practice, 89(2), 97-102. doi:10.1016/j.diabres.2010.04.019 Estruch, R., Ros, E., Salas-Salvado, J., Covas, M. I., Corella, D., Aros, F., . . . PREDIMED Study Investigators. (2018). Primary prevention of cardiovascular disease with a Mediterranean diet supplemented with extra-virgin olive oil or nuts. New England Journal of Medicine, 378(25), e34. doi:10.1056/NEJMoa1800389 Freedman, L. S., Commins, J. M., Moler, J. E., Willett, W., Tinker, L. F., Subar, A. F., . . . Prentice, R. L. (2015). Pooled results from 5 validation studies of dietary self- report instruments using recovery biomarkers for potassium and sodium intake. American Journal of Epidemiology, 181(7), 473-487. doi:10.1093/aje/kwu325 Gardener, H., Wright, C. B., Gu, Y., Demmer, R. T., Boden-Albala, B., Elkind, M. S., . . . Scarmeas, N. (2011). Mediterranean-style diet and risk of ischemic stroke, myocardial infarction, and vascular death: The Northern Manhattan Study. American Journal of Clinical Nutrition, 94(6), 1458-1464. doi:10.3945/ajcn.111.012799

298 Gates, N. J., Valenzuela, M., Sachdev, P. S., Singh, N. A., Baune, B. T., Brodaty, H., . . . Singh, M. A. F. (2011). Study of Mental Activity and Regular Training (SMART) in at risk individuals: A randomised double blind, sham controlled, longitudinal trial. BMC Geriatrics, 11, 19. doi:10.1186/1471-2318-11-19 George, E. S., Kucianski, T., Mayr, H. L., Moschonis, G., Tierney, A. C., & Itsiopoulos, C. (2018). A Mediterranean diet model in Australia: Strategies for translating the traditional Mediterranean diet into a multicultural setting. Nutrients, 10(4), 465. doi:10.3390/nu10040465 Hebestreit, K., Yahiaoui-Doktor, M., Engel, C., Vetter, W., Siniatchkin, M., Erickson, N., . . . Bischoff, S. C. (2017). Validation of the German version of the Mediterranean Diet Adherence Screener (MEDAS) questionnaire. BMC Cancer, 17, 341. doi:10.1186/s12885-017-3337-y Hernandez-Galiot, A., & Goni, I. (2017). Adherence to the Mediterranean diet pattern, cognitive status and depressive symptoms in an elderly non-institutionalized population. Nutricion Hospitalaria, 34(2), 338-344. doi:10.20960/nh.360 Hindmarch, I., Lehfeld, H., de Jongh, P., & Erzigkeit, H. (1998). The Bayer Activities of Daily Living Scale (B-ADL). Dementia and Geriatric Cognitive Disorders, 9(Suppl 2), 20-26. doi:10.1159/000051195 Hoffman, R., & Gerber, M. (2013). Evaluating and adapting the Mediterranean diet for non-Mediterranean populations: A critical appraisal. Nutrition Reviews, 71(9), 573-584. doi:10.1111/nure.12040 Huybrechts, I., Vereecken, C., De Bacquer, D., Vandevijvere, S., Van Oyen, H., Maes, L., . . . De Henauw, S. (2010). Reproducibility and validity of a diet quality index for children assessed using a FFQ. British Journal of Nutrition, 104(1), 135-144. doi:10.1017/S0007114510000231 Klipstein-Grobusch, K., Den Breeijen, J. H., Goldbohm, R. A., Geleijnse, J. M., Hofman, A., Grobbee, D. E., & Witteman, J. C. M. (1998). Dietary assessment in the elderly: Validation of a semiquantitative food frequency questionnaire. European Journal of Clinical Nutrition, 52(8), 588-596. doi:10.1038/sj.ejcn.1600611 Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33(1), 159-174. doi:10.2307/2529310

299 Livingston, G., Sommerlad, A., Orgeta, V., Costafreda, S. G., Huntley, J., Ames, D., . . . Mukadam, N. (2017). Dementia prevention, intervention, and care. The Lancet, 390(10113), 2673-2734. doi:10.1016/S0140-6736(17)31363-6 Margetts, B. M., & Nelson, M. D. (1997). Design concepts in nutritional epidemiology (2nd ed.). New York, USA: Oxford University Press. Martinez-Gonzalez, M. A., Fernandez-Jarne, E., Serrano-Martinez, M., Marti, A., Martinez, J. A., & Martin-Moreno, J. M. (2002). Mediterranean diet and reduction in the risk of a first acute myocardial infarction: An operational healthy dietary score. European Journal of Nutrition, 41(4), 153-160. doi:10.1007/s00394-002-0370-6 Martinez-Gonzalez, M. A., & Bes-Rastrollo, M. (2014). Dietary patterns, Mediterranean diet, and cardiovascular disease. Current Opinion in Lipidology, 25(1), 20-26. doi:10.1097/MOL.0000000000000044 Martínez-González, M. A., García-Arellano, A., Toledo, E., Salas-Salvado, J., Buil- Cosiales, P., Corella, D., . . . Gómez-Gracia, E. (2012). A 14-item Mediterranean diet assessment tool and obesity indexes among high-risk subjects: The PREDIMED trial. PLoS ONE, 7(8), e43134. doi:10.1371/journal.pone.0043134 Milà-Villarroel, R., Bach-Faig, A., Puig, J., Puchal, A., Farran, A., Serra-Majem, L., & Carrasco, J. L. (2011). Comparison and evaluation of the reliability of indexes of adherence to the Mediterranean diet. Public Health Nutrition, 14(12A), 2338- 2345. doi:10.1017/S1368980011002606 National Health and Medical Research Council. (2013). Australian dietary guidelines. Canberra, Australia: National Health and Medical Research Council. NHS. (2017). What is a Mediterranean diet? Retrieved 21st December 2018 from https://www.nhs.uk/live-well/eat-well/what-is-a-mediterranean-diet/ Oldways. (2008). Mediterranean diet pyramid. Retrieved 18th May 2018 from http://oldwayspt.org/resources/heritage-pyramids/mediterranean- pyramid/overview

300 Panagiotakos, D., Kalogeropoulos, N., Pitsavos, C., Roussinou, G., Palliou, K., Chrysohoou, C., & Stefanadis, C. (2009). Validation of the MedDietScore via the determination of plasma fatty acids. International Journal of Food Sciences and Nutrition, 60(Suppl5), 168-180. doi:10.1080/09637480902810338 Panagiotakos, D. B., Pitsavos, C., & Stefanadis, C. (2006). Dietary patterns: A Mediterranean diet score and its relation to clinical and biological markers of cardiovascular disease risk. Nutrition, Metabolism and Cardiovascular Diseases, 16(8), 559-568. doi:10.1016/j.numecd.2005.08.006 Peat, J. (2001). Health science research: A handbook of quantitative methods (1st ed.). Crows Nest, Australia: Allen & Unwin. Petersen, R. C., Smith, G. E., Waring, S. C., Ivnik, R. J., Tangalos, E. G., & Kokmen, E. (1999). Mild cognitive impairment: Clinical characterization and outcome. Archives of Neurology, 56(3), 303-308. doi:10.1001/archneur.56.3.303 Petersen, R. C. (2004). Mild cognitive impairment as a diagnostic entity. Journal of Internal Medicine, 256(3), 183-194. doi:10.1111/j.1365-2796.2004.01388.x Psaltopoulou, T., Sergentanis, T. N., Panagiotakos, D. B., Sergentanis, I. N., Kosti, R., & Scarmeas, N. (2013). Mediterranean diet, stroke, cognitive impairment, and depression: A meta-analysis. Annals of Neurology, 74(4), 580-591. doi:10.1002/ana.23944 Radd-Vagenas, S., Kouris-Blazos, A., Fiatarone Singh, M., & Flood, V. M. (2017). Evolution of Mediterranean diets and cuisine: Concepts and definitions. Asia Pacific Journal of Clinical Nutrition, 26(5), 749-763. doi:10.6133/apjcn.082016.06 Rosen, W. G., Mohs, R. C., & Davis, K. L. (1984). A new rating scale for Alzheimer's disease. American Journal of Psychiatry, 141(11), 1356-1364. doi:10.1176/ajp.141.11.1356 Rosner, B. (2015). Fundamentals of biostatistics (8th ed.). Boston, USA: Cengage Learning. Ruiz, A. H., Garcia-Villanova, B., Hernandez, E. J. G., Amiano, P., Azpiri, M., & Montes, E. M. (2015). Description of indexes based on the adherence to the Mediterranean Dietary Pattern: A review. Nutricion Hospitalaria, 32(5), 1872- 1884. doi:10.3305/nh.2015.32.5.9629

301 Salas-Salvadó, J., Bulló, M., Estruch, R., Ros, E., Covas, M.-I., Ibarrola-Jurado, N., . . . Martínez-González, M. A. (2014). Prevention of diabetes with Mediterranean diets: A subgroup analysis of a randomized trial. Annals of Internal Medicine, 160(1), 1-10. doi:10.7326/M13-1725 Sánchez-Taínta, A., Estruch, R., Bullo, M., Corella, D., Gomez-Gracia, E., Fiol, M., . . . Zazpe, I. (2008). Adherence to a Mediterranean-type diet and reduced prevalence of clustered cardiovascular risk factors in a cohort of 3204 high-risk patients. European Journal of Cardiovascular Prevention and Rehabilitation, 15(5), 589-593. doi:10.1097/HJR.0b013e328308ba61 Scarmeas, N., Stern, Y., Mayeux, R., & Luchsinger, J. A. (2006). Mediterranean diet, Alzheimer disease, and vascular mediation. Archives of Neurology, 63(12), 1709-1717. doi:10.1001/archneur.63.12.noc60109 Scarmeas, N., Stern, Y., Mayeux, R., Manly, J. J., Schupf, N., & Luchsinger, J. A. (2009). Mediterranean diet and mild cognitive impairment. Archives of Neurology, 66(2), 216-225. doi:10.1001/archneurol.2008.536 Schroder, H., Fito, M., Estruch, R., Martinez-Gonzalez, M. A., Corella, D., Salas- Salvado, J., . . . Covas, M. I. (2011). A short screener is valid for assessing Mediterranean diet adherence among older Spanish men and women. Journal of Nutrition, 141(6), 1140-1145. doi:10.3945/jn.110.135566 Schröder, H., Marrugat, J., Vila, J., Covas, M. I., & Elosua, R. (2004). Adherence to the traditional Mediterranean diet is inversely associated with body mass index and obesity in a Spanish population. The Journal of Nutrition, 134(12), 3355-3361. doi:10.1093/jn/134.12.3355 Singh, M. A. F., Gates, N., Saigal, N., Wilson, G. C., Meiklejohn, J., Brodaty, H., . . . Suo, C. (2014). The Study of Mental and Resistance Training (SMART) study— resistance training and/or cognitive training in mild cognitive impairment: A randomized, double-blind, double-sham controlled trial. Journal of the American Medical Directors Association, 15(12), 873-880. doi:10.1016/j.jamda.2014.09.010

302 Sofi, F., Macchi, C., Abbate, R., Gensini, G. F., & Casini, A. (2014). Mediterranean diet and health status: An updated meta-analysis and a proposal for a literature-based adherence score. Public Health Nutrition, 17(12), 2769-2782. doi:10.1017/S1368980013003169 Sotos-Prieto, M., Santos-Beneit, G., Bodega, P., Pocock, S., Mattei, J., & Penalvo, J. L. (2015). Validation of a questionnaire to measure overall Mediterranean lifestyle habits for research application: The MEDiterranean LIFEstyle index (MEDLIFE). Nutricion Hospitalaria, 32(3), 1153-1163. doi:10.3305/nh.2015.32.3.9387 Tangney, C. C., Kwasny, M. J., Li, H., Wilson, R. S., Evans, D. A., & Morris, M. C. (2011). Adherence to a Mediterranean-type dietary pattern and cognitive decline in a community population. American Journal of Clinical Nutrition, 93(3), 601- 607. doi:10.3945/ajcn.110.007369 Thompson, F. E., & Subar, A. F. (2017). Chapter 1 - Dietary assessment methodology. In A. M. Coulston, C. J. Boushey, M. G. Ferruzzi, & L. M. Delahanty (Eds.), Nutrition in the prevention and treatment of disease (4th ed., pp. 5-48). London, UK: Academic Press. Toledo, E., Salas-Salvadó, J., Donat-Vargas, C., Buil-Cosiales, P., Estruch, R., Ros, E., . . . Martínez-González, M. A. (2015). Mediterranean diet and invasive breast cancer risk among women at high cardiovascular risk in the PREDIMED trial: Randomized clinical trial. JAMA Internal Medicine, 175(11), 1752-1760. doi:10.1001/jamainternmed.2015.4838 Tong, T. Y. N., Wareham, N. J., Khaw, K.-T., Imamura, F., & Forouhi, N. G. (2016). Prospective association of the Mediterranean diet with cardiovascular disease incidence and mortality and its population impact in a non-Mediterranean population: the EPIC-Norfolk study. BMC Medicine, 14(1), 135. doi:10.1186/s12916-016-0677-4 Trichopoulou, A., Kouris-Blazos, A., Wahlqvist, M. L., Gnardellis, C., Lagiou, P., Polychronopoulos, E., . . . Trichopoulos, D. (1995). Diet and overall survival in elderly people. British Medical Journal, 311(7018), 1457-1460. doi:10.1136/bmj.311.7018.1457

303 Trichopoulou, A., Costacou, T., Bamia, C., & Trichopoulos, D. (2003). Adherence to a Mediterranean diet and survival in a Greek population. New England Journal of Medicine, 348(26), 2599-2608. doi:10.1056/nejmoa025039 US Department of Health and Human Services and US Department of Agriculture. (2015). 2015-2020 Dietary guidelines for Americans, 8th Edition. Washington DC: US Government Printing Office. Valls-Pedret, C., Sala-Vila, A., Serra-Mir, M., Corella, D., de la Torre, R., Martínez- González, M. Á., . . . Salas-Salvadó, J. (2015). Mediterranean diet and age- related cognitive decline: A randomized clinical trial. JAMA Internal Medicine, 175(7), 1094-1103. doi:10.1001/jamainternmed.2015.1668 Willett, W. (1998). Nutritional epidemiology (2nd ed. Vol. 30). New York, USA: Oxford University Press. Willett, W. (2001). Invited commentary: A further look at dietary questionnaire validation. American Journal of Epidemiology, 154(12), 1100-1102. doi:10.1093/aje/154.12.1100 Willett, W. C., Sacks, F., Trichopoulou, A., Drescher, G., Ferro-Luzzi, A., Helsing, E., & Trichopoulos, D. (1995). Mediterranean diet pyramid: A cultural model for healthy eating. American Journal of Clinical Nutrition, 61(6 Suppl), 1402S- 1406S. doi:10.1093/ajcn/61.6.1402S World Health Organization. (2017). Rates of dementia. Retrieved 25th May 2018 from http://www.who.int/en/news-room/fact-sheets/detail/dementia Zuniga, K., & McAuley, E. (2015). Considerations in selection of diet assessment methods for examining the effect of nutrition on cognition. Journal of Nutrition Health & Aging, 19(3), 333-340. doi:10.1007/s12603-014-0566-5

304 List of tables and figures

Table 6.1. Characteristics of study participants Table 6.2. Mean difference from paired samples t-tests for Mediterranean diet and culinary index (MediCul) scores from surveys A, B, mean AB versus 3-day food record Table 6.3. Nutrient intakes compared with tertiles of Mediterranean diet and culinary index (MediCul) score Figure 6.1. Participant flow chart Figure 6.2. Percentage of participants who reached Mediterranean diet thresholds according to 3-day food records Figure 6.3. Bland-Altman plot of the difference between Mediterranean diet and culinary index (MediCul) score measured by survey A and 3-day food record (FR) and the mean MediCul score of the two methods

305 Table 6.1. Characteristics of study participants (n=68)

(Mean values and standard deviations; medians, ranges and percentages)

Mean SD

Age (years)* 75.9 6.6 Females (%) 64.7 -

BMI (kg/m2)* 27.3 5.2 Education level (years)† 13.2 3.7 Married/de facto (%)† 55.9 - Number of chronic diseases‡ 2.8 1.6

Primary cook at home (%)* 70.6 - ADAS-Cog score (0-70)‡ 5.2 2.5 ADAS-&RJ 0.0 - Katz ADL score (0-12)‡ Median 0.0 Range 0.0-0.0 Bayer-IADL score (1-10)‡§ Median 0.1 Range 0.0-0.2 Bayer-IADL >3 (%) 0.0

ADAS-Cog, Alzheimer’s Disease Assessment Scale-Cognitive subscale, used as the primary test for global cognitive function in Study of Mental and Resistance Training (SMART, higher scores indicate more impairment; cut-off for GHPHQWLDLV ; Katz ADL, Katz Index of Independence in Activities of Daily Living (higher scores indicate more impairment; cut-off for significant functional impairment is >0); Bayer-IADL, Bayer Informant Activities of Daily Living (higher scores indicate more impairment; cut- off for significant functional impairment is >3).

* Assessed at time point A of validity study.

306 † Assessed at baseline of SMART study (Singh et al., 2014), on average, 78 months before validity study.

‡ Assessed at 18 months of SMART study (Singh et al., 2014), on average, 59 months before validity study with n=67 as missing tests for one participant.

§ Participant score substituted for informant score for n=3.

307 Table 6.2. Mean difference from paired samples t-tests for Mediterranean diet and culinary index (MediCul) scores from surveys A, B, mean AB versus 3-day food record (FR) (n=65)

(Mean difference and 95 % confidence intervals)

Mean difference 95% CI P value

A versus FR 5.95 3.85, 8.05 <0.0001 B versus FR 6.29 3.95, 8.64 <0.0001 AB versus FR 6.12 3.97, 8.28 <0.0001

A, first administration of MediCul; B, second administration of MediCul; AB, mean of A and B MediCul administrations.

308 Table 6.3. Nutrient intakes compared with tertiles of Mediterranean diet and culinary index (MediCul) score (n=65)*

(Mean values and standard deviations; medians and interquartile ranges (IQR))

Energy (kJ/d) FR 8317 2362 8413 2286 8571 2355 0.973 0.722 Survey 8326 2632 8043 2079 8932 2167 0.442 0.400 Protein (g/d) FR 85 19 85 19 82 19 0.855 0.695 Survey 85 21 84 17 83 19 0.884 0.622 Protein (% energy) FR 0.516 0.310 Median 18 17 16 IQR 16-20 15-19 15-18 Survey 0.057 Ļ Median 18 18 15 IQR 15-20 15-20 14-18 Fat (g/d) FR 75 31 82 25 97 44 0.109 Ĺ Survey 75 33 79 30 100 38 0.041 Ĺ

309 Table 6.3. - cont.

Nutrients from food Source of Nutrient intake for tertile 1 Nutrient intake for tertile 2 of Nutrient intake for tertile 3 Comparison Test for record MediCul score of MediCul score MediCul score of MediCul score of nutrient trend P tertile cut-off intakes value, points across direction tertiles of ĹĻ MediCul score P Mean SD Mean SD Mean SD value Fat (% energy) FR 33§ 7 37 10 40 10 0.027 Ĺ Survey 32§ 7 36 8 41 12 0.008 Ĺ SFA (g/d) FR 32 15 27 11 28 13 0.377 0.292 Survey 32 16 26 11 30 12 0.418 0.757 SFA (% energy) FR 14 4 12 5 12 3 0.156 0.081 Survey 14 4 12 3 13 4 0.367 0.421 SFA (% fat) FR 47‡§ 8 36 7 33 7 <0.0001 Ļ Survey 46‡§ 10 37 7 34 6 <0.0001 Ļ PUFA (g/d) FR 0.001 Ĺ Median 9‡§ 14 17 IQR 7-14 11-15 11-25 Survey 0.006 Ĺ Median 11§ 11 14 IQR 6-15 9-15 11-20 PUFA (% fat) FR 15§ 5 19 6 21 6 0.003 Ĺ Survey 17 7 19 5 19 5 0.341 0.214

310 Table 6.3. - cont.

Nutrients from food Source of Nutrient intake for tertile 1 Nutrient intake for tertile 2 of Nutrient intake for tertile 3 Comparison Test for record MediCul score of MediCul score MediCul score of MediCul score of nutrient trend P tertile cut-off intakes value, points across direction tertiles of ĹĻ MediCul score P Mean SD Mean SD Mean SD value MUFA (g/d) FR 26§ 11 34 13 42 22 0.006 Ĺ Survey 25§ 12 32 15 43 19 0.001 Ĺ MUFA (% fat) FR 38‡§ 5 45 7 46 6 <0.0001 Ĺ Survey 37‡§ 5 44 6 47 6 <0.0001 Ĺ n-3 LC PUFA FR 0.075 Ĺ (mg/d) Median 133 270 469 IQR 83-379 153-887 215-1228 Survey 0.778 0.484 Median 159 379 286 IQR 109-600 76-1348 176-872 ALA (mg/d) FR 0.273 0.108 Median 1017 1202 1842 IQR 820-1866 1001-2230 1094-2382 Survey 0.151 0.082 Median 1074 1828 1492 IQR 863-1642 921-2524 1001-1854

311 Table 6.3. - cont.

312 Table 6.3. - cont.

313 Table 6.3. - cont.

Nutrients from food Source of Nutrient intake for tertile 1 Nutrient intake for tertile 2 of Nutrient intake for tertile 3 Comparison Test for record MediCul score of MediCul score MediCul score of MediCul score of nutrient trend P tertile cut-off intakes value, points across direction tertiles of ĹĻ MediCul score P Mean SD Mean SD Mean SD value Cholesterol (mg/d) FR 288 101 260 84 261 135 0.613 0.397 Survey 297 89 265 125 249 103 0.334 0.149 Carbohydrate (g/d) FR 211 59 203 96 182 46 0.376 0.176 Survey 215 68 183 58 200 83 0.329 0.459 Carbohydrate (% FR 42 6 38 11 36 8 0.057 Ļ energy) Survey 42 7 38 8 36 11 0.050 Ļ Sugars (g/d) FR 109 35 96 51 99 29 0.480 0.348 Survey 113 42 88 26 104 45 0.097 0.046 Alcohol (g/d) FR 0.505 0.294 Median 3.6 5.4 0.0 IQR 0.0-17.4 0.0-10.9 0.0-10.6 Survey 0.472 0.757 Median 0.0 6.6 0.5 IQR 0.0-18.1 0.0-16.6 0.0-8.0 Water (g/d) FR 2420 592 2547 560 2649 715 0.479 0.228 Survey 2412 665 2427 446 2766 687 0.108 0.063

314 Table 6.3. - cont.

Nutrients from food Source of Nutrient intake for tertile 1 Nutrient intake for tertile 2 of Nutrient intake for tertile 3 Comparison Test for record MediCul score of MediCul score MediCul score of MediCul score of nutrient trend P tertile cut-off intakes value, points across direction tertiles of ĹĻ MediCul score P Mean SD Mean SD Mean SD value Dietary fibre (g/d) FR 25§ 8 30 11 34 8 0.004 Ĺ Survey 24§ 7 30 8 35 11 0.001 Ĺ Vitamin C (mg/d) FR 106‡§ 57 160 85 170 57 0.004 Ĺ Survey 103§ 62 141 64 186 69 <0.001 Ĺ Vitamin E (mg/d) FR <0.0001 Ĺ 19 ۅ §Median 11 IQR 8-13 11-18 15-32 Survey <0.0001 Ĺ Median 11‡§ 15 16 IQR 8-12 10-21 14-24 9LWDPLQ% ȝJG FR 0.588 0.337 Median 4.3 4.1 4.1 IQR 3.5-5.2 3.1-4.9 3.4-5.2 Survey 0.768 0.588 Median 4.3 4.5 4.1 IQR 3.4-4.9 3.3-5.0 3.3-5.2

315 Table 6.3. - cont.

Nutrients from food Source of Nutrient intake for tertile 1 Nutrient intake for tertile 2 of Nutrient intake for tertile 3 Comparison Test for record MediCul score of MediCul score MediCul score of MediCul score of nutrient trend P tertile cut-off intakes value, points across direction tertiles of ĹĻ MediCul score P Mean SD Mean SD Mean SD value )RODWHWRWDO ȝJG FR 0.236 0.110 Median 665 575 595 IQR 576-797 498-757 459-665 Survey 0.496 0.395 Median 631 641 590 IQR 574-743 489-789 439-684 9LWDPLQ$ ȝJG FR 1184 785 1019 582 1242 370 0.484 0.803 Survey 1044 606 1249 723 1154 498 0.550 0.554 ȕ-&DURWHQH ȝJG FR 4473 3582 4521 3457 5955 2389 0.248 0.144 Survey 3957 2858 5496 3732 5401 2961 0.215 0.142 Na (mg/d) FR 2308 842 2656 1029 1994 677 0.056 0.032 Survey 2354 851 2403 859 2209 989 0.766 0.605 K (mg/d) FR 3128 662 3264 945 3516 800 0.285 0.120 Survey 3117 756 3128 605 3646 954 0.048 Ĺ Mg (mg/d) FR 332§ 90 358 103 449 160 0.006 Ĺ Survey 335 102 371 131 426 136 0.061 Ĺ

316 Table 6.3. - cont.

Nutrients from food Source of Nutrient intake for tertile 1 Nutrient intake for tertile 2 of Nutrient intake for tertile 3 Comparison Test for record MediCul score of MediCul score MediCul score of MediCul score of nutrient trend P tertile cut-off intakes value, points across direction tertiles of ĹĻ MediCul score P Mean SD Mean SD Mean SD value Fe (mg/d) FR 11.2 2.5 12.8 4.9 12.9 4.4 0.287 0.162 Survey 10.9 2.4 12.2 3.6 13.7 5.3 0.076 Ĺ Zn (mg/d) FR 10.7 3.1 10.6 3.5 11.2 4.1 0.850 0.647 Survey 10.8 3.3 10.5 3.6 11.3 3.7 0.770 0.664 6H ȝJG FR 0.120 Ĺ Median 72 89 91 IQR 57-89 68-97 69-124 Survey 0.504 0.771 Median 82 81 90 IQR 62-101 64-92 66-119 n, number of participants; SD, standard deviation; IQR, interquartile range; kJ, kilojoules; d, day; FR, food record; g, grams; %, percentage; SFA, saturated fatty acids; PUFA, polyunsaturated fatty acids; MUFA, monounsaturated fatty acids; n-3, omega 3; LC, long chain; mg, milligrams; ALA, alpha linolenic acid; EPA, eicosapentaenoic acid; DPA, docosapentaenoic acid; DHA, docosahexaenoic acid; mcg, micrograms.

* Tertiles are derived for MediCul index scores from both the FR and survey A (first administration of MediCul). For survey A, the cut-offs for tertiles 2 and 3 were 47.0 and 58.0, respectively. Values are presented as means and standard deviations for normally distributed data or medians and IQR for non-normally distributed data. These data were normalised by logarithmic transformation for use in ANOVA models with the exception

317 of alcohol and DHA, which were not able to be normalized and were therefore analysed using Kruskal-Wallis model. When ANOVA F ratio was significant, variances were checked for equality, and Bonferroni was applied for equal variances or Games-Howell post hoc t-test for unequal variances. First (linear)- and second-(quadratic) order polynomial contrasts were applied to test for trends across tertiles, as well as line of best fit.

In all cases linear trends were significant, and there were no significant deviations from normality, except for sodium when compared to tertiles from the FR and sugars when compared to tertiles from survey A, where the quadratic trend was positive and the most significant.

7UHQGGLUHFWLRQLQGLFDWHGDVĹ LQFUHDVLQJ RUĻ GHFUHDVLQJ

‡ Significant differences between tertiles 1 versus 2.

§ Significant differences between tertiles 1 versus 3.

.6LJQLILFDQWGLIIHUHQces between tertiles 2 versus 3ۅ

318 SMART recruitment pool n=2094 Ineligible n=56 On hold n=200 Not interested n=1582 No contact n=61

Assessed for eligibility n=195 Ineligible n=60 On hold n=17 Withdrawals n=17

Baseline assessment n=101 Ineligible n=1 (medical) Withdrawals n=0 On hold n=0

Randomised to SMART RCT n=100 Deceased n=2 Dementia n=7 Drop out n=12 Not MCI n=1 Pilot for tool n=3 Recruited to validity study n=68 No contact n=7

)LJXUH3DUWLFLSDQWIORZFKDUW7KH6WXG\RI0HQWDODQG5HVLVWDQFH7UDLQLQJ 60$57 IURP ZKLFKSDUWLFLSDQWVZHUHUHFUXLWHGWRWKH0HGL&XOYDOLGLW\VWXG\QQXPEHURISDUWLFLSDQWV5&7 UDQGRPLVHGFRQWUROOHGWULDO0&,PLOGFRJQLWLYHLPSDLUPHQW

319 Olive oil ≥4 TBSP/d 2% Legumes ≥3 serves/week 3% Sofrito ≥2 times/week 11% Fruit ≥3 serves/d 14% Water ≥5 cups/d 15% Lemon or vinegar in food prep ≥4 times/week 22% Vegetables ≥5 serves/d 23% Nuts ≥5 serves/week 29% High lutein vegetables ≥4 times/week 29% Biscuits & cakes <1 times/week 29% Vegetable variety ≥10 types/week 42% Fish or shellfish ≥3 serves/week 42% Processed meat <0.5 serves/week 43% Red meat ≤1 serve/week 46% Moist cooking methods ≥4 times/week 48% Herbs & spices ≥4 times/week 55% Snacking ≤2 times/d 58% Dairy products ≤2 serves/d 60% Eggs ≤4/week 71% Raw vegetables ≥4 times/week 72% Meals home cooked ≥5 times/week 74% Sugary drinks <1 cup/week 89% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

)LJXUH3HUFHQWDJHRISDUWLFLSDQWVZKRUHDFKHG0HGLWHUUDQHDQGLHWWKUHVKROGVDFFRUGLQJWRGD\IRRGUHFRUGV Q 7%63WDEOHVSRRQSUHSSUHSDUDWLRQ

320 30.0

Upper LOA=22.6

20.0 urvey A-FR) s

10.0

Mean=6.0

-10.0 Mean difference in MediCul score ( Lower LOA=-10.7

-20.0

20.0 30.0 40.0 50.0 60.0 70.0 80.0 Mean MediCul score [(survey A+FR)]

)LJXUH%ODQG$OWPDQSORWRIWKHGLIIHUHQFHEHWZHHQ0HGLWHUUDQHDQGLHWDQGFXOLQDU\ LQGH[ 0HGL&XO VFRUHPHDVXUHGE\VXUYH\$ ILUVWDGPLQLVWUDWLRQRI0HGL&XO DQGGD\ IRRGUHFRUG )5 DQGWKHPHDQ0HGL&XOVFRUHRIWKHWZRPHWKRGV Q 7KHVROLGOLQH LQWKHFHQWUHLQGLFDWHVWKHPHDQGLIIHUHQFHEHWZHHQWKHWZRPHWKRGVDQGWKHGRWWHGOLQHV DERYH DQG EHORZ LQGLFDWH WKH OLPLWV ZLWKLQ ZKLFK RI WKH GLIIHUHQFHV EHWZHHQ WKH PHWKRGV DUH H[SHFWHG WR IDOO 6' DERYH DQG EHORZ WKH PHDQ GLIIHUHQFH 7KH ILWWHG UHJUHVVLRQ OLQH LV \ í [ &, í S LQGLFDWLQJ QR V\VWHPDWLFELDV6'VWDQGDUGGHYLDWLRQ/2$OLPLWVRIDJUHHPHQW

321 SUPPLEMENTARY MATERIAL

322 Supplementary Material 6.1. - MediCul tool elements, cut-off points, scoring and rationale

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

1. Olive oil Q25 Fat types Days/wk 0 = No added 6 Olive oil was the primary culinary fat used in the frequency fats used ‘traditional’ Med diet (Radd-Vagenas, Kouris- 0 = Total other Blazos, Fiatarone Singh, & Flood, 2017). fats > olive oil The relative frequency of exposure to various fat 3 = Total other types is easier to rate than quantifying amounts. fats = olive oil This also avoids scoring for MediCul using the 6 = Total other leading MEDAS Q (Schroder et al., 2011) fats < olive oil Q26 Butter/cream Serves/d N/R N/R As below for Q27 (Zheng et al., 2016). intake Q included for MEDAS score only Q27 Margarine Serves/d N/R N/R Not scored in MediCul as difficult to accurately intake Q included for quantitate serves. MEDAS score For purposes of MEDAS scoring, MediCul defines only 1 serve=1 tsp (5g), and makes an adjustment (as MEDAS 1 serve=12g), as this amount is easier to visualise and more consistent with typical food portions consumed by Australian adults (5-7g margarine for 31-71+ years) (Zheng et al., 2016). Q28 Olive oil TBSP/d 0 = 0-1.99 6 High olive oil intake (i.e., 60g/d) is consistent with quantity (1 Australian TBSP/d ‘traditional’ Cretan use (e.g., 95g/d edible fats in 3 = 2-3.99 1960s; mostly olive oil) (Kromhout et al., 1989) and

323 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

TBSP = 15 g or TBSP/d meets or exceeds MEDAS cut-off SRLQW7%63G 20 ml) 7%63G (Kromhout et al., 1989). Note MEDAS 1 TBSP=13.5 g, hence 53 g/d required. Q29a Olive oil as Subjective N/R N/R Not scored in MediCul as leading Q. main culinary fat response as to Q included for Original MEDAS Q (Schroder et al., 2011), slightly whether olive MEDAS score improved for inclusion within MediCul by oil is main fat in only prompting to consider use of all fats/oils, including diet: no or yes spreads. Also, this Q follows others on various fat serves consumed, to help participant identify more correctly whether olive oil is the main culinary fat. Q29b Olive oil N/R 0 = lite, classic, 4 Traditionally, olive oil was all extra virgin, which type other contains polyphenols. Research shows polyphenols 4 = extra virgin are key to the efficacy of olive oil (Casamenti & Stefani, 2017; Cicerale, Conlan, Sinclair, & Keast, 2009). Many olive oils are refined and do not contain significant polyphenols (Aiello, Guccione, Accardi, & Caruso, 2015) so it is important to know which type is used. No other Med diet tools assess this, likely because in the past the role of the minor constituents of olive oil was underappreciated. 2. Vegetables Q1 Vegetable Serves/d 0 = 0-2.99 2 Based on validated Q in Australia for national intake serves/d monitoring. Measures all vegetables/salads 1 = 3-4.99 including potato, which was included in original

324 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

serves/d Med diet index tool as traditionally consumed VHUYHVG cooked as a vegetable (not fast food) (Trichopoulou et al., 1995). Consistent with AGTHE serves (National Health and Medical Research Council, 2013). AGTHE 1 VHUYH òFXS J KHQFHVHUYHVG J MEDAS cut-off SRLQWVHUYHVG ZKHUHVHUYH JKHQFHJG AGTHE recommends 375 g/d, which approximates 400 g/d cut-off point in MEDAS (National Health and Medical Research Council, 2013; Schroder et al., 2011). Q2 Vegetable Types/wk 0 = <10 1 Dietary guidelines promote increased dietary variety types/wk variety. Increased vegetable variety is associated with higher diet quality (Keim, Forester, Lyly, types/wk Aaron, & Townsend, 2014). Vegetable variety is also associated more strongly with reduced cancer risk than vegetable quantity (Jansen et al., 2004). Simply assessing vegetable quantity may therefore not fully capture the mechanisms whereby vegetables decrease disease risk. Keim et al. (2014) used cut-off SRLQWGLIIHUent vegetables/wk for high variety. CSIRO Healthy Diet Score includes multiple Qs on vegetable variety (Hendrie & Noakes, 2017). No other Med diet tools assess this.

325 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

Q3 Vegetable raw Times/wk 0 = 0-3.99 1 The ‘traditional’ Med diet regularly included raw frequency times/wk vegetables in the form of salads, e.g., cabbage WLPHVZN salad, village salad (mixed), dakos (grated tomato on rusks with EVOO). Hence frequency of raw vegetable exposure over most days per week is assessed. Research suggests raw vs. cooked vegetables are associated with better mental health and reduced mortality (Brookie, Best, & Conner, 2018; Masala et al., 2007). Ciccarone et al. (2003) in their Med tool used cut- off SRLQWIRUH[SRVXUHWRUDZYHJHWDEOHVRI times/wk. CSIRO Healthy Diet Score assesses salad vegetables (Hendrie & Noakes, 2017). Higher lutein Times/wk 0 = 0-1.99 2 Dark green leafy vegetables were consumed in the vegetable times/wk ‘traditional’ Med diet, mostly picked wild from the frequency: 1 = 2-3.99 mountains (Allbaugh, 1953; Keys, 1995). These Q4a Other green times/wk are rich in lutein, nitrate and vitamin K, which vegetables WLPHVZN provide vascular and cognitive benefits (Bolton- Q4b Dark green Smith, Price, Fenton, Harrington, & Shearer, 2000; leafy vegetables Lidder & Webb, 2013; Mohn & Johnson, 2017). Higher plasma lutein levels and green leafy vegetable intake is associated with decreased dementia risk, slowing of cognitive decline and age- related macular degeneration (Morris et al., 2015a; Morris et al., 2015b; Tan et al., 2008).

326 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

Two questions are asked about different types of green vegetables to minimise response bias. Cut- off point relates to frequency of exposure for most days per week, which is simpler to estimate than quantity consumed. MIND study (Morris et al., 2015a) cut-off SRLQWLV 6 serves/wk. No other Med diet tools assess this except Italian Mediterranean Index (IMI) (Agnoli et al., 2011). Q5 Onions/garlic Times/wk 0 = 0-3.99 1 The allium family of vegetables is regularly frequency times/wk consumed (mostly daily) in the ‘traditional’ Med WLPHVZN diet in salads and cooked dishes and contributes to polyphenol intake (Fidanza & Alberti, 2005). Cut- off point relates to frequency of exposure for most days per week, which is simpler to estimate than quantity consumed. No other Med diet tools assess this except Italian Mediterranean Index (IMI) (Agnoli et al., 2011). Q8 Grow own N/R 0 = No 1 Having a home garden was traditional and is vegetables 1 = Yes positively correlated with a high Med diet score in the Australian arm of the MEDIS Study (Darmos- Thodis, 2013). Growing any vegetables may promote an increased intake, physical activity and higher vitamin D levels (Kouris-Blazos & Itsiopoulos, 2014). No other Med diet tools assess this.

327 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

3. Fruit Q12 Fruit intake Serves/d 0 = 0-0.99 6 Q based on standard Q from national surveys in serves/d Australia. Examples of dried fruit also provided 3 = 1-2.99 since consumed in the Mediterranean (Kromhout et serves/d al., 1989). VHUYHVG Cut-off point is consistent with MEDAS (Schroder et al., 2011) G EXW0('$6DOVRcounts fruit juice), Greek dietary guidelines (Ministry of Health and Welfare: Supreme Scientific Health Council, 1999) and high traditional (1960s) Cretan intake of 464 g/d (Kromhout et al., 1989). Higher fruit intake is associated with reduced stroke risk (Dauchet, Amouyel, & Dallongeville, 2005; Hu, Huang, Wang, Zhang, & Qu, 2014; Mizrahi et al., 2009). MediCul cut-off point exceeds 2 serves/d recommended by AGTHE (National Health and Medical Research Council, 2013). MediCul excludes juice as this increases the glycaemic load, thereby providing different effects to whole fruit. Also, juice was not regularly consumed in the ‘traditional’ Med diet. 4. Nuts Q24 Nuts intake Serves/wk 0 = 0-0.99 6 Nuts were regularly consumed in the ‘traditional’ (1 serve = 30g serves/wk diet (Allbaugh, 1953). Observational studies show or 1.5 TBSP 2 = 1-2.99 greater CVD proWHFWLRQZLWKH[SRVXUHWRQXWV nut/seed butter) serves/wk times/wk (Aune et al., 2016; Sabaté & Ang, 2009). 4 = 3-4.99 Clinical and observational studies show improved serves/wk cognition or reduced cognitive decline with age

328 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

with nut consumption (Barbour, Howe, Buckley, serves/wk Bryan, & Coates, 2014; O'Brien et al., 2014; Pribis et al., 2011). MediCul uses wider range of cut-off points then MEDAS for increased discernment. MEDAS (Schroder et al., 2011) cut-off SRLQW serves/wk. Goulet, Nadeau, and Lemieux (2003) cut-off point >2 portions/d is for legumes and nuts/seeds combined. 5. Wholegrains Q20 Frequency 0 = No bread 6 Grain foods were an important part of the Wholegrain/lower used ‘traditional’ Med diet and mostly unrefined (intact, GI bread frequency 0 = White > cracked or coarsely milled) (Trichopoulou & (wholegrain + Lagiou, 1997). Bread was made using a sourdough wholemeal + rye culture (Pes et al., 2015). Such forms result in a + sourdough) lower GI, which is associated with health benefits 3 = White = (Barclay et al., 2008; Brand-Miller, Hayne, Petocz, (wholegrain + & Colagiuri, 2003; Goff, Cowland, Hooper, & wholemeal + rye Frost, 2013). + sourdough) This Q uses bread as a proxy for quality of grains 6 = White < most frequently consumed. (wholegrain + MIND diet (Morris et al., 2015a) uses cut-off point wholemeal + rye VHUYHVGEXWVHUYHVIRUGLIIHUHQWJUDLQIRRGVYDU\ + sourdough) and are difficult to estimate, hence not used in MediCul. MEDLIFE criteria (Sotos-Prieto et al., 2014) for

329 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

wholegrains = fibre >25 g/d, which requires nutritional calculations. 6. Legumes Q19 Legume Serves/wk 0 = 0-0.99 6 Legumes were a core food in ‘traditional’ peasant intake (1 serve = 1 cup serves/wk Med diets in the 19th century, used as a meat cooked dry 3 = 1-2.99 replacement (Matalas, 2006; Ministry of Health and beans or 150g) serves/wk Welfare: Supreme Scientific Health Council, 1999). MediCul cut-off point is consistent with MEDAS serves/wk (Schroder et al., 2011) VHUYHVZNDQGVHUYHVL]H The AGTHE protein serve (1 cup cooked), rather than vegetable serve (1/2 cup cooked), is used as more consistent with ‘traditional’ food use (National Health and Medical Research Council, 2013). GLNC recommends legumes are consumed 2-3 times/wk (Grains & Legumes Nutrition Council, 2013). 7. Fish/shellfish Q16 Fish/shellfish Serves/wk 0 = 0-0.99 6 Fish was part of the ‘traditional’ Med diet but intake (1 serve= 1 serves/wk intake depended on proximity to the sea (Radd- small fish fillet 3 = 1-2.99 Vagenas et al., 2017). Research shows fish may or 1 small can serves/wk contribute to brain structure benefits and of fish or 200 g consumption is associated with reduced risk of shellfish) serves/wk AD/dementia (Barberger-Gateau et al., 2007; Gu et al., 2015). MediCul is consistent with MEDAS (Schroder et al., 2011) serves and cut-off SRLQWVHUYHVZN The Heart Foundation recommends 2-3 serves oily fish/wk for all Australians in their position

330 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

statement on fish and CV health (Heart Foundation, 2015). 8. Eggs Q18 Egg intake Eggs/wk HJJVZN 2 Eggs were consumed in moderation in the 1 = >4-6.99 ‘traditional’ Med diet (Radd-Vagenas et al., 2017). eggs/wk Some studies have associated egg intake with a 2 = 0-4 eggs/wk higher risk of type 2 diabetes, and stroke in those with existing type 2 diabetes (Djoussé & Gaziano, 2008; Djoussé, Khawaja, & Gaziano, 2016). MediCul cut-off point is based on the original Med diet pyramid (Willett et al., 1995). Goulet et al. (2003) cut-off point 0-4 eggs/wk with D]HURVFRUHIRUHJJVZN Ciccarone et al. (2003) cut-off point 0-2 eggs/wk. MEDLIFE (Sotos-Prieto et al., 2014) cut-off point 2-4 eggs/wk. Spanish Med pyramid cut-off point 2-4/wk (Bach- Faig et al., 2011). Eggs are not commonly assessed by Med diet index tools. 9. Dairy products Q21 Dairy product Serves/d 0 = >2 serves/d 2 Dairy intake was moderately low in the ‘traditional’ intake (1 serve = half a 2 = 0-2 serves/d diet and did not focus on low fat products, popular cup (120 g) in Western countries (Ministry of Health and ricotta/cottage Welfare: Supreme Scientific Health Council, 1999; cheese, 50 g Pes et al., 2015; Willett et al., 1995). fetta, ¾ cup MediCul cut-off point relates to maximum exposure (200 g) yoghurt, for total dairy of any type, i.e., cow, sheep, goat, 2 slices (40 g) full fat, low fat, reduced fat.

331 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

hard yellow Serve size for fetta is provided based on calcium cheese) equivalence. Q22 Milk type Response to N/R N/R Not used in MediCul scoring. which type of Q included for milk usually qualitative consumed purpose only 10. White meat Q15 White meat Serves/wk 0 = Q13 + Q14 2 Exposure to meat was low in the ‘traditional’ Med calculated serves/wk > diet (Radd-Vagenas et al., 2017). Red meat in preference Q15 + Q16 particular is associated with chronic disease risk, serves/wk and may be related to risk of dementia (Bellavia, 1 = Q13 + Q14 Stilling, & Wolk, 2016; Giem, Beeson, & Fraser, serves/wk = 1993). Q15 + Q16 MEDAS (Schroder et al., 2011) includes a leading serves/wk Q to assess white meat preference. While that 2 = Q15 + Q16 MEDAS Q is embedded within MediCul (Q17) for serves/wk > the purpose of calculating a MEDAS score, this Q13 + Q14 particular MediCul Q is designed to assess whether serves/wk OR serves of white meat are consumed more frequently Q13 + Q14 + than serves of red meat. MediCul does not score Q15 + Q16 for number of chicken serves. serves/wk = 0 (vegetarian) Q17 White meat Subjective N/R N/R Included to derive MEDAS (Schroder et al., 2011) subjective response as to Q included for score, but not scored in MediCul as leading Q. preference which is MEDAS score preferentially only

332 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

eaten: white, red or no meats 11. Q13 Red meat Serves/wk 0 = >3 3 Exposure to red meat was low in the ‘traditional’ Red/processed intake (1 serve= 100- serves/wk Med diet (Martinez-Gonzalez, Hershey, Zazpe, & 150 g) 1.5 = >1-3 Trichopoulou, 2017; Ministry of Health and serves/wk Welfare: Supreme Scientific Health Council, 1999). 3 = 0-1 serve/wk Red meat in particular is associated with chronic disease risk, and may be related to risk of dementia (Bellavia et al., 2016; Giem et al., 1993). This Q is based on MEDAS (Schroder et al., 2011) and the original Willett Med pyramid (Willett et al., 1995). However, serves are based on commonly consumed portions in Australia rather than 65 g as defined in AGTHE (Zheng et al., 2016). MedDietScore (Panagiotakos et al., 2009) uses cut- off SRLQWZN MEDLIFE (Sotos-Prieto et al., 2014) uses cut-off point <2/wk. MIND diet (Morris et al., 2015a) allows higher exposure to red meat and meat products <4 meals/wk. AGTHE (National Health and Medical Research Council, 2013) recommends not > 455 g/wk. Q14 Processed Serves/wk 0 = >1 serve/wk 3 Exposure to processed meat was low in the meat intake (1 serve= 1 ½ 1.5 = 0.5-1 ‘traditional’ Med diet (Radd-Vagenas et al., 2017). thick or 2 serve/wk Processed meat is strongly associated with chronic thinner disease risk and mortality (Larsson & Orsini, 2014).

333 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

sausages; 2 3 = 0-<0.5 Q is based on MEDAS (Schroder et al., 2011) Q. rashers bacon; 4 serve/wk However, cut-off point is consistent with slices processed MEDLIFE (Sotos-Prieto et al., 2014) VHUYHZN meats (100 g); and Ciccarone et al. (2003) 0 times/wk, which is 1 meat pie, closer to traditional intakes. This Q uses a similar pastie, sausage serve size to that for fresh red meat so responses roll; 6 chicken can be summed for MEDAS scoring. nuggets) MEDAS cut-off point <7 serves/wk. Goulet et al. (2003) cut-off point <1 portion (50- 100g)/wk. MIND diet (Morris et al., 2015a) allows higher exposure to red meat and meat products <4 meals/wk. 12. Sweets & Q31 Biscuits/cakes Times/wk WLPHVZN 2 Sugar containing foods were rare in the ‘traditional’ sugary drinks frequency 1 = 1-1.99 Med diet (Radd-Vagenas et al., 2017; Willett et al., times/wk 1995). High sugar foods can increase insulin 2 = 0-0.99 resistance, which is associated with significantly times/wk lower regional cerebral glucose metabolism and Alzheimer’s disease (Willette et al., 2015). This Q is based on MEDAS (Schroder et al., 2011) Q but includes examples of more typical sweets in Australia and includes all sweets, not just ‘commercial’ types as in MEDAS. It excludes custard. MEDLIFE (Sotos-Prieto et al., 2014) cut-off point VHUYHVZNIRUDOOVZHHWV

334 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

MEDAS cut-off point <3 times/wk for all sweets/pastries. MediCul scoring takes into account the effect of added sugars from additional sources assessed in other relevant Qs, i.e., Q32b, Q33. Q32a Custard Times/wk N/R N/R This is not scored in MediCul but included to obtain frequency Q included for a MEDAS (Schroder et al., 2011) score for MEDAS score sweets/pastries. MEDAS includes custard as part of only sweets/pastries. Q32b Ice cream Times/wk WLPHVZN 2 Ice cream was not part of the ‘traditional’ Med diet frequency 1= 1-1.99 but is a popular dessert or snack in Australia times/wk (Willett et al., 1995). 2 = 0-0.99 MediCul scoring takes into account the effect of times/wk added sugars from additional sources assessed in other relevant Qs, i.e., Q31, Q33. No other Med diet tools assess this. Q33 Sugary drinks Cups/wk FXSVZN 2 Sugary drinks were rare in the ‘traditional’ Med diet intake 1 = 1-1.99 (Radd-Vagenas et al., 2017). They can increase cups/wk insulin resistance, which is associated with 2 = 0-0.99 significantly lower regional cerebral glucose cups/wk metabolism and Alzheimer’s disease (Willette et al., 2015). Q based on MEDAS (Schroder et al., 2011) Q but provides standardised serves, whereas MEDAS simply asks re number of sugary drinks consumed per day.

335 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

MediCul scoring takes into account the effect of added sugars from additional sources assessed in other relevant Qs, i.e., Q31, Q32b. Q34 Fruit juice Cups/wk 0 = >7 cups/wk 2 Intake of fruit juice was rare in the ‘traditional’ intake 2 = 0-7 cups/wk Med diet, although some fruit syrups were homemade to offer to guests, diluted with water (Radd- Vagenas et al., 2017). Q is included to assess excessive exposure to fruit juice which, like other sugary drinks, may increase the glycaemic load and promote weight gain and insulin resistance (Sharma, Chung, Kim, & Hong, 2016; Xi et al., 2014). Q also allows fruit juice to be counted in total serves of fruit to derive a MEDAS (Schroder et al., 2011) score. Q35 Fruit juice Response to N/R N/R This Q is not scored in MediCul. type type usually Q included for Type required for MEDAS scoring only, since consumed: MEDAS score ‘natural fruit juices’ counted as fruit. Natural fruit commercial or only juices assumed to be freshly squeezed juices. freshly squeezed 13. Takeaway Q7 Hot chips Times/wk WLPHVZN 2 Hot chips were not part of the ‘traditional’ Med diet frequency 1 = 1-1.99 but are common in Western diets. Hot chips may times/wk be a proxy for takeaway (fast food) consumption. 2 = 0-0.99 Unlike boiled potato, fried potato contains times/wk significant acrylamide and AGEs content,

336 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

associated with inflammation and negative health outcomes, such as cancer and neurotoxicity (Cai et al., 2014; Kellow & Coughlan, 2015; Shojaee- Aliabadi et al., 2013; Uribarri et al., 2010). Previous Med diet tools have grouped all potato products either with vegetables or grains. No other Med diet tools assess this. Q30 Takeaway Times/wk WLPHVZN 2 The ‘traditional’ Med diet did not include takeaway frequency 1 = 1-1.99 (fast food) yet this is common in modern society times/wk and associated with negative health outcomes 2 = 0-0.99 (Cahill et al., 2014; Martinez-Gonzalez et al., times/wk 2017). MIND diet cut-off point <1 times/wk for fast fried foods (Morris et al., 2015a). 14. Water Q36 Pure water Cups/d 0 = 0-2.99 4 Water was the main drink consumed traditionally intake cups/d (Radd-Vagenas et al., 2017). Adequate intake is 2 = 3-4.99 associated with reduced risk of fatal coronary heart cups/d disease (Chan, Knutsen, Blix, Lee, & Fraser, 2002) FXSVG and dehydration with impaired cognitive performance (Lieberman, 2007; Wilson & Morley, 2003). Upper MediCul cut-off point is based on observational data and exceeds findings from MEDIS cross sectional analysis where estimated plain water intake by Greek-born Australians (from Med islands) is 947 ml/d (Tsindos, Itsiopoulos, &

337 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

Kouris-Blazos, 2015). No other Med diet tools assess this. Q38 Herbal tea Cups/d 0 = Other teas or 1 Wild herbs were picked from the mountains and intake herbal tea <4 herbal tea was regularly consumed in the cups/wk ‘traditional’ diet (Matalas, 2006). Greek mountain 1 KHUEDOWHD tea is rich in polyphenols and may benefit cognition cups/wk (Hofrichter et al., 2016; Samanidou, Tsagiannidis, & Sarakatsianos, 2012). Green/black tea was not traditionally consumed so it is not scored, despite providing polyphenols. No other Med diet tools assess this. Q39 Tea type Response to tea N/R N/R Included for qualitative purposes and to inform type mostly scoring for Q38. consumed Although green/black tea contain polyphenols, they were not part of the ‘traditional’ Med diet (Radd- Vagenas et al., 2017). 15. Alcohol Q40 Alcohol Days/wk 0 = >5 days/wk 2 Alcohol was consumed in the ‘traditional’ Med diet drinking days 2 = 0-5 days/wk but only with meals and in low amounts (usually frequency diluted with water) (Trichopoulou et al., 1995). MediCul cut-off point is based on NHMRC guidelines (National Health and Medical Research Council, 2001) which recommend 2 alcohol free days per week, as frequency is associated with increased intake. Alcohol standard Standard 0 = >14 std 2 Alcohol is associated with cancer risk, all-cause drinks intake: drinks/wk drinks/wk mortality and promotes excessive energy intake

338 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

Q41a F/S beer 1 = >7-14 std (Knott, Coombs, Stamatakis, & Biddulph, 2015; Q41b Light beer drinks/wk National Health and Medical Research Council, Q41c Wine 2 = 0-7 std 2009; Winstanley et al., 2011). Q41d Spirits drinks/wk Q assesses whether total alcohol intake is low, consistent with traditional intakes (Trichopoulou et al., 1995). The upper cut-off point is deliberately conservative compared to other Med diet index tools as the negative effects of even small amounts of alcohol are now better appreciated, especially for cancer (Bagnardi et al., 2013). Most alcohol is N/R 0 = Wine not 2 Drinking wine only with meals was customary in wine and main drink or the Med diet (Martinez-Gonzalez & Martin-Calvo, consumed only not exclusively 2016) and may reduce negative effects of alcohol with meals: consumed with by slowing absorption rate. Both red and white Q42a Wine type meals wines were traditionally consumed and contain Q42b With meals 2 = Wine main antioxidants (Trichopoulou, Bamia, & drink & Trichopoulos, 2009). Type is assessed in MediCul exclusively for qualitative purposes. consumed with MEDLIFE (Sotos-Prieto et al., 2014) cut-off point main meals 1-2 serves/d (1 serve=1 cup); also awards points for both red/white wine. MEDAS (Schroder et al., 2011) only scores for red wine and does not require consumption with meals. 16. Coffee Q37a Coffee intake Cups/d 0 = >2 cups/d 1 Caffeinated coffee was part of the ‘traditional’ Med unless de- diet but at lower levels than commonly believed caffeinated since coffee beans were not grown in the region and FXSVGRU expensive to buy (Allbaugh, 1953; Matalas, 2006).

339 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

any quantity de- This Q is a crude measure of coffee/caffeine caffeinated exposure since it is difficult to assess intake without multiple Qs to determine variety of beans, roasting and brewing type, and volume of beverage consumed. Although some observational studies suggest higher intakes of coffee may benefit cognition and dementia risk, there is large methodological heterogeneity across studies precluding conclusions for these outcomes (Santos, Costa, Santos, Vaz- Carneiro, & Lunet, 2010). MediCul uses a conservative cut-off point as higher intakes of coffee are associated with increased myocardial infarction risk in individuals with the commonly observed risk variant of the CYP1A2 gene (Cornelis, El-Sohemy, Kabagambe, & Campos, 2006). No other Med diet tools assess this. Q37b Caffeinated Response to N/R N/R The response for this Q is included to inform vs. non-caffeinated type of coffee scoring for Q37a. mostly consumed 17. Cuisine & Q6 Sofrito Times/wk 0 = 0-0.99 2 Sofrito is a combination of tomato, olive oil and lifestyle frequency times/wk onion/garlic (Estruch et al., 2013), used as basis of 1 = 1-1.99 multiple Med dishes. It is high in phytonutrients times/wk including quercetin and lycopene, important for WLPHVZN vascular function (Rodriguez-Rodriguez et al.,

340 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

2017). Because it is in a lipid matrix, sofrito facilitates increased bioavailability of phytonutrients contained in the vegetables and EVOO. Increased plasma lycopene is associated with decreased plasma and CSF IL-6 (an inflammatory marker) (Guest et al., 2014). Q based on MEDAS (Schroder et al., 2011) Q with cut-off SRLQWWLPHVZN Q9 Herbs and Times/wk 0 = 0-3.99 1 Seasonings, particularly herbs, were used in spices frequency times/wk ‘traditional’ Med cuisine (Radd-Vagenas et al., WLPHVZN 2017). These provide antioxidants and anti- inflammatory phytonutrients and may reduce the formation of noxious chemicals during high temperature cooking, e.g., AGEs, HCAs (Dearlove, Greenspan, Hartle, Swanson, & Hargrove, 2008; Murkovic, Steinberger, & Pfannhauser, 1998). MediCul scoring rewards if exposure occurs on more than half the days per week. MEDLIFE (Sotos-Prieto et al., 2014) cut-off point VHUYHVGEXWDOVRLQFOXGHVRQLRQDQGJDUOLF Q10 Times/wk 0 = 0-3.99 1 These condiments were regularly used in Lemon/vinegar times/wk ‘traditional’ cuisine. They may reduce the frequency WLPHVZN glycaemic load and retard formation of AGEs during/after cooking (Chen, Chen, Giudici, & Chen, 2016; Uribarri et al., 2010), as well as positively influence the microbiome to retard formation of TMA, a novel CVD risk factor (Wang et al., 2015).

341 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

A score for exposure to more than half the days per week is awarded. No other Med diet tools assess this. Fermented foods: Times/wk 0 = Q11 + Q23a 1 Fermented foods were regularly consumed as part Q11 Olives + Q23b = 0-3.99 of Med cuisine and may provide benefits for the frequency times/wk microbiome (Fidanza & Alberti, 2005; Iriti & Q23a Yoghurt 1 = Q11 + Q23a Vitalini, 2012; Kushi, Lenart, & Willett, 1995; frequency 4E Peres, Peres, Hernández-Mendoza, & Malcata, Q23b Fetta cheese times/wk 2012). frequency A novel combined exposure score for three fermented foods is calculated. MIND diet (Morris et al., 2015a) criteria for total cheese <1 serve/wk. No other Med diet tools assess exposure to fermented foods. Q43 Main meal Times/wk 0 = <5 1 The ‘traditional’ Med cuisine involved home home cooked Times/wk cooked meals (Radd-Vagenas et al., 2017). In frequency Western society, diet quality improves when Times/wk cooking at home more often (Monsivais, Aggarwal, & Drewnowski, 2014; Tiwari, Aggarwal, Tang, & Drewnowski, 2017). No other Med diet tools assess this. Q44 Main meal Times/wk 1 Main meals were usually eaten in company, which eaten alone Times/wk provides social connectedness (Altomare et al., frequency 1 = <5 2013; Serra-Majem, Bach-Faig, & Raido-Quintana, Times/wk 2012). This Q rewards when main meals for most

342 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

days per week are consumed in company rather than alone. No other Med diet tools assess this. Lower vs. higher Frequency 0 = Total higher 1 Traditional cooking methods generally used high AGEs cooking AGEs (Q45 + moisture and lower temperature (Radd-Vagenas et methods 4 WRWDO al., 2017). High temperature, dry heat, cooking frequency: lower AGEs methods common in Western countries promote Q45 Warmer (Q45 + Q46) formation of AGEs (Uribarri et al., 2010). AGEs weather 1 = Total higher are associated with dementia and other chronic Q46 Cooler AGEs (Q45 + diseases and their complications (Kellow & weather Q46) < total Coughlan, 2015; Takeuchi & Yamagishi, 2008). lower AGEs This novel Q assesses frequency of high vs. low (Q45 + Q46) AGEs cooking methods during warmer and cooler weather and awards a score when the majority of cooking is done using lower AGEs methods. Higher AGEs methods = a) grill, BBQ, dry fry b) shallow/deep fry c) roast/bake. Lower AGEs methods = d) boil/stew e) steam f) stir fry. Q47 Snacking Times/d 0 = >2 Times/d 1 Frequent snacking was uncommon in the frequency 7LPHVG ‘traditional’ Med diet. In Western countries frequent snacking is associated with increased energy intake and obesity (Hoffman & Gerber, 2013; Murakami & Livingstone, 2015). Q48 Fasting Days/yr 0 = no fasting 1 The ‘traditional’ Med diet was frugal and fasting frequency 1 = any days/yr from animal foods was regularly undertaken due to

343 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

religious reasons (Kafatos, Verhagen, Moschandreas, Apostolaki, & Westerop, 2000; Sarri, Linardakis, Bervanaki, Tzanakis, & Kafatos, 2004). The main meal was taken at lunch and eating late at night was uncommon, supporting a more time-restricted feeding pattern, which aligns better with circadian rhythms and may provide metabolic advantages (Sutton et al., 2018). Increasing dietary fibre from plant foods and limiting exposure to animal products may beneficially influence the microbiome (Kouris- Blazos & Itsiopoulos, 2014). Energy restriction is associated with improvement of multiple risk factors, e.g., insulin resistance (Mattson, Longo, & Harvie, 2017). Q assesses any deliberate fasting which may involve energy restriction or avoidance of animal products. No other Med diet tools assess this. Q49 Fasting type Response to N/R N/R Included for qualitative purposes only. indicate type As regular fasting is generally uncommon in usually Western countries, no scoring is provided for type practised of fasting. No other Med diet tools assess this. Q50a Napping Days/wk N/R N/R Napping after the midday meal was common in the frequency traditional lifestyle and some studies in Mediterranean countries suggest short naps reduce

344 Supplementary Material 6.1. - cont.

Elements Sub-elements Description of Cut-off Points Total Rationale Measure Score

risk of CVD (Naska, Oikonomou, Trichopoulou, Psaltopoulou, & Trichopoulos, 2007; Trichopoulos et al., 1988). Q is included for qualitative purposes and not scored in MediCul. No other Med diet tools assess this. Q50b Napping Less than 30 N/R N/R Q is included for qualitative purposes and not duration mins scored in MediCul. 30 mins or No other Med diet tools assess this. longer Q, question; wk, week; Med, Mediterranean; MEDAS, Mediterranean Diet Adherence Screener used in PREDIMED study; d, day; N/R, not relevant or not scored for in MediCul; TBSP, tablespoon; g, grams; ml, millilitres; AGTHE, Australian Guide to Healthy Eating; CSIRO, Commonwealth Scientific and Industrial Research Organisation; EVOO, extra virgin olive oil; MIND, Mediterranean-DASH diet intervention for neurodegenerative delay diet; IMI, Italian Mediterranean Index; MEDIS, MEDiterranean ISlands study; CVD, cardiovascular disease; GI, glycaemic index; MEDLIFE, MEDiterranean LIFEstyle index; GLNC, Grains Legumes Nutrition Council; AD, Alzheimer’s disease; MedDietScore, Mediterranean diet score; F/S, full strength; NHMRC, National Health Medical Research Council; std, standard; CSF, cerebrospinal fluid; TMA, trimethylamine; AGEs, advanced glycation end products; HCAs, heterocyclic amines.

345

References for Supplementary Material 6.1

Agnoli, C., Krogh, V., Grioni, S., Sieri, S., Palli, D., Masala, G., . . . Panico, S. (2011). A priori-defined dietary patterns are associated with reduced risk of stroke in a large Italian cohort. Journal of Nutrition, 141(8), 1552-1558. doi:10.3945/jn.111.140061 Aiello, A., Guccione, G. D., Accardi, G., & Caruso, C. (2015). What olive oil for healthy ageing? Maturitas, 80(2), 117-118. doi:10.1016/j.maturitas.2014.10.016 Allbaugh, L. G. (1953). Crete: a case study of an underdeveloped area. New Jersey: Princeton University Press. Altomare, R., Cacciabaudo, F., Damiano, G., Palumbo, V. D., Gioviale, M. C., Bellavia, M., . . . Lo Monte, A. I. (2013, May). The Mediterranean diet: A history of health. Iranian Journal of Public Health. Retrieved 23rd December 2018 from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3684452/ Aune, D., Keum, N., Giovannucci, E., Fadnes, L. T., Boffetta, P., Greenwood, D. C., . . . Norat, T. (2016). Nut consumption and risk of cardiovascular disease, total cancer, all-cause and cause-specific mortality: A systematic review and dose- response meta-analysis of prospective studies. BMC Medicine, 14(1). doi:10.1186/s12916-016-0730-3 Bach-Faig, A., Berry, E. M., Lairon, D., Reguant, J., Trichopoulou, A., Dernini, S., . . . Mediterranean Diet Foundation Expert Group. (2011). Mediterranean diet pyramid today: Science and cultural updates. Public Health Nutrition, 14(12A), 2274-2284. doi:10.1017/S1368980011002515 Bagnardi, V., Rota, M., Botteri, E., Tramacere, I., Islami, F., Fedirko, V., . . . La Vecchia, C. (2013). Light alcohol drinking and cancer: A meta-analysis. Annals of Oncology, 24(2), 301-308. doi:10.1093/annonc/mds337 Barberger-Gateau, P., Raffaitin, C., Letenneur, L., Berr, C., Tzourio, C., Dartigues, J.- F., & Alpérovitch, A. (2007). Dietary patterns and risk of dementia: The Three- City cohort study. Neurology, 69(20), 1921-1930. doi:10.1212/01.wnl.0000278116.37320.52 Barbour, J. A., Howe, P. R. C., Buckley, J. D., Bryan, J., & Coates, A. M. (2014). Nut consumption for vascular health and cognitive function. Nutrition Research Reviews, 27(1), 131-158. doi:10.1017/S0954422414000079

346

Barclay, A. W., Petocz, P., McMillan-Price, J., Flood, V. M., Prvan, T., Mitchell, P., & Brand-Miller, J. C. (2008). Glycemic index, glycemic load, and chronic disease risk - A meta-analysis of observational studies. American Journal of Clinical Nutrition, 87(3), 627-637. doi:10.1093/ajcn/87.3.627 Bellavia, A., Stilling, F., & Wolk, A. (2016). High red meat intake and all-cause cardiovascular and cancer mortality: Is the risk modified by fruit and vegetable intake? American Journal of Clinical Nutrition, 104(4), 1137-1143. doi:10.3945/ajcn.116.135335 Bolton-Smith, C., Price, R. J. G., Fenton, S. T., Harrington, D. J., & Shearer, M. J. (2000). Compilation of a provisional UK database for the phylloquinone (vitamin K1) content of foods. British Journal of Nutrition, 83(4), 389-399. doi:10.1017/S0007114500000490 Brand-Miller, J., Hayne, S., Petocz, P., & Colagiuri, S. (2003). Low–glycemic index diets in the management of diabetes. Diabetes Care, 26(8), 2261-2267. doi:10.2337/diacare.26.8.2261 Brookie, K. L., Best, G. I., & Conner, T. S. (2018). Intake of raw fruits and vegetables is associated with better mental health than intake of processed fruits and vegetables. Frontiers in Psychology, 9, 487. doi:10.3389/fpsyg.2018.00487 Cahill, L. E., Pan, A., Chiuve, S. E., Sun, Q., Willett, W. C., Hu, F. B., & Rimm, E. B. (2014). Fried-food consumption and risk of type 2 diabetes and coronary artery disease: A prospective study in 2 cohorts of US women and men. American Journal of Clinical Nutrition, 100(2), 667-675. doi:10.3945/ajcn.114.084129 Cai, W., Uribarri, J., Zhu, L., Chen, X., Swamy, S., Zhao, Z., . . . Vlassara, H. (2014). Oral glycotoxins are a modifiable cause of dementia and the metabolic syndrome in mice and humans. Proceedings of the National Academy of Sciences of the United States of America, 111(13), 4940-4945. doi:10.1073/pnas.1316013111 Casamenti, F., & Stefani, M. (2017). Olive polyphenols: New promising agents to combat aging-associated neurodegeneration. Expert Review of Neurotherapeutics, 17(4), 345-358. doi:10.1080/14737175.2017.1245617

347

Chan, J., Knutsen, S. F., Blix, G. G., Lee, J. W., & Fraser, G. E. (2002). Water, other fluids, and fatal coronary heart disease: The Adventist Health Study. American Journal of Epidemiology, 155(9), 827-833. doi:10.1093/aje/155.9.827 Chen, H., Chen, T., Giudici, P., & Chen, F. (2016). Vinegar functions on health: Constituents, sources, and formation mechanisms. Comprehensive Reviews in Food Science and Food Safety, 15(6), 1124-1138. doi:10.1111/1541-4337.12228 Ciccarone, E., Di Castelnuovo, A., Salcuni, M., Siani, A., Giacco, A., Donati, M. B., . . . Gendiabe, I. (2003). A high-score Mediterranean dietary pattern is associated with a reduced risk of peripheral arterial disease in Italian patients with Type 2 diabetes. Journal of Thrombosis and Haemostasis, 1(8), 1744-1752. doi:10.1046/j.1538-7836.2003.00323.x Cicerale, S., Conlan, X. A., Sinclair, A. J., & Keast, R. S. (2009). Chemistry and health of olive oil phenolics. Critical Reviews in Food Science and Nutrition, 49(3), 218-236. doi:10.1080/10408390701856223 Cornelis, M. C., El-Sohemy, A., Kabagambe, E. K., & Campos, H. (2006). Coffee, CYP1A2 genotype, and risk of myocardial infarction. JAMA, 295(10), 1135- 1141. doi:10.1001/jama.295.10.1135 Darmos-Thodis, T. (2013). Depressive symptoms and the prevalence of cardiovascular risk factors among Greek Australians: Preliminary findings from the Mediterranean Islands (MEDIS-Australia) study. Retrieved 23rd December 2018 from www.atiner.gr/papers/PSY2013-0470.pdf Dauchet, L., Amouyel, P., & Dallongeville, J. (2005). Fruit and vegetable consumption and risk of stroke: A meta-analysis of cohort studies. Neurology, 65(8), 1193- 1197. doi:10.1212/01.wnl.0000180600.09719.53 Dearlove, R. P., Greenspan, P., Hartle, D. K., Swanson, R. B., & Hargrove, J. L. (2008). Inhibition of protein glycation by extracts of culinary herbs and spices. Journal of Medicinal Food, 11(2), 275-281. doi:10.1089/jmf.2007.536 Djoussé, L., & Gaziano, J. M. (2008). Egg consumption in relation to cardiovascular disease and mortality: The Physicians' Health Study. American Journal of Clinical Nutrition, 87(4), 964-969. doi:10.1093/ajcn/87.4.964

348

Djoussé, L., Khawaja, O. A., & Gaziano, J. M. (2016). Egg consumption and risk of type 2 diabetes: A meta-analysis of prospective studies. American Journal of Clinical Nutrition, 103(2), 474-480. doi:10.3945/ajcn.115.119933 Estruch, R., Ros, E., Salas-Salvado, J., Covas, M. I., Corella, D., Aros, F., . . . for the PREDIMED Study Investigators. (2013). Primary prevention of cardiovascular disease with a Mediterranean diet. New England Journal of Medicine, 368(14), 1279-1290. doi:10.1056/NEJMoa1200303 Fidanza, F., & Alberti, A. (2005). The healthy Italian Mediterranean diet temple food guide. Nutrition Today, 40(2), 71-78. doi:10.1097/00017285-200503000-00005 Giem, P., Beeson, W. L., & Fraser, G. E. (1993). The incidence of dementia and intake of animal products: Preliminary findings from the Adventist Health Study. Neuroepidemiology, 12(1), 28-36. doi:10.1159/000110296 Goff, L. M., Cowland, D. E., Hooper, L., & Frost, G. S. (2013). Low glycaemic index diets and blood lipids: A systematic review and meta-analysis of randomised controlled trials. Nutrition, Metabolism and Cardiovascular Diseases, 23(1), 1- 10. doi:10.1016/j.numecd.2012.06.002 Goulet, J., Nadeau, G., & Lemieux, S. (2003). Effect of a nutritional intervention promoting the Mediterranean food pattern on plasma lipids, lipoproteins and body weight in healthy French-Canadian women. Atherosclerosis, 170(1), 115- 124. doi:10.1016/s0021-9150(03)00243-0 Grains & Legumes Nutrition Council. (2013). Lifting the lid on legumes: The benefits of choosing legumes. Retrieved 18th May 2018 from https://www.glnc.org.au/resources/healthcare-professionals/ Gu, Y., Brickman, A. M., Stern, Y., Habeck, C. G., Razlighi, Q. R., Luchsinger, J. A., . . . Scarmeas, N. (2015). Mediterranean diet and brain structure in a multiethnic elderly cohort. Neurology, 85(20), 1774-1751. doi:10.1212/WNL.0000000000002121 Guest, J., Grant, R., Garg, M., Mori, T. A., Croft, K. D., & Bilgin, A. (2014). Cerebrospinal fluid levels of inflammation, oxidative stress and NAD+ are linked to differences in plasma carotenoid concentrations. Journal of Neuroinflammation, 11(1), 117. doi:10.1186/1742-2094-11-117

349

Heart Foundation. (2015). Healthy hearts position statement: Fish and seafood. Retrieved 18th May 2018 from https://www.heartfoundation.org.au/for- professionals/food-and-nutrition/position-statements Hendrie, G., & Noakes, M. (2017). Fruit, vegetables and diet score. Retrieved 18th May 2018 from https://my.totalwellbeingdiet.com/healthy-diet-score Hoffman, R., & Gerber, M. (2013). Evaluating and adapting the Mediterranean diet for non-Mediterranean populations: A critical appraisal. Nutrition Reviews, 71(9), 573-584. doi:10.1111/nure.12040 Hofrichter, J., Krohn, M., Schumacher, T., Lange, C., Feistel, B., Walbroel, B., & Pahnke, J. (2016). Sideritis spp. extracts enhance memory and learning in $O]KHLPHU Vȕ-amyloidosis mouse models and aged C57Bl/6 mice. Journal of Alzheimer's Disease, 53(3), 967-980. doi:10.3233/JAD-160301 Hu, D., Huang, J., Wang, Y., Zhang, D., & Qu, Y. (2014). Fruits and vegetables consumption and risk of stroke: A meta-analysis of prospective cohort studies. Stroke, 45(6), 1613-1619. doi:10.1161/STROKEAHA.114.004836 Iriti, M., & Vitalini, S. (2012). Health-promoting effects of traditional Mediterranean diets: A review. Polish Journal of Food and Nutrition Sciences, 62(2), 71-76. doi:10.2478/v10222-011-0047-z Jansen, M. C. J. F., Bueno-de-Mesquita, H. B., Feskens, E. J. M., Streppel, M. T., Kok, F. J., & Kromhout, D. (2004). Reports: Quantity and variety of fruit and vegetable consumption and cancer risk. Nutrition and Cancer, 48(2), 142-148. doi:10.1207/s15327914nc4802_3 Kafatos, A., Verhagen, H., Moschandreas, J., Apostolaki, I., & Westerop, J. J. M. V. (2000). Mediterranean diet of Crete: Foods and nutrient content. Journal of the American Dietetic Association, 100(12), 1487-1493. doi:10.1016/S0002- 8223(00)00416-8 Keim, N. L., Forester, S. M., Lyly, M., Aaron, G. J., & Townsend, M. S. (2014). Vegetable variety is a key to improved diet quality in low-income women in California. Journal of the Academy of Nutrition and Dietetics, 114(3), 430-435. doi:10.1016/j.jand.2013.07.026

350

Kellow, N. J., & Coughlan, M. T. (2015). Effect of diet-derived advanced glycation end products on inflammation. Nutrition Reviews, 73(11), 737-759. doi:10.1093/nutrit/nuv030 Keys, A. (1995). Mediterranean diet and public health: Personal reflections. American Journal of Clinical Nutrition, 61(6), 1321S-1323S. doi:10.1093/ajcn/61.6.1321S Knott, C. S., Coombs, N., Stamatakis, E., & Biddulph, J. P. (2015). All cause mortality and the case for age specific alcohol consumption guidelines: Pooled analyses of up to 10 population based cohorts. BMJ (Clinical Research Edition), 350, h384. doi:10.1136/bmj.h384 Kouris-Blazos, A., & Itsiopoulos, C. (2014). Low all-cause mortality despite high cardiovascular risk in elderly Greek-born Australians: Attenuating potential of diet? Asia Pacific Journal of Clinical Nutrition, 23(4), 532-544. doi:10.6133/apjcn.2014.23.4.16 Kromhout, D., Keys, A., Aravanis, C., Buzina, R., Fidanza, F., Giampaoli, S., . . . Pekkarinen, M. (1989). Food consumption patterns in the 1960s in seven countries. American Journal of Clinical Nutrition, 49(5), 889-894. doi:10.1093/ajcn/49.5.889 Kushi, L. H., Lenart, E. B., & Willett, W. C. (1995). Health implications of Mediterranean diets in light of contemporary knowledge. 1. Plant foods and dairy products. American Journal of Clinical Nutrition, 61(6 Suppl), 1407S- 1415S. doi:10.1093/ajcn/61.6.1407S Larsson, S. C., & Orsini, N. (2014). Red meat and processed meat consumption and all- cause mortality: A meta-analysis. American Journal of Epidemiology, 179(3), 282-289. doi:10.1093/aje/kwt261 Lidder, S., & Webb, A. J. (2013). Vascular effects of dietary nitrate (as found in green leafy vegetables and beetroot) via the nitrate-nitrite-nitric oxide pathway. British Journal of Clinical Pharmacology, 75(3), 677-696. doi:10.1111/j.1365- 2125.2012.04420.x Lieberman, H. R. (2007). Hydration and cognition: A critical review and recommendations for future research. Journal of the American College of Nutrition, 26(Supplement 5), 555S-561S. doi:10.1080/07315724.2007.10719658

351

Martinez-Gonzalez, M. A., & Martin-Calvo, N. (2016). Mediterranean diet and life expectancy; beyond olive oil, fruits, and vegetables. Current Opinion in Clinical Nutrition and Metabolic Care, 19(6), 401-407. doi:10.1097/MCO.0000000000000316 Martinez-Gonzalez, M. A., Hershey, M. S., Zazpe, I., & Trichopoulou, A. (2017). Transferability of the Mediterranean diet to non-Mediterranean countries: What is and what is not the Mediterranean diet. Nutrients, 9(11), 1226. doi:10.3390/nu9111226 Masala, G., Ceroti, M., Pala, V., Krogh, V., Vineis, P., Sacerdote, C., . . . Palli, D. (2007). A dietary pattern rich in olive oil and raw vegetables is associated with lower mortality in Italian elderly subjects. British Journal of Nutrition, 98(2), 406-415. doi:10.1017/S0007114507704981 Matalas, A.-L. (2006). Disparities within traditional Mediterranean food patterns: An historical approach of the Greek diet. International Journal of Food Sciences and Nutrition, 57(7-8), 529-536. doi:10.1080/09637480601041037 Mattson, M. P., Longo, V. D., & Harvie, M. (2017). Impact of intermittent fasting on health and disease processes. Ageing Research Reviews, 39, 46-58. doi:10.1016/j.arr.2016.10.005 Ministry of Health and Welfare: Supreme Scientific Health Council. (1999). Dietary guidelines for adults in Greece. Archives of Hellenic Medicine. Retrieved 19th December 2018 from http://www.mednet.gr/archives/1999-5/pdf/516.pdf Mizrahi, A., Knekt, P., Montonen, J., Laaksonen, M. A., Heliovaara, M., & Jarvinen, R. (2009). Plant foods and the risk of cerebrovascular diseases: A potential protection of fruit consumption. British Journal of Nutrition, 102(7), 1075-1083. doi:10.1017/S0007114509359097 Mohn, E. S., & Johnson, E. J. (2017). Lutein and cognition across the lifespan. Nutrition Today, 52(4), 183-189. doi:10.1097/NT.0000000000000226 Monsivais, P., Aggarwal, A., & Drewnowski, A. (2014). Time spent on home food preparation and indicators of healthy eating. American Journal of Preventive Medicine, 47(6), 796-802. doi:10.1016/j.amepre.2014.07.033

352

Morris, M. C., Tangney, C. C., Wang, Y., Sacks, F. M., Barnes, L. L., Bennett, D. A., & Aggarwal, N. T. (2015a). MIND diet slows cognitive decline with aging. Alzheimer's & Dementia, 11(9), 1015-1022. doi:10.1016/j.jalz.2015.04.011 Morris, M. C., Tangney, C. C., Wang, Y., Sacks, F. M., Bennett, D. A., & Aggarwal, N. T. (2015b). MIND diet associated with reduced incidence of Alzheimer's disease. Alzheimer's & Dementia, 11, 1007-1014. doi:10.1016/j.jalz.2014.11.009 Murakami, K., & Livingstone, M. B. E. (2015). Eating frequency is positively associated with overweight and central obesity in US adults. Journal of Nutrition, 145(12), 2715-2724. doi:10.3945/jn.115.219808 Murkovic, M., Steinberger, D., & Pfannhauser, W. (1998). Antioxidant spices reduce the formation of heterocyclic amines in fried meat. Zeitschrift für Lebensmitteluntersuchung und -Forschung A, 207(6), 477-480. doi:10.1007/s002170050364 Naska, A., Oikonomou, E., Trichopoulou, A., Psaltopoulou, T., & Trichopoulos, D. (2007). Siesta in healthy adults and coronary mortality in the general population. Archives of Internal Medicine, 167(3), 296-301. doi:10.1001/archinte.167.3.296 National Health and Medical Research Council. (2001). Australian alcohol guidelines: Health risks and benefits. Canberra, Australia: Commonwealth of Australia. National Health and Medical Research Council. (2009). Australian guidelines to reduce health risks from drinking alcohol. Canberra, Australia: Commonwealth of Australia. National Health and Medical Research Council. (2013). Australian dietary guidelines. Canberra, Australia: National Health and Medical Research Council. O'Brien, J., Okereke, O., Devore, E., Rosner, B., Breteler, M., & Grodstein, F. (2014). Long-term intake of nuts in relation to cognitive function in older women. Journal of Nutrition Health & Aging, 18(5), 496-502. doi:10.1007/s12603-014- 0014-6 Panagiotakos, D., Kalogeropoulos, N., Pitsavos, C., Roussinou, G., Palliou, K., Chrysohoou, C., & Stefanadis, C. (2009). Validation of the MedDietScore via the determination of plasma fatty acids. International Journal of Food Sciences and Nutrition, 60(Suppl5), 168-180. doi:10.1080/09637480902810338

353

Peres, C. M., Peres, C., Hernández-Mendoza, A., & Malcata, F. X. (2012). Review on fermented plant materials as carriers and sources of potentially probiotic lactic acid bacteria – With an emphasis on table olives. Trends in Food Science & Technology, 26(1), 31-42. doi:10.1016/j.tifs.2012.01.006 Pes, G. M., Tolu, F., Dore, M. P., Sechi, G. P., Errigo, A., Canelada, A., & Poulain, M. (2015). Male longevity in Sardinia, a review of historical sources supporting a causal link with dietary factors. European Journal of Clinical Nutrition, 69(4), 411-418. doi:10.1038/ejcn.2014.230 Pribis, P., Bailey, R. N., Russell, A. A., Kilsby, M. A., Hernandez, M., Craig, W. J., . . . Sabatè, J. (2011). Effects of walnut consumption on cognitive performance in young adults. British Journal of Nutrition, 107(9), 1-9. doi:10.1017/S0007114511004302 Radd-Vagenas, S., Kouris-Blazos, A., Fiatarone Singh, M., & Flood, V. M. (2017). Evolution of Mediterranean diets and cuisine: Concepts and definitions. Asia Pacific Journal of Clinical Nutrition, 26(5), 749-763. doi:10.6133/apjcn.082016.06 Rodriguez-Rodriguez, R., Jiménez-Altayó, F., Alsina, L., Onetti, Y., Rinaldi de Alvarenga, J. F., Claro, C., . . . Lamuela-Raventos, R. M. (2017). Mediterranean tomato-based sofrito protects against vascular alterations in obese Zucker rats by preserving NO bioavailability. Molecular Nutrition and Food Research, 61(9), 1601010. doi:10.1002/mnfr.201601010 Sabaté, J., & Ang, Y. (2009). Nuts and health outcomes: New epidemiologic evidence. American Journal of Clinical Nutrition, 89(5), S1643-S1648. doi:10.3945/ajcn.2009.26736Q Samanidou, V., Tsagiannidis, A., & Sarakatsianos, I. (2012). Simultaneous determination of polyphenols and major purine alkaloids in Greek Sideritis species, herbal extracts, green tea, black tea, and coffee by high-performance liquid chromatography-diode array detection. Journal of Separation Science, 35(4), 608-615. doi:10.1002/jssc.201100894 Santos, C., Costa, J., Santos, J., Vaz-Carneiro, A., & Lunet, N. (2010). Caffeine intake and dementia: Systematic review and meta-analysis. Journal of Alzheimer's Disease, 20(1), S187-S204. doi:10.3233/JAD-2010-091387

354

Sarri, K. O., Linardakis, M. K., Bervanaki, F. N., Tzanakis, N. E., & Kafatos, A. G. (2004). Greek Orthodox fasting rituals: A hidden characteristic of the Mediterranean diet of Crete. British Journal of Nutrition, 92(2), 277-284. doi:10.1079/BJN20041197 Schroder, H., Fito, M., Estruch, R., Martinez-Gonzalez, M. A., Corella, D., Salas- Salvado, J., . . . Covas, M. I. (2011). A short screener is valid for assessing Mediterranean diet adherence among older Spanish men and women. Journal of Nutrition, 141(6), 1140-1145. doi:10.3945/jn.110.135566 Serra-Majem, L., Bach-Faig, A., & Raido-Quintana, B. (2012). Nutritional and cultural aspects of the Mediterranean diet. International Journal for Vitamin and Nutrition Research, 82(3), 157-162. doi:10.1024/0300-9831/a000106 Sharma, S. P., Chung, H. J., Kim, H. J., & Hong, S. T. (2016). Paradoxical effects of fruit on obesity. Nutrients, 8(10), 633. doi:10.3390/nu8100633 Shojaee-Aliabadi, S., Nikoopour, H., Kobarfard, F., Parsapour, M., Moslehishad, M., Hassanabadi, H., . . . Dahaghin, E. (2013). Acrylamide reduction in potato chips by selection of potato variety grown in Iran and processing conditions. Journal of the Science of Food and Agriculture, 93(10), 2556-2561. doi:10.1002/jsfa.6076 Sotos-Prieto, M., Moreno-Franco, B., Ordovás, J. M., León, M., Casasnovas, J. A., & Peñalvo, J. L. (2014). Design and development of an instrument to measure overall lifestyle habits for epidemiological research: The Mediterranean lifestyle (MEDLIFE) index. Public Health Nutrition, 18(6), 959-967. doi:10.1017/S1368980014001360 Sutton, E. F., Beyl, R., Early, K. S., Cefalu, W. T., Ravussin, E., & Peterson, C. M. (2018). Early time-restricted feeding improves insulin sensitivity, blood pressure, and oxidative stress even without weight loss in men with prediabetes. Cell Metabolism, 27(6), 1212-1221. doi:10.1016/j.cmet.2018.04.010 Takeuchi, M., & Yamagishi, S.-i. (2008). Possible involvement of advanced glycation end-products (AGEs) in the pathogenesis of Alzheimers disease. Current Pharmaceutical Design, 14(10), 973-978. doi:10.2174/138161208784139693

355

Tan, J. S. L., Wang, J. J., Flood, V., Rochtchina, E., Smith, W., & Mitchell, P. (2008). Dietary antioxidants and the long-term incidence of age-related macular degeneration: The Blue Mountains Eye Study. Ophthalmology, 115(2), 334-341. doi:10.1016/j.ophtha.2007.03.083 Tiwari, A., Aggarwal, A., Tang, W., & Drewnowski, A. (2017). Cooking at home: A strategy to comply with U.S. dietary guidelines at no extra cost. American Journal of Preventive Medicine, 52(5), 616-624. doi:10.1016/j.amepre.2017.01.017 Trichopoulos, D., Trichopoulos, D., Tzonou, A., Tzonou, A., Christopoulos, C., Christopoulos, C., . . . Trichopoulou, A. (1988). Siesta and risk of coronary heart disease. Stress Medicine, 4(3), 143-148. doi:10.1002/smi.2460040306 Trichopoulou, A., Kouris-Blazos, A., Wahlqvist, M. L., Gnardellis, C., Lagiou, P., Polychronopoulos, E., . . . Trichopoulos, D. (1995). Diet and overall survival in elderly people. British Medical Journal, 311(7018), 1457-1460. doi:10.1136/bmj.311.7018.1457 Trichopoulou, A., & Lagiou, P. (1997). Healthy traditional Mediterranean diet: An expression of culture, history, and lifestyle. Nutrition Reviews, 55(11), 383-389. doi:10.1111/j.1753-4887.1997.tb01578.x Trichopoulou, A., Bamia, C., & Trichopoulos, D. (2009). Anatomy of health effects of Mediterranean diet: Greek EPIC prospective cohort study. BMJ (Clinical Research Edition), 338, b2337. doi:10.1136/bmj.b2337 Tsindos, P. S., Itsiopoulos, C., & Kouris-Blazos, A. (2015). Investigation into water consumption and its influence on depression, memory problems and constipation in older persons. Journal of Aging Research & Clinical Practice. Retrieved 23rd December 2018 from http://www.jarcp.com/all- issues.html?search Uribarri, J., Woodruff, S., Goodman, S., Cai, W., Chen, X., Pyzik, R., . . . Vlassara, H. (2010). Advanced glycation end products in foods and a practical guide to their reduction in the diet. Journal of the American Dietetic Association, 110(6), 911- 916. doi:10.1016/j.jada.2010.03.018

356

Wang, Z., Roberts, Adam B., Buffa, Jennifer A., Levison, Bruce S., Zhu, W., Org, E., . . . Lusis, Aldons J. (2015). Non-lethal inhibition of gut microbial trimethylamine production for the treatment of atherosclerosis. Cell, 163(7), 1585-1595. doi:10.1016/j.cell.2015.11.055 Willett, W. C., Sacks, F., Trichopoulou, A., Drescher, G., Ferro-Luzzi, A., Helsing, E., & Trichopoulos, D. (1995). Mediterranean diet pyramid: A cultural model for healthy eating. American Journal of Clinical Nutrition, 61(6 Suppl), 1402S- 1406S. doi:10.1093/ajcn/61.6.1402S Willette, A. A., Bendlin, B. B., Starks, E. J., Birdsill, A. C., Johnson, S. C., Christian, B. T., . . . Asthana, S. (2015). Association of insulin resistance with cerebral glucose uptake in late middle–aged adults at risk for Alzheimer disease. JAMA Neurology, 72(9), 1013-1020. doi:10.1001/jamaneurol.2015.0613 Wilson, M.-M. G., & Morley, J. E. (2003). Impaired cognitive function and mental performance in mild dehydration. European Journal of Clinical Nutrition, 57(S2), S24-S29. doi:10.1038/sj.ejcn.1601898 Winstanley, M. H., Pratt, I. S., Chapman, K., Griffin, H. J., Croager, E. J., Olver, I. N., . . . Slevin, T. J. (2011). Alcohol and cancer: A position statement from Cancer Council Australia. Medical Journal of Australia. Retrieved 23rd December 2018 from https://www.cancer.org.au/policy-and-advocacy/position- statements/alcohol-and-cancer/ Xi, B., Li, S., Liu, Z., Tian, H., Yin, X., Huai, P., . . . Steffen, L. M. (2014). Intake of fruit juice and incidence of type 2 diabetes: A systematic review and meta- analysis. PLoS ONE, 9(3), e93471. doi:10.1371/journal.pone.0093471 Zheng, M., Wu, J. H. Y., Louie, J. C. Y., Flood, V. M., Gill, T., Thomas, B., . . . Rangan, A. (2016). Typical food portion sizes consumed by Australian adults: Results from the 2011-12 Australian National Nutrition and Physical Activity Survey. Scientific Reports, 6, 19596. doi:10.1038/srep19596

357 Supplementary Material 6.2. MediCul index tool Name:______

Assessment:______

Date:______

Administrator:______

Short Dietary Survey

This is a survey about the food you eat. Read it carefully and identify the amounts that best describe your usual intake over the last six months. It’s important that the answers you provide reflect what you personally eat, rather than what you think you should or shouldn’t be having, or what someone else wants you to eat. When there are options, please choose one response most relevant to you. This survey will take around 20 minutes to complete.

But first, here are two sample questions with sample answers to give you an idea of how the survey works.

Sample Question 1: How often do you eat jelly beans? (Jelly beans of all colours are included).

Sample Response 1: If you usually don’t eat any eat jelly beans, or do this rarely, you would skip the first two options and tick the last option box, like this.

______times per day

______times per week

I don’t eat jelly beans

358 Supplementary Material 6.2. MediCul index tool

Sample Question 2: How many serves specifically of Lebanese bread do you usually eat each day? (1 serve is one quarter of a large Lebanese bread, which is the size of a dinner plate).

Sample Response 2: If you usually eat one whole Lebanese bread for lunch and another half a Lebanese bread for dinner, you would write 6 serves in the first option, like this.

___6____ serves per day

______serves per week

______serves per month

I don’t eat Lebanese bread

359 Supplementary Material 6.2. MediCul index tool

Now it’s over to you. Please start with the first question below.

Remember, your answers should represent your usual intake over the last six months.

1. How many serves of vegetables do you usually eat each day? (1 serve is ½ cup cooked vegetables or 1 cup of salad vegetables). Please choose one response most relevant to you.

______serves per day

OR ______serves per week OR ______serves per month

I don’t eat vegetables or salad

2. How many different types of vegetables do you usually eat in one week? Count each type only once. ______different vegetables are eaten over the week

3. How often do you usually eat raw vegetables such as salads, carrot sticks, and sprouts? Don’t count small garnishes. Please choose one response most relevant to you. ______times per day OR ______times per week OR ______times per month OR I don’t eat raw vegetables

360 Supplementary Material 6.2. MediCul index tool 4. How many times per week do you usually eat raw or cooked green vegetables? Write a number next to each of the two groups below. Place a ‘0’ (zero) on the line if you don’t usually eat any greens on a weekly basis. SEE PICTURE BELOW FOR EXAMPLES OF DARK GREEN LEAFY VEGETABLES. ______times per week of broccoli, peas, beans, zucchini, Brussels sprouts, cabbage, bok choy AND ______times per week of dark green leafy varieties such as kale, spinach, silverbeet, amaranth, dandelion, chicory, endive, rocket

5. How often do you usually eat onions, garlic, spring onions or shallots? Count those used in cooking and eaten raw in salads. Please choose one response most relevant to you. ______times per day OR ______times per week OR ______times per month OR I don’t eat onions, garlic, spring onions or shallots

6. How many times per week do you usually eat dishes cooked in a sauce made with tomato and onion simmered in olive oil? The sauce may also include garlic and herbs. (Exclude canned/bottled tomato sauces if onion and olive oil are not used). Please choose one response most relevant to you. ______times per week OR ______times per month OR

I don’t eat dishes cooked in a sauce made with tomato and onion simmered in olive oil

361 Supplementary Material 6.2. MediCul index tool

7. How often do you usually eat hot chips, French fries, wedges or fried potatoes? Please choose one response most relevant to you. ______times per day OR ______times per week OR ______times per month OR I don’t eat hot chips, French fries wedges or fried potatoes

8. Do you grow any of your own vegetables? No OR Yes

9. How often do you usually use herbs or spices? For example, in cooking, salad or dessert. This includes fresh or dried varieties such as parsley, oregano, cinnamon, cumin, pepper etc. Please choose one response most relevant to you. ______times per day OR ______times per week

OR ______times per month

I don’t use herbs or spices

362 Supplementary Material 6.2. MediCul index tool

10. How often do you usually use lemon or vinegar when preparing food? For example, to make salad dressing, stirred into soup or for basting meat or seasoning vegetables before roasting. Please choose one response most relevant to you. ______times per day OR ______times per week OR ______times per month OR I don’t use lemon or vinegar when preparing food

11. How often do you usually eat olives? This includes black, green, kalamata or stuffed olives and tapenade (a paste) made from olives. Please choose one response most relevant to you. ______times per day OR ______times per week OR ______times per month OR I don’t eat olives or tapenade

12. How many serves of fruit do you usually eat each day? Do not count juice. (1 serve is 1 medium piece or 2 small pieces of fruit or 1 cup of diced/canned fruit or 30 g dried fruit e.g. 4 dried apricot halves, 1 ½ tablespoons sultanas). Please choose one response most relevant to you. ______serves per day OR ______serves per week OR ______serves per month OR I don’t eat fruit

363 Supplementary Material 6.2. MediCul index tool 13. How many serves of red meat such as beef, veal, lamb, kangaroo or pork do you usually eat each day? Include all steaks, chops, roasts, mince, stir-fries and casseroles. (1 serve equals 100-150 g). Please choose one response most relevant to you.

______serves per day

OR ______serves per week

OR ______serves per month

OR I don’t eat red meat

14. How many serves of processed meat such as sausages, bacon, ham, devon, frankfurts, salami, luncheon meats or meat pies do you usually eat each day? (1 serve equals 1 ½ thick or 2 thinner sausages, 2 rashers bacon, 4 slices processed meats (100 g), 1 meat pie/pastie/sausage roll, 6 chicken nuggets). Please choose one response most relevant to you. ______serves per day OR ______serves per week OR ______serves per month OR I don’t eat processed meat

15. How many serves of white meat such as chicken, turkey or rabbit do you usually eat each day? Include all fillets, pieces, roasts, mince, stir-fries and casseroles. (1 serve is 100-150 g). Please choose one response most relevant to you.

______serves per day

OR ______serves per week

OR ______serves per month

OR I don’t eat white meat

364 Supplementary Material 6.2. MediCul index tool

16. How many serves of fish or shellfish do you usually eat each week? Include fresh and canned. (1 serve is 1 small fish fillet or 1 small can of fish or 200 g shellfish). Please choose one response most relevant to you. ______serves per day OR ______serves per week OR ______serves per month OR

I don’t eat fish or shellfish

17. Which do you usually eat more often? Please choose one response most relevant to you. Chicken, turkey or rabbit OR Beef, pork, hamburgers or sausages OR I don’t eat chicken or meat

18. How many eggs do you usually eat each day? Please choose one response most relevant to you. ______per day OR ______per week OR ______per month OR I don’t eat eggs

365 Supplementary Material 6.2. MediCul index tool

19. How many serves of legumes do you usually eat each day? Legumes include baked beans, canned 4- bean mix, lentils, split peas, chickpeas and any other canned or dried beans. (1 serve is 1 cup (150 g) cooked or canned beans). They do not include fresh peas and green beans. Please choose one response most relevant to you. SEE PICTURE BELOW FOR EXAMPLES. ______serves per day OR ______serves per week OR ______serves per month OR I don’t eat legumes

20. How many times per week do you usually eat each of the following types of bread or wraps? Don’t worry about amounts. Place a ‘0’ (zero) on the line if you don’t usually eat certain types of bread or wraps. ______times per week white e.g. Tip Top, Wonder White, Molenberg ______times per week wholegrain e.g. Burgen, Helga’s, Schwob’s Swiss Bakery ______times per week wholemeal e.g. Buttercup, Glicks, Bill’s, Lawson’s ______times per week rye e.g. Country Life, Abbott’s Village Bakery, Van Der Meulin ______times per week sourdough e.g. Coles Bakery, Bill’s, Macro, Woolworths

366 Supplementary Material 6.2. MediCul index tool 21. How many serves of dairy products do you usually eat each day? (1 serve is 1 cup milk (250 ml), 2 slices hard cheese (40 g), ½ cup (120 g) ricotta/cottage, 80 g fetta cheese or 200 g (¾ cup) yoghurt). Don’t count dairy alternatives such as rice milk or soy yoghurt. Please choose one response most relevant to you. ______serves per day OR ______serves per week OR ______serves per month OR I don’t eat dairy products

22. What type of milk do you usually have? Please choose one response most relevant to you. Regular dairy milk (whole or full cream) OR Low or reduced fat dairy milk OR Skim dairy milk OR Other (please specify) ______OR I don’t have milk

23. How often do you usually eat the fermented dairy products below? Please choose one response most relevant to you. a) Yoghurt? This includes low fat, full cream, Greek yoghurt, probiotic yoghurt, fruit yoghurt and kefir. ______times per day OR ______times per week

OR ______times per month

I don’t eat yoghurt

367 Supplementary Material 6.2. MediCul index tool

b) Fetta cheese? ______times per day OR ______times per week OR ______times per month OR I don’t eat fetta cheese

24. How many serves of nuts do you usually consume per day or per week? Nuts include peanuts, walnuts, Brazil nuts, cashews etc. (1 serve is 30 g nuts or a small handful, or 1 ½ tablespoons nut paste/peanut butter). SEE PICTURE FOR EXAMPLES. ______serves per day

OR ______serves per week

OR ______serves per month

OR I don’t eat nuts

25. How many days per week do you usually use each of the following fats and oils? Fats and oils may be used in your cooking, as spreads or on salads. Don’t worry about amounts. Place a ‘0’ (zero) on the line if you don’t usually eat certain types of fats/oils. ______days per week butter ______days per week margarine ______days per week mayonnaise ______days per week vegetable oil e.g. sunflower, grapeseed, canola, rice bran ______days per week olive oil

368 Supplementary Material 6.2. MediCul index tool

26. How many serves of butter or cream do you usually eat each day? (1 serve is 1 teaspoon). Please choose one response most relevant to you. ______serves per day OR ______serves per week OR ______serves per month OR I don’t eat butter or cream

27. How many serves of margarine do you usually eat each day? (1 serve is 1 teaspoon). This includes all types/brands of margarine such as those formulated with olive oil, plant sterols and omega-3. For example, Flora, Meadow Lea, Olive Grove, Bertolli, Gold N Canola, Logical. Please choose one response most relevant to you. ______serves per day OR ______serves per week OR ______serves per month OR I don’t eat margarine

28. How many tablespoons of olive oil do you usually eat each day? This includes oil used in cooking, drizzled on salads or bread and food eaten away from home)? (1 tablespoon = 20 ml). Please choose one response most relevant to you.

______tablespoons per day

OR ______tablespoons per week

OR _____ tablespoons per month

I don’t use olive oil

369 Supplementary Material 6.2. MediCul index tool

29. a) Do you use olive oil as the main fat in your diet when considering all the types of fats/oils/spreads used in your cooking, food preparation and on your bread?

No OR Yes

b) If yes, what type of olive oil do you usually use? Please choose one response most relevant to you.

Light OR Classic/Mild/Pure OR Extra Virgin OR Other (please specify)______

You are over half way through the survey – we really appreciate your time -

30. How often do you usually have meals or snacks from takeaway food stores? Examples include McDonalds, Hungry Jacks, Pizza Hut, KFC, Red Rooster, fish/chicken shop or local take away food places and foods such as burgers, pizza, hot dogs, battered chicken or fish and chips.

______times per day

OR ______times per week

OR ______times per month

I don’t eat takeaway foods

370 Supplementary Material 6.2. MediCul index tool

31. How many times per day do you eat biscuits or cakes of any type? This includes sweet biscuits, muffins, sponge cakes, sweet buns, doughnuts and Danish pastries. Please choose one response most relevant to you.

______times per day

______times per week

OR ______times per month

I don’t eat biscuits or cakes

32. How many times per week do you usually consume custard or ice cream? Place a ‘0’ (zero) on the line if you don’t usually consume any. ______times per week custard ______times per week ice cream

33. How many cups of sugar sweetened/carbonated beverages do you usually drink each day? This includes soft drink, cordial, sports drink, vitamin water and energy drink. Don’t forget any used to mix with spirits. (1 cup is 250 ml, 1 can of soft drink is 1.5 cups). Do not count ‘diet’ drinks. Please choose one response most relevant to you.

______cups per day

OR ______cups per week OR ______cups per month OR I don’t drink sugar sweetened/carbonated beverages

371 Supplementary Material 6.2. MediCul index tool 34. How many cups of fruit juice do you usually drink each day? Fruit juice includes all types of fruit juices, fresh or commercial. (1 cup is 250 ml). Please choose one response most relevant to you. ______cups per day OR ______cups per week OR ______cups per month OR I don’t drink fruit juice (if so, skip the next question)

35. What type of fruit juice do you usually drink? Please choose one response most relevant to you. Fruit juice commercially packaged in bottles or tetra paks. OR Freshly squeezed fruit juice

36. How many cups of water do you usually drink each day? (1 cup is 250 ml; 1 litre equals 4 cups). Please choose one response most relevant to you. ______cups per day OR ______cups per week OR ______cups per month OR I don’t drink water

37. How many cups of coffee do you usually drink each day? Please choose one response most relevant to you. ______cups per day OR ______cups per week OR ______cups per month OR I don’t drink coffee

372 Supplementary Material 6.2. MediCul index tool

b) If you drink coffee, which type do you mostly drink? Caffeinated OR De-caffeinated

38. How many cups of tea do you usually drink each day? Please choose one response most relevant to you. ______cups per day OR ______cups per week OR ______cups per month OR I don’t drink tea (if so, skip the next question)

39. What type of tea do you mostly drink? Please choose one response most relevant to you. Black tea, with milk OR Black tea, no milk OR Green tea OR Herbal tea (this contains no caffeine)

40. How often do you usually drink alcoholic beverages? Please choose one response most relevant to you. ______days per week OR ______days per month OR I don’t drink alcoholic beverages (if so, skip the next two questions)

373 Supplementary Material 6.2. MediCul index tool

41. a) If you drink full strength beer, how many stubbies/cans (375 ml) do you usually have? Please choose one response most relevant to you. ______stubbies/cans per day OR ______stubbies/cans per week OR ______stubbies/cans per month OR I don’t drink full strength beer

b) If you drink lite beer, how many stubbies/cans (375 ml) do you usually have? Please choose one response most relevant to you. ______stubbies/cans per day OR ______stubbies/cans per week OR ______stubbies/cans per month OR I don’t drink lite beer

c) If you drink wine, how many glasses (150 ml) do you usually have? Please choose one response most relevant to you. ______glasses per day OR ______glasses per week OR ______glasses per month OR I don’t drink wine

374 Supplementary Material 6.2. MediCul index tool

d) If you drink spirits, how many pub-sized nips (30 ml) do you usually have? Please choose one response most relevant to you. ______nips per day OR ______nips per week OR ______nips per month OR I don’t drink spirits

42. a) If you drink wine, what type do you usually have? (If not, skip this question). White wine OR Red wine

b) When do you usually drink the wine? Only with main meals OR With meals and /or at other times, outside of main meal occasions

Finally, we’d like to ask you a few questions about your personal eating and lifestyle habits as well as the cooking methods used to prepare your food.

43. How often is your main or evening meal cooked at home? This may be prepared by you, your family or your friends. ______times per week OR ______times per month OR I don’t eat home cooked meals

44. How many of your main or evening meals do you eat alone (without company)? ______meals per week

375 Supplementary Material 6.2. MediCul index tool

45. During the warmer weather how many times per week do you usually eat foods/meals prepared using each of the following cooking methods? Place a ‘0’ (zero) on the line if you don’t use a certain cooking method. ______times per week prepared by grilling, BBQing or dry frying in a pan ______times per week prepared by shallow or deep frying ______times per week prepared by roasting or baking ______times per week prepared by boiling or stewing ______times per week prepared by steaming ______times per week prepared by stir frying

46. During the cooler weather how many times per week do you usually eat foods/meals prepared using each of the following cooking methods? Place a ‘0’ (zero) on the line if you don’t use a certain cooking method. ______times per week prepared by grilling, BBQing or dry frying in a pan ______times per week prepared by shallow or deep frying ______times per week prepared by roasting or baking ______times per week prepared by boiling or stewing ______times per week prepared by steaming ______times per week prepared by stir frying

47. a). How many times per day do you usually snack? Snacking is an eating occasion that occurs between main meals. Count snacks such as morning/afternoon tea, supper, eating while driving or while watching TV. If you snack multiple times between meals, count each occasion once if it is separated by 15 minutes. ______times per day OR I don’t usually snack between meals or I snack less often than daily

b). If you do snack daily, name the three most frequent types of snacks you usually have. ______

376 Supplementary Material 6.2. MediCul index tool

48. How often do you usually fast? Fasting means deliberately abstaining from eating all foods or avoiding certain types of foods for given periods. For example Lent, Ramadan, 5:2 diet. It does not mean occasionally skipping meals or missing breakfast. ______days per week OR ______days per month OR ______days per year OR I don’t fast (if so, skip the next question and go straight to the last question)

49. What option would best describe the type of fasting you usually practise? Please choose one response most relevant to you. I avoid certain types of foods when I fast e.g. avoid meat and dairy OR I restrict the amount of food for a given period e.g. reduce portion sizes or calories OR I avoid all foods for a given time period e.g. don’t eat at all during the day OR Other (please specify)______

50. a). How many days per week do you usually take a nap after lunch? ______days per week OR I don’t take a nap after lunch (if so, you are finished the survey)

b). If you nap after lunch three days per week or more often, how long do you usually nap for? Please pick one option. Less than 30 minutes OR 30 minutes or longer

Thank you for completing this survey -

377 Supplementary Material 6.2. MediCul index tool

© Sue Radd-Vagenas, Advanced APD 2018. The MediCul tool may be used, reproduced and distributed without permission. It should be made available free of charge to patients. Written permission is required for any form of commercial or pharma use.

Acknowledgement: Nuts for Life for image of handful of nuts www.nutsforlife.com.au; National Health and Medical Research Council for images of standard drinks of alcohol https://www.nhmrc.gov.au/health- topics/alcohol-guidelines.

378 Supplementary Material 6.3. Kappa statistics for percent agreement within the same category of 17 MediCul elements between survey A and B

MediCul elements Percent (%) Kappa value Level of agreement agreement within same (Landis & Koch, category 1977) 1. Olive oil 66.2% 0.53 (p<0.0001) Moderate 2. Vegetables 44.1% 0.35 (p<0.0001) Fair 3. Fruit 82.4% 0.71 (p<0.0001) Substantial 4. Nuts 72.1% 0.61 (p<0.0001) Substantial 5. Wholegrains 98.5% 0.93 (p<0.0001) Almost perfect 6. Legumes 52.9% 0.27 (p=0.002) Fair 7. Fish/shellfish 76.5% 0.62 (p<0.0001) Substantial 8. Eggs 86.8% 0.66 (p<0.0001) Substantial 9. Dairy products 83.8% 0.46 (p<0.0001) Moderate 10. White meat preference 79.4% 0.63 (p<0.0001) Substantial 11. Red/processed meat 61.8% 0.51 (p<0.0001) Moderate 12. Sweets and sugary drinks 73.5% 0.50 (p<0.0001) Moderate 13. Takeaway 70.6% 0.39 (p<0.0001) Fair 14. Water 82.4% 0.74 (p<0.0001) Substantial 15. Alcohol 83.8% 0.79 (p<0.0001) Substantial 16. Coffee 97.1% 0.86 (p<0.0001) Almost perfect 17. Cuisine 41.2% 0.31 (p<0.0001) Fair survey A, first administration of MediCul; survey B, second administration of MediCul.

Reference for interpretation of Kappa values Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33(1), 159-174. doi:10.2307/2529310

379 CHAPTER SEVEN

Discussion and conclusions

380 7.1 Overview and outcomes

This thesis set out to improve understanding of the ‘traditional’ Mediterranean cuisine; investigate the relationship of a Mediterranean diet to cognitive health; and to advance the measurement of adherence to the ‘traditional’ Mediterranean dietary pattern and aspects of cuisine for use in future studies within Western populations. This final chapter discusses the key findings and their implications. It describes the contribution this thesis makes to the body of knowledge and delineates remaining gaps in the evidence base.

7.1.1 Improved understanding of ‘traditional’ Mediterranean cuisine

The findings in this thesis support the premise that the ‘traditional’ Mediterranean diet is more complex than just a list of foods – it is a cuisine rooted in a way of life (Trichopoulou & Lagiou, 1997; United Nations Educational Scientific and Cultural Organization, 2010; Wahlqvist, Kouris-Blazos, Trichopoulou, & Polychronopoulos, 1991). The Mediterranean diet terminology came into existence after the famous Seven Countries Study (SCS) (Hatzis, Sifaki-Pistolla, & Kafatos, 2015). Therefore, it is generally agreed that the archetypal ‘traditional’ diet relates to dietary intakes, as well as cooking and related lifestyle habits, from rural villages in Southern Italy and Crete, Greece, during the late 1950s and early 1960s. These cohorts from the SCS were found to have the lowest chronic disease risks. Their ‘traditional’ diet differed considerably from that of some other Mediterranean countries and time periods. For example, it included a higher intake of extra virgin olive oil and fish, and a relatively low meat intake facilitated by regular religious fasting from animal products (Hatzis et al., 2015).

Chapter Two reviewed the evidence base and described the evolution of Mediterranean diets from ancient to ‘traditional’ and modern times. This was required as a broad range of definitions exist for the diet, impeding synthesis across trials. Confusion about the Mediterranean diet is also evident in the public domain. For example, a recently published consumer book, The Pioppi Diet (Malhotra & O'Neill, 2017), forbids eating bread and grain foods yet extols the virtues of coconut fat and recommends dark chocolate daily. This is a significant deviation from the ‘traditional’ Mediterranean diet of Pioppi,

381 Southern Italy, where Ancel Keys, who initiated the SCS from the University of Minnesota, eventually chose to reside. The ‘traditional’ Mediterranean diet has a solid biological foundation and does not represent a transient fashion (Trichopoulou & Vasilopoulou, 2000).

Chapter Two confirmed the unprocessed, plant-based and frugal nature of the ‘traditional’ Mediterranean diet. It also identified some unique elements relating to ‘traditional’ cooking and eating habits. These elements have not been collectively studied or well characterised by previous diet index tools, but may provide synergies with the foods consumed. They include home cooked meals; use of moist, lower temperature, cooking methods; eating main meals in company; reduced snacking occasions; fasting practice; cultivation of a vegetable garden; use of traditional foods and food combinations; and napping after the midday meal.

Therefore, this thesis has contributed to an improved understanding of the ‘traditional’ Mediterranean cuisine. Greater understanding of this cuisine may assist with future trial design, as it is often erroneously believed that any Mediterranean diet intervention represents the ‘traditional’ diet, which may have different health effects. It is also important for new tool development, or adaptation of existing tools, so that greater discernment is possible with other healthy plant-based dietary patterns.

7.1.2 Evidence for cognitive benefits warranting additional trials

Chapter Three described the results of the first systematic review on the effect of the Mediterranean diet on cognition and brain morphology and function, exclusively within randomised controlled trials. This review included a particular focus on the quality of interventions, and how the prescribed diets related to minimum food criteria for the ‘traditional’ dietary pattern, which has not been done previously.

The review found that the Mediterranean diets used in the included clinical trials varied considerably and did not necessarily represent the ‘traditional’ diet. It showed mostly

382 non-significant and small effect sizes of a Mediterranean diet on cognitive function, and no evidence for brain morphology or connectivity outcomes across the limited experimental evidence. However, the largest and most robustly-designed trial included in the review provided evidence for attenuation of cognitive decline over four years of ageing, compared to a healthy lower fat diet. This finding persisted, even with the recently published corrections within the original paper for a sub-cohort of this study (Valls-Pedret et al., 2015). Therefore, taken in conjunction with the considerable evidence from observational studies, this systematic review indicated great scope for additional clinical trials in this area. It has also contributed to the body of knowledge by providing detailed recommendations for methodological improvements in future trials.

7.1.3 Development of a new empirically-based adherence tool

Chapter Four reviewed the limitations of existing Mediterranean diet index tools. It also described the method for development of a new tool to address these limitations and assess adherence to the ‘traditional’ dietary pattern, Mediterranean foods and aspects of cuisine (e.g., limited snacking, fasting) within Western populations. While no currently available dietary assessment method is perfect or ideal for all circumstances, and difficulties remain in accurately assessing specific food intakes (e.g., vegetables), a diet index tool was selected as the best fit for purpose to assess the unique pattern and cuisine aspects previously missed by other tools. This novel tool was named Mediterranean Diet and Culinary Index (MediCul) to signify the inclusion of cuisine elements, some of which are novel.

MediCul is a 50-item survey tool spanning 17 main dietary elements, within which foods and cuisine aspects have been categorised using professional judgement. These include nine desirable elements of the ‘traditional’ Mediterranean diet and four undesirable elements of a Western diet. A measure of exposure to non-traditional foods is important for a tool to be used within Westernised countries. MediCul is scored between 0 and 100 with higher scores representing greater adherence to the dietary pattern.

383 Importantly, the elements and cut-off points used in MediCul were informed by empirical data on the ‘traditional’ Mediterranean diet and cuisine (Chapter Two). Therefore, MediCul enables assessment of adherence to a ‘traditional’ Mediterranean dietary pattern, to provide greater discernment in future with other plant-based dietary patterns. A practical feature of MediCul is that a score for the popular Spanish Mediterranean Diet Adherence Screener (MEDAS) can also be derived. This may be useful for comparison with previous or future studies using this screener.

A major contribution of this thesis is to make MediCul freely available online as Supplementary Material published in the British Journal of Nutrition, for use in research and education initiatives to help progress the field (Chapter Four).

7.1.4 Tool validation in two cohorts for online and in-person use

While most Mediterranean diet index scores are derived indirectly from a food frequency questionnaire (FFQ), scoring for MediCul is derived directly from the MediCul tool. This new tool has also been validated for online (Chapter Five) and in-person (Chapter Six) use and found to have very good reliability and moderate validity. This was done with two cohorts representing: a) middle-aged and older Australians having risk factors for cognitive decline, but no current diagnosis of dementia or neurodegenerative disease and, b) older Australians diagnosed with mild cognitive impairment (MCI). Therefore, MediCul has been tested and is suitable for use among middle-aged and older individuals within a Western country, including those at elevated risk of dementia. Although dementia has been identified as a global health priority (Shah et al., 2016) and mounting evidence suggests adoption of a Mediterranean diet may benefit cognition, to the author’s knowledge, no other published Mediterranean diet index tool, needed to assess adherence to the diet, has been validated for use within cognitively at-risk populations.

In a systematic review of 28 Mediterranean diet indexes up until December 2015, the most deficient quality identified related to their test-retest reliability (Zaragoza-Martí, Cabañero-Martínez, Hurtado-Sánchez, Laguna-Pérez, & Ferrer-Cascales, 2018).

384 Chapters Five and Six of this thesis confirmed that MediCul has very good test-retest reliability in both cohorts studied as assessed by using intra-class correlation coefficients (ICC), Bland-Altman plots and Kappa agreement within the 17 main MediCul elements. While studies often report a Pearson’s correlation coefficient, a high correlation does not necessarily mean good agreement (Bland & Altman, 1986). The ICC is a more acceptable correlation coefficient because a pooled mean and standard deviation of the data is derived and used (Bountziouka & Panagiotakos, 2010). The ICC results obtained were similar for the two cohorts as can be seen in Table 7.1. Using Bland-Altman plots, which assess actual agreement, only a minor mean difference was identified between the scores at two time points. Across both studies, 95-97% of participants fell within or on the limits of agreement (LOA) representing ±2 standard deviations from the mean difference in scores. Also, there was no systematic bias detected across the two time points. Percent agreement within the same dietary element category ranged from almost perfect to fair, depending on the element, with no elements having poor agreement in either study.

Table 7.1. Comparison of intra-class correlation coefficients (ICC) from two cohorts for test-retest reliability of the Mediterranean Diet and Culinary Index (MediCul)

ICC

MYB cohort of healthy participants ICC=0.86, 95% CI: 0.789, 0.910, p<0.0001

SMART cohort of MCI participants ICC=0·93, 95% CI: 0·884, 0·954, p<0·0001

MYB, Maintain Your Brain study; SMART, Study of Mental and Resistance Training; MCI, mild cognitive impairment.

Concurrent validity of MediCul was found to be moderate in both studies (Chapters Five and Six). The mean difference from the paired samples t-tests for MediCul scores from the new tool versus a 3-day food record (FR) was similar for the two cohorts as shown in Table 7.2. Both studies found the new tool overestimates the mean total MediCul score relative to a FR by 6% with similar LOA, ranging from under- and overestimates of 11 to 25%, respectively. It is unknown whether the approximately 20 point maximum difference between the new tool and the FR derived MediCul score may be clinically

385 meaningful. However, a certain degree of disagreement can be expected when validating a tool reflecting usual intake against a reference tool of actual intake over a limited number of days (Cade, Thompson, Burley, & Warm, 2002). Despite being considered a ‘gold standard’ method, food records may also under-report intake or include atypical days (Thompson & Subar, 2017).

Table 7.2. Comparison of the mean difference from paired samples t-tests for Mediterranean Diet and Culinary Index (MediCul)* scores from the tool versus a 3-day food record (FR) in two cohorts

(Mean differences and 95% confidence intervals (CI))

Mean 95% CI of the P value difference difference MYB cohort MediCul tool versus FR† 5.32 3.03, 7.62 <0.0001

SMART cohort MediCul tool versus FR‡ 5.95 3.85, 8.05 <0.0001

MYB, Maintain Your Brain study; SMART, Study of Mental and Resistance Training. *Represents first administration of MediCul to the study participants; †n=68; ‡n=65.

Nevertheless, scores derived from the MediCul tool were closer to the reference method used in validity testing than what has been reported for MEDAS, a popular short screener (Schroder et al., 2011). When validated against a FFQ within a Spanish population, MEDAS was found to under- and overestimate scores by 43% and 53%, respectively. When validated against a FR within a UK cohort, the MEDAS tool under- and overestimated scores by 16% and 36%, respectively (Papadaki et al., 2018). By comparison, the MediCul index tool score deviations from the FR were only 11 to 25%.

Validity of the MediCul tool was further tested indirectly by linear nutrient trends, which were observed in the expected direction for nutrients from the FR and tertiles of the MediCul score from the tool or FR (Chapters Five and Six). For example, in both studies,

386 saturated fatty acids as % fat and the unsaturated to saturated fat ratio from the FR, increased with increasing tertiles of MediCul score derived from both the tool and the FR. There was also good discrimination between the first and third tertiles of the MediCul score for the unsaturated to saturated fat ratio in both studies. Hence, a higher MediCul score reflects the more unsaturated nature of the fat in the Mediterranean diet, as compared to a Western diet. Finally, a higher MediCul score was associated with a higher quality diet, based on various nutrient trends and intakes of healthy and unhealthy foods.

7.2 Limitations and strengths

Certain limitations of this thesis deserve consideration. Dietary, brain, and cardiometabolic outcomes in relation to the MediCul tool are lacking to support predictive validity. However, these data were either unavailable or beyond the scope of the work possible for this thesis. Examination of some outcomes is already planned in other research. For example, the main Maintain Your Brain (MYB) study (Heffernan et al., 2018), a 3-year randomised controlled trial with 6236 participants, will be utilising MediCul at baseline and annual follow up. This study will generate data for multiple outcomes, including chronic disease risk factors, which will be analysed in relation to MediCul scoring.

Different modes of administration were used for the MediCul index tool in the two cohorts. The MediCul tool was administered in-person to participants from the Study of Mental and Resistance Training (SMART), allowing for greater engagement and the ability to ask questions of the dietitian and clarify meaning, whereas it was administered online to MYB participants, without a researcher present. Online use of MediCul may reduce accountability and the effort taken to carefully respond to questions. Further, for the MYB cohort, MediCul was administered as part of a suite of online surveys used in the MYB validation study. This may have caused fatigue and influenced the results. While dietitian administration of MediCul may be best for certain sub-groups, future research may be needed to directly compare in-person and online administration within the same individuals in order to investigate potential differences in scoring due to the mode of

387 administration. Therefore, until more information is available, only one mode of administration should be recommended within a single study or practice setting.

It is unknown how well MediCul may perform across various MCI cohorts, as well as poorly educated or ethnically diverse groups. Individuals with MCI are a clinically and neuropathologically heterogeneous group (DeCarli, 2003). Hence, the SMART cohort investigated within this thesis may have responded better than some other MCI cohorts. While poorly educated individuals were not intentionally excluded, the included participants in both studies had more than primary school education, which may be the maximum education level in some other cohorts. The effect of ethnicity on response outcomes was also not studied. Although Australia is a multi-cultural society, the majority of volunteers for the validation studies were Caucasian, precluding any analysis of responses stratified by ethnicity. Future studies targeting multiple ethnic groups and geographical locations are necessary to determine the performance characteristics of MediCul within various populations.

The participants involved in both validation studies consumed predominantly Western diets, as indicated by their moderately low MediCul scores (MYB=56.1/100.0; SMART=54.6/100.0) and poor compliance with thresholds for healthy Mediterranean foods. Hence, there was a relative lack of scoring for MediCul at the higher end. Although it was inevitable that some participants would score poorly for questions relating to foods atypical in the Australian diet, e.g., extra virgin olive oil, it may be expected that such scoring would increase during an intervention as participants learned to adopt the Mediterranean diet. Either way, it would be useful in future to test the performance of MediCul within a Mediterranean population as some participants would presumably achieve scores at the higher end. However, adherence to the Mediterranean diet has been declining in Mediterranean countries (León-Muñoz et al., 2012), although this trend may have stabilised in some regions (Bibiloni et al., 2017). A wider range of responses to MediCul may also be provided by Mediterranean migrants who have settled more recently in Australia.

388 The validity of a derived MEDAS score from the MediCul tool, which is not based on the ‘traditional’ Mediterranean diet discussed in this thesis but widely used in various geographical locations as it is a simple 14-item screener for a Mediterranean diet, was not tested as this was not an aim of the thesis. Interestingly, MEDAS scores derived from validated FFQs have previously been reported in relation to health outcomes without further validity testing (Alvarez-Alvarez et al., 2018; Livingstone et al., 2016).

Finally, MediCul is not designed to assess nutrient intakes or broader lifestyle factors, such as the Mediterranean Lifestyle (MEDLIFE) index (Sotos-Prieto et al., 2014) or the Total Lifestyle Index (TLI) (Anastasiou et al., 2018), although it includes some cultural and lifestyle elements related to diet and eating behaviour. The total Mediterranean lifestyle may be even more important for health and wellbeing than the diet alone.

Strengths of this thesis include a review of historical data sources to describe the ‘traditional’ Mediterranean cuisine. Also, the use of a robust and pre-registered protocol to conduct the systematic review on the effect of the Mediterranean diet on cognition and brain morphology and function within clinical trials. This review has been published in a leading nutrition journal for wide knowledge dissemination (Radd-Vagenas et al., 2018).

It has been suggested that a more precise definition of the Mediterranean diet is needed to be able to find consistent health effects across studies (Struijk, Guallar-Castillón, Rodríguez-Artalejo, & López-García, 2018). Also, since all the features of the ‘traditional’ cuisine may contribute to maintenance of cognitive function (Barberger-Gateau, Samieri, & Féart, 2015), that an attempt should be made to assess these. The new MediCul tool was developed using elements and cut-off points based on ‘traditional’ intakes, eating habits and cooking methods using professional judgement based on historical data rather than a posteriori statistical derivations, which may be population-specific and not reflect the ‘traditional’ diet. MediCul can therefore provide a more detailed and precise assessment of adherence to the ‘traditional’ diet and aspects of cuisine, including elements not previously studied, compared to existing tools. 389 MediCul is quicker to use as a tool than a typical FFQ but more comprehensive than a short screener. While short screeners may be useful to elicit behaviour change within interventions, they may not capture extreme levels of intake, leading to overestimation of associations with health outcomes (Panagiotakos, Pitsavos, & Stefanadis, 2006). When validated in an UK cohort, MEDAS was found to be limited in elements assessing exposures to common and harmful Western foods (Papadaki et al., 2018). Larger scale diet index tools may therefore provide increased diagnostic accuracy compared to smaller scale tools (Kourlaba & Panagiotakos, 2009).

Compared to the Mediterranean Diet Score (MedDietScore) (Panagiotakos et al., 2006), which has mostly been calculated by extracting data from independent FFQs, use of the MediCul tool enables scores to be directly derived. Direct score derivation may avoid the introduction of errors with various assumptions needing to be made when converting data captured in a FFQ. Calculating a Mediterranean score directly from a tool that supplies the questions of interest has been associated with improved repeatability, whereas an indirectly calculated score could result in overestimation of adherence (Bountziouka et al., 2012). However, the use of FFQs to derive a Mediterranean diet adherence score is widespread. It is also unlikely to diminish in the near future given the availability of data sets from existing FFQs for secondary analysis and the popularity of certain FFQs for use in observational studies. Thus, indirectly derived Mediterranean diet adherence scores will continue to play a role but their limitations should be acknowledged.

Finally, the use of two cohorts and two different modes of administration to test the reliability and validity of MediCul enhances the importance of the results and enables the tool to be used in future studies with people who have risk factors for dementia or a clinical diagnosis of MCI.

390 7.3 Implications

Various plant-based dietary patterns (Hillesund, Bere, Haugen, & Øverby, 2014; Kwan et al., 2013; Martinez-Gonzalez et al., 2014; Morris et al., 2015; Willett et al., 2019) and those based on dietary guidelines (Croll et al., 2018; McCullough et al., 2002; Zang et al., 2018) may be beneficial for healthier ageing, with less chronic disease, including for the brain. However, to date, most evidence exists to support a Mediterranean dietary pattern (Chen, Maguire, Brodaty, & O'Leary, 2019; Scarmeas, Anastasiou, & Yannakoulia, 2018), although the ‘traditional’ Mediterranean diet has not been adequately studied. Therefore, a validated Mediterranean diet index tool, that has overcome common limitations of existing tools and assesses adherence to a ‘traditional’ dietary pattern and aspects of cuisine, could be useful and a timely contribution for future research, especially within Western countries.

7.3.1 Future research

Further trials are warranted on the effect of the Mediterranean diet in preventing cognitive decline as well as metabolic and morphological changes in the brain associated with dementia. These trials should include testing of the ‘traditional’ Mediterranean diet (Trichopoulou et al., 2014) as an intervention. Until more objective biomarkers of adherence to this diet and cuisine are available (Playdon et al., 2017), such as a Mediterranean food metabolome mentioned in Chapter Three, various outcomes could be examined in relation to changes in MediCul scoring, as has been done for the MEDAS screener (Valls-Pedret et al., 2015). Future studies should also further investigate the mechanisms of action for the Mediterranean diet as discussed in Chapter Three, particularly in relation to the microbiome (Jin et al., 2019).

The MediCul tool may be attractive for multi-domain interventions incorporating the Mediterranean diet, in order to reduce participant and researcher burden since multiple assessments are required in such studies. In addition, while many Mediterranean diet index tools have been developed specifically for epidemiological purposes (Vitale et al., 2019), MediCul may be utilised in both clinical and observational settings.

391 7.3.2 Public health and clinical significance

While more research on the Mediterranean diet is required before conclusive statements can be made in relation to brain health, this diet is already relevant within public health and clinical settings. It is promoted by the Australian and other dietary guidelines (National Health and Medical Research Council, 2013; US Department of Health and Human Services and US Department of Agriculture, 2015) as a healthy eating pattern to reduce chronic disease risk, which increases with age, based on high level evidence for cardiovascular disease (National Health and Medical Research Council, 2013). It is therefore reasonable for clinicians to recommend it to improve vascular risk factors for dementia (Zhao et al., 2018). Indeed, the World Health Organization has most recently started promoting a Mediterranean-like diet for risk reduction of cognitive decline and dementia (World Health Organization, 2019).

The Mediterranean diet has also been identified as being of particular significance for older people and an ageing population, such as in Australia, since the confirmed benefits are relevant to key areas of morbidity (Nowson, Service, Appleton, & Grieger, 2018) and survival (Bonaccio et al., 2018). In a meta-analysis of seven prospective studies of older people, a one-point increase in the Mediterranean diet score used was associated with a 5% (range: 4-19%) lower risk of overall mortality (Bonaccio et al., 2018). According to the Australian Imaging, Biomarkers and Lifestyle (AIBL) Study of Ageing, higher adherence to a Mediterranean diet may also help maintain brain health in cognitively normal older people by slowing down the rate of amyloid pathology accumulation (Rainey-Smith et al., 2018).

Notably, the middle-aged and older participants included in both validity studies within this thesis did not have high adherence to this dietary pattern, based on their MediCul and derived MEDAS scores. Also, few reached ‘traditional’ Mediterranean intake thresholds for protective foods and food combinations, such as extra virgin olive oil, legumes, fruit and sofrito. It is well-known that Australians eat an unhealthy Western diet, confirmed recently by another government report (Australian Institute of Health and Welfare, 2018). Yet Western diets are associated with increased risk of cognitive decline and dementia.

392 In one study of people aged 65 years and older, a Western dietary pattern was associated with an 8-fold increase in cognitive decline over eight years (Tsai, 2015). In another longitudinal community study of older people from Australia, a Western dietary pattern was associated with a smaller left hippocampal volume (Jacka, Cherbuin, Anstey, Sachdev, & Butterworth, 2015). This finding is concerning since the hippocampus is important for memory and learning and one of only two areas within the brain where adult neurogenesis occurs, with the left hippocampus being particularly prone to neurodegeneration. Such findings present an opportunity for nutrition education utilising the Mediterranean diet to reduce chronic disease risk generally, and possibly preserve brain structure and slow the rate of cognitive decline among middle-aged and older adults with ageing.

Therefore, development of Mediterranean food guidelines for brain health (Scarmeas et al., 2018) and practical instructions such as cooking lessons (Martinez-Gonzalez, Hershey, Zazpe, & Trichopoulou, 2017) may be beneficial for the increasingly ageing Australian and world population. Preserving cognition for as long as possible is crucial because age-related cognitive decline, even without clinical symptoms or progression to dementia, is associated with increased mortality risk, falls with serious injury (Muir, Gopaul, & Montero Odasso, 2012) and decreased quality of life (Pusswald et al., 2015).

Importantly, evidence exists that the Mediterranean diet is transferable to Australians (Murphy & Parletta, 2018). This includes men with coronary heart disease representing multiple ethnicities (Mayr, Tierney, Kucianski, Thomas, & Itsiopoulos, 2019). A Mediterranean eating pattern is also feasible for adoption by older people (Martinez- Gonzalez et al., 2017; Yannakoulia et al., 2018), and those with MCI and dementia (Hassan et al., 2018). The provision of Mediterranean meals could be of interest in the care of individuals with Alzheimer’s disease (AD) as higher adherence to the Mediterranean diet has been associated with better survival in AD (Opie, Ralston, & Walker, 2013). In one study from Spain, participants with AD were found to have poor adherence to the diet, despite residing in a Mediterranean country (Rocaspana-García, Blanco-Blanco, Arias-Pastor, Gea-Sánchez, & Piñol-Ripoll, 2018). While cost may be a

393 perceived barrier, the data suggest a Mediterranean diet is no more expensive (Lara et al., 2015), or even cheaper than a typical Australian diet (Chatterton et al., 2018; Opie et al., 2015). It could also offer a good return on investment to manage chronic disease using group sessions, such as Mediterranean cooking workshops (Segal et al., 2018).

Finally, the MediCul tool may be useful to help train individuals to increasingly adopt Mediterranean eating habits by providing feedback on current choices. It could be integrated into electronic medical records (Rasmussen et al., 2018), or practice software, to screen patients who should be routinely referred to a dietitian if they have a low MediCul score. This could save on assessment time by clinicians so that consultations/group sessions can focus more on dietary education. Tools that evaluate adherence in patients may also indirectly benefit health professionals through a learning effect. In another study from Spain, primary care physicians were found to have poor adherence to the Mediterranean diet (Sentenach-Carbo et al., 2018), which is also likely to be the case for health professionals in Western countries. Additionally, MediCul could be transformed into a web-based or mobile application to provide personalised feedback in real time with dietary tips, recipes and menus. Such an application of the MediCul tool could likewise be utilised in clinical trials (Cattaneo et al., 2018).

7.4 Conclusions

Promotion of lifestyle interventions for healthy brain ageing is critical since there is no pharmacologic treatment to prevent, delay or reverse the course of dementia. This thesis has identified promising scope for the Mediterranean diet as a treatment modality based on a systematic literature review, warranting additional, well-designed trials to examine the relationship between the Mediterranean diet and brain health. In particular, the ‘traditional’ diet should be further studied as this was originally associated with the lowest rates of chronic disease. The thesis has improved understanding of ‘traditional’ Mediterranean cuisine, underpinning the development of a novel diet index tool called MediCul that has overcome common limitations of existing tools. MediCul provides comprehensive assessment of adherence to the ‘traditional’ dietary pattern, including important and unique Mediterranean foods and cultural/lifestyle aspects related to eating

394 that have previously been missed, to enable greater discernment also with other types of plant-based dietary patterns. MediCul has been tested in two Western cohorts at increased risk of dementia, for online and in-person administration, and shown to be suitable for use among such populations that have previously been poorly studied within clinical settings, with respect to a Mediterranean diet and cognition. This new tool is now freely available for future research and education purposes. It is already being used in the large Maintain Your Brain (MYB) study for examination of the role of implementing this dietary pattern to attenuate cognitive decline over time. Availability of validated tools such as MediCul may advance understanding of the role of diet in healthy brain ageing, which is urgently required.

395 7.5 References

Alvarez-Alvarez, I., Zazpe, I., Pérez de Rojas, J., Bes-Rastrollo, M., Ruiz-Canela, M., Fernandez-Montero, A., . . . Martínez-González, M. A. (2018). Mediterranean diet, physical activity and their combined effect on all-cause mortality: The Seguimiento Universidad de Navarra (SUN) cohort. Preventive Medicine, 106, 45-52. doi:10.1016/j.ypmed.2017.09.021 Anastasiou, C., Yannakoulia, M., Kontogianni, M., Kosmidis, M., Mamalaki, E., Dardiotis, E., . . . Scarmeas, N. (2018). Mediterranean lifestyle in relation to cognitive health: Results from the HELIAD study. Nutrients, 10(10), 1557. doi:10.3390/nu10101557 Australian Institute of Health and Welfare. (2018). Nutrition across the life stages. Cat. no. PHE 227. Canberra: AIHW. Barberger-Gateau, P., Samieri, C., & Féart, C. (2015). Can the Mediterranean diet slow down age-related cognitive decline? Phytothérapie, 13(2), 96-103. doi:10.1007/s10298-015-0943-1 Bibiloni, M. D. M., González, M., Julibert, A., Llompart, I., Pons, A., & Tur, J. A. (2017). Ten-year trends (1999-2010) of adherence to the Mediterranean diet among the Balearic Islands' adult population. Nutrients 9(7), 749. doi:10.3390/nu9070749 Bland, J. M., & Altman, D. G. (1986). Statistical methods for assessing agreement between two methods of clinical measurement. The Lancet, 327(8476), 307-310. doi:10.1016/S0140-6736(86)90837-8 Bonaccio, M., Di Castelnuovo, A., Costanzo, S., Gialluisi, A., Persichillo, M., Cerletti, C., . . . Iacoviello, L. (2018). Mediterranean diet and mortality in the elderly: A prospective cohort study and a meta-analysis. The British Journal of Nutrition, 120(8), 841-854. doi:10.1017/S0007114518002179 Bountziouka, V., & Panagiotakos, D. B. (2010). Statistical methods used for the evaluation of reliability and validity of nutrition assessment tools used in medical research. Current Pharmaceutical Design, 16(34), 3770-3775. doi:10.2174/138161210794455102

396 Bountziouka, V., Bathrellou, E., Zazpe, I., Ezquer, L., Martínez-González, M.-A., & Panagiotakos, D. B. (2012). Repeatability of food frequency assessment tools in relation to the number of items and response categories included. Food and Nutrition Bulletin, 33(4), 288-295. doi:10.1177/156482651203300409 Cade, J., Thompson, R., Burley, V., & Warm, D. (2002). Development, validation and utilisation of food-frequency questionnaires – A review. Public Health Nutrition, 5(4), 567-587. doi:10.1079/PHN2001318 Cattaneo, G., Bartres-Faz, D., Morris, T. P., Sanchez, J. S., Macia, D., Tarrero, C., . . . Pascual-Leone, A. (2018). The Barcelona Brain Health Initiative: A cohort study to define and promote determinants of brain health. Frontiers in Aging Neuroscience, 10, 321. doi:10.3389/fnagi.2018.00321 Chatterton, M. L., Mihalopoulos, C., O’Neil, A., Itsiopoulos, C., Opie, R., Castle, D., . . . Jacka, F. (2018). Economic evaluation of a dietary intervention for adults with major depression (the “SMILES” trial). BMC Public Health, 18(1), 1-11. doi:10.1186/s12889-018-5504-8 Chen, X., Maguire, B., Brodaty, H., & O'Leary, F. (2019). Dietary patterns and cognitive health in older adults: A systematic review. Journal of Alzheimer's Disease, 67(2), 583-619. doi:10.3233/JAD-180468 Croll, P. H., Voortman, T., Ikram, M. A., Franco, O. H., Schoufour, J. D., Bos, D., & Vernooij, M. W. (2018). Better diet quality relates to larger brain tissue volumes The Rotterdam Study. Neurology, 90(24), E2166-E2173. doi:10.1212/WNL.0000000000005691 DeCarli, C. (2003). Mild cognitive impairment: Prevalence, prognosis, aetiology, and treatment. Lancet Neurology, 2(1), 15-21. doi:10.1016/S1474-4422(03)00262-X Hassan, S., Aguirre, E., Betz, A., Robertson, S., Sankhla, D., & Cooper, C. (2018). Evaluating the effect of Brainfood groups for people with mild cognitive impairment and mild dementia: Preliminary mixed-methodology study. BJPsych Open, 4(4), 208-214. doi:10.1192/bjo.2018.29 Hatzis, C. M., Sifaki-Pistolla, D., & Kafatos, A. G. (2015). History of the Cretan cohort of the Seven Countries Study. Hormones. Retrieved 21st December 2018 from http://www.hormones.gr/5/search.html?s=a+50-year+follow- up+of+the+Seven+countries+study&flds=all&do=search&rpp=20&x=0&y=0

397 Heffernan, M., Andrews, G., Fiatarone Singh, M. A., Valenzuela, M., Anstey, K. J., Maeder, A., . . . Brodaty, H. (2018). Maintain Your Brain: Protocol of a 3-year randomized controlled trial of a personalized multi-modal digital health intervention to prevent cognitive decline among community dwelling 55 to 77 year olds. Journal of Alzheimer's Disease, Advance online publication. doi:10.3233/JAD-180572 Hillesund, E. R., Bere, E., Haugen, M., & Øverby, N. C. (2014). Development of a New Nordic Diet score and its association with gestational weight gain and fetal growth - A study performed in the Norwegian Mother and Child Cohort Study (MoBa). Public Health Nutrition, 17(9), 1909-1918. doi:10.1017/S1368980014000421 Jacka, F. N., Cherbuin, N., Anstey, K. J., Sachdev, P., & Butterworth, P. (2015). Western diet is associated with a smaller hippocampus: A longitudinal investigation. BMC Medicine, 13(1), 215. doi:10.1186/s12916-015-0461-x Jin, Q., Black, A., Kales, S. N., Vattem, D., Ruiz-Canela, M., & Sotos-Prieto, M. (2019). Metabolomics and microbiomes as potential tools to evaluate the effects of the Mediterranean diet. Nutrients, 11(1), 207. doi:10.3390/nu11010207 Kourlaba, G., & Panagiotakos, D. (2009). The number of index components affects the diagnostic accuracy of a diet quality index: The role of intracorrelation and intercorrelation structure of the components. Annals of Epidemiology, 19(10), 692-700. doi:10.1016/j.annepidem.2009.03.019 Kwan, M. W.-M., Wong, M. C.-S., Wang, H. H.-X., Liu, K. Q.-L., Lee, C. L.-S., Yan, B. P.-Y., . . . Griffiths, S. M. (2013). Compliance with the Dietary Approaches to Stop Hypertension (DASH) diet: A systematic review. PLoS ONE, 8(10), e78412. doi:10.1371/journal.pone.0078412 Lara, J., Turbett, E., McKevic, A., Rudgard, K., Hearth, H., & Mathers, J. C. (2015). The Mediterranean diet among British older adults: Its understanding, acceptability and the feasibility of a randomised brief intervention with two levels of dietary advice. Maturitas, 82(4), 387-393. doi:10.1016/j.maturitas.2015.07.029

398 León-Muñoz, L. M., Guallar-Castillón, P., Graciani, A., López-García, E., Mesas, A. E., Aguilera, M. T., . . . Rodríguez-Artalejo, F. (2012). Adherence to the Mediterranean diet pattern has declined in Spanish adults. The Journal of Nutrition, 142(10), 1843-1850. doi:10.3945/jn.112.164616 Livingstone, K. M., Celis-Morales, C., Navas-Carretero, S., San-Cristobal, R., Macready, A. L., Fallaize, R., . . . on behalf of the Food4Me Study. (2016). Effect of an Internet-based, personalized nutrition randomized trial on dietary changes associated with the Mediterranean diet: The Food4Me Study. American Journal of Clinical Nutrition, 104(2), 288-297. doi:10.3945/ajcn.115.129049 Malhotra, A., & O'Neill, D. (2017). The Pioppi diet: A 21-day lifestyle plan (1st ed.). Great Britain: Penguin Books Ltd. Martinez-Gonzalez, M. A., Sánchez-Tainta, A., Corella, D., Salas-Salvadó, J., Ros, E., Arós, F., . . . for the PREDIMED Group. (2014). A provegetarian food pattern and reduction in total mortality in the Prevención con Dieta Mediterránea (PREDIMED) study. American Journal of Clinical Nutrition, 100(1), 320S- 328S. doi:10.3945/ajcn.113.071431 Martinez-Gonzalez, M. A., Hershey, M. S., Zazpe, I., & Trichopoulou, A. (2017). Transferability of the Mediterranean diet to non-Mediterranean countries: What is and what is not the Mediterranean diet. Nutrients, 9(11), 1226. doi:10.3390/nu9111226 Mayr, H. L., Tierney, A. C., Kucianski, T., Thomas, C. J., & Itsiopoulos, C. (2019). Australian patients with coronary heart disease achieve high adherence to 6- month Mediterranean diet intervention: Preliminary results of the AUSMED Heart Trial. Nutrition, 61, 21-31. doi:10.1016/j.nut.2018.10.027 McCullough, M. L., Feskanich, D., Stampfer, M. J., Giovannucci, E. L., Rimm, E. B., Hu, F. B., . . . Willett, W. C. (2002). Diet quality and major chronic disease risk in men and women: Moving toward improved dietary guidance. American Journal of Clinical Nutrition, 76(6), 1261-1271. doi:10.1093/ajcn/76.6.1261 Morris, M. C., Tangney, C. C., Wang, Y., Sacks, F. M., Barnes, L. L., Bennett, D. A., & Aggarwal, N. T. (2015). MIND diet slows cognitive decline with aging. Alzheimer's & Dementia, 11(9), 1015-1022. doi:10.1016/j.jalz.2015.04.011

399 Muir, S. W., Gopaul, K., & Montero Odasso, M. M. (2012). The role of cognitive impairment in fall risk among older adults: A systematic review and meta- analysis. Age and Ageing, 41(3), 299-308. doi:10.1093/ageing/afs012 Murphy, K. J., & Parletta, N. (2018). Implementing a Mediterranean-style diet outside the Mediterranean region. Current Atherosclerosis Reports, 20(6), 1-10. doi:10.1007/s11883-018-0732-z National Health and Medical Research Council. (2013). Australian dietary guidelines. Canberra, Australia: National Health and Medical Research Council. Nowson, C. A., Service, C., Appleton, J., & Grieger, J. A. (2018). The impact of dietary factors on indices of chronic disease in older people: A systematic review. The Journal of Nutrition, Health & Aging, 22(2), 282-296. doi:10.1007/s12603-017- 0920-5 Opie, R. S., Ralston, R. A., & Walker, K. Z. (2013). Adherence to a Mediterranean- style diet can slow the rate of cognitive decline and decrease the risk of dementia: A systematic review. Nutrition & Dietetics, 70(3), 206-217. doi:10.1111/1747-0080.12016 Opie, R. S., Segal, L., Jacka, F. N., Nicholls, L., Dash, S., Pizzinga, J., & Itsiopoulos, C. (2015). Assessing healthy diet affordability in a cohort with major depressive disorders. Journal of Public Health and Epidemiology, 5(5), 159-169. doi:10.5897/JPHE2014.0668 Panagiotakos, D. B., Pitsavos, C., & Stefanadis, C. (2006). Dietary patterns: A Mediterranean diet score and its relation to clinical and biological markers of cardiovascular disease risk. Nutrition, Metabolism and Cardiovascular Diseases, 16(8), 559-568. doi:10.1016/j.numecd.2005.08.006 Papadaki, A., Johnson, L., Toumpakari, Z., England, C., Rai, M., Toms, S., . . . Feder, G. (2018). Validation of the English version of the 14-item Mediterranean Diet Adherence Screener of the PREDIMED study, in people at high cardiovascular risk in the UK. Nutrients, 10(2), 138. doi:10.3390/nu10020138 Playdon, M. C., Moore, S. C., Derkach, A., Reedy, J., Subar, A. F., Sampson, J. N., . . . Stolzenberg-Solomon, R. (2017). Identifying biomarkers of dietary patterns by using metabolomics. The American Journal of Clinical Nutrition, 105(2), 450- 465. doi:10.3945/ajcn.116.144501

400 Pusswald, G., Tropper, E., Kryspin-Exner, I., Moser, D., Klug, S., Auff, E., . . . Lehrner, J. (2015). Health-related quality of life in patients with subjective cognitive decline and mild cognitive impairment and its relation to activities of daily living. Journal of Alzheimer's Disease, 47(2), 479-486. doi:10.3233/JAD- 150284 Radd-Vagenas, S., Duffy, S. L., Naismith, S. L., Brew, B. J., Flood, V. M., & Fiatarone Singh, M. A. (2018). Effect of the Mediterranean diet on cognition and brain morphology and function: A systematic review of randomized controlled trials. American Journal of Clinical Nutrition, 107(3), 389-404. doi:10.1093/ajcn/nqx070 Rainey-Smith, S. R., Gu, Y., Gardener, S. L., Doecke, J. D., Villemagne, V. L., Brown, B. M., . . . Martins, R. N. (2018). Mediterranean diet adherence and rate of cerebral Aβ-amyloid accumulation: Data from the Australian Imaging, Biomarkers and Lifestyle Study of Ageing. Translational Psychiatry, 8(1), 1-7. doi:10.1038/s41398-018-0293-5 Rasmussen, E., Fosnacht Morgan, A. M., Munson, R., Ong, A., Patel, V., Patel, S., . . . Maraganore, D. M. (2018). Use of an electronic medical record to track adherence to the Mediterranean diet in a US neurology clinical practice. Mayo Clinic Proceedings: Innovations, Quality & Outcomes, 2(1), 49-59. doi:10.1016/j.mayocpiqo.2017.12.003 Rocaspana-García, M., Blanco-Blanco, J., Arias-Pastor, A., Gea-Sánchez, M., & Piñol- Ripoll, G. (2018). Study of community-living Alzheimer's patients' adherence to the Mediterranean diet and risks of malnutrition at different disease stages. PeerJ, 6, e5150. doi:10.7717/peerj.5150 Scarmeas, N., Anastasiou, C. A., & Yannakoulia, M. (2018). Nutrition and prevention of cognitive impairment. The Lancet Neurology, 17(11), 1006-1015. doi:10.1016/S1474-4422(18)30338-7 Schroder, H., Fito, M., Estruch, R., Martinez-Gonzalez, M. A., Corella, D., Salas- Salvado, J., . . . Covas, M. I. (2011). A short screener is valid for assessing Mediterranean diet adherence among older Spanish men and women. Journal of Nutrition, 141(6), 1140-1145. doi:10.3945/jn.110.135566

401 Segal, L., Twizeyemariya, A., Zarnowiecki, D., Niyonsenga, T., Bogomolova, S., Wilson, A., . . . Parletta, N. (2018). Cost effectiveness and cost-utility analysis of a group-based diet intervention for treating major depression – the HELFIMED trial. Nutritional Neuroscience, Advance online publication. doi:10.1080/1028415X.2018.1556896 Sentenach-Carbo, A., Batlle, C., Franquesa, M., García-Fernandez, E., Rico, L., Shamirian-Pulido, L., . . . Bach-Faig, A. (2018). Adherence of Spanish primary physicians and clinical practise to the Mediterranean diet. European Journal of Clinical Nutrition, Advance online publication. doi:10.1038/s41430-018-0314-8 Shah, H., Albanese, E., Duggan, C., Rudan, I., Langa, K. M., Carrillo, M. C., . . . Dua, T. (2016). Research priorities to reduce the global burden of dementia by 2025. The Lancet Neurology, 15(12), 1285-1294. doi:10.1016/S1474-4422(16)30235-6 Sotos-Prieto, M., Moreno-Franco, B., Ordovás, J. M., León, M., Casasnovas, J. A., & Peñalvo, J. L. (2014). Design and development of an instrument to measure overall lifestyle habits for epidemiological research: The Mediterranean lifestyle (MEDLIFE) index. Public Health Nutrition, 18(6), 959-967. doi:10.1017/S1368980014001360 Struijk, E. A., Guallar-Castillón, P., Rodríguez-Artalejo, F., & López-García, E. (2018). Mediterranean dietary patterns and impaired physical function in older adults. The Journals of Gerontology. Series A, 73(3), 333-339. doi:10.1093/gerona/glw208 Thompson, F. E., & Subar, A. F. (2017). Chapter 1 - Dietary assessment methodology. In A. M. Coulston, C. J. Boushey, M. G. Ferruzzi, & L. M. Delahanty (Eds.), Nutrition in the prevention and treatment of disease (4th ed., pp. 5-48). London, UK: Academic Press. Trichopoulou, A., & Lagiou, P. (1997). Healthy traditional Mediterranean diet: An expression of culture, history, and lifestyle. Nutrition Reviews, 55(11), 383-389. doi:10.1111/j.1753-4887.1997.tb01578.x Trichopoulou, A., & Vasilopoulou, E. (2000). Mediterranean diet and longevity. British Journal of Nutrition, 84(S2), S205-S209. doi:10.1079/09658219738855

402 Trichopoulou, A., Martinez-Gonzalez, M. A., Tong, T. Y. N., Forouhi, N. G., Khandelwal, S., Prabhakaran, D., . . . de Lorgeril, M. (2014). Definitions and potential health benefits of the Mediterranean diet: Views from experts around the world. BMC Medicine, 12(112). doi:10.1186/1741-7015-12-112 Tsai, H.-J. (2015). Dietary patterns and cognitive decline in Taiwanese aged 65 years and older: Dietary patterns and cognitive decline. International Journal of Geriatric Psychiatry, 30(5), 523-530. doi:10.1002/gps.4176 United Nations Educational Scientific and Cultural Organization. (2010). Convention for the safeguarding of the intangible cultural heritage. Retrieved 21st December 2018 from www.unisob.na.it/ateneo/c002/unesco_nomination_file_rl2010.pdf US Department of Health and Human Services and US Department of Agriculture. (2015). 2015-2020 Dietary guidelines for Americans, 8th Edition. Washington DC: US Government Printing Office. Valls-Pedret, C., Sala-Vila, A., Serra-Mir, M., Corella, D., de la Torre, R., Martínez- González, M. Á., . . . Salas-Salvadó, J. (2015). Mediterranean diet and age- related cognitive decline: A randomized clinical trial. JAMA Internal Medicine, 175(7), 1094-1103. doi:10.1001/jamainternmed.2015.1668 Vitale, M., Racca, E., Izzo, A., Giacco, A., Parente, E., Riccardi, G., & Giacco, R. (2019). Adherence to the traditional Mediterranean diet in a population of South of Italy: Factors involved and proposal of an educational field-based survey tool. International Journal of Food Sciences and Nutrition, 70(2), 197-203. doi:10.1080/09637486.2018.1481202 Wahlqvist, M. L., Kouris-Blazos, A., Trichopoulou, A., & Polychronopoulos, E. (1991). The wisdom of the Greek cuisine and way of life: Comparison of the food and health beliefs of elderly Greeks in Greece and Australia. Age and Nutrition. Retrieved 23rd December 2018 from http://apjcn.nhri.org.tw/server/markwpapers/papers.htm Willett, W., Rockström, J., Loken, B., Springmann, M., Lang, T., Vermeulen, S., . . . Murray, C. J. L. (2019). Food in the Anthropocene: the EAT–Lancet Commission on healthy diets from sustainable food systems. The Lancet, 393(10170), 447-492. doi:10.1016/S0140-6736(18)31788-4

403 World Health Organization. (2019). Risk reduction of cognitive decline and dementia: WHO guidelines. Retrieved 7th June 2019 from https://www.who.int/mental_health/neurology/dementia/guidelines_risk_reducti on/en/ Yannakoulia, M., Mamalaki, E., Anastasiou, C. A., Mourtzi, N., Lambrinoudaki, I., & Scarmeas, N. (2018). Eating habits and behaviors of older people: Where are we now and where should we go? Maturitas, 114, 14-21. doi:10.1016/j.maturitas.2018.05.001 Zang, J., Guo, C., Wang, Z., Cheng, Y., Jin, W., Zhu, Z., . . . Wu, F. (2018). Is adherence to the Chinese Dietary Guidelines associated with better self-reported health?: The Chinese dietary guidelines adherence score. Asia Pacific Journal of Clinical Nutrition, 27(4), 914-924. doi:10.6133/apjcn.072018.02 Zaragoza-Martí, A., Cabañero-Martínez, M. J., Hurtado-Sánchez, J. A., Laguna-Pérez, A., & Ferrer-Cascales, R. (2018). Evaluation of Mediterranean diet adherence scores: A systematic review. BMJ Open, 8(2), e019033. doi:10.1136/bmjopen- 2017-019033 Zhao, C., Noble, J. M., Marder, K., Hartman, J. S., Gu, Y., & Scarmeas, N. (2018). Dietary patterns, physical activity, sleep, and risk for dementia and cognitive decline. Current Nutrition Reports, 7(4), 335-345. doi:10.1007/s13668-018- 0247-9

404 APPENDIX

405 LIST OF APPENDICES

APPENDIX ONE: Conference posters and proceedings…………………………..408 Joint Annual Scientific Meeting of the Nutrition Society of New Zealand and the Nutrition Society of Australia, 2015…………………………………………………409 10th Asia-Pacific Conference on Clinical Nutrition, 2017………………………...... 410 7th International Congress on Vegetarian Nutrition, 2018………………………...... 411 Dietitians Association of Australia 35th National Conference, 2018………………...412 Nutrition Society of Australia 42nd Annual Scientific Meeting, 2018……………….413

APPENDIX TWO: Additional works not forming part of this thesis…………...... 414 Peer-reviewed paper……………………………………………………………...... 415 Published abstract……………………………………………………………...…….432

APPENDIX THREE: ……………………………………………………………...... 433 PROSPERO protocol for systematic review………………………………………....434

APPENDIX FOUR: Corrections within PREDIMED for cognitive outcomes…...440 Table 4.1. Comparison of baseline cognitive test scores and changes by study group – original values……………………………………………………………………….441 Table 4.2. Comparison of baseline cognitive test scores and changes by study group – corrected values……………………………………………………………………...445 Table 4.3. Comparison of composite cognitive domains – original and corrected values………………………………………………………………………………...449

APPENDIX FIVE: ………………………………………………………………...... 451 Comparison of selected Mediterranean diet index tools………………………….....452

APPENDIX SIX: …………………………………………………………………….493 Mediterranean Diet Adherence Screener (MEDAS)………………………………...494

APPENDIX SEVEN: ………………………………………………………………..497 Screen shots of sample MediCul questions formatted for online administration……..498

406 APPENDIX EIGHT: Ethics letters………………………………………………….500 MYB trial………………………………………………………………………….....501 SMART trial………………………………………………………………………....503

APPENDIX NINE: Research Food Diary app instructions……………………...... 504 Instructions on how to keep a 3 day food diary……………………………………….505

APPENDIX TEN: ……………………………………………………………………524 Dietitian assisted food record checklist………………………………………………525

APPENDIX ELEVEN: ……………………………………………………………...527 Front sheet for validation and reliability testing of MediCul………………………....528

APPENDIX TWELVE: 3-day paper food record template……………………...... 529 3-day paper food record……………………………………………………………...530

APPENDIX THIRTEEN: FAQ on completing the 3-day food record..…………...539 Frequently asked questions on completing the 3-day food record…………………...540

APPENDIX FOURTEEN: Challis Bequest Society lunch……………………...... 542 Challis Bequest Society lunch program...... …………………………….……………543

407 APPENDIX ONE

Conference posters and proceedings

408 What is the traditional Mediterranean diet?

1S Radd-Vagenas, 2A Kouris-Blazos, 1M Fiatarone Singh, 1,3V Flood 1Faculty of Health Sciences, University of Sydney, 2Department of Dietetics and Human Nutrition, La Trobe University, 3St Vincent’s Hospital, Sydney [email protected] @CulinaryMed

Background: The Mediterranean diet is beneficial for health and used in research and clinical practice. It is associated with a reduced risk of chronic disease, including CVD, cancer and neurodegenerative disease and increased survival.1 However, this dietary pattern has proven difficult to define as it is surprisingly complex and varies between countries and time periods. Variations exist in definitions provided by educational tools such as Mediterranean diet pyramids and index tools used to measure adherence to this diet.

Aim: The aim of this study is to describe the ‘traditional’ Mediterranean diet and identify additional elements relating to cuisine and eating habits not covered by existing educational and index tools.

Results: • Existing Mediterranean diet pattern index tools have been used in observational studies and clinical trials • Mediterranean diet definitions vary in the literature Method: to assess the effect of dietary patterns on disease and focus mostly on the relative proportion of outcomes but most only include criteria relating to • We conducted a literature review of Mediterranean characterising foods. Disparities exist in the specific foods and, in some cases, nutrients. diet definitions, including seminal research studies criteria, cut offs and scoring system used by • Mediterranean Diet Adherence Screener (MEDAS) such as the Seven Countries Study and educational common Mediterranean diet pattern index tools. used in the landmark PREDIMED trial is an and index tools.2 • Most researchers agree the ‘traditional’ advanced version of earlier tools incorporating • Papers were sourced from nine databases searched Mediterranean diet is a plant-based dietary pattern elements of cuisine. However, it measures a limited for a systematic review on a related topic from common to the olive growing regions of the number of traditional foods and does not assess inception to July 2015. Mediterranean documented in the late 1950s and frequency of home cooked meals, cooking methods, early 1960s on the island of Crete, Greece and • The definition of the Mediterranean diet in the eating in company, snacking occasions, fasting, rural villages in Southern Italy such as Nicotera, 5 Candidature Dossier submitted to United Nations home garden use and napping. See Table 2. which were part of the Seven Countries Study.2 Educational, Scientific and Cultural Organization Table 2: Components assessed in the 14-point MEDAS tool (UNESCO) was also reviewed.3 • This pattern is high in fruits, vegetables, cereals, legumes, nuts/seeds and extra virgin olive oil; 1. Olive oil asmainfat 8. Wine 2. Oliveoilquantity 9. Legumes moderate in fish/shellfish and red wine; and low in 3. Vegetables 10. Fish/seafood meat, dairy, eggs and animal fats. 4. Fruit 11. Sweets 5. Red&processedmeat 12. Nuts • UNESCO supports a broader definition of the 6. Butter,margarine&cream 13. Whitemeatpreference Mediterranean diet incorporating culinary practices 7. Sugarydrinks 14. Sofrito and consumption habits since ‘diet’ (diaita) in the Greek language means a ‘way of life’. • MEDLIFE is the most holistic index tool identified, based on the updated Mediterranean diet pyramid. • We identified additional elements of the traditional However, it lacks the same elements as the Mediterranean diet, which have not been educational tool which it is based on.6 consistently or collectively documented in previous tools. See Table 1.

Table 1: Elements of the traditional Mediterranean diet absent in existing tools Conclusions: Practices relating to cuisine and eating The traditional Mediterranean diet is much broader than just a collection of foods. It includes a cuisine 1. Frequent intake of home cooked meals with unique recipes and cooking methods and other 2. Moist, lower temperature, cooking psycho-social factors that may additionally impact on methods food and nutrient intake. Scope exists for improved 3. Eating main meals in company educational and index tools for use by researchers and Figure 1: Updated Mediterranean diet pyramid, Mediterranean 4. Reduced snacking occasions clinicians. Existing tools should be reviewed to Diet Foundation, Barcelona, Spain4 5. Regular fasting incorporate broader elements of the Mediterranean 6. Home garden to grow vegetables ‘way of life’ that may influence dietary adherence and/or provide independent health effects. 7. Use of unique traditional foods 8. Napping after the midday meal References: 1. Sofi F et al. (2014) Mediterranean diet and health status: an updated meta-analysis and a Did you know? proposal for a literature-based adherence score. Public Health Nutr, 17:2769-2782. 2. Kromhout D et al. (1989) Food consumption patterns in the 1960s in seven countries. Am J • The recently updated Mediterranean diet pyramid Clin Nutr, 49:889-894. The Mediterranean diet is the only considers a wide range of lifestyle elements but 3. UNESCO (2010) Convention for the safeguarding of the intangible cultural heritage. www.unisob.na.it/ateneo/c002/unesco_nomination_file_rl2010.pdf dietary pattern protected by UNESCO. this educational tool does not emphasise the 4. Bach-Faig A et al. (2011) Mediterranean diet pyramid today. Science and cultural updates. potentially important role of type of cooking Public Health Nutr, 14:2274-2284. In 2010, UNESCO recognised the 5. Schroder H et al. (2011) A short screener is valid for assessing Mediterranean diet methods, fasting habits and ownership of a home adherence among older Spanish men and women. J Nutr, 141:1140-1145. Mediterranean diet as an Intangible garden. See Figure 1. 6. Sotos-Prieto M et al. (2014) Design and development of an instrument to measure overall lifestyle habits for epidemiological research: the Mediterranean Lifestyle (MEDLIFE) index. Cultural Heritage. Public Health Nutr, 18:959-967. 409 Proceedings from the 10th Asia-Pacific Conference on Clinical Nutrition, Adelaide, 26- 29th November, 2017.

Citation:

Radd-Vagenas, S., Duffy, S. L., Naismith, S. L., Brew, B. J., Flood, V. M., & Fiatarone Singh, M. A. (2018). Effect of the Mediterranean diet on cognition and brain morphology/function: A systematic review and meta-analysis of RCTs. Proceedings, 2(573), 38. doi:10.3390/proceedings2120573

410 Could a plant based diet affect cognition? Findings from a systematic review of randomized controlled trials on the Mediterranean diet

Sue Radd-Vagenas, Shantel L Duffy, Sharon L Naismith, Bruce J Brew, Victoria M Flood, Maria Fiatarone Singh [email protected] @CulinaryMed www.foodasmedicine.cooking

Background Results – study characteristics Every 3 seconds somebody in the world Table 1. Study characteristics at baseline Study Study design & n (% F) Mean Cognitive Health status Comparative dies from dementia. Dementia is a leading duration age (y) status at condition/s (SD) baseline cause of death and disability without 1. Wardle (2000), Parallel, 12 wk 176 53.01 Not specified CV risk: total Waiting list & Britain (52) (10.1) ĐŚŽůĞƐƚĞƌŽůшϱ͘ϮŵD ǆƉĞƌŝŵĞŶƚĂů treatment to prevent, slow or reverse its low fat diet 2. McMillan Parallel, 10 d 25 21.10 Not specified Healthy Usual diet course. Observational studies suggest (2011), Australia (100) (3.3) 3. PREDIMED sub-studies (2011-2015) plant-based diets may benefit cognition. In - Sanchez- Post-test 243 67.85 Not specified CV ƌŝƐŬ͗ƚǇƉĞϮDŽƌLow fat diet Villegas (2011), ĐŽŵƉĂƌŝƐŽŶ at 3 y (51) (6.0) шϯŵĂũŽƌsƌŝƐŬ particular, a Mediterranean (Med) style diet Spain factors -Martinez- Post-test 285 67.34 Not specified CV ƌŝƐŬ͗ƚǇƉĞϮDŽƌLow fat diet has been associated with a reduced risk of Lapiscina ĐŽŵƉĂƌŝƐŽŶ at 6.5 y (55) (5.7) шϯŵĂũŽƌsƌŝƐŬ cognitive decline, dementia, mild cognitive (2013a), Spain factors -Martinez- Post-test 522 67.38 Not specified CV ƌŝƐŬ͗ƚǇƉĞϮDŽƌLow fat diet impairment (MCI), conversion of MCI to Lapiscina ĐŽŵƉĂƌŝƐŽŶ at 6.5 y (55) (5.7) шϯŵĂũŽƌsƌŝƐŬ (2013b), Spain factors dementia and a reduced risk of mortality -Martinez- Post-test 522 67.00 Not specified CV ƌŝƐŬ͗ƚǇƉĞϮDŽƌLow fat diet Lapiscina (2014), ĐŽŵƉĂƌŝƐŽŶ at 6.5 y (56) (6.0) шϯŵĂũŽƌsƌŝƐŬ among those with Alzheimer’s disease. Spain factors - Valls-Pedret Parallel, 4.1 y 447 66.73 EŽƌŵĂů CV ƌŝƐŬ͗ƚǇƉĞϮDŽƌLow fat diet (2015), Spain (51) (5.5) шϯŵĂũŽƌsƌŝƐŬ factors 4. Lee (2015), Cross-over, 10 d 24 25.60 EŽƌŵĂů Healthy Usual diet Australia (100) (5.1) 5. Knight (2016), 2-cohort parallel, 166 72.10 EŽƌŵĂů Healthy Usual diet Australia 24 wk (53) (5.0)

Results – study quality • 0RGHUDWHRYHUDOOZLWKVWXGLHVVFRULQJ on modified PEDro scale (range: 4-8/10). Results – brain morphology/function • The Mediterranean diets varied considerably • No studies reported brain morphology between studies, and in important ways, outcomes. including a) mentioning b) advising a quantity • Post-test comparison at 3 years in and c) meeting minimum criteria for key PREDIMED found a 78% reduced risk of elements that describe the ‘traditional’ dietary low plasma BDNF – a neuroplasticity pattern (Table 1). Y=yes; N=no. biomarker – in the Med diet group. Table 2. Examples of variable intervention elements Aim Element Wardle (2000) McMillan (2011) PREDIMED Study Lee (2015) Knight (2016) Conclusions To perform the first published systematic review, to our knowledge, of randomized controlled trials 1. Inconsistent empirical evidence exists for DĞŶƚŝŽŶĞĚ Qty Advised DĞƚƌŝƚĞƌŝĂ DĞŶƚŝŽŶĞĚ Qty Advised DĞƚƌŝƚĞƌŝĂ DĞŶƚŝŽŶĞĚ Qty Advised DĞƚƌŝƚĞƌŝĂ DĞŶƚŝŽŶĞĚ Qty Advised DĞƚƌŝƚĞƌŝĂ DĞŶƚŝŽŶĞĚ Qty Advised DĞƚƌŝƚĞƌŝĂ significant benefits of a Med diet for (RCTs) investigating the effects of a Med diet on Olive oil N NNNNNYY Y NNNY Y Y cognitive function across trials. cognition or brain morphology/function, with Nuts NNNY NNY Y Y Y NNY Y Y additional focus on how interventions relate to the Legumes NNNNNNY Y Y NNNY Y N 2. No empirical data has been reported for ‘traditional’ Med dietary pattern, which was Veges YNNYNNYYYYNNYYY brain morphology. predominantly plant-based. Fruit Y NNY NNY Y Y Y NNY Y Y 3. But significant and clinically meaningful

Grains NNNY NNNNNNNNY Y Y ESs have been found for cognitive domain composites in the most robustly designed Method Fish Y NNY NNY YY Y NNY Y Y trial (PREDIMED). • Search: 9 databases (inception to Dec 2017), using Water NNNNNNNNNNNNNNN pre-defined protocol registered with PROSPERO. 4. Further well designed RCTs are required to test the effect of a ‘traditional’ Med diet • Inclusion: RCT design or post-test comparisons within an RCT; Med diet compared to alternate diet; Results - cognition (high plant to animal ratio) using 10 days duration; any aspect of cognition or brain standardized neuropsychological tests, morphology/function. • Analyses based on 66 cognitive tests biomarkers and neuroimaging. (n=1888) from 5 included studies. • Exclusion: acute exposures; single foods; partial 5. In the absence of treatment/cure for Med diet. • Only 8/66 (12.1%) of individual outcomes at dementia, reducing lifestyle and metabolic • Quality: modified PEDro scale + 7 additional trial level were significant in favour of Med risk factors takes on added importance. indicators, including 19 elements describing the diet [ESs ranged from small (0.32) to large ‘traditional’ Med diet. 6. The Med diet is already proven to reduce (1.66)]; 2/66 favoured control. vascular risk factors and primary CVD, and • Statistics: between group Effect Sizes (ESs). • Data limitations precluded a meta-analysis. recommended by various dietary guidelines as a healthy eating pattern to prevent and • Two studies reported composite z scores for Figure 1: PRISMA flow diagram of study selection procedure better manage chronic disease. 8 domains with all 3/8 from PREDIMED Outcome Terms being significant: brain-derived neurotrophic factor, CT scan*, executive function, gray matter, grey matter, lumbar puncture, magnetic resonance imaging, mild cognitive impairment, Intervention Terms neurological test*, neuropsychological test*, post- Radd-Vagenas S, Duffy SL, Naismith SL et al, Effect Cretan diet, Mediterranean diet, mortem brain examination, problem solving, reaction • Memory (ES=0.39 for Med Nuts), Mediterranean dietary pattern*, time, spinal puncture, spinal tap, white matter, Mediterranean type diet, Alzheimer, BDNF, brain, cerebral, cerebrovascular of the Mediterranean diet on cognition and brain MedDiet, MeDi, MeDiet. cognition, cognitive, cortical, dement*, fMRS, intellect*, • Frontal cognition (ES=1.02 for Med EVOO intelligen*, IQ, memory, mental, MRI, neural, neuro*, morphology and function: a systematic review of neurologic, neurotransmitter*, thinking, tomography, visuo-spatial. & 0.72 for Med Nuts), randomized controlled trials. AJCN 2018 (in press).

Search results n • Global cognition (ES=1.29 for Med EVOO = 6385 Duplicates removed n & 1.12 for Med Nuts). = 1752 References After removal of duplicates n = 4633 • Post-test comparison at second PREDIMED 1. tĂƌĚůĞ:͕ZŽŐĞƌƐW͕:ƵĚĚWĞƚĂů͕ŵ:DĞĚϮϬϬϬ͖ϭϬϴ͗ϱϰϳ-53. Excluded on basis of title or abstract 2. DĐDŝůůĂŶ>͕KǁĞŶ>͕<ƌĂƐDĞƚĂů͕ƉƉĞƚŝƚĞϮϬϭϭ͖ϱϲ͗ϭϰϯ-47. n = 4603 site showed mixed results for diet-gene Retrieved for evaluation n 3. Sanchez-sŝůůĞŐĂƐ͕'ĂůďĞƚĞ͕DĂƌƚŝŶĞǌ-'ŽŶǌĂůĞǌDĞƚĂů͕EƵƚƌEĞƵƌŽƐĐŝ =30 ϮϬϭϭ͖ϭϰ͗ϭϵϱ-201. Added from hand interactions and cognition at 6.5 years. searching 4. DĂƌƚŝŶĞǌ->ĂƉŝƐĐŝŶĂ,͕dŽůĞĚŽ͕ůĂǀĞƌŽWĞƚĂů͕ŶŶEƵƚƌDĞƚĂď n=1 ϮϬϭϯ͖ϲϮ͗ϯϲ͘ References fully assessed n =31 5. DĂƌƚŝŶĞǌ-Lapiscina EH, Clavero P, Toledo E et al, J Neurol Neurosurg Excluded on basis of WƐǇĐŚŝĂƚƌǇϮϬϭϯ͖ϴϰ͗ϭϯϭϴ-25. eligibility criteria BDNF=Brain-derived Neurotrophic Factor; n=22 6. DĂƌƚŝŶĞǌ-Lapiscina EH, Galbete C, Corella D et al, Genes EƵƚƌϮϬϭϰ͖ϵ͗ϯϵϯ͘ Included articles n =9 CVD=cardiovascular disease; EVOO=extra 7. Valls-Pedret C, Sala-Vila A, Serra-DŝƌDĞƚĂů, :D/ŶƚĞƌŶDĞĚ (representing 5 trials) Total n included=1888 ϮϬϭϱ͖ϭϳϱ͗ϭϬϵϰ-103. 411 virgin olive oil; PREDIMED=Prevencion con 8. >ĞĞ:͕WĂƐĞD͕WŝƉŝŶŐĂƐĞƚĂů͕EƵƚƌŝƚŝŽŶϮϬϭϱ͖ϯϭ͗ϲϰϳ-52. Dieta Mediterranea. ϵ͘ Knight A, ƌǇĂŶ:͕tŝůƐŽŶĞƚĂů͕EƵƚƌŝĞŶƚƐϮϬϭϲ͖ϴ͗ϱϳϵ͘ Proceedings from the Dietitians Association of Australia 35th National Conference, Sydney, 17-19th May, 2018.

Citation:

Radd-Vagenas, S., Daniel, K., Noble, Y., Fiatarone Singh, M., & Flood, V. (2018). Reliability of new index tool to assess adherence to Mediterranean diet among people with mild cognitive impairment. Nutrition & Dietetics, 75 (Suppl 1), 11. doi:10.1111/1747-0080.12426

412 Proceedings from the Nutrition Society of Australia 42nd Annual Scientific Meeting, Canberra, 27-30th November, 2018.

Citation:

Radd-Vagenas, S., Fiatarone Singh, M. A., Daniel, K., Noble, Y., O'Leary, F., Mavros, Y., . . . Flood, V. M. (in press). Reliability and validity of MediCul (Mediterranean Diet & Culinary Index) in older Australian adults. Proceedings.

Reliability and validity of MediCul (Mediterranean Diet & Culinary Index) in older Australian adults Radd-Vagenas, S1, Fiatarone Singh, MA1,2, Daniel, K1, Noble, Y1, O’Leary, F3, Mavros, Y1, Brodaty, H4 and Flood, VM1,5 1 The University of Sydney, Physical Activity, Lifestyle, Ageing and Wellbeing Research Group, Faculty of Health Sciences, Lidcombe, NSW, Australia 2 Hebrew SeniorLife and Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, MA, USA 3 The University of Sydney, Nutrition and Dietetics Group, School of Life and Environmental Science, Faculty of Science & The Charles Perkins Centre, Camperdown, NSW, Australia 4 Centre for Healthy Brain Ageing, University of New South Wales, Sydney, NSW, Australia 5 Western Sydney Local Health District, Westmead Hospital, Westmead, NSW, Australia

A Mediterranean diet has been associated with multiple health benefits for chronic disease. Yet no index tool has been developed for, and/or validated in, a Western population to assess adherence to the ‘traditional’ dietary pattern and aspects of cuisine. We aimed to test the reliability and validity of the 50-item MediCul among older Australian adults. Participants were recruited January to March 2017 from the 45 and Up Study, as part of the Maintain Your Brain validation study, completing MediCul online on two occasions. Reliability was assessed using the intraclass coefficient (ICC) and Bland-Altman plots. Validity was tested against a 3-day food record (FR) using the Research Food Diary app. Participants (n=84; 59.5% female) were aged 65.4 years (SD=5.9) and overweight (BMI 26.1; SD=4.0). Mean MediCul scores at two time points (n=74) were 55.9/100.0 and 56.5/100.0, respectively. MediCul had very good reliability according to the ICC (ICC=0.87, 95% CI: 0.796, 0.914, p<0.0001) and Bland-Altman plots. For validity, the Bland-Altman indicated MediCul over-reported the score by 5.4 points versus FR, but with no indication of systematic bias (y=8.46-0.06*x) (95% CI: - 0.275, 0.155, p=0.582). MediCul has very good reliability and moderate validity for assessing adherence to a ‘traditional’ Mediterranean pattern among older adults.

413 APPENDIX TWO

Additional works not forming part of thesis

414 Journal of Alzheimer’s Disease xx (20xx) x–xx 1 DOI 10.3233/JAD-180572 IOS Press

1 Maintain Your Brain: Protocol

2 of a 3-Year Randomized Controlled

3 Trial of a Personalized Multi-Modal

4 Digital Health Intervention to Prevent

5 Cognitive Decline Among Community

6 Dwelling 55 to 77 Year Olds

a,∗ b c,d e 7 Megan Heffernan , Gavin Andrews , Maria A. Fiatarone Singh , Michael Valenzuela , f g h i j 8 Kaarin J. Anstey , Anthony Maeder , John McNeil , Louisa Jorm , Nicola Lautenschlager , a k b k a 9 Perminder Sachdev , Anupama Ginige , Megan Hobbs , Christos Boulamatsis , Tiffany Chau , l m n n,o d j 10 Lynne Cobiac , Kay Cox , Kenneth Daniel , Victoria M. Flood , Yareni Guerrero , Jane Gunn , d a e,j d d d 11 Nidhi Jain , Nicole A. Kochan , Amit Lampit , Yorgi Mavros , Jacinda Meiklejohn , Yian Noble , p n e 12 Fiona O’Leary , Sue Radd-Vagenas , Courtney Walton , Maintain Your Brain Collaborative a 13 Team and Henry Brodaty a 14 Centre for Healthy Brain Ageing, University of New South Wales, Australia b 15 Clinical Research Unit for Anxiety and Depression, School of Psychiatry, University of New South Wales, 16 Australia c 17 Sydney Medical School, University of Sydney, Australia d 18 Physical Activity, Lifestyle, Ageing & Wellness research group, University of Sydney, Australia e 19 Brain and Mind Centre, University of Sydney, Australia f 20 School of Psychology, University of New South Wales, Australia g 21 College of Nursing & Health Sciences, Flinders University, Australia h 22 Monash University, Australia i 23 Centre for Big Data Research in Health, University of New South Wales, Australia j 24 University of Melbourne, Australia k 25 Western Sydney University, Australia l 26 CSIRO Health and Biosecurity m 27 Medical School, University of Western Australia, Perth, Australia n 28 Faculty of Health Sciences, University of Sydney, Australia o 2930 Western Sydney Local Health District, Westmead Hospital, Australia p 31 Nutrition and Dietetics Group, School of Life and Environmental Science, Faculty of Science & The Charles Perkins Centre, University of Sydney, Australia

Accepted 10 September 2018 Uncorrected Author Proof

32 . ∗Correspondence to: Megan Heffernan, Dementia Centre for Research Collaboration, AGSM Building, The University of New South Wales,Sydney, NSW,2052, Australia. Tel.: +612 9385 3098; Fax: +612 9385 2200; E-mail: [email protected].

ISSN 1387-2877/18/$35.00 © 2018 – IOS Press and the authors. All rights reserved This article is published online with Open Access and distributed under the terms of the Creative Commons Attribution Non-Commercial License (CC BY-NC 4.0). 415 2 M. Heffernan et al. / Maintain Your Brain protocol

33 Abstract: 34 Background: Maintain Your Brain (MYB) is a randomized controlled trial of an online multi-modal lifestyle intervention 35 targeting modiﬁable dementia risk factors with its primary aim being to reduce cognitive decline in an older age cohort. 36 Methods: MYB aims to recruit 8,500 non-demented community dwelling 55 to 77 year olds from the Sax Institute’s 45 and 37 Up Study in New South Wales, Australia. Participants will be screened for risk factors related to four modules that comprise 38 the MYB intervention: physical activity, nutrition, mental health, and cognitive training. Targeting risk factors will enable 39 interventions to be personalized so that participants receive the most appropriate modules. MYB will run for three years and 40 up to four modules will be delivered sequentially each quarter during year one. Upon completing a module, participants will 41 continue to receive less frequent booster activities for their eligible modules (except for the mental health module) until the 42 end of the trial. 43 Discussion: MYB will be the largest trial to attempt to prevent cognitive decline and potentially dementia. If successful, 44 MYB will provide a model for not just effective intervention among older adults, but an intervention that is scalable for broad 45 use.

Keywords: Alzheimer’s disease, clinical trial, cognitive decline, cognitive training, dementia, depression, non- 46 pharmacological, physical activity, nutrition, randomized controlled trial 47

48 Trial registration: ACTRN 12618000851268. Registered on the Australian New Zealand Clinical Trials Registry (http://

49 www.anzctr.org.au).

33 INTRODUCTION results on clinically meaningful outcomes be found. 65 Internet delivered interventions have the potential 66 34 Dementia affects approximately 44 million peo- to overcome some of the limitations of traditional 67 35 ple worldwide [1] and assuming the status quo is face-to-face interventions, in that the former are scal- 68 36 projected to expand to over 136 million by 2050 able, can reach geographically isolated individuals, 69 37 [1, 2]. Globally, it is a leading cause of disability are convenient (e.g., can be undertaken at home), and 70 38 with costs estimated to exceed 800 billion dollars are cost effective. 71 39 [1]. In the next 40 years, these costs are projected We therefore designed an internet-delivered multi- 72 40 to exceed those of all other chronic diseases [3]. In domain preventative intervention—Maintain Your 73 41 light of the modest efficacy of symptomatic drugs for Brain (MYB)—to target multiple risk factors for 74 42 Alzheimer’s disease and the lack of success in tri- dementia in general and AD in particular. Tar- 75 43 als of disease-modifying medications, interventions geted risk factors are physical inactivity, sub-optimal 76 44 to delay onset and possibly prevent dementia have dietary patterns, cognitive inactivity, depression, anx- 77 45 gained momentum [4]. Given that dementia has a iety, and health-related issues implicated in dementia 78 46 multifactorial etiology, multi-domain interventions risk such as overweight, obesity, excess alcohol con- 79 47 are arguably more likely to be effective in delay- sumption, smoking, and chronic health conditions. 80 48 ing onset [5, 6]. However, evidence for the benefits The MYB program is organized into four differ- 81 49 of multiple combined interventions such as physical ent modules—Physical Activity, Nutrition, Peace of 82 50 activity and cognitive training is limited [7, 8]. Mind. and Brain Training—each of which is cus- 83 51 In the FINGER (Finnish Geriatric Intervention tomized to a person’s individual risks. MYB builds 84 52 Study to Prevent Cognitive Impairment and Dis- upon previous trials but introduces novel elements 85 53 ability) trial a multi-modal intervention (physical and state-of-the-art concepts in mode of delivery and 86 54 activity, diet, and cognitive training) was effective behavioral change theory to make it unique. Other 87 55 at reducing cognitive decline in a group of 60 to advantages of MYB are that it is comprehensive 88 56 77 year olds at risk of dementia and with average in addressing risk factors and it will be one of the 89 57 or below cognitive scores [9]. Two other trials, the first prevention trials to address depression. Modules 90 58 Multidomain Alzheimer Preventive Trial (MAPT) are also responsive to individual participant progress 91 59 and the Prevention of Dementia byUncorrected Intensive Vas- rather than aAuthor one-size-fits-all approach. Proof The program 92 60 cular care (PreDIVA), reported encouraging results will initially deploy interventions sequentially and 93 61 in post-hoc or subgroup analyses [10, 11]. Impor- then build up booster sessions over time, minimiz- 94 62 tantly, a limitation of both FINGER and PreDIVA ing the possibility of interference or over-dose effects 95 63 was their requirement for clinic attendance that [12, 13]; it introduces innovations in computer-based 96 64 would restrict large scale deployment should positive cognitive training including insights from gaming, 97

416 M. Heffernan et al. / Maintain Your Brain protocol 3

98 remote supervision, social support, and coaching vidual’s risk factors. In practice, this will translate 147 99 when required. Finally, the randomized controlled to a minimum of two modules and a maximum of 148 100 trial (RCT) will be larger than previous studies allow- four modules. Upon completing a module, booster 149 101 ing for sub-analyses of subsidiary questions sessions (specific to each module) and ongoing mon- 150 102 MYB aims to: 1) in the short term, develop itoring will then continue for up to three years. 151 103 a new MYB digital platform that delivers multi- Follow-up assessments measuring these risk factors 152 104 modal, personalized, and sequential modules that and cognition will be completed annually for three 153 105 target dementia risk factors via the Internet; 2) in the years. The control group (“information” group) will 154 106 medium term, evaluate the efficacy of the MYB dig- receive basic psychoeducation about dementia risk 155 107 ital platform to reduce the rate of cognitive decline factors accessed through the MYB digital platform. 156 108 in non-demented community dwelling persons aged Set activities will be organized by corresponding 157 109 55 to 77 years; 3) to examine the impact on reduc- module eligibility and also themed into the same 158 110 tion in risk factors, improvement in specific module modules of physical activity, nutrition, peace of mind, 159 111 targets (physical activity, nutrition, depression, indi- and brain training. Control participants will not be 160 112 vidual cognitive domains); and 4) evaluate the relative receiving any module-related tailored advice (i.e., the 161 113 cost effectiveness of the platform. “coaching” component in intervention). They will 162 114 The primary hypothesis of MYB is that over otherwise undertake the same enrolment, baseline 163 115 three years of intervention participants engaging with and annual follow-up assessments. An overview of 164 116 coaching (intervention) modules on the MYB digi- participant flow through the trial is provided in Fig. 1. 165 117 tal health platform will have less cognitive decline Owing to the nature of the intervention, par- 166 118 than the information only control group. Secondary ticipants and research staff involved in delivering 167 119 hypotheses are that there will be less incident all- modules will not be blind to allocation. However, 168 120 cause dementia than the control group; a relationship the MYB digital platform is an automated system 169 121 between intervention fidelity (compliance and adher- with all activities and assessments delivered elec- 170 122 ence) and cognitive outcomes; there will be less tronically without researcher intervention (unless a 171 123 categorical cognitive impairment in the intervention participant has requested to be withdrawn or for 172 124 group compared to the control; dementia risk profile safety). Researchers responsible for analyses will be 173 125 will improve significantly more in the intervention blinded to group allocation and analyses will be com- 174 126 group; and such a risk profile improvement will cor- pleted (as much as possible) without the use of data 175 127 relate with better cognitive outcomes. that may incidentally unblind the analysis (e.g., mod- 176 ule allocation, types of completed activities). 177

128 METHODS Setting 178 129 Design MYB is online and all study procedures will be 179 130 MYB is an online single-blind randomized con- delivered and managed via the MYB digital platform. 180 131 trolled trial (Australian New Zealand Clinical Trials Participants will access all trial materials and associ- 181 132 Registry ACTRN 12618000851268) of a person- ated help by logging into the MYB digital platform. 182 133 alized, multi-modal intervention designed to target Personalized activities, with due dates, reminders, 183 134 modifiable risk factors for Alzheimer’s disease and and progress will be available through this online plat- 184 135 dementia. Risk factors to be addressed include form. MYB will be supported by a team of researchers 185 136 physical inactivity, poor diet, cognitive inactivity, who will provide technical assistance to participants. 186 137 depression/anxiety, and chronic diseases and habits Each module is led and supported by experienced 187 138 (alcohol, smoking) linked to dementia risk. Up to researchers with relevant expertise in delivering their 188 139 four intervention modules (physical activity, nutri- respective interventions. 189 140 tion, peace of mind, and brain training) will be 141 administered based on individual riskUncorrected profiles. Each Recruitment Author and selection of participantsProof 190 142 module will be initially delivered using the MYB 143 digital platform over 10 weeks. In the first year of Participants will be recruited from the Sax Insti- 191 144 the RCT, participants will complete their assigned tute’s 45 and Up Study [14]. From January 2006 to 192 145 modules sequentially, noting that the total number December 2009, men and women from the general 193 146 of modules varies depending on the respective indi- population of New South Wales (NSW), Australia 194

417 4 M. Heffernan et al. / Maintain Your Brain protocol

Fig. 1. MYB trial ﬂowchart. Uncorrected Author Proof 195 were sampled through the Department of Human linkage to population health databases. A total of 200 196 Services (formerly Medicare Australia) database 267,153 people, about 1 in every 10 NSW men and 201 197 and joined the study by completing a postal ques- women aged over 45, are participating in the 45 202 198 tionnaire. Participants provided written informed and Up Study, making it the largest ongoing study 203 199 consent for follow-up through repeated contact and of healthy aging in the Southern Hemisphere. Now 204

418 M. Heffernan et al. / Maintain Your Brain protocol 5

205 8–11 years after recruitment, participants are aged require continuing chemotherapy), or non-metastatic 257 206 53 years and up. The 45 and Up Study is among prostate or skin cancer; have ever been diagnosed 258 207 the most heterogeneous large-scale cohort studies with bipolar disorder or schizophrenia; have severe 259 208 internationally, with participants from diverse socio- depression (Patient Health Questionnaire [15], PHQ9 260 209 demographic, geographic, and cultural backgrounds, total score >19); report having suicidal intent (PHQ-9 261 210 which will increase the likelihood that any cogni- item 9 >1 and Beck Depression Inventory [16] item 262 211 tive effects or risk reduction observed in the trial are 9 >1); are unable to write/converse in English; do not 263 212 generalizable to other populations. have at least two unique risk factors and are not eligi- 264 213 Invitations will be sent by the 45 and Up Study ble for at least two modules (risk factors determined 265 214 electronically and by mail to potential MYB par- after baseline assessment). 266 215 ticipants. Participants will not receive both types of Participants who are eligible for the trial will then 267 216 invitations. As MYB, being fully online, requires par- complete baseline assessments used to assess their 268 217 ticipants to have an email address to be able to login risk factors and module eligibility. Participants will 269 218 and receive notiﬁcations; electronic invitations will be enrolled if they complete all baseline assessments 270 219 be prioritized over physical mail. Invitations sent by (by the MYB start date) and have at least two unique 271 220 email or mail will be identical, except that a unique risk factors and are eligible for at least two modules. 272 221 registration link cannot be sent with the physical mail.

222 Participants receiving mailed invitations will need Randomization 273 223 to enter an email address to be able to receive the 224 unique registration link. All subsequent registration Participants who provide informed consent, meet 274 225 procedures will be identical. trial inclusion criteria and none of the exclusion cri- 275 226 Participants will be eligible to receive invitations if teria, and complete all baseline assessments will be 276 227 they are enrolled in the 45 and Up Study, have agreed enrolled in the trial. Participants agreeing to be in 277 228 to be contacted about further studies, are aged 55 to MYB will be randomized (1:1 allocation ratio) using 278 229 77 years on the ﬁrst day of the year recruitment com- the minimization technique [17] based on the follow- 279 230 mences, do not have dementia, Parkinson’s disease, ing stratiﬁcation variables: 280 231 or multiple sclerosis (based on available 45 and Up • Age (55–65 years; 66–77 years) as of 281 232 records) and did not participate in any previous MYB 01/01/2018 282 233 validation or pilot studies. • Gender (male, female) 283 234 Participants who provide consent and register on • Dementia risk as measured by Australian 284 235 the MYB digital platform will complete a brief National University Alzheimer’s Disease Risk 285 236 screening tool developed for the MYB trial. Par- Index Short Form [18] (tertiles total risk score) 286 237 ticipants will be excluded if they: do not respond • Module eligibility (1 module; 2 modules; 3 mod- 287 238 to the invitation (one reminder may be sent); are ules; 4 modules) 288 239 unwilling or unable to access trial tasks on the

240 MYB digital platform; are involved in another inter- Per the minimization technique, a weighted- 289 241 vention trial involving exercise, brain training, diet, coefﬁcient variable is used to assign the participants 290 242 depression/anxiety treatment, or any drug/nutritional within each proﬁle into intervention and control 291 243 supplement trial; do not have home internet or groups (1:1 allocation ratio). Randomization will 292 244 computer access (desktop or laptop); have vision occur after baseline assessments have been com- 293 245 problems which prevent them from reading a completed and therefore after module eligibility has 294 246 puter screen or are otherwise unable to use a been determined. We will conceal allocation with 295 247 mouse/keyboard without assistance; have a life randomization performed centrally by a computer- 296 248 threatening condition; have any dementia, Parkin- ized process and activities are then automatically 297 249 son’s disease, or multiple sclerosis; have ever been delivered to participants through the MYB digital 298 250 prescribed a medication used to treat Alzheimer’s dis- platform. 299 ≥ 251 ease; have been hospitalized overnightUncorrected ( 24 h) more Author Proof 252 than 6 times in the past 12 months; report having had Module eligibility 300 253 a stroke or heart attack in the previous 3 months; have 254 been diagnosed with brain cancer; have any form of Participants who are otherwise eligible for the trial 301 255 current malignancy other than cancer that has been in will be assessed for module eligibility using baseline 302 256 remission for at least 5 years (and does not currently assessments (i.e., prior to randomization). Each mod- 303

419 6 M. Heffernan et al. / Maintain Your Brain protocol

304 ule is designed to address a known dementia risk and Mediterranean Diet and Culinary Index (MediCul). 353 305 therefore criteria are unique to each of the modules Medicul is a newly created diet assessment for the 354 306 (see below). Module rules are built into the MYB Mediterranean diet [23] from which a score for the 355 307 platform and applied automatically. Researchers can- Mediterranean Diet Adherence Screener (MEDAS) 356 308 not override this pre-set decision algorithm except in [24] is also derived. MEDAS was selected as it is used 357 309 cases of safety where participants may be removed in PREDIMED (Prevencion´ con Dieta Mediterranea),´ 358 310 from a module should they reveal, as the trial pro- the largest clinical trial on the Mediterranean diet, 359 311 gresses, a new health issue not disclosed or present and has been associated with cognitive outcomes 360 312 at baseline, or they withdraw consent. [25, 26]. A MEDAS score of nine or less indi- 361 313 Participants with only one unique risk factor and cates low adherence with a Mediterranean dietary 362 314 hence eligible for less than two modules will be pattern. Other risks identiﬁed include high blood 363 315 allowed to enter in MYB, however, only participants pressure, high cholesterol, diabetes, heart related 364 316 with at least two unique risk factors and eligible for issues, stroke, TIAs, peripheral vascular disease, 365 317 at least two modules will be included in the planned BMI, waist circumference, and excessive alcohol 366 318 primary MYB analyses. This means if a participant consumption. 367 319 has only one risk factor, such as high blood pressure,

320 which makes them eligible for two modules (such Peace of mind 368 321 as physical activity and nutrition) they will not be Participants who currently meet probable criteria 369 322 included in the primary analysis. for unipolar depression and/or an anxiety disorder as 370 assessed by their total Patient Health Questionnaire 371 323 Individual module eligibility (PHQ-9) [15] and/or Generalized Anxiety Disorder- 372 7 (GAD-7) [27] scores; and/or report having been 373 324 Physical activity (PA) diagnosed and/or treated for unipolar depression 374 325 Participants will be eligible for the PA module or their anxiety during their lives (MYB Medi- 375 326 based on health conditions identiﬁed in the Aus- cal History) will be eligible for the “ThisWayUp 376 327 tralian National University Alzheimer’s Disease Risk Mixed Depression and Anxiety” course (e.g., see 377 328 Index Short Form (ANU-ADRI-SF [18, 19], medical https://thiswayup.org.au). 378 329 history (MYB Medical History), and MYB Physi-

330 cal Assessment battery (includes measures of body Brain training 379 331 composition, Harvard Alumni Questionnaire [20], Participants will be eligible for Brain Training 380 332 International Physical Activity Questionnaire [21] based on years of education (ANU-ADRI-SF), or 381 333 sitting, and Virtual Short Physical Performance Bat- lower than average late life mental activity (Life 382 334 tery Protocol [22]). Risks identiﬁed include low levels Experiences Questionnaire, LEQ late life subscale 383 335 of physical activity (aerobic and/or resistance train- subthreshold [28]) or mental activity during mid-life 384 336 ing), high levels of sitting, high blood pressure, (LEQ midlife sub scores subthreshold). 385 337 dyslipidemia, diabetes, heart related issues, stroke,

338 transient ischemic attack (TIA), high body mass Sequencing of modules 386 339 index (BMI), high waist circumference, and active

340 smoking status. Participants will be excluded from The order that participants will receive modules 387 341 the PA module (but not the trial) if they report is determined by a pseudo-random process (Sup- 388 342 any health concerns that would make continuing plementary Material 1). This process is carried out 389 343 in this module unsafe. This includes being advised automatically by the MYB digital platform once a 390 344 not to exercise by a health professional, worsening participant has completed all baseline assessments, 391 345 angina in the previous three months, angina at times module eligibility has been determined and prior to 392 346 other than moderate to vigorous activity, unrepaired randomization. 393 347 aneurysms and 12 or more seizures in the previous 348 12 months. UncorrectedInterventions Author Proof 394

349 Nutrition Coaching (intervention) 395 350 Participants will be eligible for the Nutrition mod- The MYB intervention comprises up to four 396 351 ule based on health conditions identiﬁed in the MYB 10-week modules delivered quarterly in the ﬁrst 397 352 Medical History or a dietary assessment using the 12 months of the trial, with monthly boosters 398

420 M. Heffernan et al. / Maintain Your Brain protocol 7

399 for modules where appropriate thereafter. These the target volume within the first month. Once the 450 400 modules are Physical Activity, Nutrition, Peace of desired volume is reached, intensity goals will be 451 401 Mind, and Brain Training. All procedures described addressed aiming to reach these by the end of the 452 402 will be accessed through the MYB digital plat- 10-week module. Progression from moderate to vig- 453 403 form. All modules will begin with a brief module orous aerobic and resistance training activities will be 454 404 expectations questionnaire and end with a brief ques- encouraged as feasible based on their logged behav- 455 405 tionnaire measuring participants’ appraisal of the ior which includes modality, volume, frequency, and 456 406 module. perceived intensity (Borg Scale of Perceived Exer- 457 tion). Participants will be encouraged to add dual 458 407 Physical activity (PA) tasking to their balance exercises once they have 459 408 The basic PA prescription is designed to increase achieved daily balance practice and reached the most 460 409 aerobic and resistance training to levels with a demon- challenging level of each exercise. Balance intensity 461 410 strated positive impact on cognition and other health is advanced by gradual reduction of hand support 462 411 outcomes [29, 30]. Where feasible, the module will and visual input during static and dynamic exercises. 463 412 support progression to higher exercise intensities After the initial 10 weeks of the intervention, booster 464 413 as this is linked to improved risk reduction, fitness or “refresher” sessions will be added. Booster ses- 465 414 outcomes and cognition in observational and exper- sions are identical to module activities but occur less 466 415 imental studies [31, 32]. In addition, daily balance frequently (every four weeks for up to three years 467 416 training exercises and information on smoking ces- follow-up) to allow internalization of monitoring and 468 417 sation will be included. A behavioral change program progression. 469 418 [33] is integrated within the PA module by providing All materials are developed by qualified exercise 470 419 personalized goals and feedback, and by encouraging physiologists and/or physiotherapists and provided 471 420 activity logging. on the MYB digital platform. Content can be accessed 472 421 The PA module is fully online and provided as many times as the participant wishes and will also 473 422 through the PA portal on the MYB digital platform. cover topics related to behavioral change, chronic dis- 474 423 All goals, feedback, and materials (e.g., videos, fact- eases, relapse prevention, progression, variety, and 475 424 sheets) are arranged around three exercise modalities: long-term adherence. All materials can be updated, 476 425 aerobic exercise, resistance training, and balance removed, and new materials added. The PA module 477 426 training. Information is provided to support achieve- is designed to be self-administered, with participants 478 427 ment of the long-term goal of 300 min/week of able to engage with materials in their own time. 479 428 moderate to vigorous intensity aerobic activity, three During the module and follow-up, participants will 480 429 days/week of vigorous intensity progressive resis- also be able to contact “module trainers” (qualified 481 430 tance training, and challenging balance training every exercise physiologists or physiotherapist) with ques- 482 431 day. Behavioral change is integrated within the mod- tions or issues related to the module. This contact 483 432 ule through updating feedback and goals. These are will initially be via the online interface, with follow- 484 433 personalized based on activity logged by the partici- up contact as appropriate (e.g., email and phone). 485 434 pant (weekly self-report via Short Physical Activity Module trainers will also update “Frequently asked 486 435 Assessment). Feedback is responsive to both partic- questions” factsheets for triaging common problems 487 436 ipants who are improving (for example, providing and questions arising during exercise adoption, refer 488 437 updated goals) and participants who are not pro- participants to their GP for suspected exercise-related 489 438 gressing or are regressing. If a lack of progression is or other injury, recurrent falls, potential medication- 490 439 detected, the module will direct participants to infor- related side effects, or instability /progression of 491 440 mation on barriers and how to overcome them, as underlying chronic diseases such as osteoarthritis, 492 441 well as how exercise can be modified for pain or angina, etc. 493 442 injury. Participants are also encouraged to use the The PA module will be adjusted if a participant 494 443 online Action Plan to write how their goals will be reports being diagnosed with osteoporosis, or a high 495 444 achieved. Feedback on progress is displayedUncorrected to par- number of fallsAuthor in the past 12 months, Proof or scores poorly 496 445 ticipants visually (via a virtual “staircase”) for each in the virtual Short Physical Performance Battery 497 446 of the three exercise modalities. Short term goals (vSPPB) [22] by removing any goals and feedback 498 447 and feedback will follow the general approach of related to aerobic exercise as walking may increase 499 448 starting exercise at low frequencies and intensities risk of falls until strength and balance are improved 500 449 for those not currently exercising and then reaching [34, 35]. During annual assessments, participants 501

421 8 M. Heffernan et al. / Maintain Your Brain protocol

502 could be reallocated to the main PA module if appro- that participants are aiming to achieve (via an ideal 553 503 priate. Initially, baseline assessments will be used “Mediterranean diet food pyramid”) based on their 554 504 to exclude any participant from PA owing to safety self-reported food intake. Food intake will be via the 555 505 concerns (for example, being advised by a medical MiniCul, a screener form of the MediCul [23] that is 556 506 professional not to exercise). Participants who are used to prompt behavior change within the module. 557 507 eligible for PA at baseline may subsequently report The components of the Mediterranean diet food pyra- 558 508 a health concern that requires adjustment to their PA mid will be completed in sections, with personalized 559 509 module. The aerobic component of the module can messages and feedback provided for food groups 560 510 be removed based on pre-set rules applied automati- that require improvement. Although personalized, the 561 511 cally by the MYB digital platform (for example too overall approach of dietary messages and feedback is 562 512 many reported falls). Their PA module can also be to target the most poorly scored and crucial changes 563 513 temporarily blocked (or ‘unblocked’ if recovered) on first. Additionally, participants will only receive a 564 514 a case-by-case basis by qualified PA module staff manageable number of changes each week (maxi- 565 515 (e.g., clinicians and exercise physiologists). For any mum four messages). If a participant does not make 566 516 health-related concern that is not planned or reported any changes for two weeks (as measured by weekly 567 517 via the MYB platform, researchers may inform the online food intake logging), messages will move onto 568 518 MYB administrator to remove participants from the other poorly scored food groups and cycle back to 569 519 module entirely. the unchanged groups if time is available. There- 570 520 Physical activity logged via the Short Physical fore, even a participant who needs to make changes 571 521 Activity Assessment (weekly in the first 10 weeks, across all 10 food groups but shows no change in 572 522 then monthly) will be used to measure PA/sedentary their weekly log, will receive advice and recommen- 573 523 behavior levels during the module. These weekly log- dations for all food groups. If a participant reaches 574 524 ging tasks, email notifications about new tasks as well the ‘ideal’ Mediterranean diet food pyramid they 575 525 as intermittent newsletters will be used to support will be encouraged to maintain their Mediterranean 576 526 adherence to the module. Whether a participant com- diet. 577 527 pletes activity logs and action plans, and how many The initial Nutrition module is delivered over 10 578 528 times they visit the PA module can be used to assess weeks during the first 12 months of MYB. During the 579 529 engagement with the module. first 10 weeks participants are encouraged to log their 580 adherence to the Mediterranean diet weekly (self- 581 530 Nutrition report via MiniCul). Participants are also encouraged 582 531 The Nutrition module is designed to encourage to fill out a weekly action plan for how they will 583 532 adoption of a Mediterranean diet acknowledged in improve their Mediterranean diet score for each of the 584 533 dietary guidelines as a healthy eating pattern for suggested food changes. After the initial 10 weeks of 585 534 chronic disease [36, 37]. The rationale for this diet intervention, booster sessions will be added. Proce- 586 535 is based on its relevance to risk factors for cognitive dures and materials for booster sessions are identical 587 536 decline and RCT and prospective cohort data for a to the initial Nutrition module. However, booster 588 537 Mediterranean diet [38–41]. At this time, the Mediter- sessions and therefore diet logging will occur less 589 538 ranean diet appears to be the most strongly supported frequently (every 4 weeks). 590 539 dietary pattern for cognitive benefits in clinical trials All materials are designed by qualified dietitians 591 540 [41]. and can be accessed as many times as the par- 592 541 The Nutrition module is fully online and pro- ticipant wishes. Materials include recipes, videos, 593 542 vided through the Nutrition portal on the MYB shopping lists, and factsheets. Topics covered by 594 543 digital platform. All dietary messages, feedback, these materials will be related to the recommended 595 544 and materials (such as cooking videos, meal plans, diet as well as areas such as behavioral change, sci- 596 545 recipes, shopping lists, health tips, factsheets, and entific literature, special dietary considerations and 597 546 behavioral change resources) are arranged around food allergies, intolerances and drug interactions. All 598 547 10 food groups (discretionary foods,Uncorrected animal pro- materials can Author be updated, removed,and Proof new materials 599 548 tein, extra virgin olive oil, nuts, vegetables, fruit, can be uploaded by the Nutrition module staff (i.e., 600 549 legumes, wholegrains, fish and seafood, and water) qualified dietitians). 601 550 within the Mediterranean diet to provide smaller The Nutrition module is designed to be self- 602 551 manageable targets for change. The Nutrition portal administered with participants able to engage with 603 552 provides a visual display of the Mediterranean diet the materials in their own time. During the module 604

422 M. Heffernan et al. / Maintain Your Brain protocol 9

605 and follow-up, participants will also be able to (CRUfAD) at UNSW/St Vincent’s Hospital and the 656 606 contact “module trainers” (qualified dietitians) with efficacy and effectiveness of the program reported 657 607 questions or issues related to the module. This previously [42–44]. During the module, participants 658 608 contact will initially be via the online interface, with will be able to contact CRUfAD staff with clinical 659 609 follow-up contact as appropriate (e.g., email and and technical questions. 660 610 phone). Module trainers will also provide updates to The Mixed Depression and Anxiety course 661 611 “Frequently asked questions” and refer participants includes six lessons delivered over 10 weeks. These 662 612 to a GP if necessary. lessons include: 1) Psychoeducation about anxiety 663 613 There are two planned modifications to the Nutri- and depression, identifying symptoms, the fight or 664 614 tion module based on alcohol consumption and being flight response, controlled breathing and physical 665 615 underweight. If participants report a high rate of activity/exercise; 2) Cognitive therapy components: 666 616 alcohol consumption, diet messages related to alco- education about the cognitive model, cognitive dis- 667 617 hol will be prioritized and remain until participants tortions, and introduction to thought monitoring; 668 618 report a reduced alcohol intake. Participants report- activity planning; 3) Thought challenging/cognitive 669 619 ing low body weight will receive modified advice restructuring; challenging positive & negative meta- 670 620 that encourages adapted serving sizes compared to cognitive beliefs about repetitive thinking; shifting 671 621 the main Nutrition module. After the 10-week inter- attention, hunt for positives; 4) Education about 672 622 vention, such participants could be reallocated to avoidance and safety behaviors; graded expo- 673 623 the main Nutrition module if goals are achieved. sure and structured problem solving; 5) Advanced 674 624 Likewise, participants receiving the main Nutrition graded exposure (imaginal exposure, interoceptive 675 625 module could be reallocated to the modified alcohol exposure); troubleshooting difficulties with graded 676 626 or underweight Nutrition module after the 10-week exposure; and 6) Relapse prevention. There are 677 627 intervention if appropriate. These modifications are no boosters associated with the Peace of Mind 678 628 pre-set and the rules are applied automatically by module. 679 629 the MYB digital platform. For any other diet-related Participants’ completion of each lesson will be 680 630 concerns, such as medical advice not to alter their used to assess adherence to the module. During 681 631 diet, researchers may remove participants from the the module, participants will complete measures of: 682 632 module entirely if it becomes unsafe for them to con- psychological distress at the start of each lesson 683 633 tinue. However, there is flexibility to allow for food (Kessler Psychological Distress Scale, K10) [45]; 684 634 preferences, financial circumstances, special dietary and depression (PHQ-9) [15] and anxiety symp- 685 635 requirements, food allergies, intolerances, and ethnic tom severity (GAD-7) [27] prior to lessons 1, 4, 686 636 backgrounds. and 6. 687 637 The MiniCul is used weekly in the first 10 638 weeks, then monthly, to track progress with adopt- Brain training 688 639 ing a Mediterranean diet during the intervention. Participants with an at-risk cognitive profile will 689 640 In addition, self-reported body measurements (waist be eligible for the Brain Training module. Partici- 690 641 circumference, weight) are also checked during pants will complete a personalized cognitive training 691 642 this module. Weekly email reminders to complete program that is vertically and horizontally adap- 692 643 MiniCul and intermittent newsletters to support tive called the “Brain Training System” (BTS). That 693 644 adherence to the module are also provided. Engage- is, the training regimen is defined on participant’s 694 645 ment with the Nutrition module is tracked by the baseline cognitive profile and continues to evolve in 695 646 number of completed MiniCuls, action plans to response to within-training task performance. Indi- 696 647 achieve Mediterranean goals and the number of times vidual cognitive exercises are run as Flash files as 697 648 participants login. provided by our commercial partner (Synaptikon, 698 trading as NeuroNation, Germany); more than 30 699 649 Peace of mind are available at the outset and will be supplemented 700 650 Participants who are currently depressedUncorrected and/or throughout Author the trial. BTS has Proof otherwise been orig- 701 651 anxious; and/or have ever been diagnosed/treated inally designed, including the approach to exercise 702 652 for depression and/or anxiety will have access to selection, streaming, feedback, coaching, and social 703 653 the “ThisWayUp Mixed Depression and Anxiety” support. 704 654 course. This course has been developed by the A participant’s cognitive profile is generated 705 655 Clinical Research Unit for Anxiety and Depression based on cognitive tests completed at baseline. 706

423 10 M. Heffernan et al. / Maintain Your Brain protocol

707 These proﬁles will rank performance across cogni- program adherence (completed sessions) and social 759 708 tive domains (i.e., verbal executive, speed, verbal platform engagement. 760 709 memory, visual executive, visual memory, visual

710 attention) in descending order. The initial exercise Information (Control) 761 711 selection and streaming order will reflect this profile, 712 preferentially targeting areas of weakness in the mid- Participants in the control group will receive basic 762 713 dle of the session. Exercise selection and streaming health information organized in the same module 763 714 will be recalibrated during training several times, in topics as the intervention (Physical Activity, Nutri- 764 715 this case based on degree of improvement on exer- tion, Peace of Mind, and Brain Training). Participants 765 716 cises at the domain-level. will have access to these information-based modules 766 717 Exercises will be presented as three sessions per based on their module eligibility and in a pseudo- 767 718 week, translating to 30 sessions over 10 weeks. random order as described in the previous section. 768 719 Each session will last 45 minutes and comprise 17 Although control participants receive modules based 769 720 exercises. If participants miss sessions, they will on their individual risk profile they will not be pro- 770 721 remain available until completed. During follow-up, vided with any tailored advice (i.e., the ‘coaching’ 771 722 booster exercises will be offered once a month (up intervention component) about how to improve their 772 723 to three years follow-up). Cognitive training mate- risk factors. The control group will receive infor- 773 724 rials (e.g., factsheets) will also be provided and mation monthly for up to four 10-week modules in 774 725 can be accessed as many times as the participant the first 12 months. All procedures described will be 775 726 wishes. delivered online and accessed through the MYB dig- 776 727 After completing the first five sessions, partic- ital platform. All modules will begin with a brief 777 728 ipants will receive a feedback performance graph module expectations questionnaire and end with a 778 729 that summarizes their cognitive performance on brief questionnaire measuring participants’ appraisal 779 730 each cognitive domain trained thus far. This perfor- of the module. 780 731 mance graph will also indicate the range of scores For the Physical Activity, Nutrition, and Peace of 781 732 for the top 25% performers in the module based Mind control modules, the information provided will 782 733 on previous sessions. Participants’ individual per- be focused on existing health guidelines and gen- 783 734 formance will be updated after completing each eral advice for achieving these targets. For example, 784 735 session and the top 25% range will be updated providing the Australian Dietary Guidelines from 785 736 weekly. the Australian Government Department of Health 786 737 In order to maximize adherence and motivation, and Ageing without specifying how participants can 787 738 BTS will deploy several socialization and gaming achieve these guidelines. The Brain Training control 788 739 strategies. This includes access to an online Trainer, module will involve a brief task (such as showing 789 740 prioritized for participants ‘red-flagged’ as strug- videos with a brief quiz). These tasks have been used 790 741 gling with engagement or performance. Video chat in previous trials and have been shown to be enjoyable 791 742 will be preferred for these interactions, but partici- with minimal impact on cognition [46]. Participants 792 743 pants will also be able to contact module trainers via will access the 10-week modules via the MYB dig- 793 744 MYB’s online interface. Participants can also nomi- ital platform with new module information or tasks 794 745 nate friends or family members who will be advised provided once a month. Upon completing the Physi- 795 746 via email when the participant reaches certain adher- cal Activity, Nutrition, and Brain Training modules, 796 747 ence milestones, with the aim of providing social these activities are then reduced to quarterly (i.e., 797 748 support and encouragement and establishing a com- boosters) for up to three years follow-up. As with 798 749 munity of practice. intervention, there will be no boosters for the Peace 799 750 The Brain Training module is therefore designed to of Mind module. 800 751 be automatically adaptive to a participant’s baseline The control group will receive periodic emails 801 752 score and progress during training. Safety issues are reminding them to login to the MYB digital 802 753 not foreseen for this module. Uncorrectedplatform to Author access module information. Proof All partic- 803 754 Participants will receive email notifications when ipants in a control module will receive the same 804 755 a new activity is available as well as when perfor- basic health information or task (with no tailored 805 756 mance problems are noted to support adherence. The advice). However, they may be removed from a 806 757 Brain Training module will measure within-exercise module for safety reasons (e.g., a previously undis- 807 758 improvement, domain-level performance changes, closed/undiagnosed health issue that might prevent 808

424 M. Heffernan et al. / Maintain Your Brain protocol 11

809 safe exercise). Engagement with the module can • Participants are screened for dementia at base- 856 810 be measured by the amount of times a partici- line via self-report questions. However, if 857 811 pant logs into the MYB eHealth platform and the baseline dementia is subsequently identified dur- 858 812 amount of times they visit the module information ing baseline assessment via any of the processes 859 813 webpage. below, participants will be retrospectively cen- 860 sored as they cannot meet the definition of 861 incident dementia. 862 814 Outcome measures • Dementia is assumed if there is any self-reported 863 dementia to MYB via the medical history 864 815 Primary outcome questionnaire (diagnosis or dementia medica- 865 816 The primary outcome is change in cognition from tions), stated as a reason to withdraw from 866 817 baseline to three years. We will test whether the the trial, or it is otherwise reported to MYB 867 818 groups show differential change over the trial as that dementia has been confirmed by a medical 868 819 indicated by the group by time interaction. In addi- professional. 869 820 tion, we will test the group difference at three years • If there is no self-report, dementia will be 870 821 as a planned comparison. The outcome variables assumed if linked data contain any of the fol- 871 822 are the global cognition composite domain scores, lowing: 872 823 at baseline and three years, as measured by the ◦ Pharmaceutical Benefits Scheme (PBS) 873 824 MYB online cognitive test battery (“MYB Battery”). claims for dispensing of Alzheimer’s dis- 874 825 Domains and tests in MYB are defined as: 1) Complex ease medications (donepezil, galantamine, 875 826 attention (Cogstate Detection and Cogstate Identifi- or rivastigmine or memantine). 876 827 cation); 2) Executive Function (Cogstate One Back, ◦ Dementia listed as a diagnosis in hospital 877 828 Cambridge Brain Sciences Spatial (Tokens) Search records. 878 829 and Cambridge Brain Sciences Grammatical Reason- ◦ Dementia listed as a cause of death in mor- 879 830 ing); and 3) Learning and Memory (Cogstate One tality records. 880 831 Card Learning and Cambridge Brain Sciences Paired • If none of the above indicates dementia, DSM-5 881 832 Associates). Two valid test scores per domain are criteria [47] requiring cognitive impairment and 882 833 required to generate a participant’s domain score at functional impairment will be applied. 883 834 each time point. Individual test scores will be con- ◦ Cognitive impairment is measured by the 884 835 verted to standard scores (z-scores) using the means MYB Battery and an additional auto- 885 836 and standard deviations (SD) of the MYB baseline mated telephone-based verbal memory test 886 837 sample. Domain scores will be calculated by first (LOGOS), developed and validated for 887 838 averaging the z-scores of the component tests, and MYB. 888 839 then transforming the composite scores using the ◦ Cognitive impairment is defined as ≥2SD 889 840 means and SDs of the MYB baseline sample. A below the mean of the baseline sample on 890 841 global cognition score will be calculated by aver- any individual test (stratified for age, sex, 891 842 aging across the three standardized domain scores, education) OR ≥1.5 SD below the mean of 892 843 and then the score will be similarly transformed to z- the baseline sample in ≥1 domain. (strati- 893 844 scores using the means and SDs of the MYB baseline fied for age, sex, education). 894 845 sample. ◦ Functional impairment is defined by the 895 Amsterdam-Instrumental Activity of Daily 896 Living-Short Form (A-IADL-SF; cut-off is 897 846 Secondary outcomes less than 51.4 [48]). 898 847 Secondary outcomes related to cognitive impair- • If there are no cognitive data, dementia will be 899 848 ment and dementia are: 1) incident dementia at three “undetermined”. 900 849 years; 2) change in dementia risk (ANU-ADRI-SF • If there is evidence for cognitive impairment 901 850 total score); and 3) using the method described for Uncorrected(defined Author above) but no data Proof on function, partic- 902 851 the primary outcome, change in cognitive domain ipants will be classified as “impaired cognition, 903 852 scores and individual cognitive tests that form the function undetermined”. 904 853 MYB Battery as well as LOGOS.

854 Dementia in MYB will be determined according Secondary outcomes related to the overall impact 905 855 to the following: of the intervention and participant experience are: 1) 906

425 12 M. Heffernan et al. / Maintain Your Brain protocol

907 service utilization identiﬁed through routinely col- for tasks such as responding to participant enquiries, 958 908 lected linked data: hospital admissions, emergency while ensuring trial-wide clinical data are not acces- 959 909 department presentations, medical and social care sible through the same pathway. 960 910 services, and prescribed medications; 2) costs of pro-

911 gram as determined by quality adjusted life years Assessments and linked data 961 912 (QALYs) based on MYB used resources (e.g., staff 913 time) and intervention effect; 3) change in module MYB assessments to be completed by all partici- 962 914 expectations as measured by the Module Expecta- pants (self-report), module specific assessments and 963 915 tions questionnaire (total; adapted from Rabipour & data sources for linked health data are summarized 964 916 Davidson [49]); 4) compliance with the intervention in Table 1. Medicare Benefits Schedule (MBS) and 965 917 and for each module undertaken (percentage activi- Pharmaceutical Benefits Scheme (PBS) data will be 966 918 ties complete and accessed per module; MYB digital accessed from the Department of Human Services 967 919 platform login data); and 5) number of adverse events (DHS) and linkage completed by the Sax Institute 968 920 reported to MYB via the medical history and medical using a unique identifier that was provided by DHS. 969 921 events forms. Events include hospitalizations, muscle Other linkage data (Table 1) will be performed by 970 922 soreness and food allergies. the Centre for Health Record Linkage (CHeReL, 971 923 Secondary outcomes related to risk factors targeted www.cherel.org.au) using probabilistic record link- 972 924 by the four module interventions are: 1) change in age techniques and ChoiceMaker software. 973 925 waist circumference and BMI if overweight/obese 926 (self-report in MYB Physical Activity questionnaire); Statistics 974 927 2) change in level of physical activity (HAQ energy 928 expenditure, stair climbing, walking. IPAQ sitting Sample size 975 929 time and volume and intensity of aerobic exercise, We will have 80% power to detect significant group 976 930 resistance training, and balance exercise/week); 3) by time interaction in the MYB Battery primary out- 977 931 new chronic conditions (count based on medical his- come, assuming linear change and a final difference 978 932 tory and medical events forms); 4) change in physical of 0.1 SD if we have 1,714 people in each group 979 933 functional performance during the PA module as (total N = 3,428) with type 1 error rate of 0.05. This 980 934 measured by the virtual Short Physical Performance assumes a within-subjects correlation of 0.6 and a 981 935 Battery Score; 5) change in adherence to the Mediter- Lear deterioration rate of 0.1. If we assume a 20% 982 936 ranean diet (MediCul score and derived MEDAS drop out rate over the trial, we will need to recruit 983 937 score); 6) change in exposure to foods/culinary prac- at least 2,143 people in each arm. With 2,143 par- 984 938 tices assessed via MediCul; 7) change in alcohol ticipants per arm (4,285 in total), we will also have 985 939 (self-report in ADRI-SF and MediCul) and smoking 80% power to detect differences of at least 0.1 SD for 986 940 status (self-report in ADRI-SF); 8) change in within continuous secondary outcomes. 987 941 task cognitive training performance in the Brain Sample size estimates were obtained using 988 942 Training module (percentage correct per session); GLIMMPSE 2.0.0 [50] and G*Power 3.1 [51]. MYB 989 943 9) change in mental activity levels (LEQ late-life has the capacity to enroll up to 16,000 persons if it is 990 944 subscale); 10) engagement during Brain Training oversubscribed. 991 945 module (count use of social platform); and 11) change 946 from baseline to three years in psychological distress Analysis 992 947 (K10). Participants with at least two unique risk fac- 993 tors and eligibility for two modules or more will 994 948 Data management be included in the MYB analysis. The trial will be 995 949 Electronic files collected by MYB that are not reported in accordance with the CONSORT guide- 996 950 linked data will be stored on a secure server. The des- lines. We will analyze all outcomes by the arm to 997 951 ignated MYB data custodian (MYB Data Manager) which participants were allocated (i.e., intention to 998 952 will be responsible for maintaining dataUncorrected security and treat analyses). Author We will analyze Proof the primary outcome, 999 953 assigning access controls. MYB has implemented change in cognition, using linear mixed models to 1000 954 a role-based access mechanism that controls the examine the treatment group by time interaction and 1001 955 access to research data. The MYB platform can allow will obtain a test of the difference between the groups 1002 956 authorized researchers, responsible for the day-to- at the end of the trial. We will also conduct anal- 1003 957 day running of the trial, to access information needed yses adjusting for baseline demographic variables; 1004

426 M. Heffernan et al. / Maintain Your Brain protocol 13

Table 1 MYB assessments and timing of administration Assessment/data source Measures Timing MYB Eligibility 1a Trial eligibility Screening only ANU-ADRI SF Dementia risk Baseline and annual follow-up MYB Batteryb Cognitive ability Baseline, every 3 months in the first 12 months and then annually LOGOSc Verbal memory Baseline, every 3 months in the first 12 months and then annually Medical History Baseline Medical history Baseline only Medical History Follow-up Medical history Annual follow-up Physical Assessment batteryd Body composition, physical Baseline and annual follow-up activity, sitting time and functional mobility performance MediCul and derived MEDASe Dietary pattern Baseline and annual follow-up Wellbeingf Psychological distress, Baseline and annual follow-up depression and anxiety LEQ Cognitive lifestyle Baseline only LEQ– Late life Late life subscale of LEQ Annual follow-up A-IADL SF g Daily function Baseline and annual follow-up Module Expectations Module expectations Pre/post modules only Medical Events form Adverse events Every 3 months Short Physical Activity measure (HAQ) Weekly physical activity Intervention Physical Activity module only - module baseline and weekly during the module; monthly during follow-up MiniCul, weight and waist circumference Weekly nutrition Intervention Nutrition module only - module baseline and weekly during the module; monthly during follow-up K10 Psychological distress Intervention Peace of Mind module only - start of each lesson (6 over 10 weeks) PHQ-9 Depression Intervention Peace of Mind module only - start of lessons 1, 4 and 6 GAD-7 Anxiety Intervention Peace of Mind module only - start of lessons 1, 4 and 6 Medicare Benefits Schedule (MBS)h Primary health care service use Available data at baseline to follow-up Pharmaceutical Benefits Scheme (PBS)h Prescribed medicines Available data at baseline to follow-up NSW Admitted Patient Data Collection (APDC)h Hospital use Available data at baseline to follow-up NSW Emergency Department Data Collection Emergency department Available data at baseline to follow-up (EDDC)h admissions NSW RBDM Death Registrationsh Deaths Available data at baseline to follow-up Cause of Death Unit Record File (COD URF)h Deaths Available data at baseline to follow-up NSW Mental Health Ambulatory Data Collectionh Service utilization Available data at baseline to follow-up ANU-ADRI SF, ANU Alzheimer’s Disease Risk Index Short Form [18]; A-IADl-SF, Amsterdam IADL Questionnaire Short Form [48]; LEQ, Lifetime of Experiences Questionnaire [28]; K10, Kessler Psychological Distress Scale [45]; GAD-7, Generalized Anxiety Disorder 7-item [27]; PHQ9, Patient Health Questionnaire [15]. aMYB Eligibility 1 includes baseline Patient Health Questionnaire (PHQ-9). bThe MYB Battery comprises tasks from the Cogstate Brief Battery (Detection, Identification, One Card and One Back) and Cambridge Brain Sciences (Spatial (Tokens) Search, Grammatical Reasoning and Paired Associates). cLOGOS: an automated telephone-based test of episodic verbal memory developed and validated for MYB. dThe Physical Assessment battery includes measures of body composition (self-reported standing height, body mass, waist circumference); physical activity (the Harvard Alumni Questionnaire, HAQ [20]); sitting time (item from the International Physical Activity Questionnaire, IPAQ [21]); and a measure of functional mobility and performance (the Virtual Short Physical Performance Battery Protocol, vSPPB [22]). eAdherence to the Mediterranean dietary patterns include MediCul (Mediterranean Diet & Culinary Index [23]) and MEDAS (Mediterranean Diet Adherence Screener [24]). f Wellbeing assessment includes the Generalized Anxiety Disorder 7-item (GAD-7), PHQ-9a, and Kessler Psychological Distress Scale (K10). gA-IADL-SF to be completed by an informant for all participants (non-compulsory assessment). hLinked data.

1005 additional adjustment for differentialUncorrected expectations tors). We expectAuthor change in cognitionProof to show a 1011 1006 between arms or modules will be applied as neces- normal distribution. However, if there are viola- 1012 1007 sary. In addition, we will conduct subgroup analyses tions of the assumptions of linear mixed models, 1013 1008 on the primary outcome to determine if there is appropriate transformations or a generalized linear 1014 1009 effect modiﬁcation by the number of risk factors mixed model with an appropriate distribution will be 1015 1010 at baseline (3 or 4 risk factors versus 2 risk fac- used. 1016

427 14 M. Heffernan et al. / Maintain Your Brain protocol

1017 Secondary outcomes will be analyzed using linear Health Services Ethics Committee (#2016/03/636). 1064 1018 mixed models for continuous normally distributed The conduct of the 45 and Up Study was approved by 1065 1019 outcomes or chi-squared tests for categorical out- the University of New South Wales Human Research 1066 1020 comes. We will use nonparametric tests for analyzing Ethics Committee (HREC). 1067 1021 continuous outcomes if the assumption of normality 1022 is violated. DISCUSSION 1068

1023 Safety Maintain Your Brain targets known modifiable 1069 cognitive decline and dementia risk factors through 1070 1024 Data safety and monitoring board (DSMB) a personalized, multimodal online intervention. The 1071 1025 A DSMB will be convened for this trial. It is expected benefits of this approach are that MYB 1072 1026 an independent group of experts that advises MYB has the capacity to target a number of risks in a 1073 1027 investigators. No member of the DSMB will have large number of participants. A range of modifi- 1074 1028 direct involvement in the conduct of the study; able risk factors for Alzheimer’s disease and vascular 1075 1029 financial, proprietary, professional, or other interests dementia including vascular, cognitive, lifestyle, and 1076 1030 that may affect independent decision-making by the psychosocial factors, account for up to one third of 1077 1031 DSMB. the population attributable risk for Alzheimer’s dis- 1078 ease [52] and are major risk factors for vascular 1079 1032 Adverse event and serious adverse events dementia. In Australia, there is an even higher esti- 1080 1033 All MYB participants will be routinely logging mated population attributable risk for dementia of 1081 1034 into the MYB platform (both control and intervention 48% owing to the distribution of these risk factors in 1082 1035 group) for trial procedures. During these, they will the Australian population [53]. Should Maintain Your 1083 1036 be prompted every 3 months by an on-screen ‘pop- Brain be successful in addressing cognitive decline 1084 1037 up’ to report any adverse event(s) in both a check by targeting multiple risk factors, it could provide a 1085 1038 box and text box format, which will be followed by model for interventions suitable for use in the broader 1086 1039 specific advice to continue/start exercises or dietary community. 1087 1040 changes or not, or to see their health care provider 1041 if indicated. All participants can also report adverse 1042 events at any time of the trial (ad hoc reporting). TRIAL STATUS 1088 1043 Adverse events during the Physical Activity and/or 1044 Nutrition modules that are potentially related to the The MYB trial (Protocol v1.29, 03/05/18) plans to 1089 1045 interventions (e.g., muscles soreness, bloating) will recruit from June to October 2018. 1090 1046 be proactively defined via factsheets and instructions 1047 within each module, and recorded by the participants 1048 on the web interface as they occur. These logged ACKNOWLEDGMENTS 1091 1049 events will be triaged by the Physical Activity and/or 1050 Nutrition module leader and trainers as appropriate This trial was funded by a National Health and 1092 1051 for follow-up, referral to GP or study physicians, or Medical Research (NHMRC) Dementia Research 1093 1052 reporting to ethics. Team Grant (APP1095097). MV was supported 1094 1053 A serious adverse event (SAE) is death, a life- by a NHMRC career development fellowship 1095 1054 threatening adverse event, inpatient hospitalization or (APP1112813). Participants will be recruited from 1096 1055 prolongation of existing hospitalization, a persistent the 45 and Up Study (http://www.saxinstitute.org.au). 1097 1056 or significant incapacity or substantial disruption of The 45 and Up Study is managed by the Sax Institute 1098 1057 the ability to conduct normal life functions. All SAEs in collaboration with major partner Cancer Council 1099 1058 will be reported to the relevant ethics committees for NSW; and partners: The National Heart Founda- 1100 1059 this trial. tion of Australia (NSW Division); NSW Ministry 1101 Uncorrectedof Health; NSWAuthor Government FamilyProof & Community 1102 1060 Ethics Services – Ageing, Carers and the Disability Council 1103 NSW; and the Australian Red Cross Blood Service. 1104 1061 All trial procedures have been approved by the We thank the many thousands of people participat- 1105 1062 University of New South Wales Human Research ing in the 45 and Up Study. We acknowledge the 1106 1063 Ethics Committee (#16252) and NSW Population and contributions of the Maintain Your Brain Collabora- 1107

428 M. Heffernan et al. / Maintain Your Brain protocol 15

1108 tive Team: Nan Bosler, Brigid Betz-Stablein, Nancy [9] Ngandu T, Lehtisalo J, Solomon A, Levalahti¨ E, Ahtiluoto S, 1162 1109 Briggs, Helen Christensen, Igor Chuprov, Andrew Antikainen R, Backman¨ L, Hanninen¨ T, Jula A, Laatikainen 1163 T, Lindstrom¨ J, Mangialasche F, Paajanen T, Pajala S, Pel- 1164 1110 Hayen, Miia Kivipelto, Helen Lapsley, Sally Red- tonen M, Rauramaa R, Stigsdotter-Neely A, Strandberg T, 1165 1111 man, Juan Carlo San Jose, and Heidi Welberry; and Tuomilehto J, Soininen H, Kivipelto M (2015) A 2 year 1166 1112 the MYB Data Safety Monitoring Board. MYB uses multidomain intervention of diet, exercise, cognitive train- 1167 ing, and vascular risk monitoring versus control to prevent 1168 1113 tasks developed by Cogstate Ltd, Cambridge Brain cognitive decline in at-risk elderly people (FINGER): A 1169 1114 Sciences, and NeuroNation. MYB thanks Signs Pub- randomised controlled trial. Lancet 385, 2255-2263. 1170 1115 lishing for a non-exclusive license to use traditional [10] Andrieu S, Guyonnet S, Coley N, Cantet C, Bonnefoy M, 1171 1116 Mediterranean recipes from Food as Medicine: Cook- Bordes S, Bories L, Cufi M-N, Dantoine T, Dartigues J-F, 1172 Desclaux F, Gabelle A, Gasnier Y, Pesce A, Sudres K, Tou- 1173 1117 ing for Your Best Health by Sue Radd and Adventist chon J, Robert P, Rouaud O, Legrand P, Payoux P, Caubere 1174 1118 Media Network for production and use of cooking J-P,Weiner M, Carrie´ I, Ousset P-J, VellasB (2017) Effect of 1175 1119 videos. long-term omega 3 polyunsaturated fatty acid supplementa- 1176 1120 Authors’ disclosures available online (https:// tion with or without multidomain intervention on cognitive 1177 function in elderly adults with memory complaints (MAPT): 1178 1121 www.j-alz.com/manuscript-disclosures/18-0572r1). A randomised, placebo-controlled trial. Lancet Neurol 16, 1179 377-389. 1180 [11] van Charante EPM, Richard E, Eurelings LS, van Dalen 1181 1122 SUPPLEMENTARY MATERIAL J-W, Ligthart SA, van Bussel EF, Hoevenaar-Blom MP,Ver- 1182 meulen M, van Gool WA (2016) Effectiveness of a 6-year 1183 1123 The supplementary material is available in the multidomain vascular care intervention to prevent dementia 1184 (preDIVA): A cluster-randomised controlled trial. Lancet 1185 1124 electronic version of this article: http://dx.doi.org/ 388, 797-805. 1186 1125 10.3233/JAD-180572. [12] Lampit A, Hallock H, Valenzuela M (2014) Computerized 1187 cognitive training in cognitively healthy older adults: A sys- 1188 tematic review and meta-analysis of effect modifiers. PLoS 1189 1126 REFERENCES Med 11, e1001756. 1190 [13] Lampit A, Valenzuela M (2015) Pointing the FINGER at 1191 1127 [1] Wortmann M (2014) World Alzheimer report 2014: multimodal dementia prevention. Lancet 386, 1625-1626. 1192 1128 Dementia and risk reduction. https://www.alz.co.uk/ [14] 45 and Up Study Collaborators (2008) Cohort profile: The 1193 1129 research/world-report-2014 45 and Up Study. Int J Epidemiol 37, 941-947. 1194 1130 [2] Ferri CP, Prince M, Brayne C, Brodaty H, Fratiglioni L, [15] Kroenke K, Spitzer RL, Williams JBW (2001) The PHQ-9: 1195 1131 Ganguli M, Hall K, Hasegawa K, Hendrie H, Huang Y,Jorm Validity of a brief depression severity measure. J Gen Intern 1196 1132 A, Mathers C, Menezes PR, Rimmer E, Scazufca M (2005) Med 16, 606-613. 1197 1133 Global prevalence of dementia: A Delphi consensus study. [16] Beck AT, Ward CH, Mendelson MM, Mock JJ, Erbaugh JJ 1198 1134 Lancet 366, 2112-2117. (1961) An inventory for measuring depression. Arch Gen 1199 1135 [3] Access Economics (2009) Keeping dementia front of Psychiatry 4, 561-571. 1200 1136 mind: Incidence and prevalence 2009-2050, Report by [17] Altman DG, Bland JM (2005) Treatment allocation by min- 1201 1137 Access Economics Pty Limited for Alzheimer’s Australia. imisation. BMJ 330, 843. 1202 1138 https://www.dementia.org.au/sites/default/files/20090800 [18] Kim S, Cherbuin N, Anstey KJ (2016) Assessing reliability 1203 1139 Nat AE FullKeepDemFrontMind.pdf of short and tick box forms of the ANU-ADRI: Conve- 1204 1140 [4] Barnes DE, Yaffe K (2011) The projected effect of risk factor nient alternatives of a self-report Alzheimer’s disease risk 1205 1141 reduction on Alzheimer’s disease prevalence. Lancet Neurol assessment. Alzheimers Dement (N Y) 2, 93-98. 1206 1142 10, 819-828. [19] Anstey KJ, Cherbuin N, Herath PM (2013) Development of 1207 1143 [5] Imtiaz B, Tolppanen AM, Kivipelto M, Soininen H (2014) a new method for assessing global risk of Alzheimer’s dis- 1208 1144 Future directions in Alzheimer’s disease from risk factors ease for use in population health approaches to prevention. 1209 1145 to prevention. Biochem Pharmacol 88, 661-670. Prev Sci 14, 411-421. 1210 1146 [6] Richard E, Andrieu S, Solomon A, Mangialasche F, Ahtilu- [20] Pereira MA, FitzerGerald SJ, Gregg EW, Joswiak ML, Ryan 1211 1147 oto S, Moll van Charante EP, Coley N, Fratiglioni L, Neely WJ, Suminski RR, Utter AC, Zmuda JM (1997) A collec- 1212 1148 AS, Vellas B, van Gool WA, Kivipelto M (2012) Method- tion of Physical Activity Questionnaires for health-related 1213 1149 ological challenges in designing dementia prevention trials - research. Med Sci Sports Exerc 29, S1-205. 1214 1150 the European Dementia Prevention Initiative (EDPI). J Neu- [21] Hagstromer¨ M, Oja P, Sjostr¨ om¨ M (2007) The Interna- 1215 1151 rol Sci 322, 64-70. tional Physical Activity Questionnaire (IPAQ): A study of 1216 1152 [7] Law LLF, Barnett F, Yau MK, Gray MA (2014) Effects of concurrent and construct validity. Public Health Nutr 9, 1217 1153 combined cognitive and exercise interventions on cognition 755-762. 1218 1154 in older adults with and without cognitive impairment: A [22] Marsh AP, Wrights AP, Haakonssen EH, Dobrosielski MA, 1219 1155 systematic review. Ageing Res Rev 15,Uncorrected 61-75. Chmelo Author EA, Barnard RT, Pecorella Proof A, Ip EH, Rejeski WJ 1220 1156 [8] Karssemeijer EGA, Aaronson JA, Bossers WJ, Smits T, (2015) The Virtual Short Physical Performance Battery. 1221 1157 Olde Rikkert MGM, Kessels RPC (2017) Positive effects J Gerontol A Biol Sci Med Sci 70, 1233-1241. 1222 1158 of combined cognitive and physical exercise training on [23] Radd-Vagenas S, Daniel K, Noble Y, Fiatarone Singh MA, 1223 1159 cognitive function in older adults with mild cognitive Flood VM (2018) Reliability of new index tool to assess 1224 1160 impairment or dementia: A meta-analysis. Ageing Res Rev adherence to Mediterranean diet among people with mild 1225 1161 40, 75-83. cognitive impairment. Nutr Diet 75, 11. 1226

429 16 M. Heffernan et al. / Maintain Your Brain protocol

1227 [24] Schroder H, Fito M, Estruch R, Martinez-Gonzalez MA, [36] National Health and Medical Research Council 1291 1228 Corella D, Salas-Salvado J, Lamuela-Raventos R, Ros E, (2013) Australian Dietary Guidelines, Canberra. 1292 1229 Salaverria I, Fiol M, Lapetra J, Vinyoles E, Gomez-Gracia https://www.nhmrc.gov.au/guidelines-publications/n55 1293 1230 E, Lahoz C, Serra-Majem L, Pinto X, Ruiz-Gutierrez V, [37] US Department of Health and Human Services Dietary 1294 1231 Covas MI (2011) A short screener is valid for assessing Guidelines for Americans, 2015-2020, U.S. Department of 1295 1232 Mediterranean diet adherence among older Spanish men Health and Human Services. 1296 1233 and women. J Nutr 141, 1140-1145. [38] Sofi F, Macchi C, Abbate R, Gensini GF, Casini A (2013) 1297 1234 [25] Estruch R, Ros E, Salas-Salvado J, Covas MI, Corella D, Mediterranean diet and health status: An updated meta- 1298 1235 Aros F, Gomez-Gracia E, Ruiz-Gutierrez V, Fiol M, Lape- analysis and a proposal for a literature-based adherence 1299 1236 tra J, Lamuela-Raventos RM, Serra-Majem L, Pinto X, score. Public Health Nutr 17, 2769-2782. 1300 1237 Basora J, Munoz MA, Sorli JV, Martinez JA, Martinez- [39] Psaltopoulou T, Sergentanis TN, Panagiotakos DB, Ser- 1301 1238 Gonzalez MA (2013) Primary prevention of cardiovascular gentanis IN, Kosti R, Scarmeas N (2013) Mediterranean 1302 1239 disease with a Mediterranean diet. N Engl J Med 368, diet, stroke, cognitive impairment, and depression: A meta- 1303 1240 1279-1290. analysis. Ann Neurol 74, 580-591. 1304 1241 [26] Valls-Pedret C, Sala-Vila A, Serra-Mir M, Corella D, de [40] Opie RS, Ralston RA, Walker KZ (2013) Adherence to 1305 1242 la Torre R, Martinez-Gonzalez MA, Martinez-Lapiscina a Mediterranean-style diet can slow the rate of cognitive 1306 1243 EH, Fito M, Perez-Heras A, Salas-Salvado J, Estruch R, decline and decrease the risk of dementia: A systematic 1307 1244 Ros E (2015) Mediterranean diet and age-related cognitive review. Nutr Diet 70, 206-217. 1308 1245 decline: A randomized clinical trial. JAMA Intern Med 175, [41] Radd-Vagenas S, Duffy SL, Naismith SL, Brew BJ, Flood 1309 1246 1094-1103. VM, Fiatarone Singh MA (2018) Effect of the Mediter- 1310 1247 [27] Spitzer RL, Kroenke K, Williams JB, Lowe B (2006) A ranean diet on cognition and brain morphology and function: 1311 1248 brief measure for assessing generalized anxiety disorder: A systematic review of randomized controlled trials. Am J 1312 1249 The GAD-7. Arch Intern Med 166, 1092-1097. Clin Nutr 107, 389-404. 1313 1250 [28] Valenzuela MJ, Sachdev P (2007) Assessment of complex [42] Newby JM, Mackenzie A, Williams AD, McIntyre K, 1314 1251 mental activity across the lifespan: Development of the Life- Watts S, Wong N, Andrews G (2013) Internet cognitive 1315 1252 time of Experiences Questionnaire (LEQ). Psychol Med 37, behavioural therapy for mixed anxiety and depression: A 1316 1253 1015-1025. randomized controlled trial and evidence of effectiveness in 1317 1254 [29] Chodzko-Zajko WJ, Proctor DN, Fiatarone Singh MA, Min- primary care. Psychol Med 43, 2635-2648. 1318 1255 son CT, Nigg CR, Salem GJ, Skinner JS (2009) American [43] Newby JM, Mewton L, Andrews G (2017) Transdiag- 1319 1256 College of Sports Medicine position stand. Exercise and nostic versus disorder-specific internet-delivered cognitive 1320 1257 physical activity for older adults. Med Sci Sports Exerc 41, behaviour therapy for anxiety and depression in primary 1321 1258 1510-1530. care. J Anxiety Disord 46, 25-34. 1322 1259 [30] Livingston G, Sommerlad A, Orgeta V, Costafreda SG, [44] Newby JM, Mewton L, Williams AD, Andrews G 1323 1260 Huntley J, Ames D, Ballard C, Banerjee S, Burns A, Cohen- (2014) Effectiveness of transdiagnostic internet cognitive 1324 1261 Mansfield J, Cooper C, Fox N, Gitlin LN, Howard R, Kales behavioural treatment for mixed anxiety and depression in 1325 1262 HC, Larson EB, Ritchie K, Rockwood K, Sampson EL, primary care. J Affect Disord 165, 45-52. 1326 1263 Samus Q, Schneider LS, Selbaek G, Teri L, Mukadam N [45] Kessler RC, Andrews G, Colpe LJ, Hiripi E, Mroczek DK, 1327 1264 (2017) Dementia prevention, intervention, and care. Lancet Normand SL, Walters EE, Zaslavsky AM (2002) Short 1328 1265 390, 2673-2734. screening scales to monitor population prevalences and 1329 1266 [31] Brown BM, Peiffer JJ, Sohrabi HR, Mondal A, Gupta VB, trends in non-specific psychological distress. Psychol Med 1330 1267 Rainey-Smith SR, Taddei K, Burnham S, Ellis KA, Szoeke 32, 959-976. 1331 1268 C, Masters CL, Ames D, Rowe CC, Martins RN; AIBL [46] Fiatarone Singh MA, Gates N, Saigal N, Wilson GC, 1332 1269 research group (2012) Intense physical activity is associated Meiklejohn J, Brodaty H, Wen W, Singh N, Baune BT, 1333 1270 with cognitive performance in the elderly. Transl Psychiatry Suo C, Baker MK, Foroughi N, Wang Y, Sachdev PS, 1334 1271 2, e191. Valenzuela M (2014) The Study of Mental and Resistance 1335 1272 [32] Mavros Y, Gates N, Wilson GC, Jain N, Meiklejohn J, Bro- Training (SMART) Study—Resistance training and/or cog- 1336 1273 daty H, Wen W, Singh N, Baune BT, Suo C, Baker MK, nitive training in mild cognitive impairment: A randomized, 1337 1274 Foroughi N, Wang Y, Sachdev PS, Valenzuela M, Fiatarone double-blind, double-sham controlled trial. J Am Med Dir 1338 1275 SMA (2017) Mediation of cognitive function improvements Assoc 15, 873-880. 1339 1276 by strength gains after resistance training in older adults with [47] American Psychiatric Association (2013) Diagnostic and 1340 1277 mild cognitive impairment: Outcomes of the study of mental statistical manual of mental disorders : DSM-5. American 1341 1278 and resistance training. J Am Geriatr Soc 65, 550-559. Psychiatric Association, Arlington, VA. 1342 1279 [33] Geidl W, Semrau J, Pfeifer K (2014) Health behaviour [48] Sikkes SA, Pijnenburg YA,Knol DL, de Lange-de Klerk ES, 1343 1280 change theories: Contributions to an ICF-based behavioural Scheltens P, Uitdehaag BM (2013) Assessment of instru- 1344 1281 exercise therapy for individuals with chronic diseases. Dis- mental activities of daily living in dementia: Diagnostic 1345 1282 abil Rehabil 36, 2091-2100. value of the Amsterdam Instrumental Activities of Daily 1346 1283 [34] Ebrahim S, Thompson PW, Baskaran V, Evans K (1997) Living Questionnaire. J Geriatr Psychiatry Neurol 26, 244- 1347 1284 Randomized placebo-controlled trial of brisk walking in the 250. 1348 1285 prevention of postmenopausal osteoporosis.UncorrectedAge Ageing 26, [49] Rabipour Author S, Davidson PSR (2015)Proof Do you believe in 1349 1286 253-260. brain training? A questionnaire about expectations of 1350 1287 [35] Preisinger E, Alacamlioglu Y, Pils K, Bosina E, Metka M, computerised cognitive training. Behav Brain Res 295, 1351 1288 Schneider B, Ernst E (1996) Exercise therapy for osteoporo- 64-70. 1352 1289 sis: Results of a randomised controlled trial. Br J Sports Med [50] Kreidler SM, Muller KE, Grunwald GK, Ringham BM, 1353 1290 30, 209-212. Coker-Dukowitz ZT, Sakhadeo UR, Baron´ AE, Glueck DH 1354

430 M. Heffernan et al. / Maintain Your Brain protocol 17

1355 (2013) GLIMMPSE: Online power computation for linear [52] Norton S, Matthews FE, Barnes DE, Yaffe K, Brayne C 1362 1356 models with and without a baseline covariate. J Stat Softw (2014) Potential for primary prevention of Alzheimer’s dis- 1363 1357 54, i10. ease: An analysis of population-based data. Lancet Neurol 1364 1358 [51] Faul F, Erdfelder E, Lang AG, Buchner A (2007) G*Power 13, 788-794. 1365 1359 3: A ﬂexible statistical power analysis program for the [53] Ashby-Mitchell K, Burns R, Shaw J, Anstey KJ (2017) 1366 1360 social, behavioral, and biomedical sciences. Behav Res Proportion of dementia in Australia explained by common 1367 1361 Methods 39, 175-191. modiﬁable risk factors. Alzheimers Res Ther 9, 11. 1368

Uncorrected Author Proof

431 Trends in Legume Consumption Among Ethnically Diverse Adults in a Longitudinal Cohort Study in Australia

VM Flood1,2,3, J Russell3, S Radd1

1Faculty of Health Sciences, University of Sydney, Australia

2St Vincent’s Hospital Sydney, Australia

3School of Health and Society, University of Wollongong, Australia

The aim of this study was to investigate changes in legume consumption among a population of ethnically diverse middle- aged adults in an urban area of Australia. The Melbourne Collaborative Cohort Study is a prospective cohort study of people aged 40- 69 years at baseline (1990-1994) (n=41,514 at baseline), with follow up 13 years later (2003-2007). One quarter of the participants were migrants from Southern Europe, in particular Greece and Italy. Diet was assessed using a 121 item Food Frequency Questionnaire (FFQ), including 2 questions about legume consumption. The FFQ was previously validated in a population of ethnically diverse participants. At baseline 34% of all participants reported never consuming legumes and this reduced to 12.4% at follow up. Greek and Italian born participants had the highest baseline intake of legume consumption (2.56 and 1.87serves per week, respectively) and these groups also reported the highest negative change in weekly legume consumption over 13 years of follow-up (- 1.29 and -0.65 times per week respectively, p < 0.05). These findings suggest that as people from Southern Europe reside longer in Australia, there is a tendency towards acculturation, with people reporting less adherence to the traditional high legume consumption observed in Mediterranean countries.

Flood, V., Russell, J., & Radd, S. (2015). Trends in legume consumption among ethnically diverse adults in a longitudinal cohort study in Australia. The FASEB Journal, 29(1) (abstract 381.4).

432 APPENDIX THREE

PROSPERO protocol for systematic review

433 PROSPERO International prospective register of systematic reviews

Effect of the Mediterranean diet on cognition and brain morphology and function: a systematic review of randomized controlled trials Sue Radd-Vagenas, Maria Fiatarone Singh, Victoria Flood, Shantel Duffy

Citation Sue Radd-Vagenas, Maria Fiatarone Singh, Victoria Flood, Shantel Duffy. Effect of the Mediterranean diet on cognition and brain morphology and function: a systematic review of randomized controlled trials. PROSPERO 2015 CRD42015024088 Available from: http://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42015024088

Review question The objective of this review is to assess the effect of the Mediterranean diet on cognitive functioning or any measure of brain morphology or function in humans by reviewing randomised controlled trials. Searches The search strategy included a combination of 'Intervention' AND 'Outcome' terms but did not include 'Population', 'Comparator' or 'Study design' terms in order to maximise the search sensitivity. We searched for ‘search terms’ in all fields. The Intervention terms included “cretan diet” OR “Mediterranean diet” OR “Mediterranean dietary pattern*” OR “Mediterranean type diet” OR MedDiet OR MeDi OR MeDiet. The Outcomes terms included “brain-derived neurotrophic factor” OR “CT scan*” OR “executive function” OR “gray matter” OR “grey matter” OR “lumbar puncture” OR “magnetic resonance imaging” OR “mild cognitive impairment” OR “neurological test*” OR “neuropsychological test*” OR “post-mortem brain examination” OR “problem solving” OR “reaction time” OR “spinal puncture” OR “spinal tap” OR “white matter” OR Alzheimer OR BDNF OR brain OR cerebral OR cerebrovascular OR cognition OR cognitive OR cortical OR dement* OR fMRS OR intellect* OR intelligen* OR IQ OR memory OR mental OR MRI OR neural OR neuro* OR neurologic OR neurotransmitter* OR thinking OR tomography OR visuo-spatial. Each search strategy was customised for the coding of each database used in order to encompass all fields and maximise sensitivity. The two search terms were combined with 'AND' to produce the final results pool. The systematic search was conducted across a wide range of databases from the earliest record available to the time of the search. Databases searched included: Ovid interface - MEDLINE (1946-Current), Premedline (not yet indexed in medline), AMED (1985-current), All EBM review databases including Cochrane Database of Systematic Reviews (DSR) (2005-current), ACP Journal Club (1991-current), Database of Abstracts of Reviews of Effects (DARE) (2nd quarter of 2015), Cochrane Central Register of Controlled Trials (CCTR) (June 2015), NHS Economic Evaluation Database (CLEED) (2nd quarter of 2015), Cochrane Methodology Register (CLMCR) (3rd quarter 2012-current), Health Technology Assessment (CLHTA) (2nd quarter 2015). EBSCO interface - CINAHL (1981-current), EMBASE (1947-current), PubMed (inception-current), PsychINFO (1806-July week 2, 2015), Web of Science (excluding entries indexed in MEDLINE) (1900-current). We also hand searched grey literature including CHEBA (http://cheba.unsw.edu.au/), Alzheimer’s Australia (https://fightdementia.org.au/), Alzheimer’s Association (http://www.alz.org/research/), Neuroscience Research Australia (http://www.neura.edu.au/research/themes/ageing-neurodegeneration), McCusker Alzheimer’s Research Foundation (http://alzheimers.com.au/research/), NIH Alzheimer’s Disease Research Centers (https://www.nia.nih.gov/alzheimers/alzheimers-disease-research-centers), BrightFocus Foundation (http://www.brightfocus.org/alzheimers/), Alzheimer’s Research UK (http://www.alzheimersresearchuk.org/). The section of each website labeled ‘research’ or the research related tab was screened in detail on July-Aug 2015. In addition, we queried known experts in dementia and MCI and reviewed reference lists of retrieved papers for potentially eligible articles in order to identify all possible studies. For references with missing data, we contacted corresponding authors for additional information. Types of study to be included Randomised controlled trials (RCTs) published as peer reviewed, full-length articles. All other study designs and publication formats and unpublished data were excluded. Condition or domain being studied

434 Page: 1 / 6 PROSPERO International prospective register of systematic reviews

Over 322,000 individuals were affected by dementia in Australia in 2013, with many more individuals at risk by virtue of their borderline cognitive impairment. Current estimates suggest approximately one million Australians will be affected by 2050. This disorder presents a major challenge to sufferers, carers and the healthcare system. Urgent interventions are required to delay the onset of dementia and slow the progression of cognitive decline. Current research suggests it may be too late to intervene or expect significant health outcomes once dementia has been diagnosed. A window of opportunity may exist during a pre-clinical stage known as mild cognitive impairment (MCI) to alter the course and speed of the cognitive decline. Testing various interventions among individuals with MCI is also important since the majority of such people progress to more severe forms of neurodegenerative disease within five years, at a rate of between 10-15 % per year. While no pharmacological agents have yet been identified to target the underlying pathology, recent studies suggest simple lifestyle improvements may influence memory and cognition to slow progression towards dementia and the rate of cognitive decline among those with MCI. In particular, the Mediterranean diet has been identified in epidemiological studies as a potentially beneficial modality. However, relatively few clinical trials exist testing the effect of the Mediterranean diet on cognitive functioning or brain morphology. Research into the effects of the Mediterranean diet is imperative in establishing its potential as a therapeutic tool. Participants/population No restrictions were placed on the participants in the clinical trials as long as they were human. Participants already consuming the Mediterranean diet as defined in this review were excluded if they had habitually consumed this diet within three months of enrolment on the trial. Intervention(s), exposure(s) For the purposes of this review the intervention used must have been described as a Mediterranean diet or Mediterranean-style dietary pattern. The intervention diet may have been provided within a Mediterranean country or other country. Supervision of cooking or food intake was not required. Interventions testing the short term or immediate effect of exposure to single foods or nutrients were excluded. Trials of less than 10 days of dietary intervention were excluded. Unpublished data, reviews, editorials, letters, commentaries, abstracts, news pieces, lectures, articles in lay literature were excluded. Definitions for the ‘Mediterranean diet’ for the purposes of quality review can be found in the accompanying PDF document. (See link below under "Reference and/or URL for published protocol"). The criterion are based primarily on the lowest denominator used in the PREDIMED 14-point Mediterranean Diet Adherence Screener (MEDAS) [Schroder, Fito et al. (2011)] or the new Mediterranean diet pyramid from the Mediterranean Diet Foundation, Barcelona, Spain [Bach-Faig, Berry et al. (2011)]. Serves are defined as per Australian standard serve sizes in the Australian Guide to Healthy Eating as appropriate. Comparator(s)/control The control group must have been provided or advised to consume an alternative diet to the Mediterranean diet for comparison. This may be a healthy dietary pattern as described in the literature (e.g. DASH diet, low- GI diet) or it may be the usual diet of the participant. A comparator group may also be one where additional non-dietary interventions were included, as long as these same modalities were also given to the intervention group. Context The clinical trials may have been conducted anywhere in the world and could have included either free living subjects or those in nursing homes or hospital metabolic wards or other residential or healthcare facilities. The intervention may also have been administered in various ways, including provision of printed educational materials, verbal instruction, internet- or mobile-assisted technology assisted implementation and/or provision of prepared foods or dietary constituents (such as olive oil or nuts). It may have been provided individually or within a group setting. Participants may or may not have been asked to keep records or provide biomarker evidence of their adherence to the prescribed diet. Main outcome(s) Cognitive functioning as assessed by any of the following or other validated tests in the literature: 1. Any validated test of intelligence or cognitive performance in any domain or globally (e.g. Mini-Mental State Examination (MMSE) - Folstein, Folstein et al. (1975); Alzheimer’s Disease Assessment Scale (ADAS-Cog) - Graham, Cully et al. (2004); Matrices - Kaufman and Lichtenberger (1999); Similarities - Kaufman and Lichtenberger (1999); Trail Making Test (TMT) - Bowie and Harvey (2006); Logical Memory, Benton Visual

435 Page: 2 / 6 PROSPERO International prospective register of systematic reviews

Retention Test (BVRT) - Sivan (1992); Symbol Digit Modalities Test (SDMT) - Strauss, Sherman et al. (2006); Category Fluency - Strauss, Sherman et al. (2006); Controlled Oral Word Association Test (COWAT) - Strauss, Sherman et al. (2006). 2. Subjective reporting of memory or cognitive concern, reported on validated scales such as the Memory Awareness Scale or obtained during semi-structured interview (e.g. GPCog - Brodaty, Pond et al. (2002); Subjective Memory Complaint (SMC) - Abdulrab and Heun (2008); Life Experience Questionnaire (LEQ) - Valenzuela and Sachdev (2007); Memory Awareness Rating Scale- Memory Functioning (MARS-MF) - Carter, Clare et al. (2002)). 3. Global assessment of cognitive function by semi-structured interview by a clinician (e.g. Clinical Dementia Rating (CDR) - Hughes, Berg et al. (1982); DSM-5R criteria for cognitive impairment or dementia – American Psychiatric Association (2013) etc.). 4. Self- or informant report of functional deficits related to cognition (e.g. Bayer Informant Instrumental Activities of Daily Living (B-IADLs) - Hindmarch, Lehfeld et al. (1998). AND/OR Brain morphology and/or function as assessed by any of the following or other validated tests in the literature: 1. MRI scans. 2. fMRS scans. 3. CT scans. 4. Lumbar puncture. 5. Measurement of neurotransmitters or biomarkers such as Brain-derived neurotrophic factor (BDNF) or others. 6. Post-mortem examination of brain. Timing and effect measures Measures must have been taken at baseline before randomisation to the experimental or control condition, and at least one time point after the intervention or control period began. The minimum time period between baseline and follow-up cognitive/brain assessments had to have been 10 days. Alternatively, measures could have been taken post intervention only, so long as such measures were provided for both experimental and control conditions. Additional outcome(s) The following secondary outcomes will be extracted but may not be reported in the primary systematic review, especially if unrelated to cognitive outcomes. 1. Body composition (weight, BMI, fat mass and distribution, FFM, LBM or muscle mass, bone density or mass, anthropometric measurements such as WC, WHR, skinfolds, etc.). 2. Cardiometabolic profile (FBS, insulin, insulin resistance, inflammatory markers, BP, lipid profile, coagulation factors, homocysteine, arterial stiffness etc.). 3. Functional status (ADL and IADL status, need for assistive devices or persons, need for residential care or home care support services, etc.). 4. Psychological status (depressive symptoms, anxiety, sleep disturbance, self-efficacy, wellbeing, quality of life, etc.) reported by such measures as Geriatric Depression Scale (GDS) - Brink, Yesavage et al. (1982); Depression Anxiety and Stress Scale (DASS) - Ng, Trauer et al. (2007); Scale of Psychological Wellbeing (SPWB) - Ryff and Keyes (1995); Duke Social Support (DSS) - Koenig, Westlund et al. (1993); Life Satisfaction Scale (LSS) - Andrews and Withey (1976); Physical and Mental Health Summary Scales (SF36) - Ware and Sherbourne (1992); Quality of Life Scales (QoLS) - Burckhardt and Anderson (2003), or any other validated tools in the literature. *Covariates: Age, sex, ethnicity, geographical setting, residential setting, health status (all diseases and medications, health care utilisation or costs), habitual physical activity level, sedentary time, socio-economic status, educational level, financial status. *Process measures: Adherence to the Mediterranean diet, adverse events, perceptions of burden and benefits of diet. Timing and effect measures Covariates measured at baseline, and at follow up for those that could change (e.g. health status). Secondary outcomes measured at baseline and 10 or more days after intervention has commenced. Process measures captured throughout the trial or retrospectively at trial conclusion. Data extraction (selection and coding) Screening procedure: Studies were selected for inclusion using a screening process, which progressively eliminated irrelevant articles. The following steps were followed: 1. Duplicates were removed from the reference management software (EndNote X7) by the primary author (SR-V). 2. The titles and abstracts were then screened by SR-V and those which did not satisfy the inclusion/exclusion criteria were removed, with the reason for removal being recorded. Any queries were discussed and agreed upon with the co- authors of this review by consensus. 3. The full text of potentially relevant articles was then retrieved and comprehensively reviewed by SR-V and SD any queries discussed with VF and/or MFS, if required to reach consensus. Those which did not meet inclusion/exclusion criteria were removed with the reason for removal

436 Page: 3 / 6 PROSPERO International prospective register of systematic reviews being recorded. 4. The remaining eligible articles were included in the systematic review. Data extraction procedure: Two authors (SR-V) and (SD) then independently extracted data from the included studies using a standardised, pre-designed form, which was pilot tested. Any discrepancies in the chosen data or analysis were discussed and resolved by mutual consent, or involvement of a third and/or fourth reviewer (VF and/or MFS) prior to tabulation. The data extracted were based on the following categories: 1. Study design/setting: author, study name, publication year, study year, study design, city & country, setting, study duration. 2. Study quality/risk of bias: eligibility criteria specified, random allocation of subjects, concealed allocation of subjects, similar prognostic indicators at baseline, blinding of subjects, blinding of clinicians, blinding of assessors, key outcome obtained for >85 % subjects initially allocated to groups, intention to treat analysis, between group analysis for key outcome/s, point measures provided, variability measures provided, diet designed by dietitian, minimum amounts specified for required foods, diet assessment method described, dietary compliance monitored, Mediterranean diet adherence tool validated, quality of validation study, social setting for meals reported. 3. Intervention/Control condition: Mediterranean diet definition, how diet administered/who provided education, educational tools provided, frequency of dietary intervention, individual or group intervention, provision of free food samples, diet tolerance, perception of burden or benefit of diet, comparator diet/s, method used to assess dietary intake, method used to assess adherence to Mediterranean diet, method adherence tool administered. 4. Cohort: subject number, % female, age range, mean age, ethnicity, geographical location, residential setting, health status, years since diagnosis if cognitively impaired, composite estimate of cognitive functioning at baseline, weight/BMI/other body composition, habitual physical activity level, sedentary time, size of family/residential unit, whether participant was cook for family/residential unit, SES or financial level, educational level. 5. Outcome: blinding of assessors, cognitive measurement tools/tests used, cognitive outcomes number, diagnosis criteria used for cognitive outcomes, brain morphology/function tests used, body composition tests used, cardiometabolic profile tests used, functional status tests used, psychological status tests used. 6. Statistical results: mean and standard deviation (SD), effect sizes (ES's), confidence intervals, mean differences between groups, statistical tests between and within groups, drop out rate, adjustments, missing data. Where possible, ES's not provided were calculated from extracted data for each outcome within each study or via contact with corresponding authors. 7. Other: funding source. Risk of bias (quality) assessment Two reviewers (SR-V and SD) independently assessed the quality of the design and reporting of the included studies using the PEDro scale [validated by Maher, Sherrington et al. (2003); Foley, Bhogal et al. (2006)]. Additional criterion used to further evaluate the quality of the interventions included: whether the intervention diet was designed and administered by a dietitian; whether minimum amounts for consumption of Mediterranean foods were specified to participants; whether the prescribed diet meets the defined minimum criterion (see uploaded table with Minimum Criterion for Traditional Mediterranean Diet); whether diet tolerance was reported; whether the frequency and setting for diet instruction was reported; whether diet compliance was assessed; and whether perception of diet burden or benefit was reported. However, these did not form part of the overall PEDro score. Any discrepancies were resolved by discussion with a third and/or fourth reviewer (VF and/or MFS). Strategy for data synthesis The data extracted from the included trials will be aggregated (at the level of the study, not the individual) unless individual data are available for participants in all included studies, which is not anticipated. At this time, a meta-analysis is not planned and a narrative summary is planned, due to anticipated heterogeneity in interventions and outcomes. Individual trial ESs will be calculated for all outcomes possible, as noted above. However, if sufficient homogenous data is available when the search is completed to allow a quantitative synthesis across studies for some outcomes, a meta-analysis will be added. Analysis of subgroups or subsets None planned. However, if sufficient data are available, subgroup analysis based on residential status (community dwelling vs. residential care), sex, age (young, middle-aged, old) and baseline cognitive status (intact, mild cognitive impairment, dementia), or other potentially relevant sub-groups will be conducted. Contact details for further information Sue Radd-Vagenas [email protected]

437 Page: 4 / 6 PROSPERO International prospective register of systematic reviews

Organisational affiliation of the review The University of Sydney www.sydney.edu.au Review team members and their organisational affiliations Mrs Sue Radd-Vagenas. The University of Sydney Professor Maria Fiatarone Singh. The University of Sydney Professor Victoria Flood. The University of Sydney Dr Shantel Duffy. The University of Sydney Collaborators Professor Sharon Naismith. The University of Sydney Professor Bruce Brew. The University of New South Wales Anticipated or actual start date 07 July 2015 Anticipated completion date 21 July 2017 Funding sources/sponsors The University of Sydney Conflicts of interest None known Language English Country Australia Stage of review Review_Completed_published Details of final report/publication(s) Radd-Vagenas S et al, Effect of the Mediterranean diet on cognition and brain morphology and function: a systematic review of randomized controlled trials, Am J Clin Nutr 2018;107:389–404. https://academic.oup.com/ajcn/article/107/3/389/4939347 https://doi.org/10.1093/ajcn/nqx070 Subject index terms status Subject indexing assigned by CRD Subject index terms Alzheimer Disease; Brain; Cognition; Cognition Disorders; Diet, Mediterranean; Humans Date of registration in PROSPERO 20 July 2015 Date of publication of this version 10 April 2018 Details of any existing review of the same topic by the same authors Stage of review at time of this submission

438 Page: 5 / 6 PROSPERO International prospective register of systematic reviews

Stage Started Completed Preliminary searches Yes Yes Piloting of the study selection process Yes Yes Formal screening of search results against eligibility criteria Yes Yes Data extraction Yes Yes Risk of bias (quality) assessment Yes Yes Data analysis Yes Yes

Versions 20 July 2015 29 January 2016 08 November 2016 30 January 2017 22 June 2017 10 April 2018

PROSPERO This information has been provided by the named contact for this review. CRD has accepted this information in good faith and registered the review in PROSPERO. CRD bears no responsibility or liability for the content of this registration record, any associated files or external websites.

439 Page: 6 / 6 APPENDIX FOUR

Corrections within PREDIMED for cognitive outcomes

440 Table 4.1. Comparison of baseline cognitive test scores and changes by study group - original values*

Mean (95% CI)

Mediterranean Diet Plus Mediterranean Diet Plus Nuts Control Diet Extra Virgin Olive Oil (n = 112) (n = 95) Variable P Valuea (n = 127) MMSEb

Baseline 28.01 (27.79 to 28.24) 28.11 (27.87 to 28.35) 28.38 (28.12 to 28.65) 0.10

Change 0.16 (−0.12 to 0.44) −0.07 (−0.36 to 0.23) −0.26 (−0.57 to 0.06) 0.15

RAVLT, total learningb

Baseline 39.31 (37.92 to 40.70) 39.46 (37.98 to 40.94) 39.24 (37.63 to 40.85) 0.98

Change 4.50 (3.24 to 5.77)c 4.26 (2.91 to 5.60) 2.10 (0.64 to 3.57) 0.04

RAVLT, delayed recallb

Baseline 6.61 (6.09 to 7.13) 6.48 (5.93 to 7.03) 6.37 (5.77 to 6.96) 0.83

Change 1.47 (1.02 to 1.92) 1.80 (1.32 to 2.28) 0.95 (0.43 to 1.47) 0.06

441 Mean (95% CI)

Mediterranean Diet Plus Mediterranean Diet Plus Nuts Control Diet Extra Virgin Olive Oil (n = 112) (n = 95) Variable P Valuea (n = 127) Paired associatesb

Baseline 15.25 (14.69 to 15.82) 15.25 (14.65 to 15.85) 14.89 (14.24 to 15.54) 0.66

Change 0.25 (−0.33 to 0.83) 0.41 (−0.21 to 1.03) −0.09 (−0.76 to 0.58) 0.56

Verbal fluencyd

Baseline 18.44 (17.18 to 19.70) 20.33 (18.86 to 21.79) 19.02 (17.53 to 20.51) 0.16

Change 0.53 (−0.78 to 1.84) −1.51 (−3.08 to 0.06) 0.30 (−1.27 to 1.87) 0.13

Digit Span Forwardd

Baseline 5.33 (5.03 to 5.63) 5.20 (4.85 to 5.56) 5.23 (4.89 to 5.57) 0.85

Change 0.11 (−0.16 to 0.38) 0.36 (−0.01 to 0.73) −0.08 (−0.40 to 0.25) 0.28

Digit Span Backwardd

Baseline 3.76 (3.49 to 4.02) 3.83 (3.52 to 4.14) 3.95 (3.66 to 4.25) 0.63

Change 0.25 (−0.10 to 0.59) 0.34 (−0.12 to 0.81) −0.24 (−0.64 to 0.17) 0.14

442 Mean (95% CI)

Mediterranean Diet Plus Mediterranean Diet Plus Nuts Control Diet Extra Virgin Olive Oil (n = 112) (n = 95) Variable P Valuea (n = 127) Color Trail Test part1d,e

Baseline 62.60 (56.29 to 68.91) 61.15 (53.56 to 68.74) 57.00 (49.61 to 64.38) 0.53

Change −5.77 (−11.25 to −0.28) 2.48 (−5.29 to 10.26) 4.53 (−2.11 to 11.17) 0.045

Color Trail Test part2d,e

Baseline 136.55 (123.19 to 149.90) 131.59 (115.13 to 148.05) 129.99 (114.39 to 145.60) 0.81

Change 5.66 (−10.23 to 21.55)c 24.23 (1.36 to 47.10) 37.56 (18.14 to 56.97) 0.045

MMSE, Mini-Mental State Examination; RAVLT, Rey Auditory Verbal Learning Test. * Data extracted from Table 3 in original publication (May 11, 2015) from Valls-Pedret, C., Sala-Vila, A., Serra-Mir, M., Corella, D., de la Torre, R., Martínez-González, M. Á., . . . Salas-Salvadó, J. (2015). Mediterranean diet and age-related cognitive decline: A randomized clinical trial. JAMA Internal Medicine, 175(7), 1094-1103. doi:10.1001/jamainternmed.2015.1668 a P value by analysis of covariance adjusted for sex, baseline age, years of education, marital status, APOE ε4 genotype, ever smoking, baseline body mass index, energy intake, physical activity, type 2 diabetes mellitus, hyperlipidaemia, ratio of total cholesterol to HDL-C, statin treatment, hypertension, use of anticholinergic drugs, and time of follow-up (not in baseline, only in change). b Measured in 334 participants (127, 112, and 95 participants, respectively). c Significantly different from control group (Bonferroni post hoc test).

443 d Measured in 96 participants (41, 25, and 30 participants, respectively). e Lower scores indicate improvement.

444 Table 4.2. Comparison of baseline cognitive test scores and changes by study group - corrected values*

Mean (95% CI)

Mediterranean Diet Plus Mediterranean Diet Plus Nuts Control Diet Extra Virgin Olive Oil (n = 112) (n = 95) Variable P Valuea (n = 127) MMSEb

Baseline 28.03 (27.80 to 28.25) 28.10 (27.63 to 28.35) 28.37 (28.11 to 28.63) 0.15

Change 0.15 (−0.13 to 0.42) −0.06 (−0.36 to 0.23) −0.24 (−0.56 to 0.08) 0.20

RAVLT, total learningb

Baseline 39.40 (38.01 to 40.79) 39.41 (37.94 to 40.89) 39.18 (37.57 to 40.79) 0.97

Change 4.38 (3.11 to 5.64) 4.33 (2.99 to 5.64) 2.19 (0.73 to 3.65) 0.052

RAVLT, delayed recallb

Baseline 6.62 (6.11 to 7.14) 6.46 (5.91 to 7.01) 6.37 (5.77 to 6.96) 0.81

Change 1.46 (1.01 to 1.91) 1.82 (1.34 to 2.29) 0.94 (0.42 to 1.46) 0.054

445 Mean (95% CI)

Mediterranean Diet Plus Mediterranean Diet Plus Nuts Control Diet Extra Virgin Olive Oil (n = 112) (n = 95) Variable P Valuea (n = 127) Paired associatesb

Baseline 15.27 (14.71 to 15.84) 15.24 (14.64 to 15.84) 14.87 (14.22 to 15.23) 0.62

Change 0.24 (−0.34 to 0.82) 0.42 (−0.21 to 1.04) −0.08 (−0.76 to 0.59) 0.56

Verbal fluencyc

Baseline 18.42 (17.14 to 19.70) 20.38 (18.90 to 21.86) 19.00 (17.48 to 21.86) 0.16

Change 0.53 (−0.78 to 1.84) −1.51 (−3.08 to 0.06) 0.30 (−1.27 to 1.87) 0.14

Digit Span Forwardc

Baseline 5.37 (5.08 to 5.67) 5.23 (4.88 to 5.57) 5.16 (4.83 to 5.49) 0.65

Change 0.07 (−0.20 to 0.35) 0.37 (−0.00 to 0.74) −0.04 (−0.36 to 0.28) 0.30

Digit Span Backwardc

Baseline 3.78 (3.52 to 4.05) 3.83 (3.52 to 4.14) 3.92 (3.62 to 4.22) 0.80

Change 0.23 (−0.12 to 0.58) 0.36 (−0.11 to 0.83) −0.23 (−0.64 to 0.18) 0.16

446 Mean (95% CI)

Mediterranean Diet Plus Mediterranean Diet Plus Nuts Control Diet Extra Virgin Olive Oil (n = 112) (n = 95) Variable P Valuea (n = 127) Color Trail Test part 1c,d

Baseline 62.62 (56.24 to 69.01) 60.80 (53.20 to 68.40) 57.28 (49.76 to 64.80) 0.58

Change −6.03 (−11.66 to −0.40) 2.84 (−5.07 to 10.74) 4.58 (−2.14 to 11.31) 0.043

Color Trail Test part 2c,d

Baseline 134.18 (120.79 to 147.58) 131.74 (115.44 to 148.04) 132.99 (117.29 to 148.68) 0.97

Change 5.55 (−10.94 to 21.94) 24.71 (1.29 to 48.14) 37.30 (17.63 to 56.98) 0.055

MMSE, Mini-Mental State Examination; RAVLT, Rey Auditory Verbal Learning Test. * Data extracted from Table 3 in corrected publication (November 5, 2018; corrected values bolded) from Valls-Pedret, C., Sala-Vila, A., Serra-Mir, M., Corella, D., de la Torre, R., Martínez-González, M. Á., . . . Salas-Salvadó, J. (2015). Mediterranean diet and age-related cognitive decline: A randomized clinical trial. JAMA Internal Medicine, 175(7), 1094-1103. doi:10.1001/jamainternmed.2015.1668 a P value by analysis of covariance adjusted for sex, baseline age, years of education, marital status, APOE ε4 genotype, ever smoking, baseline body mass index, energy intake, physical activity, type 2 diabetes mellitus, hyperlipidaemia, ratio of total cholesterol to HDL-C, statin treatment, hypertension, use of anticholinergic drugs, time of follow-up (not in baseline, only in change), and propensity score for group allocation (30 variables as predictors of allocation).

447 b Measured in 334 participants (127, 112, and 95 participants, respectively). c Measured in 96 participants (41, 25, and 30 participants, respectively). d Lower scores indicate improvement.

448 Table 4.3. Comparison of composite cognitive domains – original and corrected values*

Original Corrected

Composite Group Change score fully P value# Group Change score fully P value# domain adjusted model adjusted model

Memory Med + EVOO 0.04 (−0.09 to 0.18) Med + EVOO 0.04 (−0.10 to 0.17)

Med + Nuts 0.09 (−0.05 to 0.23) 0.04 Med + Nuts 0.10 (−0.04 to 0.24) 0.04 Control −0.17 (−0.32 to −0.01) Control −0.16 (−0.32 to −0.01)

Frontal Med + EVOO 0.23 (0.03 to 0.43) 0.003 Med + EVOO 0.23 (0.02 to 0.43) 0.004 Med + Nuts 0.03 (−0.25 to 0.31) Med + Nuts 0.03 (−0.26 to 0.32)

Control −0.33 (−0.57 to −0.09) Control −0.33 (−0.57 to −0.09)

Global Med + EVOO 0.05 (−0.11 to 0.21) 0.005 Med + EVOO 0.04 (−0.12 to 0.20) 0.008 Med + Nuts −0.05 (−0.27 to 0.18) Med + Nuts −0.04 (−0.27 to 0.19)

Control −0.38 (−0.57 to −0.18) Control −0.37 (−0.56 to −0.17)

Med, Mediterranean diet; EVOO, extra virgin olive oil. * Data extracted from Table 4 in original and corrected publications including abstract (May 11, 2015 and November 5, 2018; corrected values bolded) by Valls-Pedret, C., Sala-Vila, A., Serra-Mir, M., Corella, D., de la Torre, R., Martínez-González, M. Á., . . . Salas-Salvadó,

449 J. (2015). Mediterranean diet and age-related cognitive decline: A randomized clinical trial. JAMA Internal Medicine, 175(7), 1094-1103. doi:10.1001/jamainternmed.2015.1668 #significantly different from control group (Bonferroni post hoc test).

450 APPENDIX FIVE

Comparison of selected Mediterranean diet index tools

451 Original Greek index and variants

Author, year, country (Trichopoulou et al., 1995), Greece Index tool name MDS-1 original (Mediterranean Diet Score) a priori or posteriori a priori - elements only Score range 0-8 Adherence categorised as: KLJK Elements assessed 8 elements: 1. Vegetables 2. Legumes 3. Fruit & nuts 4. Cereals (includes potatoes) 5. MUFA:SFA 6. Meat & meat products 7. Dairy products 8. Ethanol Cut-off points used & Median scoring 1 point for each positive element above & each negative element below median for sex Score derivation source 190-item semi-quantitative FFQ (validated) for index Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method Validity results against N/A dietary reference method* Outcome associated with Survival in older Greeks (aged >70 years) from a tool use prospective study Comments 1. First Mediterranean dietary pattern score (also called MDPS, Mediterranean Diet Pattern Score) 2. Food groups based on Davidson, Passmore, Brock, and Truswell (1979), except for alcohol which was added 3. Cut-off points for scoring based on Mediterranean intakes of elderly Greeks from three rural villages & adjusted for energy intake for sex

452 Author, year, country (Trichopoulou, Costacou, Bamia, & Trichopoulos, 2003), Greece Index tool name MDS-2 (MDS-1 score updated by original authors) a priori or posteriori a priori - elements only Score range 0-9 Elements assessed 9 elements: 1. Vegetables 2. Legumes 3. Fruit & nuts 4. Cereals (includes potatoes) 5. MUFA:SFA 6. Meat & meat products 7. Dairy products 8. Ethanol 9. Fish Cut-off points used & Median scoring 1 point for each positive element above & each negative element below median for sex Score derivation source 150-item semi-quantitative FFQ (validated) for index Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method Validity results against N/A dietary reference method* Outcome associated with Mortality in Greek adults (aged 20-86 years) from the tool use prospective EPIC-Greece study Comments 1. Original score updated to include fish, based on benefits described by Hu et al. (2002) from NHS 2. Cut-off points for scoring based on sex specific median intakes of the Greek population studied

453 Author, year, country (Trichopoulou et al., 2005), 9 European countries Index tool name MDS-2 variant: MMD (Modified Mediterranean Diet) a priori or posteriori a priori - elements only Score range 0-9 Elements assessed 9 elements: 1. Vegetables 2. Legumes 3. Fruit & nuts 4. Cereals 5. MUFA + PUFA:SFA 6. Meat & meat products 7. Dairy products 8. Ethanol 9. Fish Cut-off points used & Median scoring 1 point for each positive element above & each negative element below median for sex Score derivation source FFQ (validated) for index 7-d or 14-d FR (in subset) 24 hr recall (in subset) Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method Validity results against N/A dietary reference method* Outcome associated with Mortality in European DGXOWV DJHG\HDUV from the tool use prospective EPIC-elderly study Comments 1. MMD was developed for EPIC-elderly study (n=74 607) 2. Swapped MUFA g/d used in original index with sum of MUFA + PUFA so could apply to non- Mediterranean populations

454 Author, year, country (Fung, Hu, McCullough, & Newby, 2006), USA Index tool name MDS-2 variant: aMED (Alternate Mediterranean Diet Score) a priori or posteriori a priori - elements only Score range 0-9 Elements assessed 9 elements: 1. Vegetables (no potatoes) 2. Legumes 3. Fruit 4. Wholegrains 5. MUFA:SFA 6. Red/processed meat 7. Nuts 8. Alcohol 9. Fish Cut-off points used & Median scoring 1 point for each positive element above & each negative element below median for sex Score derivation source FFQ (validated) for index Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method Validity results against N/A dietary reference method* Outcome associated with Estrogen receptor-negative breast cancer risk in US tool use postmenopausal women from the NHS Comments 1. MDS-2 modified by excluding potato from vegetables, separating fruit & nuts, eliminating dairy products, including only wholegrains in cereals group, including only red/processed meats in meat group & awarding 1 point for ‘moderate’ alcohol intake defined as 5-15g/d

455 Author, year, country (Agnoli et al., 2011), Italy Index tool name MDS-2 variant: IMI (Italian Mediterranean Index) a priori or posteriori a priori - elements only Score range 0-11 Elements assessed 11 elements: • 6 typical Mediterranean foods: 1. Pasta 2. Mediterranean vegetables 3. Fruit excluding FJ 4. Legumes 5. Olive oil 6. Fish • 4 non-Mediterranean foods: 7. Soft drinks 8. Butter 9. Red/processed meat 10. Potatoes •Plus 11. Alcohol Cut-off points used & Tertiles scoring 1 point if in 3rd tertile for Mediterranean foods or 1st tertile for non-Mediterranean foods; otherwise 0, except for alcohol where 1 point for 0.01-12 g/d otherwise 0 for abstainers & those consuming >12 g/d See Table 1 in paper for scoring Score derivation source 154-, 248-, & 438-item semi-quantitative FFQ (for 3 for index Italian regions) (validated) Validated against dietary Y reference method Y/N Dietary reference 1. Six plasma carotenoids (lutein, zeaxanthin, method used canthanxanthiQȕ-FU\SWR[DQWKLQĮ-carotene, & ȕ- carotene) 2. Three other index scores: Healthy Eating Index 2005 (HEI-2005), DASH & MDS-2 Statistics for dietary Spearman's correlation coefficient reference method Validity results against 1. Increased IMI score correlated with higher plasma dietary reference FDURWHQRLGVĮ-FDURWHQH ȡ S ȕ-carotene method* ȡ S OXWHLQ ȡ S ]HD[DQWKLQ ȡ S 2. IMI also correlated with three other index scores ȡ -0.58, depending on index, p<0.001) Outcome associated with 1. Stroke risk in Italian adults from the EPICOR cohort, tool use recruited to the prospective EPIC-Italy study 2. Validity within the same population Comments 1. Validated against dietary biomarkers but only using correlation coefficients

456 Author, year, country (Tognon et al., 2012), Sweden Index tool name MDS-2 variant: rMDS (Refined Modified Mediterranean Diet Index) a priori or posteriori a priori - elements only Score range 0-8 Adherence categorised as: high: >4 Elements assessed 8 elements: 1. Vegetables & potatoes 2. Fruit & juices 3. Wholegrain cereals 4. MUFA+PUFA:SFA 5. Meat, meat products & eggs 6. Dairy products 7. Alcohol 8. Fish & fish products Cut-off points used & Median scoring 1 point for each positive element above & each negative element below median for sex Score derivation source FFQ (validated) for index Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method Validity results against N/A dietary reference method* Outcome associated with Mortality in Northern Swedish (subarctic region) adults tool use from the prospective Vasterbotten Intervention Program study Comments 1. MDS-2 modified by swapping MUFA g/d with sum of MUFA + PUFA, including only wholegrains in cereals group & removing legumes & nuts

457 Author, year, country (Vitale et al., 2019), Italy Index tool name MDS-2 variant: MEDI-Quest a priori or posteriori a priori Score range 0-9 Adherence categorised as: low: 0-3 unsatisfactory: 4-5 high: 6-7 very high: 8-9 Elements assessed 9 elements: 1. Vegetables 2. Legumes & nuts 3. Fruit 4. Wholegrain cereals 5. Olive oil 6. Meat/meat products 7. Foods rich in animal fat, e.g., butter, cake, dairy (excluding milk & yoghurt) 8. Alcohol 9. Fish/fish products Cut-off points used & Absolute scoring 1-2 points awarded for each element if met criteria Values summed & total score divided by 2 Score derivation source MEDI-Quest tool for index Validated against dietary Y reference method Y/N Dietary reference MDS-2 by Trichopoulou 2003 (derived from FFQ using method used sex specific median cut-off points) Statistics for dietary 1. Spearman's correlation coefficient reference method 2. ICC 3. Percent of participants correctly categorised within same score Validity results against 1. Good correlation between total score for MEDI-Quest dietary reference & MDS- ȡ S ,&& S method* 2. Percent of participants awarded same score by the two tools ranged from 55.8%-91.3%, depending on element; for total score 74.7% received the same score Outcome associated with Validity in healthy Southern Italian adults (aged 18-85 tool use years) from a cross-sectional study Comments 1. Quick self-evaluation possible; however, MEDI- Quest tool with cut-off points for scoring not provided within paper 2. Dietary reference method limited by similar recall bias

458 Spanish PREDIMED index and variants

Author, year, country (Schroder et al., 2011), Spain Index tool name MEDAS (Mediterranean Diet Adherence Screener) - Spanish a priori or posteriori a priori Score range 0-14 Elements assessed 14 elements: • 12 assess food consumption: 1. Olive oil 2. Vegetables 3. Fruit 4. Red/processed meat 5. Butter, margarine or cream 6. Sugary drinks 7. Wine 8. Legumes 9. Fish/seafood 10. Commercial pastries 11. Nuts 12. Sofrito • 2 assess food intake habits: 13. Olive oil as primary fat 14. Preference for poultry Cut-off points used & Absolute scoring 1 point awarded for each element if met criteria See Table 2 in paper for scoring Score derivation source MEDAS tool for index Validated against dietary Y reference method Y/N Dietary reference 137-item semi-quantitative FFQ (validated) method used Statistics for dietary 1. Pearson's correlation coefficient reference method 2. ICC 3. General linear modelling for food & nutrient trends 4. Cohen's Kappa 5. Bland-Altman plot

459 Validity results against 1. Total MEDAS score from tool correlated with score dietary reference from FFQ (r=0.52, p<0.001; ICC=0.51, CI N/A, method* p<0.001) 2. MEDAS tool score was positively associated with healthy nutrients & foods from FFQ including vitamin &ȕ-carotene, folic acid, fibre, unsaturated fatty acids, vegetables, fruits, wholegrains, nuts & fish & negatively associated with sodium, saturated fat, sugary drinks & refined cereals 3. Absolute agreement between tool & reference method varied for individual elements (mean N=0.43) 4. MEDAS tool over-estimated score by 5% compared to FFQ in Bland-Altman analysis with upper & lower LOA of 53% & 43%, respectively Outcome associated with Validity in older Spanish adults (aged 55-80 years) from tool use the PREDIMED RCT Comments 1. Allows for quick scoring 2. Provides absolute cut-off points 3. Assesses exposure to some non-Mediterranean foods, i.e., sugary drinks, pastries 4. Dietary reference method limited by similar recall bias

460 Author, year, country (Hebestreit et al., 2017), Germany Index tool name MEDAS (Mediterranean Diet Adherence Screener) - German a priori or posteriori a priori Score range 0-14 Elements assessed 14 elements: • 12 assess food consumption: 1. Olive oil 2. Vegetables 3. Fruit 4. Red/processed meat 5. Butter, margarine or cream 6. Sugary drinks 7. Wine 8. Legumes 9. Fish/seafood 10. Commercial pastries 11. Nuts 12. Sofrito • 2 assess food intake habits: 13. Olive oil as primary fat 14. Preference for poultry Cut-off points used & Absolute scoring 1 point awarded for each element if met criteria See Table 1 in paper for scoring Score derivation source MEDAS tool for index Validated against dietary Y reference method Y/N Dietary reference 1. 148-item semi-quantitative FFQ (validated) method used 2. Plasma omega-6, omega-3 & omega-IDWW\DFLGVȕ- carotene; hsCRP Statistics for dietary 1. Pearson's correlation coefficient reference method 2. Cohen's Kappa 3. ICC 4. Bland-Altman plot 5. t-test for independent groups

461 Validity results against 1. Highest concordance for elements from tool vs. FFQ dietary reference at BL was found for Q1 re olive oil (r=0.70; N=0.70, method* 95% CI 0.51, 0.89; ICC=0.68, 95% CI 0.07, 1.3) & Q12 re nuts (r=0.72; N=0.70, 95% CI 0.52, 0.87; ICC=0.68, 95% CI 0.05, 1.30) 2. Highest concordance for elements from tool vs. FFQ at 3 months was found for Q9 re fish/seafood (r=0.86; N=0.85, 95% CI 0.66, 1.1; ICC=0.91, 95% CI 0.68, 1.14) & Q12 re nuts (r=0.78; N=0.77, 95% CI 0.61, 0.94; ICC=0.76, 95% CI 0.23, 1.28) 3. Total MEDAS score was over-estimated from tool compared to FFQ by 15% at BL & 23% at 3 months 4. At 3 months, intake of fish (using Q9 cut-off) was associated with lower plasma omega-6 (p=0.035); fruit (using Q3 cut-RII ZLWKKLJKHUȕ-carotene (p=0.056), at least 2 serves vegetaEOHVZLWKKLJKHUȕ-carotene (p=0.004); in multi-variate model fish (using Q9 cut- off) was associated with higher plasma omega-3 (p=0.037) & lower omega-6 levels (p=0.026) Outcome associated with Validity in German women (aged 24-72 years) from the tool use LIBRE RCT of breast cancer incidence among carriers of BRCA 1 & BRCA 2 genetic mutations Comments 1. MEDAS translated into German & validated for use in that language 2. Validation against dietary biomarkers included

462 Author, year, country (Papadaki et al., 2018), UK Index tool name MEDAS (Mediterranean Diet Adherence Screener) - English a priori or posteriori a priori Score range 0-14 Elements assessed 14 elements: • 12 assess food consumption: 1. Olive oil 2. Vegetables 3. Fruit 4. Red/processed meat 5. Butter, margarine or cream 6. Sugary drinks 7. Wine 8. Legumes 9. Fish/seafood 10. Commercial pastries 11. Nuts 12. Sofrito • 2 assess food intake habits: 13. Olive oil as primary fat 14. Preference for poultry Cut-off points used & Absolute scoring 1 point awarded for each element if met criteria See Table 1 in paper for scoring Score derivation source MEDAS tool for index Validated against dietary Y reference method Y/N Dietary reference 3-day estimated FR method used Statistics for dietary 1. Pearson's correlation coefficient reference method 2. ICC 3. Cohen's Kappa 4. Bland-Altman plot 5. General linear modelling

463 Validity results against 1. Total MEDAS score from tool correlated moderately dietary reference with FR derived score (r=0.50, p<0.001; ICC=0.53, method* 95% CI 0.07, 0.74, p<0.001) 2. Proportion of participants classified in same category for total score by tool & FR was 45.8% with borderline fair agreement between the methods (N=0.19, 95% CI 0.07, 0.31, p=0.002) 3. Total MEDAS score from tool was over-estimated by 1.47 points compared to score derived from FR (p<0.001) 4. Linear trends were found for key foods & nutrients from the FR and MEDAS tool score tertiles in the expected direction: olive oil, vegetables, red meat & sugary drinks (p=0.001); as well as carbohydrate (p=0.007), total fat (p=0.021), MUFA (p=0.002), PUFA (p=0.022) & vitamin E (p=0.001) 5. Test-retest reliability for MEDAS was good (r=0.69, p<0.001; ICC=0.69, 95% CI 0.571, 0.783, p<0.001) (N=0.38, 95% CI 0.24, 0.52, p<0.001) Outcome associated with Validity in UK adults (mean age 68.3 years) at high CV tool use risk from a cross-sectional study Comments 1. MEDAS translated into English & validated for use in that language

464 Author, year, country (Galilea-Zabalza et al., 2018), Spain Index tool name MEDAS (Mediterranean Diet Adherence Screener) variant: energy reduced a priori or posteriori a priori Score range 0-17 Adherence categorised as: low: 0-6 low to moderate: 7-8 moderate to high: 9-10 high: 11-17 Elements assessed 17 elements: 1. EVOO 2. Fruit 3. Vegetables 4. White bread 5. Wholegrain cereals & pasta 6. Red/processed meat 7. Butter, margarine, cream 8. Sugary drinks 9. Legumes 10. Fish/shellfish 11. Commercial sweets/pastries 12. Nuts 13. Preference for poultry 14. Sofrito 15. Preference for non-caloric sweetener 16. White pasta or rice 17. Wine Cut-off points used & Absolute scoring 1 point awarded for each element if met criteria See Table 1 in paper for scoring Score derivation source MEDAS variant tool for index Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method Validity results against N/A dietary reference method* Outcome associated with Quality of life in older Spanish adults (aged 55-75 tool use years) with overweight or obesity & harbouring the metabolic syndrome from the PREDIMED-Plus RCT

465 Comments 1. Designed for PREDIMED-Plus to take into consideration need for weight loss 2. Includes 4 new Qs re white bread, wholegrains, non- caloric sweeteners & white pasta or rice compared to original MEDAS

466 Author, year, country (Abu-Saad et al., 2019), Israel Index tool name MEDAS (Mediterranean Diet Adherence Screener) variant: I-MEDAS (Israeli Mediterranean diet screener) a priori or posteriori a priori Score range 0-17 Elements assessed 17 elements: 1. Preference for olive oil 2. Non-starchy vegetables including sofrito 3. Fruit (excludes FJ) 4. Red/processed meat 5. Poultry more than red meat 6. Butter/margarine 7. Sweet soft drinks 8. Alcohol 9. Legumes 10. Fish (fresh & preserved) 11. Desserts (cakes & cookies) 12. Nuts 13. Salty snacks 14. Savoury pastries 15. Wholegrains 16. Non-sweetened dairy 17. Hummus/tahini salad Cut-off points used & Absolute scoring 1 point awarded for each element if met criteria See Table 3 in paper for scoring Score derivation source 240-item FFQ for index Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method Validity results against N/A dietary reference method* Outcome associated with Mortality in Israeli adults representing multi-ethnic tool use population (aged 25-74 years) from a prospective study Comments 1. Adapted to reflect Israeli Mediterranean diet & national dietary recommendations 2. Assessment of following MEDAS elements was modified: deleted 4 TBSP EVOO/d; added savoury pastries, wholegrains & dairy; sofrito combined with non-starchy vegetables; both commercial & home-made cake types included; FJ excluded from fruit

467 Other indexes

Author, year, country (Alberti, Fruttini, & Fidanza, 2009), Italy Index tool name MAI (Mediterranean Adequacy Index) a priori or posteriori a priori - elements only Score range 0-100+, depending on food intake patterns Elements assessed 17 elements: • Carbohydrate foods: 1. Bread 2. Cereals • Protective food group: 3. Fish 4. Legumes 5. Potatoes 6. Vegetables 7. Fruit & nuts 8. Olive oil 9. Wine •Food of animal origin: 10. Milk 11. Cheese 12. Meat 13. Eggs 14. Fats & margarines • Sweets food group: 15. Sweet beverages 16. Cakes/pies/cookies 17. Sugar Cut-off points used & Score calculated either by a) dividing sum of percent scoring energy intake from Mediterranean foods (carbohydrate foods + protective food group) by sum of percent energy intake from non-Mediterranean foods (food of animal origin + sweets food group), or b) by ratio of Mediterranean foods (g/d) vs. non-Mediterranean foods, depending on data availability Score derivation source 24-hr recall for index Estimated FR Weighed FR FFQ (depending on cohort) Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method

468 Validity results against N/A dietary reference method* Outcome associated with 1. Adherence to Healthy Reference National tool use Mediterranean Diet (HRNMD) in adults from 23 cohorts in Italy, Greece, USA, Costa Rica, Chile, Spain & Germany 2. Coronary heart disease & total mortality for 16 cohorts from the SCS Comments 1. Based on Nicotera diet in the 1960s at time of SCS, as this was related to low chronic disease outcomes 2. Heavy drinkers can obtain very high MAI scores

469 Author, year, country (Ciccarone et al., 2003), Italy Index tool name Unnamed a priori or posteriori a priori Score range 0-18 Elements assessed 18 elements: 1. Cooked vegetables 2. Raw vegetables 3. Carrots 4. Fruit 5. Fish 6. Cheese 7. Eggs 8. Meat 9. Processed meat 10. Olive oil 11. Vegetable oils 12. Butter 13. Cream 14. Margarine 15. Wine 16. Beer 17. Spirits 18. Ethanol Cut-off points used & Absolute scoring 1 point for each element if met criteria See Table 2 in paper for scoring Score derivation source 24-item semi-quantitative FFQ for index Validated against dietary Y reference method Y/N Dietary reference 7-day FR method used Statistics for dietary Correlation coefficient (assumed Pearson's) reference method Validity results against 1. Good correlation between total score derived from dietary reference FFQ & 7-day FR (r=0.71, p=0.001) method* 2. Total energy underestimated by FFQ, although significant correlation coefficient observed between FFQ & 7-day FR (r=0.457, p=0.007); energy-adjusted correlation coefficients for nutrients ranged from 0.118 (protein) to 0.919 (alcohol) Outcome associated with 1. PAD in Italian adult patients with type 2 diabetes tool use from a case control study 2. Validity within the same population

470 Comments 1. Scoring based on average intake of population within Abruzzo region, Italy 2. Cereals, e.g., pasta, bread, as well as dairy, not scored as considered neutral

471 Author, year, country (Goulet, Nadeau, & Lemieux, 2003), Canada Index tool name MedScore (Mediterranean Score) a priori or posteriori a priori Score range 0-44 Adherence categorised as: low: 0-27 Elements assessed 11 elements: 1. Vegetables 2. Legumes/nuts/seeds 3. Fruit 4. Wholegrains 5. Olive oil 6. Fish/seafood 7. Poultry 8. Dairy products 9. Eggs 10. Sweets 11. Red/processed meat Cut-off points used & Absolute scoring 1-4 points for each element if met criteria See Appendix A in paper for scoring Score derivation source FFQ for index Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method Validity results against N/A dietary reference method* Outcome associated with Lipids & body weight in free-living French Canadian tool use women (aged 30-65 years) from an uncontrolled Mediterranean diet intervention Comments 1. Scoring based on Oldways Mediterranean pyramid 2. Max. 1 point awarded for refined grain products within scoring for wholegrains 3. Olive oil, margarine from olive oil & olives grouped together; max. 1 point awarded for use of canola oil or margarine made with canola within scoring for olive oil

472 Author, year, country (Panagiotakos, Pitsavos, & Stefanadis, 2006), Greece Index tool name MedDietScore (Mediterranean Diet Score) a priori or posteriori a priori Score range 0-55 Adherence categorised as: KLJK Elements assessed 11 elements: 1. Non-refined cereals 2. Potatoes 3. Fruits 4. Vegetables 5. Legumes 6. Fish 7. Red meat & products 8. Poultry 9. Full fat dairy products 10. Olive oil in cooking 11. Alcohol Cut-off points used & Absolute scoring 1-5 points awarded for each element if met criteria Score derivation source 190-item semi-quantitative FFQ for EPIC-Greece for index (validated) Validated against dietary Y reference method Y/N Dietary reference Food groups & nutrients derived from same FFQ method used Statistics for dietary One-way ANOVA reference method Validity results against 1. Highest tertile of MedDietScore associated with dietary reference increased intakes of fruits (p=0.03), vegetables method* (p=0.01), potatoes (p=0.04), non-refined cereals (p=0.02), fish (p=0.01), legumes (p=0.001) & olive oil (p=0.01); lowest tertile associated with higher intakes of red meat & products (p=0.03), poultry (p=0.03) & full fat dairy (p=0.04) 2. Significant positive association of MedDietScore with MUFA (p=0.001) & MUFA:SFA (p=0.02) Outcome associated with 1. CV risk factors in Greek adults (>18 years old) from tool use the prospective ATTICA study & CHD incidence from the CARDIO2000 case-control study 2. Validity within the same population Comments 1. Tool based on rationale of Mediterranean dietary pattern by Supreme Scientific Health Council, Greece 2. Dietary reference method limited by similar recall bias

473 Author, year, country (Panagiotakos et al., 2009), Greece Index tool name MedDietScore variant (by original authors) a priori or posteriori a priori Score range 0-130 Elements assessed 11 elements: 1. Non-refined cereals 2. Potatoes 3. Fruits 4. Vegetables 5. Legumes 6. Fish 7. Red meat & products 8. Poultry 9. Full fat dairy products 10. Olive oil in cooking 11. Alcohol Cut-off points used & Absolute scoring 1-5 points awarded for each element if met criteria Weighting then applied by multiplying scores for daily intake by 3, weekly intake by 2 & monthly intake by 1 See Table 1 in paper for scoring Score derivation source 190-item semi-quantitative FFQ for EPIC-Greece for index (validated) Validated against dietary Y reference method Y/N Dietary reference Plasma fatty acids method used Statistics for dietary 1. Spearman's correlation coefficient reference method 2. Linear regression models Validity results against 1. Correlation coefficients between MedDietScore dietary reference YDULDQW SODVPDELRPDUNHUV08)$ ȡ method* S 08)$6)$ ȡ S '+$ ȡ S Q-IDWW\DFLGV ȡ S PUFA (n-3 + n-IDWW\DFLGV ȡ íS 6)$ ȡ íS WRWDOQ-IDWW\DFLGV ȡ í p<0.001) 2. MedDietScore variant was in full agreement with SUHYLRXV0HG'LHW6FRUH ȡ S 3. Age-adjusted & sex-adjusted regression models confirmed previous unadjusted results Outcome associated with Validity in healthy Greek adults (mean age: 44 years in tool use men & 40 years in women) from the prospective ATTICA study Comments 1. Updated MedDietScore with weighting applied 2. First study to evaluate validity of diet index against plasma fatty acids as dietary biomarkers

474 Author, year, country (Rumawas et al., 2009), USA Index tool name MSDPS (Mediterranean-style Dietary Pattern Score) a priori or posteriori a priori Score range 0-100 Elements assessed 13 elements: 1. Wholegrains 2. Fruits 3. Vegetables 4. Dairy 5. Wine 6. Fish/seafood 7. Poultry 8. Olives, legumes & nuts 9. Potatoes & other starchy foods 10. Eggs 11. Sweets 12. Meat 13. Olive oil Cut-off points used & Absolute scoring 1-10 points awarded for each element if met criteria Sum of scores then standardised to a 0–100 score by dividing by 130 & multiplying by 100 Standardised score then weighted by proportion of total energy intake from Mediterranean diet pyramid foods See Table 1 in paper for scoring Score derivation source 126-item semi-quantitative FFQ (validated) for index Validated against dietary Y reference method Y/N Dietary reference 24 selected nutrients derived from same FFQ method used Statistics for dietary 1. Spearman's correlation coefficient reference method 2. Linear & logistic regression Validity results against 1. Correlation coefficients for total MSDPS score & dietary reference individual element scores were all positive & significant method* ranging IURPȡ PHDW WRȡ YHJHWDEOHV 2. MSDPS score tertiles were significantly & positively associated with fibre, n-IDWW\DFLGVȕ-carotene, lycopene, folate, vitamins C & E, Ca, Mg & K (p<0.001 for all) & inversely associated with added sugar (p=0.001), GI, saturated fat, trans fat & n-6:n-3 ratio (p<0.001 for all); a significant inverse association found PUFA & positive association found for linolenic acid 3. MSDPS score was unexpectedly positively associated with total energy & inversely associated with MUFA & oleic acid (p<0.001 for all)

475 Outcome associated with Validity in US adults from the Framingham Heart Study tool use Offspring Cohort Comments 1. Scoring based on the Greek Mediterranean diet pyramid from Ministry of Health and Welfare: Supreme Scientific Health Council (1999) 2. Complicated computation for scoring 3. Scoring for olive oil is categorical based on: exclusive use of olive oil (score 10), use of olive oil along with other vegetable oils (score 5), or no olive oil (score 0) 4. Dietary reference method limited by similar recall bias

476 Author, year, country (Naja et al., 2015), Lebanon Index tool name LMD (Traditional Lebanese Mediterranean Diet) a priori or posteriori a priori - elements only Score range 9-27 Elements assessed 9 elements: 1. Fruits 2. Dried fruits 3. Vegetables 4. Legumes 5. Dairy products 6. Burghol 7. Olive oil 8. Eggs 9. Potatoes Cut-off points used & Tertiles scoring 1-3 points awarded for each element if met criteria Score derivation source 61-item FFQ (not validated) for index Validated against dietary Y reference method Y/N Dietary reference 5 selected European Mediterranean diet indexes derived method used from same FFQ Statistics for dietary 1. Spearman's correlation coefficient reference method 2. Cohen's Kappa Validity results against 1. Significant correlation coefficients (p<0.01) for total dietary reference LMD score & scores for other indexes: Greek method* 0HG'LHW6FRUH ȡ 6SDQLVKU0(' ȡ ,WDOLDQ,0, ȡ )UHQFK0HG-'4, ȡ (3,& MDS-2 vDULDQW ȡ 2. Only ,0,KDGFRUUHODWLRQZLWK/0' 3. Highest agreement found between LMD & IMI scores with 53.17% of participants classified in same tertile & 92.38% classified in same or adjoining tertiles (N=0.49, 95% CI: 0.45, 0.52) Outcome associated with Validity in Lebanese adults (aged 20-55 years) from the tool use nation-wide nutrition & non-communicable diseases risk factors cross-sectional survey in Lebanon Comments 1. First Mediterranean diet index developed for Middle Eastern Mediterranean region with unique Lebanese foods, e.g., burghol, dried fruits 2. Alcohol not included for scoring as large Muslim population 3. Choice of elements based on previous factor analyses using same dataset 4. Dietary reference method limited by similar recall bias

477 Author, year, country (Tong, Wareham, Khaw, Imamura, & Forouhi, 2016), UK Index tool name PyrMDS (Pyramid-based Mediterranean Diet Score) a priori or posteriori a priori Score range 0-15 Elements assessed 15 elements: 1. Vegetables 2. Legumes 3. Fruits 4. Nuts 5. Cereals 6. Dairy 7. Fish 8. Red meat 9. Processed meat 10. White meat 11. Eggs 12. Potatoes 13. Sweets 14. Alcohol 15. Olive oil Cut-off points used & Absolute scoring 1 point awarded for each element if met criteria Adjusted to 2000 Cal/d for scoring See Table S2 in paper for scoring Score derivation source 130-item semi-quantitative FFQ (validated) for index Validated against dietary Y reference method Y/N Dietary reference 3 selected Mediterranean diet indexes derived from method used same FFQ: LitMDS (Literature based MDS) mMDS (median based MDS) tMDS (tertile based MDS) Statistics for dietary Spearman's correlation coefficient reference method Validity results against 1. Total score for PyrMDS correlated with total scores dietary reference IRURWKHULQGH[HVDW%/P0'6 ȡ W0'6 method* ȡ /LW0'6 ȡ SYDOXH1$ 2. Correlation coefficient for PyrMDS total score at BL (recruited 1993-19 )8 WR ZDVȡ (p value N/A) Outcome associated with 1. CVD incidence & mortality in UK adults (aged 40-79 tool use years) from the EPIC-Norfolk study 2. Validity within the same population Comments 1. Based on Mediterranean Food Guide Pyramid developed by Mediterranean Diet Foundation

478 2. Dietary reference method limited by similar recall bias

479 Author, year, country (Cerwinske, Rasmussen, Lipson, Volgman, & Tangney, 2017), USA Index tool name MEPA (Mediterranean Eating Pattern for Americans) a priori or posteriori a priori Score range 0-16 Elements assessed 16 elements: 1. Olive oil 2. Green leafy vegetables 3. Other vegetables 4. Berries 5. Other fruit 6. Meat 7. Fish 8. Chicken 9. Cheese 10. Butter/cream 11. Beans 12. Wholegrains 13. Pastries 14. Nuts 15. Fast food 16. Alcohol Cut-off points used & Absolute scoring 1 point awarded for each element if met criteria See Appendix A in paper for scoring Score derivation source MEPA tool for index Validated against dietary Y reference method Y/N Dietary reference 156-item VioScreen FFQ (validated) method used Statistics for dietary 1. Spearman's correlation reference method 2. Cohen's Kappa 3. Bland-Altman plot 4. Kruskal-Wallis & Bonferroni-adjusted Mann- Whitney post-hoc comparisons

480 Validity results against 1. Total MEPA score from tool correlated moderately dietary reference ZLWKVFRUHIURP))4 ȡ S method* 2. Percent agreement ranged from fair to moderate-to- good, depending on element; highest was for alcohol (85.7%, p<0.001), fish (84.3%, p<0.001), olive oil (82.9%, p=0.006) & wholegrains (81.4%, p=1.00) & lowest for fast food (45.7%, p=1.00) 3. Agreement for total MEPA score from tool with score from FFQ was fair (mean N=0.27); agreement was poor for other vegetables (N=0.19), meat (N=0.12), chicken (N=0.13), wholegrains (N í SDVWULHV N=0.19) & fast food (N í DJUeement was fair for olive oil (N=0.33), green leafy vegetables (N=0.36), other fruit (N=0.25), cheese (N=0.21), butter (N=0.21), & beans (N=0.29); agreement was moderate to good for berries (N=0.47), nuts (N=0.42), fish (N=0.62) & alcohol (N=0.64) 4. MEPA tool over-estimated score relative to FFQ 5. Nutrients derived from FFQ were in expected direction with MEPA tool score: K (p=0.001), Mg (p<0.0001), vitamin C (p=0.002), saturated fat (p=0.005), selected carotenoids (p=0.007 to <0.0001), folate (p=0.009) & fibre (p<0.0001) Outcome associated with Validity in US women (aged 24-79 years) from a cross- tool use sectional study Comments 1. Modified version of MEDAS screener for use within US population 2. MEPA incorporates selected elements protective for brain health 3. Dietary reference method limited by similar recall bias

481 Author, year, country (Sofi, Dinu, Pagliai, Marcucci, & Casini, 2017), Italy Index tool name MEDI-LITE (Mediterranean Diet Based on the Literature) a priori or posteriori a priori - elements only Score range 0-18 Elements assessed 9 elements: 1. Fruit 2. Vegetables 3. Cereal grains 4. Legumes 5. Fish & fish products 6. Meat & meat products 7. Dairy products 8. Alcohol 9. Olive oil Cut-off points used & Tertiles scoring For typical Mediterranean foods: 2 points awarded for highest tertile of intake, 1 for middle & 0 for lowest; vice versa for non-Mediterranean foods For alcohol: 2 points awarded for 1-2 std units/d, 1 point for 1 unit/d & 0 for highest category of intake For olive oil: 2 points awarded for regular use, 1 point for frequent use & 0 for occasional use Score derivation source MEDI-LITE tool for index Validated against dietary Y reference method Y/N Dietary reference MedDietScore method used Statistics for dietary 1. Pearson’s correlation coefficient reference method 2. Receiver operating characteristic (ROC) curve methodology for correct classification of adherence to Mediterranean diet Validity results against 1. Good correlation of MEDI-LITE with MedDietScore dietary reference for total score (r=0.70, p<0.0001) & most of 9 elements method* (r=0.56-0.90, p<0.05, except for alcohol where r=0.27) 2. MEDI-LITE had significant discriminative capacity of 85% for adherents vs. non-adherents to Mediterranean diet (AUC: 0.851, 95% CI 0.799, 0.904, p<0.0001) Outcome associated with Validity in healthy Italian adults (aged 18-80 years) tool use from a cross-sectional study Comments 1. Elements for tool originally obtained by posteriori analyses but scoring adapted according to method used by MedDietScore 2. Dietary reference method limited by similar recall bias

482 Lifestyle index

Author, year, country (Sotos-Prieto et al., 2015), Spain Index tool name MEDLIFE (MEDiterranean LIFEstyle) a priori or posteriori a priori Score range 0-28 Elements assessed 28 elements: • 15 food frequencies: 1. Pastries 2. Red meat 3. Processed meat 4. Eggs 5. Legumes 6. White meat 7. Fish/seafood 8. Potatoes 9. Low fat dairy products 10. Nuts & olives 11. Herbs & spices 12. Fruit 13. Vegetables 14. Olive oil 15. Grain products • 7 traditional dietary habits: 16. Water or tea 17. Wine 18. Added salt 19. Wholegrain preference 20. Snacking frequency 21. Snacking between meals 22. Sugary drinks • 6 lifestyle habits: 23. Physical activity 24. Siesta/nap 25. Sleep 26. TV 27. Social interaction 28. Team sports Cut-off points used & Absolute scoring 1 point for each element if met criteria See Table 1 in paper for scoring Score derivation source MEDLIFE tool for index Validated against dietary Y reference method Y/N Dietary reference 142-item FFQ (validated) method used

483 Statistics for dietary 1. Pearson's correlation coefficient reference method 2. ICC 3. Cohen's Kappa 4. Bland-Altman plot Validity results against 1. Moderate to good correlation for total MEDLIFE dietary reference score from tool & FFQ (r=0.626, p<0.05; ICC=0.544, method* 95% CI 0.3, 0.7) 2. Absolute agreement between tool & FFQ for nearly 60% of elements ranged from very good to moderate (N=0.41-1, p<0.0001) 3. Bland-$OWPDQ/2$UDQJHGIURPíWR (mean=1.40) Outcome associated with Validity in middle-aged Spanish adults (mean age 41.4 tool use years) from a cross-sectional study Comments 1. Based on Mediterranean Food Guide Pyramid developed by Mediterranean Diet Foundation 2. First tool to also assess Mediterranean lifestyle habits 3. Dietary reference method limited by similar recall bias

484 Life stage indexes

Author, year, country (Serra-Majem et al., 2004), Spain Index tool name KIDMED (Mediterranean Diet Quality Index) a priori or posteriori a priori Score range 0-12 Adherence categorised as: RSWLPDO medium:4–7 YHU\ORZ Elements assessed 16 elements: • Positive: 1. Fruit or FJ daily 2. Second fruit daily 3. Fresh or ckd veg once/d 4. Fresh or ckd veg >1/d 5. Fish 2-3x/wk 6. Legumes >1x/wk 7. Pasta or rice t5x/wk 8. Cereals or bread for BF 9. Nuts t2-3x/wk 10. Olive oil used at home 11. Has a dairy product for BF 12. Has 2 yoghurts &/or cheese daily • Negative: 13. Fast food >1x/wk 14. Skips BF 15. Has pastries for BF 16. Sweets or lollies several times daily Cut-off points used & Absolute scoring 1 point awarded for each positive element & 1 point deducted for each negative element, if met criteria Score derivation source KIDMED tool for index Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method Validity results against N/A dietary reference method* Outcome associated with Diet quality in Spanish children & adolescents (2 to 24 tool use years of age) from the EnKid study

485 Comments 1. First tool to assess Mediterranean diet adherence in children & adolescents Author, year, country (Mariscal-Arcas et al., 2009), Spain Index tool name MDS-2 variant: MDS-P (Mediterranean Diet Score for Pregnancy) a priori or posteriori a priori - elements only Score range 0-11 Adherence categorised as: SRRU adequate:5–8 KLJK Elements assessed 11 elements: 1. Vegetables 2. Fruit & nuts 3. Legumes 4. Cereals 5. Fish 6. MUFA:SFA 7. Meat 8. Dairy 9. Iron 10. Calcium 11. Folic acid Cut-off points used & Median scoring 1 point for each positive element above & each negative element below median for sex 3 points added if intake of 3 PLFURQXWULHQWV5',RU if receiving medically prescribed nutrient supplement Score derivation source FFQ for pregnant women (validated) for index Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method Validity results against N/A dietary reference method* Outcome associated with Diet quality in pregnant Spanish women based on a tool use Mediterranean-type diet, with micronutrient consideration, from a prospective study Comments 1. Alcohol group removed as avoided by pregnant women

486 Meal index

Author, year, country (Monteagudo et al., 2013), Spain Index tool name BQI (Breakfast Quality Index) a priori or posteriori a priori Score range 1-10 Adherence categorised as: SRRU medium: 4-7 DGHTXDWH Elements assessed 10 elements: 1. Cereals & derivatives (bread, biscuits, bakery products) 2. Fruit & vegetables (includes FJ) 3. Dairy products (FC & semi-skim milk, shakes) 4. Foods rich in sugars (jam, honey) <5% E 5. MUFA rich fats (olive & vegetable oil) 6. MUFA:SFA > median 7. Compliance with E intake recommendations (20-25% E) 8. Cereals + fruit + dairy in same meal 9. Calcium (200-300 mg) 10. Absence of SFA & trans rich fats (butter, margarine) Cut-off points used & Absolute scoring 1 point awarded for each element if met criteria See Table 1 in paper for scoring Score derivation source FFQ (validated) for index Validated against dietary N reference method Y/N Dietary reference N/A method used Statistics for dietary N/A reference method Validity results against N/A dietary reference method* Outcome associated with Breakfast quality in Spanish (from Granada & Balearic tool use Islands) children & adolescents (aged 8-17 years) from a cross-sectional study Comments 1. Novel tool to evaluate quality of breakfast within a Mediterranean context

&, and; ANOVA, analysis of variance; BF, breakfast; BL, baseline; BRCA 1 & BRCA 2 genetic mutations, mutations of two tumour suppressor genes; ckd, cooked; CV,

487 cardiovascular; d, day; DASH, Dietary Approaches to Stop Hypertension; E, energy; EPIC, European Prospective Investigation into Cancer and Nutrition; EPICOR, (long- tErm follow-up of antithrombotic management Patterns In acute CORonary syndrome patients) study; EVOO, extra virgin olive oil; FC, full cream; FFQ, food frequency questionnaire; FJ, fruit juice; FR, food record; HEI, Healthy Eating Index; hr, hour; hsCRP, high sensitivity C-reactive protein; ICC, intra-class correlation coefficient; K, potassium; LOA, limits of agreement; max., maximum; Mg, magnesium; min, minimum; MUFA, monounsaturated fatty acids; N, no; N/A, not available; NHS, Nurses’ Health Study; PA, physical activity; PAD, peripheral artery disease; proc, processed; PREDIMED, Prevencion con Dieta Mediterranea; PUFA, polyunsaturated fatty acids; Q, question; RCT, randomised controlled trial; RDI, Recommended Dietary Intake; SCS, Seven Countries Study; SFA, saturated fatty acids; std, standard; TBSP, tablespoon; vs., versus; wk, week; Y, yes. * Reliability data is also stated where provided.

488 References for Appendix Five

Abu-Saad, K., Endevelt, R., Goldsmith, R., Shimony, T., Nitsan, L., Shahar, D. R., . . . Kalter-Leibovici, O. (2019). Adaptation and predictive utility of a Mediterranean diet screener score. Clinical Nutrition, Advance online publication. doi:10.1016/j.clnu.2018.12.034 Agnoli, C., Krogh, V., Grioni, S., Sieri, S., Palli, D., Masala, G., . . . Panico, S. (2011). A priori-defined dietary patterns are associated with reduced risk of stroke in a large Italian cohort. Journal of Nutrition, 141(8), 1552-1558. doi:10.3945/jn.111.140061 Alberti, A., Fruttini, D., & Fidanza, F. (2009). The Mediterranean Adequacy Index: further confirming results of validity. Nutrition, Metabolism and Cardiovascular Diseases, 19(1), 61-66. doi:10.1016/j.numecd.2007.11.008 Cerwinske, L. A., Rasmussen, H. E., Lipson, S., Volgman, A. S., & Tangney, C. C. (2017). Evaluation of a dietary screener: The Mediterranean Eating Pattern for Americans tool. Journal of Human Nutrition and Dietetics, 30(5), 596-603. doi:10.1111/jhn.12451 Ciccarone, E., Di Castelnuovo, A., Salcuni, M., Siani, A., Giacco, A., Donati, M. B., . . . Gendiabe, I. (2003). A high-score Mediterranean dietary pattern is associated with a reduced risk of peripheral arterial disease in Italian patients with Type 2 diabetes. Journal of Thrombosis and Haemostasis, 1(8), 1744-1752. doi:10.1046/j.1538-7836.2003.00323.x Davidson, S., Passmore, R., Brock, J. F., & Truswell, A. S. (1979). Human nutrition and dietetics. London, New York: Churchill Livingstone. Fung, T. T., Hu, F. B., McCullough, M. L., & Newby, P. K. (2006). Diet quality is associated with the risk of estrogen receptor-negative breast cancer in postmenopausal women. The Journal of Nutrition, 136(2), 466. doi:10.1093/jn/136.2.466 Galilea-Zabalza, I., Buil-Cosiales, P., Salas-Salvadó, J., Toledo, E., Ortega-Azorín, C., Díez-Espino, J., . . . for the PREDIMED-Plus Study Investigators. (2018). Mediterranean diet and quality of life: Baseline cross-sectional analysis of the PREDIMED-PLUS trial. PLoS ONE, 13(6), e0198974. doi:10.1371/journal.pone.0198974

489 Goulet, J., Nadeau, G., & Lemieux, S. (2003). Effect of a nutritional intervention promoting the Mediterranean food pattern on plasma lipids, lipoproteins and body weight in healthy French-Canadian women. Atherosclerosis, 170(1), 115- 124. doi:10.1016/s0021-9150(03)00243-0 Hebestreit, K., Yahiaoui-Doktor, M., Engel, C., Vetter, W., Siniatchkin, M., Erickson, N., . . . Bischoff, S. C. (2017). Validation of the German version of the Mediterranean Diet Adherence Screener (MEDAS) questionnaire. BMC Cancer, 17, 341. doi:10.1186/s12885-017-3337-y Hu, F. B., Bronner, L., Willett, W. C., Stampfer, M. J., Rexrode, K. M., Albert, C. M., . . . Manson, J. E. (2002). Fish and omega-3 fatty acid intake and risk of coronary heart disease in women. JAMA, 287(14), 1815-1821. doi:10.1001/jama.287.14.1815 Mariscal-Arcas, M., Rivas, A., Monteagudo, C., Granada, A., Cerrillo, I., & Olea- Serrano, F. (2009). Proposal of a Mediterranean diet index for pregnant women. British Journal of Nutrition, 102(5), 744-749. doi:10.1017/S0007114509274769 Ministry of Health and Welfare: Supreme Scientific Health Council. (1999). Dietary guidelines for adults in Greece. Archives of Hellenic Medicine. Retrieved 19th December 2018 from http://www.mednet.gr/archives/1999-5/pdf/516.pdf Monteagudo, C., Palacín-Arce, A., Bibiloni, M. D. M., Pons, A., Tur, J. A., Olea- Serrano, F., & Mariscal-Arcas, M. (2013). Proposal for a Breakfast Quality Index (BQI) for children and adolescents. Public Health Nutrition, 16(4), 639- 644. doi:10.1017/S1368980012003175 Naja, F., Hwalla, N., Itani, L., Baalbaki, S., Sibai, A., & Nasreddine, L. (2015). A novel Mediterranean diet index from Lebanon: Comparison with Europe. European Journal of Nutrition, 54(8), 1229-1243. doi:10.1007/s00394-014-0801-1 Panagiotakos, D., Kalogeropoulos, N., Pitsavos, C., Roussinou, G., Palliou, K., Chrysohoou, C., & Stefanadis, C. (2009). Validation of the MedDietScore via the determination of plasma fatty acids. International Journal of Food Sciences and Nutrition, 60(Suppl5), 168-180. doi:10.1080/09637480902810338

490 Panagiotakos, D. B., Pitsavos, C., & Stefanadis, C. (2006). Dietary patterns: A Mediterranean diet score and its relation to clinical and biological markers of cardiovascular disease risk. Nutrition, Metabolism and Cardiovascular Diseases, 16(8), 559-568. doi:10.1016/j.numecd.2005.08.006 Papadaki, A., Johnson, L., Toumpakari, Z., England, C., Rai, M., Toms, S., . . . Feder, G. (2018). Validation of the English version of the 14-item Mediterranean Diet Adherence Screener of the PREDIMED study, in people at high cardiovascular risk in the UK. Nutrients, 10(2), 138. doi:10.3390/nu10020138 Rumawas, M. E., Dwyer, J. T., Mckeown, N. M., Meigs, J. B., Rogers, G., & Jacques, P. F. (2009). The development of the Mediterranean-style dietary pattern score and its application to the American diet in the Framingham Offspring Cohort. The Journal of Nutrition, 139(6), 1150-1156. doi:10.3945/jn.108.103424 Schroder, H., Fito, M., Estruch, R., Martinez-Gonzalez, M. A., Corella, D., Salas- Salvado, J., . . . Covas, M. I. (2011). A short screener is valid for assessing Mediterranean diet adherence among older Spanish men and women. Journal of Nutrition, 141(6), 1140-1145. doi:10.3945/jn.110.135566 Serra-Majem, L., Ribas, L., Ngo, J., Ortega, R. M., García, A., Pérez-Rodrigo, C., & Aranceta, J. (2004). Food, youth and the Mediterranean diet in Spain. Development of KIDMED, Mediterranean Diet Quality Index in children and adolescents. Public Health Nutrition, 7(7), 931-935. doi:10.1079/PHN2004556 Sofi, F., Dinu, M., Pagliai, G., Marcucci, R., & Casini, A. (2017). Validation of a literature-based adherence score to Mediterranean diet: The MEDI-LITE score. International Journal of Food Sciences and Nutrition, 68(6), 757-762. doi:10.1080/09637486.2017.1287884 Sotos-Prieto, M., Santos-Beneit, G., Bodega, P., Pocock, S., Mattei, J., & Penalvo, J. L. (2015). Validation of a questionnaire to measure overall Mediterranean lifestyle habits for research application: The MEDiterranean LIFEstyle index (MEDLIFE). Nutricion Hospitalaria, 32(3), 1153-1163. doi:10.3305/nh.2015.32.3.9387

491 Tognon, G., Nilsson, L. M., Lissner, L., Johansson, I., Hallmans, G., Lindahl, B., . . . Yrkes- och, m. (2012). The Mediterranean diet score and mortality are inversely associated in adults living in the subarctic region. The Journal of Nutrition, 142(8), 1547-1553. doi:10.3945/jn.112.160499 Tong, T. Y. N., Wareham, N. J., Khaw, K.-T., Imamura, F., & Forouhi, N. G. (2016). Prospective association of the Mediterranean diet with cardiovascular disease incidence and mortality and its population impact in a non-Mediterranean population: the EPIC-Norfolk study. BMC Medicine, 14(1), 135. doi:10.1186/s12916-016-0677-4 Trichopoulou, A., Kouris-Blazos, A., Wahlqvist, M. L., Gnardellis, C., Lagiou, P., Polychronopoulos, E., . . . Trichopoulos, D. (1995). Diet and overall survival in elderly people. British Medical Journal, 311(7018), 1457-1460. doi:10.1136/bmj.311.7018.1457 Trichopoulou, A., Costacou, T., Bamia, C., & Trichopoulos, D. (2003). Adherence to a Mediterranean diet and survival in a Greek population. New England Journal of Medicine, 348(26), 2599-2608. doi:10.1056/nejmoa025039 Trichopoulou, A., Orfanos, P., Norat, T., Bueno-de-Mesquita, B., Ocké, M. C., Peeters, P. H. M., . . . Boffetta, P. (2005). Modified Mediterranean diet and survival: EPIC-elderly prospective cohort study. BMJ (Clinical Research Edition), 330(7498), 991-995. doi:10.1136/bmj.38415.644155.8f Vitale, M., Racca, E., Izzo, A., Giacco, A., Parente, E., Riccardi, G., & Giacco, R. (2019). Adherence to the traditional Mediterranean diet in a population of South of Italy: Factors involved and proposal of an educational field-based survey tool. International Journal of Food Sciences and Nutrition, 70(2), 197-203. doi:10.1080/09637486.2018.1481202

492 APPENDIX SIX

Mediterranean Diet Adherence Screener (MEDAS)

493 MEDAS question (Schroder et al., 2011) Criteria for 1 point*

1. Do you use olive oil as the principal source of fat for Yes cooking?

2. How much olive oil do you consume per day (including 7%63 that used in frying, salads, meals eaten away from home, etc.)?

3. How many servings of vegetables do you consume per (at least 1 portion day? Count garnish and side servings as 1/2 point; a full raw or as salad) serving is 200 g.

4. How many pieces of fruit (including fresh-squeezed juice) do you consume per day?

5. How many servings of red meat, hamburger, or sausages <1 do you consume per day? A full serving is 100–150 g.

6. How many servings (12 g) of butter, margarine, or cream <1 do you consume per day?

7. How many carbonated and/or sugar-sweetened beverages <1 do you consume per day?

8. Do you drink wine? How much do you consume per week? FXSV

9. How many servings (150 g) of pulses do you consume per week?

10. How many servings of fish/seafood do you consume per week? (100–150 g of fish, 4–5 pieces or 200 g of seafood)

11. How many times do you consume commercial (not <3 (Martínez- homemade) pastry such as cookies or cake per week? González et al., 2012)

12. How many times do you consume nuts per week? (1 serving = 30 g)

494 13. Do you prefer to eat chicken, turkey or rabbit instead of Yes beef, pork, hamburgers, or sausages?

14. How many times per week do you consume boiled vegetables, pasta, rice, or other dishes with a sauce of tomato, garlic, onion, or leeks sautéed in olive oil?

TBSP, tablespoon; g, grams.

* zero points are awarded if criteria is not met.

495 References for Appendix Six

Martínez-González, M. A., García-Arellano, A., Toledo, E., Salas-Salvado, J., Buil- Cosiales, P., Corella, D., . . . Gómez-Gracia, E. (2012). A 14-item Mediterranean diet assessment tool and obesity indexes among high-risk subjects: The PREDIMED trial. PLoS ONE, 7(8), e43134. doi:10.1371/journal.pone.0043134

Schroder, H., Fito, M., Estruch, R., Martinez-Gonzalez, M. A., Corella, D., Salas- Salvado, J., . . . Covas, M. I. (2011). A short screener is valid for assessing Mediterranean diet adherence among older Spanish men and women. Journal of Nutrition, 141(6), 1140-1145. doi:10.3945/jn.110.135566

496 APPENDIX SEVEN

Screen shots of sample MediCul questions formatted

for online administration

497 498 499 APPENDIX EIGHT

Ethics letters

500 Human Research Ethics Committee (HREC) The University of New South Wales UNSW Sydney, NSW, Australia, 2052 T: +612 9385 622 or +612 9385 7257 E: [email protected] W:https://research.unsw.edu.au/human-research-ethics-home

21-Apr-2016

Dear Scientia Professor Henry Brodaty,

Project Title Maintain Your Brain (MYB) HC No HC16252 Re Notification of Ethics Approval Approval Period 21-Apr-2016 - 20-Apr-2021

Thank you for submitting the above research project to the HREC Committee A for ethical review. This project was considered by the HREC Committee A at its meeting on 19-Apr-2016.

I am pleased to advise you that the HREC Committee A has granted ethical approval of this research project, subject to the following conditions being met:

Evidence of the trial being registered on the Australian New Zealand Clinical Trials Registry.

The provision of a withdrawal of consent form following the participant information sheet.

'>Conditions of Approval Specific to Project:

Conditions of Approval - All Projects:

The Chief Investigator will immediately report anything that might warrant review of ethical approval of the project. The Chief Investigator will notify the HREC Committee A of any event that requires a modification to the protocol or other project documents and submit any required amendments in accordance with the instructions provided by the HREC Committee A. These instructions can be found at https://research.unsw.edu.au/research-ethics-and-compliance-support-recs. The Chief Investigator will submit any necessary reports related to the safety of research participants in accordance with HREC Committee A policy and procedures. These instructions can be found at https://research.unsw.edu.au/research-ethics-and-compliance-support-recs. The Chief Investigator will report to the HREC Committee A annually in the specified format and notify the HREC Committee A when the project is completed at all sites. The Chief Investigator will notify the HREC Committee A if the project is discontinued at a participating site before the expected completion date, with reasons provided. The Chief Investigator will notify the HREC Committee A of any plan to extend the duration of the project

501 past the approval period listed above and will submit any associated required documentation. Iinstructions for obtaining an extension of approval can be found at https://research.unsw.edu.au /research-ethics-and-compliance-support-recs. The Chief Investigator will notify the HREC Committee A of his or her inability to continue as Coordinating Chief Investigator including the name of and contact information for a replacement.

A copy of this ethical approval letter must be submitted to all Investigators and sites prior to commencing the project.

The HREC Committee A Terms of Reference, Standard Operating Procedures, membership and standard forms are available from https://research.unsw.edu.au/research-ethics-and-compliance-support-recs.

If you would like any assistance, or further information, please contact the ethics office on: P: +61 2 9385 6222, + 61 2 9385 7257 or + 61 2 9385 7007 E: [email protected]

The HREC Committee A wishes you continued success in your research.

Kind Regards

HREC Presiding Chairperson

This HREC is constituted and operates in accordance with the National Health and Medical Research Council’s (NHMRC) National Statement on Ethical Conduct in Human Research (2007). The processes used by this HREC to review multi-centre research proposals have been certified by the National Health and Medical Research Council.

502 503 APPENDIX NINE

Research Food Diary app instructions

504 Instructions on How to Keep a 3 Day Food Diary

Thank you for agreeing to keep a 3 day electronic food diary of your food intake. The instructions below will give you all the information you need to do this easily. They are organised in four key sections.

Please carefully read through the instruction booklet, or refer to the individual sections below. If you are still unsure about any part of this study phase, please get in touch with us via [email protected] .

Page

x Checklist Before You Start……………………………………………………….………………………….2

x How to Download ‘Research Food Diary’ and Set Up Your Diary Account …………3

- Special instructions for iPad users …………………………………………………..……………..6

x How to Use ‘Research Food Diary’ …………………………………………………..………………...9

- Adding food to your diary……………………………………………………………………………….9 - Scanning a product barcode to add a food to your diary……………………………….10 - Creating a recipe to add home made foods into your diary…………………………..12 - Editing a recipe……………………………………………………………………………………………..16 - Deleting a recipe..………………………………………………………...………………………………17 - Deleting a food item………………………………………………..……………………………………17 - Substituting a food item if its not in the Research Food Diary database………..18

x How to Submit Your 3 Day Food Diary…...…………………………………………………………19

505 Checklist Before You Start & General Tips About Logging Food Intake

1. Please record EVERYTHING you eat or drink – meals, snacks, hot drinks, alcohol and water - over three days, including one weekend day (or non-working day) and two week days.

The days do not need to be consecutive but must fall within a one week period.

2. Log foods and drinks straight away as you consume them. You will need to carry your device with you to do this using an app called ‘Research Food Diary’.

3. Don’t change your diet during this period – we are interested in your usual intake.

4. Select the options most relevant to you from within the app, including brands, formulations (e.g. full cream vs. low fat, flavoured vs. plain, added sugar vs. ‘diet’, light vs. full strength) and your serve sizes.

5. Log your serve sizes using ‘weight’ in grams if you know this. Otherwise, estimate your serve sizes using household measures such as teaspoons, tablespoons or cups. It is important to measure your foods using a measure instead of guessing portions for accuracy.

6. Ensure you record all the extras (e.g. dressing for salad; sugar in tea; milk in coffee; margarine on bread; oil, salt or herbs used in cooking).

7. For meals made with a recipe (e.g. beef and vegetable casserole), see below for how to record the recipe in the app first and then select the number of serves you personally consumed.

506 How to Download ‘Research Food Diary’ and Set Up Your Diary Account

Research Food Diary is a free iPhone app for tracking and recording your diet. This app is compatible with devices such as iPhone, iPad and iPod touch using iOS 8.0 or later.

You will need to have an iTunes account and be connected to the internet (e.g. through a WiFi network or your mobile 3G/4G network) to download the app. If you do not have an iTunes account, you will need to set one up first.

Information on how to set up an iTunes account can be found here: https://support.apple.com/en- au/HT204034

For iPhone

1. To download this app onto your iPhone, select the App Store from your home screen.

507 2. Tap on the “Search” icon and type “Research Food Diary” into the search field.

3. Select Research Food Diary from the search results. Tap the icon. An icon will then appear, tap this to download the Research Food Diary app.

508 4. You may see a screen asking you to enter your iTunes password. Enter this to allow the download to proceed. The download may take several minutes depending on the speed of your internet connection.

5. Once the download is completed the icon should appear on the main screen of your iOS device(s). Note – if you have many apps on your device it may appear on another screen so you will have to swipe between screens to find it. Tap on this icon to open and begin using the Research Food Diary app.

6. Create an account using Sign Up with Email.

7. To start adding food to your diary within the app, please see instructions below on: How to use Research Food Diary.

509 For iPad and iPod Touch

1. To download this app onto your iPad or iPod Touch, select the App Store on your device.

2. Tap on the “Search” box on the top right corner and type “Research Food Diary” into the search field. Then tap on the button.

510 3. You will be directed to a page which says ’no results for “Research Food Diary”’. Tap on ‘iPad Only’ on the top left corner of your device. This will open up a dropdown menu. Tap on ‘iPhone Only’ to display ”Research Food Diary” on the search results.

4. Select Research Food Diary from the search results. Tap the icon. An icon will then appear, tap this to download the Research Food Diary app.

511 5. You may see a screen asking you to enter your iTunes password. Enter this to allow the download to proceed. The download may take several minutes depending on the speed of your internet connection.

6. Once download is completed the icon should appear on the main screen of your iPad. Note – if you have many apps on your device it may appear on another screen so you will have to swipe between screens to find it. Tap on this icon to open and begin using the Research Food Diary app.

8. Create an account using Sign Up with Email.

7. To start adding food to your diary within the app, please see instructions below on: How to use ‘Research Food Diary’.

512 How to Use ‘Research Food Diary’

Adding food to your diary

There are three ways you can add food to your diary. First, by searching the foods database. Second, by scanning the barcode of food packaging. Third, by adding a recipe of a mixed dish to your diary and then selecting the number of serves you personally consumed from that recipe. Instructions for each method are detailed below.

To log a food by searching from the foods database:

1. On the “Diary” screen tap a meal, (e.g. “Breakfast”), to open the meal. 2. Tap magnifying glass icon at the bottom left under the meal.

3. In the “Find a food…” search field, enter food name then tap on Search. 4. Select the most appropriate item from the list. If multiple options are provided to further narrow down the type of food, please select the most relevant option for you. 5. Tap on the “Amount” box and enter the amount consumed. 6. Choose your serving size from the list (e.g. slice, cups, grams etc). 7. Tap “Add” at the top right of the screen.

513 TIP: If the food has a barcode, instead of searching, try scanning it.

To log a food by scanning its barcode:

1. On the “Diary” screen tap a meal, (e.g. “Breakfast”), to open the meal. 2. Tap the barcode icon below the meal.

*Note that you may need to authorize permission for Research Food Diary to access your camera. Tap “Yes” when prompted.

514 3. View the barcode of the food item through your device's camera.

4. Hold still while the barcode is scanned automatically. 5. Tap the “Amount” box and type the amount. 6. Choose your serving size from the list (e.g. slice, cups, grams etc.) then tap “Add” at the top of the screen. 7. To return to the meal, tap the meal name at the top left of the screen.

515 Creating a recipe

Creating a recipe while logging your foods

1. On the “Diary” screen tap a meal, (e.g. “Breakfast”), to open the meal to which you want to add the recipe. 2. Tap the magnifying glass icon at the bottom left under the meal. 3. Tap “Recipes” on the top right side of the tab. 4. Tap the (+) at the bottom of the recipe list.

5. Tap the “Recipe Name” field and enter the recipe name. 6. Tap “Add ingredients” field to add ingredients. 7. Tap the magnifying glass icon to search for an ingredient, or the barcode icon to scan the barcode of an ingredient 8. Enter the ingredients in the same way that you enter foods into your diary. 9. Tap the “Serves” box below the Ingredients box and specify the number of serves produced by that recipe, then tap on to diary now?”.

516 12. To log in the food item into your diary, tap the “Amount” box and choose your serve size from the list (e.g. slice, cups, grams etc.). 13. Tap “Add” 14. at the top of the screen. 15. Tap the meal name at the top left of the screen to return to your diary.

TIP: x The recipe is automatically added to your recipe list. x To use this recipe in your diary at another time:

1. Open the meal to which you want to add it. 2. Tap the magnifying glass icon and tap the “Recipes” tab on top of your screen. 3. Your recipe will be listed there in alphabetical order. 4. Tap the recipe to add it to your meal.

517 Creating a recipe at any time to add to your recipe list

1. Tap “Recipes” at the bottom of the screen.

2. Tap the (+) in the top right corner of the screen. 3. Type the recipe name in the “Recipe Name” box.

4. In the “Ingredients” box, tap the magnifying glass icon to search for an ingredient, or the barcode icon to scan the barcode of an ingredient. 5. Enter the ingredients in the same way that you enter foods into your diary. 6. Tap the “Serves” box and type the number of serves.

518 7. Tap “Add” on the top right corner. 8. Repeat for all the remaining ingredients for that recipe. 9. Tap “Done” at the top right of the screen.

Converting foods in my diary to a recipe

1. On the “Diary” screen tap a meal, (e.g. “Breakfast”) to open it. 2. Tap the Tick icon on the bottom right corner. 3. Select the foods that are the ingredients of the recipe.

4. Tap “Recipe” at the bottom of the screen. 5. Type a name for the recipe. 6. Edit the ingredients and serves, as required, by tapping on them individually. 7. Tap “Done” at the top of the screen.

519 Editing a recipe

1. Tap “Recipes” at the bottom of the screen. 2. Select the recipe that you would like to edit.

x To modify the recipe name tap on the name for the recipe and change it.

x To add an ingredient:

1. Tap the magnifying glass icon or the barcode icon. 2. Add a food item, just as you would add it to a meal.

x To modify quantities in a recipe:

1. Tap on the name of an ingredient in the recipe. 2. Modify the amount and measurements and tap on top left

x To modify serving yield:

1. Tap the box under the ingredients “Makes … Serve”. 2. Modify the amount and measurements (e.g. grams instead of serves) and tap on top left

x To delete an ingredient

1. Tap the Edit icon. 2. Tap the red circle beside the ingredient, then tap . 3. Tap “Done” at the top right of the screen.

520 Deleting a recipe

1. Tap “More” on the bottom right of your screen. 2. Tap the “Recipes” tab. 3. Swipe left on a recipe name and tap

– OR –

1. Tap the Edit icon at the bottom of the recipe list.

2. Tap the red circle beside the recipe, then tap 3. Tap the Edit icon again to exit Edit mode.

Deleting a food item

1. Tap a meal, (e.g. “Breakfast”) on the Diary screen to open the meal. 2. Tap the Tick icon in the bottom right corner. 3. Select all the ingredients you want to delete. 4. Tap “Delete” at the bottom left of the screen.

521 What to do if you can’t find your meal in Research Food Diary

There may be times when you are unable to find your exact meal in Research Food Diary. For example, when you are eating at a friends place or having a takeaway meal.

To record meals that are not in ‘Research Food Diary’

1. Add the meal as a new recipe. (See instructions above for ‘Creating a recipe’). 2. In “Recipe Name” specify if your meal was homemade or a takeaway (e.g. “Homemade Vegetarian Laksa” or “Takeaway Vegetarian Laksa”). 3. Looking at your meal, estimate the quantity of each ingredient and log this (e.g. noodles 1 cup, bean sprouts ½ cup, vegetable broth 1 ½ cups). 4. Don’t forget to include all the extras such as sauces, herb garnishes, fried onion or whatever you can see in the dish.

522 How to Submit Your 3 Day Food Diary

1. Once you have completed your 3-day food diary, please review each day in your diary first, to ensure you have not missed logging any meals, snacks or drinks.

2. To email your diary to the Dietitian tap on the “More” icon at the bottom right corner of the screen. Then select “Settings” and tap on “Email Diary to Your Dietitian”. Do not click on ‘Connect with Health’ as this won’t send your diary to us.

3. When prompted please select the “Last month”. Please address your diary to [email protected] and type your full name on the subject line as below. Once done, tap on “Send” on the top right corner.

523 APPENDIX TEN

Dietitian assisted food record checklist

524 Dietitian assisted food record checklist

Purpose: to ask about forgotten foods/drinks in 3-day food record; what and how much consumed; brands (where relevant as FoodWorks maps to generic foods except for sodium which may be brand specific) and cooking methods; query and record serves/omissions by probing with open ended questions.

Prior to starting interview, for context, ask participants whether a) any measuring devices used e.g. kitchen scales, b) recorded throughout or end of day, c) typical or atypical days reported, d) electronic or paper record used.

Occasion Context specific prompts Breakfast x Water - lemon x Cold drink – juice (100%/fruit juice drink), smoothie (dairy/alternative) x Hot drink – sweetener, milk/type, coffee style, caffeinated, flavouring x Cereal – RTE/cooked, sugar/honey, nut/seed sprinkles, dairy/alternative, fruit (if canned syrup type/water) x Bread/toast – yellow fat spread, other spreads x Protein (eggs/meat/baked beans) – fat/oil, salt, spices/herbs x Yoghurt – plain/flavoured/diet/fat content/non-dairy x Weekend - bacon/sausages, vegetables, pastries Lunch x Sandwich/wrap – spread, sauce/mayo, filling x Soup – bread/roll, spread x Salad – protein (drainage for canned fish), carbs, nuts/croutons, oil/acid/mayo, salt, herbs x Leftovers – as for dinner x Yoghurt (plain/flavoured/diet/fat content/non-dairy), fruit, cake (icing/dried fruit), biscuits (savoury/sweet) x Cold drink – canned/bottled x Hot drink – sweetener, milk/type, coffee style, caffeinated, flavouring Dinner x Soup – bread/roll, spread x Protein – meat cut/fat trimming, oil/fat, crumbed/battered/cooking method, sauce, salt x Carbs (rice/pasta/grains) – oil/fat, cooking method, sauce, salt x Vegetables/salad – dressing/mayo, butter/oil, salt, acid, nut/seed sprinkles x Bread/rolls – spread

525 x Mixed dishes – check ingredients, cooking method, fats/oils, paste, salt, serves x Dessert – ice cream/custard (flavour), fruit (if canned, syrup type/water) x Cheese (cream/fat content) and biscuits/crackers x Cold drink - soft drink/cordial/energy drinks/juice/milk x Hot drink – sweetener, milk/type, coffee style, caffeinated, flavouring x Alcohol - 150 ml wine, 375 ml beer stubbie/can, nip spirits Snacks x Crisps/hot chips (prompt x Biscuits/muffins/pastries after dinner x Cheese and biscuits/crackers e.g. TV x Lollies/chocolates viewing) x Nuts (salt)/dried fruits/fresh fruit/canned fruit (syrup type/water) Throughout x Water – tap/mineral/flavoured the day x Soft drinks/cordials/energy drinks/juice/milk x Hot drink – sweetener, milk/type, coffee style, caffeinated, flavouring Other x Brands/names of takeaway outlets/restaurants information x Herbs/spices for breakfast/lunch/dinner x Quantity if portions appear too small/large x Check composite dishes for added fats/oils, sugar, salt, cream etc. x If composite dishes are vague and guessed, re-enter using representative recipes with total serves RTE, ready to eat.

526 APPENDIX ELEVEN

Front sheet for validation and reliability testing

of MediCul

527 Name: Gender: M OR F DOB: Age: Height: (Ax B) Weight: (Ax B) Actigraph: Y OR N Address: Phone (H): Mobile: Email: Study ID: Date:

FRONT SHEET: VALIDATION & RELIABILITY OF MEDICUL 1. Are you the person responsible for most of the food shopping in your household? NOR YOR shared If no or shared, who else does the grocery shopping for you? ______

2. Are you the person responsible for most of the cooking in your household? NOR YORshared If no or shared, who else does the cooking for you? ______

3. Are you currently following any type of special diet? N OR Y If yes, please tick. Diabetes Cholesterol lowering Gluten free/Wheat free Allergy/Food Intolerance Weight loss Vegetarian/Vegan Low salt Other ______

4. Do you have any specific nutritional concerns or goals? N OR Y If yes, what are they? ______

528 APPENDIX TWELVE

3-day paper food record template

529 Name:______

Study ID:______

Date:______

Administrator:______

3-day Food Record

Instructions

x Please fill in this food intake record on 3 days, including 2 weekdays and 1 weekend day. x The days in which food intake is recorded do not need to be consecutive, but need to be within the same 1 week period. x You should fill in the details immediately after you eat or drink, for each meal or snack. Do not wait till the end of the day to remember everything. x Detailed directions for how to fill in the food record are at the top of each day. An example of one day is on the following page.

When you have completed the food record please return it to the research dietitian Sue Radd. Or call Michael Inskip on his office number 9351 9138 or mobile 0423 471 510 for alternate instructions on how to return this form.

If you have any questions about how you should complete the food record, please contact one of the study dietitian’s: Prof Vicki Flood on 0412 118 977 or email Sue Radd on [email protected]

530 Instructions for completing the food record x Write down everything that you eat and drink over one day from waking to going to sleep. It is important that you do this straight away after you eat or drink, rather than waiting till the end of the day. This also includes any snacks, water and vitamin and mineral supplements you have. x Use a new line for each food, drink or supplement. x Record the type of eating occasion in the appropriate column. For example, breakfast, lunch, dinner, morning tea, afternoon tea or snack. x Record each food individually. For example, while you may state ‘tuna sandwich’ in the first row for lunch, you should then list each of the individual foods contained in the sandwich in the rows below, together with their amounts. x Include the amounts in household measures or natural portion sizes. For example, 2 slices of bread, ½ cup rice, ¼ cup peas. Please use the set of standard spoon, cup and jug measures we have provided to assist in recording this information. o If you wish to use abbreviations for spoon measures please use the following: 1 teaspoon = 1 tsp 1 tablespoon = 1 TBSP x Record the cooking method used, where applicable. For example, grilled, BBQ, dry fried, deep fried, baked, boiled, steamed, stir fried etc. x Give a detailed description of the food or drink and include brand names where possible. For example, Arnott's Milk Arrowroot® biscuit, Yalla Humus or Cobram Estate Extra Virgin Olive Oil robust flavour. x Don't forget to include any sauces, mayonnaise, dressings or gravies that are used. We are interested to find out about your usual eating patterns, so please keep your food intake as usual. x If you record a day that is not typical, please indicate in the box at the end of each food record day and tell us how it differs from your usual intake. For example, I attended a wedding.

531 Example of one day

Date: 22/03/2015 Day: Friday

Location Time Meal/Eating Foods/Drinks/Water/Supplements Cooking Amount/Size Occasion Method (where EATEN applicable)

Home 7am Breakfast Skim milk (Dairy Farmers) - 1 cup

Weet-bix - 2 Weet-bix

White sugar - 1 tsp

Toast – white bread (Tip Top) - 2 slices

Margarine (Gold’n Canola) - 2 tsp

Strawberry jam (Coles) - 2 tsp

Home 9am Supplement Berocca Performance - 1 tablet

Neighbours 10am Morning tea Coffee, instant - 1 cup place

HeartActive milk (Dairy Farmer’s) - 1 TBSP

Doughnut, home made Deep fried 1

Home Cheese and salad sandwich: White bread (Tip Top) 2 slices Margarine (Meadow-lea) 2 tsp Tasty cheese (Coon) 1 slice 12md Lunch - Lettuce, shredded 1 leaf Tomato, sliced 2 slices Carrot, grated 1 TBSP Beetroot, sliced 1 large slice

Home 1pm Water 1 cup

Home 3pm Snack Tim Tams chocolate biscuits (Arnott’s) - 2

Orange cordial (Home Brand) - 1 cup

At the Club 6:30pm Dinner 2 thin beef sausages Grilled 2 thin, length 10 cm

Potato, peeled Steamed 1 medium

Carrots, peeled Steamed ¼ cup

Green peas, frozen Boiled ¼ cup

Red wine, shiraz (Penfolds) - 300 ml

Vanilla ice cream (Dairy Farmer’s) - 2 small scoops

Chocolate sauce (Cottees) - 2 tablespoons

Home 8pm Snack Banana - 1 medium

Home 9pm Supper HeartActive milk (Dairy Farmer’s) - 1 cup

Was your intake unusual in any way? No Yes ¥

If yes, in what way? __Had dinner at the Club, which I do once per month__

532 DAY 1 FOOD INTAKE RECORD (WEEKDAY)

9 Write down everything that you eat and drink over one day from waking to going to sleep. It is important that you do this straight away after you eat or drink, rather than waiting till the end of the day. This also includes any snacks, water and vitamin and mineral supplements you have. 9 Use a new line for each food, drink or supplement. 9 Record the type of eating occasion in the appropriate column. For example, breakfast, lunch, dinner, morning tea, afternoon tea or snack. 9 Record each food individually. For example, while you may state ‘tuna sandwich’ in the first row for lunch, you should then list each of the individual foods contained in the sandwich in the rows below, together with their amounts. 9 Include the amounts in household measures or natural portion sizes. For example: 2 slices of bread, ½ cup rice, ¼ cup peas. Please use the set of standard spoon, cup and jug measures we have provided. 9 Record the cooking method used, where applicable. For example, grilled, BBQ, dry fried, deep fried, baked, boiled, steamed, stir fried etc. 9 Give a detailed description of the food or drink and include brand names where possible. For example, Arnott's Milk Arrowroot® biscuit, biscuit, Yalla Humus or Cobram Estate Extra Virgin Olive Oil robust flavour. 9 Don't forget to include any sauces, mayonnaise, dressings or gravies that are used. We are interested to find out about your usual eating patterns, so please keep your food intake as usual. 9 If you record a day that is not typical, please indicate in the box at the end of each food record day and tell us how it differs from your usual intake. For example, I attended a wedding.

Date: Day:

Location Time Meal/Eating Foods/Drinks/Water/Supplements Cooking method Amount/Size occasion (where EATEN applicable)

533 Was your intake unusual in any way? No Yes

If yes, in what way? ______

534 DAY 2 FOOD INTAKE RECORD (WEEKDAY)

Date: Day:

Location Time Meal/Eating Foods/Drinks/Water/Supplements Cooking method Amount/Size occasion (where EATEN applicable)

535 Was your intake unusual in any way? No Yes

If yes, in what way? ______

536 DAY 3 FOOD INTAKE RECORD (WEEKEND)

Date: Day:

Location Time Meal/Eating Foods/Drinks/Water/Supplements Cooking method Amount/Size occasion (where EATEN applicable)

537 Was your intake unusual in any way? No Yes

If yes, in what way? ______

538 APPENDIX THIRTEEN

FAQ on completing the 3-day food record

539 Frequently Asked Questions on Completing the 3-day Food Record

Do the days need to be consecutive?

No, they can be any days of your choice, so long as they occur in the same 1 week period.

Should I pick my ‘best’ days to record?

No. It’s important that you pick 3 typical days where you record your usual intake, not what you think might impress the researchers or is easiest to record. Otherwise, your response will not reflect usual intake.

Do I really need to measure everything I eat and drink?

Yes. It is important once you serve or pour out the food/drink you intend you eat, that you first measure the amounts before eating, using the equipment provided. You should then record these measurements on your food record. If you go for second helpings, don’t forget to also measure these.

Measure everything by using the cups, spoons and jug provided. If items are difficult to measure this way, you can use the kitchen scales provided to weigh foods. These are optional.

Record everything immediately after eating your meals/snacks to ensure it is accurate. Don’t wait till the evening to try and remember all the foods and amounts you had during the day.

What is the easiest way to measure various foods and drinks?

We recommend that you first plate up what you are intending to eat (or pour out your drink), just like you would normally do. You can then measure each item in turn and transfer to another plate/cup from which you will consume it. Please note, it is important that you portion out foods and amounts as you would normally eat them. You should not change these while keeping the 3-day food record.

If you are using the scales, it is important to zero the reading once you place your plate on top and before you add any food for weighing.

How do I measure dishes that contain a mixture of foods?

This can be tricky. Do your best to ‘pull apart’ the cooked ingredients, where possible, and measure each in turn. For example, if you were going to eat a beef and vegetable curry, you would first

540

portion out your usual amount onto a plate. You could then pick out the meat pieces and weigh these before transferring onto another plate. This would be followed by measuring each vegetable contained in the curry, in turn, and also transferring this to the other plate ready for eating.

Alternatively, you could provide the recipe for the dish to the dietitian and advise how much of the recipe e.g. serves, you personally consumed at a given meal.

If it is too difficult to measure the individual ingredients within a mixed dish/recipe, don’t worry. Just measure/weigh the total mixed dish you intend to eat and provide a description of what it contained.

What if I feel like I might be recording too much detail?

We love details. It’s best to record everything you know about the food, drink or supplement you may be taking at the time you consume it. Otherwise, we might need to come back and ask you for the details later.

What if I am out and about on one of the days?

We suggest you take the cups and spoons with you and measure things as best you can. You might need to ask for a separate plate to transfer foods once you measure them, before eating.

How do I record my nutritional supplements?

It’s important that you accurately record the type and dose of any supplements you are taking. We suggest you bring the bottles with you when you hand in your food record so the dietitian can check. Alternatively, you can take a photo on your phone of the front and back of each supplement bottle, where the details of the formulation are printed.

What if I run out of space to record my food items?

Just attach an additional paper to the back and clearly mark which day it relates to.

Do I need to return the measuring equipment to The University of Sydney?

The measuring spoons, cups and jug are yours to keep after you have finished keeping the 3-day food record. The kitchen scales, however, need to be returned for future use by other people.

541 APPENDIX FOURTEEN

Challis Bequest Society lunch

542 543 Welcome

Dear friends,

We’d like to welcome you most warmly to the 2018 Challis Bequest Society lunch, a wonderful opportunity to celebrate your generosity and thank you for making the University of Sydney a part of your lives. Bequests, both large and small, play a vital role in shaping the future careers of our academics and students.

Great things are achieved when passion is matched by opportunity. Take, for instance, the story of Dr Angela Crean. When she’d finished her PhD in animal reproduction, Angela struggled to find a long-term position as a scientist, and she was about to walk away from the career she loved. Opportunity came in the form of a scholarship for early-career researchers that came about from a bequest from the late Mabs Melville.

For Angela, winning that scholarship marked the beginning of a spectacular rise that has seen her work acclaimed internationally – in 2016, she won the $25,000 L’Oreal-UNESCO for Women in Science Fellowship, and in 2017, she took out the Young Tall Poppy Science Award in NSW. Her research into reproduction, both animal and human, has the potential to change lives. Without Mabs Melville’s generous bequest, Angela’s passion for scientific research would have gone unfulfilled.

Like Dr Crean, everyone in this room shares a passion for changing lives. But you have all taken the next step by choosing to create opportunity for others. And in so doing, you are leaving a lasting legacy. We thank you most sincerely.

Belinda Hutchinson AM Dr Michael Spence AC Chancellor Vice-Chancellor and Principal 544 Order of proceedings

11.45am Carillon recital Amy Johansen, University Organist and Carillonist 12pm Drinks and canapés Heathcliffe Auchinachie, Sydney Conservatorium of Music

12.35pm Welcome, Master of Ceremonies Alexandra Miller (BAS ’10, MPACS ’11) Associate Director, Planned Giving, Division of Alumni and Development

Acknowledgement of Country Liam Coe, Bachelor of Architecture and Environments, third year

Entertainment Showcase Kirralee Elliott, Sydney Conservatorium of Music Jack Dawson, Sydney Conservatorium of Music

12.50pm Main course

1.30pm Keynote speaker Dr Craig Barker (BA(Hons) ’96 PhD’ 05) Manager, Education and Public Programs, Sydney University Museums

1.45pm Vice-Chancellor and Principal’s address Dr Michael Spence AC (BA ’85 LLB ’87)

2.15pm Dessert, tea and coffee

3pm Event concludes

A hearing loop is available for today’s event. Hearing aid users will need to use their T-switch (telecoil). Please let us know if you are experiencing difficulties.

545 Today’s menu is inspired by the Mediterranean diet Canapés Lamb dolmades, beetroot goats curd (gf) Mushroom baklava, almond cream (vegan, df) Olive crostini, heirloom tomatoes, Persian feta, oregano (v, gf) Grilled haloumi, honey and mint (v, gf) Pita bread, taramasalata, salmon roe Goat and oregano mini pies

Main course (alternate serve) Sous vide lamb rump, ancient grains, preserved lemon dressing, eggplant chutney, moussaka, cherry vine tomatoes (gf) or Roast La Ionica chicken crown, beetroot leaf and fetta pie, green beans, walnuts

Side dishes Greek salad (v, gf) Roast potatoes, olive and lemon (vegan, gf, df) Sonoma sourdough breads, Cobram Estate Olive Oil

Dessert (two mini desserts per person) Mini baklava, walnut crumb Rice pudding, rose water & lemon (gf) Loukoumades, Greek doughnuts, honey syrup

Coffee and tea Crave coffee and variety of T2 teas Chocolate truffles

Beverages Marchand and Burch Cremant de Bourgogne, Burgundy, France Teusner Woodside Adelaide Hills Sauvignon Blanc, Barossa Valley, SA Teusner Joshua Grenache Mataro Shiraz, Barossa Valley, SA v - vegetarian, gf - gluten-free, df - dairy free If you have notified us of dietary requirements, we will cater for these separately, please advise your wait staff. 546 Research shows that a Mediterranean diet can protect against metabolic syndrome, helps prevent and control diabetes, may reverse fatty liver disease, and possibly slash the risk of Alzheimer's disease. The more closely you adhere to this diet, the stronger your protection from these health issues will be. So, what constitutes a Mediterranean diet? Health Sciences PhD researcher Sue Radd (AdvAPD AdvAN) lists 10 ways you can incorporate this prevention-style diet into your daily routine:

1. Olive oil daily Ditch the margarine and vegetable oils. Enjoy extra virgin olive oil on your salads and in cooking. It enhances the absorption of antioxidants into the body and has anti-inflammatory effects to protect your health.

2. Legumes at least twice weekly You should get a good serving of beans, such as chickpeas or lentils, to reduce saturated fat and increase your fibre intake. Legumes protect against multiple chronic diseases and are a great alternative to meat.

3. Meat in very small amounts Research shows that people who eat red meat are at an increased risk of early death from heart disease, stroke or diabetes. Limit your meat intake to small amounts monthly, and where possible, replace meat with legumes.

4. Vegetables every day Include a half cup of tomatoes and a half cup of leafy greens, plus at least two cups of other coloured vegetables. Vegetables provide a myriad of antioxidants and anti-inflammatory phytonutrients for good health.

5. Go wholegrain Intact, cracked or coarsely milled flours and their products - think dark, dense, grainy bread - were staple foods in Greece. Dietary fibre from wholegrains and other plant foods, may help reduce blood cholesterol levels and may lower risk of heart disease, obesity, and type 2 diabetes.

547 6. Seasoning is key Incorporate onion, garlic, herbs and spices - not only will they make your food taste great, they amplify the antioxidants delivered to your plate. 7. No day without fresh fruit Fruit is the perfect snack. Unlike cakes and biscuits, it’s a whole food packed with fibre and phytonutrients, not just a source of sugar. Eaten at the end of a meal, as was done traditionally, fruit may help dampen the inflammation that can occur in the lining of your blood vessels following a meal.

8. Snack on nuts and seeds Nuts and seeds contain healthful mono and polyunsaturated fats, as well as minerals and multiple phytonutrients. They help lower high cholesterol levels and a Mediterranean diet including daily nuts has been shown to protect against having a first heart attack and stroke! 9. Dairy in moderation Dairy was enjoyed in moderation in the traditional Mediterranean diet - about two serves per day - and mainly eaten by adults as a fermented food, such as yoghurt and soft white cheese, e.g. ricotta, fetta. Milk was reserved for children and the elderly.

10. Drink water Humans are made up of 70 percent water, and our blood is 90 percent water. Water should be your main source of hydration. Aim for at least five cups per day (1.25 litres) of pure water in addition to other drinks. To function properly, all the cells and organs of the body need water. It is also used to lubricate the joints, protect the spinal cord and other sensitive tissues, regulate body temperature, and assist the passage of food through the intestines.

You will find a Mediterranean recipe card within your program compliments of Sue Radd.

548 549